]>
Commit | Line | Data |
---|---|---|
d77de738 ML |
1 | \input texinfo.tex @c -*-texinfo-*- |
2 | @c @ifnothtml | |
3 | @c %**start of header | |
4 | @setfilename gccinstall.info | |
5 | @setchapternewpage odd | |
6 | @c %**end of header | |
7 | @c @end ifnothtml | |
8 | ||
9 | @include gcc-common.texi | |
10 | ||
11 | @c Specify title for specific html page | |
12 | @ifset indexhtml | |
13 | @settitle Installing GCC | |
14 | @end ifset | |
15 | @ifset specifichtml | |
16 | @settitle Host/Target specific installation notes for GCC | |
17 | @end ifset | |
18 | @ifset prerequisiteshtml | |
19 | @settitle Prerequisites for GCC | |
20 | @end ifset | |
21 | @ifset downloadhtml | |
22 | @settitle Downloading GCC | |
23 | @end ifset | |
24 | @ifset configurehtml | |
25 | @settitle Installing GCC: Configuration | |
26 | @end ifset | |
27 | @ifset buildhtml | |
28 | @settitle Installing GCC: Building | |
29 | @end ifset | |
30 | @ifset testhtml | |
31 | @settitle Installing GCC: Testing | |
32 | @end ifset | |
33 | @ifset finalinstallhtml | |
34 | @settitle Installing GCC: Final installation | |
35 | @end ifset | |
36 | @ifset binarieshtml | |
37 | @settitle Installing GCC: Binaries | |
38 | @end ifset | |
39 | @ifset gfdlhtml | |
40 | @settitle Installing GCC: GNU Free Documentation License | |
41 | @end ifset | |
42 | ||
74d5206f | 43 | @c Copyright (C) 1988-2023 Free Software Foundation, Inc. |
d77de738 ML |
44 | @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com |
45 | ||
46 | @c IMPORTANT: whenever you modify this file, run `install.texi2html' to | |
47 | @c test the generation of HTML documents for the gcc.gnu.org web pages. | |
48 | @c | |
49 | @c Do not use @footnote{} in this file as it breaks install.texi2html! | |
50 | ||
51 | @c Include everything if we're not making html | |
52 | @ifnothtml | |
53 | @set indexhtml | |
54 | @set specifichtml | |
55 | @set prerequisiteshtml | |
56 | @set downloadhtml | |
57 | @set configurehtml | |
58 | @set buildhtml | |
59 | @set testhtml | |
60 | @set finalinstallhtml | |
61 | @set binarieshtml | |
62 | @set gfdlhtml | |
63 | @end ifnothtml | |
64 | ||
65 | @c Part 2 Summary Description and Copyright | |
66 | @copying | |
74d5206f | 67 | Copyright @copyright{} 1988-2023 Free Software Foundation, Inc. |
d77de738 ML |
68 | @sp 1 |
69 | Permission is granted to copy, distribute and/or modify this document | |
70 | under the terms of the GNU Free Documentation License, Version 1.3 or | |
71 | any later version published by the Free Software Foundation; with no | |
72 | Invariant Sections, the Front-Cover texts being (a) (see below), and | |
73 | with the Back-Cover Texts being (b) (see below). A copy of the | |
74 | license is included in the section entitled ``@uref{./gfdl.html,,GNU | |
75 | Free Documentation License}''. | |
76 | ||
77 | (a) The FSF's Front-Cover Text is: | |
78 | ||
79 | A GNU Manual | |
80 | ||
81 | (b) The FSF's Back-Cover Text is: | |
82 | ||
83 | You have freedom to copy and modify this GNU Manual, like GNU | |
84 | software. Copies published by the Free Software Foundation raise | |
85 | funds for GNU development. | |
86 | @end copying | |
87 | @ifinfo | |
88 | @insertcopying | |
89 | @end ifinfo | |
90 | @dircategory Software development | |
91 | @direntry | |
92 | * gccinstall: (gccinstall). Installing the GNU Compiler Collection. | |
93 | @end direntry | |
94 | ||
95 | @c Part 3 Titlepage and Copyright | |
96 | @titlepage | |
97 | @title Installing GCC | |
98 | @versionsubtitle | |
99 | ||
100 | @c The following two commands start the copyright page. | |
101 | @page | |
102 | @vskip 0pt plus 1filll | |
103 | @insertcopying | |
104 | @end titlepage | |
105 | ||
106 | @c Part 4 Top node, Master Menu, and/or Table of Contents | |
107 | @ifinfo | |
108 | @node Top, , , (dir) | |
109 | @comment node-name, next, Previous, up | |
110 | ||
111 | @menu | |
112 | * Installing GCC:: This document describes the generic installation | |
113 | procedure for GCC as well as detailing some target | |
114 | specific installation instructions. | |
115 | ||
116 | * Specific:: Host/target specific installation notes for GCC. | |
117 | * Binaries:: Where to get pre-compiled binaries. | |
118 | ||
119 | * GNU Free Documentation License:: How you can copy and share this manual. | |
120 | * Concept Index:: This index has two entries. | |
121 | @end menu | |
122 | @end ifinfo | |
123 | ||
124 | @iftex | |
125 | @contents | |
126 | @end iftex | |
127 | ||
128 | @c Part 5 The Body of the Document | |
129 | @c ***Installing GCC********************************************************** | |
130 | @ifnothtml | |
131 | @comment node-name, next, previous, up | |
132 | @node Installing GCC, Binaries, , Top | |
133 | @end ifnothtml | |
134 | @ifset indexhtml | |
135 | @ifnothtml | |
136 | @chapter Installing GCC | |
137 | @end ifnothtml | |
138 | ||
139 | The latest version of this document is always available at | |
140 | @uref{https://gcc.gnu.org/install/,,https://gcc.gnu.org/install/}. | |
141 | It refers to the current development sources, instructions for | |
142 | specific released versions are included with the sources. | |
143 | ||
144 | This document describes the generic installation procedure for GCC as well | |
145 | as detailing some target specific installation instructions. | |
146 | ||
147 | GCC includes several components that previously were separate distributions | |
148 | with their own installation instructions. This document supersedes all | |
149 | package-specific installation instructions. | |
150 | ||
151 | @emph{Before} starting the build/install procedure please check the | |
152 | @ifnothtml | |
153 | @ref{Specific, host/target specific installation notes}. | |
154 | @end ifnothtml | |
155 | @ifhtml | |
156 | @uref{specific.html,,host/target specific installation notes}. | |
157 | @end ifhtml | |
158 | We recommend you browse the entire generic installation instructions before | |
159 | you proceed. | |
160 | ||
161 | Lists of successful builds for released versions of GCC are | |
162 | available at @uref{https://gcc.gnu.org/buildstat.html}. | |
163 | These lists are updated as new information becomes available. | |
164 | ||
165 | The installation procedure itself is broken into five steps. | |
166 | ||
167 | @ifinfo | |
168 | @menu | |
169 | * Prerequisites:: | |
170 | * Downloading the source:: | |
171 | * Configuration:: | |
172 | * Building:: | |
173 | * Testing:: (optional) | |
174 | * Final install:: | |
175 | @end menu | |
176 | @end ifinfo | |
177 | @ifhtml | |
178 | @enumerate | |
179 | @item | |
180 | @uref{prerequisites.html,,Prerequisites} | |
181 | @item | |
182 | @uref{download.html,,Downloading the source} | |
183 | @item | |
184 | @uref{configure.html,,Configuration} | |
185 | @item | |
186 | @uref{build.html,,Building} | |
187 | @item | |
188 | @uref{test.html,,Testing} (optional) | |
189 | @item | |
190 | @uref{finalinstall.html,,Final install} | |
191 | @end enumerate | |
192 | @end ifhtml | |
193 | ||
194 | Please note that GCC does not support @samp{make uninstall} and probably | |
195 | won't do so in the near future as this would open a can of worms. Instead, | |
196 | we suggest that you install GCC into a directory of its own and simply | |
197 | remove that directory when you do not need that specific version of GCC | |
198 | any longer, and, if shared libraries are installed there as well, no | |
199 | more binaries exist that use them. | |
200 | ||
201 | @html | |
202 | <hr /> | |
203 | <p> | |
204 | @end html | |
205 | @ifhtml | |
206 | @uref{./index.html,,Return to the GCC Installation page} | |
207 | ||
208 | @insertcopying | |
209 | @end ifhtml | |
210 | @end ifset | |
211 | ||
212 | @c ***Prerequisites************************************************** | |
213 | @ifnothtml | |
214 | @comment node-name, next, previous, up | |
215 | @node Prerequisites, Downloading the source, , Installing GCC | |
216 | @end ifnothtml | |
217 | @ifset prerequisiteshtml | |
218 | @ifnothtml | |
219 | @chapter Prerequisites | |
220 | @end ifnothtml | |
221 | @cindex Prerequisites | |
222 | ||
223 | GCC requires that various tools and packages be available for use in the | |
224 | build procedure. Modifying GCC sources requires additional tools | |
225 | described below. | |
226 | ||
227 | @heading Tools/packages necessary for building GCC | |
228 | @table @asis | |
229 | @item ISO C++11 compiler | |
b97c33fb RB |
230 | Necessary to bootstrap GCC. GCC 4.8.3 or newer has sufficient |
231 | support for used C++11 features, with earlier GCC versions you | |
232 | might run into implementation bugs. | |
d77de738 ML |
233 | |
234 | Versions of GCC prior to 11 also allow bootstrapping with an ISO C++98 | |
235 | compiler, versions of GCC prior to 4.8 also allow bootstrapping with a | |
236 | ISO C89 compiler, and versions of GCC prior to 3.4 also allow | |
237 | bootstrapping with a traditional (K&R) C compiler. | |
238 | ||
239 | To build all languages in a cross-compiler or other configuration where | |
240 | 3-stage bootstrap is not performed, you need to start with an existing | |
b97c33fb | 241 | GCC binary (version 4.8.3 or later) because source code for language |
d77de738 ML |
242 | frontends other than C might use GCC extensions. |
243 | ||
244 | @item C standard library and headers | |
245 | ||
246 | In order to build GCC, the C standard library and headers must be present | |
247 | for all target variants for which target libraries will be built (and not | |
248 | only the variant of the host C++ compiler). | |
249 | ||
250 | This affects the popular @samp{x86_64-pc-linux-gnu} platform (among | |
251 | other multilib targets), for which 64-bit (@samp{x86_64}) and 32-bit | |
252 | (@samp{i386}) libc headers are usually packaged separately. If you do a | |
253 | build of a native compiler on @samp{x86_64-pc-linux-gnu}, make sure you | |
254 | either have the 32-bit libc developer package properly installed (the exact | |
255 | name of the package depends on your distro) or you must build GCC as a | |
256 | 64-bit only compiler by configuring with the option | |
257 | @option{--disable-multilib}. Otherwise, you may encounter an error such as | |
258 | @samp{fatal error: gnu/stubs-32.h: No such file} | |
259 | ||
260 | @item @anchor{GNAT-prerequisite}GNAT | |
261 | ||
262 | In order to build GNAT, the Ada compiler, you need a working GNAT | |
263 | compiler (GCC version 5.1 or later). | |
264 | ||
265 | This includes GNAT tools such as @command{gnatmake} and | |
266 | @command{gnatlink}, since the Ada front end is written in Ada and | |
267 | uses some GNAT-specific extensions. | |
268 | ||
269 | In order to build a cross compiler, it is strongly recommended to install | |
270 | the new compiler as native first, and then use it to build the cross | |
271 | compiler. Other native compiler versions may work but this is not guaranteed and | |
272 | will typically fail with hard to understand compilation errors during the | |
273 | build. | |
274 | ||
275 | Similarly, it is strongly recommended to use an older version of GNAT to build | |
276 | GNAT. More recent versions of GNAT than the version built are not guaranteed | |
277 | to work and will often fail during the build with compilation errors. | |
278 | ||
279 | Note that @command{configure} does not test whether the GNAT installation works | |
280 | and has a sufficiently recent version; if too old a GNAT version is | |
281 | installed and @option{--enable-languages=ada} is used, the build will fail. | |
282 | ||
283 | @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables | |
284 | must not be set when building the Ada compiler, the Ada tools, or the | |
285 | Ada runtime libraries. You can check that your build environment is clean | |
286 | by verifying that @samp{gnatls -v} lists only one explicit path in each | |
287 | section. | |
288 | ||
289 | @item @anchor{GDC-prerequisite}GDC | |
290 | ||
291 | In order to build GDC, the D compiler, you need a working GDC | |
05b7cf52 | 292 | compiler (GCC version 9.4 or later) and D runtime library, |
d77de738 ML |
293 | @samp{libphobos}, as the D front end is written in D. |
294 | ||
295 | Versions of GDC prior to 12 can be built with an ISO C++11 compiler, which can | |
296 | then be installed and used to bootstrap newer versions of the D front end. | |
297 | ||
298 | It is strongly recommended to use an older version of GDC to build GDC. More | |
299 | recent versions of GDC than the version built are not guaranteed to work and | |
300 | will often fail during the build with compilation errors relating to | |
301 | deprecations or removed features. | |
302 | ||
303 | Note that @command{configure} does not test whether the GDC installation works | |
304 | and has a sufficiently recent version. Though the implementation of the D | |
305 | front end does not make use of any GDC-specific extensions, or novel features | |
306 | of the D language, if too old a GDC version is installed and | |
307 | @option{--enable-languages=d} is used, the build will fail. | |
308 | ||
309 | On some targets, @samp{libphobos} isn't enabled by default, but compiles | |
310 | and works if @option{--enable-libphobos} is used. Specifics are | |
311 | documented for affected targets. | |
312 | ||
1eee94d3 GM |
313 | @item @anchor{GM2-prerequisite}GM2 |
314 | ||
315 | Python3 is required if you want to build the complete Modula-2 | |
316 | documentation including the target @code{SYSTEM} definition module. | |
317 | If Python3 is unavailable Modula-2 documentation will include a target | |
318 | independent version of the SYSTEM modules. | |
319 | ||
d77de738 ML |
320 | @item A ``working'' POSIX compatible shell, or GNU bash |
321 | ||
322 | Necessary when running @command{configure} because some | |
323 | @command{/bin/sh} shells have bugs and may crash when configuring the | |
324 | target libraries. In other cases, @command{/bin/sh} or @command{ksh} | |
325 | have disastrous corner-case performance problems. This | |
326 | can cause target @command{configure} runs to literally take days to | |
327 | complete in some cases. | |
328 | ||
329 | So on some platforms @command{/bin/ksh} is sufficient, on others it | |
330 | isn't. See the host/target specific instructions for your platform, or | |
331 | use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your | |
332 | environment to your ``good'' shell prior to running | |
333 | @command{configure}/@command{make}. | |
334 | ||
335 | @command{zsh} is not a fully compliant POSIX shell and will not | |
336 | work when configuring GCC@. | |
337 | ||
338 | @item A POSIX or SVR4 awk | |
339 | ||
340 | Necessary for creating some of the generated source files for GCC@. | |
341 | If in doubt, use a recent GNU awk version, as some of the older ones | |
342 | are broken. GNU awk version 3.1.5 is known to work. | |
343 | ||
344 | @item GNU binutils | |
345 | ||
346 | Necessary in some circumstances, optional in others. See the | |
347 | host/target specific instructions for your platform for the exact | |
348 | requirements. | |
349 | ||
350 | Note binutils 2.35 or newer is required for LTO to work correctly | |
351 | with GNU libtool that includes doing a bootstrap with LTO enabled. | |
352 | ||
353 | @item gzip version 1.2.4 (or later) or | |
354 | @itemx bzip2 version 1.0.2 (or later) | |
355 | ||
356 | Necessary to uncompress GCC @command{tar} files when source code is | |
357 | obtained via HTTPS mirror sites. | |
358 | ||
359 | @item GNU make version 3.80 (or later) | |
360 | ||
361 | You must have GNU make installed to build GCC@. | |
362 | ||
363 | @item GNU tar version 1.14 (or later) | |
364 | ||
365 | Necessary (only on some platforms) to untar the source code. Many | |
366 | systems' @command{tar} programs will also work, only try GNU | |
367 | @command{tar} if you have problems. | |
368 | ||
369 | @item Perl version between 5.6.1 and 5.6.24 | |
370 | ||
371 | Necessary when targeting Darwin, building @samp{libstdc++}, | |
372 | and not using @option{--disable-symvers}. | |
373 | Necessary when targeting Solaris 2 with Solaris @command{ld} and not using | |
374 | @option{--disable-symvers}. | |
375 | ||
376 | Necessary when regenerating @file{Makefile} dependencies in libiberty. | |
377 | Necessary when regenerating @file{libiberty/functions.texi}. | |
378 | Necessary when generating manpages from Texinfo manuals. | |
379 | Used by various scripts to generate some files included in the source | |
380 | repository (mainly Unicode-related and rarely changing) from source | |
381 | tables. | |
382 | ||
383 | Used by @command{automake}. | |
384 | ||
385 | @end table | |
386 | ||
387 | Several support libraries are necessary to build GCC, some are required, | |
388 | others optional. While any sufficiently new version of required tools | |
389 | usually work, library requirements are generally stricter. Newer | |
390 | versions may work in some cases, but it's safer to use the exact | |
391 | versions documented. We appreciate bug reports about problems with | |
392 | newer versions, though. If your OS vendor provides packages for the | |
393 | support libraries then using those packages may be the simplest way to | |
394 | install the libraries. | |
395 | ||
396 | @table @asis | |
397 | @item GNU Multiple Precision Library (GMP) version 4.3.2 (or later) | |
398 | ||
399 | Necessary to build GCC@. If a GMP source distribution is found in a | |
400 | subdirectory of your GCC sources named @file{gmp}, it will be built | |
401 | together with GCC. Alternatively, if GMP is already installed but it | |
402 | is not in your library search path, you will have to configure with the | |
