1This is ld.info, produced by makeinfo version 4.8 from ld.texinfo. 2 3INFO-DIR-SECTION Software development 4START-INFO-DIR-ENTRY 5* Ld: (ld). The GNU linker. 6END-INFO-DIR-ENTRY 7 8 This file documents the GNU linker LD (GNU Binutils) version 2.26. 9 10 Copyright (C) 1991-2015 Free Software Foundation, Inc. 11 12 Permission is granted to copy, distribute and/or modify this document 13under the terms of the GNU Free Documentation License, Version 1.3 or 14any later version published by the Free Software Foundation; with no 15Invariant Sections, with no Front-Cover Texts, and with no Back-Cover 16Texts. A copy of the license is included in the section entitled "GNU 17Free Documentation License". 18 19 20File: ld.info, Node: Top, Next: Overview, Up: (dir) 21 22LD 23** 24 25This file documents the GNU linker ld (GNU Binutils) version 2.26. 26 27 This document is distributed under the terms of the GNU Free 28Documentation License version 1.3. A copy of the license is included 29in the section entitled "GNU Free Documentation License". 30 31* Menu: 32 33* Overview:: Overview 34* Invocation:: Invocation 35* Scripts:: Linker Scripts 36 37* Machine Dependent:: Machine Dependent Features 38 39* BFD:: BFD 40 41* Reporting Bugs:: Reporting Bugs 42* MRI:: MRI Compatible Script Files 43* GNU Free Documentation License:: GNU Free Documentation License 44* LD Index:: LD Index 45 46 47File: ld.info, Node: Overview, Next: Invocation, Prev: Top, Up: Top 48 491 Overview 50********** 51 52`ld' combines a number of object and archive files, relocates their 53data and ties up symbol references. Usually the last step in compiling 54a program is to run `ld'. 55 56 `ld' accepts Linker Command Language files written in a superset of 57AT&T's Link Editor Command Language syntax, to provide explicit and 58total control over the linking process. 59 60 This version of `ld' uses the general purpose BFD libraries to 61operate on object files. This allows `ld' to read, combine, and write 62object files in many different formats--for example, COFF or `a.out'. 63Different formats may be linked together to produce any available kind 64of object file. *Note BFD::, for more information. 65 66 Aside from its flexibility, the GNU linker is more helpful than other 67linkers in providing diagnostic information. Many linkers abandon 68execution immediately upon encountering an error; whenever possible, 69`ld' continues executing, allowing you to identify other errors (or, in 70some cases, to get an output file in spite of the error). 71 72 73File: ld.info, Node: Invocation, Next: Scripts, Prev: Overview, Up: Top 74 752 Invocation 76************ 77 78The GNU linker `ld' is meant to cover a broad range of situations, and 79to be as compatible as possible with other linkers. As a result, you 80have many choices to control its behavior. 81 82* Menu: 83 84* Options:: Command Line Options 85* Environment:: Environment Variables 86 87 88File: ld.info, Node: Options, Next: Environment, Up: Invocation 89 902.1 Command Line Options 91======================== 92 93 The linker supports a plethora of command-line options, but in actual 94practice few of them are used in any particular context. For instance, 95a frequent use of `ld' is to link standard Unix object files on a 96standard, supported Unix system. On such a system, to link a file 97`hello.o': 98 99 ld -o OUTPUT /lib/crt0.o hello.o -lc 100 101 This tells `ld' to produce a file called OUTPUT as the result of 102linking the file `/lib/crt0.o' with `hello.o' and the library `libc.a', 103which will come from the standard search directories. (See the 104discussion of the `-l' option below.) 105 106 Some of the command-line options to `ld' may be specified at any 107point in the command line. However, options which refer to files, such 108as `-l' or `-T', cause the file to be read at the point at which the 109option appears in the command line, relative to the object files and 110other file options. Repeating non-file options with a different 111argument will either have no further effect, or override prior 112occurrences (those further to the left on the command line) of that 113option. Options which may be meaningfully specified more than once are 114noted in the descriptions below. 115 116 Non-option arguments are object files or archives which are to be 117linked together. They may follow, precede, or be mixed in with 118command-line options, except that an object file argument may not be 119placed between an option and its argument. 120 121 Usually the linker is invoked with at least one object file, but you 122can specify other forms of binary input files using `-l', `-R', and the 123script command language. If _no_ binary input files at all are 124specified, the linker does not produce any output, and issues the 125message `No input files'. 126 127 If the linker cannot recognize the format of an object file, it will 128assume that it is a linker script. A script specified in this way 129augments the main linker script used for the link (either the default 130linker script or the one specified by using `-T'). This feature 131permits the linker to link against a file which appears to be an object 132or an archive, but actually merely defines some symbol values, or uses 133`INPUT' or `GROUP' to load other objects. Specifying a script in this 134way merely augments the main linker script, with the extra commands 135placed after the main script; use the `-T' option to replace the 136default linker script entirely, but note the effect of the `INSERT' 137command. *Note Scripts::. 138 139 For options whose names are a single letter, option arguments must 140either follow the option letter without intervening whitespace, or be 141given as separate arguments immediately following the option that 142requires them. 143 144 For options whose names are multiple letters, either one dash or two 145can precede the option name; for example, `-trace-symbol' and 146`--trace-symbol' are equivalent. Note--there is one exception to this 147rule. Multiple letter options that start with a lower case 'o' can 148only be preceded by two dashes. This is to reduce confusion with the 149`-o' option. So for example `-omagic' sets the output file name to 150`magic' whereas `--omagic' sets the NMAGIC flag on the output. 151 152 Arguments to multiple-letter options must either be separated from 153the option name by an equals sign, or be given as separate arguments 154immediately following the option that requires them. For example, 155`--trace-symbol foo' and `--trace-symbol=foo' are equivalent. Unique 156abbreviations of the names of multiple-letter options are accepted. 157 158 Note--if the linker is being invoked indirectly, via a compiler 159driver (e.g. `gcc') then all the linker command line options should be 160prefixed by `-Wl,' (or whatever is appropriate for the particular 161compiler driver) like this: 162 163 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group 164 165 This is important, because otherwise the compiler driver program may 166silently drop the linker options, resulting in a bad link. Confusion 167may also arise when passing options that require values through a 168driver, as the use of a space between option and argument acts as a 169separator, and causes the driver to pass only the option to the linker 170and the argument to the compiler. In this case, it is simplest to use 171the joined forms of both single- and multiple-letter options, such as: 172 173 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map 174 175 Here is a table of the generic command line switches accepted by the 176GNU linker: 177 178`@FILE' 179 Read command-line options from FILE. The options read are 180 inserted in place of the original @FILE option. If FILE does not 181 exist, or cannot be read, then the option will be treated 182 literally, and not removed. 183 184 Options in FILE are separated by whitespace. A whitespace 185 character may be included in an option by surrounding the entire 186 option in either single or double quotes. Any character 187 (including a backslash) may be included by prefixing the character 188 to be included with a backslash. The FILE may itself contain 189 additional @FILE options; any such options will be processed 190 recursively. 191 192`-a KEYWORD' 193 This option is supported for HP/UX compatibility. The KEYWORD 194 argument must be one of the strings `archive', `shared', or 195 `default'. `-aarchive' is functionally equivalent to `-Bstatic', 196 and the other two keywords are functionally equivalent to 197 `-Bdynamic'. This option may be used any number of times. 198 199`--audit AUDITLIB' 200 Adds AUDITLIB to the `DT_AUDIT' entry of the dynamic section. 201 AUDITLIB is not checked for existence, nor will it use the 202 DT_SONAME specified in the library. If specified multiple times 203 `DT_AUDIT' will contain a colon separated list of audit interfaces 204 to use. If the linker finds an object with an audit entry while 205 searching for shared libraries, it will add a corresponding 206 `DT_DEPAUDIT' entry in the output file. This option is only 207 meaningful on ELF platforms supporting the rtld-audit interface. 208 209`-A ARCHITECTURE' 210`--architecture=ARCHITECTURE' 211 In the current release of `ld', this option is useful only for the 212 Intel 960 family of architectures. In that `ld' configuration, the 213 ARCHITECTURE argument identifies the particular architecture in 214 the 960 family, enabling some safeguards and modifying the 215 archive-library search path. *Note `ld' and the Intel 960 family: 216 i960, for details. 217 218 Future releases of `ld' may support similar functionality for 219 other architecture families. 220 221`-b INPUT-FORMAT' 222`--format=INPUT-FORMAT' 223 `ld' may be configured to support more than one kind of object 224 file. If your `ld' is configured this way, you can use the `-b' 225 option to specify the binary format for input object files that 226 follow this option on the command line. Even when `ld' is 227 configured to support alternative object formats, you don't 228 usually need to specify this, as `ld' should be configured to 229 expect as a default input format the most usual format on each 230 machine. INPUT-FORMAT is a text string, the name of a particular 231 format supported by the BFD libraries. (You can list the 232 available binary formats with `objdump -i'.) *Note BFD::. 233 234 You may want to use this option if you are linking files with an 235 unusual binary format. You can also use `-b' to switch formats 236 explicitly (when linking object files of different formats), by 237 including `-b INPUT-FORMAT' before each group of object files in a 238 particular format. 239 240 The default format is taken from the environment variable 241 `GNUTARGET'. *Note Environment::. You can also define the input 242 format from a script, using the command `TARGET'; see *Note Format 243 Commands::. 244 245`-c MRI-COMMANDFILE' 246`--mri-script=MRI-COMMANDFILE' 247 For compatibility with linkers produced by MRI, `ld' accepts script 248 files written in an alternate, restricted command language, 249 described in *Note MRI Compatible Script Files: MRI. Introduce 250 MRI script files with the option `-c'; use the `-T' option to run 251 linker scripts written in the general-purpose `ld' scripting 252 language. If MRI-CMDFILE does not exist, `ld' looks for it in the 253 directories specified by any `-L' options. 254 255`-d' 256`-dc' 257`-dp' 258 These three options are equivalent; multiple forms are supported 259 for compatibility with other linkers. They assign space to common 260 symbols even if a relocatable output file is specified (with 261 `-r'). The script command `FORCE_COMMON_ALLOCATION' has the same 262 effect. *Note Miscellaneous Commands::. 263 264`--depaudit AUDITLIB' 265`-P AUDITLIB' 266 Adds AUDITLIB to the `DT_DEPAUDIT' entry of the dynamic section. 267 AUDITLIB is not checked for existence, nor will it use the 268 DT_SONAME specified in the library. If specified multiple times 269 `DT_DEPAUDIT' will contain a colon separated list of audit 270 interfaces to use. This option is only meaningful on ELF 271 platforms supporting the rtld-audit interface. The -P option is 272 provided for Solaris compatibility. 273 274`-e ENTRY' 275`--entry=ENTRY' 276 Use ENTRY as the explicit symbol for beginning execution of your 277 program, rather than the default entry point. If there is no 278 symbol named ENTRY, the linker will try to parse ENTRY as a number, 279 and use that as the entry address (the number will be interpreted 280 in base 10; you may use a leading `0x' for base 16, or a leading 281 `0' for base 8). *Note Entry Point::, for a discussion of defaults 282 and other ways of specifying the entry point. 283 284`--exclude-libs LIB,LIB,...' 285 Specifies a list of archive libraries from which symbols should 286 not be automatically exported. The library names may be delimited 287 by commas or colons. Specifying `--exclude-libs ALL' excludes 288 symbols in all archive libraries from automatic export. This 289 option is available only for the i386 PE targeted port of the 290 linker and for ELF targeted ports. For i386 PE, symbols 291 explicitly listed in a .def file are still exported, regardless of 292 this option. For ELF targeted ports, symbols affected by this 293 option will be treated as hidden. 294 295`--exclude-modules-for-implib MODULE,MODULE,...' 296 Specifies a list of object files or archive members, from which 297 symbols should not be automatically exported, but which should be 298 copied wholesale into the import library being generated during 299 the link. The module names may be delimited by commas or colons, 300 and must match exactly the filenames used by `ld' to open the 301 files; for archive members, this is simply the member name, but 302 for object files the name listed must include and match precisely 303 any path used to specify the input file on the linker's 304 command-line. This option is available only for the i386 PE 305 targeted port of the linker. Symbols explicitly listed in a .def 306 file are still exported, regardless of this option. 307 308`-E' 309`--export-dynamic' 310`--no-export-dynamic' 311 When creating a dynamically linked executable, using the `-E' 312 option or the `--export-dynamic' option causes the linker to add 313 all symbols to the dynamic symbol table. The dynamic symbol table 314 is the set of symbols which are visible from dynamic objects at 315 run time. 316 317 If you do not use either of these options (or use the 318 `--no-export-dynamic' option to restore the default behavior), the 319 dynamic symbol table will normally contain only those symbols 320 which are referenced by some dynamic object mentioned in the link. 321 322 If you use `dlopen' to load a dynamic object which needs to refer 323 back to the symbols defined by the program, rather than some other 324 dynamic object, then you will probably need to use this option when 325 linking the program itself. 326 327 You can also use the dynamic list to control what symbols should 328 be added to the dynamic symbol table if the output format supports 329 it. See the description of `--dynamic-list'. 330 331 Note that this option is specific to ELF targeted ports. PE 332 targets support a similar function to export all symbols from a 333 DLL or EXE; see the description of `--export-all-symbols' below. 334 335`-EB' 336 Link big-endian objects. This affects the default output format. 337 338`-EL' 339 Link little-endian objects. This affects the default output 340 format. 341 342`-f NAME' 343`--auxiliary=NAME' 344 When creating an ELF shared object, set the internal DT_AUXILIARY 345 field to the specified name. This tells the dynamic linker that 346 the symbol table of the shared object should be used as an 347 auxiliary filter on the symbol table of the shared object NAME. 348 349 If you later link a program against this filter object, then, when 350 you run the program, the dynamic linker will see the DT_AUXILIARY 351 field. If the dynamic linker resolves any symbols from the filter 352 object, it will first check whether there is a definition in the 353 shared object NAME. If there is one, it will be used instead of 354 the definition in the filter object. The shared object NAME need 355 not exist. Thus the shared object NAME may be used to provide an 356 alternative implementation of certain functions, perhaps for 357 debugging or for machine specific performance. 358 359 This option may be specified more than once. The DT_AUXILIARY 360 entries will be created in the order in which they appear on the 361 command line. 362 363`-F NAME' 364`--filter=NAME' 365 When creating an ELF shared object, set the internal DT_FILTER 366 field to the specified name. This tells the dynamic linker that 367 the symbol table of the shared object which is being created 368 should be used as a filter on the symbol table of the shared 369 object NAME. 370 371 If you later link a program against this filter object, then, when 372 you run the program, the dynamic linker will see the DT_FILTER 373 field. The dynamic linker will resolve symbols according to the 374 symbol table of the filter object as usual, but it will actually 375 link to the definitions found in the shared object NAME. Thus the 376 filter object can be used to select a subset of the symbols 377 provided by the object NAME. 378 379 Some older linkers used the `-F' option throughout a compilation 380 toolchain for specifying object-file format for both input and 381 output object files. The GNU linker uses other mechanisms for 382 this purpose: the `-b', `--format', `--oformat' options, the 383 `TARGET' command in linker scripts, and the `GNUTARGET' 384 environment variable. The GNU linker will ignore the `-F' option 385 when not creating an ELF shared object. 386 387`-fini=NAME' 388 When creating an ELF executable or shared object, call NAME when 389 the executable or shared object is unloaded, by setting DT_FINI to 390 the address of the function. By default, the linker uses `_fini' 391 as the function to call. 392 393`-g' 394 Ignored. Provided for compatibility with other tools. 395 396`-G VALUE' 397`--gpsize=VALUE' 398 Set the maximum size of objects to be optimized using the GP 399 register to SIZE. This is only meaningful for object file formats 400 such as MIPS ELF that support putting large and small objects into 401 different sections. This is ignored for other object file formats. 402 403`-h NAME' 404`-soname=NAME' 405 When creating an ELF shared object, set the internal DT_SONAME 406 field to the specified name. When an executable is linked with a 407 shared object which has a DT_SONAME field, then when the 408 executable is run the dynamic linker will attempt to load the 409 shared object specified by the DT_SONAME field rather than the 410 using the file name given to the linker. 411 412`-i' 413 Perform an incremental link (same as option `-r'). 414 415`-init=NAME' 416 When creating an ELF executable or shared object, call NAME when 417 the executable or shared object is loaded, by setting DT_INIT to 418 the address of the function. By default, the linker uses `_init' 419 as the function to call. 420 421`-l NAMESPEC' 422`--library=NAMESPEC' 423 Add the archive or object file specified by NAMESPEC to the list 424 of files to link. This option may be used any number of times. 425 If NAMESPEC is of the form `:FILENAME', `ld' will search the 426 library path for a file called FILENAME, otherwise it will search 427 the library path for a file called `libNAMESPEC.a'. 428 429 On systems which support shared libraries, `ld' may also search for 430 files other than `libNAMESPEC.a'. Specifically, on ELF and SunOS 431 systems, `ld' will search a directory for a library called 432 `libNAMESPEC.so' before searching for one called `libNAMESPEC.a'. 433 (By convention, a `.so' extension indicates a shared library.) 434 Note that this behavior does not apply to `:FILENAME', which 435 always specifies a file called FILENAME. 436 437 The linker will search an archive only once, at the location where 438 it is specified on the command line. If the archive defines a 439 symbol which was undefined in some object which appeared before 440 the archive on the command line, the linker will include the 441 appropriate file(s) from the archive. However, an undefined 442 symbol in an object appearing later on the command line will not 443 cause the linker to search the archive again. 444 445 See the `-(' option for a way to force the linker to search 446 archives multiple times. 447 448 You may list the same archive multiple times on the command line. 449 450 This type of archive searching is standard for Unix linkers. 451 However, if you are using `ld' on AIX, note that it is different 452 from the behaviour of the AIX linker. 453 454`-L SEARCHDIR' 455`--library-path=SEARCHDIR' 456 Add path SEARCHDIR to the list of paths that `ld' will search for 457 archive libraries and `ld' control scripts. You may use this 458 option any number of times. The directories are searched in the 459 order in which they are specified on the command line. 460 Directories specified on the command line are searched before the 461 default directories. All `-L' options apply to all `-l' options, 462 regardless of the order in which the options appear. `-L' options 463 do not affect how `ld' searches for a linker script unless `-T' 464 option is specified. 465 466 If SEARCHDIR begins with `=', then the `=' will be replaced by the 467 "sysroot prefix", controlled by the `--sysroot' option, or 468 specified when the linker is configured. 469 470 The default set of paths searched (without being specified with 471 `-L') depends on which emulation mode `ld' is using, and in some 472 cases also on how it was configured. *Note Environment::. 473 474 The paths can also be specified in a link script with the 475 `SEARCH_DIR' command. Directories specified this way are searched 476 at the point in which the linker script appears in the command 477 line. 478 479`-m EMULATION' 480 Emulate the EMULATION linker. You can list the available 481 emulations with the `--verbose' or `-V' options. 482 483 If the `-m' option is not used, the emulation is taken from the 484 `LDEMULATION' environment variable, if that is defined. 485 486 Otherwise, the default emulation depends upon how the linker was 487 configured. 488 489`-M' 490`--print-map' 491 Print a link map to the standard output. A link map provides 492 information about the link, including the following: 493 494 * Where object files are mapped into memory. 495 496 * How common symbols are allocated. 497 498 * All archive members included in the link, with a mention of 499 the symbol which caused the archive member to be brought in. 500 501 * The values assigned to symbols. 502 503 Note - symbols whose values are computed by an expression 504 which involves a reference to a previous value of the same 505 symbol may not have correct result displayed in the link map. 506 This is because the linker discards intermediate results and 507 only retains the final value of an expression. Under such 508 circumstances the linker will display the final value 509 enclosed by square brackets. Thus for example a linker 510 script containing: 511 512 foo = 1 513 foo = foo * 4 514 foo = foo + 8 515 516 will produce the following output in the link map if the `-M' 517 option is used: 518 519 0x00000001 foo = 0x1 520 [0x0000000c] foo = (foo * 0x4) 521 [0x0000000c] foo = (foo + 0x8) 522 523 See *Note Expressions:: for more information about 524 expressions in linker scripts. 525 526`-n' 527`--nmagic' 528 Turn off page alignment of sections, and disable linking against 529 shared libraries. If the output format supports Unix style magic 530 numbers, mark the output as `NMAGIC'. 531 532`-N' 533`--omagic' 534 Set the text and data sections to be readable and writable. Also, 535 do not page-align the data segment, and disable linking against 536 shared libraries. If the output format supports Unix style magic 537 numbers, mark the output as `OMAGIC'. Note: Although a writable 538 text section is allowed for PE-COFF targets, it does not conform 539 to the format specification published by Microsoft. 540 541`--no-omagic' 542 This option negates most of the effects of the `-N' option. It 543 sets the text section to be read-only, and forces the data segment 544 to be page-aligned. Note - this option does not enable linking 545 against shared libraries. Use `-Bdynamic' for this. 546 547`-o OUTPUT' 548`--output=OUTPUT' 549 Use OUTPUT as the name for the program produced by `ld'; if this 550 option is not specified, the name `a.out' is used by default. The 551 script command `OUTPUT' can also specify the output file name. 552 553`-O LEVEL' 554 If LEVEL is a numeric values greater than zero `ld' optimizes the 555 output. This might take significantly longer and therefore 556 probably should only be enabled for the final binary. At the 557 moment this option only affects ELF shared library generation. 558 Future releases of the linker may make more use of this option. 559 Also currently there is no difference in the linker's behaviour 560 for different non-zero values of this option. Again this may 561 change with future releases. 562 563`--push-state' 564 The `--push-state' allows to preserve the current state of the 565 flags which govern the input file handling so that they can all be 566 restored with one corresponding `--pop-state' option. 567 568 The option which are covered are: `-Bdynamic', `-Bstatic', `-dn', 569 `-dy', `-call_shared', `-non_shared', `-static', `-N', `-n', 570 `--whole-archive', `--no-whole-archive', `-r', `-Ur', 571 `--copy-dt-needed-entries', `--no-copy-dt-needed-entries', 572 `--as-needed', `--no-as-needed', and `-a'. 573 574 One target for this option are specifications for `pkg-config'. 575 When used with the `--libs' option all possibly needed libraries 576 are listed and then possibly linked with all the time. It is 577 better to return something as follows: 578 579 -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state 580 581 Undoes the effect of -push-state, restores the previous values of 582 the flags governing input file handling. 583 584`-q' 585`--emit-relocs' 586 Leave relocation sections and contents in fully linked executables. 587 Post link analysis and optimization tools may need this 588 information in order to perform correct modifications of 589 executables. This results in larger executables. 590 591 This option is currently only supported on ELF platforms. 592 593`--force-dynamic' 594 Force the output file to have dynamic sections. This option is 595 specific to VxWorks targets. 596 597`-r' 598`--relocatable' 599 Generate relocatable output--i.e., generate an output file that 600 can in turn serve as input to `ld'. This is often called "partial 601 linking". As a side effect, in environments that support standard 602 Unix magic numbers, this option also sets the output file's magic 603 number to `OMAGIC'. If this option is not specified, an absolute 604 file is produced. When linking C++ programs, this option _will 605 not_ resolve references to constructors; to do that, use `-Ur'. 606 607 When an input file does not have the same format as the output 608 file, partial linking is only supported if that input file does 609 not contain any relocations. Different output formats can have 610 further restrictions; for example some `a.out'-based formats do 611 not support partial linking with input files in other formats at 612 all. 613 614 This option does the same thing as `-i'. 615 616`-R FILENAME' 617`--just-symbols=FILENAME' 618 Read symbol names and their addresses from FILENAME, but do not 619 relocate it or include it in the output. This allows your output 620 file to refer symbolically to absolute locations of memory defined 621 in other programs. You may use this option more than once. 622 623 For compatibility with other ELF linkers, if the `-R' option is 624 followed by a directory name, rather than a file name, it is 625 treated as the `-rpath' option. 626 627`-s' 628`--strip-all' 629 Omit all symbol information from the output file. 630 631`-S' 632`--strip-debug' 633 Omit debugger symbol information (but not all symbols) from the 634 output file. 635 636`-t' 637`--trace' 638 Print the names of the input files as `ld' processes them. 639 640`-T SCRIPTFILE' 641`--script=SCRIPTFILE' 642 Use SCRIPTFILE as the linker script. This script replaces `ld''s 643 default linker script (rather than adding to it), so COMMANDFILE 644 must specify everything necessary to describe the output file. 645 *Note Scripts::. If SCRIPTFILE does not exist in the current 646 directory, `ld' looks for it in the directories specified by any 647 preceding `-L' options. Multiple `-T' options accumulate. 648 649`-dT SCRIPTFILE' 650`--default-script=SCRIPTFILE' 651 Use SCRIPTFILE as the default linker script. *Note Scripts::. 652 653 This option is similar to the `--script' option except that 654 processing of the script is delayed until after the rest of the 655 command line has been processed. This allows options placed after 656 the `--default-script' option on the command line to affect the 657 behaviour of the linker script, which can be important when the 658 linker command line cannot be directly controlled by the user. 659 (eg because the command line is being constructed by another tool, 660 such as `gcc'). 661 662`-u SYMBOL' 663`--undefined=SYMBOL' 664 Force SYMBOL to be entered in the output file as an undefined 665 symbol. Doing this may, for example, trigger linking of additional 666 modules from standard libraries. `-u' may be repeated with 667 different option arguments to enter additional undefined symbols. 668 This option is equivalent to the `EXTERN' linker script command. 669 670 If this option is being used to force additional modules to be 671 pulled into the link, and if it is an error for the symbol to 672 remain undefined, then the option `--require-defined' should be 673 used instead. 674 675`--require-defined=SYMBOL' 676 Require that SYMBOL is defined in the output file. This option is 677 the same as option `--undefined' except that if SYMBOL is not 678 defined in the output file then the linker will issue an error and 679 exit. The same effect can be achieved in a linker script by using 680 `EXTERN', `ASSERT' and `DEFINED' together. This option can be 681 used multiple times to require additional symbols. 682 683`-Ur' 684 For anything other than C++ programs, this option is equivalent to 685 `-r': it generates relocatable output--i.e., an output file that 686 can in turn serve as input to `ld'. When linking C++ programs, 687 `-Ur' _does_ resolve references to constructors, unlike `-r'. It 688 does not work to use `-Ur' on files that were themselves linked 689 with `-Ur'; once the constructor table has been built, it cannot 690 be added to. Use `-Ur' only for the last partial link, and `-r' 691 for the others. 692 693`--orphan-handling=MODE' 694 Control how orphan sections are handled. An orphan section is one 695 not specifically mentioned in a linker script. *Note Orphan 696 Sections::. 697 698 MODE can have any of the following values: 699 700 `place' 701 Orphan sections are placed into a suitable output section 702 following the strategy described in *Note Orphan Sections::. 703 The option `--unique' also effects how sections are placed. 704 705 `discard' 706 All orphan sections are discarded, by placing them in the 707 `/DISCARD/' section (*note Output Section Discarding::). 708 709 `warn' 710 The linker will place the orphan section as for `place' and 711 also issue a warning. 712 713 `error' 714 The linker will exit with an error if any orphan section is 715 found. 716 717 The default if `--orphan-handling' is not given is `place'. 718 719`--unique[=SECTION]' 720 Creates a separate output section for every input section matching 721 SECTION, or if the optional wildcard SECTION argument is missing, 722 for every orphan input section. An orphan section is one not 723 specifically mentioned in a linker script. You may use this option 724 multiple times on the command line; It prevents the normal 725 merging of input sections with the same name, overriding output 726 section assignments in a linker script. 727 728`-v' 729`--version' 730`-V' 731 Display the version number for `ld'. The `-V' option also lists 732 the supported emulations. 733 734`-x' 735`--discard-all' 736 Delete all local symbols. 737 738`-X' 739`--discard-locals' 740 Delete all temporary local symbols. (These symbols start with 741 system-specific local label prefixes, typically `.L' for ELF 742 systems or `L' for traditional a.out systems.) 743 744`-y SYMBOL' 745`--trace-symbol=SYMBOL' 746 Print the name of each linked file in which SYMBOL appears. This 747 option may be given any number of times. On many systems it is 748 necessary to prepend an underscore. 749 750 This option is useful when you have an undefined symbol in your 751 link but don't know where the reference is coming from. 752 753`-Y PATH' 754 Add PATH to the default library search path. This option exists 755 for Solaris compatibility. 756 757`-z KEYWORD' 758 The recognized keywords are: 759 `combreloc' 760 Combines multiple reloc sections and sorts them to make 761 dynamic symbol lookup caching possible. 762 763 `defs' 764 Disallows undefined symbols in object files. Undefined 765 symbols in shared libraries are still allowed. 766 767 `execstack' 768 Marks the object as requiring executable stack. 769 770 `global' 771 This option is only meaningful when building a shared object. 772 It makes the symbols defined by this shared object available 773 for symbol resolution of subsequently loaded libraries. 774 775 `initfirst' 776 This option is only meaningful when building a shared object. 777 It marks the object so that its runtime initialization will 778 occur before the runtime initialization of any other objects 779 brought into the process at the same time. Similarly the 780 runtime finalization of the object will occur after the 781 runtime finalization of any other objects. 782 783 `interpose' 784 Marks the object that its symbol table interposes before all 785 symbols but the primary executable. 786 787 `lazy' 788 When generating an executable or shared library, mark it to 789 tell the dynamic linker to defer function call resolution to 790 the point when the function is called (lazy binding), rather 791 than at load time. Lazy binding is the default. 792 793 `loadfltr' 794 Marks the object that its filters be processed immediately at 795 runtime. 796 797 `muldefs' 798 Allows multiple definitions. 799 800 `nocombreloc' 801 Disables multiple reloc sections combining. 802 803 `nocopyreloc' 804 Disable linker generated .dynbss variables used in place of 805 variables defined in shared libraries. May result in dynamic 806 text relocations. 807 808 `nodefaultlib' 809 Marks the object that the search for dependencies of this 810 object will ignore any default library search paths. 811 812 `nodelete' 813 Marks the object shouldn't be unloaded at runtime. 814 815 `nodlopen' 816 Marks the object not available to `dlopen'. 817 818 `nodump' 819 Marks the object can not be dumped by `dldump'. 820 821 `noexecstack' 822 Marks the object as not requiring executable stack. 823 824 `text' 825 Treat DT_TEXTREL in shared object as error. 826 827 `notext' 828 Don't treat DT_TEXTREL in shared object as error. 829 830 `textoff' 831 Don't treat DT_TEXTREL in shared object as error. 832 833 `norelro' 834 Don't create an ELF `PT_GNU_RELRO' segment header in the 835 object. 836 837 `now' 838 When generating an executable or shared library, mark it to 839 tell the dynamic linker to resolve all symbols when the 840 program is started, or when the shared library is linked to 841 using dlopen, instead of deferring function call resolution 842 to the point when the function is first called. 843 844 `origin' 845 Marks the object may contain $ORIGIN. 846 847 `relro' 848 Create an ELF `PT_GNU_RELRO' segment header in the object. 849 850 `max-page-size=VALUE' 851 Set the emulation maximum page size to VALUE. 852 853 `common-page-size=VALUE' 854 Set the emulation common page size to VALUE. 855 856 `stack-size=VALUE' 857 Specify a stack size for in an ELF `PT_GNU_STACK' segment. 858 Specifying zero will override any default non-zero sized 859 `PT_GNU_STACK' segment creation. 860 861 `bndplt' 862 Always generate BND prefix in PLT entries. Supported for 863 Linux/x86_64. 864 865 `noextern-protected-data' 866 Don't treat protected data symbol as external when building 867 shared library. This option overrides linker backend 868 default. It can be used to workaround incorrect relocations 869 against protected data symbols generated by compiler. 870 Updates on protected data symbols by another module aren't 871 visible to the resulting shared library. Supported for i386 872 and x86-64. 873 874 `call-nop=prefix-addr' 875 `call-nop=prefix-nop' 876 `call-nop=suffix-nop' 877 `call-nop=prefix-BYTE' 878 `call-nop=suffix-BYTE' 879 Specify the 1-byte `NOP' padding when transforming indirect 880 call to a locally defined function, foo, via its GOT slot. 881 `call-nop=prefix-addr' generates `0x67 call foo'. 882 `call-nop=prefix-nop' generates `0x90 call foo'. 883 `call-nop=suffix-nop' generates `call foo 0x90'. 884 `call-nop=prefix-BYTE' generates `BYTE call foo'. 885 `call-nop=suffix-BYTE' generates `call foo BYTE'. Supported 886 for i386 and x86_64. 887 888 889 Other keywords are ignored for Solaris compatibility. 890 891`-( ARCHIVES -)' 892`--start-group ARCHIVES --end-group' 893 The ARCHIVES should be a list of archive files. They may be 894 either explicit file names, or `-l' options. 895 896 The specified archives are searched repeatedly until no new 897 undefined references are created. Normally, an archive is 898 searched only once in the order that it is specified on the 899 command line. If a symbol in that archive is needed to resolve an 900 undefined symbol referred to by an object in an archive that 901 appears later on the command line, the linker would not be able to 902 resolve that reference. By grouping the archives, they all be 903 searched repeatedly until all possible references are resolved. 904 905 Using this option has a significant performance cost. It is best 906 to use it only when there are unavoidable circular references 907 between two or more archives. 908 909`--accept-unknown-input-arch' 910`--no-accept-unknown-input-arch' 911 Tells the linker to accept input files whose architecture cannot be 912 recognised. The assumption is that the user knows what they are 913 doing and deliberately wants to link in these unknown input files. 914 This was the default behaviour of the linker, before release 915 2.14. The default behaviour from release 2.14 onwards is to 916 reject such input files, and so the `--accept-unknown-input-arch' 917 option has been added to restore the old behaviour. 918 919`--as-needed' 920`--no-as-needed' 921 This option affects ELF DT_NEEDED tags for dynamic libraries 922 mentioned on the command line after the `--as-needed' option. 923 Normally the linker will add a DT_NEEDED tag for each dynamic 924 library mentioned on the command line, regardless of whether the 925 library is actually needed or not. `--as-needed' causes a 926 DT_NEEDED tag to only be emitted for a library that _at that point 927 in the link_ satisfies a non-weak undefined symbol reference from 928 a regular object file or, if the library is not found in the 929 DT_NEEDED lists of other needed libraries, a non-weak undefined 930 symbol reference from another needed dynamic library. Object 931 files or libraries appearing on the command line _after_ the 932 library in question do not affect whether the library is seen as 933 needed. This is similar to the rules for extraction of object 934 files from archives. `--no-as-needed' restores the default 935 behaviour. 936 937`--add-needed' 938`--no-add-needed' 939 These two options have been deprecated because of the similarity of 940 their names to the `--as-needed' and `--no-as-needed' options. 941 They have been replaced by `--copy-dt-needed-entries' and 942 `--no-copy-dt-needed-entries'. 943 944`-assert KEYWORD' 945 This option is ignored for SunOS compatibility. 946 947`-Bdynamic' 948`-dy' 949`-call_shared' 950 Link against dynamic libraries. This is only meaningful on 951 platforms for which shared libraries are supported. This option 952 is normally the default on such platforms. The different variants 953 of this option are for compatibility with various systems. You 954 may use this option multiple times on the command line: it affects 955 library searching for `-l' options which follow it. 956 957`-Bgroup' 958 Set the `DF_1_GROUP' flag in the `DT_FLAGS_1' entry in the dynamic 959 section. This causes the runtime linker to handle lookups in this 960 object and its dependencies to be performed only inside the group. 961 `--unresolved-symbols=report-all' is implied. This option is only 962 meaningful on ELF platforms which support shared libraries. 963 964`-Bstatic' 965`-dn' 966`-non_shared' 967`-static' 968 Do not link against shared libraries. This is only meaningful on 969 platforms for which shared libraries are supported. The different 970 variants of this option are for compatibility with various 971 systems. You may use this option multiple times on the command 972 line: it affects library searching for `-l' options which follow 973 it. This option also implies `--unresolved-symbols=report-all'. 974 This option can be used with `-shared'. Doing so means that a 975 shared library is being created but that all of the library's 976 external references must be resolved by pulling in entries from 977 static libraries. 978 979`-Bsymbolic' 980 When creating a shared library, bind references to global symbols 981 to the definition within the shared library, if any. Normally, it 982 is possible for a program linked against a shared library to 983 override the definition within the shared library. This option 984 can also be used with the `--export-dynamic' option, when creating 985 a position independent executable, to bind references to global 986 symbols to the definition within the executable. This option is 987 only meaningful on ELF platforms which support shared libraries 988 and position independent executables. 989 990`-Bsymbolic-functions' 991 When creating a shared library, bind references to global function 992 symbols to the definition within the shared library, if any. This 993 option can also be used with the `--export-dynamic' option, when 994 creating a position independent executable, to bind references to 995 global function symbols to the definition within the executable. 996 This option is only meaningful on ELF platforms which support 997 shared libraries and position independent executables. 998 999`--dynamic-list=DYNAMIC-LIST-FILE' 1000 Specify the name of a dynamic list file to the linker. This is 1001 typically used when creating shared libraries to specify a list of 1002 global symbols whose references shouldn't be bound to the 1003 definition within the shared library, or creating dynamically 1004 linked executables to specify a list of symbols which should be 1005 added to the symbol table in the executable. This option is only 1006 meaningful on ELF platforms which support shared libraries. 1007 1008 The format of the dynamic list is the same as the version node 1009 without scope and node name. See *Note VERSION:: for more 1010 information. 1011 1012`--dynamic-list-data' 1013 Include all global data symbols to the dynamic list. 1014 1015`--dynamic-list-cpp-new' 1016 Provide the builtin dynamic list for C++ operator new and delete. 1017 It is mainly useful for building shared libstdc++. 1018 1019`--dynamic-list-cpp-typeinfo' 1020 Provide the builtin dynamic list for C++ runtime type 1021 identification. 1022 1023`--check-sections' 1024`--no-check-sections' 1025 Asks the linker _not_ to check section addresses after they have 1026 been assigned to see if there are any overlaps. Normally the 1027 linker will perform this check, and if it finds any overlaps it 1028 will produce suitable error messages. The linker does know about, 1029 and does make allowances for sections in overlays. The default 1030 behaviour can be restored by using the command line switch 1031 `--check-sections'. Section overlap is not usually checked for 1032 relocatable links. You can force checking in that case by using 1033 the `--check-sections' option. 1034 1035`--copy-dt-needed-entries' 1036`--no-copy-dt-needed-entries' 1037 This option affects the treatment of dynamic libraries referred to 1038 by DT_NEEDED tags _inside_ ELF dynamic libraries mentioned on the 1039 command line. Normally the linker won't add a DT_NEEDED tag to the 1040 output binary for each library mentioned in a DT_NEEDED tag in an 1041 input dynamic library. With `--copy-dt-needed-entries' specified 1042 on the command line however any dynamic libraries that follow it 1043 will have their DT_NEEDED entries added. The default behaviour 1044 can be restored with `--no-copy-dt-needed-entries'. 1045 1046 This option also has an effect on the resolution of symbols in 1047 dynamic libraries. With `--copy-dt-needed-entries' dynamic 1048 libraries mentioned on the command line will be recursively 1049 searched, following their DT_NEEDED tags to other libraries, in 1050 order to resolve symbols required by the output binary. With the 1051 default setting however the searching of dynamic libraries that 1052 follow it will stop with the dynamic library itself. No DT_NEEDED 1053 links will be traversed to resolve symbols. 1054 1055`--cref' 1056 Output a cross reference table. If a linker map file is being 1057 generated, the cross reference table is printed to the map file. 1058 Otherwise, it is printed on the standard output. 1059 1060 The format of the table is intentionally simple, so that it may be 1061 easily processed by a script if necessary. The symbols are 1062 printed out, sorted by name. For each symbol, a list of file 1063 names is given. If the symbol is defined, the first file listed 1064 is the location of the definition. If the symbol is defined as a 1065 common value then any files where this happens appear next. 1066 Finally any files that reference the symbol are listed. 1067 1068`--no-define-common' 1069 This option inhibits the assignment of addresses to common symbols. 1070 The script command `INHIBIT_COMMON_ALLOCATION' has the same effect. 1071 *Note Miscellaneous Commands::. 1072 1073 The `--no-define-common' option allows decoupling the decision to 1074 assign addresses to Common symbols from the choice of the output 1075 file type; otherwise a non-Relocatable output type forces 1076 assigning addresses to Common symbols. Using `--no-define-common' 1077 allows Common symbols that are referenced from a shared library to 1078 be assigned addresses only in the main program. This eliminates 1079 the unused duplicate space in the shared library, and also 1080 prevents any possible confusion over resolving to the wrong 1081 duplicate when there are many dynamic modules with specialized 1082 search paths for runtime symbol resolution. 