403 | @option{--with-gmp} configure option. See also @option{--with-gmp-lib} | |
404 | and @option{--with-gmp-include}. | |
405 | The in-tree build is only supported with the GMP version that | |
406 | download_prerequisites installs. | |
407 | ||
408 | @item MPFR Library version 3.1.0 (or later) | |
409 | ||
410 | Necessary to build GCC@. It can be downloaded from | |
411 | @uref{https://www.mpfr.org}. If an MPFR source distribution is found | |
412 | in a subdirectory of your GCC sources named @file{mpfr}, it will be | |
413 | built together with GCC. Alternatively, if MPFR is already installed | |
414 | but it is not in your default library search path, the | |
415 | @option{--with-mpfr} configure option should be used. See also | |
416 | @option{--with-mpfr-lib} and @option{--with-mpfr-include}. | |
417 | The in-tree build is only supported with the MPFR version that | |
418 | download_prerequisites installs. | |
419 | ||
420 | @item MPC Library version 1.0.1 (or later) | |
421 | ||
422 | Necessary to build GCC@. It can be downloaded from | |
423 | @uref{https://www.multiprecision.org/mpc/}. If an MPC source distribution | |
424 | is found in a subdirectory of your GCC sources named @file{mpc}, it | |
425 | will be built together with GCC. Alternatively, if MPC is already | |
426 | installed but it is not in your default library search path, the | |
427 | @option{--with-mpc} configure option should be used. See also | |
428 | @option{--with-mpc-lib} and @option{--with-mpc-include}. | |
429 | The in-tree build is only supported with the MPC version that | |
430 | download_prerequisites installs. | |
431 | ||
432 | @item isl Library version 0.15 or later. | |
433 | ||
434 | Necessary to build GCC with the Graphite loop optimizations. | |
435 | It can be downloaded from @uref{https://gcc.gnu.org/pub/gcc/infrastructure/}. | |
436 | If an isl source distribution is found | |
437 | in a subdirectory of your GCC sources named @file{isl}, it will be | |
438 | built together with GCC. Alternatively, the @option{--with-isl} configure | |
439 | option should be used if isl is not installed in your default library | |
440 | search path. | |
441 | ||
442 | @item zstd Library. | |
443 | ||
444 | Necessary to build GCC with zstd compression used for LTO bytecode. | |
445 | The library is searched in your default library patch search. | |
446 | Alternatively, the @option{--with-zstd} configure option should be used. | |
447 | ||
1eee94d3 GM |
448 | @item Python3 modules |
449 | ||
450 | The complete list of Python3 modules broken down by GCC subcomponent | |
451 | is shown below: | |
452 | ||
453 | @table @asis | |
454 | @item internal debugging in gdbhooks | |
455 | @code{gdb}, @code{gdb.printing}, @code{gdb.types}, | |
456 | @code{os.path}, @code{re}, @code{sys} and @code{tempfile}, | |
457 | ||
458 | @item g++ testsuite | |
459 | @code{gcov}, @code{gzip}, @code{json}, @code{os} and @code{pytest}. | |
460 | ||
461 | @item c++ cxx api generation | |
462 | @code{csv}, @code{os}, @code{sys} and @code{time}. | |
463 | ||
464 | @item modula-2 documentation | |
465 | @code{argparse}, @code{os}, @code{pathlib}, @code{shutil} and | |
466 | @code{sys}. | |
467 | ||
468 | @item git developer tools | |
469 | @code{os} and @code{sys}. | |
470 | ||
471 | @item ada documentation | |
472 | @code{latex_elements}, @code{os}, @code{pygments}, @code{re}, | |
473 | @code{sys} and @code{time}. | |
474 | @end table | |
475 | ||
d77de738 ML |
476 | @end table |
477 | ||
478 | @heading Tools/packages necessary for modifying GCC | |
479 | @table @asis | |
480 | @item autoconf version 2.69 | |
481 | @itemx GNU m4 version 1.4.6 (or later) | |
482 | ||
483 | Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@: | |
484 | to regenerate @file{configure} and @file{config.in} files. | |
485 | ||
486 | @item automake version 1.15.1 | |
487 | ||
488 | Necessary when modifying a @file{Makefile.am} file to regenerate its | |
489 | associated @file{Makefile.in}. | |
490 | ||
491 | Much of GCC does not use automake, so directly edit the @file{Makefile.in} | |
492 | file. Specifically this applies to the @file{gcc}, @file{intl}, | |
493 | @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well | |
494 | as any of their subdirectories. | |
495 | ||
496 | For directories that use automake, GCC requires the latest release in | |
497 | the 1.15 series, which is currently 1.15.1. When regenerating a directory | |
498 | to a newer version, please update all the directories using an older 1.15 | |
499 | to the latest released version. | |
500 | ||
501 | @item gettext version 0.14.5 (or later) | |
502 | ||
503 | Needed to regenerate @file{gcc.pot}. | |
504 | ||
505 | @item gperf version 2.7.2 (or later) | |
506 | ||
507 | Necessary when modifying @command{gperf} input files, e.g.@: | |
508 | @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@: | |
509 | @file{gcc/cp/cfns.h}. | |
510 | ||
511 | @item DejaGnu version 1.5.3 (or later) | |
512 | @itemx Expect | |
513 | @itemx Tcl | |
514 | @c Once Tcl 8.5 or higher is required, remove any obsolete | |
515 | @c compatibility workarounds: | |
516 | @c git grep 'compatibility with earlier Tcl releases' | |
517 | ||
518 | Necessary to run the GCC testsuite; see the section on testing for | |
519 | details. | |
520 | ||
521 | @item autogen version 5.5.4 (or later) and | |
522 | @itemx guile version 1.4.1 (or later) | |
523 | ||
524 | Necessary to regenerate @file{fixinc/fixincl.x} from | |
525 | @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}. | |
526 | ||
527 | Necessary to run @samp{make check} for @file{fixinc}. | |
528 | ||
529 | Necessary to regenerate the top level @file{Makefile.in} file from | |
530 | @file{Makefile.tpl} and @file{Makefile.def}. | |
531 | ||
532 | @item Flex version 2.5.4 (or later) | |
533 | ||
534 | Necessary when modifying @file{*.l} files. | |
535 | ||
536 | Necessary to build GCC during development because the generated output | |
537 | files are not included in the version-controlled source repository. | |
538 | They are included in releases. | |
539 | ||
540 | @item Texinfo version 4.7 (or later) | |
541 | ||
542 | Necessary for running @command{makeinfo} when modifying @file{*.texi} | |
543 | files to test your changes. | |
544 | ||
545 | Necessary for running @command{make dvi} or @command{make pdf} to | |
546 | create printable documentation in DVI or PDF format. Texinfo version | |
547 | 4.8 or later is required for @command{make pdf}. | |
548 | ||
549 | Necessary to build GCC documentation during development because the | |
550 | generated output files are not included in the repository. They are | |
551 | included in releases. | |
552 | ||
553 | @item @TeX{} (any working version) | |
554 | ||
555 | Necessary for running @command{texi2dvi} and @command{texi2pdf}, which | |
556 | are used when running @command{make dvi} or @command{make pdf} to create | |
557 | DVI or PDF files, respectively. | |
558 | ||
559 | @item Sphinx version 1.0 (or later) | |
560 | ||
561 | Necessary to regenerate @file{jit/docs/_build/texinfo} from the @file{.rst} | |
562 | files in the directories below @file{jit/docs}. | |
563 | ||
564 | @item git (any version) | |
565 | @itemx SSH (any version) | |
566 | ||
567 | Necessary to access the source repository. Public releases and weekly | |
568 | snapshots of the development sources are also available via HTTPS@. | |
569 | ||
570 | @item GNU diffutils version 2.7 (or later) | |
571 | ||
572 | Useful when submitting patches for the GCC source code. | |
573 | ||
574 | @item patch version 2.5.4 (or later) | |
575 | ||
576 | Necessary when applying patches, created with @command{diff}, to one's | |
577 | own sources. | |
578 | ||
579 | @end table | |
580 | ||
581 | @html | |
582 | <hr /> | |
583 | <p> | |
584 | @end html | |
585 | @ifhtml | |
586 | @uref{./index.html,,Return to the GCC Installation page} | |
587 | @end ifhtml | |
588 | @end ifset | |
589 | ||
590 | @c ***Downloading the source************************************************** | |
591 | @ifnothtml | |
592 | @comment node-name, next, previous, up | |
593 | @node Downloading the source, Configuration, Prerequisites, Installing GCC | |
594 | @end ifnothtml | |
595 | @ifset downloadhtml | |
596 | @ifnothtml | |
597 | @chapter Downloading GCC | |
598 | @end ifnothtml | |
599 | @cindex Downloading GCC | |
600 | @cindex Downloading the Source | |
601 | ||
602 | GCC is distributed via @uref{https://gcc.gnu.org/git.html,,git} and via | |
603 | HTTPS as tarballs compressed with @command{gzip} or @command{bzip2}. | |
604 | ||
605 | Please refer to the @uref{https://gcc.gnu.org/releases.html,,releases web page} | |
606 | for information on how to obtain GCC@. | |
607 | ||
608 | The source distribution includes the C, C++, Objective-C, Fortran, | |
609 | and Ada (in the case of GCC 3.1 and later) compilers, as well as | |
610 | runtime libraries for C++, Objective-C, and Fortran. | |
611 | For previous versions these were downloadable as separate components such | |
612 | as the core GCC distribution, which included the C language front end and | |
613 | shared components, and language-specific distributions including the | |
614 | language front end and the language runtime (where appropriate). | |
615 | ||
616 | If you also intend to build binutils (either to upgrade an existing | |
617 | installation or for use in place of the corresponding tools of your | |
618 | OS), unpack the binutils distribution either in the same directory or | |
619 | a separate one. In the latter case, add symbolic links to any | |
620 | components of the binutils you intend to build alongside the compiler | |
621 | (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld}, | |
622 | @file{opcodes}, @dots{}) to the directory containing the GCC sources. | |
623 | ||
624 | Likewise the GMP, MPFR and MPC libraries can be automatically built | |
625 | together with GCC. You may simply run the | |
626 | @command{contrib/download_prerequisites} script in the GCC source directory | |
627 | to set up everything. | |
628 | Otherwise unpack the GMP, MPFR and/or MPC source | |
629 | distributions in the directory containing the GCC sources and rename | |
630 | their directories to @file{gmp}, @file{mpfr} and @file{mpc}, | |
631 | respectively (or use symbolic links with the same name). | |
632 | ||
633 | @html | |
634 | <hr /> | |
635 | <p> | |
636 | @end html | |
637 | @ifhtml | |
638 | @uref{./index.html,,Return to the GCC Installation page} | |
639 | @end ifhtml | |
640 | @end ifset | |
641 | ||
642 | @c ***Configuration*********************************************************** | |
643 | @ifnothtml | |
644 | @comment node-name, next, previous, up | |
645 | @node Configuration, Building, Downloading the source, Installing GCC | |
646 | @end ifnothtml | |
647 | @ifset configurehtml | |
648 | @ifnothtml | |
649 | @chapter Installing GCC: Configuration | |
650 | @end ifnothtml | |
651 | @cindex Configuration | |
652 | @cindex Installing GCC: Configuration | |
653 | ||
654 | Like most GNU software, GCC must be configured before it can be built. | |
655 | This document describes the recommended configuration procedure | |
656 | for both native and cross targets. | |
657 | ||
658 | We use @var{srcdir} to refer to the toplevel source directory for | |
659 | GCC; we use @var{objdir} to refer to the toplevel build/object directory. | |
660 | ||
661 | If you obtained the sources by cloning the repository, @var{srcdir} | |
662 | must refer to the top @file{gcc} directory, the one where the | |
663 | @file{MAINTAINERS} file can be found, and not its @file{gcc} | |
664 | subdirectory, otherwise the build will fail. | |
665 | ||
666 | If either @var{srcdir} or @var{objdir} is located on an automounted NFS | |
667 | file system, the shell's built-in @command{pwd} command will return | |
668 | temporary pathnames. Using these can lead to various sorts of build | |
669 | problems. To avoid this issue, set the @env{PWDCMD} environment | |
670 | variable to an automounter-aware @command{pwd} command, e.g., | |
671 | @command{pawd} or @samp{amq -w}, during the configuration and build | |
672 | phases. | |
673 | ||
674 | First, we @strong{highly} recommend that GCC be built into a | |
675 | separate directory from the sources which does @strong{not} reside | |
676 | within the source tree. This is how we generally build GCC; building | |
677 | where @var{srcdir} == @var{objdir} should still work, but doesn't | |
678 | get extensive testing; building where @var{objdir} is a subdirectory | |
679 | of @var{srcdir} is unsupported. | |
680 | ||
681 | If you have previously built GCC in the same directory for a | |
682 | different target machine, do @samp{make distclean} to delete all files | |
683 | that might be invalid. One of the files this deletes is @file{Makefile}; | |
684 | if @samp{make distclean} complains that @file{Makefile} does not exist | |
685 | or issues a message like ``don't know how to make distclean'' it probably | |
686 | means that the directory is already suitably clean. However, with the | |
687 | recommended method of building in a separate @var{objdir}, you should | |
688 | simply use a different @var{objdir} for each target. | |
689 | ||
690 | Second, when configuring a native system, either @command{cc} or | |
691 | @command{gcc} must be in your path or you must set @env{CC} in | |
692 | your environment before running configure. Otherwise the configuration | |
693 | scripts may fail. | |
694 | ||
695 | @ignore | |
696 | Note that the bootstrap compiler and the resulting GCC must be link | |
697 | compatible, else the bootstrap will fail with linker errors about | |
698 | incompatible object file formats. Several multilibed targets are | |
699 | affected by this requirement, see | |
700 | @ifnothtml | |
701 | @ref{Specific, host/target specific installation notes}. | |
702 | @end ifnothtml | |
703 | @ifhtml | |
704 | @uref{specific.html,,host/target specific installation notes}. | |
705 | @end ifhtml | |
706 | @end ignore | |
707 | ||
708 | To configure GCC: | |
709 | ||
710 | @smallexample | |
711 | % mkdir @var{objdir} | |
712 | % cd @var{objdir} | |
713 | % @var{srcdir}/configure [@var{options}] [@var{target}] | |
714 | @end smallexample | |
715 | ||
716 | @heading Distributor options | |
717 | ||
718 | If you will be distributing binary versions of GCC, with modifications | |
719 | to the source code, you should use the options described in this | |
720 | section to make clear that your version contains modifications. | |
721 | ||
722 | @table @code | |
723 | @item --with-pkgversion=@var{version} | |
724 | Specify a string that identifies your package. You may wish | |
725 | to include a build number or build date. This version string will be | |
726 | included in the output of @command{gcc --version}. This suffix does | |
727 | not replace the default version string, only the @samp{GCC} part. | |
728 | ||
729 | The default value is @samp{GCC}. | |
730 | ||
731 | @item --with-bugurl=@var{url} | |
732 | Specify the URL that users should visit if they wish to report a bug. | |
733 | You are of course welcome to forward bugs reported to you to the FSF, | |
734 | if you determine that they are not bugs in your modifications. | |
735 | ||
736 | The default value refers to the FSF's GCC bug tracker. | |
737 | ||
738 | @item --with-documentation-root-url=@var{url} | |
739 | Specify the URL root that contains GCC option documentation. The @var{url} | |
740 | should end with a @code{/} character. | |
741 | ||
742 | The default value is @uref{https://gcc.gnu.org/onlinedocs/,,https://gcc.gnu.org/onlinedocs/}. | |
743 | ||
744 | @item --with-changes-root-url=@var{url} | |
745 | Specify the URL root that contains information about changes in GCC | |
746 | releases like @code{gcc-@var{version}/changes.html}. | |
747 | The @var{url} should end with a @code{/} character. | |
748 | ||
749 | The default value is @uref{https://gcc.gnu.org/,,https://gcc.gnu.org/}. | |
750 | ||
751 | @end table | |
752 | ||
753 | @heading Host, Build and Target specification | |
754 | ||
755 | Specify the host, build and target machine configurations. You do this | |
756 | when you run the @file{configure} script. | |
757 | ||
758 | The @dfn{build} machine is the system which you are using, the | |
759 | @dfn{host} machine is the system where you want to run the resulting | |
760 | compiler (normally the build machine), and the @dfn{target} machine is | |
761 | the system for which you want the compiler to generate code. | |
762 | ||
763 | If you are building a compiler to produce code for the machine it runs | |
764 | on (a native compiler), you normally do not need to specify any operands | |
765 | to @file{configure}; it will try to guess the type of machine you are on | |
766 | and use that as the build, host and target machines. So you don't need | |
767 | to specify a configuration when building a native compiler unless | |
768 | @file{configure} cannot figure out what your configuration is or guesses | |
769 | wrong. | |
770 | ||
771 | In those cases, specify the build machine's @dfn{configuration name} | |
772 | with the @option{--host} option; the host and target will default to be | |
773 | the same as the host machine. | |
774 | ||
775 | Here is an example: | |
776 | ||
777 | @smallexample | |
778 | ./configure --host=x86_64-pc-linux-gnu | |
779 | @end smallexample | |
780 | ||
781 | A configuration name may be canonical or it may be more or less | |
782 | abbreviated (@file{config.sub} script produces canonical versions). | |
783 | ||
784 | A canonical configuration name has three parts, separated by dashes. | |
785 | It looks like this: @samp{@var{cpu}-@var{company}-@var{system}}. | |
786 | ||
787 | Here are the possible CPU types: | |
788 | ||
789 | @quotation | |
790 | aarch64, aarch64_be, alpha, alpha64, amdgcn, arc, arceb, arm, armeb, avr, bfin, | |
791 | bpf, cris, csky, epiphany, fido, fr30, frv, ft32, h8300, hppa, hppa2.0, | |
792 | hppa64, i486, i686, ia64, iq2000, lm32, loongarch64, m32c, m32r, m32rle, m68k, | |
793 | mcore, microblaze, microblazeel, mips, mips64, mips64el, mips64octeon, | |
794 | mips64orion, mips64vr, mipsel, mipsisa32, mipsisa32r2, mipsisa64, mipsisa64r2, | |
795 | mipsisa64r2el, mipsisa64sb1, mipsisa64sr71k, mipstx39, mmix, mn10300, moxie, | |
796 | msp430, nds32be, nds32le, nios2, nvptx, or1k, pdp11, powerpc, powerpc64, | |
797 | powerpc64le, powerpcle, pru, riscv32, riscv32be, riscv64, riscv64be, rl78, rx, | |
798 | s390, s390x, sh, shle, sparc, sparc64, tic6x, v850, | |
799 | v850e, v850e1, vax, visium, x86_64, xstormy16, xtensa | |
800 | @end quotation | |
801 | ||
802 | Here is a list of system types: | |
803 | ||
804 | @quotation | |
805 | aix@var{version}, amdhsa, aout, cygwin, darwin@var{version}, | |
806 | eabi, eabialtivec, eabisim, eabisimaltivec, elf, elf32, | |
807 | elfbare, elfoabi, freebsd@var{version}, gnu, hpux, hpux@var{version}, | |
808 | kfreebsd-gnu, kopensolaris-gnu, linux-androideabi, linux-gnu, | |
809 | linux-gnu_altivec, linux-musl, linux-uclibc, lynxos, mingw32, mingw32crt, | |
810 | mmixware, msdosdjgpp, netbsd, netbsdelf@var{version}, nto-qnx, openbsd, | |
811 | rtems, solaris@var{version}, symbianelf, tpf, uclinux, uclinux_eabi, vms, | |
812 | vxworks, vxworksae, vxworksmils | |
813 | @end quotation | |
814 | ||
815 | @heading Options specification | |
816 | ||
817 | Use @var{options} to override several configure time options for | |
818 | GCC@. A list of supported @var{options} follows; @samp{configure | |
819 | --help} may list other options, but those not listed below may not | |
820 | work and should not normally be used. | |
821 | ||
822 | Note that each @option{--enable} option has a corresponding | |
823 | @option{--disable} option and that each @option{--with} option has a | |
824 | corresponding @option{--without} option. | |
825 | ||
826 | @table @code | |
827 | @item --prefix=@var{dirname} | |
828 | Specify the toplevel installation | |
829 | directory. This is the recommended way to install the tools into a directory | |
830 | other than the default. The toplevel installation directory defaults to | |
831 | @file{/usr/local}. | |
832 | ||
833 | We @strong{highly} recommend against @var{dirname} being the same or a | |
834 | subdirectory of @var{objdir} or vice versa. If specifying a directory | |
835 | beneath a user's home directory tree, some shells will not expand | |
836 | @var{dirname} correctly if it contains the @samp{~} metacharacter; use | |
837 | @env{$HOME} instead. | |
838 | ||
839 | The following standard @command{autoconf} options are supported. Normally you | |
840 | should not need to use these options. | |
841 | @table @code | |
842 | @item --exec-prefix=@var{dirname} | |
843 | Specify the toplevel installation directory for architecture-dependent | |
844 | files. The default is @file{@var{prefix}}. | |
845 | ||
846 | @item --bindir=@var{dirname} | |
847 | Specify the installation directory for the executables called by users | |
848 | (such as @command{gcc} and @command{g++}). The default is | |
849 | @file{@var{exec-prefix}/bin}. | |
850 | ||
851 | @item --libdir=@var{dirname} | |
852 | Specify the installation directory for object code libraries and | |
853 | internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}. | |
854 | ||
855 | @item --libexecdir=@var{dirname} | |
856 | Specify the installation directory for internal executables of GCC@. | |
857 | The default is @file{@var{exec-prefix}/libexec}. | |
858 | ||
859 | @item --with-slibdir=@var{dirname} | |
860 | Specify the installation directory for the shared libgcc library. The | |
861 | default is @file{@var{libdir}}. | |
862 | ||
863 | @item --datarootdir=@var{dirname} | |
864 | Specify the root of the directory tree for read-only architecture-independent | |
865 | data files referenced by GCC@. The default is @file{@var{prefix}/share}. | |
866 | ||
867 | @item --infodir=@var{dirname} | |
868 | Specify the installation directory for documentation in info format. | |
869 | The default is @file{@var{datarootdir}/info}. | |
870 | ||
871 | @item --datadir=@var{dirname} | |
872 | Specify the installation directory for some architecture-independent | |
873 | data files referenced by GCC@. The default is @file{@var{datarootdir}}. | |
874 | ||
875 | @item --docdir=@var{dirname} | |
876 | Specify the installation directory for documentation files (other | |
877 | than Info) for GCC@. The default is @file{@var{datarootdir}/doc}. | |
878 | ||
879 | @item --htmldir=@var{dirname} | |
880 | Specify the installation directory for HTML documentation files. | |
881 | The default is @file{@var{docdir}}. | |
882 | ||
883 | @item --pdfdir=@var{dirname} | |
884 | Specify the installation directory for PDF documentation files. | |
885 | The default is @file{@var{docdir}}. | |
886 | ||
887 | @item --mandir=@var{dirname} | |
888 | Specify the installation directory for manual pages. The default is | |
889 | @file{@var{datarootdir}/man}. (Note that the manual pages are only extracts | |
890 | from the full GCC manuals, which are provided in Texinfo format. The manpages | |
891 | are derived by an automatic conversion process from parts of the full | |
892 | manual.) | |
893 | ||
894 | @item --with-gxx-include-dir=@var{dirname} | |
895 | Specify | |
896 | the installation directory for G++ header files. The default depends | |
897 | on other configuration options, and differs between cross and native | |
898 | configurations. | |
899 | ||
900 | @item --with-specs=@var{specs} | |
901 | Specify additional command line driver SPECS. | |
902 | This can be useful if you need to turn on a non-standard feature by | |
903 | default without modifying the compiler's source code, for instance | |
904 | @option{--with-specs=%@{!fcommon:%@{!fno-common:-fno-common@}@}}. | |
905 | @ifnothtml | |
906 | @xref{Spec Files,, Specifying subprocesses and the switches to pass to them, | |
907 | gcc, Using the GNU Compiler Collection (GCC)}, | |
908 | @end ifnothtml | |
909 | @ifhtml | |
910 | See ``Spec Files'' in the main manual | |
911 | @end ifhtml | |
912 | ||
913 | @end table | |
914 | ||
915 | @item --program-prefix=@var{prefix} | |
916 | GCC supports some transformations of the names of its programs when | |
917 | installing them. This option prepends @var{prefix} to the names of | |
918 | programs to install in @var{bindir} (see above). For example, specifying | |
919 | @option{--program-prefix=foo-} would result in @samp{gcc} | |
920 | being installed as @file{/usr/local/bin/foo-gcc}. | |
921 | ||
922 | @item --program-suffix=@var{suffix} | |
923 | Appends @var{suffix} to the names of programs to install in @var{bindir} | |
924 | (see above). For example, specifying @option{--program-suffix=-3.1} | |
925 | would result in @samp{gcc} being installed as | |
926 | @file{/usr/local/bin/gcc-3.1}. | |
927 | ||
928 | @item --program-transform-name=@var{pattern} | |
929 | Applies the @samp{sed} script @var{pattern} to be applied to the names | |
930 | of programs to install in @var{bindir} (see above). @var{pattern} has to | |
931 | consist of one or more basic @samp{sed} editing commands, separated by | |
932 | semicolons. For example, if you want the @samp{gcc} program name to be | |
933 | transformed to the installed program @file{/usr/local/bin/myowngcc} and | |
934 | the @samp{g++} program name to be transformed to | |
935 | @file{/usr/local/bin/gspecial++} without changing other program names, | |
936 | you could use the pattern | |
937 | @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'} | |
938 | to achieve this effect. | |
939 | ||
940 | All three options can be combined and used together, resulting in more | |
941 | complex conversion patterns. As a basic rule, @var{prefix} (and | |
942 | @var{suffix}) are prepended (appended) before further transformations | |
943 | can happen with a special transformation script @var{pattern}. | |
944 | ||
945 | As currently implemented, this option only takes effect for native | |
946 | builds; cross compiler binaries' names are not transformed even when a | |
947 | transformation is explicitly asked for by one of these options. | |
948 | ||
949 | For native builds, some of the installed programs are also installed | |
950 | with the target alias in front of their name, as in | |
951 | @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen | |
952 | before the target alias is prepended to the name---so, specifying | |
953 | @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the | |
954 | resulting binary would be installed as | |
955 | @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}. | |
956 | ||
957 | As a last shortcoming, none of the installed Ada programs are | |
958 | transformed yet, which will be fixed in some time. | |
959 | ||
960 | @item --with-local-prefix=@var{dirname} | |
961 | Specify the | |
962 | installation directory for local include files. The default is | |
963 | @file{/usr/local}. Specify this option if you want the compiler to | |
964 | search directory @file{@var{dirname}/include} for locally installed | |
965 | header files @emph{instead} of @file{/usr/local/include}. | |
966 | ||
967 | You should specify @option{--with-local-prefix} @strong{only} if your | |
968 | site has a different convention (not @file{/usr/local}) for where to put | |
969 | site-specific files. | |
970 | ||
971 | The default value for @option{--with-local-prefix} is @file{/usr/local} | |
972 | regardless of the value of @option{--prefix}. Specifying | |
973 | @option{--prefix} has no effect on which directory GCC searches for | |
974 | local header files. This may seem counterintuitive, but actually it is | |
975 | logical. | |
976 | ||
977 | The purpose of @option{--prefix} is to specify where to @emph{install | |
978 | GCC}. The local header files in @file{/usr/local/include}---if you put | |
979 | any in that directory---are not part of GCC@. They are part of other | |
980 | programs---perhaps many others. (GCC installs its own header files in | |
981 | another directory which is based on the @option{--prefix} value.) | |
982 | ||
983 | Both the local-prefix include directory and the GCC-prefix include | |
984 | directory are part of GCC's ``system include'' directories. Although these | |
985 | two directories are not fixed, they need to be searched in the proper | |
986 | order for the correct processing of the include_next directive. The | |
987 | local-prefix include directory is searched before the GCC-prefix | |
988 | include directory. Another characteristic of system include directories | |
989 | is that pedantic warnings are turned off for headers in these directories. | |
990 | ||
991 | Some autoconf macros add @option{-I @var{directory}} options to the | |
992 | compiler command line, to ensure that directories containing installed | |
993 | packages' headers are searched. When @var{directory} is one of GCC's | |
994 | system include directories, GCC will ignore the option so that system | |
995 | directories continue to be processed in the correct order. This | |
996 | may result in a search order different from what was specified but the | |
997 | directory will still be searched. | |
998 | ||
999 | GCC automatically searches for ordinary libraries using | |
1000 | @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is | |
1001 | used for both GCC and packages, GCC will automatically search for | |
1002 | both headers and libraries. This provides a configuration that is | |
1003 | easy to use. GCC behaves in a manner similar to that when it is | |
1004 | installed as a system compiler in @file{/usr}. | |
1005 | ||
1006 | Sites that need to install multiple versions of GCC may not want to | |
1007 | use the above simple configuration. It is possible to use the | |
1008 | @option{--program-prefix}, @option{--program-suffix} and | |
1009 | @option{--program-transform-name} options to install multiple versions | |
1010 | into a single directory, but it may be simpler to use different prefixes | |
1011 | and the @option{--with-local-prefix} option to specify the location of the | |
1012 | site-specific files for each version. It will then be necessary for | |
1013 | users to specify explicitly the location of local site libraries | |
1014 | (e.g., with @env{LIBRARY_PATH}). | |
1015 | ||
1016 | The same value can be used for both @option{--with-local-prefix} and | |
1017 | @option{--prefix} provided it is not @file{/usr}. This can be used | |
1018 | to avoid the default search of @file{/usr/local/include}. | |
1019 | ||
1020 | @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}! | |
1021 | The directory you use for @option{--with-local-prefix} @strong{must not} | |
1022 | contain any of the system's standard header files. If it did contain | |
1023 | them, certain programs would be miscompiled (including GNU Emacs, on | |
1024 | certain targets), because this would override and nullify the header | |
1025 | file corrections made by the @command{fixincludes} script. | |
1026 | ||
1027 | Indications are that people who use this option use it based on mistaken | |
1028 | ideas of what it is for. People use it as if it specified where to | |
1029 | install part of GCC@. Perhaps they make this assumption because | |
1030 | installing GCC creates the directory. | |
1031 | ||
1032 | @item --with-gcc-major-version-only | |
1033 | Specifies that GCC should use only the major number rather than | |
1034 | @var{major}.@var{minor}.@var{patchlevel} in filesystem paths. | |
1035 | ||
1036 | @item --with-native-system-header-dir=@var{dirname} | |
1037 | Specifies that @var{dirname} is the directory that contains native system | |
1038 | header files, rather than @file{/usr/include}. This option is most useful | |
1039 | if you are creating a compiler that should be isolated from the system | |
1040 | as much as possible. It is most commonly used with the | |
1041 | @option{--with-sysroot} option and will cause GCC to search | |
1042 | @var{dirname} inside the system root specified by that option. | |
1043 | ||
1044 | @item --enable-shared[=@var{package}[,@dots{}]] | |
1045 | Build shared versions of libraries, if shared libraries are supported on | |
1046 | the target platform. Unlike GCC 2.95.x and earlier, shared libraries | |
1047 | are enabled by default on all platforms that support shared libraries. | |
1048 | ||
1049 | If a list of packages is given as an argument, build shared libraries | |
1050 | only for the listed packages. For other packages, only static libraries | |
1051 | will be built. Package names currently recognized in the GCC tree are | |
1052 | @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not | |
1053 | @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc}, | |
1054 | @samp{ada}, @samp{libada}, @samp{libgo}, @samp{libobjc}, and @samp{libphobos}. | |
1055 | Note @samp{libiberty} does not support shared libraries at all. | |
1056 | ||
1057 | Use @option{--disable-shared} to build only static libraries. Note that | |
1058 | @option{--disable-shared} does not accept a list of package names as | |
1059 | argument, only @option{--enable-shared} does. | |
1060 | ||
1061 | Contrast with @option{--enable-host-shared}, which affects @emph{host} | |
1062 | code. | |
1063 | ||
1064 | @item --enable-host-shared | |
1065 | Specify that the @emph{host} code should be built into position-independent | |
04711f51 MP |
1066 | machine code (with -fPIC), allowing it to be used within shared libraries, |
1067 | but yielding a slightly slower compiler. | |
d77de738 ML |
1068 | |
1069 | This option is required when building the libgccjit.so library. | |
1070 | ||
1071 | Contrast with @option{--enable-shared}, which affects @emph{target} | |
1072 | libraries. | |
1073 | ||
1074 | @item @anchor{with-gnu-as}--with-gnu-as | |
1075 | Specify that the compiler should assume that the | |
1076 | assembler it finds is the GNU assembler. However, this does not modify | |
1077 | the rules to find an assembler and will result in confusion if the | |
1078 | assembler found is not actually the GNU assembler. (Confusion may also | |
1079 | result if the compiler finds the GNU assembler but has not been | |
1080 | configured with @option{--with-gnu-as}.) If you have more than one | |
1081 | assembler installed on your system, you may want to use this option in | |
1082 | connection with @option{--with-as=@var{pathname}} or | |
1083 | @option{--with-build-time-tools=@var{pathname}}. | |
1084 | ||
1085 | The following systems are the only ones where it makes a difference | |
1086 | whether you use the GNU assembler. On any other system, | |
1087 | @option{--with-gnu-as} has no effect. | |
1088 | ||
1089 | @itemize @bullet | |
1090 | @item @samp{hppa1.0-@var{any}-@var{any}} | |
1091 | @item @samp{hppa1.1-@var{any}-@var{any}} | |
1092 | @item @samp{sparc-sun-solaris2.@var{any}} | |
1093 | @item @samp{sparc64-@var{any}-solaris2.@var{any}} | |
1094 | @end itemize | |
1095 | ||
1096 | @item @anchor{with-as}--with-as=@var{pathname} | |
1097 | Specify that the compiler should use the assembler pointed to by | |
1098 | @var{pathname}, rather than the one found by the standard rules to find | |
1099 | an assembler, which are: | |
1100 | @itemize @bullet | |
1101 | @item | |
1102 | Unless GCC is being built with a cross compiler, check the | |
1103 | @file{@var{libexec}/gcc/@var{target}/@var{version}} directory. | |