1083 1084`--defsym=SYMBOL=EXPRESSION' 1085 Create a global symbol in the output file, containing the absolute 1086 address given by EXPRESSION. You may use this option as many 1087 times as necessary to define multiple symbols in the command line. 1088 A limited form of arithmetic is supported for the EXPRESSION in 1089 this context: you may give a hexadecimal constant or the name of 1090 an existing symbol, or use `+' and `-' to add or subtract 1091 hexadecimal constants or symbols. If you need more elaborate 1092 expressions, consider using the linker command language from a 1093 script (*note Assignments::). _Note:_ there should be no white 1094 space between SYMBOL, the equals sign ("<=>"), and EXPRESSION. 1095 1096`--demangle[=STYLE]' 1097`--no-demangle' 1098 These options control whether to demangle symbol names in error 1099 messages and other output. When the linker is told to demangle, 1100 it tries to present symbol names in a readable fashion: it strips 1101 leading underscores if they are used by the object file format, 1102 and converts C++ mangled symbol names into user readable names. 1103 Different compilers have different mangling styles. The optional 1104 demangling style argument can be used to choose an appropriate 1105 demangling style for your compiler. The linker will demangle by 1106 default unless the environment variable `COLLECT_NO_DEMANGLE' is 1107 set. These options may be used to override the default. 1108 1109`-IFILE' 1110`--dynamic-linker=FILE' 1111 Set the name of the dynamic linker. This is only meaningful when 1112 generating dynamically linked ELF executables. The default dynamic 1113 linker is normally correct; don't use this unless you know what 1114 you are doing. 1115 1116`--no-dynamic-linker' 1117 When producing an executable file, omit the request for a dynamic 1118 linker to be used at load-time. This is only meaningful for ELF 1119 executables that contain dynamic relocations, and usually requires 1120 entry point code that is capable of processing these relocations. 1121 1122`--fatal-warnings' 1123`--no-fatal-warnings' 1124 Treat all warnings as errors. The default behaviour can be 1125 restored with the option `--no-fatal-warnings'. 1126 1127`--force-exe-suffix' 1128 Make sure that an output file has a .exe suffix. 1129 1130 If a successfully built fully linked output file does not have a 1131 `.exe' or `.dll' suffix, this option forces the linker to copy the 1132 output file to one of the same name with a `.exe' suffix. This 1133 option is useful when using unmodified Unix makefiles on a 1134 Microsoft Windows host, since some versions of Windows won't run 1135 an image unless it ends in a `.exe' suffix. 1136 1137`--gc-sections' 1138`--no-gc-sections' 1139 Enable garbage collection of unused input sections. It is ignored 1140 on targets that do not support this option. The default behaviour 1141 (of not performing this garbage collection) can be restored by 1142 specifying `--no-gc-sections' on the command line. Note that 1143 garbage collection for COFF and PE format targets is supported, 1144 but the implementation is currently considered to be experimental. 1145 1146 `--gc-sections' decides which input sections are used by examining 1147 symbols and relocations. The section containing the entry symbol 1148 and all sections containing symbols undefined on the command-line 1149 will be kept, as will sections containing symbols referenced by 1150 dynamic objects. Note that when building shared libraries, the 1151 linker must assume that any visible symbol is referenced. Once 1152 this initial set of sections has been determined, the linker 1153 recursively marks as used any section referenced by their 1154 relocations. See `--entry' and `--undefined'. 1155 1156 This option can be set when doing a partial link (enabled with 1157 option `-r'). In this case the root of symbols kept must be 1158 explicitly specified either by an `--entry' or `--undefined' 1159 option or by a `ENTRY' command in the linker script. 1160 1161`--print-gc-sections' 1162`--no-print-gc-sections' 1163 List all sections removed by garbage collection. The listing is 1164 printed on stderr. This option is only effective if garbage 1165 collection has been enabled via the `--gc-sections') option. The 1166 default behaviour (of not listing the sections that are removed) 1167 can be restored by specifying `--no-print-gc-sections' on the 1168 command line. 1169 1170`--print-output-format' 1171 Print the name of the default output format (perhaps influenced by 1172 other command-line options). This is the string that would appear 1173 in an `OUTPUT_FORMAT' linker script command (*note File 1174 Commands::). 1175 1176`--print-memory-usage' 1177 Print used size, total size and used size of memory regions 1178 created with the *Note MEMORY:: command. This is useful on 1179 embedded targets to have a quick view of amount of free memory. 1180 The format of the output has one headline and one line per region. 1181 It is both human readable and easily parsable by tools. Here is 1182 an example of an output: 1183 1184 Memory region Used Size Region Size %age Used 1185 ROM: 256 KB 1 MB 25.00% 1186 RAM: 32 B 2 GB 0.00% 1187 1188`--help' 1189 Print a summary of the command-line options on the standard output 1190 and exit. 1191 1192`--target-help' 1193 Print a summary of all target specific options on the standard 1194 output and exit. 1195 1196`-Map=MAPFILE' 1197 Print a link map to the file MAPFILE. See the description of the 1198 `-M' option, above. 1199 1200`--no-keep-memory' 1201 `ld' normally optimizes for speed over memory usage by caching the 1202 symbol tables of input files in memory. This option tells `ld' to 1203 instead optimize for memory usage, by rereading the symbol tables 1204 as necessary. This may be required if `ld' runs out of memory 1205 space while linking a large executable. 1206 1207`--no-undefined' 1208`-z defs' 1209 Report unresolved symbol references from regular object files. 1210 This is done even if the linker is creating a non-symbolic shared 1211 library. The switch `--[no-]allow-shlib-undefined' controls the 1212 behaviour for reporting unresolved references found in shared 1213 libraries being linked in. 1214 1215`--allow-multiple-definition' 1216`-z muldefs' 1217 Normally when a symbol is defined multiple times, the linker will 1218 report a fatal error. These options allow multiple definitions and 1219 the first definition will be used. 1220 1221`--allow-shlib-undefined' 1222`--no-allow-shlib-undefined' 1223 Allows or disallows undefined symbols in shared libraries. This 1224 switch is similar to `--no-undefined' except that it determines 1225 the behaviour when the undefined symbols are in a shared library 1226 rather than a regular object file. It does not affect how 1227 undefined symbols in regular object files are handled. 1228 1229 The default behaviour is to report errors for any undefined symbols 1230 referenced in shared libraries if the linker is being used to 1231 create an executable, but to allow them if the linker is being 1232 used to create a shared library. 1233 1234 The reasons for allowing undefined symbol references in shared 1235 libraries specified at link time are that: 1236 1237 * A shared library specified at link time may not be the same 1238 as the one that is available at load time, so the symbol 1239 might actually be resolvable at load time. 1240 1241 * There are some operating systems, eg BeOS and HPPA, where 1242 undefined symbols in shared libraries are normal. 1243 1244 The BeOS kernel for example patches shared libraries at load 1245 time to select whichever function is most appropriate for the 1246 current architecture. This is used, for example, to 1247 dynamically select an appropriate memset function. 1248 1249`--no-undefined-version' 1250 Normally when a symbol has an undefined version, the linker will 1251 ignore it. This option disallows symbols with undefined version 1252 and a fatal error will be issued instead. 1253 1254`--default-symver' 1255 Create and use a default symbol version (the soname) for 1256 unversioned exported symbols. 1257 1258`--default-imported-symver' 1259 Create and use a default symbol version (the soname) for 1260 unversioned imported symbols. 1261 1262`--no-warn-mismatch' 1263 Normally `ld' will give an error if you try to link together input 1264 files that are mismatched for some reason, perhaps because they 1265 have been compiled for different processors or for different 1266 endiannesses. This option tells `ld' that it should silently 1267 permit such possible errors. This option should only be used with 1268 care, in cases when you have taken some special action that 1269 ensures that the linker errors are inappropriate. 1270 1271`--no-warn-search-mismatch' 1272 Normally `ld' will give a warning if it finds an incompatible 1273 library during a library search. This option silences the warning. 1274 1275`--no-whole-archive' 1276 Turn off the effect of the `--whole-archive' option for subsequent 1277 archive files. 1278 1279`--noinhibit-exec' 1280 Retain the executable output file whenever it is still usable. 1281 Normally, the linker will not produce an output file if it 1282 encounters errors during the link process; it exits without 1283 writing an output file when it issues any error whatsoever. 1284 1285`-nostdlib' 1286 Only search library directories explicitly specified on the 1287 command line. Library directories specified in linker scripts 1288 (including linker scripts specified on the command line) are 1289 ignored. 1290 1291`--oformat=OUTPUT-FORMAT' 1292 `ld' may be configured to support more than one kind of object 1293 file. If your `ld' is configured this way, you can use the 1294 `--oformat' option to specify the binary format for the output 1295 object file. Even when `ld' is configured to support alternative 1296 object formats, you don't usually need to specify this, as `ld' 1297 should be configured to produce as a default output format the most 1298 usual format on each machine. OUTPUT-FORMAT is a text string, the 1299 name of a particular format supported by the BFD libraries. (You 1300 can list the available binary formats with `objdump -i'.) The 1301 script command `OUTPUT_FORMAT' can also specify the output format, 1302 but this option overrides it. *Note BFD::. 1303 1304`-pie' 1305`--pic-executable' 1306 Create a position independent executable. This is currently only 1307 supported on ELF platforms. Position independent executables are 1308 similar to shared libraries in that they are relocated by the 1309 dynamic linker to the virtual address the OS chooses for them 1310 (which can vary between invocations). Like normal dynamically 1311 linked executables they can be executed and symbols defined in the 1312 executable cannot be overridden by shared libraries. 1313 1314`-qmagic' 1315 This option is ignored for Linux compatibility. 1316 1317`-Qy' 1318 This option is ignored for SVR4 compatibility. 1319 1320`--relax' 1321`--no-relax' 1322 An option with machine dependent effects. This option is only 1323 supported on a few targets. *Note `ld' and the H8/300: H8/300. 1324 *Note `ld' and the Intel 960 family: i960. *Note `ld' and Xtensa 1325 Processors: Xtensa. *Note `ld' and the 68HC11 and 68HC12: 1326 M68HC11/68HC12. *Note `ld' and the Altera Nios II: Nios II. 1327 *Note `ld' and PowerPC 32-bit ELF Support: PowerPC ELF32. 1328 1329 On some platforms the `--relax' option performs target specific, 1330 global optimizations that become possible when the linker resolves 1331 addressing in the program, such as relaxing address modes, 1332 synthesizing new instructions, selecting shorter version of current 1333 instructions, and combining constant values. 1334 1335 On some platforms these link time global optimizations may make 1336 symbolic debugging of the resulting executable impossible. This 1337 is known to be the case for the Matsushita MN10200 and MN10300 1338 family of processors. 1339 1340 On platforms where this is not supported, `--relax' is accepted, 1341 but ignored. 1342 1343 On platforms where `--relax' is accepted the option `--no-relax' 1344 can be used to disable the feature. 1345 1346`--retain-symbols-file=FILENAME' 1347 Retain _only_ the symbols listed in the file FILENAME, discarding 1348 all others. FILENAME is simply a flat file, with one symbol name 1349 per line. This option is especially useful in environments (such 1350 as VxWorks) where a large global symbol table is accumulated 1351 gradually, to conserve run-time memory. 1352 1353 `--retain-symbols-file' does _not_ discard undefined symbols, or 1354 symbols needed for relocations. 1355 1356 You may only specify `--retain-symbols-file' once in the command 1357 line. It overrides `-s' and `-S'. 1358 1359`-rpath=DIR' 1360 Add a directory to the runtime library search path. This is used 1361 when linking an ELF executable with shared objects. All `-rpath' 1362 arguments are concatenated and passed to the runtime linker, which 1363 uses them to locate shared objects at runtime. The `-rpath' 1364 option is also used when locating shared objects which are needed 1365 by shared objects explicitly included in the link; see the 1366 description of the `-rpath-link' option. If `-rpath' is not used 1367 when linking an ELF executable, the contents of the environment 1368 variable `LD_RUN_PATH' will be used if it is defined. 1369 1370 The `-rpath' option may also be used on SunOS. By default, on 1371 SunOS, the linker will form a runtime search path out of all the 1372 `-L' options it is given. If a `-rpath' option is used, the 1373 runtime search path will be formed exclusively using the `-rpath' 1374 options, ignoring the `-L' options. This can be useful when using 1375 gcc, which adds many `-L' options which may be on NFS mounted file 1376 systems. 1377 1378 For compatibility with other ELF linkers, if the `-R' option is 1379 followed by a directory name, rather than a file name, it is 1380 treated as the `-rpath' option. 1381 1382`-rpath-link=DIR' 1383 When using ELF or SunOS, one shared library may require another. 1384 This happens when an `ld -shared' link includes a shared library 1385 as one of the input files. 1386 1387 When the linker encounters such a dependency when doing a 1388 non-shared, non-relocatable link, it will automatically try to 1389 locate the required shared library and include it in the link, if 1390 it is not included explicitly. In such a case, the `-rpath-link' 1391 option specifies the first set of directories to search. The 1392 `-rpath-link' option may specify a sequence of directory names 1393 either by specifying a list of names separated by colons, or by 1394 appearing multiple times. 1395 1396 This option should be used with caution as it overrides the search 1397 path that may have been hard compiled into a shared library. In 1398 such a case it is possible to use unintentionally a different 1399 search path than the runtime linker would do. 1400 1401 The linker uses the following search paths to locate required 1402 shared libraries: 1403 1. Any directories specified by `-rpath-link' options. 1404 1405 2. Any directories specified by `-rpath' options. The difference 1406 between `-rpath' and `-rpath-link' is that directories 1407 specified by `-rpath' options are included in the executable 1408 and used at runtime, whereas the `-rpath-link' option is only 1409 effective at link time. Searching `-rpath' in this way is 1410 only supported by native linkers and cross linkers which have 1411 been configured with the `--with-sysroot' option. 1412 1413 3. On an ELF system, for native linkers, if the `-rpath' and 1414 `-rpath-link' options were not used, search the contents of 1415 the environment variable `LD_RUN_PATH'. 1416 1417 4. On SunOS, if the `-rpath' option was not used, search any 1418 directories specified using `-L' options. 1419 1420 5. For a native linker, search the contents of the environment 1421 variable `LD_LIBRARY_PATH'. 1422 1423 6. For a native ELF linker, the directories in `DT_RUNPATH' or 1424 `DT_RPATH' of a shared library are searched for shared 1425 libraries needed by it. The `DT_RPATH' entries are ignored if 1426 `DT_RUNPATH' entries exist. 1427 1428 7. The default directories, normally `/lib' and `/usr/lib'. 1429 1430 8. For a native linker on an ELF system, if the file 1431 `/etc/ld.so.conf' exists, the list of directories found in 1432 that file. 1433 1434 If the required shared library is not found, the linker will issue 1435 a warning and continue with the link. 1436 1437`-shared' 1438`-Bshareable' 1439 Create a shared library. This is currently only supported on ELF, 1440 XCOFF and SunOS platforms. On SunOS, the linker will 1441 automatically create a shared library if the `-e' option is not 1442 used and there are undefined symbols in the link. 1443 1444`--sort-common' 1445`--sort-common=ascending' 1446`--sort-common=descending' 1447 This option tells `ld' to sort the common symbols by alignment in 1448 ascending or descending order when it places them in the 1449 appropriate output sections. The symbol alignments considered are 1450 sixteen-byte or larger, eight-byte, four-byte, two-byte, and 1451 one-byte. This is to prevent gaps between symbols due to alignment 1452 constraints. If no sorting order is specified, then descending 1453 order is assumed. 1454 1455`--sort-section=name' 1456 This option will apply `SORT_BY_NAME' to all wildcard section 1457 patterns in the linker script. 1458 1459`--sort-section=alignment' 1460 This option will apply `SORT_BY_ALIGNMENT' to all wildcard section 1461 patterns in the linker script. 1462 1463`--split-by-file[=SIZE]' 1464 Similar to `--split-by-reloc' but creates a new output section for 1465 each input file when SIZE is reached. SIZE defaults to a size of 1466 1 if not given. 1467 1468`--split-by-reloc[=COUNT]' 1469 Tries to creates extra sections in the output file so that no 1470 single output section in the file contains more than COUNT 1471 relocations. This is useful when generating huge relocatable 1472 files for downloading into certain real time kernels with the COFF 1473 object file format; since COFF cannot represent more than 65535 1474 relocations in a single section. Note that this will fail to work 1475 with object file formats which do not support arbitrary sections. 1476 The linker will not split up individual input sections for 1477 redistribution, so if a single input section contains more than 1478 COUNT relocations one output section will contain that many 1479 relocations. COUNT defaults to a value of 32768. 1480 1481`--stats' 1482 Compute and display statistics about the operation of the linker, 1483 such as execution time and memory usage. 1484 1485`--sysroot=DIRECTORY' 1486 Use DIRECTORY as the location of the sysroot, overriding the 1487 configure-time default. This option is only supported by linkers 1488 that were configured using `--with-sysroot'. 1489 1490`--traditional-format' 1491 For some targets, the output of `ld' is different in some ways from 1492 the output of some existing linker. This switch requests `ld' to 1493 use the traditional format instead. 1494 1495 For example, on SunOS, `ld' combines duplicate entries in the 1496 symbol string table. This can reduce the size of an output file 1497 with full debugging information by over 30 percent. 1498 Unfortunately, the SunOS `dbx' program can not read the resulting 1499 program (`gdb' has no trouble). The `--traditional-format' switch 1500 tells `ld' to not combine duplicate entries. 1501 1502`--section-start=SECTIONNAME=ORG' 1503 Locate a section in the output file at the absolute address given 1504 by ORG. You may use this option as many times as necessary to 1505 locate multiple sections in the command line. ORG must be a 1506 single hexadecimal integer; for compatibility with other linkers, 1507 you may omit the leading `0x' usually associated with hexadecimal 1508 values. _Note:_ there should be no white space between 1509 SECTIONNAME, the equals sign ("<=>"), and ORG. 1510 1511`-Tbss=ORG' 1512`-Tdata=ORG' 1513`-Ttext=ORG' 1514 Same as `--section-start', with `.bss', `.data' or `.text' as the 1515 SECTIONNAME. 1516 1517`-Ttext-segment=ORG' 1518 When creating an ELF executable, it will set the address of the 1519 first byte of the text segment. 1520 1521`-Trodata-segment=ORG' 1522 When creating an ELF executable or shared object for a target where 1523 the read-only data is in its own segment separate from the 1524 executable text, it will set the address of the first byte of the 1525 read-only data segment. 1526 1527`-Tldata-segment=ORG' 1528 When creating an ELF executable or shared object for x86-64 medium 1529 memory model, it will set the address of the first byte of the 1530 ldata segment. 1531 1532`--unresolved-symbols=METHOD' 1533 Determine how to handle unresolved symbols. There are four 1534 possible values for `method': 1535 1536 `ignore-all' 1537 Do not report any unresolved symbols. 1538 1539 `report-all' 1540 Report all unresolved symbols. This is the default. 1541 1542 `ignore-in-object-files' 1543 Report unresolved symbols that are contained in shared 1544 libraries, but ignore them if they come from regular object 1545 files. 1546 1547 `ignore-in-shared-libs' 1548 Report unresolved symbols that come from regular object 1549 files, but ignore them if they come from shared libraries. 1550 This can be useful when creating a dynamic binary and it is 1551 known that all the shared libraries that it should be 1552 referencing are included on the linker's command line. 1553 1554 The behaviour for shared libraries on their own can also be 1555 controlled by the `--[no-]allow-shlib-undefined' option. 1556 1557 Normally the linker will generate an error message for each 1558 reported unresolved symbol but the option 1559 `--warn-unresolved-symbols' can change this to a warning. 1560 1561`--dll-verbose' 1562`--verbose[=NUMBER]' 1563 Display the version number for `ld' and list the linker emulations 1564 supported. Display which input files can and cannot be opened. 1565 Display the linker script being used by the linker. If the 1566 optional NUMBER argument > 1, plugin symbol status will also be 1567 displayed. 1568 1569`--version-script=VERSION-SCRIPTFILE' 1570 Specify the name of a version script to the linker. This is 1571 typically used when creating shared libraries to specify 1572 additional information about the version hierarchy for the library 1573 being created. This option is only fully supported on ELF 1574 platforms which support shared libraries; see *Note VERSION::. It 1575 is partially supported on PE platforms, which can use version 1576 scripts to filter symbol visibility in auto-export mode: any 1577 symbols marked `local' in the version script will not be exported. 1578 *Note WIN32::. 1579 1580`--warn-common' 1581 Warn when a common symbol is combined with another common symbol 1582 or with a symbol definition. Unix linkers allow this somewhat 1583 sloppy practice, but linkers on some other operating systems do 1584 not. This option allows you to find potential problems from 1585 combining global symbols. Unfortunately, some C libraries use 1586 this practice, so you may get some warnings about symbols in the 1587 libraries as well as in your programs. 1588 1589 There are three kinds of global symbols, illustrated here by C 1590 examples: 1591 1592 `int i = 1;' 1593 A definition, which goes in the initialized data section of 1594 the output file. 1595 1596 `extern int i;' 1597 An undefined reference, which does not allocate space. There 1598 must be either a definition or a common symbol for the 1599 variable somewhere. 1600 1601 `int i;' 1602 A common symbol. If there are only (one or more) common 1603 symbols for a variable, it goes in the uninitialized data 1604 area of the output file. The linker merges multiple common 1605 symbols for the same variable into a single symbol. If they 1606 are of different sizes, it picks the largest size. The 1607 linker turns a common symbol into a declaration, if there is 1608 a definition of the same variable. 1609 1610 The `--warn-common' option can produce five kinds of warnings. 1611 Each warning consists of a pair of lines: the first describes the 1612 symbol just encountered, and the second describes the previous 1613 symbol encountered with the same name. One or both of the two 1614 symbols will be a common symbol. 1615 1616 1. Turning a common symbol into a reference, because there is 1617 already a definition for the symbol. 1618 FILE(SECTION): warning: common of `SYMBOL' 1619 overridden by definition 1620 FILE(SECTION): warning: defined here 1621 1622 2. Turning a common symbol into a reference, because a later 1623 definition for the symbol is encountered. This is the same 1624 as the previous case, except that the symbols are encountered 1625 in a different order. 1626 FILE(SECTION): warning: definition of `SYMBOL' 1627 overriding common 1628 FILE(SECTION): warning: common is here 1629 1630 3. Merging a common symbol with a previous same-sized common 1631 symbol. 1632 FILE(SECTION): warning: multiple common 1633 of `SYMBOL' 1634 FILE(SECTION): warning: previous common is here 1635 1636 4. Merging a common symbol with a previous larger common symbol. 1637 FILE(SECTION): warning: common of `SYMBOL' 1638 overridden by larger common 1639 FILE(SECTION): warning: larger common is here 1640 1641 5. Merging a common symbol with a previous smaller common 1642 symbol. This is the same as the previous case, except that 1643 the symbols are encountered in a different order. 1644 FILE(SECTION): warning: common of `SYMBOL' 1645 overriding smaller common 1646 FILE(SECTION): warning: smaller common is here 1647 1648`--warn-constructors' 1649 Warn if any global constructors are used. This is only useful for 1650 a few object file formats. For formats like COFF or ELF, the 1651 linker can not detect the use of global constructors. 1652 1653`--warn-multiple-gp' 1654 Warn if multiple global pointer values are required in the output 1655 file. This is only meaningful for certain processors, such as the 1656 Alpha. Specifically, some processors put large-valued constants 1657 in a special section. A special register (the global pointer) 1658 points into the middle of this section, so that constants can be 1659 loaded efficiently via a base-register relative addressing mode. 1660 Since the offset in base-register relative mode is fixed and 1661 relatively small (e.g., 16 bits), this limits the maximum size of 1662 the constant pool. Thus, in large programs, it is often necessary 1663 to use multiple global pointer values in order to be able to 1664 address all possible constants. This option causes a warning to 1665 be issued whenever this case occurs. 1666 1667`--warn-once' 1668 Only warn once for each undefined symbol, rather than once per 1669 module which refers to it. 1670 1671`--warn-section-align' 1672 Warn if the address of an output section is changed because of 1673 alignment. Typically, the alignment will be set by an input 1674 section. The address will only be changed if it not explicitly 1675 specified; that is, if the `SECTIONS' command does not specify a 1676 start address for the section (*note SECTIONS::). 1677 1678`--warn-shared-textrel' 1679 Warn if the linker adds a DT_TEXTREL to a shared object. 1680 1681`--warn-alternate-em' 1682 Warn if an object has alternate ELF machine code. 1683 1684`--warn-unresolved-symbols' 1685 If the linker is going to report an unresolved symbol (see the 1686 option `--unresolved-symbols') it will normally generate an error. 1687 This option makes it generate a warning instead. 1688 1689`--error-unresolved-symbols' 1690 This restores the linker's default behaviour of generating errors 1691 when it is reporting unresolved symbols. 1692 1693`--whole-archive' 1694 For each archive mentioned on the command line after the 1695 `--whole-archive' option, include every object file in the archive 1696 in the link, rather than searching the archive for the required 1697 object files. This is normally used to turn an archive file into 1698 a shared library, forcing every object to be included in the 1699 resulting shared library. This option may be used more than once. 1700 1701 Two notes when using this option from gcc: First, gcc doesn't know 1702 about this option, so you have to use `-Wl,-whole-archive'. 1703 Second, don't forget to use `-Wl,-no-whole-archive' after your 1704 list of archives, because gcc will add its own list of archives to 1705 your link and you may not want this flag to affect those as well. 1706 1707`--wrap=SYMBOL' 1708 Use a wrapper function for SYMBOL. Any undefined reference to 1709 SYMBOL will be resolved to `__wrap_SYMBOL'. Any undefined 1710 reference to `__real_SYMBOL' will be resolved to SYMBOL. 1711 1712 This can be used to provide a wrapper for a system function. The 1713 wrapper function should be called `__wrap_SYMBOL'. If it wishes 1714 to call the system function, it should call `__real_SYMBOL'. 1715 1716 Here is a trivial example: 1717 1718 void * 1719 __wrap_malloc (size_t c) 1720 { 1721 printf ("malloc called with %zu\n", c); 1722 return __real_malloc (c); 1723 } 1724 1725 If you link other code with this file using `--wrap malloc', then 1726 all calls to `malloc' will call the function `__wrap_malloc' 1727 instead. The call to `__real_malloc' in `__wrap_malloc' will call 1728 the real `malloc' function. 1729 1730 You may wish to provide a `__real_malloc' function as well, so that 1731 links without the `--wrap' option will succeed. If you do this, 1732 you should not put the definition of `__real_malloc' in the same 1733 file as `__wrap_malloc'; if you do, the assembler may resolve the 1734 call before the linker has a chance to wrap it to `malloc'. 1735 1736`--eh-frame-hdr' 1737 Request creation of `.eh_frame_hdr' section and ELF 1738 `PT_GNU_EH_FRAME' segment header. 1739 1740`--no-ld-generated-unwind-info' 1741 Request creation of `.eh_frame' unwind info for linker generated 1742 code sections like PLT. This option is on by default if linker 1743 generated unwind info is supported. 1744 1745`--enable-new-dtags' 1746`--disable-new-dtags' 1747 This linker can create the new dynamic tags in ELF. But the older 1748 ELF systems may not understand them. If you specify 1749 `--enable-new-dtags', the new dynamic tags will be created as 1750 needed and older dynamic tags will be omitted. If you specify 1751 `--disable-new-dtags', no new dynamic tags will be created. By 1752 default, the new dynamic tags are not created. Note that those 1753 options are only available for ELF systems. 1754 1755`--hash-size=NUMBER' 1756 Set the default size of the linker's hash tables to a prime number 1757 close to NUMBER. Increasing this value can reduce the length of 1758 time it takes the linker to perform its tasks, at the expense of 1759 increasing the linker's memory requirements. Similarly reducing 1760 this value can reduce the memory requirements at the expense of 1761 speed. 1762 1763`--hash-style=STYLE' 1764 Set the type of linker's hash table(s). STYLE can be either 1765 `sysv' for classic ELF `.hash' section, `gnu' for new style GNU 1766 `.gnu.hash' section or `both' for both the classic ELF `.hash' and 1767 new style GNU `.gnu.hash' hash tables. The default is `sysv'. 1768 1769`--compress-debug-sections=none' 1770`--compress-debug-sections=zlib' 1771`--compress-debug-sections=zlib-gnu' 1772`--compress-debug-sections=zlib-gabi' 1773 On ELF platforms , these options control how DWARF debug sections 1774 are compressed using zlib. `--compress-debug-sections=none' 1775 doesn't compress DWARF debug sections. 1776 `--compress-debug-sections=zlib-gnu' compresses DWARF debug 1777 sections and rename debug section names to begin with `.zdebug' 1778 instead of `.debug'. `--compress-debug-sections=zlib' and 1779 `--compress-debug-sections=zlib-gabi' compress DWARF debug 1780 sections with SHF_COMPRESSED from the ELF ABI. The default 1781 behaviour varies depending upon the target involved and the 1782 configure options used to build the toolchain. The default can be 1783 determined by examing the output from the linker's `--help' option. 1784 1785`--reduce-memory-overheads' 1786 This option reduces memory requirements at ld runtime, at the 1787 expense of linking speed. This was introduced to select the old 1788 O(n^2) algorithm for link map file generation, rather than the new 1789 O(n) algorithm which uses about 40% more memory for symbol storage. 1790 1791 Another effect of the switch is to set the default hash table size 1792 to 1021, which again saves memory at the cost of lengthening the 1793 linker's run time. This is not done however if the `--hash-size' 1794 switch has been used. 1795 1796 The `--reduce-memory-overheads' switch may be also be used to 1797 enable other tradeoffs in future versions of the linker. 1798 1799`--build-id' 1800`--build-id=STYLE' 1801 Request the creation of a `.note.gnu.build-id' ELF note section or 1802 a `.buildid' COFF section. The contents of the note are unique 1803 bits identifying this linked file. STYLE can be `uuid' to use 128 1804 random bits, `sha1' to use a 160-bit SHA1 hash on the normative 1805 parts of the output contents, `md5' to use a 128-bit MD5 hash on 1806 the normative parts of the output contents, or `0xHEXSTRING' to 1807 use a chosen bit string specified as an even number of hexadecimal 1808 digits (`-' and `:' characters between digit pairs are ignored). 1809 If STYLE is omitted, `sha1' is used. 1810 1811 The `md5' and `sha1' styles produces an identifier that is always 1812 the same in an identical output file, but will be unique among all 1813 nonidentical output files. It is not intended to be compared as a 1814 checksum for the file's contents. A linked file may be changed 1815 later by other tools, but the build ID bit string identifying the 1816 original linked file does not change. 1817 1818 Passing `none' for STYLE disables the setting from any 1819 `--build-id' options earlier on the command line. 1820 18212.1.1 Options Specific to i386 PE Targets 1822----------------------------------------- 1823 1824The i386 PE linker supports the `-shared' option, which causes the 1825output to be a dynamically linked library (DLL) instead of a normal 1826executable. You should name the output `*.dll' when you use this 1827option. In addition, the linker fully supports the standard `*.def' 1828files, which may be specified on the linker command line like an object 1829file (in fact, it should precede archives it exports symbols from, to 1830ensure that they get linked in, just like a normal object file). 1831 1832 In addition to the options common to all targets, the i386 PE linker 1833support additional command line options that are specific to the i386 1834PE target. Options that take values may be separated from their values 1835by either a space or an equals sign. 1836 1837`--add-stdcall-alias' 1838 If given, symbols with a stdcall suffix (@NN) will be exported 1839 as-is and also with the suffix stripped. [This option is specific 1840 to the i386 PE targeted port of the linker] 1841 1842`--base-file FILE' 1843 Use FILE as the name of a file in which to save the base addresses 1844 of all the relocations needed for generating DLLs with `dlltool'. 1845 [This is an i386 PE specific option] 1846 1847`--dll' 1848 Create a DLL instead of a regular executable. You may also use 1849 `-shared' or specify a `LIBRARY' in a given `.def' file. [This 1850 option is specific to the i386 PE targeted port of the linker] 1851 1852`--enable-long-section-names' 1853`--disable-long-section-names' 1854 The PE variants of the COFF object format add an extension that 1855 permits the use of section names longer than eight characters, the 1856 normal limit for COFF. By default, these names are only allowed 1857 in object files, as fully-linked executable images do not carry 1858 the COFF string table required to support the longer names. As a 1859 GNU extension, it is possible to allow their use in executable 1860 images as well, or to (probably pointlessly!) disallow it in 1861 object files, by using these two options. Executable images 1862 generated with these long section names are slightly non-standard, 1863 carrying as they do a string table, and may generate confusing 1864 output when examined with non-GNU PE-aware tools, such as file 1865 viewers and dumpers. However, GDB relies on the use of PE long 1866 section names to find Dwarf-2 debug information sections in an 1867 executable image at runtime, and so if neither option is specified 1868 on the command-line, `ld' will enable long section names, 1869 overriding the default and technically correct behaviour, when it 1870 finds the presence of debug information while linking an executable 1871 image and not stripping symbols. [This option is valid for all PE 1872 targeted ports of the linker] 1873 1874`--enable-stdcall-fixup' 1875`--disable-stdcall-fixup' 1876 If the link finds a symbol that it cannot resolve, it will attempt 1877 to do "fuzzy linking" by looking for another defined symbol that 1878 differs only in the format of the symbol name (cdecl vs stdcall) 1879 and will resolve that symbol by linking to the match. For 1880 example, the undefined symbol `_foo' might be linked to the 1881 function `_foo@12', or the undefined symbol `_bar@16' might be 1882 linked to the function `_bar'. When the linker does this, it 1883 prints a warning, since it normally should have failed to link, 1884 but sometimes import libraries generated from third-party dlls may 1885 need this feature to be usable. If you specify 1886 `--enable-stdcall-fixup', this feature is fully enabled and 1887 warnings are not printed. If you specify 1888 `--disable-stdcall-fixup', this feature is disabled and such 1889 mismatches are considered to be errors. [This option is specific 1890 to the i386 PE targeted port of the linker] 1891 1892`--leading-underscore' 1893`--no-leading-underscore' 1894 For most targets default symbol-prefix is an underscore and is 1895 defined in target's description. By this option it is possible to 1896 disable/enable the default underscore symbol-prefix. 1897 1898`--export-all-symbols' 1899 If given, all global symbols in the objects used to build a DLL 1900 will be exported by the DLL. Note that this is the default if 1901 there otherwise wouldn't be any exported symbols. When symbols are 1902 explicitly exported via DEF files or implicitly exported via 1903 function attributes, the default is to not export anything else 1904 unless this option is given. Note that the symbols `DllMain@12', 1905 `DllEntryPoint@0', `DllMainCRTStartup@12', and `impure_ptr' will 1906 not be automatically exported. Also, symbols imported from other 1907 DLLs will not be re-exported, nor will symbols specifying the 1908 DLL's internal layout such as those beginning with `_head_' or 1909 ending with `_iname'. In addition, no symbols from `libgcc', 1910 `libstd++', `libmingw32', or `crtX.o' will be exported. Symbols 1911 whose names begin with `__rtti_' or `__builtin_' will not be 1912 exported, to help with C++ DLLs. Finally, there is an extensive 1913 list of cygwin-private symbols that are not exported (obviously, 1914 this applies on when building DLLs for cygwin targets). These 1915 cygwin-excludes are: `_cygwin_dll_entry@12', 1916 `_cygwin_crt0_common@8', `_cygwin_noncygwin_dll_entry@12', 1917 `_fmode', `_impure_ptr', `cygwin_attach_dll', `cygwin_premain0', 1918 `cygwin_premain1', `cygwin_premain2', `cygwin_premain3', and 1919 `environ'. [This option is specific to the i386 PE targeted port 1920 of the linker] 1921 1922`--exclude-symbols SYMBOL,SYMBOL,...' 1923 Specifies a list of symbols which should not be automatically 1924 exported. The symbol names may be delimited by commas or colons. 1925 [This option is specific to the i386 PE targeted port of the 1926 linker] 1927 1928`--exclude-all-symbols' 1929 Specifies no symbols should be automatically exported. [This 1930 option is specific to the i386 PE targeted port of the linker] 1931 1932`--file-alignment' 1933 Specify the file alignment. Sections in the file will always 1934 begin at file offsets which are multiples of this number. This 1935 defaults to 512. [This option is specific to the i386 PE targeted 1936 port of the linker] 1937 1938`--heap RESERVE' 1939`--heap RESERVE,COMMIT' 1940 Specify the number of bytes of memory to reserve (and optionally 1941 commit) to be used as heap for this program. The default is 1MB 1942 reserved, 4K committed. [This option is specific to the i386 PE 1943 targeted port of the linker] 1944 1945`--image-base VALUE' 1946 Use VALUE as the base address of your program or dll. This is the 1947 lowest memory location that will be used when your program or dll 1948 is loaded. To reduce the need to relocate and improve performance 1949 of your dlls, each should have a unique base address and not 1950 overlap any other dlls. The default is 0x400000 for executables, 1951 and 0x10000000 for dlls. [This option is specific to the i386 PE 1952 targeted port of the linker] 1953 1954`--kill-at' 1955 If given, the stdcall suffixes (@NN) will be stripped from symbols 1956 before they are exported. [This option is specific to the i386 PE 1957 targeted port of the linker] 1958 1959`--large-address-aware' 1960 If given, the appropriate bit in the "Characteristics" field of 1961 the COFF header is set to indicate that this executable supports 1962 virtual addresses greater than 2 gigabytes. This should be used 1963 in conjunction with the /3GB or /USERVA=VALUE megabytes switch in 1964 the "[operating systems]" section of the BOOT.INI. Otherwise, 1965 this bit has no effect. [This option is specific to PE targeted 1966 ports of the linker] 1967 1968`--disable-large-address-aware' 1969 Reverts the effect of a previous `--large-address-aware' option. 1970 This is useful if `--large-address-aware' is always set by the 1971 compiler driver (e.g. Cygwin gcc) and the executable does not 1972 support virtual addresses greater than 2 gigabytes. [This option 1973 is specific to PE targeted ports of the linker] 1974 1975`--major-image-version VALUE' 1976 Sets the major number of the "image version". Defaults to 1. 1977 [This option is specific to the i386 PE targeted port of the 1978 linker] 1979 1980`--major-os-version VALUE' 1981 Sets the major number of the "os version". Defaults to 4. [This 1982 option is specific to the i386 PE targeted port of the linker] 1983 1984`--major-subsystem-version VALUE' 1985 Sets the major number of the "subsystem version". Defaults to 4. 1986 [This option is specific to the i386 PE targeted port of the 1987 linker] 1988 1989`--minor-image-version VALUE' 1990 Sets the minor number of the "image version". Defaults to 0. 1991 [This option is specific to the i386 PE targeted port of the 1992 linker] 1993 1994`--minor-os-version VALUE' 1995 Sets the minor number of the "os version". Defaults to 0. [This 1996 option is specific to the i386 PE targeted port of the linker] 1997 1998`--minor-subsystem-version VALUE' 1999 Sets the minor number of the "subsystem version". Defaults to 0. 2000 [This option is specific to the i386 PE targeted port of the 2001 linker] 2002 2003`--output-def FILE' 2004 The linker will create the file FILE which will contain a DEF file 2005 corresponding to the DLL the linker is generating. This DEF file 2006 (which should be called `*.def') may be used to create an import 2007 library with `dlltool' or may be used as a reference to 2008 automatically or implicitly exported symbols. [This option is 2009 specific to the i386 PE targeted port of the linker] 2010 2011`--out-implib FILE' 2012 The linker will create the file FILE which will contain an import 2013 lib corresponding to the DLL the linker is generating. This import 2014 lib (which should be called `*.dll.a' or `*.a' may be used to link 2015 clients against the generated DLL; this behaviour makes it 2016 possible to skip a separate `dlltool' import library creation step. 2017 [This option is specific to the i386 PE targeted port of the 2018 linker] 2019 2020`--enable-auto-image-base' 2021`--enable-auto-image-base=VALUE' 2022 Automatically choose the image base for DLLs, optionally starting 2023 with base VALUE, unless one is specified using the `--image-base' 2024 argument. By using a hash generated from the dllname to create 2025 unique image bases for each DLL, in-memory collisions and 2026 relocations which can delay program execution are avoided. [This 2027 option is specific to the i386 PE targeted port of the linker] 2028 2029`--disable-auto-image-base' 2030 Do not automatically generate a unique image base. If there is no 2031 user-specified image base (`--image-base') then use the platform 2032 default. [This option is specific to the i386 PE targeted port of 2033 the linker] 2034 2035`--dll-search-prefix STRING' 2036 When linking dynamically to a dll without an import library, 2037 search for `<string><basename>.dll' in preference to 2038 `lib<basename>.dll'. This behaviour allows easy distinction 2039 between DLLs built for the various "subplatforms": native, cygwin, 2040 uwin, pw, etc. For instance, cygwin DLLs typically use 2041 `--dll-search-prefix=cyg'. [This option is specific to the i386 2042 PE targeted port of the linker] 2043 2044`--enable-auto-import' 2045 Do sophisticated linking of `_symbol' to `__imp__symbol' for DATA 2046 imports from DLLs, and create the necessary thunking symbols when 2047 building the import libraries with those DATA exports. Note: Use 2048 of the 'auto-import' extension will cause the text section of the 2049 image file to be made writable. This does not conform to the 2050 PE-COFF format specification published by Microsoft. 2051 2052 Note - use of the 'auto-import' extension will also cause read only 2053 data which would normally be placed into the .rdata section to be 2054 placed into the .data section instead. This is in order to work 2055 around a problem with consts that is described here: 2056 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html 2057 2058 Using 'auto-import' generally will 'just work' - but sometimes you 2059 may see this message: 2060 2061 "variable '<var>' can't be auto-imported. Please read the 2062 documentation for ld's `--enable-auto-import' for details." 