1104 | @var{libexec} defaults to @file{@var{exec-prefix}/libexec}; | |
1105 | @var{exec-prefix} defaults to @var{prefix}, which | |
1106 | defaults to @file{/usr/local} unless overridden by the | |
1107 | @option{--prefix=@var{pathname}} switch described above. @var{target} | |
1108 | is the target system triple, such as @samp{sparc-sun-solaris2.7}, and | |
1109 | @var{version} denotes the GCC version, such as 3.0. | |
1110 | ||
1111 | @item | |
1112 | If the target system is the same that you are building on, check | |
1113 | operating system specific directories (e.g.@: @file{/usr/ccs/bin} on | |
1114 | Solaris 2). | |
1115 | ||
1116 | @item | |
1117 | Check in the @env{PATH} for a tool whose name is prefixed by the | |
1118 | target system triple. | |
1119 | ||
1120 | @item | |
1121 | Check in the @env{PATH} for a tool whose name is not prefixed by the | |
1122 | target system triple, if the host and target system triple are | |
1123 | the same (in other words, we use a host tool if it can be used for | |
1124 | the target as well). | |
1125 | @end itemize | |
1126 | ||
1127 | You may want to use @option{--with-as} if no assembler | |
1128 | is installed in the directories listed above, or if you have multiple | |
1129 | assemblers installed and want to choose one that is not found by the | |
1130 | above rules. | |
1131 | ||
1132 | @item @anchor{with-gnu-ld}--with-gnu-ld | |
1133 | Same as @uref{#with-gnu-as,,@option{--with-gnu-as}} | |
1134 | but for the linker. | |
1135 | ||
1136 | @item --with-ld=@var{pathname} | |
1137 | Same as @uref{#with-as,,@option{--with-as}} | |
1138 | but for the linker. | |
1139 | ||
1140 | @item --with-dsymutil=@var{pathname} | |
1141 | Same as @uref{#with-as,,@option{--with-as}} | |
1142 | but for the debug linker (only used on Darwin platforms so far). | |
1143 | ||
1144 | @item --with-tls=@var{dialect} | |
1145 | Specify the default TLS dialect, for systems were there is a choice. | |
1146 | For ARM targets, possible values for @var{dialect} are @code{gnu} or | |
1147 | @code{gnu2}, which select between the original GNU dialect and the GNU TLS | |
1148 | descriptor-based dialect. | |
1149 | ||
1150 | @item --enable-multiarch | |
1151 | Specify whether to enable or disable multiarch support. The default is | |
1152 | to check for glibc start files in a multiarch location, and enable it | |
1153 | if the files are found. The auto detection is enabled for native builds, | |
1154 | and for cross builds configured with @option{--with-sysroot}, and without | |
1155 | @option{--with-native-system-header-dir}. | |
1156 | More documentation about multiarch can be found at | |
1157 | @uref{https://wiki.debian.org/Multiarch}. | |
1158 | ||
1159 | @item --enable-sjlj-exceptions | |
1160 | Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions. | |
1161 | @samp{configure} ordinarily picks the correct value based on the platform. | |
1162 | Only use this option if you are sure you need a different setting. | |
1163 | ||
1164 | @item --enable-vtable-verify | |
1165 | Specify whether to enable or disable the vtable verification feature. | |
1166 | Enabling this feature causes libstdc++ to be built with its virtual calls | |
1167 | in verifiable mode. This means that, when linked with libvtv, every | |
1168 | virtual call in libstdc++ will verify the vtable pointer through which the | |
1169 | call will be made before actually making the call. If not linked with libvtv, | |
1170 | the verifier will call stub functions (in libstdc++ itself) and do nothing. | |
1171 | If vtable verification is disabled, then libstdc++ is not built with its | |
1172 | virtual calls in verifiable mode at all. However the libvtv library will | |
1173 | still be built (see @option{--disable-libvtv} to turn off building libvtv). | |
1174 | @option{--disable-vtable-verify} is the default. | |
1175 | ||
1176 | @item --disable-gcov | |
1177 | Specify that the run-time library used for coverage analysis | |
1178 | and associated host tools should not be built. | |
1179 | ||
1180 | @item --disable-multilib | |
1181 | Specify that multiple target | |
1182 | libraries to support different target variants, calling | |
1183 | conventions, etc.@: should not be built. The default is to build a | |
1184 | predefined set of them. | |
1185 | ||
1186 | Some targets provide finer-grained control over which multilibs are built | |
1187 | (e.g., @option{--disable-softfloat}): | |
1188 | @table @code | |
1189 | @item arm-*-* | |
1190 | fpu, 26bit, underscore, interwork, biendian, nofmult. | |
1191 | ||
1192 | @item m68*-*-* | |
1193 | softfloat, m68881, m68000, m68020. | |
1194 | ||
1195 | @item mips*-*-* | |
1196 | single-float, biendian, softfloat. | |
1197 | ||
1198 | @item msp430-*-* | |
1199 | no-exceptions | |
1200 | ||
1201 | @item powerpc*-*-*, rs6000*-*-* | |
1202 | aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian, | |
1203 | sysv, aix. | |
1204 | ||
1205 | @end table | |
1206 | ||
1207 | @item --with-multilib-list=@var{list} | |
1208 | @itemx --without-multilib-list | |
1209 | Specify what multilibs to build. @var{list} is a comma separated list of | |
1210 | values, possibly consisting of a single value. Currently only implemented | |
1211 | for aarch64*-*-*, arm*-*-*, loongarch64-*-*, riscv*-*-*, sh*-*-* and | |
1212 | x86-64-*-linux*. The accepted values and meaning for each target is given | |
1213 | below. | |
1214 | ||
1215 | @table @code | |
1216 | @item aarch64*-*-* | |
1217 | @var{list} is a comma separated list of @code{ilp32}, and @code{lp64} | |
1218 | to enable ILP32 and LP64 run-time libraries, respectively. If | |
1219 | @var{list} is empty, then there will be no multilibs and only the | |
1220 | default run-time library will be built. If @var{list} is | |
1221 | @code{default} or --with-multilib-list= is not specified, then the | |
1222 | default set of libraries is selected based on the value of | |
1223 | @option{--target}. | |
1224 | ||
1225 | @item arm*-*-* | |
1226 | @var{list} is a comma separated list of @code{aprofile} and | |
1227 | @code{rmprofile} to build multilibs for A or R and M architecture | |
1228 | profiles respectively. Note that, due to some limitation of the current | |
1229 | multilib framework, using the combined @code{aprofile,rmprofile} | |
1230 | multilibs selects in some cases a less optimal multilib than when using | |
1231 | the multilib profile for the architecture targetted. The special value | |
1232 | @code{default} is also accepted and is equivalent to omitting the | |
1233 | option, i.e., only the default run-time library will be enabled. | |
1234 | ||
1235 | @var{list} may instead contain @code{@@name}, to use the multilib | |
1236 | configuration Makefile fragment @file{name} in @file{gcc/config/arm} in | |
1237 | the source tree (it is part of the corresponding sources, after all). | |
1238 | It is recommended, but not required, that files used for this purpose to | |
1239 | be named starting with @file{t-ml-}, to make their intended purpose | |
1240 | self-evident, in line with GCC conventions. Such files enable custom, | |
1241 | user-chosen multilib lists to be configured. Whether multiple such | |
1242 | files can be used together depends on the contents of the supplied | |
1243 | files. See @file{gcc/config/arm/t-multilib} and its supplementary | |
1244 | @file{gcc/config/arm/t-*profile} files for an example of what such | |
1245 | Makefile fragments might look like for this version of GCC. The macros | |
1246 | expected to be defined in these fragments are not stable across GCC | |
1247 | releases, so make sure they define the @code{MULTILIB}-related macros | |
1248 | expected by the version of GCC you are building. | |
1249 | @ifnothtml | |
1250 | @xref{Target Fragment,, Target Makefile Fragments, gccint, GNU Compiler | |
1251 | Collection (GCC) Internals}. | |
1252 | @end ifnothtml | |
1253 | @ifhtml | |
1254 | See ``Target Makefile Fragments'' in the internals manual. | |
1255 | @end ifhtml | |
1256 | ||
1257 | The table below gives the combination of ISAs, architectures, FPUs and | |
1258 | floating-point ABIs for which multilibs are built for each predefined | |
1259 | profile. The union of these options is considered when specifying both | |
1260 | @code{aprofile} and @code{rmprofile}. | |
1261 | ||
1262 | @multitable @columnfractions .15 .28 .30 | |
1263 | @item Option @tab aprofile @tab rmprofile | |
1264 | @item ISAs | |
1265 | @tab @code{-marm} and @code{-mthumb} | |
1266 | @tab @code{-mthumb} | |
1267 | @item Architectures@*@*@*@*@*@* | |
1268 | @tab default architecture@* | |
1269 | @code{-march=armv7-a}@* | |
1270 | @code{-march=armv7ve}@* | |
1271 | @code{-march=armv8-a}@*@*@* | |
1272 | @tab default architecture@* | |
1273 | @code{-march=armv6s-m}@* | |
1274 | @code{-march=armv7-m}@* | |
1275 | @code{-march=armv7e-m}@* | |
1276 | @code{-march=armv8-m.base}@* | |
1277 | @code{-march=armv8-m.main}@* | |
1278 | @code{-march=armv7} | |
1279 | @item FPUs@*@*@*@*@* | |
1280 | @tab none@* | |
1281 | @code{-mfpu=vfpv3-d16}@* | |
1282 | @code{-mfpu=neon}@* | |
1283 | @code{-mfpu=vfpv4-d16}@* | |
1284 | @code{-mfpu=neon-vfpv4}@* | |
1285 | @code{-mfpu=neon-fp-armv8} | |
1286 | @tab none@* | |
1287 | @code{-mfpu=vfpv3-d16}@* | |
1288 | @code{-mfpu=fpv4-sp-d16}@* | |
1289 | @code{-mfpu=fpv5-sp-d16}@* | |
1290 | @code{-mfpu=fpv5-d16}@* | |
1291 | @item floating-point@/ ABIs@*@* | |
1292 | @tab @code{-mfloat-abi=soft}@* | |
1293 | @code{-mfloat-abi=softfp}@* | |
1294 | @code{-mfloat-abi=hard} | |
1295 | @tab @code{-mfloat-abi=soft}@* | |
1296 | @code{-mfloat-abi=softfp}@* | |
1297 | @code{-mfloat-abi=hard} | |
1298 | @end multitable | |
1299 | ||
1300 | @item loongarch*-*-* | |
1301 | @var{list} is a comma-separated list of the following ABI identifiers: | |
1302 | @code{lp64d[/base]} @code{lp64f[/base]} @code{lp64d[/base]}, where the | |
1303 | @code{/base} suffix may be omitted, to enable their respective run-time | |
1304 | libraries. If @var{list} is empty or @code{default}, | |
1305 | or if @option{--with-multilib-list} is not specified, then the default ABI | |
1306 | as specified by @option{--with-abi} or implied by @option{--target} is selected. | |
1307 | ||
1308 | @item riscv*-*-* | |
1309 | @var{list} is a single ABI name. The target architecture must be either | |
1310 | @code{rv32gc} or @code{rv64gc}. This will build a single multilib for the | |
1311 | specified architecture and ABI pair. If @code{--with-multilib-list} is not | |
1312 | given, then a default set of multilibs is selected based on the value of | |
1313 | @option{--target}. This is usually a large set of multilibs. | |
1314 | ||
1315 | @item sh*-*-* | |
1316 | @var{list} is a comma separated list of CPU names. These must be of the | |
1317 | form @code{sh*} or @code{m*} (in which case they match the compiler option | |
1318 | for that processor). The list should not contain any endian options - | |
1319 | these are handled by @option{--with-endian}. | |
1320 | ||
1321 | If @var{list} is empty, then there will be no multilibs for extra | |
1322 | processors. The multilib for the secondary endian remains enabled. | |
1323 | ||
1324 | As a special case, if an entry in the list starts with a @code{!} | |
1325 | (exclamation point), then it is added to the list of excluded multilibs. | |
1326 | Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES} | |
1327 | (once the leading @code{!} has been stripped). | |
1328 | ||
1329 | If @option{--with-multilib-list} is not given, then a default set of | |
1330 | multilibs is selected based on the value of @option{--target}. This is | |
1331 | usually the complete set of libraries, but some targets imply a more | |
1332 | specialized subset. | |
1333 | ||
1334 | Example 1: to configure a compiler for SH4A only, but supporting both | |
1335 | endians, with little endian being the default: | |
1336 | @smallexample | |
1337 | --with-cpu=sh4a --with-endian=little,big --with-multilib-list= | |
1338 | @end smallexample | |
1339 | ||
1340 | Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with | |
1341 | only little endian SH4AL: | |
1342 | @smallexample | |
1343 | --with-cpu=sh4a --with-endian=little,big \ | |
1344 | --with-multilib-list=sh4al,!mb/m4al | |
1345 | @end smallexample | |
1346 | ||
1347 | @item x86-64-*-linux* | |
1348 | @var{list} is a comma separated list of @code{m32}, @code{m64} and | |
1349 | @code{mx32} to enable 32-bit, 64-bit and x32 run-time libraries, | |
1350 | respectively. If @var{list} is empty, then there will be no multilibs | |
1351 | and only the default run-time library will be enabled. | |
1352 | ||
1353 | If @option{--with-multilib-list} is not given, then only 32-bit and | |
1354 | 64-bit run-time libraries will be enabled. | |
1355 | @end table | |
1356 | ||
1357 | @item --with-multilib-generator=@var{config} | |
1358 | Specify what multilibs to build. @var{config} is a semicolon separated list of | |
1359 | values, possibly consisting of a single value. Currently only implemented | |
1360 | for riscv*-*-elf*. The accepted values and meanings are given below. | |
1361 | ||
1362 | ||
1363 | Every config is constructed with four components: architecture string, ABI, | |
1364 | reuse rule with architecture string and reuse rule with sub-extension. | |
1365 | ||
1366 | Example 1: Add multi-lib suppport for rv32i with ilp32. | |
1367 | @smallexample | |
1368 | rv32i-ilp32-- | |
1369 | @end smallexample | |
1370 | ||
1371 | Example 2: Add multi-lib suppport for rv32i with ilp32 and rv32imafd with ilp32. | |
1372 | @smallexample | |
1373 | rv32i-ilp32--;rv32imafd-ilp32-- | |
1374 | @end smallexample | |
1375 | ||
1376 | Example 3: Add multi-lib suppport for rv32i with ilp32; rv32im with ilp32 and | |
1377 | rv32ic with ilp32 will reuse this multi-lib set. | |
1378 | @smallexample | |
1379 | rv32i-ilp32-rv32im-c | |
1380 | @end smallexample | |
1381 | ||
1382 | Example 4: Add multi-lib suppport for rv64ima with lp64; rv64imaf with lp64, | |
1383 | rv64imac with lp64 and rv64imafc with lp64 will reuse this multi-lib set. | |
1384 | @smallexample | |
1385 | rv64ima-lp64--f,c,fc | |
1386 | @end smallexample | |
1387 | ||
1388 | @option{--with-multilib-generator} have an optional configuration argument | |
1389 | @option{--cmodel=val} for code model, this option will expand with other | |
1390 | config options, @var{val} is a comma separated list of possible code model, | |
1391 | currently we support medlow and medany. | |
1392 | ||
1393 | Example 5: Add multi-lib suppport for rv64ima with lp64; rv64ima with lp64 and | |
1394 | medlow code model | |
1395 | @smallexample | |
1396 | rv64ima-lp64--;--cmodel=medlow | |
1397 | @end smallexample | |
1398 | ||
1399 | Example 6: Add multi-lib suppport for rv64ima with lp64; rv64ima with lp64 and | |
1400 | medlow code model; rv64ima with lp64 and medany code model | |
1401 | @smallexample | |
1402 | rv64ima-lp64--;--cmodel=medlow,medany | |
1403 | @end smallexample | |
1404 | ||
1405 | @item --with-endian=@var{endians} | |
1406 | Specify what endians to use. | |
1407 | Currently only implemented for sh*-*-*. | |
1408 | ||
1409 | @var{endians} may be one of the following: | |
1410 | @table @code | |
1411 | @item big | |
1412 | Use big endian exclusively. | |
1413 | @item little | |
1414 | Use little endian exclusively. | |
1415 | @item big,little | |
1416 | Use big endian by default. Provide a multilib for little endian. | |
1417 | @item little,big | |
1418 | Use little endian by default. Provide a multilib for big endian. | |
1419 | @end table | |
1420 | ||
1421 | @item --enable-threads | |
1422 | Specify that the target | |
1423 | supports threads. This affects the Objective-C compiler and runtime | |
1424 | library, and exception handling for other languages like C++. | |
1425 | On some systems, this is the default. | |
1426 | ||
1427 | In general, the best (and, in many cases, the only known) threading | |
1428 | model available will be configured for use. Beware that on some | |
1429 | systems, GCC has not been taught what threading models are generally | |
1430 | available for the system. In this case, @option{--enable-threads} is an | |
1431 | alias for @option{--enable-threads=single}. | |
1432 | ||
1433 | @item --disable-threads | |
1434 | Specify that threading support should be disabled for the system. | |
1435 | This is an alias for @option{--enable-threads=single}. | |
1436 | ||
1437 | @item --enable-threads=@var{lib} | |
1438 | Specify that | |
1439 | @var{lib} is the thread support library. This affects the Objective-C | |
1440 | compiler and runtime library, and exception handling for other languages | |
1441 | like C++. The possibilities for @var{lib} are: | |
1442 | ||
1443 | @table @code | |
1444 | @item aix | |
1445 | AIX thread support. | |
1446 | @item dce | |
1447 | DCE thread support. | |
1448 | @item lynx | |
1449 | LynxOS thread support. | |
1450 | @item mipssde | |
1451 | MIPS SDE thread support. | |
1452 | @item no | |
1453 | This is an alias for @samp{single}. | |
1454 | @item posix | |
1455 | Generic POSIX/Unix98 thread support. | |
1456 | @item rtems | |
1457 | RTEMS thread support. | |
1458 | @item single | |
1459 | Disable thread support, should work for all platforms. | |
1460 | @item tpf | |
1461 | TPF thread support. | |
1462 | @item vxworks | |
1463 | VxWorks thread support. | |
1464 | @item win32 | |
1465 | Microsoft Win32 API thread support. | |
1466 | @end table | |
1467 | ||
1468 | @item --enable-tls | |
1469 | Specify that the target supports TLS (Thread Local Storage). Usually | |
1470 | configure can correctly determine if TLS is supported. In cases where | |