2063 2064 This message occurs when some (sub)expression accesses an address 2065 ultimately given by the sum of two constants (Win32 import tables 2066 only allow one). Instances where this may occur include accesses 2067 to member fields of struct variables imported from a DLL, as well 2068 as using a constant index into an array variable imported from a 2069 DLL. Any multiword variable (arrays, structs, long long, etc) may 2070 trigger this error condition. However, regardless of the exact 2071 data type of the offending exported variable, ld will always 2072 detect it, issue the warning, and exit. 2073 2074 There are several ways to address this difficulty, regardless of 2075 the data type of the exported variable: 2076 2077 One way is to use -enable-runtime-pseudo-reloc switch. This leaves 2078 the task of adjusting references in your client code for runtime 2079 environment, so this method works only when runtime environment 2080 supports this feature. 2081 2082 A second solution is to force one of the 'constants' to be a 2083 variable - that is, unknown and un-optimizable at compile time. 2084 For arrays, there are two possibilities: a) make the indexee (the 2085 array's address) a variable, or b) make the 'constant' index a 2086 variable. Thus: 2087 2088 extern type extern_array[]; 2089 extern_array[1] --> 2090 { volatile type *t=extern_array; t[1] } 2091 2092 or 2093 2094 extern type extern_array[]; 2095 extern_array[1] --> 2096 { volatile int t=1; extern_array[t] } 2097 2098 For structs (and most other multiword data types) the only option 2099 is to make the struct itself (or the long long, or the ...) 2100 variable: 2101 2102 extern struct s extern_struct; 2103 extern_struct.field --> 2104 { volatile struct s *t=&extern_struct; t->field } 2105 2106 or 2107 2108 extern long long extern_ll; 2109 extern_ll --> 2110 { volatile long long * local_ll=&extern_ll; *local_ll } 2111 2112 A third method of dealing with this difficulty is to abandon 2113 'auto-import' for the offending symbol and mark it with 2114 `__declspec(dllimport)'. However, in practice that requires using 2115 compile-time #defines to indicate whether you are building a DLL, 2116 building client code that will link to the DLL, or merely 2117 building/linking to a static library. In making the choice 2118 between the various methods of resolving the 'direct address with 2119 constant offset' problem, you should consider typical real-world 2120 usage: 2121 2122 Original: 2123 --foo.h 2124 extern int arr[]; 2125 --foo.c 2126 #include "foo.h" 2127 void main(int argc, char **argv){ 2128 printf("%d\n",arr[1]); 2129 } 2130 2131 Solution 1: 2132 --foo.h 2133 extern int arr[]; 2134 --foo.c 2135 #include "foo.h" 2136 void main(int argc, char **argv){ 2137 /* This workaround is for win32 and cygwin; do not "optimize" */ 2138 volatile int *parr = arr; 2139 printf("%d\n",parr[1]); 2140 } 2141 2142 Solution 2: 2143 --foo.h 2144 /* Note: auto-export is assumed (no __declspec(dllexport)) */ 2145 #if (defined(_WIN32) || defined(__CYGWIN__)) && \ 2146 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC)) 2147 #define FOO_IMPORT __declspec(dllimport) 2148 #else 2149 #define FOO_IMPORT 2150 #endif 2151 extern FOO_IMPORT int arr[]; 2152 --foo.c 2153 #include "foo.h" 2154 void main(int argc, char **argv){ 2155 printf("%d\n",arr[1]); 2156 } 2157 2158 A fourth way to avoid this problem is to re-code your library to 2159 use a functional interface rather than a data interface for the 2160 offending variables (e.g. set_foo() and get_foo() accessor 2161 functions). [This option is specific to the i386 PE targeted port 2162 of the linker] 2163 2164`--disable-auto-import' 2165 Do not attempt to do sophisticated linking of `_symbol' to 2166 `__imp__symbol' for DATA imports from DLLs. [This option is 2167 specific to the i386 PE targeted port of the linker] 2168 2169`--enable-runtime-pseudo-reloc' 2170 If your code contains expressions described in -enable-auto-import 2171 section, that is, DATA imports from DLL with non-zero offset, this 2172 switch will create a vector of 'runtime pseudo relocations' which 2173 can be used by runtime environment to adjust references to such 2174 data in your client code. [This option is specific to the i386 PE 2175 targeted port of the linker] 2176 2177`--disable-runtime-pseudo-reloc' 2178 Do not create pseudo relocations for non-zero offset DATA imports 2179 from DLLs. [This option is specific to the i386 PE targeted port 2180 of the linker] 2181 2182`--enable-extra-pe-debug' 2183 Show additional debug info related to auto-import symbol thunking. 2184 [This option is specific to the i386 PE targeted port of the 2185 linker] 2186 2187`--section-alignment' 2188 Sets the section alignment. Sections in memory will always begin 2189 at addresses which are a multiple of this number. Defaults to 2190 0x1000. [This option is specific to the i386 PE targeted port of 2191 the linker] 2192 2193`--stack RESERVE' 2194`--stack RESERVE,COMMIT' 2195 Specify the number of bytes of memory to reserve (and optionally 2196 commit) to be used as stack for this program. The default is 2MB 2197 reserved, 4K committed. [This option is specific to the i386 PE 2198 targeted port of the linker] 2199 2200`--subsystem WHICH' 2201`--subsystem WHICH:MAJOR' 2202`--subsystem WHICH:MAJOR.MINOR' 2203 Specifies the subsystem under which your program will execute. The 2204 legal values for WHICH are `native', `windows', `console', 2205 `posix', and `xbox'. You may optionally set the subsystem version 2206 also. Numeric values are also accepted for WHICH. [This option 2207 is specific to the i386 PE targeted port of the linker] 2208 2209 The following options set flags in the `DllCharacteristics' field 2210 of the PE file header: [These options are specific to PE targeted 2211 ports of the linker] 2212 2213`--high-entropy-va' 2214 Image is compatible with 64-bit address space layout randomization 2215 (ASLR). 2216 2217`--dynamicbase' 2218 The image base address may be relocated using address space layout 2219 randomization (ASLR). This feature was introduced with MS Windows 2220 Vista for i386 PE targets. 2221 2222`--forceinteg' 2223 Code integrity checks are enforced. 2224 2225`--nxcompat' 2226 The image is compatible with the Data Execution Prevention. This 2227 feature was introduced with MS Windows XP SP2 for i386 PE targets. 2228 2229`--no-isolation' 2230 Although the image understands isolation, do not isolate the image. 2231 2232`--no-seh' 2233 The image does not use SEH. No SE handler may be called from this 2234 image. 2235 2236`--no-bind' 2237 Do not bind this image. 2238 2239`--wdmdriver' 2240 The driver uses the MS Windows Driver Model. 2241 2242`--tsaware' 2243 The image is Terminal Server aware. 2244 2245`--insert-timestamp' 2246`--no-insert-timestamp' 2247 Insert a real timestamp into the image. This is the default 2248 behaviour as it matches legacy code and it means that the image 2249 will work with other, proprietary tools. The problem with this 2250 default is that it will result in slightly different images being 2251 produced each time the same sources are linked. The option 2252 `--no-insert-timestamp' can be used to insert a zero value for the 2253 timestamp, this ensuring that binaries produced from identical 2254 sources will compare identically. 2255 22562.1.2 Options specific to C6X uClinux targets 2257--------------------------------------------- 2258 2259The C6X uClinux target uses a binary format called DSBT to support 2260shared libraries. Each shared library in the system needs to have a 2261unique index; all executables use an index of 0. 2262 2263`--dsbt-size SIZE' 2264 This option sets the number of entries in the DSBT of the current 2265 executable or shared library to SIZE. The default is to create a 2266 table with 64 entries. 2267 2268`--dsbt-index INDEX' 2269 This option sets the DSBT index of the current executable or 2270 shared library to INDEX. The default is 0, which is appropriate 2271 for generating executables. If a shared library is generated with 2272 a DSBT index of 0, the `R_C6000_DSBT_INDEX' relocs are copied into 2273 the output file. 2274 2275 The `--no-merge-exidx-entries' switch disables the merging of 2276 adjacent exidx entries in frame unwind info. 2277 2278 22792.1.3 Options specific to Motorola 68HC11 and 68HC12 targets 2280------------------------------------------------------------ 2281 2282The 68HC11 and 68HC12 linkers support specific options to control the 2283memory bank switching mapping and trampoline code generation. 2284 2285`--no-trampoline' 2286 This option disables the generation of trampoline. By default a 2287 trampoline is generated for each far function which is called 2288 using a `jsr' instruction (this happens when a pointer to a far 2289 function is taken). 2290 2291`--bank-window NAME' 2292 This option indicates to the linker the name of the memory region 2293 in the `MEMORY' specification that describes the memory bank 2294 window. The definition of such region is then used by the linker 2295 to compute paging and addresses within the memory window. 2296 2297 22982.1.4 Options specific to Motorola 68K target 2299--------------------------------------------- 2300 2301The following options are supported to control handling of GOT 2302generation when linking for 68K targets. 2303 2304`--got=TYPE' 2305 This option tells the linker which GOT generation scheme to use. 2306 TYPE should be one of `single', `negative', `multigot' or 2307 `target'. For more information refer to the Info entry for `ld'. 2308 2309 23102.1.5 Options specific to MIPS targets 2311-------------------------------------- 2312 2313The following options are supported to control microMIPS instruction 2314generation when linking for MIPS targets. 2315 2316`--insn32' 2317`--no-insn32' 2318 These options control the choice of microMIPS instructions used in 2319 code generated by the linker, such as that in the PLT or lazy 2320 binding stubs, or in relaxation. If `--insn32' is used, then the 2321 linker only uses 32-bit instruction encodings. By default or if 2322 `--no-insn32' is used, all instruction encodings are used, 2323 including 16-bit ones where possible. 2324 2325 2326 2327File: ld.info, Node: Environment, Prev: Options, Up: Invocation 2328 23292.2 Environment Variables 2330========================= 2331 2332You can change the behaviour of `ld' with the environment variables 2333`GNUTARGET', `LDEMULATION' and `COLLECT_NO_DEMANGLE'. 2334 2335 `GNUTARGET' determines the input-file object format if you don't use 2336`-b' (or its synonym `--format'). Its value should be one of the BFD 2337names for an input format (*note BFD::). If there is no `GNUTARGET' in 2338the environment, `ld' uses the natural format of the target. If 2339`GNUTARGET' is set to `default' then BFD attempts to discover the input 2340format by examining binary input files; this method often succeeds, but 2341there are potential ambiguities, since there is no method of ensuring 2342that the magic number used to specify object-file formats is unique. 2343However, the configuration procedure for BFD on each system places the 2344conventional format for that system first in the search-list, so 2345ambiguities are resolved in favor of convention. 2346 2347 `LDEMULATION' determines the default emulation if you don't use the 2348`-m' option. The emulation can affect various aspects of linker 2349behaviour, particularly the default linker script. You can list the 2350available emulations with the `--verbose' or `-V' options. If the `-m' 2351option is not used, and the `LDEMULATION' environment variable is not 2352defined, the default emulation depends upon how the linker was 2353configured. 2354 2355 Normally, the linker will default to demangling symbols. However, if 2356`COLLECT_NO_DEMANGLE' is set in the environment, then it will default 2357to not demangling symbols. This environment variable is used in a 2358similar fashion by the `gcc' linker wrapper program. The default may 2359be overridden by the `--demangle' and `--no-demangle' options. 2360 2361 2362File: ld.info, Node: Scripts, Next: Machine Dependent, Prev: Invocation, Up: Top 2363 23643 Linker Scripts 2365**************** 2366 2367Every link is controlled by a "linker script". This script is written 2368in the linker command language. 2369 2370 The main purpose of the linker script is to describe how the 2371sections in the input files should be mapped into the output file, and 2372to control the memory layout of the output file. Most linker scripts 2373do nothing more than this. However, when necessary, the linker script 2374can also direct the linker to perform many other operations, using the 2375commands described below. 2376 2377 The linker always uses a linker script. If you do not supply one 2378yourself, the linker will use a default script that is compiled into the 2379linker executable. You can use the `--verbose' command line option to 2380display the default linker script. Certain command line options, such 2381as `-r' or `-N', will affect the default linker script. 2382 2383 You may supply your own linker script by using the `-T' command line 2384option. When you do this, your linker script will replace the default 2385linker script. 2386 2387 You may also use linker scripts implicitly by naming them as input 2388files to the linker, as though they were files to be linked. *Note 2389Implicit Linker Scripts::. 2390 2391* Menu: 2392 2393* Basic Script Concepts:: Basic Linker Script Concepts 2394* Script Format:: Linker Script Format 2395* Simple Example:: Simple Linker Script Example 2396* Simple Commands:: Simple Linker Script Commands 2397* Assignments:: Assigning Values to Symbols 2398* SECTIONS:: SECTIONS Command 2399* MEMORY:: MEMORY Command 2400* PHDRS:: PHDRS Command 2401* VERSION:: VERSION Command 2402* Expressions:: Expressions in Linker Scripts 2403* Implicit Linker Scripts:: Implicit Linker Scripts 2404 2405 2406File: ld.info, Node: Basic Script Concepts, Next: Script Format, Up: Scripts 2407 24083.1 Basic Linker Script Concepts 2409================================ 2410 2411We need to define some basic concepts and vocabulary in order to 2412describe the linker script language. 2413 2414 The linker combines input files into a single output file. The 2415output file and each input file are in a special data format known as an 2416"object file format". Each file is called an "object file". The 2417output file is often called an "executable", but for our purposes we 2418will also call it an object file. Each object file has, among other 2419things, a list of "sections". We sometimes refer to a section in an 2420input file as an "input section"; similarly, a section in the output 2421file is an "output section". 2422 2423 Each section in an object file has a name and a size. Most sections 2424also have an associated block of data, known as the "section contents". 2425A section may be marked as "loadable", which means that the contents 2426should be loaded into memory when the output file is run. A section 2427with no contents may be "allocatable", which means that an area in 2428memory should be set aside, but nothing in particular should be loaded 2429there (in some cases this memory must be zeroed out). A section which 2430is neither loadable nor allocatable typically contains some sort of 2431debugging information. 2432 2433 Every loadable or allocatable output section has two addresses. The 2434first is the "VMA", or virtual memory address. This is the address the 2435section will have when the output file is run. The second is the 2436"LMA", or load memory address. This is the address at which the 2437section will be loaded. In most cases the two addresses will be the 2438same. An example of when they might be different is when a data section 2439is loaded into ROM, and then copied into RAM when the program starts up 2440(this technique is often used to initialize global variables in a ROM 2441based system). In this case the ROM address would be the LMA, and the 2442RAM address would be the VMA. 2443 2444 You can see the sections in an object file by using the `objdump' 2445program with the `-h' option. 2446 2447 Every object file also has a list of "symbols", known as the "symbol 2448table". A symbol may be defined or undefined. Each symbol has a name, 2449and each defined symbol has an address, among other information. If 2450you compile a C or C++ program into an object file, you will get a 2451defined symbol for every defined function and global or static 2452variable. Every undefined function or global variable which is 2453referenced in the input file will become an undefined symbol. 2454 2455 You can see the symbols in an object file by using the `nm' program, 2456or by using the `objdump' program with the `-t' option. 2457 2458 2459File: ld.info, Node: Script Format, Next: Simple Example, Prev: Basic Script Concepts, Up: Scripts 2460 24613.2 Linker Script Format 2462======================== 2463 2464Linker scripts are text files. 2465 2466 You write a linker script as a series of commands. Each command is 2467either a keyword, possibly followed by arguments, or an assignment to a 2468symbol. You may separate commands using semicolons. Whitespace is 2469generally ignored. 2470 2471 Strings such as file or format names can normally be entered 2472directly. If the file name contains a character such as a comma which 2473would otherwise serve to separate file names, you may put the file name 2474in double quotes. There is no way to use a double quote character in a 2475file name. 2476 2477 You may include comments in linker scripts just as in C, delimited by 2478`/*' and `*/'. As in C, comments are syntactically equivalent to 2479whitespace. 2480 2481 2482File: ld.info, Node: Simple Example, Next: Simple Commands, Prev: Script Format, Up: Scripts 2483 24843.3 Simple Linker Script Example 2485================================ 2486 2487Many linker scripts are fairly simple. 2488 2489 The simplest possible linker script has just one command: 2490`SECTIONS'. You use the `SECTIONS' command to describe the memory 2491layout of the output file. 2492 2493 The `SECTIONS' command is a powerful command. Here we will describe 2494a simple use of it. Let's assume your program consists only of code, 2495initialized data, and uninitialized data. These will be in the 2496`.text', `.data', and `.bss' sections, respectively. Let's assume 2497further that these are the only sections which appear in your input 2498files. 2499 2500 For this example, let's say that the code should be loaded at address 25010x10000, and that the data should start at address 0x8000000. Here is a 2502linker script which will do that: 2503 SECTIONS 2504 { 2505 . = 0x10000; 2506 .text : { *(.text) } 2507 . = 0x8000000; 2508 .data : { *(.data) } 2509 .bss : { *(.bss) } 2510 } 2511 2512 You write the `SECTIONS' command as the keyword `SECTIONS', followed 2513by a series of symbol assignments and output section descriptions 2514enclosed in curly braces. 2515 2516 The first line inside the `SECTIONS' command of the above example 2517sets the value of the special symbol `.', which is the location 2518counter. If you do not specify the address of an output section in some 2519other way (other ways are described later), the address is set from the 2520current value of the location counter. The location counter is then 2521incremented by the size of the output section. At the start of the 2522`SECTIONS' command, the location counter has the value `0'. 2523 2524 The second line defines an output section, `.text'. The colon is 2525required syntax which may be ignored for now. Within the curly braces 2526after the output section name, you list the names of the input sections 2527which should be placed into this output section. The `*' is a wildcard 2528which matches any file name. The expression `*(.text)' means all 2529`.text' input sections in all input files. 2530 2531 Since the location counter is `0x10000' when the output section 2532`.text' is defined, the linker will set the address of the `.text' 2533section in the output file to be `0x10000'. 2534 2535 The remaining lines define the `.data' and `.bss' sections in the 2536output file. The linker will place the `.data' output section at 2537address `0x8000000'. After the linker places the `.data' output 2538section, the value of the location counter will be `0x8000000' plus the 2539size of the `.data' output section. The effect is that the linker will 2540place the `.bss' output section immediately after the `.data' output 2541section in memory. 2542 2543 The linker will ensure that each output section has the required 2544alignment, by increasing the location counter if necessary. In this 2545example, the specified addresses for the `.text' and `.data' sections 2546will probably satisfy any alignment constraints, but the linker may 2547have to create a small gap between the `.data' and `.bss' sections. 2548 2549 That's it! That's a simple and complete linker script. 2550 2551 2552File: ld.info, Node: Simple Commands, Next: Assignments, Prev: Simple Example, Up: Scripts 2553 25543.4 Simple Linker Script Commands 2555================================= 2556 2557In this section we describe the simple linker script commands. 2558 2559* Menu: 2560 2561* Entry Point:: Setting the entry point 2562* File Commands:: Commands dealing with files 2563 2564* Format Commands:: Commands dealing with object file formats 2565 2566* REGION_ALIAS:: Assign alias names to memory regions 2567* Miscellaneous Commands:: Other linker script commands 2568 2569 2570File: ld.info, Node: Entry Point, Next: File Commands, Up: Simple Commands 2571 25723.4.1 Setting the Entry Point 2573----------------------------- 2574 2575The first instruction to execute in a program is called the "entry 2576point". You can use the `ENTRY' linker script command to set the entry 2577point. The argument is a symbol name: 2578 ENTRY(SYMBOL) 2579 2580 There are several ways to set the entry point. The linker will set 2581the entry point by trying each of the following methods in order, and 2582stopping when one of them succeeds: 2583 * the `-e' ENTRY command-line option; 2584 2585 * the `ENTRY(SYMBOL)' command in a linker script; 2586 2587 * the value of a target specific symbol, if it is defined; For many 2588 targets this is `start', but PE and BeOS based systems for example 2589 check a list of possible entry symbols, matching the first one 2590 found. 2591 2592 * the address of the first byte of the `.text' section, if present; 2593 2594 * The address `0'. 2595 2596 2597File: ld.info, Node: File Commands, Next: Format Commands, Prev: Entry Point, Up: Simple Commands 2598 25993.4.2 Commands Dealing with Files 2600--------------------------------- 2601 2602Several linker script commands deal with files. 2603 2604`INCLUDE FILENAME' 2605 Include the linker script FILENAME at this point. The file will 2606 be searched for in the current directory, and in any directory 2607 specified with the `-L' option. You can nest calls to `INCLUDE' 2608 up to 10 levels deep. 2609 2610 You can place `INCLUDE' directives at the top level, in `MEMORY' or 2611 `SECTIONS' commands, or in output section descriptions. 2612 2613`INPUT(FILE, FILE, ...)' 2614`INPUT(FILE FILE ...)' 2615 The `INPUT' command directs the linker to include the named files 2616 in the link, as though they were named on the command line. 2617 2618 For example, if you always want to include `subr.o' any time you do 2619 a link, but you can't be bothered to put it on every link command 2620 line, then you can put `INPUT (subr.o)' in your linker script. 2621 2622 In fact, if you like, you can list all of your input files in the 2623 linker script, and then invoke the linker with nothing but a `-T' 2624 option. 2625 2626 In case a "sysroot prefix" is configured, and the filename starts 2627 with the `/' character, and the script being processed was located 2628 inside the "sysroot prefix", the filename will be looked for in 2629 the "sysroot prefix". Otherwise, the linker will try to open the 2630 file in the current directory. If it is not found, the linker 2631 will search through the archive library search path. The "sysroot 2632 prefix" can also be forced by specifying `=' as the first 2633 character in the filename path. See also the description of `-L' 2634 in *Note Command Line Options: Options. 2635 2636 If you use `INPUT (-lFILE)', `ld' will transform the name to 2637 `libFILE.a', as with the command line argument `-l'. 2638 2639 When you use the `INPUT' command in an implicit linker script, the 2640 files will be included in the link at the point at which the linker 2641 script file is included. This can affect archive searching. 2642 2643`GROUP(FILE, FILE, ...)' 2644`GROUP(FILE FILE ...)' 2645 The `GROUP' command is like `INPUT', except that the named files 2646 should all be archives, and they are searched repeatedly until no 2647 new undefined references are created. See the description of `-(' 2648 in *Note Command Line Options: Options. 2649 2650`AS_NEEDED(FILE, FILE, ...)' 2651`AS_NEEDED(FILE FILE ...)' 2652 This construct can appear only inside of the `INPUT' or `GROUP' 2653 commands, among other filenames. The files listed will be handled 2654 as if they appear directly in the `INPUT' or `GROUP' commands, 2655 with the exception of ELF shared libraries, that will be added only 2656 when they are actually needed. This construct essentially enables 2657 `--as-needed' option for all the files listed inside of it and 2658 restores previous `--as-needed' resp. `--no-as-needed' setting 2659 afterwards. 2660 2661`OUTPUT(FILENAME)' 2662 The `OUTPUT' command names the output file. Using 2663 `OUTPUT(FILENAME)' in the linker script is exactly like using `-o 2664 FILENAME' on the command line (*note Command Line Options: 2665 Options.). If both are used, the command line option takes 2666 precedence. 2667 2668 You can use the `OUTPUT' command to define a default name for the 2669 output file other than the usual default of `a.out'. 2670 2671`SEARCH_DIR(PATH)' 2672 The `SEARCH_DIR' command adds PATH to the list of paths where `ld' 2673 looks for archive libraries. Using `SEARCH_DIR(PATH)' is exactly 2674 like using `-L PATH' on the command line (*note Command Line 2675 Options: Options.). If both are used, then the linker will search 2676 both paths. Paths specified using the command line option are 2677 searched first. 2678 2679`STARTUP(FILENAME)' 2680 The `STARTUP' command is just like the `INPUT' command, except 2681 that FILENAME will become the first input file to be linked, as 2682 though it were specified first on the command line. This may be 2683 useful when using a system in which the entry point is always the 2684 start of the first file. 2685 2686 2687File: ld.info, Node: Format Commands, Next: REGION_ALIAS, Prev: File Commands, Up: Simple Commands 2688 26893.4.3 Commands Dealing with Object File Formats 2690----------------------------------------------- 2691 2692A couple of linker script commands deal with object file formats. 2693 2694`OUTPUT_FORMAT(BFDNAME)' 2695`OUTPUT_FORMAT(DEFAULT, BIG, LITTLE)' 2696 The `OUTPUT_FORMAT' command names the BFD format to use for the 2697 output file (*note BFD::). Using `OUTPUT_FORMAT(BFDNAME)' is 2698 exactly like using `--oformat BFDNAME' on the command line (*note 2699 Command Line Options: Options.). If both are used, the command 2700 line option takes precedence. 2701 2702 You can use `OUTPUT_FORMAT' with three arguments to use different 2703 formats based on the `-EB' and `-EL' command line options. This 2704 permits the linker script to set the output format based on the 2705 desired endianness. 2706 2707 If neither `-EB' nor `-EL' are used, then the output format will 2708 be the first argument, DEFAULT. If `-EB' is used, the output 2709 format will be the second argument, BIG. If `-EL' is used, the 2710 output format will be the third argument, LITTLE. 2711 2712 For example, the default linker script for the MIPS ELF target 2713 uses this command: 2714 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips) 2715 This says that the default format for the output file is 2716 `elf32-bigmips', but if the user uses the `-EL' command line 2717 option, the output file will be created in the `elf32-littlemips' 2718 format. 2719 2720`TARGET(BFDNAME)' 2721 The `TARGET' command names the BFD format to use when reading input 2722 files. It affects subsequent `INPUT' and `GROUP' commands. This 2723 command is like using `-b BFDNAME' on the command line (*note 2724 Command Line Options: Options.). If the `TARGET' command is used 2725 but `OUTPUT_FORMAT' is not, then the last `TARGET' command is also 2726 used to set the format for the output file. *Note BFD::. 2727 2728 2729File: ld.info, Node: REGION_ALIAS, Next: Miscellaneous Commands, Prev: Format Commands, Up: Simple Commands 2730 27313.4.4 Assign alias names to memory regions 2732------------------------------------------ 2733 2734Alias names can be added to existing memory regions created with the 2735*Note MEMORY:: command. Each name corresponds to at most one memory 2736region. 2737 2738 REGION_ALIAS(ALIAS, REGION) 2739 2740 The `REGION_ALIAS' function creates an alias name ALIAS for the 2741memory region REGION. This allows a flexible mapping of output sections 2742to memory regions. An example follows. 2743 2744 Suppose we have an application for embedded systems which come with 2745various memory storage devices. All have a general purpose, volatile 2746memory `RAM' that allows code execution or data storage. Some may have 2747a read-only, non-volatile memory `ROM' that allows code execution and 2748read-only data access. The last variant is a read-only, non-volatile 2749memory `ROM2' with read-only data access and no code execution 2750capability. We have four output sections: 2751 2752 * `.text' program code; 2753 2754 * `.rodata' read-only data; 2755 2756 * `.data' read-write initialized data; 2757 2758 * `.bss' read-write zero initialized data. 2759 2760 The goal is to provide a linker command file that contains a system 2761independent part defining the output sections and a system dependent 2762part mapping the output sections to the memory regions available on the 2763system. Our embedded systems come with three different memory setups 2764`A', `B' and `C': 2765Section Variant A Variant B Variant C 2766.text RAM ROM ROM 2767.rodata RAM ROM ROM2 2768.data RAM RAM/ROM RAM/ROM2 2769.bss RAM RAM RAM 2770 The notation `RAM/ROM' or `RAM/ROM2' means that this section is 2771loaded into region `ROM' or `ROM2' respectively. Please note that the 2772load address of the `.data' section starts in all three variants at the 2773end of the `.rodata' section. 2774 2775 The base linker script that deals with the output sections follows. 2776It includes the system dependent `linkcmds.memory' file that describes 2777the memory layout: 2778 INCLUDE linkcmds.memory 2779 2780 SECTIONS 2781 { 2782 .text : 2783 { 2784 *(.text) 2785 } > REGION_TEXT 2786 .rodata : 2787 { 2788 *(.rodata) 2789 rodata_end = .; 2790 } > REGION_RODATA 2791 .data : AT (rodata_end) 2792 { 2793 data_start = .; 2794 *(.data) 2795 } > REGION_DATA 2796 data_size = SIZEOF(.data); 2797 data_load_start = LOADADDR(.data); 2798 .bss : 2799 { 2800 *(.bss) 2801 } > REGION_BSS 2802 } 2803 2804 Now we need three different `linkcmds.memory' files to define memory 2805regions and alias names. The content of `linkcmds.memory' for the three 2806variants `A', `B' and `C': 2807`A' 2808 Here everything goes into the `RAM'. 2809 MEMORY 2810 { 2811 RAM : ORIGIN = 0, LENGTH = 4M 2812 } 2813 2814 REGION_ALIAS("REGION_TEXT", RAM); 2815 REGION_ALIAS("REGION_RODATA", RAM); 2816 REGION_ALIAS("REGION_DATA", RAM); 2817 REGION_ALIAS("REGION_BSS", RAM); 2818 2819`B' 2820 Program code and read-only data go into the `ROM'. Read-write 2821 data goes into the `RAM'. An image of the initialized data is 2822 loaded into the `ROM' and will be copied during system start into 2823 the `RAM'. 2824 MEMORY 2825 { 2826 ROM : ORIGIN = 0, LENGTH = 3M 2827 RAM : ORIGIN = 0x10000000, LENGTH = 1M 2828 } 2829 2830 REGION_ALIAS("REGION_TEXT", ROM); 2831 REGION_ALIAS("REGION_RODATA", ROM); 2832 REGION_ALIAS("REGION_DATA", RAM); 2833 REGION_ALIAS("REGION_BSS", RAM); 2834 2835`C' 2836 Program code goes into the `ROM'. Read-only data goes into the 2837 `ROM2'. Read-write data goes into the `RAM'. An image of the 2838 initialized data is loaded into the `ROM2' and will be copied 2839 during system start into the `RAM'. 2840 MEMORY 2841 { 2842 ROM : ORIGIN = 0, LENGTH = 2M 2843 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M 2844 RAM : ORIGIN = 0x20000000, LENGTH = 1M 2845 } 2846 2847 REGION_ALIAS("REGION_TEXT", ROM); 2848 REGION_ALIAS("REGION_RODATA", ROM2); 2849 REGION_ALIAS("REGION_DATA", RAM); 2850 REGION_ALIAS("REGION_BSS", RAM); 2851 2852 It is possible to write a common system initialization routine to 2853copy the `.data' section from `ROM' or `ROM2' into the `RAM' if 2854necessary: 2855 #include <string.h> 2856 2857 extern char data_start []; 2858 extern char data_size []; 2859 extern char data_load_start []; 2860 2861 void copy_data(void) 2862 { 2863 if (data_start != data_load_start) 2864 { 2865 memcpy(data_start, data_load_start, (size_t) data_size); 2866 } 2867 } 2868 2869 2870File: ld.info, Node: Miscellaneous Commands, Prev: REGION_ALIAS, Up: Simple Commands 2871 28723.4.5 Other Linker Script Commands 2873---------------------------------- 2874 2875There are a few other linker scripts commands. 2876 2877`ASSERT(EXP, MESSAGE)' 2878 Ensure that EXP is non-zero. If it is zero, then exit the linker 2879 with an error code, and print MESSAGE. 2880 2881 Note that assertions are checked before the final stages of linking 2882 take place. This means that expressions involving symbols PROVIDEd 2883 inside section definitions will fail if the user has not set values 2884 for those symbols. The only exception to this rule is PROVIDEd 2885 symbols that just reference dot. Thus an assertion like this: 2886 2887 .stack : 2888 { 2889 PROVIDE (__stack = .); 2890 PROVIDE (__stack_size = 0x100); 2891 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack"); 2892 } 2893 2894 will fail if `__stack_size' is not defined elsewhere. Symbols 2895 PROVIDEd outside of section definitions are evaluated earlier, so 2896 they can be used inside ASSERTions. Thus: 2897 2898 PROVIDE (__stack_size = 0x100); 2899 .stack : 2900 { 2901 PROVIDE (__stack = .); 2902 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack"); 2903 } 2904 2905 will work. 2906 2907`EXTERN(SYMBOL SYMBOL ...)' 2908 Force SYMBOL to be entered in the output file as an undefined 2909 symbol. Doing this may, for example, trigger linking of additional 2910 modules from standard libraries. You may list several SYMBOLs for 2911 each `EXTERN', and you may use `EXTERN' multiple times. This 2912 command has the same effect as the `-u' command-line option. 2913 2914`FORCE_COMMON_ALLOCATION' 2915 This command has the same effect as the `-d' command-line option: 2916 to make `ld' assign space to common symbols even if a relocatable 2917 output file is specified (`-r'). 2918 2919`INHIBIT_COMMON_ALLOCATION' 2920 This command has the same effect as the `--no-define-common' 2921 command-line option: to make `ld' omit the assignment of addresses 2922 to common symbols even for a non-relocatable output file. 2923 2924`INSERT [ AFTER | BEFORE ] OUTPUT_SECTION' 2925 This command is typically used in a script specified by `-T' to 2926 augment the default `SECTIONS' with, for example, overlays. It 2927 inserts all prior linker script statements after (or before) 2928 OUTPUT_SECTION, and also causes `-T' to not override the default 2929 linker script. The exact insertion point is as for orphan 2930 sections. *Note Location Counter::. The insertion happens after 2931 the linker has mapped input sections to output sections. Prior to 2932 the insertion, since `-T' scripts are parsed before the default 2933 linker script, statements in the `-T' script occur before the 2934 default linker script statements in the internal linker 2935 representation of the script. In particular, input section 2936 assignments will be made to `-T' output sections before those in 2937 the default script. Here is an example of how a `-T' script using 2938 `INSERT' might look: 2939 2940 SECTIONS 2941 { 2942 OVERLAY : 2943 { 2944 .ov1 { ov1*(.text) } 2945 .ov2 { ov2*(.text) } 2946 } 2947 } 2948 INSERT AFTER .text; 2949 2950`NOCROSSREFS(SECTION SECTION ...)' 2951 This command may be used to tell `ld' to issue an error about any 2952 references among certain output sections. 2953 2954 In certain types of programs, particularly on embedded systems when 2955 using overlays, when one section is loaded into memory, another 2956 section will not be. Any direct references between the two 2957 sections would be errors. For example, it would be an error if 2958 code in one section called a function defined in the other section. 2959 2960 The `NOCROSSREFS' command takes a list of output section names. If 2961 `ld' detects any cross references between the sections, it reports 2962 an error and returns a non-zero exit status. Note that the 2963 `NOCROSSREFS' command uses output section names, not input section 2964 names. 2965 2966`OUTPUT_ARCH(BFDARCH)' 2967 Specify a particular output machine architecture. The argument is 2968 one of the names used by the BFD library (*note BFD::). You can 2969 see the architecture of an object file by using the `objdump' 2970 program with the `-f' option. 2971 2972`LD_FEATURE(STRING)' 2973 This command may be used to modify `ld' behavior. If STRING is 2974 `"SANE_EXPR"' then absolute symbols and numbers in a script are 2975 simply treated as numbers everywhere. *Note Expression Section::. 2976 2977 2978File: ld.info, Node: Assignments, Next: SECTIONS, Prev: Simple Commands, Up: Scripts 2979 29803.5 Assigning Values to Symbols 2981=============================== 2982 2983You may assign a value to a symbol in a linker script. This will define 2984the symbol and place it into the symbol table with a global scope. 2985 2986* Menu: 2987 2988* Simple Assignments:: Simple Assignments 2989* HIDDEN:: HIDDEN 2990* PROVIDE:: PROVIDE 2991* PROVIDE_HIDDEN:: PROVIDE_HIDDEN 2992* Source Code Reference:: How to use a linker script defined symbol in source code 2993 2994 2995File: ld.info, Node: Simple Assignments, Next: HIDDEN, Up: Assignments 2996 29973.5.1 Simple Assignments 2998------------------------ 2999 3000You may assign to a symbol using any of the C assignment operators: 3001 3002`SYMBOL = EXPRESSION ;' 3003`SYMBOL += EXPRESSION ;' 3004`SYMBOL -= EXPRESSION ;' 3005`SYMBOL *= EXPRESSION ;' 3006`SYMBOL /= EXPRESSION ;' 3007`SYMBOL <<= EXPRESSION ;' 3008`SYMBOL >>= EXPRESSION ;' 3009`SYMBOL &= EXPRESSION ;' 3010`SYMBOL |= EXPRESSION ;' 3011 3012 The first case will define SYMBOL to the value of EXPRESSION. In 3013the other cases, SYMBOL must already be defined, and the value will be 3014adjusted accordingly. 3015 3016 The special symbol name `.' indicates the location counter. You may 3017only use this within a `SECTIONS' command. *Note Location Counter::. 3018 3019 The semicolon after EXPRESSION is required. 3020 3021 Expressions are defined below; see *Note Expressions::. 3022 3023 You may write symbol assignments as commands in their own right, or 3024as statements within a `SECTIONS' command, or as part of an output 3025section description in a `SECTIONS' command. 3026 3027 The section of the symbol will be set from the section of the 3028expression; for more information, see *Note Expression Section::. 3029 3030 Here is an example showing the three different places that symbol 3031assignments may be used: 3032 3033 floating_point = 0; 3034 SECTIONS 3035 { 3036 .text : 3037 { 3038 *(.text) 3039 _etext = .; 3040 } 3041 _bdata = (. + 3) & ~ 3; 3042 .data : { *(.data) } 3043 } 3044 In this example, the symbol `floating_point' will be defined as 3045zero. The symbol `_etext' will be defined as the address following the 3046last `.text' input section. The symbol `_bdata' will be defined as the 3047address following the `.text' output section aligned upward to a 4 byte 3048boundary. 3049 3050 3051File: ld.info, Node: HIDDEN, Next: PROVIDE, Prev: Simple Assignments, Up: Assignments 3052 30533.5.2 HIDDEN 3054------------ 3055 3056For ELF targeted ports, define a symbol that will be hidden and won't be 3057exported. The syntax is `HIDDEN(SYMBOL = EXPRESSION)'. 3058 3059 Here is the example from *Note Simple Assignments::, rewritten to use 3060`HIDDEN': 3061 3062 HIDDEN(floating_point = 0); 3063 SECTIONS 3064 { 3065 .text : 3066 { 3067 *(.text) 3068 HIDDEN(_etext = .); 3069 } 3070 HIDDEN(_bdata = (. + 3) & ~ 3); 3071 .data : { *(.data) } 3072 } 3073 In this case none of the three symbols will be visible outside this 3074module. 3075 3076 3077File: ld.info, Node: PROVIDE, Next: PROVIDE_HIDDEN, Prev: HIDDEN, Up: Assignments 3078 30793.5.3 PROVIDE 3080------------- 3081 3082In some cases, it is desirable for a linker script to define a symbol 3083only if it is referenced and is not defined by any object included in 3084the link. For example, traditional linkers defined the symbol `etext'. 3085However, ANSI C requires that the user be able to use `etext' as a 3086function name without encountering an error. The `PROVIDE' keyword may 3087be used to define a symbol, such as `etext', only if it is referenced 3088but not defined. The syntax is `PROVIDE(SYMBOL = EXPRESSION)'. 3089 3090 Here is an example of using `PROVIDE' to define `etext': 3091 SECTIONS 3092 { 3093 .text : 3094 { 3095 *(.text) 3096 _etext = .; 3097 PROVIDE(etext = .); 3098 } 3099 } 3100 3101 In this example, if the program defines `_etext' (with a leading 3102underscore), the linker will give a multiple definition error. If, on 3103the other hand, the program defines `etext' (with no leading 3104underscore), the linker will silently use the definition in the program. 3105If the program references `etext' but does not define it, the linker 3106will use the definition in the linker script. 3107 3108 3109File: ld.info, Node: PROVIDE_HIDDEN, Next: Source Code Reference, Prev: PROVIDE, Up: Assignments 3110 31113.5.4 PROVIDE_HIDDEN 3112-------------------- 3113 3114Similar to `PROVIDE'. For ELF targeted ports, the symbol will be 3115hidden and won't be exported. 3116 3117 3118File: ld.info, Node: Source Code Reference, Prev: PROVIDE_HIDDEN, Up: Assignments 3119 31203.5.5 Source Code Reference 3121--------------------------- 3122 3123Accessing a linker script defined variable from source code is not 3124intuitive. In particular a linker script symbol is not equivalent to a 3125variable declaration in a high level language, it is instead a symbol 3126that does not have a value. 3127 3128 Before going further, it is important to note that compilers often 3129transform names in the source code into different names when they are 3130stored in the symbol table. For example, Fortran compilers commonly 3131prepend or append an underscore, and C++ performs extensive `name 3132mangling'. Therefore there might be a discrepancy between the name of 3133a variable as it is used in source code and the name of the same 3134variable as it is defined in a linker script. For example in C a 3135linker script variable might be referred to as: 3136 3137 extern int foo; 3138 3139 But in the linker script it might be defined as: 3140 3141 _foo = 1000; 3142 3143 In the remaining examples however it is assumed that no name 3144transformation has taken place. 3145 3146 When a symbol is declared in a high level language such as C, two 3147things happen. The first is that the compiler reserves enough space in 3148the program's memory to hold the _value_ of the symbol. The second is 3149that the compiler creates an entry in the program's symbol table which 3150holds the symbol's _address_. ie the symbol table contains the address 3151of the block of memory holding the symbol's value. So for example the 3152following C declaration, at file scope: 3153 3154 int foo = 1000; 3155 3156 creates an entry called `foo' in the symbol table. This entry holds 3157the address of an `int' sized block of memory where the number 1000 is 3158initially stored. 3159 3160 When a program references a symbol the compiler generates code that 3161first accesses the symbol table to find the address of the symbol's 3162memory block and then code to read the value from that memory block. 3163So: 3164 3165 foo = 1; 3166 3167 looks up the symbol `foo' in the symbol table, gets the address 3168associated with this symbol and then writes the value 1 into that 3169address. Whereas: 3170 3171 int * a = & foo; 3172 3173 looks up the symbol `foo' in the symbol table, gets its address and 3174then copies this address into the block of memory associated with the 3175variable `a'. 3176 3177 Linker scripts symbol declarations, by contrast, create an entry in 3178the symbol table but do not assign any memory to them. Thus they are 3179an address without a value. So for example the linker script 3180definition: 3181 3182 foo = 1000; 3183 3184 creates an entry in the symbol table called `foo' which holds the 3185address of memory location 1000, but nothing special is stored at 3186address 1000. This means that you cannot access the _value_ of a 3187linker script defined symbol - it has no value - all you can do is 3188access the _address_ of a linker script defined symbol. 3189 3190 Hence when you are using a linker script defined symbol in source 3191code you should always take the address of the symbol, and never 3192attempt to use its value. For example suppose you want to copy the 3193contents of a section of memory called .ROM into a section called 3194.FLASH and the linker script contains these declarations: 3195 3196 start_of_ROM = .ROM; 3197 end_of_ROM = .ROM + sizeof (.ROM); 3198 start_of_FLASH = .FLASH; 3199 3200 Then the C source code to perform the copy would be: 3201 3202 extern char start_of_ROM, end_of_ROM, start_of_FLASH; 3203 3204 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM); 3205 3206 Note the use of the `&' operators. These are correct. 3207Alternatively the symbols can be treated as the names of vectors or 3208arrays and then the code will again work as expected: 3209 3210 extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[]; 3211 3212 memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM); 3213 3214 Note how using this method does not require the use of `&' operators. 3215 3216 3217File: ld.info, Node: SECTIONS, Next: MEMORY, Prev: Assignments, Up: Scripts 3218 32193.6 SECTIONS Command 3220==================== 3221 3222The `SECTIONS' command tells the linker how to map input sections into 3223output sections, and how to place the output sections in memory. 3224 3225 The format of the `SECTIONS' command is: 3226 SECTIONS 3227 { 3228 SECTIONS-COMMAND 3229 SECTIONS-COMMAND 3230 ... 3231 } 3232 3233 Each SECTIONS-COMMAND may of be one of the following: 3234 3235 * an `ENTRY' command (*note Entry command: Entry Point.) 3236 3237 * a symbol assignment (*note Assignments::) 3238 3239 * an output section description 3240 3241 * an overlay description 3242 3243 The `ENTRY' command and symbol assignments are permitted inside the 3244`SECTIONS' command for convenience in using the location counter in 3245those commands. This can also make the linker script easier to 3246understand because you can use those commands at meaningful points in 3247the layout of the output file. 3248 3249 Output section descriptions and overlay descriptions are described 3250below. 3251 3252 If you do not use a `SECTIONS' command in your linker script, the 3253linker will place each input section into an identically named output 3254section in the order that the sections are first encountered in the 3255input files. If all input sections are present in the first file, for 3256example, the order of sections in the output file will match the order 3257in the first input file. The first section will be at address zero. 3258 3259* Menu: 3260 3261* Output Section Description:: Output section description 3262* Output Section Name:: Output section name 3263* Output Section Address:: Output section address 3264* Input Section:: Input section description 3265* Output Section Data:: Output section data 3266* Output Section Keywords:: Output section keywords 3267* Output Section Discarding:: Output section discarding 3268* Output Section Attributes:: Output section attributes 3269* Overlay Description:: Overlay description 3270 3271 3272File: ld.info, Node: Output Section Description, Next: Output Section Name, Up: SECTIONS 3273 32743.6.1 Output Section Description 3275-------------------------------- 3276 3277The full description of an output section looks like this: 3278 SECTION [ADDRESS] [(TYPE)] : 3279 [AT(LMA)] 3280 [ALIGN(SECTION_ALIGN) | ALIGN_WITH_INPUT] 3281 [SUBALIGN(SUBSECTION_ALIGN)] 3282 [CONSTRAINT] 3283 { 3284 OUTPUT-SECTION-COMMAND 3285 OUTPUT-SECTION-COMMAND 3286 ... 3287 } [>REGION] [AT>LMA_REGION] [:PHDR :PHDR ...] [=FILLEXP] [,] 3288 3289 Most output sections do not use most of the optional section 3290attributes. 3291 3292 The whitespace around SECTION is required, so that the section name 3293is unambiguous. The colon and the curly braces are also required. The 3294comma at the end may be required if a FILLEXP is used and the next 3295SECTIONS-COMMAND looks like a continuation of the expression. The line 3296breaks and other white space are optional. 3297 3298 Each OUTPUT-SECTION-COMMAND may be one of the following: 3299 3300 * a symbol assignment (*note Assignments::) 3301 3302 * an input section description (*note Input Section::) 3303 3304 * data values to include directly (*note Output Section Data::) 3305 3306 * a special output section keyword (*note Output Section Keywords::) 3307 3308 3309File: ld.info, Node: Output Section Name, Next: Output Section Address, Prev: Output Section Description, Up: SECTIONS 3310 33113.6.2 Output Section Name 3312------------------------- 3313 3314The name of the output section is SECTION. SECTION must meet the 3315constraints of your output format. In formats which only support a 3316limited number of sections, such as `a.out', the name must be one of 3317the names supported by the format (`a.out', for example, allows only 3318`.text', `.data' or `.bss'). If the output format supports any number 3319of sections, but with numbers and not names (as is the case for Oasys), 3320the name should be supplied as a quoted numeric string. A section name 3321may consist of any sequence of characters, but a name which contains 3322any unusual characters such as commas must be quoted. 3323 3324 The output section name `/DISCARD/' is special; *Note Output Section 3325Discarding::. 3326 3327 3328File: ld.info, Node: Output Section Address, Next: Input Section, Prev: Output Section Name, Up: SECTIONS 3329 33303.6.3 Output Section Address 3331---------------------------- 3332 3333The ADDRESS is an expression for the VMA (the virtual memory address) 3334of the output section. This address is optional, but if it is provided 3335then the output address will be set exactly as specified. 3336 3337 If the output address is not specified then one will be chosen for 3338the section, based on the heuristic below. This address will be 3339adjusted to fit the alignment requirement of the output section. The 3340alignment requirement is the strictest alignment of any input section 3341contained within the output section. 3342 3343 The output section address heuristic is as follows: 3344 3345 * If an output memory REGION is set for the section then it is added 3346 to this region and its address will be the next free address in 3347 that region. 3348 3349 * If the MEMORY command has been used to create a list of memory 3350 regions then the first region which has attributes compatible with 3351 the section is selected to contain it. The section's output 3352 address will be the next free address in that region; *Note 3353 MEMORY::. 3354 3355 * If no memory regions were specified, or none match the section then 3356 the output address will be based on the current value of the 3357 location counter. 3358 3359For example: 3360 3361 .text . : { *(.text) } 3362 3363and 3364 3365 .text : { *(.text) } 3366 3367are subtly different. The first will set the address of the `.text' 3368output section to the current value of the location counter. The 3369second will set it to the current value of the location counter aligned 3370to the strictest alignment of any of the `.text' input sections. 3371 3372 The ADDRESS may be an arbitrary expression; *Note Expressions::. 3373For example, if you want to align the section on a 0x10 byte boundary, 3374so that the lowest four bits of the section address are zero, you could 3375do something like this: 3376 .text ALIGN(0x10) : { *(.text) } 3377 This works because `ALIGN' returns the current location counter 3378aligned upward to the specified value. 3379 3380 Specifying ADDRESS for a section will change the value of the 3381location counter, provided that the section is non-empty. (Empty 3382sections are ignored). 3383 3384 3385File: ld.info, Node: Input Section, Next: Output Section Data, Prev: Output Section Address, Up: SECTIONS 3386 33873.6.4 Input Section Description 3388------------------------------- 3389 3390The most common output section command is an input section description. 3391 3392 The input section description is the most basic linker script 3393operation. You use output sections to tell the linker how to lay out 3394your program in memory. You use input section descriptions to tell the 3395linker how to map the input files into your memory layout. 3396 3397* Menu: 3398 3399* Input Section Basics:: Input section basics 3400* Input Section Wildcards:: Input section wildcard patterns 3401* Input Section Common:: Input section for common symbols 3402* Input Section Keep:: Input section and garbage collection 3403* Input Section Example:: Input section example 3404 3405 3406File: ld.info, Node: Input Section Basics, Next: Input Section Wildcards, Up: Input Section 3407 34083.6.4.1 Input Section Basics 3409............................ 3410 3411An input section description consists of a file name optionally followed 3412by a list of section names in parentheses. 3413 3414 The file name and the section name may be wildcard patterns, which we 3415describe further below (*note Input Section Wildcards::). 3416 3417 The most common input section description is to include all input 3418sections with a particular name in the output section. For example, to 3419include all input `.text' sections, you would write: 3420 *(.text) 3421 Here the `*' is a wildcard which matches any file name. To exclude 3422a list of files from matching the file name wildcard, EXCLUDE_FILE may 3423be used to match all files except the ones specified in the 3424EXCLUDE_FILE list. For example: 3425 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors) 3426 will cause all .ctors sections from all files except `crtend.o' and 3427`otherfile.o' to be included. 3428 3429 There are two ways to include more than one section: 3430 *(.text .rdata) 3431 *(.text) *(.rdata) 3432 The difference between these is the order in which the `.text' and 3433`.rdata' input sections will appear in the output section. In the 3434first example, they will be intermingled, appearing in the same order as 3435they are found in the linker input. In the second example, all `.text' 3436input sections will appear first, followed by all `.rdata' input 3437sections. 3438 3439 You can specify a file name to include sections from a particular 3440file. You would do this if one or more of your files contain special 3441data that needs to be at a particular location in memory. For example: 3442 data.o(.data) 3443 3444 To refine the sections that are included based on the section flags 3445of an input section, INPUT_SECTION_FLAGS may be used. 3446 3447 Here is a simple example for using Section header flags for ELF 3448sections: 3449 3450 SECTIONS { 3451 .text : { INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) } 3452 .text2 : { INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) } 3453 } 3454 3455 In this example, the output section `.text' will be comprised of any 3456input section matching the name *(.text) whose section header flags 3457`SHF_MERGE' and `SHF_STRINGS' are set. The output section `.text2' 3458will be comprised of any input section matching the name *(.text) whose 3459section header flag `SHF_WRITE' is clear. 3460 3461 You can also specify files within archives by writing a pattern 3462matching the archive, a colon, then the pattern matching the file, with 3463no whitespace around the colon. 3464 3465`archive:file' 3466 matches file within archive 3467 3468`archive:' 3469 matches the whole archive 3470 3471`:file' 3472 matches file but not one in an archive 3473 3474 Either one or both of `archive' and `file' can contain shell 3475wildcards. On DOS based file systems, the linker will assume that a 3476single letter followed by a colon is a drive specifier, so `c:myfile.o' 3477is a simple file specification, not `myfile.o' within an archive called 3478`c'. `archive:file' filespecs may also be used within an 3479`EXCLUDE_FILE' list, but may not appear in other linker script 3480contexts. For instance, you cannot extract a file from an archive by 3481using `archive:file' in an `INPUT' command. 3482 3483 If you use a file name without a list of sections, then all sections 3484in the input file will be included in the output section. This is not 3485commonly done, but it may by useful on occasion. For example: 3486 data.o 3487 3488 When you use a file name which is not an `archive:file' specifier 3489and does not contain any wild card characters, the linker will first 3490see if you also specified the file name on the linker command line or 3491in an `INPUT' command. If you did not, the linker will attempt to open 3492the file as an input file, as though it appeared on the command line. 3493Note that this differs from an `INPUT' command, because the linker will 3494not search for the file in the archive search path. 3495 3496 3497File: ld.info, Node: Input Section Wildcards, Next: Input Section Common, Prev: Input Section Basics, Up: Input Section 3498 34993.6.4.2 Input Section Wildcard Patterns 3500....................................... 3501 3502In an input section description, either the file name or the section 3503name or both may be wildcard patterns. 3504 3505 The file name of `*' seen in many examples is a simple wildcard 3506pattern for the file name. 3507 3508 The wildcard patterns are like those used by the Unix shell. 3509 3510`*' 3511 matches any number of characters 3512 3513`?' 3514 matches any single character 3515 3516`[CHARS]' 3517 matches a single instance of any of the CHARS; the `-' character 3518 may be used to specify a range of characters, as in `[a-z]' to 3519 match any lower case letter 3520 3521`\' 3522 quotes the following character 3523 3524 When a file name is matched with a wildcard, the wildcard characters 3525will not match a `/' character (used to separate directory names on 3526Unix). A pattern consisting of a single `*' character is an exception; 3527it will always match any file name, whether it contains a `/' or not. 3528In a section name, the wildcard characters will match a `/' character. 3529 3530 File name wildcard patterns only match files which are explicitly 3531specified on the command line or in an `INPUT' command. The linker 3532does not search directories to expand wildcards. 3533 3534 If a file name matches more than one wildcard pattern, or if a file 3535name appears explicitly and is also matched by a wildcard pattern, the 3536linker will use the first match in the linker script. For example, this 3537sequence of input section descriptions is probably in error, because the 3538`data.o' rule will not be used: 3539 .data : { *(.data) } 3540 .data1 : { data.o(.data) } 3541 3542 Normally, the linker will place files and sections matched by 3543wildcards in the order in which they are seen during the link. You can 3544change this by using the `SORT_BY_NAME' keyword, which appears before a 3545wildcard pattern in parentheses (e.g., `SORT_BY_NAME(.text*)'). When 3546the `SORT_BY_NAME' keyword is used, the linker will sort the files or 3547sections into ascending order by name before placing them in the output 3548file. 3549 3550 `SORT_BY_ALIGNMENT' is very similar to `SORT_BY_NAME'. The 3551difference is `SORT_BY_ALIGNMENT' will sort sections into descending 3552order by alignment before placing them in the output file. Larger 3553alignments are placed before smaller alignments in order to reduce the 3554amount of padding necessary. 3555 3556 `SORT_BY_INIT_PRIORITY' is very similar to `SORT_BY_NAME'. The 3557difference is `SORT_BY_INIT_PRIORITY' will sort sections into ascending 3558order by numerical value of the GCC init_priority attribute encoded in 3559the section name before placing them in the output file. 3560 3561 `SORT' is an alias for `SORT_BY_NAME'. 3562 3563 When there are nested section sorting commands in linker script, 3564there can be at most 1 level of nesting for section sorting commands. 3565 3566 1. `SORT_BY_NAME' (`SORT_BY_ALIGNMENT' (wildcard section pattern)). 3567 It will sort the input sections by name first, then by alignment 3568 if two sections have the same name. 3569 3570 2. `SORT_BY_ALIGNMENT' (`SORT_BY_NAME' (wildcard section pattern)). 3571 It will sort the input sections by alignment first, then by name 3572 if two sections have the same alignment. 3573 3574 3. `SORT_BY_NAME' (`SORT_BY_NAME' (wildcard section pattern)) is 3575 treated the same as `SORT_BY_NAME' (wildcard section pattern). 3576 3577 4. `SORT_BY_ALIGNMENT' (`SORT_BY_ALIGNMENT' (wildcard section 3578 pattern)) is treated the same as `SORT_BY_ALIGNMENT' (wildcard 3579 section pattern). 3580 3581 5. All other nested section sorting commands are invalid. 3582 3583 When both command line section sorting option and linker script 3584section sorting command are used, section sorting command always takes 3585precedence over the command line option. 3586 3587 If the section sorting command in linker script isn't nested, the 3588command line option will make the section sorting command to be treated 3589as nested sorting command. 3590 3591 1. `SORT_BY_NAME' (wildcard section pattern ) with `--sort-sections 3592 alignment' is equivalent to `SORT_BY_NAME' (`SORT_BY_ALIGNMENT' 3593 (wildcard section pattern)). 3594 3595 2. `SORT_BY_ALIGNMENT' (wildcard section pattern) with 3596 `--sort-section name' is equivalent to `SORT_BY_ALIGNMENT' 3597 (`SORT_BY_NAME' (wildcard section pattern)). 3598 3599 If the section sorting command in linker script is nested, the 3600command line option will be ignored. 3601 3602 `SORT_NONE' disables section sorting by ignoring the command line 3603section sorting option. 3604 3605 If you ever get confused about where input sections are going, use 3606the `-M' linker option to generate a map file. The map file shows 3607precisely how input sections are mapped to output sections. 3608 3609 This example shows how wildcard patterns might be used to partition 3610files. This linker script directs the linker to place all `.text' 3611sections in `.text' and all `.bss' sections in `.bss'. The linker will 3612place the `.data' section from all files beginning with an upper case 3613character in `.DATA'; for all other files, the linker will place the 3614`.data' section in `.data'. 3615 SECTIONS { 3616 .text : { *(.text) } 3617 .DATA : { [A-Z]*(.data) } 3618 .data : { *(.data) } 3619 .bss : { *(.bss) } 3620 } 3621 3622 3623File: ld.info, Node: Input Section Common, Next: Input Section Keep, Prev: Input Section Wildcards, Up: Input Section 3624 36253.6.4.3 Input Section for Common Symbols 3626........................................ 3627 3628A special notation is needed for common symbols, because in many object 3629file formats common symbols do not have a particular input section. The 3630linker treats common symbols as though they are in an input section 3631named `COMMON'. 3632 3633 You may use file names with the `COMMON' section just as with any 3634other input sections. You can use this to place common symbols from a 3635particular input file in one section while common symbols from other 3636input files are placed in another section. 3637 3638 In most cases, common symbols in input files will be placed in the 3639`.bss' section in the output file. For example: 3640 .bss { *(.bss) *(COMMON) } 3641 3642 Some object file formats have more than one type of common symbol. 3643For example, the MIPS ELF object file format distinguishes standard 3644common symbols and small common symbols. In this case, the linker will 3645use a different special section name for other types of common symbols. 3646In the case of MIPS ELF, the linker uses `COMMON' for standard common 3647symbols and `.scommon' for small common symbols. This permits you to 3648map the different types of common symbols into memory at different 3649locations. 3650 3651 You will sometimes see `[COMMON]' in old linker scripts. This 3652notation is now considered obsolete. It is equivalent to `*(COMMON)'. 3653 3654 3655File: ld.info, Node: Input Section Keep, Next: Input Section Example, Prev: Input Section Common, Up: Input Section 3656 36573.6.4.4 Input Section and Garbage Collection 3658............................................ 3659 3660When link-time garbage collection is in use (`--gc-sections'), it is 3661often useful to mark sections that should not be eliminated. This is 3662accomplished by surrounding an input section's wildcard entry with 3663`KEEP()', as in `KEEP(*(.init))' or `KEEP(SORT_BY_NAME(*)(.ctors))'. 3664 3665 3666File: ld.info, Node: Input Section Example, Prev: Input Section Keep, Up: Input Section 3667 36683.6.4.5 Input Section Example 3669............................. 3670 3671The following example is a complete linker script. It tells the linker 3672to read all of the sections from file `all.o' and place them at the 3673start of output section `outputa' which starts at location `0x10000'. 3674All of section `.input1' from file `foo.o' follows immediately, in the 3675same output section. All of section `.input2' from `foo.o' goes into 3676output section `outputb', followed by section `.input1' from `foo1.o'. 3677All of the remaining `.input1' and `.input2' sections from any files 3678are written to output section `outputc'. 3679 3680 SECTIONS { 3681 outputa 0x10000 : 3682 { 3683 all.o 3684 foo.o (.input1) 3685 } 3686 outputb : 3687 { 3688 foo.o (.input2) 3689 foo1.o (.input1) 3690 } 3691 outputc : 3692 { 3693 *(.input1) 3694 *(.input2) 3695 } 3696 } 3697 3698 3699File: ld.info, Node: Output Section Data, Next: Output Section Keywords, Prev: Input Section, Up: SECTIONS 3700 37013.6.5 Output Section Data 3702------------------------- 3703 3704You can include explicit bytes of data in an output section by using 3705`BYTE', `SHORT', `LONG', `QUAD', or `SQUAD' as an output section 3706command. Each keyword is followed by an expression in parentheses 3707providing the value to store (*note Expressions::). The value of the 3708expression is stored at the current value of the location counter. 3709 3710 The `BYTE', `SHORT', `LONG', and `QUAD' commands store one, two, 3711four, and eight bytes (respectively). After storing the bytes, the 3712location counter is incremented by the number of bytes stored. 3713 3714 For example, this will store the byte 1 followed by the four byte 3715value of the symbol `addr': 3716 BYTE(1) 3717 LONG(addr) 3718 3719 When using a 64 bit host or target, `QUAD' and `SQUAD' are the same; 3720they both store an 8 byte, or 64 bit, value. When both host and target 3721are 32 bits, an expression is computed as 32 bits. In this case `QUAD' 3722stores a 32 bit value zero extended to 64 bits, and `SQUAD' stores a 32 3723bit value sign extended to 64 bits. 3724 3725 If the object file format of the output file has an explicit 3726endianness, which is the normal case, the value will be stored in that 3727endianness. When the object file format does not have an explicit 3728endianness, as is true of, for example, S-records, the value will be 3729stored in the endianness of the first input object file. 3730 3731 Note--these commands only work inside a section description and not 3732between them, so the following will produce an error from the linker: 3733 SECTIONS { .text : { *(.text) } LONG(1) .data : { *(.data) } } 3734 whereas this will work: 3735 SECTIONS { .text : { *(.text) ; LONG(1) } .data : { *(.data) } } 3736 3737 You may use the `FILL' command to set the fill pattern for the 3738current section. It is followed by an expression in parentheses. Any 3739otherwise unspecified regions of memory within the section (for example, 3740gaps left due to the required alignment of input sections) are filled 3741with the value of the expression, repeated as necessary. A `FILL' 3742statement covers memory locations after the point at which it occurs in 3743the section definition; by including more than one `FILL' statement, 3744you can have different fill patterns in different parts of an output 3745section. 3746 3747 This example shows how to fill unspecified regions of memory with the 3748value `0x90': 3749 FILL(0x90909090) 3750 3751 The `FILL' command is similar to the `=FILLEXP' output section 3752attribute, but it only affects the part of the section following the 3753`FILL' command, rather than the entire section. If both are used, the 3754`FILL' command takes precedence. *Note Output Section Fill::, for 3755details on the fill expression. 3756 3757 3758File: ld.info, Node: Output Section Keywords, Next: Output Section Discarding, Prev: Output Section Data, Up: SECTIONS 3759 37603.6.6 Output Section Keywords 3761----------------------------- 3762 3763There are a couple of keywords which can appear as output section 3764commands. 3765 3766`CREATE_OBJECT_SYMBOLS' 3767 The command tells the linker to create a symbol for each input 3768 file. The name of each symbol will be the name of the 3769 corresponding input file. The section of each symbol will be the 3770 output section in which the `CREATE_OBJECT_SYMBOLS' command 3771 appears. 3772 3773 This is conventional for the a.out object file format. It is not 3774 normally used for any other object file format. 3775 3776`CONSTRUCTORS' 3777 When linking using the a.out object file format, the linker uses an 3778 unusual set construct to support C++ global constructors and 3779 destructors. When linking object file formats which do not support 3780 arbitrary sections, such as ECOFF and XCOFF, the linker will 3781 automatically recognize C++ global constructors and destructors by 3782 name. For these object file formats, the `CONSTRUCTORS' command 3783 tells the linker to place constructor information in the output 3784 section where the `CONSTRUCTORS' command appears. The 3785 `CONSTRUCTORS' command is ignored for other object file formats. 3786 3787 The symbol `__CTOR_LIST__' marks the start of the global 3788 constructors, and the symbol `__CTOR_END__' marks the end. 3789 Similarly, `__DTOR_LIST__' and `__DTOR_END__' mark the start and 3790 end of the global destructors. The first word in the list is the 3791 number of entries, followed by the address of each constructor or 3792 destructor, followed by a zero word. The compiler must arrange to 3793 actually run the code. For these object file formats GNU C++ 3794 normally calls constructors from a subroutine `__main'; a call to 3795 `__main' is automatically inserted into the startup code for 3796 `main'. GNU C++ normally runs destructors either by using 3797 `atexit', or directly from the function `exit'. 3798 3799 For object file formats such as `COFF' or `ELF' which support 3800 arbitrary section names, GNU C++ will normally arrange to put the 3801 addresses of global constructors and destructors into the `.ctors' 3802 and `.dtors' sections. Placing the following sequence into your 3803 linker script will build the sort of table which the GNU C++ 3804 runtime code expects to see. 3805 3806 __CTOR_LIST__ = .; 3807 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2) 3808 *(.ctors) 3809 LONG(0) 3810 __CTOR_END__ = .; 3811 __DTOR_LIST__ = .; 3812 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2) 3813 *(.dtors) 3814 LONG(0) 3815 __DTOR_END__ = .; 3816 3817 If you are using the GNU C++ support for initialization priority, 3818 which provides some control over the order in which global 3819 constructors are run, you must sort the constructors at link time 3820 to ensure that they are executed in the correct order. When using 3821 the `CONSTRUCTORS' command, use `SORT_BY_NAME(CONSTRUCTORS)' 3822 instead. When using the `.ctors' and `.dtors' sections, use 3823 `*(SORT_BY_NAME(.ctors))' and `*(SORT_BY_NAME(.dtors))' instead of 3824 just `*(.ctors)' and `*(.dtors)'. 3825 3826 Normally the compiler and linker will handle these issues 3827 automatically, and you will not need to concern yourself with 3828 them. However, you may need to consider this if you are using C++ 3829 and writing your own linker scripts. 3830 3831 3832 3833File: ld.info, Node: Output Section Discarding, Next: Output Section Attributes, Prev: Output Section Keywords, Up: SECTIONS 3834 38353.6.7 Output Section Discarding 3836------------------------------- 3837 3838The linker will not normally create output sections with no contents. 3839This is for convenience when referring to input sections that may or 3840may not be present in any of the input files. For example: 3841 .foo : { *(.foo) } 3842 will only create a `.foo' section in the output file if there is a 3843`.foo' section in at least one input file, and if the input sections 3844are not all empty. Other link script directives that allocate space in 3845an output section will also create the output section. So too will 3846assignments to dot even if the assignment does not create space, except 3847for `. = 0', `. = . + 0', `. = sym', `. = . + sym' and `. = ALIGN (. != 38480, expr, 1)' when `sym' is an absolute symbol of value 0 defined in the 3849script. This allows you to force output of an empty section with `. = 3850.'. 3851 3852 The linker will ignore address assignments (*note Output Section 3853Address::) on discarded output sections, except when the linker script 3854defines symbols in the output section. In that case the linker will 3855obey the address assignments, possibly advancing dot even though the 3856section is discarded. 3857 3858 The special output section name `/DISCARD/' may be used to discard 3859input sections. Any input sections which are assigned to an output 3860section named `/DISCARD/' are not included in the output file. 3861 3862 3863File: ld.info, Node: Output Section Attributes, Next: Overlay Description, Prev: Output Section Discarding, Up: SECTIONS 3864 38653.6.8 Output Section Attributes 3866------------------------------- 3867 3868We showed above that the full description of an output section looked 3869like this: 3870 3871 SECTION [ADDRESS] [(TYPE)] : 3872 [AT(LMA)] 3873 [ALIGN(SECTION_ALIGN)] 3874 [SUBALIGN(SUBSECTION_ALIGN)] 3875 [CONSTRAINT] 3876 { 3877 OUTPUT-SECTION-COMMAND 3878 OUTPUT-SECTION-COMMAND 3879 ... 3880 } [>REGION] [AT>LMA_REGION] [:PHDR :PHDR ...] [=FILLEXP] 3881 3882 We've already described SECTION, ADDRESS, and 3883OUTPUT-SECTION-COMMAND. In this section we will describe the remaining 3884section attributes. 3885 3886* Menu: 3887 3888* Output Section Type:: Output section type 3889* Output Section LMA:: Output section LMA 3890* Forced Output Alignment:: Forced Output Alignment 3891* Forced Input Alignment:: Forced Input Alignment 3892* Output Section Constraint:: Output section constraint 3893* Output Section Region:: Output section region 3894* Output Section Phdr:: Output section phdr 3895* Output Section Fill:: Output section fill 3896 3897 3898File: ld.info, Node: Output Section Type, Next: Output Section LMA, Up: Output Section Attributes 3899 39003.6.8.1 Output Section Type 3901........................... 3902 3903Each output section may have a type. The type is a keyword in 3904parentheses. The following types are defined: 3905 3906`NOLOAD' 3907 The section should be marked as not loadable, so that it will not 3908 be loaded into memory when the program is run. 3909 3910`DSECT' 3911`COPY' 3912`INFO' 3913`OVERLAY' 3914 These type names are supported for backward compatibility, and are 3915 rarely used. They all have the same effect: the section should be 3916 marked as not allocatable, so that no memory is allocated for the 3917 section when the program is run. 3918 3919 The linker normally sets the attributes of an output section based on 3920the input sections which map into it. You can override this by using 3921the section type. For example, in the script sample below, the `ROM' 3922section is addressed at memory location `0' and does not need to be 3923loaded when the program is run. 3924 SECTIONS { 3925 ROM 0 (NOLOAD) : { ... } 3926 ... 3927 } 3928 3929 3930File: ld.info, Node: Output Section LMA, Next: Forced Output Alignment, Prev: Output Section Type, Up: Output Section Attributes 3931 39323.6.8.2 Output Section LMA 3933.......................... 3934 3935Every section has a virtual address (VMA) and a load address (LMA); see 3936*Note Basic Script Concepts::. The virtual address is specified by the 3937*note Output Section Address:: described earlier. The load address is 3938specified by the `AT' or `AT>' keywords. Specifying a load address is 3939optional. 3940 3941 The `AT' keyword takes an expression as an argument. This specifies 3942the exact load address of the section. The `AT>' keyword takes the 3943name of a memory region as an argument. *Note MEMORY::. The load 3944address of the section is set to the next free address in the region, 3945aligned to the section's alignment requirements. 3946 3947 If neither `AT' nor `AT>' is specified for an allocatable section, 3948the linker will use the following heuristic to determine the load 3949address: 3950 3951 * If the section has a specific VMA address, then this is used as 3952 the LMA address as well. 3953 3954 * If the section is not allocatable then its LMA is set to its VMA. 3955 3956 * Otherwise if a memory region can be found that is compatible with 3957 the current section, and this region contains at least one 3958 section, then the LMA is set so the difference between the VMA and 3959 LMA is the same as the difference between the VMA and LMA of the 3960 last section in the located region. 3961 3962 * If no memory regions have been declared then a default region that 3963 covers the entire address space is used in the previous step. 3964 3965 * If no suitable region could be found, or there was no previous 3966 section then the LMA is set equal to the VMA. 3967 3968 This feature is designed to make it easy to build a ROM image. For 3969example, the following linker script creates three output sections: one 3970called `.text', which starts at `0x1000', one called `.mdata', which is 3971loaded at the end of the `.text' section even though its VMA is 3972`0x2000', and one called `.bss' to hold uninitialized data at address 3973`0x3000'. The symbol `_data' is defined with the value `0x2000', which 3974shows that the location counter holds the VMA value, not the LMA value. 3975 3976 SECTIONS 3977 { 3978 .text 0x1000 : { *(.text) _etext = . ; } 3979 .mdata 0x2000 : 3980 AT ( ADDR (.text) + SIZEOF (.text) ) 3981 { _data = . ; *(.data); _edata = . ; } 3982 .bss 0x3000 : 3983 { _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;} 3984 } 3985 3986 The run-time initialization code for use with a program generated 3987with this linker script would include something like the following, to 3988copy the initialized data from the ROM image to its runtime address. 3989Notice how this code takes advantage of the symbols defined by the 3990linker script. 3991 3992 extern char _etext, _data, _edata, _bstart, _bend; 3993 char *src = &_etext; 3994 char *dst = &_data; 3995 3996 /* ROM has data at end of text; copy it. */ 3997 while (dst < &_edata) 3998 *dst++ = *src++; 3999 4000 /* Zero bss. */ 4001 for (dst = &_bstart; dst< &_bend; dst++) 4002 *dst = 0; 4003 4004 4005File: ld.info, Node: Forced Output Alignment, Next: Forced Input Alignment, Prev: Output Section LMA, Up: Output Section Attributes 4006 40073.6.8.3 Forced Output Alignment 4008............................... 4009 4010You can increase an output section's alignment by using ALIGN. As an 4011alternative you can enforce that the difference between the VMA and LMA 4012remains intact throughout this output section with the ALIGN_WITH_INPUT 4013attribute. 4014 4015 4016File: ld.info, Node: Forced Input Alignment, Next: Output Section Constraint, Prev: Forced Output Alignment, Up: Output Section Attributes 4017 40183.6.8.4 Forced Input Alignment 4019.............................. 4020 4021You can force input section alignment within an output section by using 4022SUBALIGN. The value specified overrides any alignment given by input 4023sections, whether larger or smaller. 4024 4025 4026File: ld.info, Node: Output Section Constraint, Next: Output Section Region, Prev: Forced Input Alignment, Up: Output Section Attributes 4027 40283.6.8.5 Output Section Constraint 4029................................. 4030 4031You can specify that an output section should only be created if all of 4032its input sections are read-only or all of its input sections are 4033read-write by using the keyword `ONLY_IF_RO' and `ONLY_IF_RW' 4034respectively. 4035 4036 4037File: ld.info, Node: Output Section Region, Next: Output Section Phdr, Prev: Output Section Constraint, Up: Output Section Attributes 4038 40393.6.8.6 Output Section Region 4040............................. 4041 4042You can assign a section to a previously defined region of memory by 4043using `>REGION'. *Note MEMORY::. 4044 4045 Here is a simple example: 4046 MEMORY { rom : ORIGIN = 0x1000, LENGTH = 0x1000 } 4047 SECTIONS { ROM : { *(.text) } >rom } 4048 4049 4050File: ld.info, Node: Output Section Phdr, Next: Output Section Fill, Prev: Output Section Region, Up: Output Section Attributes 4051 40523.6.8.7 Output Section Phdr 4053........................... 4054 4055You can assign a section to a previously defined program segment by 4056using `:PHDR'. *Note PHDRS::. If a section is assigned to one or more 4057segments, then all subsequent allocated sections will be assigned to 4058those segments as well, unless they use an explicitly `:PHDR' modifier. 4059You can use `:NONE' to tell the linker to not put the section in any 4060segment at all. 4061 4062 Here is a simple example: 4063 PHDRS { text PT_LOAD ; } 4064 SECTIONS { .text : { *(.text) } :text } 4065 4066 4067File: ld.info, Node: Output Section Fill, Prev: Output Section Phdr, Up: Output Section Attributes 4068 40693.6.8.8 Output Section Fill 4070........................... 4071 4072You can set the fill pattern for an entire section by using `=FILLEXP'. 4073FILLEXP is an expression (*note Expressions::). Any otherwise 4074unspecified regions of memory within the output section (for example, 4075gaps left due to the required alignment of input sections) will be 4076filled with the value, repeated as necessary. If the fill expression 4077is a simple hex number, ie. a string of hex digit starting with `0x' 4078and without a trailing `k' or `M', then an arbitrarily long sequence of 4079hex digits can be used to specify the fill pattern; Leading zeros 4080become part of the pattern too. For all other cases, including extra 4081parentheses or a unary `+', the fill pattern is the four least 4082significant bytes of the value of the expression. In all cases, the 4083number is big-endian. 4084 4085 You can also change the fill value with a `FILL' command in the 4086output section commands; (*note Output Section Data::). 4087 4088 Here is a simple example: 4089 SECTIONS { .text : { *(.text) } =0x90909090 } 4090 4091 4092File: ld.info, Node: Overlay Description, Prev: Output Section Attributes, Up: SECTIONS 4093 40943.6.9 Overlay Description 4095------------------------- 4096 4097An overlay description provides an easy way to describe sections which 4098are to be loaded as part of a single memory image but are to be run at 4099the same memory address. At run time, some sort of overlay manager will 4100copy the overlaid sections in and out of the runtime memory address as 4101required, perhaps by simply manipulating addressing bits. This approach 4102can be useful, for example, when a certain region of memory is faster 4103than another. 4104 4105 Overlays are described using the `OVERLAY' command. The `OVERLAY' 4106command is used within a `SECTIONS' command, like an output section 4107description. The full syntax of the `OVERLAY' command is as follows: 4108 OVERLAY [START] : [NOCROSSREFS] [AT ( LDADDR )] 4109 { 4110 SECNAME1 4111 { 4112 OUTPUT-SECTION-COMMAND 4113 OUTPUT-SECTION-COMMAND 4114 ... 4115 } [:PHDR...] [=FILL] 4116 SECNAME2 4117 { 4118 OUTPUT-SECTION-COMMAND 4119 OUTPUT-SECTION-COMMAND 4120 ... 4121 } [:PHDR...] [=FILL] 4122 ... 4123 } [>REGION] [:PHDR...] [=FILL] [,] 4124 4125 Everything is optional except `OVERLAY' (a keyword), and each 4126section must have a name (SECNAME1 and SECNAME2 above). The section 4127definitions within the `OVERLAY' construct are identical to those 4128within the general `SECTIONS' construct (*note SECTIONS::), except that 4129no addresses and no memory regions may be defined for sections within 4130an `OVERLAY'. 4131 4132 The comma at the end may be required if a FILL is used and the next 4133SECTIONS-COMMAND looks like a continuation of the expression. 4134 4135 The sections are all defined with the same starting address. The 4136load addresses of the sections are arranged such that they are 4137consecutive in memory starting at the load address used for the 4138`OVERLAY' as a whole (as with normal section definitions, the load 4139address is optional, and defaults to the start address; the start 4140address is also optional, and defaults to the current value of the 4141location counter). 4142 4143 If the `NOCROSSREFS' keyword is used, and there are any references 4144among the sections, the linker will report an error. Since the 4145sections all run at the same address, it normally does not make sense 4146for one section to refer directly to another. *Note NOCROSSREFS: 4147Miscellaneous Commands. 4148 4149 For each section within the `OVERLAY', the linker automatically 4150provides two symbols. The symbol `__load_start_SECNAME' is defined as 4151the starting load address of the section. The symbol 4152`__load_stop_SECNAME' is defined as the final load address of the 4153section. Any characters within SECNAME which are not legal within C 4154identifiers are removed. C (or assembler) code may use these symbols 4155to move the overlaid sections around as necessary. 4156 4157 At the end of the overlay, the value of the location counter is set 4158to the start address of the overlay plus the size of the largest 4159section. 4160 4161 Here is an example. Remember that this would appear inside a 4162`SECTIONS' construct. 4163 OVERLAY 0x1000 : AT (0x4000) 4164 { 4165 .text0 { o1/*.o(.text) } 4166 .text1 { o2/*.o(.text) } 4167 } 4168This will define both `.text0' and `.text1' to start at address 41690x1000. `.text0' will be loaded at address 0x4000, and `.text1' will 4170be loaded immediately after `.text0'. The following symbols will be 4171defined if referenced: `__load_start_text0', `__load_stop_text0', 4172`__load_start_text1', `__load_stop_text1'. 4173 4174 C code to copy overlay `.text1' into the overlay area might look 4175like the following. 4176 4177 extern char __load_start_text1, __load_stop_text1; 4178 memcpy ((char *) 0x1000, &__load_start_text1, 4179 &__load_stop_text1 - &__load_start_text1); 4180 4181 Note that the `OVERLAY' command is just syntactic sugar, since 4182everything it does can be done using the more basic commands. The above 4183example could have been written identically as follows. 4184 4185 .text0 0x1000 : AT (0x4000) { o1/*.o(.text) } 4186 PROVIDE (__load_start_text0 = LOADADDR (.text0)); 4187 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0)); 4188 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) { o2/*.o(.text) } 4189 PROVIDE (__load_start_text1 = LOADADDR (.text1)); 4190 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1)); 4191 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1)); 4192 4193 4194File: ld.info, Node: MEMORY, Next: PHDRS, Prev: SECTIONS, Up: Scripts 4195 41963.7 MEMORY Command 4197================== 4198 4199The linker's default configuration permits allocation of all available 4200memory. You can override this by using the `MEMORY' command. 4201 4202 The `MEMORY' command describes the location and size of blocks of 4203memory in the target. You can use it to describe which memory regions 4204may be used by the linker, and which memory regions it must avoid. You 4205can then assign sections to particular memory regions. The linker will 4206set section addresses based on the memory regions, and will warn about 4207regions that become too full. The linker will not shuffle sections 4208around to fit into the available regions. 4209 4210 A linker script may contain many uses of the `MEMORY' command, 4211however, all memory blocks defined are treated as if they were 4212specified inside a single `MEMORY' command. The syntax for `MEMORY' is: 4213 MEMORY 4214 { 4215 NAME [(ATTR)] : ORIGIN = ORIGIN, LENGTH = LEN 4216 ... 4217 } 4218 4219 The NAME is a name used in the linker script to refer to the region. 4220The region name has no meaning outside of the linker script. Region 4221names are stored in a separate name space, and will not conflict with 4222symbol names, file names, or section names. Each memory region must 4223have a distinct name within the `MEMORY' command. However you can add 4224later alias names to existing memory regions with the *Note 4225REGION_ALIAS:: command. 4226 4227 The ATTR string is an optional list of attributes that specify 4228whether to use a particular memory region for an input section which is 4229not explicitly mapped in the linker script. As described in *Note 4230SECTIONS::, if you do not specify an output section for some input 4231section, the linker will create an output section with the same name as 4232the input section. If you define region attributes, the linker will use 4233them to select the memory region for the output section that it creates. 4234 4235 The ATTR string must consist only of the following characters: 4236`R' 4237 Read-only section 4238 4239`W' 4240 Read/write section 4241 4242`X' 4243 Executable section 4244 4245`A' 4246 Allocatable section 4247 4248`I' 4249 Initialized section 4250 4251`L' 4252 Same as `I' 4253 4254`!' 4255 Invert the sense of any of the attributes that follow 4256 4257 If a unmapped section matches any of the listed attributes other than 4258`!', it will be placed in the memory region. The `!' attribute 4259reverses this test, so that an unmapped section will be placed in the 4260memory region only if it does not match any of the listed attributes. 4261 4262 The ORIGIN is an numerical expression for the start address of the 4263memory region. The expression must evaluate to a constant and it 4264cannot involve any symbols. The keyword `ORIGIN' may be abbreviated to 4265`org' or `o' (but not, for example, `ORG'). 4266 4267 The LEN is an expression for the size in bytes of the memory region. 4268As with the ORIGIN expression, the expression must be numerical only 4269and must evaluate to a constant. The keyword `LENGTH' may be 4270abbreviated to `len' or `l'. 4271 4272 In the following example, we specify that there are two memory 4273regions available for allocation: one starting at `0' for 256 kilobytes, 4274and the other starting at `0x40000000' for four megabytes. The linker 4275will place into the `rom' memory region every section which is not 4276explicitly mapped into a memory region, and is either read-only or 4277executable. The linker will place other sections which are not 4278explicitly mapped into a memory region into the `ram' memory region. 4279 4280 MEMORY 4281 { 4282 rom (rx) : ORIGIN = 0, LENGTH = 256K 4283 ram (!rx) : org = 0x40000000, l = 4M 4284 } 4285 4286 Once you define a memory region, you can direct the linker to place 4287specific output sections into that memory region by using the `>REGION' 4288output section attribute. For example, if you have a memory region 4289named `mem', you would use `>mem' in the output section definition. 4290*Note Output Section Region::. If no address was specified for the 4291output section, the linker will set the address to the next available 4292address within the memory region. If the combined output sections 4293directed to a memory region are too large for the region, the linker 4294will issue an error message. 4295 4296 It is possible to access the origin and length of a memory in an 4297expression via the `ORIGIN(MEMORY)' and `LENGTH(MEMORY)' functions: 4298 4299 _fstack = ORIGIN(ram) + LENGTH(ram) - 4; 4300 4301 4302File: ld.info, Node: PHDRS, Next: VERSION, Prev: MEMORY, Up: Scripts 4303 43043.8 PHDRS Command 4305================= 4306 4307The ELF object file format uses "program headers", also knows as 4308"segments". The program headers describe how the program should be 4309loaded into memory. You can print them out by using the `objdump' 4310program with the `-p' option. 4311 4312 When you run an ELF program on a native ELF system, the system loader 4313reads the program headers in order to figure out how to load the 4314program. This will only work if the program headers are set correctly. 4315This manual does not describe the details of how the system loader 4316interprets program headers; for more information, see the ELF ABI. 4317 4318 The linker will create reasonable program headers by default. 4319However, in some cases, you may need to specify the program headers more 4320precisely. You may use the `PHDRS' command for this purpose. When the 4321linker sees the `PHDRS' command in the linker script, it will not 4322create any program headers other than the ones specified. 