1471 | it guesses incorrectly, TLS can be explicitly enabled or disabled with | |
1472 | @option{--enable-tls} or @option{--disable-tls}. This can happen if | |
1473 | the assembler supports TLS but the C library does not, or if the | |
1474 | assumptions made by the configure test are incorrect. | |
1475 | ||
1476 | @item --disable-tls | |
1477 | Specify that the target does not support TLS. | |
1478 | This is an alias for @option{--enable-tls=no}. | |
1479 | ||
1480 | @item --disable-tm-clone-registry | |
1481 | Disable TM clone registry in libgcc. It is enabled in libgcc by default. | |
1482 | This option helps to reduce code size for embedded targets which do | |
1483 | not use transactional memory. | |
1484 | ||
1485 | @item --with-cpu=@var{cpu} | |
1486 | @itemx --with-cpu-32=@var{cpu} | |
1487 | @itemx --with-cpu-64=@var{cpu} | |
1488 | Specify which cpu variant the compiler should generate code for by default. | |
1489 | @var{cpu} will be used as the default value of the @option{-mcpu=} switch. | |
1490 | This option is only supported on some targets, including ARC, ARM, i386, M68k, | |
1491 | PowerPC, and SPARC@. It is mandatory for ARC@. The @option{--with-cpu-32} and | |
1492 | @option{--with-cpu-64} options specify separate default CPUs for | |
1493 | 32-bit and 64-bit modes; these options are only supported for aarch64, i386, | |
1494 | x86-64, PowerPC, and SPARC@. | |
1495 | ||
1496 | @item --with-schedule=@var{cpu} | |
1497 | @itemx --with-arch=@var{cpu} | |
1498 | @itemx --with-arch-32=@var{cpu} | |
1499 | @itemx --with-arch-64=@var{cpu} | |
1500 | @itemx --with-tune=@var{cpu} | |
1501 | @itemx --with-tune-32=@var{cpu} | |
1502 | @itemx --with-tune-64=@var{cpu} | |
1503 | @itemx --with-abi=@var{abi} | |
1504 | @itemx --with-fpu=@var{type} | |
1505 | @itemx --with-float=@var{type} | |
1506 | These configure options provide default values for the @option{-mschedule=}, | |
1507 | @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=} | |
1508 | options and for @option{-mhard-float} or @option{-msoft-float}. As with | |
1509 | @option{--with-cpu}, which switches will be accepted and acceptable values | |
1510 | of the arguments depend on the target. | |
1511 | ||
1512 | @item --with-mode=@var{mode} | |
1513 | Specify if the compiler should default to @option{-marm} or @option{-mthumb}. | |
1514 | This option is only supported on ARM targets. | |
1515 | ||
1516 | @item --with-stack-offset=@var{num} | |
1517 | This option sets the default for the -mstack-offset=@var{num} option, | |
1518 | and will thus generally also control the setting of this option for | |
1519 | libraries. This option is only supported on Epiphany targets. | |
1520 | ||
1521 | @item --with-fpmath=@var{isa} | |
1522 | This options sets @option{-mfpmath=sse} by default and specifies the default | |
1523 | ISA for floating-point arithmetics. You can select either @samp{sse} which | |
1524 | enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default. | |
1525 | This option is only supported on i386 and x86-64 targets. | |
1526 | ||
1527 | @item --with-fp-32=@var{mode} | |
1528 | On MIPS targets, set the default value for the @option{-mfp} option when using | |
1529 | the o32 ABI. The possibilities for @var{mode} are: | |
1530 | @table @code | |
1531 | @item 32 | |
1532 | Use the o32 FP32 ABI extension, as with the @option{-mfp32} command-line | |
1533 | option. | |
1534 | @item xx | |
1535 | Use the o32 FPXX ABI extension, as with the @option{-mfpxx} command-line | |
1536 | option. | |
1537 | @item 64 | |
1538 | Use the o32 FP64 ABI extension, as with the @option{-mfp64} command-line | |
1539 | option. | |
1540 | @end table | |
1541 | In the absence of this configuration option the default is to use the o32 | |
1542 | FP32 ABI extension. | |
1543 | ||
1544 | @item --with-odd-spreg-32 | |
1545 | On MIPS targets, set the @option{-modd-spreg} option by default when using | |
1546 | the o32 ABI. | |
1547 | ||
1548 | @item --without-odd-spreg-32 | |
1549 | On MIPS targets, set the @option{-mno-odd-spreg} option by default when using | |
1550 | the o32 ABI. This is normally used in conjunction with | |
1551 | @option{--with-fp-32=64} in order to target the o32 FP64A ABI extension. | |
1552 | ||
1553 | @item --with-nan=@var{encoding} | |
1554 | On MIPS targets, set the default encoding convention to use for the | |
1555 | special not-a-number (NaN) IEEE 754 floating-point data. The | |
1556 | possibilities for @var{encoding} are: | |
1557 | @table @code | |
1558 | @item legacy | |
1559 | Use the legacy encoding, as with the @option{-mnan=legacy} command-line | |
1560 | option. | |
1561 | @item 2008 | |
1562 | Use the 754-2008 encoding, as with the @option{-mnan=2008} command-line | |
1563 | option. | |
1564 | @end table | |
1565 | To use this configuration option you must have an assembler version | |
1566 | installed that supports the @option{-mnan=} command-line option too. | |
1567 | In the absence of this configuration option the default convention is | |
1568 | the legacy encoding, as when neither of the @option{-mnan=2008} and | |
1569 | @option{-mnan=legacy} command-line options has been used. | |
1570 | ||
1571 | @item --with-divide=@var{type} | |
1572 | Specify how the compiler should generate code for checking for | |
1573 | division by zero. This option is only supported on the MIPS target. | |
1574 | The possibilities for @var{type} are: | |
1575 | @table @code | |
1576 | @item traps | |
1577 | Division by zero checks use conditional traps (this is the default on | |
1578 | systems that support conditional traps). | |
1579 | @item breaks | |
1580 | Division by zero checks use the break instruction. | |
1581 | @end table | |
1582 | ||
1583 | @item --with-compact-branches=@var{policy} | |
1584 | Specify how the compiler should generate branch instructions. | |
1585 | This option is only supported on the MIPS target. | |
1586 | The possibilities for @var{type} are: | |
1587 | @table @code | |
1588 | @item optimal | |
1589 | Cause a delay slot branch to be used if one is available in the | |
1590 | current ISA and the delay slot is successfully filled. If the delay slot | |
1591 | is not filled, a compact branch will be chosen if one is available. | |
1592 | @item never | |
1593 | Ensures that compact branch instructions will never be generated. | |
1594 | @item always | |
1595 | Ensures that a compact branch instruction will be generated if available. | |
1596 | If a compact branch instruction is not available, | |
1597 | a delay slot form of the branch will be used instead. | |
1598 | This option is supported from MIPS Release 6 onwards. | |
1599 | For pre-R6/microMIPS/MIPS16, this option is just same as never/optimal. | |
1600 | @end table | |
1601 | ||
1602 | @c If you make --with-llsc the default for additional targets, | |
1603 | @c update the --with-llsc description in the MIPS section below. | |
1604 | ||
1605 | @item --with-llsc | |
1606 | On MIPS targets, make @option{-mllsc} the default when no | |
1607 | @option{-mno-llsc} option is passed. This is the default for | |
1608 | Linux-based targets, as the kernel will emulate them if the ISA does | |
1609 | not provide them. | |
1610 | ||
1611 | @item --without-llsc | |
1612 | On MIPS targets, make @option{-mno-llsc} the default when no | |
1613 | @option{-mllsc} option is passed. | |
1614 | ||
1615 | @item --with-synci | |
1616 | On MIPS targets, make @option{-msynci} the default when no | |
1617 | @option{-mno-synci} option is passed. | |
1618 | ||
1619 | @item --without-synci | |
1620 | On MIPS targets, make @option{-mno-synci} the default when no | |
1621 | @option{-msynci} option is passed. This is the default. | |
1622 | ||
1623 | @item --with-lxc1-sxc1 | |
1624 | On MIPS targets, make @option{-mlxc1-sxc1} the default when no | |
1625 | @option{-mno-lxc1-sxc1} option is passed. This is the default. | |
1626 | ||
1627 | @item --without-lxc1-sxc1 | |
1628 | On MIPS targets, make @option{-mno-lxc1-sxc1} the default when no | |
1629 | @option{-mlxc1-sxc1} option is passed. The indexed load/store | |
1630 | instructions are not directly a problem but can lead to unexpected | |
1631 | behaviour when deployed in an application intended for a 32-bit address | |
1632 | space but run on a 64-bit processor. The issue is seen because all | |
1633 | known MIPS 64-bit Linux kernels execute o32 and n32 applications | |
1634 | with 64-bit addressing enabled which affects the overflow behaviour | |
1635 | of the indexed addressing mode. GCC will assume that ordinary | |
1636 | 32-bit arithmetic overflow behaviour is the same whether performed | |
1637 | as an @code{addu} instruction or as part of the address calculation | |
1638 | in @code{lwxc1} type instructions. This assumption holds true in a | |
1639 | pure 32-bit environment and can hold true in a 64-bit environment if | |
1640 | the address space is accurately set to be 32-bit for o32 and n32. | |
1641 | ||
1642 | @item --with-madd4 | |
1643 | On MIPS targets, make @option{-mmadd4} the default when no | |
1644 | @option{-mno-madd4} option is passed. This is the default. | |
1645 | ||
1646 | @item --without-madd4 | |
1647 | On MIPS targets, make @option{-mno-madd4} the default when no | |
1648 | @option{-mmadd4} option is passed. The @code{madd4} instruction | |
1649 | family can be problematic when targeting a combination of cores that | |
1650 | implement these instructions differently. There are two known cores | |
1651 | that implement these as fused operations instead of unfused (where | |
1652 | unfused is normally expected). Disabling these instructions is the | |
1653 | only way to ensure compatible code is generated; this will incur | |
1654 | a performance penalty. | |
1655 | ||
1656 | @item --with-mips-plt | |
1657 | On MIPS targets, make use of copy relocations and PLTs. | |
1658 | These features are extensions to the traditional | |
1659 | SVR4-based MIPS ABIs and require support from GNU binutils | |
1660 | and the runtime C library. | |
1661 | ||
1662 | @item --with-stack-clash-protection-guard-size=@var{size} | |
1663 | On certain targets this option sets the default stack clash protection guard | |
1664 | size as a power of two in bytes. On AArch64 @var{size} is required to be either | |
1665 | 12 (4KB) or 16 (64KB). | |
1666 | ||
1667 | @item --with-isa-spec=@var{ISA-spec-string} | |
1668 | On RISC-V targets specify the default version of the RISC-V Unprivileged | |
1669 | (formerly User-Level) ISA specification to produce code conforming to. | |
1670 | The possibilities for @var{ISA-spec-string} are: | |
1671 | @table @code | |
1672 | @item 2.2 | |
1673 | Produce code conforming to version 2.2. | |
1674 | @item 20190608 | |
1675 | Produce code conforming to version 20190608. | |
1676 | @item 20191213 | |
1677 | Produce code conforming to version 20191213. | |
1678 | @end table | |
1679 | In the absence of this configuration option the default version is 20191213. | |
1680 | ||
1681 | @item --enable-__cxa_atexit | |
1682 | Define if you want to use __cxa_atexit, rather than atexit, to | |
1683 | register C++ destructors for local statics and global objects. | |
1684 | This is essential for fully standards-compliant handling of | |
1685 | destructors, but requires __cxa_atexit in libc. This option is currently | |
1686 | only available on systems with GNU libc. When enabled, this will cause | |
1687 | @option{-fuse-cxa-atexit} to be passed by default. | |
1688 | ||
1689 | @item --enable-gnu-indirect-function | |
1690 | Define if you want to enable the @code{ifunc} attribute. This option is | |
1691 | currently only available on systems with GNU libc on certain targets. | |
1692 | ||
1693 | @item --enable-target-optspace | |
1694 | Specify that target | |
1695 | libraries should be optimized for code space instead of code speed. | |
1696 | This is the default for the m32r platform. | |
1697 | ||
1698 | @item --with-cpp-install-dir=@var{dirname} | |
1699 | Specify that the user visible @command{cpp} program should be installed | |
1700 | in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}. | |
1701 | ||
1702 | @item --enable-comdat | |
1703 | Enable COMDAT group support. This is primarily used to override the | |
1704 | automatically detected value. | |
1705 | ||
1706 | @item --enable-initfini-array | |
1707 | Force the use of sections @code{.init_array} and @code{.fini_array} | |
1708 | (instead of @code{.init} and @code{.fini}) for constructors and | |
1709 | destructors. Option @option{--disable-initfini-array} has the | |
1710 | opposite effect. If neither option is specified, the configure script | |
1711 | will try to guess whether the @code{.init_array} and | |
1712 | @code{.fini_array} sections are supported and, if they are, use them. | |
1713 | ||
1714 | @item --enable-link-mutex | |
1715 | When building GCC, use a mutex to avoid linking the compilers for | |
1716 | multiple languages at the same time, to avoid thrashing on build | |
1717 | systems with limited free memory. The default is not to use such a mutex. | |
1718 | ||
1719 | @item --enable-link-serialization | |
1720 | When building GCC, use make dependencies to serialize linking the compilers for | |
1721 | multiple languages, to avoid thrashing on build | |
1722 | systems with limited free memory. The default is not to add such | |
1723 | dependencies and thus with parallel make potentially link different | |
1724 | compilers concurrently. If the argument is a positive integer, allow | |
1725 | that number of concurrent link processes for the large binaries. | |
1726 | ||
1727 | @item --enable-maintainer-mode | |
1728 | The build rules that regenerate the Autoconf and Automake output files as | |
1729 | well as the GCC master message catalog @file{gcc.pot} are normally | |
1730 | disabled. This is because it can only be rebuilt if the complete source | |
1731 | tree is present. If you have changed the sources and want to rebuild the | |
1732 | catalog, configuring with @option{--enable-maintainer-mode} will enable | |
1733 | this. Note that you need a recent version of the @code{gettext} tools | |
1734 | to do so. | |
1735 | ||
1736 | @item --disable-bootstrap | |
1737 | For a native build, the default configuration is to perform | |
1738 | a 3-stage bootstrap of the compiler when @samp{make} is invoked, | |
1739 | testing that GCC can compile itself correctly. If you want to disable | |
1740 | this process, you can configure with @option{--disable-bootstrap}. | |
1741 | ||
1742 | @item --enable-bootstrap | |
1743 | In special cases, you may want to perform a 3-stage build | |
1744 | even if the target and host triplets are different. | |
1745 | This is possible when the host can run code compiled for | |
1746 | the target (e.g.@: host is i686-linux, target is i486-linux). | |
1747 | Starting from GCC 4.2, to do this you have to configure explicitly | |
1748 | with @option{--enable-bootstrap}. | |
1749 | ||
1750 | @item --enable-generated-files-in-srcdir | |
1751 | Neither the .c and .h files that are generated from Bison and flex nor the | |
1752 | info manuals and man pages that are built from the .texi files are present | |
1753 | in the repository development tree. When building GCC from that development tree, | |
1754 | or from one of our snapshots, those generated files are placed in your | |
1755 | build directory, which allows for the source to be in a readonly | |
1756 | directory. | |
1757 | ||
1758 | If you configure with @option{--enable-generated-files-in-srcdir} then those | |
1759 | generated files will go into the source directory. This is mainly intended | |
1760 | for generating release or prerelease tarballs of the GCC sources, since it | |
1761 | is not a requirement that the users of source releases to have flex, Bison, | |
1762 | or makeinfo. | |
1763 | ||
1764 | @item --enable-version-specific-runtime-libs | |
1765 | Specify | |
1766 | that runtime libraries should be installed in the compiler specific | |
1767 | subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In | |
1768 | addition, @samp{libstdc++}'s include files will be installed into | |
1769 | @file{@var{libdir}} unless you overruled it by using | |
1770 | @option{--with-gxx-include-dir=@var{dirname}}. Using this option is | |
1771 | particularly useful if you intend to use several versions of GCC in | |
1772 | parallel. The default is @samp{yes} for @samp{libada}, and @samp{no} for | |
1773 | the remaining libraries. | |
1774 | ||
1775 | @item @anchor{WithAixSoname}--with-aix-soname=@samp{aix}, @samp{svr4} or @samp{both} | |
1776 | Traditional AIX shared library versioning (versioned @code{Shared Object} | |
1777 | files as members of unversioned @code{Archive Library} files named | |
1778 | @samp{lib.a}) causes numerous headaches for package managers. However, | |
1779 | @code{Import Files} as members of @code{Archive Library} files allow for | |
1780 | @strong{filename-based versioning} of shared libraries as seen on Linux/SVR4, | |
1781 | where this is called the "SONAME". But as they prevent static linking, | |
1782 | @code{Import Files} may be used with @code{Runtime Linking} only, where the | |
1783 | linker does search for @samp{libNAME.so} before @samp{libNAME.a} library | |
1784 | filenames with the @samp{-lNAME} linker flag. | |
1785 | ||
1786 | @anchor{AixLdCommand}For detailed information please refer to the AIX | |
1787 | @uref{https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22,,ld | |
1788 | Command} reference. | |
1789 | ||
1790 | As long as shared library creation is enabled, upon: | |
1791 | @table @code | |
1792 | @item --with-aix-soname=aix | |
1793 | @item --with-aix-soname=both | |