4323 4324 The linker only pays attention to the `PHDRS' command when 4325generating an ELF output file. In other cases, the linker will simply 4326ignore `PHDRS'. 4327 4328 This is the syntax of the `PHDRS' command. The words `PHDRS', 4329`FILEHDR', `AT', and `FLAGS' are keywords. 4330 4331 PHDRS 4332 { 4333 NAME TYPE [ FILEHDR ] [ PHDRS ] [ AT ( ADDRESS ) ] 4334 [ FLAGS ( FLAGS ) ] ; 4335 } 4336 4337 The NAME is used only for reference in the `SECTIONS' command of the 4338linker script. It is not put into the output file. Program header 4339names are stored in a separate name space, and will not conflict with 4340symbol names, file names, or section names. Each program header must 4341have a distinct name. The headers are processed in order and it is 4342usual for them to map to sections in ascending load address order. 4343 4344 Certain program header types describe segments of memory which the 4345system loader will load from the file. In the linker script, you 4346specify the contents of these segments by placing allocatable output 4347sections in the segments. You use the `:PHDR' output section attribute 4348to place a section in a particular segment. *Note Output Section 4349Phdr::. 4350 4351 It is normal to put certain sections in more than one segment. This 4352merely implies that one segment of memory contains another. You may 4353repeat `:PHDR', using it once for each segment which should contain the 4354section. 4355 4356 If you place a section in one or more segments using `:PHDR', then 4357the linker will place all subsequent allocatable sections which do not 4358specify `:PHDR' in the same segments. This is for convenience, since 4359generally a whole set of contiguous sections will be placed in a single 4360segment. You can use `:NONE' to override the default segment and tell 4361the linker to not put the section in any segment at all. 4362 4363 You may use the `FILEHDR' and `PHDRS' keywords after the program 4364header type to further describe the contents of the segment. The 4365`FILEHDR' keyword means that the segment should include the ELF file 4366header. The `PHDRS' keyword means that the segment should include the 4367ELF program headers themselves. If applied to a loadable segment 4368(`PT_LOAD'), all prior loadable segments must have one of these 4369keywords. 4370 4371 The TYPE may be one of the following. The numbers indicate the 4372value of the keyword. 4373 4374`PT_NULL' (0) 4375 Indicates an unused program header. 4376 4377`PT_LOAD' (1) 4378 Indicates that this program header describes a segment to be 4379 loaded from the file. 4380 4381`PT_DYNAMIC' (2) 4382 Indicates a segment where dynamic linking information can be found. 4383 4384`PT_INTERP' (3) 4385 Indicates a segment where the name of the program interpreter may 4386 be found. 4387 4388`PT_NOTE' (4) 4389 Indicates a segment holding note information. 4390 4391`PT_SHLIB' (5) 4392 A reserved program header type, defined but not specified by the 4393 ELF ABI. 4394 4395`PT_PHDR' (6) 4396 Indicates a segment where the program headers may be found. 4397 4398EXPRESSION 4399 An expression giving the numeric type of the program header. This 4400 may be used for types not defined above. 4401 4402 You can specify that a segment should be loaded at a particular 4403address in memory by using an `AT' expression. This is identical to the 4404`AT' command used as an output section attribute (*note Output Section 4405LMA::). The `AT' command for a program header overrides the output 4406section attribute. 4407 4408 The linker will normally set the segment flags based on the sections 4409which comprise the segment. You may use the `FLAGS' keyword to 4410explicitly specify the segment flags. The value of FLAGS must be an 4411integer. It is used to set the `p_flags' field of the program header. 4412 4413 Here is an example of `PHDRS'. This shows a typical set of program 4414headers used on a native ELF system. 4415 4416 PHDRS 4417 { 4418 headers PT_PHDR PHDRS ; 4419 interp PT_INTERP ; 4420 text PT_LOAD FILEHDR PHDRS ; 4421 data PT_LOAD ; 4422 dynamic PT_DYNAMIC ; 4423 } 4424 4425 SECTIONS 4426 { 4427 . = SIZEOF_HEADERS; 4428 .interp : { *(.interp) } :text :interp 4429 .text : { *(.text) } :text 4430 .rodata : { *(.rodata) } /* defaults to :text */ 4431 ... 4432 . = . + 0x1000; /* move to a new page in memory */ 4433 .data : { *(.data) } :data 4434 .dynamic : { *(.dynamic) } :data :dynamic 4435 ... 4436 } 4437 4438 4439File: ld.info, Node: VERSION, Next: Expressions, Prev: PHDRS, Up: Scripts 4440 44413.9 VERSION Command 4442=================== 4443 4444The linker supports symbol versions when using ELF. Symbol versions are 4445only useful when using shared libraries. The dynamic linker can use 4446symbol versions to select a specific version of a function when it runs 4447a program that may have been linked against an earlier version of the 4448shared library. 4449 4450 You can include a version script directly in the main linker script, 4451or you can supply the version script as an implicit linker script. You 4452can also use the `--version-script' linker option. 4453 4454 The syntax of the `VERSION' command is simply 4455 VERSION { version-script-commands } 4456 4457 The format of the version script commands is identical to that used 4458by Sun's linker in Solaris 2.5. The version script defines a tree of 4459version nodes. You specify the node names and interdependencies in the 4460version script. You can specify which symbols are bound to which 4461version nodes, and you can reduce a specified set of symbols to local 4462scope so that they are not globally visible outside of the shared 4463library. 4464 4465 The easiest way to demonstrate the version script language is with a 4466few examples. 4467 4468 VERS_1.1 { 4469 global: 4470 foo1; 4471 local: 4472 old*; 4473 original*; 4474 new*; 4475 }; 4476 4477 VERS_1.2 { 4478 foo2; 4479 } VERS_1.1; 4480 4481 VERS_2.0 { 4482 bar1; bar2; 4483 extern "C++" { 4484 ns::*; 4485 "f(int, double)"; 4486 }; 4487 } VERS_1.2; 4488 4489 This example version script defines three version nodes. The first 4490version node defined is `VERS_1.1'; it has no other dependencies. The 4491script binds the symbol `foo1' to `VERS_1.1'. It reduces a number of 4492symbols to local scope so that they are not visible outside of the 4493shared library; this is done using wildcard patterns, so that any 4494symbol whose name begins with `old', `original', or `new' is matched. 4495The wildcard patterns available are the same as those used in the shell 4496when matching filenames (also known as "globbing"). However, if you 4497specify the symbol name inside double quotes, then the name is treated 4498as literal, rather than as a glob pattern. 4499 4500 Next, the version script defines node `VERS_1.2'. This node depends 4501upon `VERS_1.1'. The script binds the symbol `foo2' to the version 4502node `VERS_1.2'. 4503 4504 Finally, the version script defines node `VERS_2.0'. This node 4505depends upon `VERS_1.2'. The scripts binds the symbols `bar1' and 4506`bar2' are bound to the version node `VERS_2.0'. 4507 4508 When the linker finds a symbol defined in a library which is not 4509specifically bound to a version node, it will effectively bind it to an 4510unspecified base version of the library. You can bind all otherwise 4511unspecified symbols to a given version node by using `global: *;' 4512somewhere in the version script. Note that it's slightly crazy to use 4513wildcards in a global spec except on the last version node. Global 4514wildcards elsewhere run the risk of accidentally adding symbols to the 4515set exported for an old version. That's wrong since older versions 4516ought to have a fixed set of symbols. 4517 4518 The names of the version nodes have no specific meaning other than 4519what they might suggest to the person reading them. The `2.0' version 4520could just as well have appeared in between `1.1' and `1.2'. However, 4521this would be a confusing way to write a version script. 4522 4523 Node name can be omitted, provided it is the only version node in 4524the version script. Such version script doesn't assign any versions to 4525symbols, only selects which symbols will be globally visible out and 4526which won't. 4527 4528 { global: foo; bar; local: *; }; 4529 4530 When you link an application against a shared library that has 4531versioned symbols, the application itself knows which version of each 4532symbol it requires, and it also knows which version nodes it needs from 4533each shared library it is linked against. Thus at runtime, the dynamic 4534loader can make a quick check to make sure that the libraries you have 4535linked against do in fact supply all of the version nodes that the 4536application will need to resolve all of the dynamic symbols. In this 4537way it is possible for the dynamic linker to know with certainty that 4538all external symbols that it needs will be resolvable without having to 4539search for each symbol reference. 4540 4541 The symbol versioning is in effect a much more sophisticated way of 4542doing minor version checking that SunOS does. The fundamental problem 4543that is being addressed here is that typically references to external 4544functions are bound on an as-needed basis, and are not all bound when 4545the application starts up. If a shared library is out of date, a 4546required interface may be missing; when the application tries to use 4547that interface, it may suddenly and unexpectedly fail. With symbol 4548versioning, the user will get a warning when they start their program if 4549the libraries being used with the application are too old. 4550 4551 There are several GNU extensions to Sun's versioning approach. The 4552first of these is the ability to bind a symbol to a version node in the 4553source file where the symbol is defined instead of in the versioning 4554script. This was done mainly to reduce the burden on the library 4555maintainer. You can do this by putting something like: 4556 __asm__(".symver original_foo,foo@VERS_1.1"); 4557 in the C source file. This renames the function `original_foo' to 4558be an alias for `foo' bound to the version node `VERS_1.1'. The 4559`local:' directive can be used to prevent the symbol `original_foo' 4560from being exported. A `.symver' directive takes precedence over a 4561version script. 4562 4563 The second GNU extension is to allow multiple versions of the same 4564function to appear in a given shared library. In this way you can make 4565an incompatible change to an interface without increasing the major 4566version number of the shared library, while still allowing applications 4567linked against the old interface to continue to function. 4568 4569 To do this, you must use multiple `.symver' directives in the source 4570file. Here is an example: 4571 4572 __asm__(".symver original_foo,foo@"); 4573 __asm__(".symver old_foo,foo@VERS_1.1"); 4574 __asm__(".symver old_foo1,foo@VERS_1.2"); 4575 __asm__(".symver new_foo,foo@@VERS_2.0"); 4576 4577 In this example, `foo@' represents the symbol `foo' bound to the 4578unspecified base version of the symbol. The source file that contains 4579this example would define 4 C functions: `original_foo', `old_foo', 4580`old_foo1', and `new_foo'. 4581 4582 When you have multiple definitions of a given symbol, there needs to 4583be some way to specify a default version to which external references to 4584this symbol will be bound. You can do this with the `foo@@VERS_2.0' 4585type of `.symver' directive. You can only declare one version of a 4586symbol as the default in this manner; otherwise you would effectively 4587have multiple definitions of the same symbol. 4588 4589 If you wish to bind a reference to a specific version of the symbol 4590within the shared library, you can use the aliases of convenience 4591(i.e., `old_foo'), or you can use the `.symver' directive to 4592specifically bind to an external version of the function in question. 4593 4594 You can also specify the language in the version script: 4595 4596 VERSION extern "lang" { version-script-commands } 4597 4598 The supported `lang's are `C', `C++', and `Java'. The linker will 4599iterate over the list of symbols at the link time and demangle them 4600according to `lang' before matching them to the patterns specified in 4601`version-script-commands'. The default `lang' is `C'. 4602 4603 Demangled names may contains spaces and other special characters. As 4604described above, you can use a glob pattern to match demangled names, 4605or you can use a double-quoted string to match the string exactly. In 4606the latter case, be aware that minor differences (such as differing 4607whitespace) between the version script and the demangler output will 4608cause a mismatch. As the exact string generated by the demangler might 4609change in the future, even if the mangled name does not, you should 4610check that all of your version directives are behaving as you expect 4611when you upgrade. 4612 4613 4614File: ld.info, Node: Expressions, Next: Implicit Linker Scripts, Prev: VERSION, Up: Scripts 4615 46163.10 Expressions in Linker Scripts 4617================================== 4618 4619The syntax for expressions in the linker script language is identical to 4620that of C expressions. All expressions are evaluated as integers. All 4621expressions are evaluated in the same size, which is 32 bits if both the 4622host and target are 32 bits, and is otherwise 64 bits. 4623 4624 You can use and set symbol values in expressions. 4625 4626 The linker defines several special purpose builtin functions for use 4627in expressions. 4628 4629* Menu: 4630 4631* Constants:: Constants 4632* Symbolic Constants:: Symbolic constants 4633* Symbols:: Symbol Names 4634* Orphan Sections:: Orphan Sections 4635* Location Counter:: The Location Counter 4636* Operators:: Operators 4637* Evaluation:: Evaluation 4638* Expression Section:: The Section of an Expression 4639* Builtin Functions:: Builtin Functions 4640 4641 4642File: ld.info, Node: Constants, Next: Symbolic Constants, Up: Expressions 4643 46443.10.1 Constants 4645---------------- 4646 4647All constants are integers. 4648 4649 As in C, the linker considers an integer beginning with `0' to be 4650octal, and an integer beginning with `0x' or `0X' to be hexadecimal. 4651Alternatively the linker accepts suffixes of `h' or `H' for 4652hexadecimal, `o' or `O' for octal, `b' or `B' for binary and `d' or `D' 4653for decimal. Any integer value without a prefix or a suffix is 4654considered to be decimal. 4655 4656 In addition, you can use the suffixes `K' and `M' to scale a 4657constant by `1024' or `1024*1024' respectively. For example, the 4658following all refer to the same quantity: 4659 4660 _fourk_1 = 4K; 4661 _fourk_2 = 4096; 4662 _fourk_3 = 0x1000; 4663 _fourk_4 = 10000o; 4664 4665 Note - the `K' and `M' suffixes cannot be used in conjunction with 4666the base suffixes mentioned above. 4667 4668 4669File: ld.info, Node: Symbolic Constants, Next: Symbols, Prev: Constants, Up: Expressions 4670 46713.10.2 Symbolic Constants 4672------------------------- 4673 4674It is possible to refer to target specific constants via the use of the 4675`CONSTANT(NAME)' operator, where NAME is one of: 4676 4677`MAXPAGESIZE' 4678 The target's maximum page size. 4679 4680`COMMONPAGESIZE' 4681 The target's default page size. 4682 4683 So for example: 4684 4685 .text ALIGN (CONSTANT (MAXPAGESIZE)) : { *(.text) } 4686 4687 will create a text section aligned to the largest page boundary 4688supported by the target. 4689 4690 4691File: ld.info, Node: Symbols, Next: Orphan Sections, Prev: Symbolic Constants, Up: Expressions 4692 46933.10.3 Symbol Names 4694------------------- 4695 4696Unless quoted, symbol names start with a letter, underscore, or period 4697and may include letters, digits, underscores, periods, and hyphens. 4698Unquoted symbol names must not conflict with any keywords. You can 4699specify a symbol which contains odd characters or has the same name as a 4700keyword by surrounding the symbol name in double quotes: 4701 "SECTION" = 9; 4702 "with a space" = "also with a space" + 10; 4703 4704 Since symbols can contain many non-alphabetic characters, it is 4705safest to delimit symbols with spaces. For example, `A-B' is one 4706symbol, whereas `A - B' is an expression involving subtraction. 4707 4708 4709File: ld.info, Node: Orphan Sections, Next: Location Counter, Prev: Symbols, Up: Expressions 4710 47113.10.4 Orphan Sections 4712---------------------- 4713 4714Orphan sections are sections present in the input files which are not 4715explicitly placed into the output file by the linker script. The 4716linker will still copy these sections into the output file, but it has 4717to guess as to where they should be placed. The linker uses a simple 4718heuristic to do this. It attempts to place orphan sections after 4719non-orphan sections of the same attribute, such as code vs data, 4720loadable vs non-loadable, etc. If there is not enough room to do this 4721then it places at the end of the file. 4722 4723 For ELF targets, the attribute of the section includes section type 4724as well as section flag. 4725 4726 The command line options `--orphan-handling' and `--unique' (*note 4727Command Line Options: Options.) can be used to control which output 4728sections an orphan is placed in. 4729 4730 If an orphaned section's name is representable as a C identifier then 4731the linker will automatically *note PROVIDE:: two symbols: 4732__start_SECNAME and __stop_SECNAME, where SECNAME is the name of the 4733section. These indicate the start address and end address of the 4734orphaned section respectively. Note: most section names are not 4735representable as C identifiers because they contain a `.' character. 4736 4737 4738File: ld.info, Node: Location Counter, Next: Operators, Prev: Orphan Sections, Up: Expressions 4739 47403.10.5 The Location Counter 4741--------------------------- 4742 4743The special linker variable "dot" `.' always contains the current 4744output location counter. Since the `.' always refers to a location in 4745an output section, it may only appear in an expression within a 4746`SECTIONS' command. The `.' symbol may appear anywhere that an 4747ordinary symbol is allowed in an expression. 4748 4749 Assigning a value to `.' will cause the location counter to be 4750moved. This may be used to create holes in the output section. The 4751location counter may not be moved backwards inside an output section, 4752and may not be moved backwards outside of an output section if so doing 4753creates areas with overlapping LMAs. 4754 4755 SECTIONS 4756 { 4757 output : 4758 { 4759 file1(.text) 4760 . = . + 1000; 4761 file2(.text) 4762 . += 1000; 4763 file3(.text) 4764 } = 0x12345678; 4765 } 4766 In the previous example, the `.text' section from `file1' is located 4767at the beginning of the output section `output'. It is followed by a 47681000 byte gap. Then the `.text' section from `file2' appears, also 4769with a 1000 byte gap following before the `.text' section from `file3'. 4770The notation `= 0x12345678' specifies what data to write in the gaps 4771(*note Output Section Fill::). 4772 4773 Note: `.' actually refers to the byte offset from the start of the 4774current containing object. Normally this is the `SECTIONS' statement, 4775whose start address is 0, hence `.' can be used as an absolute address. 4776If `.' is used inside a section description however, it refers to the 4777byte offset from the start of that section, not an absolute address. 4778Thus in a script like this: 4779 4780 SECTIONS 4781 { 4782 . = 0x100 4783 .text: { 4784 *(.text) 4785 . = 0x200 4786 } 4787 . = 0x500 4788 .data: { 4789 *(.data) 4790 . += 0x600 4791 } 4792 } 4793 4794 The `.text' section will be assigned a starting address of 0x100 and 4795a size of exactly 0x200 bytes, even if there is not enough data in the 4796`.text' input sections to fill this area. (If there is too much data, 4797an error will be produced because this would be an attempt to move `.' 4798backwards). The `.data' section will start at 0x500 and it will have 4799an extra 0x600 bytes worth of space after the end of the values from 4800the `.data' input sections and before the end of the `.data' output 4801section itself. 4802 4803 Setting symbols to the value of the location counter outside of an 4804output section statement can result in unexpected values if the linker 4805needs to place orphan sections. For example, given the following: 4806 4807 SECTIONS 4808 { 4809 start_of_text = . ; 4810 .text: { *(.text) } 4811 end_of_text = . ; 4812 4813 start_of_data = . ; 4814 .data: { *(.data) } 4815 end_of_data = . ; 4816 } 4817 4818 If the linker needs to place some input section, e.g. `.rodata', not 4819mentioned in the script, it might choose to place that section between 4820`.text' and `.data'. You might think the linker should place `.rodata' 4821on the blank line in the above script, but blank lines are of no 4822particular significance to the linker. As well, the linker doesn't 4823associate the above symbol names with their sections. Instead, it 4824assumes that all assignments or other statements belong to the previous 4825output section, except for the special case of an assignment to `.'. 4826I.e., the linker will place the orphan `.rodata' section as if the 4827script was written as follows: 4828 4829 SECTIONS 4830 { 4831 start_of_text = . ; 4832 .text: { *(.text) } 4833 end_of_text = . ; 4834 4835 start_of_data = . ; 4836 .rodata: { *(.rodata) } 4837 .data: { *(.data) } 4838 end_of_data = . ; 4839 } 4840 4841 This may or may not be the script author's intention for the value of 4842`start_of_data'. One way to influence the orphan section placement is 4843to assign the location counter to itself, as the linker assumes that an 4844assignment to `.' is setting the start address of a following output 4845section and thus should be grouped with that section. So you could 4846write: 4847 4848 SECTIONS 4849 { 4850 start_of_text = . ; 4851 .text: { *(.text) } 4852 end_of_text = . ; 4853 4854 . = . ; 4855 start_of_data = . ; 4856 .data: { *(.data) } 4857 end_of_data = . ; 4858 } 4859 4860 Now, the orphan `.rodata' section will be placed between 4861`end_of_text' and `start_of_data'. 4862 4863 4864File: ld.info, Node: Operators, Next: Evaluation, Prev: Location Counter, Up: Expressions 4865 48663.10.6 Operators 4867---------------- 4868 4869The linker recognizes the standard C set of arithmetic operators, with 4870the standard bindings and precedence levels: 4871 precedence associativity Operators Notes 4872 (highest) 4873 1 left ! - ~ (1) 4874 2 left * / % 4875 3 left + - 4876 4 left >> << 4877 5 left == != > < <= >= 4878 6 left & 4879 7 left | 4880 8 left && 4881 9 left || 4882 10 right ? : 4883 11 right &= += -= *= /= (2) 4884 (lowest) 4885 Notes: (1) Prefix operators (2) *Note Assignments::. 4886 4887 4888File: ld.info, Node: Evaluation, Next: Expression Section, Prev: Operators, Up: Expressions 4889 48903.10.7 Evaluation 4891----------------- 4892 4893The linker evaluates expressions lazily. It only computes the value of 4894an expression when absolutely necessary. 4895 4896 The linker needs some information, such as the value of the start 4897address of the first section, and the origins and lengths of memory 4898regions, in order to do any linking at all. These values are computed 4899as soon as possible when the linker reads in the linker script. 4900 4901 However, other values (such as symbol values) are not known or needed 4902until after storage allocation. Such values are evaluated later, when 4903other information (such as the sizes of output sections) is available 4904for use in the symbol assignment expression. 4905 4906 The sizes of sections cannot be known until after allocation, so 4907assignments dependent upon these are not performed until after 4908allocation. 4909 4910 Some expressions, such as those depending upon the location counter 4911`.', must be evaluated during section allocation. 4912 4913 If the result of an expression is required, but the value is not 4914available, then an error results. For example, a script like the 4915following 4916 SECTIONS 4917 { 4918 .text 9+this_isnt_constant : 4919 { *(.text) } 4920 } 4921will cause the error message `non constant expression for initial 4922address'. 4923 4924 4925File: ld.info, Node: Expression Section, Next: Builtin Functions, Prev: Evaluation, Up: Expressions 4926 49273.10.8 The Section of an Expression 4928----------------------------------- 4929 4930Addresses and symbols may be section relative, or absolute. A section 4931relative symbol is relocatable. If you request relocatable output 4932using the `-r' option, a further link operation may change the value of 4933a section relative symbol. On the other hand, an absolute symbol will 4934retain the same value throughout any further link operations. 4935 4936 Some terms in linker expressions are addresses. This is true of 4937section relative symbols and for builtin functions that return an 4938address, such as `ADDR', `LOADADDR', `ORIGIN' and `SEGMENT_START'. 4939Other terms are simply numbers, or are builtin functions that return a 4940non-address value, such as `LENGTH'. One complication is that unless 4941you set `LD_FEATURE ("SANE_EXPR")' (*note Miscellaneous Commands::), 4942numbers and absolute symbols are treated differently depending on their 4943location, for compatibility with older versions of `ld'. Expressions 4944appearing outside an output section definition treat all numbers as 4945absolute addresses. Expressions appearing inside an output section 4946definition treat absolute symbols as numbers. If `LD_FEATURE 4947("SANE_EXPR")' is given, then absolute symbols and numbers are simply 4948treated as numbers everywhere. 4949 4950 In the following simple example, 4951 4952 SECTIONS 4953 { 4954 . = 0x100; 4955 __executable_start = 0x100; 4956 .data : 4957 { 4958 . = 0x10; 4959 __data_start = 0x10; 4960 *(.data) 4961 } 4962 ... 4963 } 4964 4965 both `.' and `__executable_start' are set to the absolute address 49660x100 in the first two assignments, then both `.' and `__data_start' 4967are set to 0x10 relative to the `.data' section in the second two 4968assignments. 4969 4970 For expressions involving numbers, relative addresses and absolute 4971addresses, ld follows these rules to evaluate terms: 4972 4973 * Unary operations on an absolute address or number, and binary 4974 operations on two absolute addresses or two numbers, or between one 4975 absolute address and a number, apply the operator to the value(s). 4976 4977 * Unary operations on a relative address, and binary operations on 4978 two relative addresses in the same section or between one relative 4979 address and a number, apply the operator to the offset part of the 4980 address(es). 4981 4982 * Other binary operations, that is, between two relative addresses 4983 not in the same section, or between a relative address and an 4984 absolute address, first convert any non-absolute term to an 4985 absolute address before applying the operator. 4986 4987 The result section of each sub-expression is as follows: 4988 4989 * An operation involving only numbers results in a number. 4990 4991 * The result of comparisons, `&&' and `||' is also a number. 4992 4993 * The result of other binary arithmetic and logical operations on two 4994 relative addresses in the same section or two absolute addresses 4995 (after above conversions) is also a number. 4996 4997 * The result of other operations on relative addresses or one 4998 relative address and a number, is a relative address in the same 4999 section as the relative operand(s). 5000 5001 * The result of other operations on absolute addresses (after above 5002 conversions) is an absolute address. 5003 5004 You can use the builtin function `ABSOLUTE' to force an expression 5005to be absolute when it would otherwise be relative. For example, to 5006create an absolute symbol set to the address of the end of the output 5007section `.data': 5008 SECTIONS 5009 { 5010 .data : { *(.data) _edata = ABSOLUTE(.); } 5011 } 5012 If `ABSOLUTE' were not used, `_edata' would be relative to the 5013`.data' section. 5014 5015 Using `LOADADDR' also forces an expression absolute, since this 5016particular builtin function returns an absolute address. 5017 5018 5019File: ld.info, Node: Builtin Functions, Prev: Expression Section, Up: Expressions 5020 50213.10.9 Builtin Functions 5022------------------------ 5023 5024The linker script language includes a number of builtin functions for 5025use in linker script expressions. 5026 5027`ABSOLUTE(EXP)' 5028 Return the absolute (non-relocatable, as opposed to non-negative) 5029 value of the expression EXP. Primarily useful to assign an 5030 absolute value to a symbol within a section definition, where 5031 symbol values are normally section relative. *Note Expression 5032 Section::. 5033 5034`ADDR(SECTION)' 5035 Return the address (VMA) of the named SECTION. Your script must 5036 previously have defined the location of that section. In the 5037 following example, `start_of_output_1', `symbol_1' and `symbol_2' 5038 are assigned equivalent values, except that `symbol_1' will be 5039 relative to the `.output1' section while the other two will be 5040 absolute: 5041 SECTIONS { ... 5042 .output1 : 5043 { 5044 start_of_output_1 = ABSOLUTE(.); 5045 ... 5046 } 5047 .output : 5048 { 5049 symbol_1 = ADDR(.output1); 5050 symbol_2 = start_of_output_1; 5051 } 5052 ... } 5053 5054`ALIGN(ALIGN)' 5055`ALIGN(EXP,ALIGN)' 5056 Return the location counter (`.') or arbitrary expression aligned 5057 to the next ALIGN boundary. The single operand `ALIGN' doesn't 5058 change the value of the location counter--it just does arithmetic 5059 on it. The two operand `ALIGN' allows an arbitrary expression to 5060 be aligned upwards (`ALIGN(ALIGN)' is equivalent to 5061 `ALIGN(ABSOLUTE(.), ALIGN)'). 5062 5063 Here is an example which aligns the output `.data' section to the 5064 next `0x2000' byte boundary after the preceding section and sets a 5065 variable within the section to the next `0x8000' boundary after the 5066 input sections: 5067 SECTIONS { ... 5068 .data ALIGN(0x2000): { 5069 *(.data) 5070 variable = ALIGN(0x8000); 5071 } 5072 ... } 5073 The first use of `ALIGN' in this example specifies the 5074 location of a section because it is used as the optional ADDRESS 5075 attribute of a section definition (*note Output Section 5076 Address::). The second use of `ALIGN' is used to defines the 5077 value of a symbol. 5078 5079 The builtin function `NEXT' is closely related to `ALIGN'. 5080 5081`ALIGNOF(SECTION)' 5082 Return the alignment in bytes of the named SECTION, if that 5083 section has been allocated. If the section has not been allocated 5084 when this is evaluated, the linker will report an error. In the 5085 following example, the alignment of the `.output' section is 5086 stored as the first value in that section. 5087 SECTIONS{ ... 5088 .output { 5089 LONG (ALIGNOF (.output)) 5090 ... 5091 } 5092 ... } 5093 5094`BLOCK(EXP)' 5095 This is a synonym for `ALIGN', for compatibility with older linker 5096 scripts. It is most often seen when setting the address of an 5097 output section. 5098 5099`DATA_SEGMENT_ALIGN(MAXPAGESIZE, COMMONPAGESIZE)' 5100 This is equivalent to either 5101 (ALIGN(MAXPAGESIZE) + (. & (MAXPAGESIZE - 1))) 5102 or 5103 (ALIGN(MAXPAGESIZE) 5104 + ((. + COMMONPAGESIZE - 1) & (MAXPAGESIZE - COMMONPAGESIZE))) 5105 depending on whether the latter uses fewer COMMONPAGESIZE sized 5106 pages for the data segment (area between the result of this 5107 expression and `DATA_SEGMENT_END') than the former or not. If the 5108 latter form is used, it means COMMONPAGESIZE bytes of runtime 5109 memory will be saved at the expense of up to COMMONPAGESIZE wasted 5110 bytes in the on-disk file. 5111 5112 This expression can only be used directly in `SECTIONS' commands, 5113 not in any output section descriptions and only once in the linker 5114 script. COMMONPAGESIZE should be less or equal to MAXPAGESIZE and 5115 should be the system page size the object wants to be optimized 5116 for (while still working on system page sizes up to MAXPAGESIZE). 5117 5118 Example: 5119 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000); 5120 5121`DATA_SEGMENT_END(EXP)' 5122 This defines the end of data segment for `DATA_SEGMENT_ALIGN' 5123 evaluation purposes. 5124 5125 . = DATA_SEGMENT_END(.); 5126 5127`DATA_SEGMENT_RELRO_END(OFFSET, EXP)' 5128 This defines the end of the `PT_GNU_RELRO' segment when `-z relro' 5129 option is used. When `-z relro' option is not present, 5130 `DATA_SEGMENT_RELRO_END' does nothing, otherwise 5131 `DATA_SEGMENT_ALIGN' is padded so that EXP + OFFSET is aligned to 5132 the most commonly used page boundary for particular target. If 5133 present in the linker script, it must always come in between 5134 `DATA_SEGMENT_ALIGN' and `DATA_SEGMENT_END'. Evaluates to the 5135 second argument plus any padding needed at the end of the 5136 `PT_GNU_RELRO' segment due to section alignment. 5137 5138 . = DATA_SEGMENT_RELRO_END(24, .); 5139 5140`DEFINED(SYMBOL)' 5141 Return 1 if SYMBOL is in the linker global symbol table and is 5142 defined before the statement using DEFINED in the script, otherwise 5143 return 0. You can use this function to provide default values for 5144 symbols. For example, the following script fragment shows how to 5145 set a global symbol `begin' to the first location in the `.text' 5146 section--but if a symbol called `begin' already existed, its value 5147 is preserved: 5148 5149 SECTIONS { ... 5150 .text : { 5151 begin = DEFINED(begin) ? begin : . ; 5152 ... 5153 } 5154 ... 5155 } 5156 5157`LENGTH(MEMORY)' 5158 Return the length of the memory region named MEMORY. 5159 5160`LOADADDR(SECTION)' 5161 Return the absolute LMA of the named SECTION. (*note Output 5162 Section LMA::). 5163 5164`LOG2CEIL(EXP)' 5165 Return the binary logarithm of EXP rounded towards infinity. 5166 `LOG2CEIL(0)' returns 0. 5167 5168`MAX(EXP1, EXP2)' 5169 Returns the maximum of EXP1 and EXP2. 5170 5171`MIN(EXP1, EXP2)' 5172 Returns the minimum of EXP1 and EXP2. 5173 5174`NEXT(EXP)' 5175 Return the next unallocated address that is a multiple of EXP. 5176 This function is closely related to `ALIGN(EXP)'; unless you use 5177 the `MEMORY' command to define discontinuous memory for the output 5178 file, the two functions are equivalent. 5179 5180`ORIGIN(MEMORY)' 5181 Return the origin of the memory region named MEMORY. 5182 5183`SEGMENT_START(SEGMENT, DEFAULT)' 5184 Return the base address of the named SEGMENT. If an explicit 5185 value has already been given for this segment (with a command-line 5186 `-T' option) then that value will be returned otherwise the value 5187 will be DEFAULT. At present, the `-T' command-line option can 5188 only be used to set the base address for the "text", "data", and 5189 "bss" sections, but you can use `SEGMENT_START' with any segment 5190 name. 5191 5192`SIZEOF(SECTION)' 5193 Return the size in bytes of the named SECTION, if that section has 5194 been allocated. If the section has not been allocated when this is 5195 evaluated, the linker will report an error. In the following 5196 example, `symbol_1' and `symbol_2' are assigned identical values: 5197 SECTIONS{ ... 5198 .output { 5199 .start = . ; 5200 ... 5201 .end = . ; 5202 } 5203 symbol_1 = .end - .start ; 5204 symbol_2 = SIZEOF(.output); 5205 ... } 5206 5207`SIZEOF_HEADERS' 5208`sizeof_headers' 5209 Return the size in bytes of the output file's headers. This is 5210 information which appears at the start of the output file. You 5211 can use this number when setting the start address of the first 5212 section, if you choose, to facilitate paging. 5213 5214 When producing an ELF output file, if the linker script uses the 5215 `SIZEOF_HEADERS' builtin function, the linker must compute the 5216 number of program headers before it has determined all the section 5217 addresses and sizes. If the linker later discovers that it needs 5218 additional program headers, it will report an error `not enough 5219 room for program headers'. To avoid this error, you must avoid 5220 using the `SIZEOF_HEADERS' function, or you must rework your linker 5221 script to avoid forcing the linker to use additional program 5222 headers, or you must define the program headers yourself using the 5223 `PHDRS' command (*note PHDRS::). 5224 5225 5226File: ld.info, Node: Implicit Linker Scripts, Prev: Expressions, Up: Scripts 5227 52283.11 Implicit Linker Scripts 5229============================ 5230 5231If you specify a linker input file which the linker can not recognize as 5232an object file or an archive file, it will try to read the file as a 5233linker script. If the file can not be parsed as a linker script, the 5234linker will report an error. 5235 5236 An implicit linker script will not replace the default linker script. 5237 5238 Typically an implicit linker script would contain only symbol 5239assignments, or the `INPUT', `GROUP', or `VERSION' commands. 5240 5241 Any input files read because of an implicit linker script will be 5242read at the position in the command line where the implicit linker 5243script was read. This can affect archive searching. 5244 5245 5246File: ld.info, Node: Machine Dependent, Next: BFD, Prev: Scripts, Up: Top 5247 52484 Machine Dependent Features 5249**************************** 5250 5251`ld' has additional features on some platforms; the following sections 5252describe them. Machines where `ld' has no additional functionality are 5253not listed. 5254 5255* Menu: 5256 5257 5258* H8/300:: `ld' and the H8/300 5259 5260* i960:: `ld' and the Intel 960 family 5261 5262* M68HC11/68HC12:: `ld' and the Motorola 68HC11 and 68HC12 families 5263 5264* ARM:: `ld' and the ARM family 5265 5266* HPPA ELF32:: `ld' and HPPA 32-bit ELF 5267 5268* M68K:: `ld' and the Motorola 68K family 5269 5270* MIPS:: `ld' and the MIPS family 5271 5272* MMIX:: `ld' and MMIX 5273 5274* MSP430:: `ld' and MSP430 5275 5276* NDS32:: `ld' and NDS32 5277 5278* Nios II:: `ld' and the Altera Nios II 5279 5280* PowerPC ELF32:: `ld' and PowerPC 32-bit ELF Support 5281 5282* PowerPC64 ELF64:: `ld' and PowerPC64 64-bit ELF Support 5283 5284* SPU ELF:: `ld' and SPU ELF Support 5285 5286* TI COFF:: `ld' and TI COFF 5287 5288* WIN32:: `ld' and WIN32 (cygwin/mingw) 5289 5290* Xtensa:: `ld' and Xtensa Processors 5291 5292 5293File: ld.info, Node: H8/300, Next: i960, Up: Machine Dependent 5294 52954.1 `ld' and the H8/300 5296======================= 5297 5298For the H8/300, `ld' can perform these global optimizations when you 5299specify the `--relax' command-line option. 5300 5301_relaxing address modes_ 5302 `ld' finds all `jsr' and `jmp' instructions whose targets are 5303 within eight bits, and turns them into eight-bit program-counter 5304 relative `bsr' and `bra' instructions, respectively. 5305 5306_synthesizing instructions_ 5307 `ld' finds all `mov.b' instructions which use the sixteen-bit 5308 absolute address form, but refer to the top page of memory, and 5309 changes them to use the eight-bit address form. (That is: the 5310 linker turns `mov.b `@'AA:16' into `mov.b `@'AA:8' whenever the 5311 address AA is in the top page of memory). 5312 5313 `ld' finds all `mov' instructions which use the register indirect 5314 with 32-bit displacement addressing mode, but use a small 5315 displacement inside 16-bit displacement range, and changes them to 5316 use the 16-bit displacement form. (That is: the linker turns 5317 `mov.b `@'D:32,ERx' into `mov.b `@'D:16,ERx' whenever the 5318 displacement D is in the 16 bit signed integer range. Only 5319 implemented in ELF-format ld). 5320 5321_bit manipulation instructions_ 5322 `ld' finds all bit manipulation instructions like `band, bclr, 5323 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, 5324 bxor' which use 32 bit and 16 bit absolute address form, but refer 5325 to the top page of memory, and changes them to use the 8 bit 5326 address form. (That is: the linker turns `bset #xx:3,`@'AA:32' 5327 into `bset #xx:3,`@'AA:8' whenever the address AA is in the top 5328 page of memory). 5329 5330_system control instructions_ 5331 `ld' finds all `ldc.w, stc.w' instructions which use the 32 bit 5332 absolute address form, but refer to the top page of memory, and 5333 changes them to use 16 bit address form. (That is: the linker 5334 turns `ldc.w `@'AA:32,ccr' into `ldc.w `@'AA:16,ccr' whenever the 5335 address AA is in the top page of memory). 5336 5337 5338File: ld.info, Node: i960, Next: M68HC11/68HC12, Prev: H8/300, Up: Machine Dependent 5339 53404.2 `ld' and the Intel 960 Family 5341================================= 5342 5343You can use the `-AARCHITECTURE' command line option to specify one of 5344the two-letter names identifying members of the 960 family; the option 5345specifies the desired output target, and warns of any incompatible 5346instructions in the input files. It also modifies the linker's search 5347strategy for archive libraries, to support the use of libraries 5348specific to each particular architecture, by including in the search 5349loop names suffixed with the string identifying the architecture. 5350 5351 For example, if your `ld' command line included `-ACA' as well as 5352`-ltry', the linker would look (in its built-in search paths, and in 5353any paths you specify with `-L') for a library with the names 5354 5355 try 5356 libtry.a 5357 tryca 5358 libtryca.a 5359 5360The first two possibilities would be considered in any event; the last 5361two are due to the use of `-ACA'. 5362 5363 You can meaningfully use `-A' more than once on a command line, since 5364the 960 architecture family allows combination of target architectures; 5365each use will add another pair of name variants to search for when `-l' 5366specifies a library. 5367 5368 `ld' supports the `--relax' option for the i960 family. If you 5369specify `--relax', `ld' finds all `balx' and `calx' instructions whose 5370targets are within 24 bits, and turns them into 24-bit program-counter 5371relative `bal' and `cal' instructions, respectively. `ld' also turns 5372`cal' instructions into `bal' instructions when it determines that the 5373target subroutine is a leaf routine (that is, the target subroutine does 5374not itself call any subroutines). 5375 5376 5377File: ld.info, Node: M68HC11/68HC12, Next: ARM, Prev: i960, Up: Machine Dependent 5378 53794.3 `ld' and the Motorola 68HC11 and 68HC12 families 5380==================================================== 5381 53824.3.1 Linker Relaxation 5383----------------------- 5384 5385For the Motorola 68HC11, `ld' can perform these global optimizations 5386when you specify the `--relax' command-line option. 5387 5388_relaxing address modes_ 5389 `ld' finds all `jsr' and `jmp' instructions whose targets are 5390 within eight bits, and turns them into eight-bit program-counter 5391 relative `bsr' and `bra' instructions, respectively. 5392 5393 `ld' also looks at all 16-bit extended addressing modes and 5394 transforms them in a direct addressing mode when the address is in 5395 page 0 (between 0 and 0x0ff). 5396 5397_relaxing gcc instruction group_ 5398 When `gcc' is called with `-mrelax', it can emit group of 5399 instructions that the linker can optimize to use a 68HC11 direct 5400 addressing mode. These instructions consists of `bclr' or `bset' 5401 instructions. 5402 5403 54044.3.2 Trampoline Generation 5405--------------------------- 5406 5407For 68HC11 and 68HC12, `ld' can generate trampoline code to call a far 5408function using a normal `jsr' instruction. The linker will also change 5409the relocation to some far function to use the trampoline address 5410instead of the function address. This is typically the case when a 5411pointer to a function is taken. The pointer will in fact point to the 5412function trampoline. 5413 5414 5415File: ld.info, Node: ARM, Next: HPPA ELF32, Prev: M68HC11/68HC12, Up: Machine Dependent 5416 54174.4 `ld' and the ARM family 5418=========================== 5419 5420For the ARM, `ld' will generate code stubs to allow functions calls 5421between ARM and Thumb code. These stubs only work with code that has 5422been compiled and assembled with the `-mthumb-interwork' command line 5423option. If it is necessary to link with old ARM object files or 5424libraries, which have not been compiled with the -mthumb-interwork 5425option then the `--support-old-code' command line switch should be 5426given to the linker. This will make it generate larger stub functions 5427which will work with non-interworking aware ARM code. Note, however, 5428the linker does not support generating stubs for function calls to 5429non-interworking aware Thumb code. 5430 5431 The `--thumb-entry' switch is a duplicate of the generic `--entry' 5432switch, in that it sets the program's starting address. But it also 5433sets the bottom bit of the address, so that it can be branched to using 5434a BX instruction, and the program will start executing in Thumb mode 5435straight away. 5436 5437 The `--use-nul-prefixed-import-tables' switch is specifying, that 5438the import tables idata4 and idata5 have to be generated with a zero 5439element prefix for import libraries. This is the old style to generate 5440import tables. By default this option is turned off. 5441 5442 The `--be8' switch instructs `ld' to generate BE8 format 5443executables. This option is only valid when linking big-endian objects 5444- ie ones which have been assembled with the `-EB' option. The 5445resulting image will contain big-endian data and little-endian code. 5446 5447 The `R_ARM_TARGET1' relocation is typically used for entries in the 5448`.init_array' section. It is interpreted as either `R_ARM_REL32' or 5449`R_ARM_ABS32', depending on the target. The `--target1-rel' and 5450`--target1-abs' switches override the default. 5451 5452 The `--target2=type' switch overrides the default definition of the 5453`R_ARM_TARGET2' relocation. Valid values for `type', their meanings, 5454and target defaults are as follows: 5455`rel' 5456 `R_ARM_REL32' (arm*-*-elf, arm*-*-eabi) 5457 5458`abs' 5459 `R_ARM_ABS32' (arm*-*-symbianelf) 5460 5461`got-rel' 5462 `R_ARM_GOT_PREL' (arm*-*-linux, arm*-*-*bsd) 5463 5464 The `R_ARM_V4BX' relocation (defined by the ARM AAELF specification) 5465enables objects compiled for the ARMv4 architecture to be 5466interworking-safe when linked with other objects compiled for ARMv4t, 5467but also allows pure ARMv4 binaries to be built from the same ARMv4 5468objects. 5469 5470 In the latter case, the switch `--fix-v4bx' must be passed to the 5471linker, which causes v4t `BX rM' instructions to be rewritten as `MOV 5472PC,rM', since v4 processors do not have a `BX' instruction. 5473 5474 In the former case, the switch should not be used, and `R_ARM_V4BX' 5475relocations are ignored. 5476 5477 Replace `BX rM' instructions identified by `R_ARM_V4BX' relocations 5478with a branch to the following veneer: 5479 5480 TST rM, #1 5481 MOVEQ PC, rM 5482 BX Rn 5483 5484 This allows generation of libraries/applications that work on ARMv4 5485cores and are still interworking safe. Note that the above veneer 5486clobbers the condition flags, so may cause incorrect program behavior 5487in rare cases. 5488 5489 The `--use-blx' switch enables the linker to use ARM/Thumb BLX 5490instructions (available on ARMv5t and above) in various situations. 5491Currently it is used to perform calls via the PLT from Thumb code using 5492BLX rather than using BX and a mode-switching stub before each PLT 5493entry. This should lead to such calls executing slightly faster. 5494 5495 This option is enabled implicitly for SymbianOS, so there is no need 5496to specify it if you are using that target. 5497 5498 The `--vfp11-denorm-fix' switch enables a link-time workaround for a 5499bug in certain VFP11 coprocessor hardware, which sometimes allows 5500instructions with denorm operands (which must be handled by support 5501code) to have those operands overwritten by subsequent instructions 5502before the support code can read the intended values. 5503 5504 The bug may be avoided in scalar mode if you allow at least one 5505intervening instruction between a VFP11 instruction which uses a 5506register and another instruction which writes to the same register, or 5507at least two intervening instructions if vector mode is in use. The bug 5508only affects full-compliance floating-point mode: you do not need this 5509workaround if you are using "runfast" mode. Please contact ARM for 5510further details. 5511 5512 If you know you are using buggy VFP11 hardware, you can enable this 5513workaround by specifying the linker option `--vfp-denorm-fix=scalar' if 5514you are using the VFP11 scalar mode only, or `--vfp-denorm-fix=vector' 5515if you are using vector mode (the latter also works for scalar code). 5516The default is `--vfp-denorm-fix=none'. 5517 5518 If the workaround is enabled, instructions are scanned for 5519potentially-troublesome sequences, and a veneer is created for each 5520such sequence which may trigger the erratum. The veneer consists of the 5521first instruction of the sequence and a branch back to the subsequent 5522instruction. The original instruction is then replaced with a branch to 5523the veneer. The extra cycles required to call and return from the veneer 5524are sufficient to avoid the erratum in both the scalar and vector cases. 5525 5526 The `--fix-arm1176' switch enables a link-time workaround for an 5527erratum in certain ARM1176 processors. The workaround is enabled by 5528default if you are targeting ARM v6 (excluding ARM v6T2) or earlier. 5529It can be disabled unconditionally by specifying `--no-fix-arm1176'. 5530 5531 Further information is available in the "ARM1176JZ-S and ARM1176JZF-S 5532Programmer Advice Notice" available on the ARM documentation website at: 5533http://infocenter.arm.com/. 5534 5535 The `--fix-stm32l4xx-629360' switch enables a link-time workaround 5536for a bug in the bus matrix / memory controller for some of the STM32 5537Cortex-M4 based products (STM32L4xx). When accessing off-chip memory 5538via the affected bus for bus reads of 9 words or more, the bus can 5539generate corrupt data and/or abort. These are only core-initiated 5540accesses (not DMA), and might affect any access: integer loads such as 5541LDM, POP and floating-point loads such as VLDM, VPOP. Stores are not 5542affected. 5543 5544 The bug can be avoided by splitting memory accesses into the 5545necessary chunks to keep bus reads below 8 words. 5546 5547 The workaround is not enabled by default, this is equivalent to use 5548`--fix-stm32l4xx-629360=none'. If you know you are using buggy 5549STM32L4xx hardware, you can enable the workaround by specifying the 5550linker option `--fix-stm32l4xx-629360', or the equivalent 5551`--fix-stm32l4xx-629360=default'. 5552 5553 If the workaround is enabled, instructions are scanned for 5554potentially-troublesome sequences, and a veneer is created for each 5555such sequence which may trigger the erratum. The veneer consists in a 5556replacement sequence emulating the behaviour of the original one and a 5557branch back to the subsequent instruction. The original instruction is 5558then replaced with a branch to the veneer. 5559 5560 The workaround does not always preserve the memory access order for 5561the LDMDB instruction, when the instruction loads the PC. 5562 5563 The workaround is not able to handle problematic instructions when 5564they are in the middle of an IT block, since a branch is not allowed 5565there. In that case, the linker reports a warning and no replacement 5566occurs. 5567 5568 The workaround is not able to replace problematic instructions with a 5569PC-relative branch instruction if the `.text' section is too large. In 5570that case, when the branch that replaces the original code cannot be 5571encoded, the linker reports a warning and no replacement occurs. 5572 5573 The `--no-enum-size-warning' switch prevents the linker from warning 5574when linking object files that specify incompatible EABI enumeration 5575size attributes. For example, with this switch enabled, linking of an 5576object file using 32-bit enumeration values with another using 5577enumeration values fitted into the smallest possible space will not be 5578diagnosed. 5579 5580 The `--no-wchar-size-warning' switch prevents the linker from 5581warning when linking object files that specify incompatible EABI 5582`wchar_t' size attributes. For example, with this switch enabled, 5583linking of an object file using 32-bit `wchar_t' values with another 5584using 16-bit `wchar_t' values will not be diagnosed. 5585 5586 The `--pic-veneer' switch makes the linker use PIC sequences for 5587ARM/Thumb interworking veneers, even if the rest of the binary is not 5588PIC. This avoids problems on uClinux targets where `--emit-relocs' is 5589used to generate relocatable binaries. 5590 5591 The linker will automatically generate and insert small sequences of 5592code into a linked ARM ELF executable whenever an attempt is made to 5593perform a function call to a symbol that is too far away. The 5594placement of these sequences of instructions - called stubs - is 5595controlled by the command line option `--stub-group-size=N'. The 5596placement is important because a poor choice can create a need for 5597duplicate stubs, increasing the code size. The linker will try to 5598group stubs together in order to reduce interruptions to the flow of 5599code, but it needs guidance as to how big these groups should be and 5600where they should be placed. 5601 5602 The value of `N', the parameter to the `--stub-group-size=' option 5603controls where the stub groups are placed. If it is negative then all 5604stubs are placed after the first branch that needs them. If it is 5605positive then the stubs can be placed either before or after the 5606branches that need them. If the value of `N' is 1 (either +1 or -1) 5607then the linker will choose exactly where to place groups of stubs, 5608using its built in heuristics. A value of `N' greater than 1 (or 5609smaller than -1) tells the linker that a single group of stubs can 5610service at most `N' bytes from the input sections. 5611 5612 The default, if `--stub-group-size=' is not specified, is `N = +1'. 5613 5614 Farcalls stubs insertion is fully supported for the ARM-EABI target 5615only, because it relies on object files properties not present 5616otherwise. 5617 5618 The `--fix-cortex-a8' switch enables a link-time workaround for an 5619erratum in certain Cortex-A8 processors. The workaround is enabled by 5620default if you are targeting the ARM v7-A architecture profile. It can 5621be enabled otherwise by specifying `--fix-cortex-a8', or disabled 5622unconditionally by specifying `--no-fix-cortex-a8'. 5623 5624 The erratum only affects Thumb-2 code. Please contact ARM for 5625further details. 5626 5627 The `--fix-cortex-a53-835769' switch enables a link-time workaround 5628for erratum 835769 present on certain early revisions of Cortex-A53 5629processors. The workaround is disabled by default. It can be enabled 5630by specifying `--fix-cortex-a53-835769', or disabled unconditionally by 5631specifying `--no-fix-cortex-a53-835769'. 5632 5633 Please contact ARM for further details. 5634 5635 The `--no-merge-exidx-entries' switch disables the merging of 5636adjacent exidx entries in debuginfo. 5637 5638 The `--long-plt' option enables the use of 16 byte PLT entries which 5639support up to 4Gb of code. The default is to use 12 byte PLT entries 5640which only support 512Mb of code. 5641 5642 5643File: ld.info, Node: HPPA ELF32, Next: M68K, Prev: ARM, Up: Machine Dependent 5644 56454.5 `ld' and HPPA 32-bit ELF Support 5646==================================== 5647 5648When generating a shared library, `ld' will by default generate import 5649stubs suitable for use with a single sub-space application. The 5650`--multi-subspace' switch causes `ld' to generate export stubs, and 5651different (larger) import stubs suitable for use with multiple 5652sub-spaces. 5653 5654 Long branch stubs and import/export stubs are placed by `ld' in stub 5655sections located between groups of input sections. `--stub-group-size' 5656specifies the maximum size of a group of input sections handled by one 5657stub section. Since branch offsets are signed, a stub section may 5658serve two groups of input sections, one group before the stub section, 5659and one group after it. However, when using conditional branches that 5660require stubs, it may be better (for branch prediction) that stub 5661sections only serve one group of input sections. A negative value for 5662`N' chooses this scheme, ensuring that branches to stubs always use a 5663negative offset. Two special values of `N' are recognized, `1' and 5664`-1'. These both instruct `ld' to automatically size input section 5665groups for the branch types detected, with the same behaviour regarding 5666stub placement as other positive or negative values of `N' respectively. 5667 5668 Note that `--stub-group-size' does not split input sections. A 5669single input section larger than the group size specified will of course 5670create a larger group (of one section). If input sections are too 5671large, it may not be possible for a branch to reach its stub. 5672 5673 5674File: ld.info, Node: M68K, Next: MIPS, Prev: HPPA ELF32, Up: Machine Dependent 5675 56764.6 `ld' and the Motorola 68K family 5677==================================== 5678 5679The `--got=TYPE' option lets you choose the GOT generation scheme. The 5680choices are `single', `negative', `multigot' and `target'. When 5681`target' is selected the linker chooses the default GOT generation 5682scheme for the current target. `single' tells the linker to generate a 5683single GOT with entries only at non-negative offsets. `negative' 5684instructs the linker to generate a single GOT with entries at both 5685negative and positive offsets. Not all environments support such GOTs. 5686`multigot' allows the linker to generate several GOTs in the output 5687file. All GOT references from a single input object file access the 5688same GOT, but references from different input object files might access 5689different GOTs. Not all environments support such GOTs. 5690 5691 5692File: ld.info, Node: MIPS, Next: MMIX, Prev: M68K, Up: Machine Dependent 5693 56944.7 `ld' and the MIPS family 5695============================ 5696 5697The `--insn32' and `--no-insn32' options control the choice of 5698microMIPS instructions used in code generated by the linker, such as 5699that in the PLT or lazy binding stubs, or in relaxation. If `--insn32' 5700is used, then the linker only uses 32-bit instruction encodings. By 5701default or if `--no-insn32' is used, all instruction encodings are used, 5702including 16-bit ones where possible. 5703 5704 5705File: ld.info, Node: MMIX, Next: MSP430, Prev: MIPS, Up: Machine Dependent 5706 57074.8 `ld' and MMIX 5708================= 5709 5710For MMIX, there is a choice of generating `ELF' object files or `mmo' 5711object files when linking. The simulator `mmix' understands the `mmo' 5712format. The binutils `objcopy' utility can translate between the two 5713formats. 5714 5715 There is one special section, the `.MMIX.reg_contents' section. 5716Contents in this section is assumed to correspond to that of global 5717registers, and symbols referring to it are translated to special 5718symbols, equal to registers. In a final link, the start address of the 5719`.MMIX.reg_contents' section corresponds to the first allocated global 5720register multiplied by 8. Register `$255' is not included in this 5721section; it is always set to the program entry, which is at the symbol 5722`Main' for `mmo' files. 5723 5724 Global symbols with the prefix `__.MMIX.start.', for example 5725`__.MMIX.start..text' and `__.MMIX.start..data' are special. The 5726default linker script uses these to set the default start address of a 5727section. 5728 5729 Initial and trailing multiples of zero-valued 32-bit words in a 5730section, are left out from an mmo file. 5731 5732 5733File: ld.info, Node: MSP430, Next: NDS32, Prev: MMIX, Up: Machine Dependent 5734 57354.9 `ld' and MSP430 5736=================== 5737 5738For the MSP430 it is possible to select the MPU architecture. The flag 5739`-m [mpu type]' will select an appropriate linker script for selected 5740MPU type. (To get a list of known MPUs just pass `-m help' option to 5741the linker). 5742 5743 The linker will recognize some extra sections which are MSP430 5744specific: 5745 5746``.vectors'' 5747 Defines a portion of ROM where interrupt vectors located. 5748 5749``.bootloader'' 5750 Defines the bootloader portion of the ROM (if applicable). Any 5751 code in this section will be uploaded to the MPU. 5752 5753``.infomem'' 5754 Defines an information memory section (if applicable). Any code in 5755 this section will be uploaded to the MPU. 5756 5757``.infomemnobits'' 5758 This is the same as the `.infomem' section except that any code in 5759 this section will not be uploaded to the MPU. 5760 5761``.noinit'' 5762 Denotes a portion of RAM located above `.bss' section. 5763 5764 The last two sections are used by gcc. 5765 5766 5767File: ld.info, Node: NDS32, Next: Nios II, Prev: MSP430, Up: Machine Dependent 5768 57694.10 `ld' and NDS32 5770=================== 5771 5772For NDS32, there are some options to select relaxation behavior. The 5773linker relaxes objects according to these options. 5774 5775``--m[no-]fp-as-gp'' 5776 Disable/enable fp-as-gp relaxation. 5777 5778``--mexport-symbols=FILE'' 5779 Exporting symbols and their address into FILE as linker script. 5780 5781``--m[no-]ex9'' 5782 Disable/enable link-time EX9 relaxation. 5783 5784``--mexport-ex9=FILE'' 5785 Export the EX9 table after linking. 5786 5787``--mimport-ex9=FILE'' 5788 Import the Ex9 table for EX9 relaxation. 5789 5790``--mupdate-ex9'' 5791 Update the existing EX9 table. 5792 5793``--mex9-limit=NUM'' 5794 Maximum number of entries in the ex9 table. 5795 5796``--mex9-loop-aware'' 5797 Avoid generating the EX9 instruction inside the loop. 5798 5799``--m[no-]ifc'' 5800 Disable/enable the link-time IFC optimization. 5801 5802``--mifc-loop-aware'' 5803 Avoid generating the IFC instruction inside the loop. 5804 5805 5806File: ld.info, Node: Nios II, Next: PowerPC ELF32, Prev: NDS32, Up: Machine Dependent 5807 58084.11 `ld' and the Altera Nios II 5809================================ 5810 5811Call and immediate jump instructions on Nios II processors are limited 5812to transferring control to addresses in the same 256MB memory segment, 5813which may result in `ld' giving `relocation truncated to fit' errors 5814with very large programs. The command-line option `--relax' enables 5815the generation of trampolines that can access the entire 32-bit address 5816space for calls outside the normal `call' and `jmpi' address range. 5817These trampolines are inserted at section boundaries, so may not 5818themselves be reachable if an input section and its associated call 5819trampolines are larger than 256MB. 5820 5821 The `--relax' option is enabled by default unless `-r' is also 5822specified. You can disable trampoline generation by using the 5823`--no-relax' linker option. You can also disable this optimization 5824locally by using the `set .noat' directive in assembly-language source 5825files, as the linker-inserted trampolines use the `at' register as a 5826temporary. 5827 5828 Note that the linker `--relax' option is independent of assembler 5829relaxation options, and that using the GNU assembler's `-relax-all' 5830option interferes with the linker's more selective call instruction 5831relaxation. 5832 5833 5834File: ld.info, Node: PowerPC ELF32, Next: PowerPC64 ELF64, Prev: Nios II, Up: Machine Dependent 5835 58364.12 `ld' and PowerPC 32-bit ELF Support 5837======================================== 5838 5839Branches on PowerPC processors are limited to a signed 26-bit 5840displacement, which may result in `ld' giving `relocation truncated to 5841fit' errors with very large programs. `--relax' enables the generation 5842of trampolines that can access the entire 32-bit address space. These 5843trampolines are inserted at section boundaries, so may not themselves 5844be reachable if an input section exceeds 33M in size. You may combine 5845`-r' and `--relax' to add trampolines in a partial link. In that case 5846both branches to undefined symbols and inter-section branches are also 5847considered potentially out of range, and trampolines inserted. 5848 5849`--bss-plt' 5850 Current PowerPC GCC accepts a `-msecure-plt' option that generates 5851 code capable of using a newer PLT and GOT layout that has the 5852 security advantage of no executable section ever needing to be 5853 writable and no writable section ever being executable. PowerPC 5854 `ld' will generate this layout, including stubs to access the PLT, 5855 if all input files (including startup and static libraries) were 5856 compiled with `-msecure-plt'. `--bss-plt' forces the old BSS PLT 5857 (and GOT layout) which can give slightly better performance. 5858 5859`--secure-plt' 5860 `ld' will use the new PLT and GOT layout if it is linking new 5861 `-fpic' or `-fPIC' code, but does not do so automatically when 5862 linking non-PIC code. This option requests the new PLT and GOT 5863 layout. A warning will be given if some object file requires the 5864 old style BSS PLT. 5865 5866`--sdata-got' 5867 The new secure PLT and GOT are placed differently relative to other 5868 sections compared to older BSS PLT and GOT placement. The 5869 location of `.plt' must change because the new secure PLT is an 5870 initialized section while the old PLT is uninitialized. The 5871 reason for the `.got' change is more subtle: The new placement 5872 allows `.got' to be read-only in applications linked with `-z 5873 relro -z now'. However, this placement means that `.sdata' cannot 5874 always be used in shared libraries, because the PowerPC ABI 5875 accesses `.sdata' in shared libraries from the GOT pointer. 5876 `--sdata-got' forces the old GOT placement. PowerPC GCC doesn't 5877 use `.sdata' in shared libraries, so this option is really only 5878 useful for other compilers that may do so. 5879 5880`--emit-stub-syms' 5881 This option causes `ld' to label linker stubs with a local symbol 5882 that encodes the stub type and destination. 5883 5884`--no-tls-optimize' 5885 PowerPC `ld' normally performs some optimization of code sequences 5886 used to access Thread-Local Storage. Use this option to disable 5887 the optimization. 5888 5889 5890File: ld.info, Node: PowerPC64 ELF64, Next: SPU ELF, Prev: PowerPC ELF32, Up: Machine Dependent 5891 58924.13 `ld' and PowerPC64 64-bit ELF Support 5893========================================== 5894 5895`--stub-group-size' 5896 Long branch stubs, PLT call stubs and TOC adjusting stubs are 5897 placed by `ld' in stub sections located between groups of input 5898 sections. `--stub-group-size' specifies the maximum size of a 5899 group of input sections handled by one stub section. Since branch 5900 offsets are signed, a stub section may serve two groups of input 5901 sections, one group before the stub section, and one group after 5902 it. However, when using conditional branches that require stubs, 5903 it may be better (for branch prediction) that stub sections only 5904 serve one group of input sections. A negative value for `N' 5905 chooses this scheme, ensuring that branches to stubs always use a 5906 negative offset. Two special values of `N' are recognized, `1' 5907 and `-1'. These both instruct `ld' to automatically size input 5908 section groups for the branch types detected, with the same 5909 behaviour regarding stub placement as other positive or negative 5910 values of `N' respectively. 5911 5912 Note that `--stub-group-size' does not split input sections. A 5913 single input section larger than the group size specified will of 5914 course create a larger group (of one section). If input sections 5915 are too large, it may not be possible for a branch to reach its 5916 stub. 5917 5918`--emit-stub-syms' 5919 This option causes `ld' to label linker stubs with a local symbol 5920 that encodes the stub type and destination. 5921 5922`--dotsyms' 5923`--no-dotsyms' 5924 These two options control how `ld' interprets version patterns in 5925 a version script. Older PowerPC64 compilers emitted both a 5926 function descriptor symbol with the same name as the function, and 5927 a code entry symbol with the name prefixed by a dot (`.'). To 5928 properly version a function `foo', the version script thus needs 5929 to control both `foo' and `.foo'. The option `--dotsyms', on by 5930 default, automatically adds the required dot-prefixed patterns. 5931 Use `--no-dotsyms' to disable this feature. 5932 5933`--save-restore-funcs' 5934`--no-save-restore-funcs' 5935 These two options control whether PowerPC64 `ld' automatically 5936 provides out-of-line register save and restore functions used by 5937 `-Os' code. The default is to provide any such referenced 5938 function for a normal final link, and to not do so for a 5939 relocatable link. 5940 5941`--no-tls-optimize' 5942 PowerPC64 `ld' normally performs some optimization of code 5943 sequences used to access Thread-Local Storage. Use this option to 5944 disable the optimization. 5945 5946`--tls-get-addr-optimize' 5947`--no-tls-get-addr-optimize' 5948 These options control whether PowerPC64 `ld' uses a special stub 5949 to call __tls_get_addr. PowerPC64 glibc 2.22 and later support an 5950 optimization that allows the second and subsequent calls to 5951 `__tls_get_addr' for a given symbol to be resolved by the special 5952 stub without calling in to glibc. By default the linker enables 5953 this option when glibc advertises the availability of 5954 __tls_get_addr_opt. Forcing this option on when using an older 5955 glibc won't do much besides slow down your applications, but may 5956 be useful if linking an application against an older glibc with 5957 the expectation that it will normally be used on systems having a 5958 newer glibc. 5959 5960`--no-opd-optimize' 5961 PowerPC64 `ld' normally removes `.opd' section entries 5962 corresponding to deleted link-once functions, or functions removed 5963 by the action of `--gc-sections' or linker script `/DISCARD/'. 5964 Use this option to disable `.opd' optimization. 5965 5966`--non-overlapping-opd' 5967 Some PowerPC64 compilers have an option to generate compressed 5968 `.opd' entries spaced 16 bytes apart, overlapping the third word, 5969 the static chain pointer (unused in C) with the first word of the 5970 next entry. This option expands such entries to the full 24 bytes. 5971 5972`--no-toc-optimize' 5973 PowerPC64 `ld' normally removes unused `.toc' section entries. 5974 Such entries are detected by examining relocations that reference 5975 the TOC in code sections. A reloc in a deleted code section marks 5976 a TOC word as unneeded, while a reloc in a kept code section marks 5977 a TOC word as needed. Since the TOC may reference itself, TOC 5978 relocs are also examined. TOC words marked as both needed and 5979 unneeded will of course be kept. TOC words without any referencing 5980 reloc are assumed to be part of a multi-word entry, and are kept or 5981 discarded as per the nearest marked preceding word. This works 5982 reliably for compiler generated code, but may be incorrect if 5983 assembly code is used to insert TOC entries. Use this option to 5984 disable the optimization. 5985 5986`--no-multi-toc' 5987 If given any toc option besides `-mcmodel=medium' or 5988 `-mcmodel=large', PowerPC64 GCC generates code for a TOC model 5989 where TOC entries are accessed with a 16-bit offset from r2. This 5990 limits the total TOC size to 64K. PowerPC64 `ld' extends this 5991 limit by grouping code sections such that each group uses less 5992 than 64K for its TOC entries, then inserts r2 adjusting stubs 5993 between inter-group calls. `ld' does not split apart input 5994 sections, so cannot help if a single input file has a `.toc' 5995 section that exceeds 64K, most likely from linking multiple files 5996 with `ld -r'. Use this option to turn off this feature. 5997 5998`--no-toc-sort' 5999 By default, `ld' sorts TOC sections so that those whose file 6000 happens to have a section called `.init' or `.fini' are placed 6001 first, followed by TOC sections referenced by code generated with 6002 PowerPC64 gcc's `-mcmodel=small', and lastly TOC sections 6003 referenced only by code generated with PowerPC64 gcc's 6004 `-mcmodel=medium' or `-mcmodel=large' options. Doing this results 6005 in better TOC grouping for multi-TOC. Use this option to turn off 6006 this feature. 6007 6008`--plt-align' 6009`--no-plt-align' 6010 Use these options to control whether individual PLT call stubs are 6011 padded so that they don't cross a 32-byte boundary, or to the 6012 specified power of two boundary when using `--plt-align='. Note 6013 that this isn't alignment in the usual sense. By default PLT call 6014 stubs are packed tightly. 6015 6016`--plt-static-chain' 6017`--no-plt-static-chain' 6018 Use these options to control whether PLT call stubs load the static 6019 chain pointer (r11). `ld' defaults to not loading the static 6020 chain since there is never any need to do so on a PLT call. 6021 6022`--plt-thread-safe' 6023`--no-thread-safe' 6024 With power7's weakly ordered memory model, it is possible when 6025 using lazy binding for ld.so to update a plt entry in one thread 6026 and have another thread see the individual plt entry words update 6027 in the wrong order, despite ld.so carefully writing in the correct 6028 order and using memory write barriers. To avoid this we need some 6029 sort of read barrier in the call stub, or use LD_BIND_NOW=1. By 6030 default, `ld' looks for calls to commonly used functions that 6031 create threads, and if seen, adds the necessary barriers. Use 6032 these options to change the default behaviour. 6033 6034 6035File: ld.info, Node: SPU ELF, Next: TI COFF, Prev: PowerPC64 ELF64, Up: Machine Dependent 6036 60374.14 `ld' and SPU ELF Support 6038============================= 6039 6040`--plugin' 6041 This option marks an executable as a PIC plugin module. 6042 6043`--no-overlays' 6044 Normally, `ld' recognizes calls to functions within overlay 6045 regions, and redirects such calls to an overlay manager via a stub. 6046 `ld' also provides a built-in overlay manager. This option turns 6047 off all this special overlay handling. 6048 6049`--emit-stub-syms' 6050 This option causes `ld' to label overlay stubs with a local symbol 6051 that encodes the stub type and destination. 6052 6053`--extra-overlay-stubs' 6054 This option causes `ld' to add overlay call stubs on all function 6055 calls out of overlay regions. Normally stubs are not added on 6056 calls to non-overlay regions. 6057 6058`--local-store=lo:hi' 6059 `ld' usually checks that a final executable for SPU fits in the 6060 address range 0 to 256k. This option may be used to change the 6061 range. Disable the check entirely with `--local-store=0:0'. 6062 6063`--stack-analysis' 6064 SPU local store space is limited. Over-allocation of stack space 6065 unnecessarily limits space available for code and data, while 6066 under-allocation results in runtime failures. If given this 6067 option, `ld' will provide an estimate of maximum stack usage. 6068 `ld' does this by examining symbols in code sections to determine 6069 the extents of functions, and looking at function prologues for 6070 stack adjusting instructions. A call-graph is created by looking 6071 for relocations on branch instructions. The graph is then searched 6072 for the maximum stack usage path. Note that this analysis does not 6073 find calls made via function pointers, and does not handle 6074 recursion and other cycles in the call graph. Stack usage may be 6075 under-estimated if your code makes such calls. Also, stack usage 6076 for dynamic allocation, e.g. alloca, will not be detected. If a 6077 link map is requested, detailed information about each function's 6078 stack usage and calls will be given. 6079 6080`--emit-stack-syms' 6081 This option, if given along with `--stack-analysis' will result in 6082 `ld' emitting stack sizing symbols for each function. These take 6083 the form `__stack_<function_name>' for global functions, and 6084 `__stack_<number>_<function_name>' for static functions. 6085 `<number>' is the section id in hex. The value of such symbols is 6086 the stack requirement for the corresponding function. The symbol 6087 size will be zero, type `STT_NOTYPE', binding `STB_LOCAL', and 6088 section `SHN_ABS'. 6089 6090 6091File: ld.info, Node: TI COFF, Next: WIN32, Prev: SPU ELF, Up: Machine Dependent 6092 60934.15 `ld''s Support for Various TI COFF Versions 6094================================================ 6095 6096The `--format' switch allows selection of one of the various TI COFF 6097versions. The latest of this writing is 2; versions 0 and 1 are also 6098supported. The TI COFF versions also vary in header byte-order format; 6099`ld' will read any version or byte order, but the output header format 6100depends on the default specified by the specific target. 6101 6102 6103File: ld.info, Node: WIN32, Next: Xtensa, Prev: TI COFF, Up: Machine Dependent 6104 61054.16 `ld' and WIN32 (cygwin/mingw) 6106================================== 6107 6108This section describes some of the win32 specific `ld' issues. See 6109*Note Command Line Options: Options. for detailed description of the 6110command line options mentioned here. 6111 6112_import libraries_ 6113 The standard Windows linker creates and uses so-called import 6114 libraries, which contains information for linking to dll's. They 6115 are regular static archives and are handled as any other static 6116 archive. The cygwin and mingw ports of `ld' have specific support 6117 for creating such libraries provided with the `--out-implib' 6118 command line option. 6119 6120_exporting DLL symbols_ 6121 The cygwin/mingw `ld' has several ways to export symbols for dll's. 6122 6123 _using auto-export functionality_ 6124 By default `ld' exports symbols with the auto-export 6125 functionality, which is controlled by the following command 6126 line options: 6127 6128 * -export-all-symbols [This is the default] 6129 6130 * -exclude-symbols 6131 6132 * -exclude-libs 6133 6134 * -exclude-modules-for-implib 6135 6136 * -version-script 6137 6138 When auto-export is in operation, `ld' will export all the 6139 non-local (global and common) symbols it finds in a DLL, with 6140 the exception of a few symbols known to belong to the 6141 system's runtime and libraries. As it will often not be 6142 desirable to export all of a DLL's symbols, which may include 6143 private functions that are not part of any public interface, 6144 the command-line options listed above may be used to filter 6145 symbols out from the list for exporting. The `--output-def' 6146 option can be used in order to see the final list of exported 6147 symbols with all exclusions taken into effect. 6148 6149 If `--export-all-symbols' is not given explicitly on the 6150 command line, then the default auto-export behavior will be 6151 _disabled_ if either of the following are true: 6152 6153 * A DEF file is used. 6154 6155 * Any symbol in any object file was marked with the 6156 __declspec(dllexport) attribute. 6157 6158 _using a DEF file_ 6159 Another way of exporting symbols is using a DEF file. A DEF 6160 file is an ASCII file containing definitions of symbols which 6161 should be exported when a dll is created. Usually it is 6162 named `<dll name>.def' and is added as any other object file 6163 to the linker's command line. The file's name must end in 6164 `.def' or `.DEF'. 6165 6166 gcc -o <output> <objectfiles> <dll name>.def 6167 6168 Using a DEF file turns off the normal auto-export behavior, 6169 unless the `--export-all-symbols' option is also used. 6170 6171 Here is an example of a DEF file for a shared library called 6172 `xyz.dll': 6173 6174 LIBRARY "xyz.dll" BASE=0x20000000 6175 6176 EXPORTS 6177 foo 6178 bar 6179 _bar = bar 6180 another_foo = abc.dll.afoo 6181 var1 DATA 6182 doo = foo == foo2 6183 eoo DATA == var1 6184 6185 This example defines a DLL with a non-default base address 6186 and seven symbols in the export table. The third exported 6187 symbol `_bar' is an alias for the second. The fourth symbol, 6188 `another_foo' is resolved by "forwarding" to another module 6189 and treating it as an alias for `afoo' exported from the DLL 6190 `abc.dll'. The final symbol `var1' is declared to be a data 6191 object. The `doo' symbol in export library is an alias of 6192 `foo', which gets the string name in export table `foo2'. The 6193 `eoo' symbol is an data export symbol, which gets in export 6194 table the name `var1'. 6195 6196 The optional `LIBRARY <name>' command indicates the _internal_ 6197 name of the output DLL. If `<name>' does not include a suffix, 6198 the default library suffix, `.DLL' is appended. 6199 6200 When the .DEF file is used to build an application, rather 6201 than a library, the `NAME <name>' command should be used 6202 instead of `LIBRARY'. If `<name>' does not include a suffix, 6203 the default executable suffix, `.EXE' is appended. 6204 6205 With either `LIBRARY <name>' or `NAME <name>' the optional 6206 specification `BASE = <number>' may be used to specify a 6207 non-default base address for the image. 6208 6209 If neither `LIBRARY <name>' nor `NAME <name>' is specified, 6210 or they specify an empty string, the internal name is the 6211 same as the filename specified on the command line. 6212 6213 The complete specification of an export symbol is: 6214 6215 EXPORTS 6216 ( ( ( <name1> [ = <name2> ] ) 6217 | ( <name1> = <module-name> . <external-name>)) 6218 [ @ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) * 6219 6220 Declares `<name1>' as an exported symbol from the DLL, or 6221 declares `<name1>' as an exported alias for `<name2>'; or 6222 declares `<name1>' as a "forward" alias for the symbol 6223 `<external-name>' in the DLL `<module-name>'. Optionally, 6224 the symbol may be exported by the specified ordinal 6225 `<integer>' alias. The optional `<name3>' is the to be used 6226 string in import/export table for the symbol. 6227 6228 The optional keywords that follow the declaration indicate: 6229 6230 `NONAME': Do not put the symbol name in the DLL's export 6231 table. It will still be exported by its ordinal alias 6232 (either the value specified by the .def specification or, 6233 otherwise, the value assigned by the linker). The symbol 6234 name, however, does remain visible in the import library (if 6235 any), unless `PRIVATE' is also specified. 6236 6237 `DATA': The symbol is a variable or object, rather than a 6238 function. The import lib will export only an indirect 6239 reference to `foo' as the symbol `_imp__foo' (ie, `foo' must 6240 be resolved as `*_imp__foo'). 6241 6242 `CONSTANT': Like `DATA', but put the undecorated `foo' as 6243 well as `_imp__foo' into the import library. Both refer to the 6244 read-only import address table's pointer to the variable, not 6245 to the variable itself. This can be dangerous. If the user 6246 code fails to add the `dllimport' attribute and also fails to 6247 explicitly add the extra indirection that the use of the 6248 attribute enforces, the application will behave unexpectedly. 6249 6250 `PRIVATE': Put the symbol in the DLL's export table, but do 6251 not put it into the static import library used to resolve 6252 imports at link time. The symbol can still be imported using 6253 the `LoadLibrary/GetProcAddress' API at runtime or by by 6254 using the GNU ld extension of linking directly to the DLL 6255 without an import library. 6256 6257 See ld/deffilep.y in the binutils sources for the full 6258 specification of other DEF file statements 6259 6260 While linking a shared dll, `ld' is able to create a DEF file 6261 with the `--output-def <file>' command line option. 6262 6263 _Using decorations_ 6264 Another way of marking symbols for export is to modify the 6265 source code itself, so that when building the DLL each symbol 6266 to be exported is declared as: 6267 6268 __declspec(dllexport) int a_variable 6269 __declspec(dllexport) void a_function(int with_args) 6270 6271 All such symbols will be exported from the DLL. If, however, 6272 any of the object files in the DLL contain symbols decorated 6273 in this way, then the normal auto-export behavior is 6274 disabled, unless the `--export-all-symbols' option is also 6275 used. 6276 6277 Note that object files that wish to access these symbols must 6278 _not_ decorate them with dllexport. Instead, they should use 6279 dllimport, instead: 6280 6281 __declspec(dllimport) int a_variable 6282 __declspec(dllimport) void a_function(int with_args) 6283 6284 This complicates the structure of library header files, 6285 because when included by the library itself the header must 6286 declare the variables and functions as dllexport, but when 6287 included by client code the header must declare them as 6288 dllimport. There are a number of idioms that are typically 6289 used to do this; often client code can omit the __declspec() 6290 declaration completely. See `--enable-auto-import' and 6291 `automatic data imports' for more information. 6292 6293_automatic data imports_ 6294 The standard Windows dll format supports data imports from dlls 6295 only by adding special decorations (dllimport/dllexport), which 6296 let the compiler produce specific assembler instructions to deal 6297 with this issue. This increases the effort necessary to port 6298 existing Un*x code to these platforms, especially for large c++ 6299 libraries and applications. The auto-import feature, which was 6300 initially provided by Paul Sokolovsky, allows one to omit the 6301 decorations to achieve a behavior that conforms to that on 6302 POSIX/Un*x platforms. This feature is enabled with the 6303 `--enable-auto-import' command-line option, although it is enabled 6304 by default on cygwin/mingw. The `--enable-auto-import' option 6305 itself now serves mainly to suppress any warnings that are 6306 ordinarily emitted when linked objects trigger the feature's use. 6307 6308 auto-import of variables does not always work flawlessly without 6309 additional assistance. Sometimes, you will see this message 6310 6311 "variable '<var>' can't be auto-imported. Please read the 6312 documentation for ld's `--enable-auto-import' for details." 6313 6314 The `--enable-auto-import' documentation explains why this error 6315 occurs, and several methods that can be used to overcome this 6316 difficulty. One of these methods is the _runtime pseudo-relocs_ 6317 feature, described below. 6318 6319 For complex variables imported from DLLs (such as structs or 6320 classes), object files typically contain a base address for the 6321 variable and an offset (_addend_) within the variable-to specify a 6322 particular field or public member, for instance. Unfortunately, 6323 the runtime loader used in win32 environments is incapable of 6324 fixing these references at runtime without the additional 6325 information supplied by dllimport/dllexport decorations. The 6326 standard auto-import feature described above is unable to resolve 6327 these references. 6328 6329 The `--enable-runtime-pseudo-relocs' switch allows these 6330 references to be resolved without error, while leaving the task of 6331 adjusting the references themselves (with their non-zero addends) 6332 to specialized code provided by the runtime environment. Recent 6333 versions of the cygwin and mingw environments and compilers 6334 provide this runtime support; older versions do not. However, the 6335 support is only necessary on the developer's platform; the 6336 compiled result will run without error on an older system. 6337 6338 `--enable-runtime-pseudo-relocs' is not the default; it must be 6339 explicitly enabled as needed. 6340 6341_direct linking to a dll_ 6342 The cygwin/mingw ports of `ld' support the direct linking, 6343 including data symbols, to a dll without the usage of any import 6344 libraries. This is much faster and uses much less memory than 6345 does the traditional import library method, especially when 6346 linking large libraries or applications. When `ld' creates an 6347 import lib, each function or variable exported from the dll is 6348 stored in its own bfd, even though a single bfd could contain many 6349 exports. The overhead involved in storing, loading, and 6350 processing so many bfd's is quite large, and explains the 6351 tremendous time, memory, and storage needed to link against 6352 particularly large or complex libraries when using import libs. 6353 6354 Linking directly to a dll uses no extra command-line switches 6355 other than `-L' and `-l', because `ld' already searches for a 6356 number of names to match each library. All that is needed from 6357 the developer's perspective is an understanding of this search, in 6358 order to force ld to select the dll instead of an import library. 6359 6360 For instance, when ld is called with the argument `-lxxx' it will 6361 attempt to find, in the first directory of its search path, 6362 6363 libxxx.dll.a 6364 xxx.dll.a 6365 libxxx.a 6366 xxx.lib 6367 cygxxx.dll (*) 6368 libxxx.dll 6369 xxx.dll 6370 6371 before moving on to the next directory in the search path. 6372 6373 (*) Actually, this is not `cygxxx.dll' but in fact is 6374 `<prefix>xxx.dll', where `<prefix>' is set by the `ld' option 6375 `--dll-search-prefix=<prefix>'. In the case of cygwin, the 6376 standard gcc spec file includes `--dll-search-prefix=cyg', so in 6377 effect we actually search for `cygxxx.dll'. 6378 6379 Other win32-based unix environments, such as mingw or pw32, may 6380 use other `<prefix>'es, although at present only cygwin makes use 6381 of this feature. It was originally intended to help avoid name 6382 conflicts among dll's built for the various win32/un*x 6383 environments, so that (for example) two versions of a zlib dll 6384 could coexist on the same machine. 6385 6386 The generic cygwin/mingw path layout uses a `bin' directory for 6387 applications and dll's and a `lib' directory for the import 6388 libraries (using cygwin nomenclature): 6389 6390 bin/ 6391 cygxxx.dll 6392 lib/ 6393 libxxx.dll.a (in case of dll's) 6394 libxxx.a (in case of static archive) 6395 6396 Linking directly to a dll without using the import library can be 6397 done two ways: 6398 6399 1. Use the dll directly by adding the `bin' path to the link line 6400 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx 6401 6402 However, as the dll's often have version numbers appended to their 6403 names (`cygncurses-5.dll') this will often fail, unless one 6404 specifies `-L../bin -lncurses-5' to include the version. Import 6405 libs are generally not versioned, and do not have this difficulty. 6406 6407 2. Create a symbolic link from the dll to a file in the `lib' 6408 directory according to the above mentioned search pattern. This 6409 should be used to avoid unwanted changes in the tools needed for 6410 making the app/dll. 6411 6412 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a] 6413 6414 Then you can link without any make environment changes. 6415 6416 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx 6417 6418 This technique also avoids the version number problems, because 6419 the following is perfectly legal 6420 6421 bin/ 6422 cygxxx-5.dll 6423 lib/ 6424 libxxx.dll.a -> ../bin/cygxxx-5.dll 6425 6426 Linking directly to a dll without using an import lib will work 6427 even when auto-import features are exercised, and even when 6428 `--enable-runtime-pseudo-relocs' is used. 6429 6430 Given the improvements in speed and memory usage, one might 6431 justifiably wonder why import libraries are used at all. There 6432 are three reasons: 6433 6434 1. Until recently, the link-directly-to-dll functionality did _not_ 6435 work with auto-imported data. 6436 6437 2. Sometimes it is necessary to include pure static objects within 6438 the import library (which otherwise contains only bfd's for 6439 indirection symbols that point to the exports of a dll). Again, 6440 the import lib for the cygwin kernel makes use of this ability, 6441 and it is not possible to do this without an import lib. 6442 6443 3. Symbol aliases can only be resolved using an import lib. This 6444 is critical when linking against OS-supplied dll's (eg, the win32 6445 API) in which symbols are usually exported as undecorated aliases 6446 of their stdcall-decorated assembly names. 6447 6448 So, import libs are not going away. But the ability to replace 6449 true import libs with a simple symbolic link to (or a copy of) a 6450 dll, in many cases, is a useful addition to the suite of tools 6451 binutils makes available to the win32 developer. Given the 6452 massive improvements in memory requirements during linking, storage 6453 requirements, and linking speed, we expect that many developers 6454 will soon begin to use this feature whenever possible. 6455 6456_symbol aliasing_ 6457 6458 _adding additional names_ 6459 Sometimes, it is useful to export symbols with additional 6460 names. A symbol `foo' will be exported as `foo', but it can 6461 also be exported as `_foo' by using special directives in the 6462 DEF file when creating the dll. This will affect also the 6463 optional created import library. Consider the following DEF 6464 file: 6465 6466 LIBRARY "xyz.dll" BASE=0x61000000 6467 6468 EXPORTS 6469 foo 6470 _foo = foo 6471 6472 The line `_foo = foo' maps the symbol `foo' to `_foo'. 6473 6474 Another method for creating a symbol alias is to create it in 6475 the source code using the "weak" attribute: 6476 6477 void foo () { /* Do something. */; } 6478 void _foo () __attribute__ ((weak, alias ("foo"))); 6479 6480 See the gcc manual for more information about attributes and 6481 weak symbols. 