1794 | A (traditional AIX) @code{Shared Archive Library} file is created: | |
1795 | @itemize @bullet | |
1796 | @item using the @samp{libNAME.a} filename scheme | |
1797 | @item with the @code{Shared Object} file as archive member named | |
1798 | @samp{libNAME.so.V} (except for @samp{libgcc_s}, where the @code{Shared | |
1799 | Object} file is named @samp{shr.o} for backwards compatibility), which | |
1800 | @itemize @minus | |
1801 | @item is used for runtime loading from inside the @samp{libNAME.a} file | |
1802 | @item is used for dynamic loading via | |
1803 | @code{dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)} | |
1804 | @item is used for shared linking | |
1805 | @item is used for static linking, so no separate @code{Static Archive | |
1806 | Library} file is needed | |
1807 | @end itemize | |
1808 | @end itemize | |
1809 | @item --with-aix-soname=both | |
1810 | @item --with-aix-soname=svr4 | |
1811 | A (second) @code{Shared Archive Library} file is created: | |
1812 | @itemize @bullet | |
1813 | @item using the @samp{libNAME.so.V} filename scheme | |
1814 | @item with the @code{Shared Object} file as archive member named | |
1815 | @samp{shr.o}, which | |
1816 | @itemize @minus | |
1817 | @item is created with the @code{-G linker flag} | |
1818 | @item has the @code{F_LOADONLY} flag set | |
1819 | @item is used for runtime loading from inside the @samp{libNAME.so.V} file | |
1820 | @item is used for dynamic loading via @code{dlopen("libNAME.so.V(shr.o)", | |
1821 | RTLD_MEMBER)} | |
1822 | @end itemize | |
1823 | @item with the @code{Import File} as archive member named @samp{shr.imp}, | |
1824 | which | |
1825 | @itemize @minus | |
1826 | @item refers to @samp{libNAME.so.V(shr.o)} as the "SONAME", to be recorded | |
1827 | in the @code{Loader Section} of subsequent binaries | |
1828 | @item indicates whether @samp{libNAME.so.V(shr.o)} is 32 or 64 bit | |
1829 | @item lists all the public symbols exported by @samp{lib.so.V(shr.o)}, | |
1830 | eventually decorated with the @code{@samp{weak} Keyword} | |
1831 | @item is necessary for shared linking against @samp{lib.so.V(shr.o)} | |
1832 | @end itemize | |
1833 | @end itemize | |
1834 | A symbolic link using the @samp{libNAME.so} filename scheme is created: | |
1835 | @itemize @bullet | |
1836 | @item pointing to the @samp{libNAME.so.V} @code{Shared Archive Library} file | |
1837 | @item to permit the @code{ld Command} to find @samp{lib.so.V(shr.imp)} via | |
1838 | the @samp{-lNAME} argument (requires @code{Runtime Linking} to be enabled) | |
1839 | @item to permit dynamic loading of @samp{lib.so.V(shr.o)} without the need | |
1840 | to specify the version number via @code{dlopen("libNAME.so(shr.o)", | |
1841 | RTLD_MEMBER)} | |
1842 | @end itemize | |
1843 | @end table | |
1844 | ||
1845 | As long as static library creation is enabled, upon: | |
1846 | @table @code | |
1847 | @item --with-aix-soname=svr4 | |
1848 | A @code{Static Archive Library} is created: | |
1849 | @itemize @bullet | |
1850 | @item using the @samp{libNAME.a} filename scheme | |
1851 | @item with all the @code{Static Object} files as archive members, which | |
1852 | @itemize @minus | |
1853 | @item are used for static linking | |
1854 | @end itemize | |
1855 | @end itemize | |
1856 | @end table | |
1857 | ||
1858 | While the aix-soname=@samp{svr4} option does not create @code{Shared Object} | |
1859 | files as members of unversioned @code{Archive Library} files any more, package | |
1860 | managers still are responsible to | |
1861 | @uref{./specific.html#TransferAixShobj,,transfer} @code{Shared Object} files | |
1862 | found as member of a previously installed unversioned @code{Archive Library} | |
1863 | file into the newly installed @code{Archive Library} file with the same | |
1864 | filename. | |
1865 | ||
1866 | @emph{WARNING:} Creating @code{Shared Object} files with @code{Runtime Linking} | |
1867 | enabled may bloat the TOC, eventually leading to @code{TOC overflow} errors, | |
1868 | requiring the use of either the @option{-Wl,-bbigtoc} linker flag (seen to | |
1869 | break with the @code{GDB} debugger) or some of the TOC-related compiler flags, | |
1870 | @ifnothtml | |
1871 | @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc, | |
1872 | Using the GNU Compiler Collection (GCC)}. | |
1873 | @end ifnothtml | |
1874 | @ifhtml | |
1875 | see ``RS/6000 and PowerPC Options'' in the main manual. | |
1876 | @end ifhtml | |
1877 | ||
1878 | @option{--with-aix-soname} is currently supported by @samp{libgcc_s} only, so | |
1879 | this option is still experimental and not for normal use yet. | |
1880 | ||
1881 | Default is the traditional behavior @option{--with-aix-soname=@samp{aix}}. | |
1882 | ||
1883 | @item --enable-languages=@var{lang1},@var{lang2},@dots{} | |
1884 | Specify that only a particular subset of compilers and | |
1885 | their runtime libraries should be built. For a list of valid values for | |
1886 | @var{langN} you can issue the following command in the | |
1887 | @file{gcc} directory of your GCC source tree:@* | |
1888 | @smallexample | |
1889 | grep ^language= */config-lang.in | |
1890 | @end smallexample | |
1891 | Currently, you can use any of the following: | |
1892 | @code{all}, @code{default}, @code{ada}, @code{c}, @code{c++}, @code{d}, | |
1eee94d3 GM |
1893 | @code{fortran}, @code{go}, @code{jit}, @code{lto}, @code{m2}, |
1894 | @code{objc}, @code{obj-c++}. | |
d77de738 ML |
1895 | Building the Ada compiler has special requirements, see below. |
1896 | If you do not pass this flag, or specify the option @code{default}, then the | |
1897 | default languages available in the @file{gcc} sub-tree will be configured. | |
1eee94d3 GM |
1898 | Ada, D, Go, Jit, Objective-C++ and Modula-2 are not default languages. |
1899 | LTO is not a | |
d77de738 ML |
1900 | default language, but is built by default because @option{--enable-lto} is |
1901 | enabled by default. The other languages are default languages. If | |
1902 | @code{all} is specified, then all available languages are built. An | |
1903 | exception is @code{jit} language, which requires | |
1904 | @option{--enable-host-shared} to be included with @code{all}. | |
1905 | ||
1906 | @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{} | |
1907 | Specify that a particular subset of compilers and their runtime | |
1908 | libraries should be built with the system C compiler during stage 1 of | |
1909 | the bootstrap process, rather than only in later stages with the | |
1910 | bootstrapped C compiler. The list of valid values is the same as for | |
1911 | @option{--enable-languages}, and the option @code{all} will select all | |
1912 | of the languages enabled by @option{--enable-languages}. This option is | |
1913 | primarily useful for GCC development; for instance, when a development | |
1914 | version of the compiler cannot bootstrap due to compiler bugs, or when | |
1915 | one is debugging front ends other than the C front end. When this | |
1916 | option is used, one can then build the target libraries for the | |
1917 | specified languages with the stage-1 compiler by using @command{make | |
1918 | stage1-bubble all-target}, or run the testsuite on the stage-1 compiler | |
1919 | for the specified languages using @command{make stage1-start check-gcc}. | |
1920 | ||
1921 | @item --disable-libada | |
1922 | Specify that the run-time libraries and tools used by GNAT should not | |
1923 | be built. This can be useful for debugging, or for compatibility with | |
1924 | previous Ada build procedures, when it was required to explicitly | |
1925 | do a @samp{make -C gcc gnatlib_and_tools}. | |
1926 | ||
1eee94d3 GM |
1927 | @item --disable-libgm2 |
1928 | Specify that the run-time libraries and tools used by Modula-2 should not | |
1929 | be built. This can be useful for debugging. | |
1930 | ||
d77de738 ML |
1931 | @item --disable-libsanitizer |
1932 | Specify that the run-time libraries for the various sanitizers should | |
1933 | not be built. | |
1934 | ||
1935 | @item --disable-libssp | |
1936 | Specify that the run-time libraries for stack smashing protection | |
1937 | should not be built or linked against. On many targets library support | |
1938 | is provided by the C library instead. | |
1939 | ||
1940 | @item --disable-libquadmath | |
1941 | Specify that the GCC quad-precision math library should not be built. | |
1942 | On some systems, the library is required to be linkable when building | |
1943 | the Fortran front end, unless @option{--disable-libquadmath-support} | |
1944 | is used. | |
1945 | ||
1946 | @item --disable-libquadmath-support | |
1947 | Specify that the Fortran front end and @code{libgfortran} do not add | |
1948 | support for @code{libquadmath} on systems supporting it. | |
1949 | ||
1950 | @item --disable-libgomp | |
1951 | Specify that the GNU Offloading and Multi Processing Runtime Library | |
1952 | should not be built. | |
1953 | ||
1954 | @item --disable-libvtv | |
1955 | Specify that the run-time libraries used by vtable verification | |
1956 | should not be built. | |
1957 | ||
1958 | @item --with-dwarf2 | |
1959 | Specify that the compiler should | |
a710f3ce EG |
1960 | use DWARF debugging information as the default; the exact |
1961 | DWARF version that is the default is target-specific. | |
d77de738 ML |
1962 | |
1963 | @item --with-advance-toolchain=@var{at} | |
1964 | On 64-bit PowerPC Linux systems, configure the compiler to use the | |
1965 | header files, library files, and the dynamic linker from the Advance | |
1966 | Toolchain release @var{at} instead of the default versions that are | |
1967 | provided by the Linux distribution. In general, this option is | |
1968 | intended for the developers of GCC, and it is not intended for general | |
1969 | use. | |
1970 | ||
1971 | @item --enable-targets=all | |
1972 | @itemx --enable-targets=@var{target_list} | |
1973 | Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers. | |
1974 | These are compilers that are able to generate either 64-bit or 32-bit | |
1975 | code. Typically, the corresponding 32-bit target, e.g.@: | |
1976 | powerpc-linux for powerpc64-linux, only generates 32-bit code. This | |
1977 | option enables the 32-bit target to be a bi-arch compiler, which is | |
1978 | useful when you want a bi-arch compiler that defaults to 32-bit, and | |
1979 | you are building a bi-arch or multi-arch binutils in a combined tree. | |
1980 | On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64), | |
1981 | defaulted to o32. | |
1982 | Currently, this option only affects sparc-linux, powerpc-linux, x86-linux, | |
1983 | mips-linux and s390-linux. | |
1984 | ||
1985 | @item --enable-default-pie | |
1986 | Turn on @option{-fPIE} and @option{-pie} by default. | |
1987 | ||
1988 | @item --enable-secureplt | |
1989 | This option enables @option{-msecure-plt} by default for powerpc-linux. | |
1990 | @ifnothtml | |
1991 | @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc, | |
1992 | Using the GNU Compiler Collection (GCC)}, | |
1993 | @end ifnothtml | |
1994 | @ifhtml | |
1995 | See ``RS/6000 and PowerPC Options'' in the main manual | |
1996 | @end ifhtml | |
1997 | ||
1998 | @item --enable-default-ssp | |
1999 | Turn on @option{-fstack-protector-strong} by default. | |
2000 | ||
2001 | @item --enable-cld | |
2002 | This option enables @option{-mcld} by default for 32-bit x86 targets. | |
2003 | @ifnothtml | |
2004 | @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc, | |
2005 | Using the GNU Compiler Collection (GCC)}, | |
2006 | @end ifnothtml | |
2007 | @ifhtml | |
2008 | See ``i386 and x86-64 Options'' in the main manual | |
2009 | @end ifhtml | |
2010 | ||
2011 | @item --enable-large-address-aware | |
2012 | The @option{--enable-large-address-aware} option arranges for MinGW | |
2013 | executables to be linked using the @option{--large-address-aware} | |
2014 | option, that enables the use of more than 2GB of memory. If GCC is | |
2015 | configured with this option, its effects can be reversed by passing the | |
2016 | @option{-Wl,--disable-large-address-aware} option to the so-configured | |
2017 | compiler driver. | |
2018 | ||
2019 | @item --enable-win32-registry | |
2020 | @itemx --enable-win32-registry=@var{key} | |
2021 | @itemx --disable-win32-registry | |
2022 | The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC | |
2023 | to look up installations paths in the registry using the following key: | |
2024 | ||
2025 | @smallexample | |
2026 | @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}} | |
2027 | @end smallexample | |
2028 | ||
2029 | @var{key} defaults to GCC version number, and can be overridden by the | |
2030 | @option{--enable-win32-registry=@var{key}} option. Vendors and distributors | |
2031 | who use custom installers are encouraged to provide a different key, | |
2032 | perhaps one comprised of vendor name and GCC version number, to | |
2033 | avoid conflict with existing installations. This feature is enabled | |
2034 | by default, and can be disabled by @option{--disable-win32-registry} | |
2035 | option. This option has no effect on the other hosts. | |
2036 | ||
2037 | @item --nfp | |
2038 | Specify that the machine does not have a floating point unit. This | |
2039 | option only applies to @samp{m68k-sun-sunos@var{n}}. On any other | |
2040 | system, @option{--nfp} has no effect. | |
2041 | ||
2042 | @item --enable-werror | |
2043 | @itemx --disable-werror | |
2044 | @itemx --enable-werror=yes | |
2045 | @itemx --enable-werror=no | |
2046 | When you specify this option, it controls whether certain files in the | |
2047 | compiler are built with @option{-Werror} in bootstrap stage2 and later. | |
2048 | If you don't specify it, @option{-Werror} is turned on for the main | |
2049 | development trunk. However it defaults to off for release branches and | |
2050 | final releases. The specific files which get @option{-Werror} are | |
2051 | controlled by the Makefiles. | |
2052 | ||
2053 | @item --enable-checking | |
2054 | @itemx --disable-checking | |
2055 | @itemx --enable-checking=@var{list} | |
2056 | This option controls performing internal consistency checks in the compiler. | |
2057 | It does not change the generated code, but adds error checking of the | |
2058 | requested complexity. This slows down the compiler and may only work | |
2059 | properly if you are building the compiler with GCC@. | |
2060 | ||
2061 | When the option is not specified, the active set of checks depends on context. | |
2062 | Namely, bootstrap stage 1 defaults to @samp{--enable-checking=yes}, builds | |
2063 | from release branches or release archives default to | |
2064 | @samp{--enable-checking=release}, and otherwise | |
2065 | @samp{--enable-checking=yes,extra} is used. When the option is | |
2066 | specified without a @var{list}, the result is the same as | |
2067 | @samp{--enable-checking=yes}. Likewise, @samp{--disable-checking} is | |
2068 | equivalent to @samp{--enable-checking=no}. | |
2069 | ||
2070 | The categories of checks available in @var{list} are @samp{yes} (most common | |
2071 | checks @samp{assert,misc,gc,gimple,rtlflag,runtime,tree,types}), @samp{no} | |
2072 | (no checks at all), @samp{all} (all but @samp{valgrind}), @samp{release} | |
2073 | (cheapest checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}). | |
2074 | @samp{release} checks are always on and to disable them | |
2075 | @samp{--disable-checking} or @samp{--enable-checking=no[,<other checks>]} | |
2076 | must be explicitly requested. Disabling assertions makes the compiler and | |
2077 | runtime slightly faster but increases the risk of undetected internal errors | |
2078 | causing wrong code to be generated. | |
2079 | ||
2080 | Individual checks can be enabled with these flags: @samp{assert}, @samp{df}, | |
2081 | @samp{extra}, @samp{fold}, @samp{gc}, @samp{gcac}, @samp{gimple}, | |
2082 | @samp{misc}, @samp{rtl}, @samp{rtlflag}, @samp{runtime}, @samp{tree}, | |
2083 | @samp{types} and @samp{valgrind}. @samp{extra} extends @samp{misc} | |
2084 | checking with extra checks that might affect code generation and should | |
2085 | therefore not differ between stage1 and later stages in bootstrap. | |
2086 | ||
2087 | The @samp{valgrind} check requires the external @command{valgrind} simulator, | |
2088 | available from @uref{https://valgrind.org}. The @samp{rtl} checks are | |
2089 | expensive and the @samp{df}, @samp{gcac} and @samp{valgrind} checks are very | |
2090 | expensive. | |
2091 | ||
2092 | @item --disable-stage1-checking | |
2093 | @itemx --enable-stage1-checking | |
2094 | @itemx --enable-stage1-checking=@var{list} | |
2095 | This option affects only bootstrap build. If no @option{--enable-checking} | |
2096 | option is specified the stage1 compiler is built with @samp{yes} checking | |
2097 | enabled, otherwise the stage1 checking flags are the same as specified by | |
2098 | @option{--enable-checking}. To build the stage1 compiler with | |
2099 | different checking options use @option{--enable-stage1-checking}. | |
2100 | The list of checking options is the same as for @option{--enable-checking}. | |
2101 | If your system is too slow or too small to bootstrap a released compiler | |
2102 | with checking for stage1 enabled, you can use @samp{--disable-stage1-checking} | |
2103 | to disable checking for the stage1 compiler. | |
2104 | ||
2105 | @item --enable-coverage | |
2106 | @itemx --enable-coverage=@var{level} | |
2107 | With this option, the compiler is built to collect self coverage | |
2108 | information, every time it is run. This is for internal development | |
2109 | purposes, and only works when the compiler is being built with gcc. The | |