6482 6483 _renaming symbols_ 6484 Sometimes it is useful to rename exports. For instance, the 6485 cygwin kernel does this regularly. A symbol `_foo' can be 6486 exported as `foo' but not as `_foo' by using special 6487 directives in the DEF file. (This will also affect the import 6488 library, if it is created). In the following example: 6489 6490 LIBRARY "xyz.dll" BASE=0x61000000 6491 6492 EXPORTS 6493 _foo = foo 6494 6495 The line `_foo = foo' maps the exported symbol `foo' to 6496 `_foo'. 6497 6498 Note: using a DEF file disables the default auto-export behavior, 6499 unless the `--export-all-symbols' command line option is used. 6500 If, however, you are trying to rename symbols, then you should list 6501 _all_ desired exports in the DEF file, including the symbols that 6502 are not being renamed, and do _not_ use the `--export-all-symbols' 6503 option. If you list only the renamed symbols in the DEF file, and 6504 use `--export-all-symbols' to handle the other symbols, then the 6505 both the new names _and_ the original names for the renamed 6506 symbols will be exported. In effect, you'd be aliasing those 6507 symbols, not renaming them, which is probably not what you wanted. 6508 6509_weak externals_ 6510 The Windows object format, PE, specifies a form of weak symbols 6511 called weak externals. When a weak symbol is linked and the 6512 symbol is not defined, the weak symbol becomes an alias for some 6513 other symbol. There are three variants of weak externals: 6514 * Definition is searched for in objects and libraries, 6515 historically called lazy externals. 6516 6517 * Definition is searched for only in other objects, not in 6518 libraries. This form is not presently implemented. 6519 6520 * No search; the symbol is an alias. This form is not presently 6521 implemented. 6522 As a GNU extension, weak symbols that do not specify an alternate 6523 symbol are supported. If the symbol is undefined when linking, 6524 the symbol uses a default value. 6525 6526_aligned common symbols_ 6527 As a GNU extension to the PE file format, it is possible to 6528 specify the desired alignment for a common symbol. This 6529 information is conveyed from the assembler or compiler to the 6530 linker by means of GNU-specific commands carried in the object 6531 file's `.drectve' section, which are recognized by `ld' and 6532 respected when laying out the common symbols. Native tools will 6533 be able to process object files employing this GNU extension, but 6534 will fail to respect the alignment instructions, and may issue 6535 noisy warnings about unknown linker directives. 6536 6537 6538 6539File: ld.info, Node: Xtensa, Prev: WIN32, Up: Machine Dependent 6540 65414.17 `ld' and Xtensa Processors 6542=============================== 6543 6544The default `ld' behavior for Xtensa processors is to interpret 6545`SECTIONS' commands so that lists of explicitly named sections in a 6546specification with a wildcard file will be interleaved when necessary to 6547keep literal pools within the range of PC-relative load offsets. For 6548example, with the command: 6549 6550 SECTIONS 6551 { 6552 .text : { 6553 *(.literal .text) 6554 } 6555 } 6556 6557`ld' may interleave some of the `.literal' and `.text' sections from 6558different object files to ensure that the literal pools are within the 6559range of PC-relative load offsets. A valid interleaving might place 6560the `.literal' sections from an initial group of files followed by the 6561`.text' sections of that group of files. Then, the `.literal' sections 6562from the rest of the files and the `.text' sections from the rest of 6563the files would follow. 6564 6565 Relaxation is enabled by default for the Xtensa version of `ld' and 6566provides two important link-time optimizations. The first optimization 6567is to combine identical literal values to reduce code size. A redundant 6568literal will be removed and all the `L32R' instructions that use it 6569will be changed to reference an identical literal, as long as the 6570location of the replacement literal is within the offset range of all 6571the `L32R' instructions. The second optimization is to remove 6572unnecessary overhead from assembler-generated "longcall" sequences of 6573`L32R'/`CALLXN' when the target functions are within range of direct 6574`CALLN' instructions. 6575 6576 For each of these cases where an indirect call sequence can be 6577optimized to a direct call, the linker will change the `CALLXN' 6578instruction to a `CALLN' instruction, remove the `L32R' instruction, 6579and remove the literal referenced by the `L32R' instruction if it is 6580not used for anything else. Removing the `L32R' instruction always 6581reduces code size but can potentially hurt performance by changing the 6582alignment of subsequent branch targets. By default, the linker will 6583always preserve alignments, either by switching some instructions 6584between 24-bit encodings and the equivalent density instructions or by 6585inserting a no-op in place of the `L32R' instruction that was removed. 6586If code size is more important than performance, the `--size-opt' 6587option can be used to prevent the linker from widening density 6588instructions or inserting no-ops, except in a few cases where no-ops 6589are required for correctness. 6590 6591 The following Xtensa-specific command-line options can be used to 6592control the linker: 6593 6594`--size-opt' 6595 When optimizing indirect calls to direct calls, optimize for code 6596 size more than performance. With this option, the linker will not 6597 insert no-ops or widen density instructions to preserve branch 6598 target alignment. There may still be some cases where no-ops are 6599 required to preserve the correctness of the code. 6600 6601 6602File: ld.info, Node: BFD, Next: Reporting Bugs, Prev: Machine Dependent, Up: Top 6603 66045 BFD 6605***** 6606 6607The linker accesses object and archive files using the BFD libraries. 6608These libraries allow the linker to use the same routines to operate on 6609object files whatever the object file format. A different object file 6610format can be supported simply by creating a new BFD back end and adding 6611it to the library. To conserve runtime memory, however, the linker and 6612associated tools are usually configured to support only a subset of the 6613object file formats available. You can use `objdump -i' (*note 6614objdump: (binutils.info)objdump.) to list all the formats available for 6615your configuration. 6616 6617 As with most implementations, BFD is a compromise between several 6618conflicting requirements. The major factor influencing BFD design was 6619efficiency: any time used converting between formats is time which 6620would not have been spent had BFD not been involved. This is partly 6621offset by abstraction payback; since BFD simplifies applications and 6622back ends, more time and care may be spent optimizing algorithms for a 6623greater speed. 6624 6625 One minor artifact of the BFD solution which you should bear in mind 6626is the potential for information loss. There are two places where 6627useful information can be lost using the BFD mechanism: during 6628conversion and during output. *Note BFD information loss::. 6629 6630* Menu: 6631 6632* BFD outline:: How it works: an outline of BFD 6633 6634 6635File: ld.info, Node: BFD outline, Up: BFD 6636 66375.1 How It Works: An Outline of BFD 6638=================================== 6639 6640When an object file is opened, BFD subroutines automatically determine 6641the format of the input object file. They then build a descriptor in 6642memory with pointers to routines that will be used to access elements of 6643the object file's data structures. 6644 6645 As different information from the object files is required, BFD 6646reads from different sections of the file and processes them. For 6647example, a very common operation for the linker is processing symbol 6648tables. Each BFD back end provides a routine for converting between 6649the object file's representation of symbols and an internal canonical 6650format. When the linker asks for the symbol table of an object file, it 6651calls through a memory pointer to the routine from the relevant BFD 6652back end which reads and converts the table into a canonical form. The 6653linker then operates upon the canonical form. When the link is finished 6654and the linker writes the output file's symbol table, another BFD back 6655end routine is called to take the newly created symbol table and 6656convert it into the chosen output format. 6657 6658* Menu: 6659 6660* BFD information loss:: Information Loss 6661* Canonical format:: The BFD canonical object-file format 6662 6663 6664File: ld.info, Node: BFD information loss, Next: Canonical format, Up: BFD outline 6665 66665.1.1 Information Loss 6667---------------------- 6668 6669_Information can be lost during output._ The output formats supported 6670by BFD do not provide identical facilities, and information which can 6671be described in one form has nowhere to go in another format. One 6672example of this is alignment information in `b.out'. There is nowhere 6673in an `a.out' format file to store alignment information on the 6674contained data, so when a file is linked from `b.out' and an `a.out' 6675image is produced, alignment information will not propagate to the 6676output file. (The linker will still use the alignment information 6677internally, so the link is performed correctly). 6678 6679 Another example is COFF section names. COFF files may contain an 6680unlimited number of sections, each one with a textual section name. If 6681the target of the link is a format which does not have many sections 6682(e.g., `a.out') or has sections without names (e.g., the Oasys format), 6683the link cannot be done simply. You can circumvent this problem by 6684describing the desired input-to-output section mapping with the linker 6685command language. 6686 6687 _Information can be lost during canonicalization._ The BFD internal 6688canonical form of the external formats is not exhaustive; there are 6689structures in input formats for which there is no direct representation 6690internally. This means that the BFD back ends cannot maintain all 6691possible data richness through the transformation between external to 6692internal and back to external formats. 6693 6694 This limitation is only a problem when an application reads one 6695format and writes another. Each BFD back end is responsible for 6696maintaining as much data as possible, and the internal BFD canonical 6697form has structures which are opaque to the BFD core, and exported only 6698to the back ends. When a file is read in one format, the canonical form 6699is generated for BFD and the application. At the same time, the back 6700end saves away any information which may otherwise be lost. If the data 6701is then written back in the same format, the back end routine will be 6702able to use the canonical form provided by the BFD core as well as the 6703information it prepared earlier. Since there is a great deal of 6704commonality between back ends, there is no information lost when 6705linking or copying big endian COFF to little endian COFF, or `a.out' to 6706`b.out'. When a mixture of formats is linked, the information is only 6707lost from the files whose format differs from the destination. 6708 6709 6710File: ld.info, Node: Canonical format, Prev: BFD information loss, Up: BFD outline 6711 67125.1.2 The BFD canonical object-file format 6713------------------------------------------ 6714 6715The greatest potential for loss of information occurs when there is the 6716least overlap between the information provided by the source format, 6717that stored by the canonical format, and that needed by the destination 6718format. A brief description of the canonical form may help you 6719understand which kinds of data you can count on preserving across 6720conversions. 6721 6722_files_ 6723 Information stored on a per-file basis includes target machine 6724 architecture, particular implementation format type, a demand 6725 pageable bit, and a write protected bit. Information like Unix 6726 magic numbers is not stored here--only the magic numbers' meaning, 6727 so a `ZMAGIC' file would have both the demand pageable bit and the 6728 write protected text bit set. The byte order of the target is 6729 stored on a per-file basis, so that big- and little-endian object 6730 files may be used with one another. 6731 6732_sections_ 6733 Each section in the input file contains the name of the section, 6734 the section's original address in the object file, size and 6735 alignment information, various flags, and pointers into other BFD 6736 data structures. 6737 6738_symbols_ 6739 Each symbol contains a pointer to the information for the object 6740 file which originally defined it, its name, its value, and various 6741 flag bits. When a BFD back end reads in a symbol table, it 6742 relocates all symbols to make them relative to the base of the 6743 section where they were defined. Doing this ensures that each 6744 symbol points to its containing section. Each symbol also has a 6745 varying amount of hidden private data for the BFD back end. Since 6746 the symbol points to the original file, the private data format 6747 for that symbol is accessible. `ld' can operate on a collection 6748 of symbols of wildly different formats without problems. 6749 6750 Normal global and simple local symbols are maintained on output, 6751 so an output file (no matter its format) will retain symbols 6752 pointing to functions and to global, static, and common variables. 6753 Some symbol information is not worth retaining; in `a.out', type 6754 information is stored in the symbol table as long symbol names. 6755 This information would be useless to most COFF debuggers; the 6756 linker has command line switches to allow users to throw it away. 6757 6758 There is one word of type information within the symbol, so if the 6759 format supports symbol type information within symbols (for 6760 example, COFF, IEEE, Oasys) and the type is simple enough to fit 6761 within one word (nearly everything but aggregates), the 6762 information will be preserved. 6763 6764_relocation level_ 6765 Each canonical BFD relocation record contains a pointer to the 6766 symbol to relocate to, the offset of the data to relocate, the 6767 section the data is in, and a pointer to a relocation type 6768 descriptor. Relocation is performed by passing messages through 6769 the relocation type descriptor and the symbol pointer. Therefore, 6770 relocations can be performed on output data using a relocation 6771 method that is only available in one of the input formats. For 6772 instance, Oasys provides a byte relocation format. A relocation 6773 record requesting this relocation type would point indirectly to a 6774 routine to perform this, so the relocation may be performed on a 6775 byte being written to a 68k COFF file, even though 68k COFF has no 6776 such relocation type. 6777 6778_line numbers_ 6779 Object formats can contain, for debugging purposes, some form of 6780 mapping between symbols, source line numbers, and addresses in the 6781 output file. These addresses have to be relocated along with the 6782 symbol information. Each symbol with an associated list of line 6783 number records points to the first record of the list. The head 6784 of a line number list consists of a pointer to the symbol, which 6785 allows finding out the address of the function whose line number 6786 is being described. The rest of the list is made up of pairs: 6787 offsets into the section and line numbers. Any format which can 6788 simply derive this information can pass it successfully between 6789 formats (COFF, IEEE and Oasys). 6790 6791 6792File: ld.info, Node: Reporting Bugs, Next: MRI, Prev: BFD, Up: Top 6793 67946 Reporting Bugs 6795**************** 6796 6797Your bug reports play an essential role in making `ld' reliable. 6798 6799 Reporting a bug may help you by bringing a solution to your problem, 6800or it may not. But in any case the principal function of a bug report 6801is to help the entire community by making the next version of `ld' work 6802better. Bug reports are your contribution to the maintenance of `ld'. 6803 6804 In order for a bug report to serve its purpose, you must include the 6805information that enables us to fix the bug. 6806 6807* Menu: 6808 6809* Bug Criteria:: Have you found a bug? 6810* Bug Reporting:: How to report bugs 6811 6812 6813File: ld.info, Node: Bug Criteria, Next: Bug Reporting, Up: Reporting Bugs 6814 68156.1 Have You Found a Bug? 6816========================= 6817 6818If you are not sure whether you have found a bug, here are some 6819guidelines: 6820 6821 * If the linker gets a fatal signal, for any input whatever, that is 6822 a `ld' bug. Reliable linkers never crash. 6823 6824 * If `ld' produces an error message for valid input, that is a bug. 6825 6826 * If `ld' does not produce an error message for invalid input, that 6827 may be a bug. In the general case, the linker can not verify that 6828 object files are correct. 6829 6830 * If you are an experienced user of linkers, your suggestions for 6831 improvement of `ld' are welcome in any case. 6832 6833 6834File: ld.info, Node: Bug Reporting, Prev: Bug Criteria, Up: Reporting Bugs 6835 68366.2 How to Report Bugs 6837====================== 6838 6839A number of companies and individuals offer support for GNU products. 6840If you obtained `ld' from a support organization, we recommend you 6841contact that organization first. 6842 6843 You can find contact information for many support companies and 6844individuals in the file `etc/SERVICE' in the GNU Emacs distribution. 6845 6846 Otherwise, send bug reports for `ld' to 6847`http://www.sourceware.org/bugzilla/'. 6848 6849 The fundamental principle of reporting bugs usefully is this: 6850*report all the facts*. If you are not sure whether to state a fact or 6851leave it out, state it! 6852 6853 Often people omit facts because they think they know what causes the 6854problem and assume that some details do not matter. Thus, you might 6855assume that the name of a symbol you use in an example does not matter. 6856Well, probably it does not, but one cannot be sure. Perhaps the bug 6857is a stray memory reference which happens to fetch from the location 6858where that name is stored in memory; perhaps, if the name were 6859different, the contents of that location would fool the linker into 6860doing the right thing despite the bug. Play it safe and give a 6861specific, complete example. That is the easiest thing for you to do, 6862and the most helpful. 6863 6864 Keep in mind that the purpose of a bug report is to enable us to fix 6865the bug if it is new to us. Therefore, always write your bug reports 6866on the assumption that the bug has not been reported previously. 6867 6868 Sometimes people give a few sketchy facts and ask, "Does this ring a 6869bell?" This cannot help us fix a bug, so it is basically useless. We 6870respond by asking for enough details to enable us to investigate. You 6871might as well expedite matters by sending them to begin with. 6872 6873 To enable us to fix the bug, you should include all these things: 6874 6875 * The version of `ld'. `ld' announces it if you start it with the 6876 `--version' argument. 6877 6878 Without this, we will not know whether there is any point in 6879 looking for the bug in the current version of `ld'. 6880 6881 * Any patches you may have applied to the `ld' source, including any 6882 patches made to the `BFD' library. 6883 6884 * The type of machine you are using, and the operating system name 6885 and version number. 6886 6887 * What compiler (and its version) was used to compile `ld'--e.g. 6888 "`gcc-2.7'". 6889 6890 * The command arguments you gave the linker to link your example and 6891 observe the bug. To guarantee you will not omit something 6892 important, list them all. A copy of the Makefile (or the output 6893 from make) is sufficient. 6894 6895 If we were to try to guess the arguments, we would probably guess 6896 wrong and then we might not encounter the bug. 6897 6898 * A complete input file, or set of input files, that will reproduce 6899 the bug. It is generally most helpful to send the actual object 6900 files provided that they are reasonably small. Say no more than 6901 10K. For bigger files you can either make them available by FTP 6902 or HTTP or else state that you are willing to send the object 6903 file(s) to whomever requests them. (Note - your email will be 6904 going to a mailing list, so we do not want to clog it up with 6905 large attachments). But small attachments are best. 6906 6907 If the source files were assembled using `gas' or compiled using 6908 `gcc', then it may be OK to send the source files rather than the 6909 object files. In this case, be sure to say exactly what version of 6910 `gas' or `gcc' was used to produce the object files. Also say how 6911 `gas' or `gcc' were configured. 6912 6913 * A description of what behavior you observe that you believe is 6914 incorrect. For example, "It gets a fatal signal." 6915 6916 Of course, if the bug is that `ld' gets a fatal signal, then we 6917 will certainly notice it. But if the bug is incorrect output, we 6918 might not notice unless it is glaringly wrong. You might as well 6919 not give us a chance to make a mistake. 6920 6921 Even if the problem you experience is a fatal signal, you should 6922 still say so explicitly. Suppose something strange is going on, 6923 such as, your copy of `ld' is out of sync, or you have encountered 6924 a bug in the C library on your system. (This has happened!) Your 6925 copy might crash and ours would not. If you told us to expect a 6926 crash, then when ours fails to crash, we would know that the bug 6927 was not happening for us. If you had not told us to expect a 6928 crash, then we would not be able to draw any conclusion from our 6929 observations. 6930 6931 * If you wish to suggest changes to the `ld' source, send us context 6932 diffs, as generated by `diff' with the `-u', `-c', or `-p' option. 6933 Always send diffs from the old file to the new file. If you even 6934 discuss something in the `ld' source, refer to it by context, not 6935 by line number. 6936 6937 The line numbers in our development sources will not match those 6938 in your sources. Your line numbers would convey no useful 6939 information to us. 6940 6941 Here are some things that are not necessary: 6942 6943 * A description of the envelope of the bug. 6944 6945 Often people who encounter a bug spend a lot of time investigating 6946 which changes to the input file will make the bug go away and which 6947 changes will not affect it. 6948 6949 This is often time consuming and not very useful, because the way 6950 we will find the bug is by running a single example under the 6951 debugger with breakpoints, not by pure deduction from a series of 6952 examples. We recommend that you save your time for something else. 6953 6954 Of course, if you can find a simpler example to report _instead_ 6955 of the original one, that is a convenience for us. Errors in the 6956 output will be easier to spot, running under the debugger will take 6957 less time, and so on. 6958 6959 However, simplification is not vital; if you do not want to do 6960 this, report the bug anyway and send us the entire test case you 6961 used. 6962 6963 * A patch for the bug. 6964 6965 A patch for the bug does help us if it is a good one. But do not 6966 omit the necessary information, such as the test case, on the 6967 assumption that a patch is all we need. We might see problems 6968 with your patch and decide to fix the problem another way, or we 6969 might not understand it at all. 6970 6971 Sometimes with a program as complicated as `ld' it is very hard to 6972 construct an example that will make the program follow a certain 6973 path through the code. If you do not send us the example, we will 6974 not be able to construct one, so we will not be able to verify 6975 that the bug is fixed. 6976 6977 And if we cannot understand what bug you are trying to fix, or why 6978 your patch should be an improvement, we will not install it. A 6979 test case will help us to understand. 6980 6981 * A guess about what the bug is or what it depends on. 6982 6983 Such guesses are usually wrong. Even we cannot guess right about 6984 such things without first using the debugger to find the facts. 6985 6986 6987File: ld.info, Node: MRI, Next: GNU Free Documentation License, Prev: Reporting Bugs, Up: Top 6988 6989Appendix A MRI Compatible Script Files 6990************************************** 6991 6992To aid users making the transition to GNU `ld' from the MRI linker, 6993`ld' can use MRI compatible linker scripts as an alternative to the 6994more general-purpose linker scripting language described in *Note 6995Scripts::. MRI compatible linker scripts have a much simpler command 6996set than the scripting language otherwise used with `ld'. GNU `ld' 6997supports the most commonly used MRI linker commands; these commands are 6998described here. 6999 7000 In general, MRI scripts aren't of much use with the `a.out' object 7001file format, since it only has three sections and MRI scripts lack some 7002features to make use of them. 7003 7004 You can specify a file containing an MRI-compatible script using the 7005`-c' command-line option. 7006 7007 Each command in an MRI-compatible script occupies its own line; each 7008command line starts with the keyword that identifies the command (though 7009blank lines are also allowed for punctuation). If a line of an 7010MRI-compatible script begins with an unrecognized keyword, `ld' issues 7011a warning message, but continues processing the script. 7012 7013 Lines beginning with `*' are comments. 7014 7015 You can write these commands using all upper-case letters, or all 7016lower case; for example, `chip' is the same as `CHIP'. The following 7017list shows only the upper-case form of each command. 7018 7019`ABSOLUTE SECNAME' 7020`ABSOLUTE SECNAME, SECNAME, ... SECNAME' 7021 Normally, `ld' includes in the output file all sections from all 7022 the input files. However, in an MRI-compatible script, you can 7023 use the `ABSOLUTE' command to restrict the sections that will be 7024 present in your output program. If the `ABSOLUTE' command is used 7025 at all in a script, then only the sections named explicitly in 7026 `ABSOLUTE' commands will appear in the linker output. You can 7027 still use other input sections (whatever you select on the command 7028 line, or using `LOAD') to resolve addresses in the output file. 7029 7030`ALIAS OUT-SECNAME, IN-SECNAME' 7031 Use this command to place the data from input section IN-SECNAME 7032 in a section called OUT-SECNAME in the linker output file. 7033 7034 IN-SECNAME may be an integer. 7035 7036`ALIGN SECNAME = EXPRESSION' 7037 Align the section called SECNAME to EXPRESSION. The EXPRESSION 7038 should be a power of two. 7039 7040`BASE EXPRESSION' 7041 Use the value of EXPRESSION as the lowest address (other than 7042 absolute addresses) in the output file. 7043 7044`CHIP EXPRESSION' 7045`CHIP EXPRESSION, EXPRESSION' 7046 This command does nothing; it is accepted only for compatibility. 7047 7048`END' 7049 This command does nothing whatever; it's only accepted for 7050 compatibility. 7051 7052`FORMAT OUTPUT-FORMAT' 7053 Similar to the `OUTPUT_FORMAT' command in the more general linker 7054 language, but restricted to one of these output formats: 7055 7056 1. S-records, if OUTPUT-FORMAT is `S' 7057 7058 2. IEEE, if OUTPUT-FORMAT is `IEEE' 7059 7060 3. COFF (the `coff-m68k' variant in BFD), if OUTPUT-FORMAT is 7061 `COFF' 7062 7063`LIST ANYTHING...' 7064 Print (to the standard output file) a link map, as produced by the 7065 `ld' command-line option `-M'. 7066 7067 The keyword `LIST' may be followed by anything on the same line, 7068 with no change in its effect. 7069 7070`LOAD FILENAME' 7071`LOAD FILENAME, FILENAME, ... FILENAME' 7072 Include one or more object file FILENAME in the link; this has the 7073 same effect as specifying FILENAME directly on the `ld' command 7074 line. 7075 7076`NAME OUTPUT-NAME' 7077 OUTPUT-NAME is the name for the program produced by `ld'; the 7078 MRI-compatible command `NAME' is equivalent to the command-line 7079 option `-o' or the general script language command `OUTPUT'. 7080 7081`ORDER SECNAME, SECNAME, ... SECNAME' 7082`ORDER SECNAME SECNAME SECNAME' 7083 Normally, `ld' orders the sections in its output file in the order 7084 in which they first appear in the input files. In an 7085 MRI-compatible script, you can override this ordering with the 7086 `ORDER' command. The sections you list with `ORDER' will appear 7087 first in your output file, in the order specified. 7088 7089`PUBLIC NAME=EXPRESSION' 7090`PUBLIC NAME,EXPRESSION' 7091`PUBLIC NAME EXPRESSION' 7092 Supply a value (EXPRESSION) for external symbol NAME used in the 7093 linker input files. 7094 7095`SECT SECNAME, EXPRESSION' 7096`SECT SECNAME=EXPRESSION' 7097`SECT SECNAME EXPRESSION' 7098 You can use any of these three forms of the `SECT' command to 7099 specify the start address (EXPRESSION) for section SECNAME. If 7100 you have more than one `SECT' statement for the same SECNAME, only 7101 the _first_ sets the start address. 7102 7103 7104File: ld.info, Node: GNU Free Documentation License, Next: LD Index, Prev: MRI, Up: Top 7105 7106Appendix B GNU Free Documentation License 7107***************************************** 7108 7109 Version 1.3, 3 November 2008 7110 7111 Copyright (C) 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc. 7112 `http://fsf.org/' 7113 7114 Everyone is permitted to copy and distribute verbatim copies 7115 of this license document, but changing it is not allowed. 7116 7117 0. PREAMBLE 7118 7119 The purpose of this License is to make a manual, textbook, or other 7120 functional and useful document "free" in the sense of freedom: to 7121 assure everyone the effective freedom to copy and redistribute it, 7122 with or without modifying it, either commercially or 7123 noncommercially. Secondarily, this License preserves for the 7124 author and publisher a way to get credit for their work, while not 7125 being considered responsible for modifications made by others. 7126 7127 This License is a kind of "copyleft", which means that derivative 7128 works of the document must themselves be free in the same sense. 7129 It complements the GNU General Public License, which is a copyleft 7130 license designed for free software. 7131 7132 We have designed this License in order to use it for manuals for 7133 free software, because free software needs free documentation: a 7134 free program should come with manuals providing the same freedoms 7135 that the software does. But this License is not limited to 7136 software manuals; it can be used for any textual work, regardless 7137 of subject matter or whether it is published as a printed book. 7138 We recommend this License principally for works whose purpose is 7139 instruction or reference. 7140 7141 1. APPLICABILITY AND DEFINITIONS 7142 7143 This License applies to any manual or other work, in any medium, 7144 that contains a notice placed by the copyright holder saying it 7145 can be distributed under the terms of this License. Such a notice 7146 grants a world-wide, royalty-free license, unlimited in duration, 7147 to use that work under the conditions stated herein. The 7148 "Document", below, refers to any such manual or work. Any member 7149 of the public is a licensee, and is addressed as "you". You 7150 accept the license if you copy, modify or distribute the work in a 7151 way requiring permission under copyright law. 7152 7153 A "Modified Version" of the Document means any work containing the 7154 Document or a portion of it, either copied verbatim, or with 7155 modifications and/or translated into another language. 7156 7157 A "Secondary Section" is a named appendix or a front-matter section 7158 of the Document that deals exclusively with the relationship of the 7159 publishers or authors of the Document to the Document's overall 7160 subject (or to related matters) and contains nothing that could 7161 fall directly within that overall subject. (Thus, if the Document 7162 is in part a textbook of mathematics, a Secondary Section may not 7163 explain any mathematics.) The relationship could be a matter of 7164 historical connection with the subject or with related matters, or 7165 of legal, commercial, philosophical, ethical or political position 7166 regarding them. 7167 7168 The "Invariant Sections" are certain Secondary Sections whose 7169 titles are designated, as being those of Invariant Sections, in 7170 the notice that says that the Document is released under this 7171 License. If a section does not fit the above definition of 7172 Secondary then it is not allowed to be designated as Invariant. 7173 The Document may contain zero Invariant Sections. If the Document 7174 does not identify any Invariant Sections then there are none. 7175 7176 The "Cover Texts" are certain short passages of text that are 7177 listed, as Front-Cover Texts or Back-Cover Texts, in the notice 7178 that says that the Document is released under this License. A 7179 Front-Cover Text may be at most 5 words, and a Back-Cover Text may 7180 be at most 25 words. 7181 7182 A "Transparent" copy of the Document means a machine-readable copy, 7183 represented in a format whose specification is available to the 7184 general public, that is suitable for revising the document 7185 straightforwardly with generic text editors or (for images 7186 composed of pixels) generic paint programs or (for drawings) some 7187 widely available drawing editor, and that is suitable for input to 7188 text formatters or for automatic translation to a variety of 7189 formats suitable for input to text formatters. A copy made in an 7190 otherwise Transparent file format whose markup, or absence of 7191 markup, has been arranged to thwart or discourage subsequent 7192 modification by readers is not Transparent. An image format is 7193 not Transparent if used for any substantial amount of text. A 7194 copy that is not "Transparent" is called "Opaque". 7195 7196 Examples of suitable formats for Transparent copies include plain 7197 ASCII without markup, Texinfo input format, LaTeX input format, 7198 SGML or XML using a publicly available DTD, and 7199 standard-conforming simple HTML, PostScript or PDF designed for 7200 human modification. Examples of transparent image formats include 7201 PNG, XCF and JPG. Opaque formats include proprietary formats that 7202 can be read and edited only by proprietary word processors, SGML or 7203 XML for which the DTD and/or processing tools are not generally 7204 available, and the machine-generated HTML, PostScript or PDF 7205 produced by some word processors for output purposes only. 7206 7207 The "Title Page" means, for a printed book, the title page itself, 7208 plus such following pages as are needed to hold, legibly, the 7209 material this License requires to appear in the title page. For 7210 works in formats which do not have any title page as such, "Title 7211 Page" means the text near the most prominent appearance of the 7212 work's title, preceding the beginning of the body of the text. 7213 7214 The "publisher" means any person or entity that distributes copies 7215 of the Document to the public. 7216 7217 A section "Entitled XYZ" means a named subunit of the Document 7218 whose title either is precisely XYZ or contains XYZ in parentheses 7219 following text that translates XYZ in another language. (Here XYZ 7220 stands for a specific section name mentioned below, such as 7221 "Acknowledgements", "Dedications", "Endorsements", or "History".) 7222 To "Preserve the Title" of such a section when you modify the 7223 Document means that it remains a section "Entitled XYZ" according 7224 to this definition. 7225 7226 The Document may include Warranty Disclaimers next to the notice 7227 which states that this License applies to the Document. These 7228 Warranty Disclaimers are considered to be included by reference in 7229 this License, but only as regards disclaiming warranties: any other 7230 implication that these Warranty Disclaimers may have is void and 7231 has no effect on the meaning of this License. 7232 7233 2. VERBATIM COPYING 7234 7235 You may copy and distribute the Document in any medium, either 7236 commercially or noncommercially, provided that this License, the 7237 copyright notices, and the license notice saying this License 7238 applies to the Document are reproduced in all copies, and that you 7239 add no other conditions whatsoever to those of this License. You 7240 may not use technical measures to obstruct or control the reading 7241 or further copying of the copies you make or distribute. However, 7242 you may accept compensation in exchange for copies. If you 7243 distribute a large enough number of copies you must also follow 7244 the conditions in section 3. 7245 7246 You may also lend copies, under the same conditions stated above, 7247 and you may publicly display copies. 7248 7249 3. COPYING IN QUANTITY 7250 7251 If you publish printed copies (or copies in media that commonly 7252 have printed covers) of the Document, numbering more than 100, and 7253 the Document's license notice requires Cover Texts, you must 7254 enclose the copies in covers that carry, clearly and legibly, all 7255 these Cover Texts: Front-Cover Texts on the front cover, and 7256 Back-Cover Texts on the back cover. Both covers must also clearly 7257 and legibly identify you as the publisher of these copies. The 7258 front cover must present the full title with all words of the 7259 title equally prominent and visible. You may add other material 7260 on the covers in addition. Copying with changes limited to the 7261 covers, as long as they preserve the title of the Document and 7262 satisfy these conditions, can be treated as verbatim copying in 7263 other respects. 7264 7265 If the required texts for either cover are too voluminous to fit 7266 legibly, you should put the first ones listed (as many as fit 7267 reasonably) on the actual cover, and continue the rest onto 7268 adjacent pages. 7269 7270 If you publish or distribute Opaque copies of the Document 7271 numbering more than 100, you must either include a 7272 machine-readable Transparent copy along with each Opaque copy, or 7273 state in or with each Opaque copy a computer-network location from 7274 which the general network-using public has access to download 7275 using public-standard network protocols a complete Transparent 7276 copy of the Document, free of added material. If you use the 7277 latter option, you must take reasonably prudent steps, when you 7278 begin distribution of Opaque copies in quantity, to ensure that 7279 this Transparent copy will remain thus accessible at the stated 7280 location until at least one year after the last time you 7281 distribute an Opaque copy (directly or through your agents or 7282 retailers) of that edition to the public. 7283 7284 It is requested, but not required, that you contact the authors of 7285 the Document well before redistributing any large number of 7286 copies, to give them a chance to provide you with an updated 7287 version of the Document. 7288 7289 4. MODIFICATIONS 7290 7291 You may copy and distribute a Modified Version of the Document 7292 under the conditions of sections 2 and 3 above, provided that you 7293 release the Modified Version under precisely this License, with 7294 the Modified Version filling the role of the Document, thus 7295 licensing distribution and modification of the Modified Version to 7296 whoever possesses a copy of it. In addition, you must do these 7297 things in the Modified Version: 7298 7299 A. Use in the Title Page (and on the covers, if any) a title 7300 distinct from that of the Document, and from those of 7301 previous versions (which should, if there were any, be listed 7302 in the History section of the Document). You may use the 7303 same title as a previous version if the original publisher of 7304 that version gives permission. 7305 7306 B. List on the Title Page, as authors, one or more persons or 7307 entities responsible for authorship of the modifications in 7308 the Modified Version, together with at least five of the 7309 principal authors of the Document (all of its principal 7310 authors, if it has fewer than five), unless they release you 7311 from this requirement. 7312 7313 C. State on the Title page the name of the publisher of the 7314 Modified Version, as the publisher. 7315 7316 D. Preserve all the copyright notices of the Document. 7317 7318 E. Add an appropriate copyright notice for your modifications 7319 adjacent to the other copyright notices. 7320 7321 F. Include, immediately after the copyright notices, a license 7322 notice giving the public permission to use the Modified 7323 Version under the terms of this License, in the form shown in 7324 the Addendum below. 7325 7326 G. Preserve in that license notice the full lists of Invariant 7327 Sections and required Cover Texts given in the Document's 7328 license notice. 7329 7330 H. Include an unaltered copy of this License. 7331 7332 I. Preserve the section Entitled "History", Preserve its Title, 7333 and add to it an item stating at least the title, year, new 7334 authors, and publisher of the Modified Version as given on 7335 the Title Page. If there is no section Entitled "History" in 7336 the Document, create one stating the title, year, authors, 7337 and publisher of the Document as given on its Title Page, 7338 then add an item describing the Modified Version as stated in 7339 the previous sentence. 7340 7341 J. Preserve the network location, if any, given in the Document 7342 for public access to a Transparent copy of the Document, and 7343 likewise the network locations given in the Document for 7344 previous versions it was based on. These may be placed in 7345 the "History" section. You may omit a network location for a 7346 work that was published at least four years before the 7347 Document itself, or if the original publisher of the version 7348 it refers to gives permission. 7349 7350 K. For any section Entitled "Acknowledgements" or "Dedications", 7351 Preserve the Title of the section, and preserve in the 7352 section all the substance and tone of each of the contributor 7353 acknowledgements and/or dedications given therein. 7354 7355 L. Preserve all the Invariant Sections of the Document, 7356 unaltered in their text and in their titles. Section numbers 7357 or the equivalent are not considered part of the section 7358 titles. 7359 7360 M. Delete any section Entitled "Endorsements". Such a section 7361 may not be included in the Modified Version. 7362 7363 N. Do not retitle any existing section to be Entitled 7364 "Endorsements" or to conflict in title with any Invariant 7365 Section. 7366 7367 O. Preserve any Warranty Disclaimers. 7368 7369 If the Modified Version includes new front-matter sections or 7370 appendices that qualify as Secondary Sections and contain no 7371 material copied from the Document, you may at your option 7372 designate some or all of these sections as invariant. To do this, 7373 add their titles to the list of Invariant Sections in the Modified 7374 Version's license notice. These titles must be distinct from any 7375 other section titles. 7376 7377 You may add a section Entitled "Endorsements", provided it contains 7378 nothing but endorsements of your Modified Version by various 7379 parties--for example, statements of peer review or that the text 7380 has been approved by an organization as the authoritative 7381 definition of a standard. 7382 7383 You may add a passage of up to five words as a Front-Cover Text, 7384 and a passage of up to 25 words as a Back-Cover Text, to the end 7385 of the list of Cover Texts in the Modified Version. Only one 7386 passage of Front-Cover Text and one of Back-Cover Text may be 7387 added by (or through arrangements made by) any one entity. If the 7388 Document already includes a cover text for the same cover, 7389 previously added by you or by arrangement made by the same entity 7390 you are acting on behalf of, you may not add another; but you may 7391 replace the old one, on explicit permission from the previous 7392 publisher that added the old one. 7393 7394 The author(s) and publisher(s) of the Document do not by this 7395 License give permission to use their names for publicity for or to 7396 assert or imply endorsement of any Modified Version. 7397 7398 5. COMBINING DOCUMENTS 7399 7400 You may combine the Document with other documents released under 7401 this License, under the terms defined in section 4 above for 7402 modified versions, provided that you include in the combination 7403 all of the Invariant Sections of all of the original documents, 7404 unmodified, and list them all as Invariant Sections of your 7405 combined work in its license notice, and that you preserve all 7406 their Warranty Disclaimers. 7407 7408 The combined work need only contain one copy of this License, and 7409 multiple identical Invariant Sections may be replaced with a single 7410 copy. If there are multiple Invariant Sections with the same name 7411 but different contents, make the title of each such section unique 7412 by adding at the end of it, in parentheses, the name of the 7413 original author or publisher of that section if known, or else a 7414 unique number. Make the same adjustment to the section titles in 7415 the list of Invariant Sections in the license notice of the 7416 combined work. 7417 7418 In the combination, you must combine any sections Entitled 7419 "History" in the various original documents, forming one section 7420 Entitled "History"; likewise combine any sections Entitled 7421 "Acknowledgements", and any sections Entitled "Dedications". You 7422 must delete all sections Entitled "Endorsements." 7423 7424 6. COLLECTIONS OF DOCUMENTS 7425 7426 You may make a collection consisting of the Document and other 7427 documents released under this License, and replace the individual 7428 copies of this License in the various documents with a single copy 7429 that is included in the collection, provided that you follow the 7430 rules of this License for verbatim copying of each of the 7431 documents in all other respects. 7432 7433 You may extract a single document from such a collection, and 7434 distribute it individually under this License, provided you insert 7435 a copy of this License into the extracted document, and follow 7436 this License in all other respects regarding verbatim copying of 7437 that document. 