2110 | @var{level} argument controls whether the compiler is built optimized or | |
2111 | not, values are @samp{opt} and @samp{noopt}. For coverage analysis you | |
2112 | want to disable optimization, for performance analysis you want to | |
2113 | enable optimization. When coverage is enabled, the default level is | |
2114 | without optimization. | |
2115 | ||
2116 | @item --enable-gather-detailed-mem-stats | |
2117 | When this option is specified more detailed information on memory | |
2118 | allocation is gathered. This information is printed when using | |
2119 | @option{-fmem-report}. | |
2120 | ||
2121 | @item --enable-valgrind-annotations | |
2122 | Mark selected memory related operations in the compiler when run under | |
2123 | valgrind to suppress false positives. | |
2124 | ||
2125 | @item --enable-nls | |
2126 | @itemx --disable-nls | |
2127 | The @option{--enable-nls} option enables Native Language Support (NLS), | |
2128 | which lets GCC output diagnostics in languages other than American | |
2129 | English. Native Language Support is enabled by default if not doing a | |
2130 | canadian cross build. The @option{--disable-nls} option disables NLS@. | |
2131 | ||
2132 | @item --with-included-gettext | |
2133 | If NLS is enabled, the @option{--with-included-gettext} option causes the build | |
2134 | procedure to prefer its copy of GNU @command{gettext}. | |
2135 | ||
2136 | @item --with-catgets | |
2137 | If NLS is enabled, and if the host lacks @code{gettext} but has the | |
2138 | inferior @code{catgets} interface, the GCC build procedure normally | |
2139 | ignores @code{catgets} and instead uses GCC's copy of the GNU | |
2140 | @code{gettext} library. The @option{--with-catgets} option causes the | |
2141 | build procedure to use the host's @code{catgets} in this situation. | |
2142 | ||
2143 | @item --with-libiconv-prefix=@var{dir} | |
2144 | Search for libiconv header files in @file{@var{dir}/include} and | |
2145 | libiconv library files in @file{@var{dir}/lib}. | |
2146 | ||
2147 | @item --enable-obsolete | |
2148 | Enable configuration for an obsoleted system. If you attempt to | |
2149 | configure GCC for a system (build, host, or target) which has been | |
2150 | obsoleted, and you do not specify this flag, configure will halt with an | |
2151 | error message. | |
2152 | ||
2153 | All support for systems which have been obsoleted in one release of GCC | |
2154 | is removed entirely in the next major release, unless someone steps | |
2155 | forward to maintain the port. | |
2156 | ||
2157 | @item --enable-decimal-float | |
2158 | @itemx --enable-decimal-float=yes | |
2159 | @itemx --enable-decimal-float=no | |
2160 | @itemx --enable-decimal-float=bid | |
2161 | @itemx --enable-decimal-float=dpd | |
2162 | @itemx --disable-decimal-float | |
2163 | Enable (or disable) support for the C decimal floating point extension | |
2164 | that is in the IEEE 754-2008 standard. This is enabled by default only | |
2165 | on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also | |
2166 | support it, but require the user to specifically enable it. You can | |
2167 | optionally control which decimal floating point format is used (either | |
2168 | @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal) | |
2169 | format is default on i386 and x86_64 systems, and the @samp{dpd} | |
2170 | (densely packed decimal) format is default on PowerPC systems. | |
2171 | ||
2172 | @item --enable-fixed-point | |
2173 | @itemx --disable-fixed-point | |
2174 | Enable (or disable) support for C fixed-point arithmetic. | |
2175 | This option is enabled by default for some targets (such as MIPS) which | |
2176 | have hardware-support for fixed-point operations. On other targets, you | |
2177 | may enable this option manually. | |
2178 | ||
2179 | @item --with-long-double-128 | |
2180 | Specify if @code{long double} type should be 128-bit by default on selected | |
2181 | GNU/Linux architectures. If using @code{--without-long-double-128}, | |
2182 | @code{long double} will be by default 64-bit, the same as @code{double} type. | |
2183 | When neither of these configure options are used, the default will be | |
2184 | 128-bit @code{long double} when built against GNU C Library 2.4 and later, | |
2185 | 64-bit @code{long double} otherwise. | |
2186 | ||
2187 | @item --with-long-double-format=ibm | |
2188 | @itemx --with-long-double-format=ieee | |
2189 | Specify whether @code{long double} uses the IBM extended double format | |
2190 | or the IEEE 128-bit floating point format on PowerPC Linux systems. | |
2191 | This configuration switch will only work on little endian PowerPC | |
2192 | Linux systems and on big endian 64-bit systems where the default cpu | |
2193 | is at least power7 (i.e.@: @option{--with-cpu=power7}, | |
2194 | @option{--with-cpu=power8}, or @option{--with-cpu=power9} is used). | |
2195 | ||
2196 | If you use the @option{--with-long-double-64} configuration option, | |
2197 | the @option{--with-long-double-format=ibm} and | |
2198 | @option{--with-long-double-format=ieee} options are ignored. | |
2199 | ||
2200 | The default @code{long double} format is to use IBM extended double. | |
2201 | Until all of the libraries are converted to use IEEE 128-bit floating | |
2202 | point, it is not recommended to use | |
2203 | @option{--with-long-double-format=ieee}. | |
2204 | ||
2205 | @item --enable-fdpic | |
2206 | On SH Linux systems, generate ELF FDPIC code. | |
2207 | ||
2208 | @item --with-gmp=@var{pathname} | |
2209 | @itemx --with-gmp-include=@var{pathname} | |
2210 | @itemx --with-gmp-lib=@var{pathname} | |
2211 | @itemx --with-mpfr=@var{pathname} | |
2212 | @itemx --with-mpfr-include=@var{pathname} | |
2213 | @itemx --with-mpfr-lib=@var{pathname} | |
2214 | @itemx --with-mpc=@var{pathname} | |
2215 | @itemx --with-mpc-include=@var{pathname} | |
2216 | @itemx --with-mpc-lib=@var{pathname} | |
2217 | If you want to build GCC but do not have the GMP library, the MPFR | |
2218 | library and/or the MPC library installed in a standard location and | |
2219 | do not have their sources present in the GCC source tree then you | |
2220 | can explicitly specify the directory where they are installed | |
2221 | (@samp{--with-gmp=@var{gmpinstalldir}}, | |
2222 | @samp{--with-mpfr=@/@var{mpfrinstalldir}}, | |
2223 | @samp{--with-mpc=@/@var{mpcinstalldir}}). The | |
2224 | @option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for | |
2225 | @option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and | |
2226 | @option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the | |
2227 | @option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for | |
2228 | @option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and | |
2229 | @option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the | |
2230 | @option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for | |
2231 | @option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and | |
2232 | @option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these | |
2233 | shorthand assumptions are not correct, you can use the explicit | |
2234 | include and lib options directly. You might also need to ensure the | |
2235 | shared libraries can be found by the dynamic linker when building and | |
2236 | using GCC, for example by setting the runtime shared library path | |
2237 | variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems). | |
2238 | ||
2239 | These flags are applicable to the host platform only. When building | |
2240 | a cross compiler, they will not be used to configure target libraries. | |
2241 | ||
2242 | @item --with-isl=@var{pathname} | |
2243 | @itemx --with-isl-include=@var{pathname} | |
2244 | @itemx --with-isl-lib=@var{pathname} | |
2245 | If you do not have the isl library installed in a standard location and you | |
2246 | want to build GCC, you can explicitly specify the directory where it is | |
2247 | installed (@samp{--with-isl=@/@var{islinstalldir}}). The | |
2248 | @option{--with-isl=@/@var{islinstalldir}} option is shorthand for | |
2249 | @option{--with-isl-lib=@/@var{islinstalldir}/lib} and | |
2250 | @option{--with-isl-include=@/@var{islinstalldir}/include}. If this | |
2251 | shorthand assumption is not correct, you can use the explicit | |
2252 | include and lib options directly. | |
2253 | ||
2254 | These flags are applicable to the host platform only. When building | |
2255 | a cross compiler, they will not be used to configure target libraries. | |
2256 | ||
2257 | @item --with-stage1-ldflags=@var{flags} | |
2258 | This option may be used to set linker flags to be used when linking | |
2259 | stage 1 of GCC. These are also used when linking GCC if configured with | |
2260 | @option{--disable-bootstrap}. If @option{--with-stage1-libs} is not set to a | |
2261 | value, then the default is @samp{-static-libstdc++ -static-libgcc}, if | |
2262 | supported. | |
2263 | ||
2264 | @item --with-stage1-libs=@var{libs} | |
2265 | This option may be used to set libraries to be used when linking stage 1 | |
2266 | of GCC. These are also used when linking GCC if configured with | |
2267 | @option{--disable-bootstrap}. | |
2268 | ||
2269 | @item --with-boot-ldflags=@var{flags} | |
2270 | This option may be used to set linker flags to be used when linking | |
2271 | stage 2 and later when bootstrapping GCC. If --with-boot-libs | |
2272 | is not is set to a value, then the default is | |
2273 | @samp{-static-libstdc++ -static-libgcc}. | |
2274 | ||
2275 | @item --with-boot-libs=@var{libs} | |
2276 | This option may be used to set libraries to be used when linking stage 2 | |
2277 | and later when bootstrapping GCC. | |
2278 | ||
2279 | @item --with-debug-prefix-map=@var{map} | |
2280 | Convert source directory names using @option{-fdebug-prefix-map} when | |
2281 | building runtime libraries. @samp{@var{map}} is a space-separated | |
2282 | list of maps of the form @samp{@var{old}=@var{new}}. | |
2283 | ||
2284 | @item --enable-linker-build-id | |
2285 | Tells GCC to pass @option{--build-id} option to the linker for all final | |
2286 | links (links performed without the @option{-r} or @option{--relocatable} | |
2287 | option), if the linker supports it. If you specify | |
2288 | @option{--enable-linker-build-id}, but your linker does not | |
2289 | support @option{--build-id} option, a warning is issued and the | |
2290 | @option{--enable-linker-build-id} option is ignored. The default is off. | |
2291 | ||
2292 | @item --with-linker-hash-style=@var{choice} | |
2293 | Tells GCC to pass @option{--hash-style=@var{choice}} option to the | |
2294 | linker for all final links. @var{choice} can be one of | |
2295 | @samp{sysv}, @samp{gnu}, and @samp{both} where @samp{sysv} is the default. | |
2296 | ||
2297 | @item --enable-gnu-unique-object | |
2298 | @itemx --disable-gnu-unique-object | |
2299 | Tells GCC to use the gnu_unique_object relocation for C++ template | |
2300 | static data members and inline function local statics. Enabled by | |
2301 | default for a toolchain with an assembler that accepts it and | |
2302 | GLIBC 2.11 or above, otherwise disabled. | |
2303 | ||
2304 | @item --with-diagnostics-color=@var{choice} | |
2305 | Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-color=} | |
2306 | option (if not used explicitly on the command line). @var{choice} | |
2307 | can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env} | |
2308 | where @samp{auto} is the default. @samp{auto-if-env} makes | |
2309 | @option{-fdiagnostics-color=auto} the default if @env{GCC_COLORS} | |
2310 | is present and non-empty in the environment of the compiler, and | |
2311 | @option{-fdiagnostics-color=never} otherwise. | |
2312 | ||
2313 | @item --with-diagnostics-urls=@var{choice} | |
2314 | Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-urls=} | |
2315 | option (if not used explicitly on the command line). @var{choice} | |
2316 | can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env} | |
2317 | where @samp{auto} is the default. @samp{auto-if-env} makes | |
2318 | @option{-fdiagnostics-urls=auto} the default if @env{GCC_URLS} | |
2319 | or @env{TERM_URLS} is present and non-empty in the environment of the | |
2320 | compiler, and @option{-fdiagnostics-urls=never} otherwise. | |
2321 | ||
2322 | @item --enable-lto | |
2323 | @itemx --disable-lto | |
2324 | Enable support for link-time optimization (LTO). This is enabled by | |
2325 | default, and may be disabled using @option{--disable-lto}. | |
2326 | ||
2327 | @item --enable-linker-plugin-configure-flags=FLAGS | |
2328 | @itemx --enable-linker-plugin-flags=FLAGS | |
2329 | By default, linker plugins (such as the LTO plugin) are built for the | |
2330 | host system architecture. For the case that the linker has a | |
2331 | different (but run-time compatible) architecture, these flags can be | |
2332 | specified to build plugins that are compatible to the linker. For | |
2333 | example, if you are building GCC for a 64-bit x86_64 | |
2334 | (@samp{x86_64-pc-linux-gnu}) host system, but have a 32-bit x86 | |
2335 | GNU/Linux (@samp{i686-pc-linux-gnu}) linker executable (which is | |
2336 | executable on the former system), you can configure GCC as follows for | |
2337 | getting compatible linker plugins: | |
2338 | ||
2339 | @smallexample | |
2340 | % @var{srcdir}/configure \ | |
2341 | --host=x86_64-pc-linux-gnu \ | |
2342 | --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \ | |
2343 | --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib' | |
2344 | @end smallexample | |
2345 | ||
2346 | @item --with-plugin-ld=@var{pathname} | |
2347 | Enable an alternate linker to be used at link-time optimization (LTO) | |
2348 | link time when @option{-fuse-linker-plugin} is enabled. | |
2349 | This linker should have plugin support such as gold starting with | |
2350 | version 2.20 or GNU ld starting with version 2.21. | |
2351 | See @option{-fuse-linker-plugin} for details. | |
2352 | ||
2353 | @item --enable-canonical-system-headers | |
2354 | @itemx --disable-canonical-system-headers | |
2355 | Enable system header path canonicalization for @file{libcpp}. This can | |
2356 | produce shorter header file paths in diagnostics and dependency output | |
2357 | files, but these changed header paths may conflict with some compilation | |
2358 | environments. Enabled by default, and may be disabled using | |
2359 | @option{--disable-canonical-system-headers}. | |
2360 | ||
2361 | @item --with-glibc-version=@var{major}.@var{minor} | |
2362 | Tell GCC that when the GNU C Library (glibc) is used on the target it | |
2363 | will be version @var{major}.@var{minor} or later. Normally this can | |
2364 | be detected from the C library's header files, but this option may be | |
2365 | needed when bootstrapping a cross toolchain without the header files | |
2366 | available for building the initial bootstrap compiler. | |
2367 | ||
2368 | If GCC is configured with some multilibs that use glibc and some that | |
2369 | do not, this option applies only to the multilibs that use glibc. | |
2370 | However, such configurations may not work well as not all the relevant | |
2371 | configuration in GCC is on a per-multilib basis. | |
2372 | ||
2373 | @item --enable-as-accelerator-for=@var{target} | |
2374 | Build as offload target compiler. Specify offload host triple by @var{target}. | |
2375 | ||
2376 | @item --enable-offload-targets=@var{target1}[=@var{path1}],@dots{},@var{targetN}[=@var{pathN}] | |
2377 | Enable offloading to targets @var{target1}, @dots{}, @var{targetN}. | |
2378 | Offload compilers are expected to be already installed. Default search | |
2379 | path for them is @file{@var{exec-prefix}}, but it can be changed by | |
2380 | specifying paths @var{path1}, @dots{}, @var{pathN}. | |
2381 | ||
2382 | @smallexample | |
2383 | % @var{srcdir}/configure \ | |
2384 | --enable-offload-targets=amdgcn-amdhsa,nvptx-none | |
2385 | @end smallexample | |
2386 | ||
2387 | @item --enable-offload-defaulted | |
2388 | ||
2389 | Tell GCC that configured but not installed offload compilers and libgomp | |
2390 | plugins are silently ignored. Useful for distribution compilers where | |
2391 | those are in separate optional packages and where the presence or absence | |
2392 | of those optional packages should determine the actual supported offloading | |
2393 | target set rather than the GCC configure-time selection. | |
2394 | ||
2395 | @item --enable-cet | |
2396 | @itemx --disable-cet | |
2397 | Enable building target run-time libraries with control-flow | |
2398 | instrumentation, see @option{-fcf-protection} option. When | |
2399 | @code{--enable-cet} is specified target libraries are configured | |
2400 | to add @option{-fcf-protection} and, if needed, other target | |
2401 | specific options to a set of building options. | |
2402 | ||
2403 | @code{--enable-cet=auto} is default. CET is enabled on Linux/x86 if | |
2404 | target binutils supports @code{Intel CET} instructions and disabled | |
2405 | otherwise. In this case, the target libraries are configured to get | |
2406 | additional @option{-fcf-protection} option. | |
2407 | ||
2408 | @item --with-riscv-attribute=@samp{yes}, @samp{no} or @samp{default} | |
2409 | Generate RISC-V attribute by default, in order to record extra build | |
2410 | information in object. | |
2411 | ||
2412 | The option is disabled by default. It is enabled on RISC-V/ELF (bare-metal) | |
2413 | target if target binutils supported. | |
2414 | ||
2415 | @item --enable-s390-excess-float-precision | |
2416 | @itemx --disable-s390-excess-float-precision | |
2417 | On s390(x) targets, enable treatment of float expressions with double precision | |