7438 7439 7. AGGREGATION WITH INDEPENDENT WORKS 7440 7441 A compilation of the Document or its derivatives with other 7442 separate and independent documents or works, in or on a volume of 7443 a storage or distribution medium, is called an "aggregate" if the 7444 copyright resulting from the compilation is not used to limit the 7445 legal rights of the compilation's users beyond what the individual 7446 works permit. When the Document is included in an aggregate, this 7447 License does not apply to the other works in the aggregate which 7448 are not themselves derivative works of the Document. 7449 7450 If the Cover Text requirement of section 3 is applicable to these 7451 copies of the Document, then if the Document is less than one half 7452 of the entire aggregate, the Document's Cover Texts may be placed 7453 on covers that bracket the Document within the aggregate, or the 7454 electronic equivalent of covers if the Document is in electronic 7455 form. Otherwise they must appear on printed covers that bracket 7456 the whole aggregate. 7457 7458 8. TRANSLATION 7459 7460 Translation is considered a kind of modification, so you may 7461 distribute translations of the Document under the terms of section 7462 4. Replacing Invariant Sections with translations requires special 7463 permission from their copyright holders, but you may include 7464 translations of some or all Invariant Sections in addition to the 7465 original versions of these Invariant Sections. You may include a 7466 translation of this License, and all the license notices in the 7467 Document, and any Warranty Disclaimers, provided that you also 7468 include the original English version of this License and the 7469 original versions of those notices and disclaimers. In case of a 7470 disagreement between the translation and the original version of 7471 this License or a notice or disclaimer, the original version will 7472 prevail. 7473 7474 If a section in the Document is Entitled "Acknowledgements", 7475 "Dedications", or "History", the requirement (section 4) to 7476 Preserve its Title (section 1) will typically require changing the 7477 actual title. 7478 7479 9. TERMINATION 7480 7481 You may not copy, modify, sublicense, or distribute the Document 7482 except as expressly provided under this License. Any attempt 7483 otherwise to copy, modify, sublicense, or distribute it is void, 7484 and will automatically terminate your rights under this License. 7485 7486 However, if you cease all violation of this License, then your 7487 license from a particular copyright holder is reinstated (a) 7488 provisionally, unless and until the copyright holder explicitly 7489 and finally terminates your license, and (b) permanently, if the 7490 copyright holder fails to notify you of the violation by some 7491 reasonable means prior to 60 days after the cessation. 7492 7493 Moreover, your license from a particular copyright holder is 7494 reinstated permanently if the copyright holder notifies you of the 7495 violation by some reasonable means, this is the first time you have 7496 received notice of violation of this License (for any work) from 7497 that copyright holder, and you cure the violation prior to 30 days 7498 after your receipt of the notice. 7499 7500 Termination of your rights under this section does not terminate 7501 the licenses of parties who have received copies or rights from 7502 you under this License. If your rights have been terminated and 7503 not permanently reinstated, receipt of a copy of some or all of 7504 the same material does not give you any rights to use it. 7505 7506 10. FUTURE REVISIONS OF THIS LICENSE 7507 7508 The Free Software Foundation may publish new, revised versions of 7509 the GNU Free Documentation License from time to time. Such new 7510 versions will be similar in spirit to the present version, but may 7511 differ in detail to address new problems or concerns. See 7512 `http://www.gnu.org/copyleft/'. 7513 7514 Each version of the License is given a distinguishing version 7515 number. If the Document specifies that a particular numbered 7516 version of this License "or any later version" applies to it, you 7517 have the option of following the terms and conditions either of 7518 that specified version or of any later version that has been 7519 published (not as a draft) by the Free Software Foundation. If 7520 the Document does not specify a version number of this License, 7521 you may choose any version ever published (not as a draft) by the 7522 Free Software Foundation. If the Document specifies that a proxy 7523 can decide which future versions of this License can be used, that 7524 proxy's public statement of acceptance of a version permanently 7525 authorizes you to choose that version for the Document. 7526 7527 11. RELICENSING 7528 7529 "Massive Multiauthor Collaboration Site" (or "MMC Site") means any 7530 World Wide Web server that publishes copyrightable works and also 7531 provides prominent facilities for anybody to edit those works. A 7532 public wiki that anybody can edit is an example of such a server. 7533 A "Massive Multiauthor Collaboration" (or "MMC") contained in the 7534 site means any set of copyrightable works thus published on the MMC 7535 site. 7536 7537 "CC-BY-SA" means the Creative Commons Attribution-Share Alike 3.0 7538 license published by Creative Commons Corporation, a not-for-profit 7539 corporation with a principal place of business in San Francisco, 7540 California, as well as future copyleft versions of that license 7541 published by that same organization. 7542 7543 "Incorporate" means to publish or republish a Document, in whole or 7544 in part, as part of another Document. 7545 7546 An MMC is "eligible for relicensing" if it is licensed under this 7547 License, and if all works that were first published under this 7548 License somewhere other than this MMC, and subsequently 7549 incorporated in whole or in part into the MMC, (1) had no cover 7550 texts or invariant sections, and (2) were thus incorporated prior 7551 to November 1, 2008. 7552 7553 The operator of an MMC Site may republish an MMC contained in the 7554 site under CC-BY-SA on the same site at any time before August 1, 7555 2009, provided the MMC is eligible for relicensing. 7556 7557 7558ADDENDUM: How to use this License for your documents 7559==================================================== 7560 7561To use this License in a document you have written, include a copy of 7562the License in the document and put the following copyright and license 7563notices just after the title page: 7564 7565 Copyright (C) YEAR YOUR NAME. 7566 Permission is granted to copy, distribute and/or modify this document 7567 under the terms of the GNU Free Documentation License, Version 1.3 7568 or any later version published by the Free Software Foundation; 7569 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover 7570 Texts. A copy of the license is included in the section entitled ``GNU 7571 Free Documentation License''. 7572 7573 If you have Invariant Sections, Front-Cover Texts and Back-Cover 7574Texts, replace the "with...Texts." line with this: 7575 7576 with the Invariant Sections being LIST THEIR TITLES, with 7577 the Front-Cover Texts being LIST, and with the Back-Cover Texts 7578 being LIST. 7579 7580 If you have Invariant Sections without Cover Texts, or some other 7581combination of the three, merge those two alternatives to suit the 7582situation. 7583 7584 If your document contains nontrivial examples of program code, we 7585recommend releasing these examples in parallel under your choice of 7586free software license, such as the GNU General Public License, to 7587permit their use in free software. 7588 7589 7590File: ld.info, Node: LD Index, Prev: GNU Free Documentation License, Up: Top 7591 7592LD Index 7593******** 7594 7595[index] 7596* Menu: 7597 7598* ": Symbols. (line 6) 7599* -(: Options. (line 804) 7600* --accept-unknown-input-arch: Options. (line 822) 7601* --add-needed: Options. (line 850) 7602* --add-stdcall-alias: Options. (line 1750) 7603* --allow-multiple-definition: Options. (line 1128) 7604* --allow-shlib-undefined: Options. (line 1134) 7605* --architecture=ARCH: Options. (line 123) 7606* --as-needed: Options. (line 832) 7607* --audit AUDITLIB: Options. (line 112) 7608* --auxiliary=NAME: Options. (line 255) 7609* --bank-window: Options. (line 2204) 7610* --base-file: Options. (line 1755) 7611* --be8: ARM. (line 28) 7612* --bss-plt: PowerPC ELF32. (line 16) 7613* --build-id: Options. (line 1712) 7614* --build-id=STYLE: Options. (line 1712) 7615* --check-sections: Options. (line 936) 7616* --compress-debug-sections=none: Options. (line 1682) 7617* --compress-debug-sections=zlib: Options. (line 1682) 7618* --compress-debug-sections=zlib-gabi: Options. (line 1682) 7619* --compress-debug-sections=zlib-gnu: Options. (line 1682) 7620* --copy-dt-needed-entries: Options. (line 948) 7621* --cref: Options. (line 968) 7622* --default-imported-symver: Options. (line 1171) 7623* --default-script=SCRIPT: Options. (line 562) 7624* --default-symver: Options. (line 1167) 7625* --defsym=SYMBOL=EXP: Options. (line 997) 7626* --demangle[=STYLE]: Options. (line 1009) 7627* --depaudit AUDITLIB: Options. (line 177) 7628* --disable-auto-image-base: Options. (line 1942) 7629* --disable-auto-import: Options. (line 2077) 7630* --disable-large-address-aware: Options. (line 1881) 7631* --disable-long-section-names: Options. (line 1765) 7632* --disable-new-dtags: Options. (line 1658) 7633* --disable-runtime-pseudo-reloc: Options. (line 2090) 7634* --disable-stdcall-fixup: Options. (line 1787) 7635* --discard-all: Options. (line 647) 7636* --discard-locals: Options. (line 651) 7637* --dll: Options. (line 1760) 7638* --dll-search-prefix: Options. (line 1948) 7639* --dotsyms: PowerPC64 ELF64. (line 33) 7640* --dsbt-index: Options. (line 2181) 7641* --dsbt-size: Options. (line 2176) 7642* --dynamic-linker=FILE: Options. (line 1022) 7643* --dynamic-list-cpp-new: Options. (line 928) 7644* --dynamic-list-cpp-typeinfo: Options. (line 932) 7645* --dynamic-list-data: Options. (line 925) 7646* --dynamic-list=DYNAMIC-LIST-FILE: Options. (line 912) 7647* --dynamicbase: Options. (line 2130) 7648* --eh-frame-hdr: Options. (line 1649) 7649* --emit-relocs: Options. (line 497) 7650* --emit-stack-syms: SPU ELF. (line 46) 7651* --emit-stub-syms <1>: PowerPC ELF32. (line 47) 7652* --emit-stub-syms <2>: SPU ELF. (line 15) 7653* --emit-stub-syms: PowerPC64 ELF64. (line 29) 7654* --enable-auto-image-base: Options. (line 1933) 7655* --enable-auto-import: Options. (line 1957) 7656* --enable-extra-pe-debug: Options. (line 2095) 7657* --enable-long-section-names: Options. (line 1765) 7658* --enable-new-dtags: Options. (line 1658) 7659* --enable-runtime-pseudo-reloc: Options. (line 2082) 7660* --enable-stdcall-fixup: Options. (line 1787) 7661* --entry=ENTRY: Options. (line 187) 7662* --error-unresolved-symbols: Options. (line 1602) 7663* --exclude-all-symbols: Options. (line 1841) 7664* --exclude-libs: Options. (line 197) 7665* --exclude-modules-for-implib: Options. (line 208) 7666* --exclude-symbols: Options. (line 1835) 7667* --export-all-symbols: Options. (line 1811) 7668* --export-dynamic: Options. (line 221) 7669* --extra-overlay-stubs: SPU ELF. (line 19) 7670* --fatal-warnings: Options. (line 1035) 7671* --file-alignment: Options. (line 1845) 7672* --filter=NAME: Options. (line 276) 7673* --fix-arm1176: ARM. (line 112) 7674* --fix-cortex-a53-835769: ARM. (line 213) 7675* --fix-cortex-a8: ARM. (line 204) 7676* --fix-stm32l4xx-629360: ARM. (line 121) 7677* --fix-v4bx: ARM. (line 50) 7678* --fix-v4bx-interworking: ARM. (line 63) 7679* --force-dynamic: Options. (line 506) 7680* --force-exe-suffix: Options. (line 1040) 7681* --forceinteg: Options. (line 2135) 7682* --format=FORMAT: Options. (line 134) 7683* --format=VERSION: TI COFF. (line 6) 7684* --gc-sections: Options. (line 1050) 7685* --got: Options. (line 2217) 7686* --got=TYPE: M68K. (line 6) 7687* --gpsize=VALUE: Options. (line 309) 7688* --hash-size=NUMBER: Options. (line 1668) 7689* --hash-style=STYLE: Options. (line 1676) 7690* --heap: Options. (line 1851) 7691* --help: Options. (line 1101) 7692* --high-entropy-va: Options. (line 2126) 7693* --image-base: Options. (line 1858) 7694* --insert-timestamp: Options. (line 2158) 7695* --insn32 <1>: MIPS. (line 6) 7696* --insn32: Options. (line 2229) 7697* --just-symbols=FILE: Options. (line 529) 7698* --kill-at: Options. (line 1867) 7699* --large-address-aware: Options. (line 1872) 7700* --ld-generated-unwind-info: Options. (line 1653) 7701* --leading-underscore: Options. (line 1805) 7702* --library-path=DIR: Options. (line 367) 7703* --library=NAMESPEC: Options. (line 334) 7704* --local-store=lo:hi: SPU ELF. (line 24) 7705* --long-plt: ARM. (line 224) 7706* --major-image-version: Options. (line 1888) 7707* --major-os-version: Options. (line 1893) 7708* --major-subsystem-version: Options. (line 1897) 7709* --merge-exidx-entries: ARM. (line 221) 7710* --minor-image-version: Options. (line 1902) 7711* --minor-os-version: Options. (line 1907) 7712* --minor-subsystem-version: Options. (line 1911) 7713* --mri-script=MRI-CMDFILE: Options. (line 158) 7714* --multi-subspace: HPPA ELF32. (line 6) 7715* --nmagic: Options. (line 439) 7716* --no-accept-unknown-input-arch: Options. (line 822) 7717* --no-add-needed: Options. (line 850) 7718* --no-allow-shlib-undefined: Options. (line 1134) 7719* --no-as-needed: Options. (line 832) 7720* --no-bind: Options. (line 2149) 7721* --no-check-sections: Options. (line 936) 7722* --no-copy-dt-needed-entries: Options. (line 948) 7723* --no-define-common: Options. (line 981) 7724* --no-demangle: Options. (line 1009) 7725* --no-dotsyms: PowerPC64 ELF64. (line 33) 7726* --no-dynamic-linker: Options. (line 1029) 7727* --no-enum-size-warning: ARM. (line 159) 7728* --no-export-dynamic: Options. (line 221) 7729* --no-fatal-warnings: Options. (line 1035) 7730* --no-fix-arm1176: ARM. (line 112) 7731* --no-fix-cortex-a53-835769: ARM. (line 213) 7732* --no-fix-cortex-a8: ARM. (line 204) 7733* --no-gc-sections: Options. (line 1050) 7734* --no-insn32 <1>: Options. (line 2230) 7735* --no-insn32: MIPS. (line 6) 7736* --no-isolation: Options. (line 2142) 7737* --no-keep-memory: Options. (line 1113) 7738* --no-leading-underscore: Options. (line 1805) 7739* --no-merge-exidx-entries <1>: Options. (line 2188) 7740* --no-merge-exidx-entries: ARM. (line 221) 7741* --no-multi-toc: PowerPC64 ELF64. (line 97) 7742* --no-omagic: Options. (line 454) 7743* --no-opd-optimize: PowerPC64 ELF64. (line 71) 7744* --no-overlays: SPU ELF. (line 9) 7745* --no-plt-align: PowerPC64 ELF64. (line 119) 7746* --no-plt-static-chain: PowerPC64 ELF64. (line 127) 7747* --no-plt-thread-safe: PowerPC64 ELF64. (line 133) 7748* --no-print-gc-sections: Options. (line 1074) 7749* --no-save-restore-funcs: PowerPC64 ELF64. (line 44) 7750* --no-seh: Options. (line 2145) 7751* --no-tls-get-addr-optimize: PowerPC64 ELF64. (line 57) 7752* --no-tls-optimize <1>: PowerPC64 ELF64. (line 52) 7753* --no-tls-optimize: PowerPC ELF32. (line 51) 7754* --no-toc-optimize: PowerPC64 ELF64. (line 83) 7755* --no-toc-sort: PowerPC64 ELF64. (line 109) 7756* --no-trampoline: Options. (line 2198) 7757* --no-undefined: Options. (line 1120) 7758* --no-undefined-version: Options. (line 1162) 7759* --no-warn-mismatch: Options. (line 1175) 7760* --no-warn-search-mismatch: Options. (line 1184) 7761* --no-wchar-size-warning: ARM. (line 166) 7762* --no-whole-archive: Options. (line 1188) 7763* --noinhibit-exec: Options. (line 1192) 7764* --non-overlapping-opd: PowerPC64 ELF64. (line 77) 7765* --nxcompat: Options. (line 2138) 7766* --oformat=OUTPUT-FORMAT: Options. (line 1204) 7767* --omagic: Options. (line 445) 7768* --orphan-handling=MODE: Options. (line 606) 7769* --out-implib: Options. (line 1924) 7770* --output-def: Options. (line 1916) 7771* --output=OUTPUT: Options. (line 460) 7772* --pic-executable: Options. (line 1217) 7773* --pic-veneer: ARM. (line 172) 7774* --plt-align: PowerPC64 ELF64. (line 119) 7775* --plt-static-chain: PowerPC64 ELF64. (line 127) 7776* --plt-thread-safe: PowerPC64 ELF64. (line 133) 7777* --plugin: SPU ELF. (line 6) 7778* --pop-state: Options. (line 494) 7779* --print-gc-sections: Options. (line 1074) 7780* --print-map: Options. (line 402) 7781* --print-memory-usage: Options. (line 1089) 7782* --print-output-format: Options. (line 1083) 7783* --push-state: Options. (line 476) 7784* --reduce-memory-overheads: Options. (line 1698) 7785* --relax: Options. (line 1233) 7786* --relax on i960: i960. (line 31) 7787* --relax on Nios II: Nios II. (line 6) 7788* --relax on PowerPC: PowerPC ELF32. (line 6) 7789* --relax on Xtensa: Xtensa. (line 27) 7790* --relocatable: Options. (line 510) 7791* --require-defined=SYMBOL: Options. (line 588) 7792* --retain-symbols-file=FILENAME: Options. (line 1259) 7793* --save-restore-funcs: PowerPC64 ELF64. (line 44) 7794* --script=SCRIPT: Options. (line 553) 7795* --sdata-got: PowerPC ELF32. (line 33) 7796* --section-alignment: Options. (line 2100) 7797* --section-start=SECTIONNAME=ORG: Options. (line 1415) 7798* --secure-plt: PowerPC ELF32. (line 26) 7799* --sort-common: Options. (line 1357) 7800* --sort-section=alignment: Options. (line 1372) 7801* --sort-section=name: Options. (line 1368) 7802* --split-by-file: Options. (line 1376) 7803* --split-by-reloc: Options. (line 1381) 7804* --stack: Options. (line 2106) 7805* --stack-analysis: SPU ELF. (line 29) 7806* --stats: Options. (line 1394) 7807* --strip-all: Options. (line 540) 7808* --strip-debug: Options. (line 544) 7809* --stub-group-size: PowerPC64 ELF64. (line 6) 7810* --stub-group-size=N <1>: HPPA ELF32. (line 12) 7811* --stub-group-size=N: ARM. (line 177) 7812* --subsystem: Options. (line 2113) 7813* --support-old-code: ARM. (line 6) 7814* --sysroot=DIRECTORY: Options. (line 1398) 7815* --target-help: Options. (line 1105) 7816* --target1-abs: ARM. (line 33) 7817* --target1-rel: ARM. (line 33) 7818* --target2=TYPE: ARM. (line 38) 7819* --thumb-entry=ENTRY: ARM. (line 17) 7820* --tls-get-addr-optimize: PowerPC64 ELF64. (line 57) 7821* --trace: Options. (line 549) 7822* --trace-symbol=SYMBOL: Options. (line 657) 7823* --traditional-format: Options. (line 1403) 7824* --tsaware: Options. (line 2155) 7825* --undefined=SYMBOL: Options. (line 575) 7826* --unique[=SECTION]: Options. (line 632) 7827* --unresolved-symbols: Options. (line 1445) 7828* --use-blx: ARM. (line 75) 7829* --use-nul-prefixed-import-tables: ARM. (line 23) 7830* --verbose[=NUMBER]: Options. (line 1474) 7831* --version: Options. (line 641) 7832* --version-script=VERSION-SCRIPTFILE: Options. (line 1482) 7833* --vfp11-denorm-fix: ARM. (line 84) 7834* --warn-alternate-em: Options. (line 1594) 7835* --warn-common: Options. (line 1493) 7836* --warn-constructors: Options. (line 1561) 7837* --warn-multiple-gp: Options. (line 1566) 7838* --warn-once: Options. (line 1580) 7839* --warn-section-align: Options. (line 1584) 7840* --warn-shared-textrel: Options. (line 1591) 7841* --warn-unresolved-symbols: Options. (line 1597) 7842* --wdmdriver: Options. (line 2152) 7843* --whole-archive: Options. (line 1606) 7844* --wrap=SYMBOL: Options. (line 1620) 7845* -A ARCH: Options. (line 122) 7846* -a KEYWORD: Options. (line 105) 7847* -assert KEYWORD: Options. (line 857) 7848* -b FORMAT: Options. (line 134) 7849* -Bdynamic: Options. (line 860) 7850* -Bgroup: Options. (line 870) 7851* -Bshareable: Options. (line 1350) 7852* -Bstatic: Options. (line 877) 7853* -Bsymbolic: Options. (line 892) 7854* -Bsymbolic-functions: Options. (line 903) 7855* -c MRI-CMDFILE: Options. (line 158) 7856* -call_shared: Options. (line 860) 7857* -d: Options. (line 168) 7858* -dc: Options. (line 168) 7859* -dn: Options. (line 877) 7860* -dp: Options. (line 168) 7861* -dT SCRIPT: Options. (line 562) 7862* -dy: Options. (line 860) 7863* -E: Options. (line 221) 7864* -e ENTRY: Options. (line 187) 7865* -EB: Options. (line 248) 7866* -EL: Options. (line 251) 7867* -f NAME: Options. (line 255) 7868* -F NAME: Options. (line 276) 7869* -fini=NAME: Options. (line 300) 7870* -g: Options. (line 306) 7871* -G VALUE: Options. (line 309) 7872* -h NAME: Options. (line 316) 7873* -i: Options. (line 325) 7874* -IFILE: Options. (line 1022) 7875* -init=NAME: Options. (line 328) 7876* -L DIR: Options. (line 367) 7877* -l NAMESPEC: Options. (line 334) 7878* -M: Options. (line 402) 7879* -m EMULATION: Options. (line 392) 7880* -Map=MAPFILE: Options. (line 1109) 7881* -n: Options. (line 439) 7882* -N: Options. (line 445) 7883* -no-relax: Options. (line 1233) 7884* -non_shared: Options. (line 877) 7885* -nostdlib: Options. (line 1198) 7886* -O LEVEL: Options. (line 466) 7887* -o OUTPUT: Options. (line 460) 7888* -P AUDITLIB: Options. (line 177) 7889* -pie: Options. (line 1217) 7890* -q: Options. (line 497) 7891* -qmagic: Options. (line 1227) 7892* -Qy: Options. (line 1230) 7893* -r: Options. (line 510) 7894* -R FILE: Options. (line 529) 7895* -rpath-link=DIR: Options. (line 1295) 7896* -rpath=DIR: Options. (line 1273) 7897* -s: Options. (line 540) 7898* -S: Options. (line 544) 7899* -shared: Options. (line 1350) 7900* -soname=NAME: Options. (line 316) 7901* -static: Options. (line 877) 7902* -t: Options. (line 549) 7903* -T SCRIPT: Options. (line 553) 7904* -Tbss=ORG: Options. (line 1424) 7905* -Tdata=ORG: Options. (line 1424) 7906* -Tldata-segment=ORG: Options. (line 1440) 7907* -Trodata-segment=ORG: Options. (line 1434) 7908* -Ttext-segment=ORG: Options. (line 1430) 7909* -Ttext=ORG: Options. (line 1424) 7910* -u SYMBOL: Options. (line 575) 7911* -Ur: Options. (line 596) 7912* -v: Options. (line 641) 7913* -V: Options. (line 641) 7914* -x: Options. (line 647) 7915* -X: Options. (line 651) 7916* -Y PATH: Options. (line 666) 7917* -y SYMBOL: Options. (line 657) 7918* -z defs: Options. (line 1120) 7919* -z KEYWORD: Options. (line 670) 7920* -z muldefs: Options. (line 1128) 7921* .: Location Counter. (line 6) 7922* /DISCARD/: Output Section Discarding. 7923 (line 26) 7924* 32-bit PLT entries: ARM. (line 224) 7925* :PHDR: Output Section Phdr. 7926 (line 6) 7927* =FILLEXP: Output Section Fill. 7928 (line 6) 7929* >REGION: Output Section Region. 7930 (line 6) 7931* [COMMON]: Input Section Common. 7932 (line 29) 7933* ABSOLUTE (MRI): MRI. (line 33) 7934* absolute and relocatable symbols: Expression Section. (line 6) 7935* absolute expressions: Expression Section. (line 6) 7936* ABSOLUTE(EXP): Builtin Functions. (line 10) 7937* ADDR(SECTION): Builtin Functions. (line 17) 7938* address, section: Output Section Address. 7939 (line 6) 7940* ALIAS (MRI): MRI. (line 44) 7941* ALIGN (MRI): MRI. (line 50) 7942* align expression: Builtin Functions. (line 38) 7943* align location counter: Builtin Functions. (line 38) 7944* ALIGN(ALIGN): Builtin Functions. (line 38) 7945* ALIGN(EXP,ALIGN): Builtin Functions. (line 38) 7946* ALIGN(SECTION_ALIGN): Forced Output Alignment. 7947 (line 6) 7948* aligned common symbols: WIN32. (line 424) 7949* ALIGNOF(SECTION): Builtin Functions. (line 64) 7950* allocating memory: MEMORY. (line 6) 7951* architecture: Miscellaneous Commands. 7952 (line 98) 7953* architectures: Options. (line 122) 7954* archive files, from cmd line: Options. (line 334) 7955* archive search path in linker script: File Commands. (line 76) 7956* arithmetic: Expressions. (line 6) 7957* arithmetic operators: Operators. (line 6) 7958* ARM interworking support: ARM. (line 6) 7959* ARM1176 erratum workaround: ARM. (line 112) 7960* AS_NEEDED(FILES): File Commands. (line 56) 7961* ASSERT: Miscellaneous Commands. 7962 (line 9) 7963* assertion in linker script: Miscellaneous Commands. 7964 (line 9) 7965* assignment in scripts: Assignments. (line 6) 7966* AT(LMA): Output Section LMA. (line 6) 7967* AT>LMA_REGION: Output Section LMA. (line 6) 7968* automatic data imports: WIN32. (line 191) 7969* back end: BFD. (line 6) 7970* BASE (MRI): MRI. (line 54) 7971* BE8: ARM. (line 28) 7972* BFD canonical format: Canonical format. (line 11) 7973* BFD requirements: BFD. (line 16) 7974* big-endian objects: Options. (line 248) 7975* binary input format: Options. (line 134) 7976* BLOCK(EXP): Builtin Functions. (line 77) 7977* bug criteria: Bug Criteria. (line 6) 7978* bug reports: Bug Reporting. (line 6) 7979* bugs in ld: Reporting Bugs. (line 6) 7980* BYTE(EXPRESSION): Output Section Data. 7981 (line 6) 7982* C++ constructors, arranging in link: Output Section Keywords. 7983 (line 19) 7984* CHIP (MRI): MRI. (line 58) 7985* COLLECT_NO_DEMANGLE: Environment. (line 29) 7986* combining symbols, warnings on: Options. (line 1493) 7987* command files: Scripts. (line 6) 7988* command line: Options. (line 6) 7989* common allocation: Options. (line 168) 7990* common allocation in linker script: Miscellaneous Commands. 7991 (line 46) 7992* common symbol placement: Input Section Common. 7993 (line 6) 7994* COMMONPAGESIZE: Symbolic Constants. (line 13) 7995* compatibility, MRI: Options. (line 158) 7996* CONSTANT: Symbolic Constants. (line 6) 7997* constants in linker scripts: Constants. (line 6) 7998* constraints on output sections: Output Section Constraint. 7999 (line 6) 8000* CONSTRUCTORS: Output Section Keywords. 8001 (line 19) 8002* constructors: Options. (line 596) 8003* constructors, arranging in link: Output Section Keywords. 8004 (line 19) 8005* Cortex-A53 erratum 835769 workaround: ARM. (line 213) 8006* Cortex-A8 erratum workaround: ARM. (line 204) 8007* crash of linker: Bug Criteria. (line 9) 8008* CREATE_OBJECT_SYMBOLS: Output Section Keywords. 8009 (line 9) 8010* creating a DEF file: WIN32. (line 158) 8011* cross reference table: Options. (line 968) 8012* cross references: Miscellaneous Commands. 8013 (line 82) 8014* current output location: Location Counter. (line 6) 8015* data: Output Section Data. 8016 (line 6) 8017* DATA_SEGMENT_ALIGN(MAXPAGESIZE, COMMONPAGESIZE): Builtin Functions. 8018 (line 82) 8019* DATA_SEGMENT_END(EXP): Builtin Functions. (line 104) 8020* DATA_SEGMENT_RELRO_END(OFFSET, EXP): Builtin Functions. (line 110) 8021* dbx: Options. (line 1408) 8022* DEF files, creating: Options. (line 1916) 8023* default emulation: Environment. (line 21) 8024* default input format: Environment. (line 9) 8025* defined symbol: Options. (line 588) 8026* DEFINED(SYMBOL): Builtin Functions. (line 123) 8027* deleting local symbols: Options. (line 647) 8028* demangling, default: Environment. (line 29) 8029* demangling, from command line: Options. (line 1009) 8030* direct linking to a dll: WIN32. (line 239) 8031* discarding sections: Output Section Discarding. 8032 (line 6) 8033* discontinuous memory: MEMORY. (line 6) 8034* DLLs, creating: Options. (line 1924) 8035* DLLs, linking to: Options. (line 1948) 8036* dot: Location Counter. (line 6) 8037* dot inside sections: Location Counter. (line 36) 8038* dot outside sections: Location Counter. (line 66) 8039* dynamic linker, from command line: Options. (line 1022) 8040* dynamic symbol table: Options. (line 221) 8041* ELF program headers: PHDRS. (line 6) 8042* emulation: Options. (line 392) 8043* emulation, default: Environment. (line 21) 8044* END (MRI): MRI. (line 62) 8045* endianness: Options. (line 248) 8046* entry point: Entry Point. (line 6) 8047* entry point, from command line: Options. (line 187) 8048* entry point, thumb: ARM. (line 17) 8049* ENTRY(SYMBOL): Entry Point. (line 6) 8050* error on valid input: Bug Criteria. (line 12) 8051* example of linker script: Simple Example. (line 6) 8052* exporting DLL symbols: WIN32. (line 19) 8053* expression evaluation order: Evaluation. (line 6) 8054* expression sections: Expression Section. (line 6) 8055* expression, absolute: Builtin Functions. (line 10) 8056* expressions: Expressions. (line 6) 8057* EXTERN: Miscellaneous Commands. 8058 (line 39) 8059* fatal signal: Bug Criteria. (line 9) 8060* file name wildcard patterns: Input Section Wildcards. 8061 (line 6) 8062* FILEHDR: PHDRS. (line 62) 8063* filename symbols: Output Section Keywords. 8064 (line 9) 8065* fill pattern, entire section: Output Section Fill. 8066 (line 6) 8067* FILL(EXPRESSION): Output Section Data. 8068 (line 39) 8069* finalization function: Options. (line 300) 8070* first input file: File Commands. (line 84) 8071* first instruction: Entry Point. (line 6) 8072* FIX_V4BX: ARM. (line 50) 8073* FIX_V4BX_INTERWORKING: ARM. (line 63) 8074* FORCE_COMMON_ALLOCATION: Miscellaneous Commands. 8075 (line 46) 8076* forcing input section alignment: Forced Input Alignment. 8077 (line 6) 8078* forcing output section alignment: Forced Output Alignment. 8079 (line 6) 8080* forcing the creation of dynamic sections: Options. (line 506) 8081* FORMAT (MRI): MRI. (line 66) 8082* functions in expressions: Builtin Functions. (line 6) 8083* garbage collection <1>: Options. (line 1050) 8084* garbage collection: Input Section Keep. (line 6) 8085* generating optimized output: Options. (line 466) 8086* GNU linker: Overview. (line 6) 8087* GNUTARGET: Environment. (line 9) 8088* GROUP(FILES): File Commands. (line 49) 8089* grouping input files: File Commands. (line 49) 8090* groups of archives: Options. (line 804) 8091* H8/300 support: H8/300. (line 6) 8092* header size: Builtin Functions. (line 191) 8093* heap size: Options. (line 1851) 8094* help: Options. (line 1101) 8095* HIDDEN: HIDDEN. (line 6) 8096* holes: Location Counter. (line 12) 8097* holes, filling: Output Section Data. 8098 (line 39) 8099* HPPA multiple sub-space stubs: HPPA ELF32. (line 6) 8100* HPPA stub grouping: HPPA ELF32. (line 12) 8101* i960 support: i960. (line 6) 8102* image base: Options. (line 1858) 8103* implicit linker scripts: Implicit Linker Scripts. 8104 (line 6) 8105* import libraries: WIN32. (line 10) 8106* INCLUDE FILENAME: File Commands. (line 9) 8107* including a linker script: File Commands. (line 9) 8108* including an entire archive: Options. (line 1606) 8109* incremental link: Options. (line 325) 8110* INHIBIT_COMMON_ALLOCATION: Miscellaneous Commands. 8111 (line 51) 8112* initialization function: Options. (line 328) 8113* initialized data in ROM: Output Section LMA. (line 39) 8114* input file format in linker script: Format Commands. (line 35) 8115* input filename symbols: Output Section Keywords. 8116 (line 9) 8117* input files in linker scripts: File Commands. (line 19) 8118* input files, displaying: Options. (line 549) 8119* input format: Options. (line 134) 8120* input object files in linker scripts: File Commands. (line 19) 8121* input section alignment: Forced Input Alignment. 8122 (line 6) 8123* input section basics: Input Section Basics. 8124 (line 6) 8125* input section wildcards: Input Section Wildcards. 8126 (line 6) 8127* input sections: Input Section. (line 6) 8128* INPUT(FILES): File Commands. (line 19) 8129* INSERT: Miscellaneous Commands. 8130 (line 56) 8131* insert user script into default script: Miscellaneous Commands. 8132 (line 56) 8133* integer notation: Constants. (line 6) 8134* integer suffixes: Constants. (line 15) 8135* internal object-file format: Canonical format. (line 11) 8136* invalid input: Bug Criteria. (line 14) 8137* K and M integer suffixes: Constants. (line 15) 8138* KEEP: Input Section Keep. (line 6) 8139* l =: MEMORY. (line 74) 8140* lazy evaluation: Evaluation. (line 6) 8141* ld bugs, reporting: Bug Reporting. (line 6) 8142* LD_FEATURE(STRING): Miscellaneous Commands. 8143 (line 104) 8144* ldata segment origin, cmd line: Options. (line 1441) 8145* LDEMULATION: Environment. (line 21) 8146* len =: MEMORY. (line 74) 8147* LENGTH =: MEMORY. (line 74) 8148* LENGTH(MEMORY): Builtin Functions. (line 140) 8149* library search path in linker script: File Commands. (line 76) 8150* link map: Options. (line 402) 8151* link-time runtime library search path: Options. (line 1295) 8152* linker crash: Bug Criteria. (line 9) 8153* linker script concepts: Basic Script Concepts. 8154 (line 6) 8155* linker script example: Simple Example. (line 6) 8156* linker script file commands: File Commands. (line 6) 8157* linker script format: Script Format. (line 6) 8158* linker script input object files: File Commands. (line 19) 8159* linker script simple commands: Simple Commands. (line 6) 8160* linker scripts: Scripts. (line 6) 8161* LIST (MRI): MRI. (line 77) 8162* little-endian objects: Options. (line 251) 8163* LOAD (MRI): MRI. (line 84) 8164* load address: Output Section LMA. (line 6) 8165* LOADADDR(SECTION): Builtin Functions. (line 143) 8166* loading, preventing: Output Section Type. 8167 (line 22) 8168* local symbols, deleting: Options. (line 651) 8169* location counter: Location Counter. (line 6) 8170* LOG2CEIL(EXP): Builtin Functions. (line 147) 8171* LONG(EXPRESSION): Output Section Data. 8172 (line 6) 8173* M and K integer suffixes: Constants. (line 15) 8174* M68HC11 and 68HC12 support: M68HC11/68HC12. (line 6) 8175* machine architecture: Miscellaneous Commands. 8176 (line 98) 8177* machine dependencies: Machine Dependent. (line 6) 8178* mapping input sections to output sections: Input Section. (line 6) 8179* MAX: Builtin Functions. (line 150) 8180* MAXPAGESIZE: Symbolic Constants. (line 10) 8181* MEMORY: MEMORY. (line 6) 8182* memory region attributes: MEMORY. (line 34) 8183* memory regions: MEMORY. (line 6) 8184* memory regions and sections: Output Section Region. 8185 (line 6) 8186* memory usage: Options. (line 1113) 8187* Merging exidx entries: ARM. (line 221) 8188* MIN: Builtin Functions. (line 153) 8189* MIPS microMIPS instruction choice selection: MIPS. (line 6) 8190* Motorola 68K GOT generation: M68K. (line 6) 8191* MRI compatibility: MRI. (line 6) 8192* MSP430 extra sections: MSP430. (line 11) 8193* NAME (MRI): MRI. (line 90) 8194* name, section: Output Section Name. 8195 (line 6) 8196* names: Symbols. (line 6) 8197* naming the output file: Options. (line 460) 8198* NEXT(EXP): Builtin Functions. (line 157) 8199* Nios II call relaxation: Nios II. (line 6) 8200* NMAGIC: Options. (line 439) 8201* NO_ENUM_SIZE_WARNING: ARM. (line 159) 8202* NO_WCHAR_SIZE_WARNING: ARM. (line 166) 8203* NOCROSSREFS(SECTIONS): Miscellaneous Commands. 8204 (line 82) 8205* NOLOAD: Output Section Type. 8206 (line 22) 8207* not enough room for program headers: Builtin Functions. (line 196) 8208* o =: MEMORY. (line 69) 8209* objdump -i: BFD. (line 6) 8210* object file management: BFD. (line 6) 8211* object files: Options. (line 29) 8212* object formats available: BFD. (line 6) 8213* object size: Options. (line 309) 8214* OMAGIC: Options. (line 454) 8215* ONLY_IF_RO: Output Section Constraint. 8216 (line 6) 8217* ONLY_IF_RW: Output Section Constraint. 8218 (line 6) 8219* opening object files: BFD outline. (line 6) 8220* operators for arithmetic: Operators. (line 6) 8221* options: Options. (line 6) 8222* ORDER (MRI): MRI. (line 95) 8223* org =: MEMORY. (line 69) 8224* ORIGIN =: MEMORY. (line 69) 8225* ORIGIN(MEMORY): Builtin Functions. (line 163) 8226* orphan: Orphan Sections. (line 6) 8227* orphan sections: Options. (line 606) 8228* output file after errors: Options. (line 1192) 8229* output file format in linker script: Format Commands. (line 10) 8230* output file name in linker script: File Commands. (line 66) 8231* output format: Options. (line 1083) 8232* output section alignment: Forced Output Alignment. 8233 (line 6) 8234* output section attributes: Output Section Attributes. 8235 (line 6) 8236* output section data: Output Section Data. 8237 (line 6) 8238* OUTPUT(FILENAME): File Commands. (line 66) 8239* OUTPUT_ARCH(BFDARCH): Miscellaneous Commands. 8240 (line 98) 8241* OUTPUT_FORMAT(BFDNAME): Format Commands. (line 10) 8242* OVERLAY: Overlay Description. 8243 (line 6) 8244* overlays: Overlay Description. 8245 (line 6) 8246* partial link: Options. (line 510) 8247* PE import table prefixing: ARM. (line 23) 8248* PHDRS: PHDRS. (line 6) 8249* PIC_VENEER: ARM. (line 172) 8250* pop state governing input file handling: Options. (line 494) 8251* position independent executables: Options. (line 1219) 8252* PowerPC ELF32 options: PowerPC ELF32. (line 16) 8253* PowerPC GOT: PowerPC ELF32. (line 33) 8254* PowerPC long branches: PowerPC ELF32. (line 6) 8255* PowerPC PLT: PowerPC ELF32. (line 16) 8256* PowerPC stub symbols: PowerPC ELF32. (line 47) 8257* PowerPC TLS optimization: PowerPC ELF32. (line 51) 8258* PowerPC64 __tls_get_addr optimization: PowerPC64 ELF64. (line 57) 8259* PowerPC64 dot symbols: PowerPC64 ELF64. (line 33) 8260* PowerPC64 ELF64 options: PowerPC64 ELF64. (line 6) 8261* PowerPC64 multi-TOC: PowerPC64 ELF64. (line 97) 8262* PowerPC64 OPD optimization: PowerPC64 ELF64. (line 71) 8263* PowerPC64 OPD spacing: PowerPC64 ELF64. (line 77) 8264* PowerPC64 PLT call stub static chain: PowerPC64 ELF64. (line 127) 8265* PowerPC64 PLT call stub thread safety: PowerPC64 ELF64. (line 133) 8266* PowerPC64 PLT stub alignment: PowerPC64 ELF64. (line 119) 8267* PowerPC64 register save/restore functions: PowerPC64 ELF64. 8268 (line 44) 8269* PowerPC64 stub grouping: PowerPC64 ELF64. (line 6) 8270* PowerPC64 stub symbols: PowerPC64 ELF64. (line 29) 8271* PowerPC64 TLS optimization: PowerPC64 ELF64. (line 52) 8272* PowerPC64 TOC optimization: PowerPC64 ELF64. (line 83) 8273* PowerPC64 TOC sorting: PowerPC64 ELF64. (line 109) 8274* precedence in expressions: Operators. (line 6) 8275* prevent unnecessary loading: Output Section Type. 8276 (line 22) 8277* program headers: PHDRS. (line 6) 8278* program headers and sections: Output Section Phdr. 8279 (line 6) 8280* program headers, not enough room: Builtin Functions. (line 196) 8281* program segments: PHDRS. (line 6) 8282* PROVIDE: PROVIDE. (line 6) 8283* PROVIDE_HIDDEN: PROVIDE_HIDDEN. (line 6) 8284* PUBLIC (MRI): MRI. (line 103) 8285* push state governing input file handling: Options. (line 476) 8286* QUAD(EXPRESSION): Output Section Data. 8287 (line 6) 8288* quoted symbol names: Symbols. (line 6) 8289* read-only text: Options. (line 439) 8290* read/write from cmd line: Options. (line 445) 8291* region alias: REGION_ALIAS. (line 6) 8292* region names: REGION_ALIAS. (line 6) 8293* REGION_ALIAS(ALIAS, REGION): REGION_ALIAS. (line 6) 8294* regions of memory: MEMORY. (line 6) 8295* relative expressions: Expression Section. (line 6) 8296* relaxing addressing modes: Options. (line 1233) 8297* relaxing on H8/300: H8/300. (line 9) 8298* relaxing on i960: i960. (line 31) 8299* relaxing on M68HC11: M68HC11/68HC12. (line 12) 8300* relaxing on NDS32: NDS32. (line 6) 8301* relaxing on Xtensa: Xtensa. (line 27) 8302* relocatable and absolute symbols: Expression Section. (line 6) 8303* relocatable output: Options. (line 510) 8304* removing sections: Output Section Discarding. 8305 (line 6) 8306* reporting bugs in ld: Reporting Bugs. (line 6) 8307* requirements for BFD: BFD. (line 16) 8308* retain relocations in final executable: Options. (line 497) 8309* retaining specified symbols: Options. (line 1259) 8310* rodata segment origin, cmd line: Options. (line 1435) 8311* ROM initialized data: Output Section LMA. (line 39) 8312* round up expression: Builtin Functions. (line 38) 8313* round up location counter: Builtin Functions. (line 38) 8314* runtime library name: Options. (line 316) 8315* runtime library search path: Options. (line 1273) 8316* runtime pseudo-relocation: WIN32. (line 217) 8317* scaled integers: Constants. (line 15) 8318* scommon section: Input Section Common. 8319 (line 20) 8320* script files: Options. (line 562) 8321* scripts: Scripts. (line 6) 8322* search directory, from cmd line: Options. (line 367) 8323* search path in linker script: File Commands. (line 76) 8324* SEARCH_DIR(PATH): File Commands. (line 76) 8325* SECT (MRI): MRI. (line 109) 8326* section address: Output Section Address. 8327 (line 6) 8328* section address in expression: Builtin Functions. (line 17) 8329* section alignment: Builtin Functions. (line 64) 8330* section alignment, warnings on: Options. (line 1584) 8331* section data: Output Section Data. 8332 (line 6) 8333* section fill pattern: Output Section Fill. 8334 (line 6) 8335* section load address: Output Section LMA. (line 6) 8336* section load address in expression: Builtin Functions. (line 143) 8337* section name: Output Section Name. 8338 (line 6) 8339* section name wildcard patterns: Input Section Wildcards. 8340 (line 6) 8341* section size: Builtin Functions. (line 175) 8342* section, assigning to memory region: Output Section Region. 8343 (line 6) 8344* section, assigning to program header: Output Section Phdr. 8345 (line 6) 8346* SECTIONS: SECTIONS. (line 6) 8347* sections, discarding: Output Section Discarding. 8348 (line 6) 8349* sections, orphan: Options. (line 606) 8350* segment origins, cmd line: Options. (line 1424) 8351* SEGMENT_START(SEGMENT, DEFAULT): Builtin Functions. (line 166) 8352* segments, ELF: PHDRS. (line 6) 8353* shared libraries: Options. (line 1352) 8354* SHORT(EXPRESSION): Output Section Data. 8355 (line 6) 8356* SIZEOF(SECTION): Builtin Functions. (line 175) 8357* SIZEOF_HEADERS: Builtin Functions. (line 191) 8358* small common symbols: Input Section Common. 8359 (line 20) 8360* SORT: Input Section Wildcards. 8361 (line 65) 8362* SORT_BY_ALIGNMENT: Input Section Wildcards. 8363 (line 54) 8364* SORT_BY_INIT_PRIORITY: Input Section Wildcards. 8365 (line 60) 8366* SORT_BY_NAME: Input Section Wildcards. 8367 (line 46) 8368* SORT_NONE: Input Section Wildcards. 8369 (line 106) 8370* SPU: SPU ELF. (line 46) 8371* SPU ELF options: SPU ELF. (line 6) 8372* SPU extra overlay stubs: SPU ELF. (line 19) 8373* SPU local store size: SPU ELF. (line 24) 8374* SPU overlay stub symbols: SPU ELF. (line 15) 8375* SPU overlays: SPU ELF. (line 9) 8376* SPU plugins: SPU ELF. (line 6) 8377* SQUAD(EXPRESSION): Output Section Data. 8378 (line 6) 8379* stack size: Options. (line 2106) 8380* standard Unix system: Options. (line 7) 8381* start of execution: Entry Point. (line 6) 8382* STARTUP(FILENAME): File Commands. (line 84) 8383* STM32L4xx erratum workaround: ARM. (line 121) 8384* strip all symbols: Options. (line 540) 8385* strip debugger symbols: Options. (line 544) 8386* stripping all but some symbols: Options. (line 1259) 8387* STUB_GROUP_SIZE: ARM. (line 177) 8388* SUBALIGN(SUBSECTION_ALIGN): Forced Input Alignment. 8389 (line 6) 8390* suffixes for integers: Constants. (line 15) 8391* symbol defaults: Builtin Functions. (line 123) 8392* symbol definition, scripts: Assignments. (line 6) 8393* symbol names: Symbols. (line 6) 8394* symbol tracing: Options. (line 657) 8395* symbol versions: VERSION. (line 6) 8396* symbol-only input: Options. (line 529) 8397* symbolic constants: Symbolic Constants. (line 6) 8398* symbols, from command line: Options. (line 997) 8399* symbols, relocatable and absolute: Expression Section. (line 6) 8400* symbols, require defined: Options. (line 588) 8401* symbols, retaining selectively: Options. (line 1259) 8402* synthesizing linker: Options. (line 1233) 8403* synthesizing on H8/300: H8/300. (line 14) 8404* TARGET(BFDNAME): Format Commands. (line 35) 8405* TARGET1: ARM. (line 33) 8406* TARGET2: ARM. (line 38) 8407* text segment origin, cmd line: Options. (line 1431) 8408* thumb entry point: ARM. (line 17) 8409* TI COFF versions: TI COFF. (line 6) 8410* traditional format: Options. (line 1403) 8411* trampoline generation on M68HC11: M68HC11/68HC12. (line 31) 8412* trampoline generation on M68HC12: M68HC11/68HC12. (line 31) 8413* unallocated address, next: Builtin Functions. (line 157) 8414* undefined symbol: Options. (line 575) 8415* undefined symbol in linker script: Miscellaneous Commands. 8416 (line 39) 8417* undefined symbols, warnings on: Options. (line 1580) 8418* uninitialized data placement: Input Section Common. 8419 (line 6) 8420* unspecified memory: Output Section Data. 8421 (line 39) 8422* usage: Options. (line 1101) 8423* USE_BLX: ARM. (line 75) 8424* using a DEF file: WIN32. (line 57) 8425* using auto-export functionality: WIN32. (line 22) 8426* Using decorations: WIN32. (line 162) 8427* variables, defining: Assignments. (line 6) 8428* verbose[=NUMBER]: Options. (line 1474) 8429* version: Options. (line 641) 8430* version script: VERSION. (line 6) 8431* version script, symbol versions: Options. (line 1482) 8432* VERSION {script text}: VERSION. (line 6) 8433* versions of symbols: VERSION. (line 6) 8434* VFP11_DENORM_FIX: ARM. (line 84) 8435* warnings, on combining symbols: Options. (line 1493) 8436* warnings, on section alignment: Options. (line 1584) 8437* warnings, on undefined symbols: Options. (line 1580) 8438* weak externals: WIN32. (line 407) 8439* what is this?: Overview. (line 6) 8440* wildcard file name patterns: Input Section Wildcards. 8441 (line 6) 8442* Xtensa options: Xtensa. (line 56) 8443* Xtensa processors: Xtensa. (line 6) 8444 8445 8446 8447Tag Table: 8448Node: Top710 8449Node: Overview1493 8450Node: Invocation2607 8451Node: Options3015 8452Node: Environment104154 8453Node: Scripts105914 8454Node: Basic Script Concepts107648 8455Node: Script Format110356 8456Node: Simple Example111219 8457Node: Simple Commands114315 8458Node: Entry Point114821 8459Node: File Commands115754 8460Node: Format Commands119874 8461Node: REGION_ALIAS121830 8462Node: Miscellaneous Commands126662 8463Node: Assignments131286 8464Node: Simple Assignments131797 8465Node: HIDDEN133532 8466Node: PROVIDE134162 8467Node: PROVIDE_HIDDEN135355 8468Node: Source Code Reference135599 8469Node: SECTIONS139515 8470Node: Output Section Description141406 8471Node: Output Section Name142650 8472Node: Output Section Address143526 8473Node: Input Section145761 8474Node: Input Section Basics146562 8475Node: Input Section Wildcards150468 8476Node: Input Section Common155675 8477Node: Input Section Keep157157 8478Node: Input Section Example157647 8479Node: Output Section Data158615 8480Node: Output Section Keywords161392 8481Node: Output Section Discarding164961 8482Node: Output Section Attributes166454 8483Node: Output Section Type167555 8484Node: Output Section LMA168626 8485Node: Forced Output Alignment171697 8486Node: Forced Input Alignment172127 8487Node: Output Section Constraint172516 8488Node: Output Section Region172944 8489Node: Output Section Phdr173377 8490Node: Output Section Fill174041 8491Node: Overlay Description175183 8492Node: MEMORY179629 8493Node: PHDRS184006 8494Node: VERSION189260 8495Node: Expressions197353 8496Node: Constants198282 8497Node: Symbolic Constants199157 8498Node: Symbols199708 8499Node: Orphan Sections200455 8500Node: Location Counter201792 8501Node: Operators206228 8502Node: Evaluation207150 8503Node: Expression Section208514 8504Node: Builtin Functions212378 8505Node: Implicit Linker Scripts220618 8506Node: Machine Dependent221393 8507Node: H8/300222517 8508Node: i960224579 8509Node: M68HC11/68HC12226275 8510Node: ARM227717 8511Node: HPPA ELF32238737 8512Node: M68K240360 8513Node: MIPS241269 8514Node: MMIX241793 8515Node: MSP430242958 8516Node: NDS32243998 8517Node: Nios II244964 8518Node: PowerPC ELF32246280 8519Node: PowerPC64 ELF64249111 8520Node: SPU ELF256443 8521Node: TI COFF259075 8522Node: WIN32259601 8523Node: Xtensa279727 8524Node: BFD282692 8525Node: BFD outline284147 8526Node: BFD information loss285433 8527Node: Canonical format287950 8528Node: Reporting Bugs292307 8529Node: Bug Criteria293001 8530Node: Bug Reporting293700 8531Node: MRI300739 8532Node: GNU Free Documentation License305382 8533Node: LD Index330538 8534 8535End Tag Table 8536