2418 | when in standards-compliant mode (e.g., when @code{--std=c99} or | |
2419 | @code{-fexcess-precision=standard} are given). | |
2420 | ||
2421 | For a native build and cross compiles that have target headers, the option's | |
2422 | default is derived from glibc's behavior. When glibc clamps float_t to double, | |
2423 | GCC follows and enables the option. For other cross compiles, the default is | |
2424 | disabled. | |
2425 | ||
2426 | @item --with-zstd=@var{pathname} | |
2427 | @itemx --with-zstd-include=@var{pathname} | |
2428 | @itemx --with-zstd-lib=@var{pathname} | |
2429 | If you do not have the @code{zstd} library installed in a standard | |
2430 | location and you want to build GCC, you can explicitly specify the | |
2431 | directory where it is installed (@samp{--with-zstd=@/@var{zstdinstalldir}}). | |
2432 | The @option{--with-zstd=@/@var{zstdinstalldir}} option is shorthand for | |
2433 | @option{--with-zstd-lib=@/@var{zstdinstalldir}/lib} and | |
2434 | @option{--with-zstd-include=@/@var{zstdinstalldir}/include}. If this | |
2435 | shorthand assumption is not correct, you can use the explicit | |
2436 | include and lib options directly. | |
2437 | ||
2438 | These flags are applicable to the host platform only. When building | |
2439 | a cross compiler, they will not be used to configure target libraries. | |
2440 | @end table | |
2441 | ||
2442 | @subheading Cross-Compiler-Specific Options | |
2443 | The following options only apply to building cross compilers. | |
2444 | ||
2445 | @table @code | |
2446 | @item --with-toolexeclibdir=@var{dir} | |
2447 | Specify the installation directory for libraries built with a cross compiler. | |
2448 | The default is @option{$@{gcc_tooldir@}/lib}. | |
2449 | ||
2450 | @item --with-sysroot | |
2451 | @itemx --with-sysroot=@var{dir} | |
2452 | Tells GCC to consider @var{dir} as the root of a tree that contains | |
2453 | (a subset of) the root filesystem of the target operating system. | |
2454 | Target system headers, libraries and run-time object files will be | |
2455 | searched for in there. More specifically, this acts as if | |
2456 | @option{--sysroot=@var{dir}} was added to the default options of the built | |
2457 | compiler. The specified directory is not copied into the | |
2458 | install tree, unlike the options @option{--with-headers} and | |
2459 | @option{--with-libs} that this option obsoletes. The default value, | |
2460 | in case @option{--with-sysroot} is not given an argument, is | |
2461 | @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a | |
2462 | subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to | |
2463 | the GCC binaries if the installation tree is moved. | |
2464 | ||
2465 | This option affects the system root for the compiler used to build | |
2466 | target libraries (which runs on the build system) and the compiler newly | |
2467 | installed with @code{make install}; it does not affect the compiler which is | |
2468 | used to build GCC itself. | |
2469 | ||
2470 | If you specify the @option{--with-native-system-header-dir=@var{dirname}} | |
2471 | option then the compiler will search that directory within @var{dirname} for | |
2472 | native system headers rather than the default @file{/usr/include}. | |
2473 | ||
2474 | @item --with-build-sysroot | |
2475 | @itemx --with-build-sysroot=@var{dir} | |
2476 | Tells GCC to consider @var{dir} as the system root (see | |
2477 | @option{--with-sysroot}) while building target libraries, instead of | |
2478 | the directory specified with @option{--with-sysroot}. This option is | |
2479 | only useful when you are already using @option{--with-sysroot}. You | |
2480 | can use @option{--with-build-sysroot} when you are configuring with | |
2481 | @option{--prefix} set to a directory that is different from the one in | |
2482 | which you are installing GCC and your target libraries. | |
2483 | ||
2484 | This option affects the system root for the compiler used to build | |
2485 | target libraries (which runs on the build system); it does not affect | |
2486 | the compiler which is used to build GCC itself. | |
2487 | ||
2488 | If you specify the @option{--with-native-system-header-dir=@var{dirname}} | |
2489 | option then the compiler will search that directory within @var{dirname} for | |
2490 | native system headers rather than the default @file{/usr/include}. | |
2491 | ||
2492 | @item --with-headers | |
2493 | @itemx --with-headers=@var{dir} | |
2494 | Deprecated in favor of @option{--with-sysroot}. | |
2495 | Specifies that target headers are available when building a cross compiler. | |
2496 | The @var{dir} argument specifies a directory which has the target include | |
2497 | files. These include files will be copied into the @file{gcc} install | |
2498 | directory. @emph{This option with the @var{dir} argument is required} when | |
2499 | building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include} | |
2500 | doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does | |
2501 | pre-exist, the @var{dir} argument may be omitted. @command{fixincludes} | |
2502 | will be run on these files to make them compatible with GCC@. | |
2503 | ||
2504 | @item --without-headers | |
2505 | Tells GCC not use any target headers from a libc when building a cross | |
2506 | compiler. When crossing to GNU/Linux, you need the headers so GCC | |
2507 | can build the exception handling for libgcc. | |
2508 | ||
2509 | @item --with-libs | |
2510 | @itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}" | |
2511 | Deprecated in favor of @option{--with-sysroot}. | |
2512 | Specifies a list of directories which contain the target runtime | |
2513 | libraries. These libraries will be copied into the @file{gcc} install | |
2514 | directory. If the directory list is omitted, this option has no | |
2515 | effect. | |
2516 | ||
2517 | @item --with-newlib | |
2518 | Specifies that @samp{newlib} is | |
2519 | being used as the target C library. This causes @code{__eprintf} to be | |
2520 | omitted from @file{libgcc.a} on the assumption that it will be provided by | |
2521 | @samp{newlib}. | |
2522 | ||
2523 | @html | |
2524 | <a name="avr"></a> | |
2525 | @end html | |
2526 | @item --with-avrlibc | |
2527 | Only supported for the AVR target. Specifies that @samp{AVR-Libc} is | |
2528 | being used as the target C@tie{} library. This causes float support | |
2529 | functions like @code{__addsf3} to be omitted from @file{libgcc.a} on | |
2530 | the assumption that it will be provided by @file{libm.a}. For more | |
2531 | technical details, cf. @uref{https://gcc.gnu.org/PR54461,,PR54461}. | |
2532 | It is not supported for | |
2533 | RTEMS configurations, which currently use newlib. The option is | |
2534 | supported since version 4.7.2 and is the default in 4.8.0 and newer. | |
2535 | ||
2536 | @item --with-double=@{32|64|32,64|64,32@} | |
2537 | @itemx --with-long-double=@{32|64|32,64|64,32|double@} | |
2538 | Only supported for the AVR target since version@tie{}10. | |
2539 | Specify the default layout available for the C/C++ @samp{double} | |
2540 | and @samp{long double} type, respectively. The following rules apply: | |
2541 | @itemize | |
2542 | @item | |
2543 | The first value after the @samp{=} specifies the default layout (in bits) | |
2544 | of the type and also the default for the @option{-mdouble=} resp. | |
2545 | @option{-mlong-double=} compiler option. | |
2546 | @item | |
2547 | If more than one value is specified, respective multilib variants are | |
2548 | available, and @option{-mdouble=} resp. @option{-mlong-double=} acts | |
2549 | as a multilib option. | |
2550 | @item | |
2551 | If @option{--with-long-double=double} is specified, @samp{double} and | |
2552 | @samp{long double} will have the same layout. | |
2553 | @item | |
2554 | The defaults are @option{--with-long-double=64,32} and | |
2555 | @option{--with-double=32,64}. The default @samp{double} layout imposed by | |
2556 | the latter is compatible with older versions of the compiler that implement | |
2557 | @samp{double} as a 32-bit type, which does not comply to the language standard. | |
2558 | @end itemize | |
2559 | Not all combinations of @option{--with-double=} and | |
2560 | @option{--with-long-double=} are valid. For example, the combination | |
2561 | @option{--with-double=32,64} @option{--with-long-double=32} will be | |
2562 | rejected because the first option specifies the availability of | |
2563 | multilibs for @samp{double}, whereas the second option implies | |
2564 | that @samp{long double} --- and hence also @samp{double} --- is always | |
2565 | 32@tie{}bits wide. | |
2566 | ||
2567 | @item --with-double-comparison=@{tristate|bool|libf7@} | |
2568 | Only supported for the AVR target since version@tie{}10. | |
2569 | Specify what result format is returned by library functions that | |
2570 | compare 64-bit floating point values (@code{DFmode}). | |
2571 | The GCC default is @samp{tristate}. If the floating point | |
2572 | implementation returns a boolean instead, set it to @samp{bool}. | |
2573 | ||
2574 | @item --with-libf7=@{libgcc|math|math-symbols|no@} | |
2575 | Only supported for the AVR target since version@tie{}10. | |
2576 | Specify to which degree code from LibF7 is included in libgcc. | |
2577 | LibF7 is an ad-hoc, AVR-specific, 64-bit floating point emulation | |
2578 | written in C and (inline) assembly. @samp{libgcc} adds support | |
2579 | for functions that one would usually expect in libgcc like double addition, | |
2580 | double comparisons and double conversions. @samp{math} also adds routines | |
2581 | that one would expect in @file{libm.a}, but with @code{__} (two underscores) | |
2582 | prepended to the symbol names as specified by @file{math.h}. | |
2583 | @samp{math-symbols} also defines weak aliases for the functions | |
2584 | declared in @file{math.h}. However, @code{--with-libf7} won't | |
2585 | install no @file{math.h} header file whatsoever, this file must come | |
2586 | from elsewhere. This option sets @option{--with-double-comparison} | |
2587 | to @samp{bool}. | |
2588 | ||
2589 | @item --with-nds32-lib=@var{library} | |
2590 | Specifies that @var{library} setting is used for building @file{libgcc.a}. | |
2591 | Currently, the valid @var{library} is @samp{newlib} or @samp{mculib}. | |
2592 | This option is only supported for the NDS32 target. | |
2593 | ||
2594 | @item --with-build-time-tools=@var{dir} | |
2595 | Specifies where to find the set of target tools (assembler, linker, etc.) | |
2596 | that will be used while building GCC itself. This option can be useful | |
2597 | if the directory layouts are different between the system you are building | |
2598 | GCC on, and the system where you will deploy it. | |
2599 | ||
2600 | For example, on an @samp{ia64-hp-hpux} system, you may have the GNU | |
2601 | assembler and linker in @file{/usr/bin}, and the native tools in a | |
2602 | different path, and build a toolchain that expects to find the | |
2603 | native tools in @file{/usr/bin}. | |
2604 | ||
2605 | When you use this option, you should ensure that @var{dir} includes | |
2606 | @command{ar}, @command{as}, @command{ld}, @command{nm}, | |
2607 | @command{ranlib} and @command{strip} if necessary, and possibly | |
2608 | @command{objdump}. Otherwise, GCC may use an inconsistent set of | |
2609 | tools. | |
2610 | @end table | |
2611 | ||
2612 | @subsubheading Overriding @command{configure} test results | |
2613 | ||
2614 | Sometimes, it might be necessary to override the result of some | |
2615 | @command{configure} test, for example in order to ease porting to a new | |
2616 | system or work around a bug in a test. The toplevel @command{configure} | |
2617 | script provides three variables for this: | |
2618 | ||
2619 | @table @code | |
2620 | ||
2621 | @item build_configargs | |
2622 | @cindex @code{build_configargs} | |
2623 | The contents of this variable is passed to all build @command{configure} | |
2624 | scripts. | |
2625 | ||
2626 | @item host_configargs | |
2627 | @cindex @code{host_configargs} | |
2628 | The contents of this variable is passed to all host @command{configure} | |
2629 | scripts. | |
2630 | ||
2631 | @item target_configargs | |
2632 | @cindex @code{target_configargs} | |
2633 | The contents of this variable is passed to all target @command{configure} | |
2634 | scripts. | |
2635 | ||
2636 | @end table | |
2637 | ||
2638 | In order to avoid shell and @command{make} quoting issues for complex | |
2639 | overrides, you can pass a setting for @env{CONFIG_SITE} and set | |
2640 | variables in the site file. | |
2641 | ||
2642 | @subheading Objective-C-Specific Options | |
2643 | ||
2644 | The following options apply to the build of the Objective-C runtime library. | |
2645 | ||
2646 | @table @code | |
2647 | @item --enable-objc-gc | |
2648 | Specify that an additional variant of the GNU Objective-C runtime library | |
2649 | is built, using an external build of the Boehm-Demers-Weiser garbage | |
2650 | collector (@uref{https://www.hboehm.info/gc/}). This library needs to be | |
2651 | available for each multilib variant, unless configured with | |
2652 | @option{--enable-objc-gc=@samp{auto}} in which case the build of the | |
2653 | additional runtime library is skipped when not available and the build | |
2654 | continues. | |
2655 | ||
2656 | @item --with-target-bdw-gc=@var{list} | |
2657 | @itemx --with-target-bdw-gc-include=@var{list} | |
2658 | @itemx --with-target-bdw-gc-lib=@var{list} | |
2659 | Specify search directories for the garbage collector header files and | |
2660 | libraries. @var{list} is a comma separated list of key value pairs of the | |
2661 | form @samp{@var{multilibdir}=@var{path}}, where the default multilib key | |
2662 | is named as @samp{.} (dot), or is omitted (e.g.@: | |
2663 | @samp{--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32}). | |
2664 | ||
2665 | The options @option{--with-target-bdw-gc-include} and | |
2666 | @option{--with-target-bdw-gc-lib} must always be specified together | |
2667 | for each multilib variant and they take precedence over | |
2668 | @option{--with-target-bdw-gc}. If @option{--with-target-bdw-gc-include} | |
2669 | is missing values for a multilib, then the value for the default | |
2670 | multilib is used (e.g.@: @samp{--with-target-bdw-gc-include=/opt/bdw-gc/include} | |
2671 | @samp{--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32}). | |
2672 | If none of these options are specified, the library is assumed in | |
2673 | default locations. | |
2674 | @end table | |
2675 | ||
2676 | @subheading D-Specific Options | |
2677 | ||
2678 | The following options apply to the build of the D runtime library. | |
2679 | ||
2680 | @table @code | |
2681 | @item --enable-libphobos-checking | |
2682 | @itemx --disable-libphobos-checking | |
2683 | @itemx --enable-libphobos-checking=@var{list} | |
2684 | This option controls whether run-time checks and contracts are compiled into | |
2685 | the D runtime library. When the option is not specified, the library is built | |
2686 | with @samp{release} checking. When the option is specified without a | |
2687 | @var{list}, the result is the same as @samp{--enable-libphobos-checking=yes}. | |
2688 | Likewise, @samp{--disable-libphobos-checking} is equivalent to | |
2689 | @samp{--enable-libphobos-checking=no}. | |
2690 | ||
2691 | The categories of checks available in @var{list} are @samp{yes} (compiles | |
2692 | libphobos with @option{-fno-release}), @samp{no} (compiles libphobos with | |
2693 | @option{-frelease}), @samp{all} (same as @samp{yes}), @samp{none} or | |
2694 | @samp{release} (same as @samp{no}). | |
2695 | ||
2696 | Individual checks available in @var{list} are @samp{assert} (compiles libphobos | |
2697 | with an extra option @option{-fassert}). | |
2698 | ||
2699 | @item --with-libphobos-druntime-only | |
2700 | @itemx --with-libphobos-druntime-only=@var{choice} | |
2701 | Specify whether to build only the core D runtime library (druntime), or both | |
2702 | the core and standard library (phobos) into libphobos. This is useful for | |
2703 | targets that have full support in druntime, but no or incomplete support | |
2704 | in phobos. @var{choice} can be one of @samp{auto}, @samp{yes}, and @samp{no} | |
2705 | where @samp{auto} is the default. | |
2706 | ||
2707 | When the option is not specified, the default choice @samp{auto} means that it | |
2708 | is inferred whether the target has support for the phobos standard library. | |
2709 | When the option is specified without a @var{choice}, the result is the same as | |
2710 | @samp{--with-libphobos-druntime-only=yes}. | |
2711 | ||
2712 | @item --with-target-system-zlib | |
2713 | Use installed @samp{zlib} rather than that included with GCC@. This needs | |
2714 | to be available for each multilib variant, unless configured with | |
2715 | @option{--with-target-system-zlib=@samp{auto}} in which case the GCC@ included | |
2716 | @samp{zlib} is only used when the system installed library is not available. | |
2717 | @end table | |
2718 | ||
2719 | @html | |
2720 | <hr /> | |
2721 | <p> | |
2722 | @end html | |
2723 | @ifhtml | |
2724 | @uref{./index.html,,Return to the GCC Installation page} | |
2725 | @end ifhtml | |
2726 | @end ifset | |
2727 | ||
2728 | @c ***Building**************************************************************** | |
2729 | @ifnothtml | |
2730 | @comment node-name, next, previous, up | |
2731 | @node Building, Testing, Configuration, Installing GCC | |
2732 | @end ifnothtml | |
2733 | @ifset buildhtml | |
2734 | @ifnothtml | |
2735 | @chapter Building | |
2736 | @end ifnothtml | |
2737 | @cindex Installing GCC: Building | |
2738 | ||
2739 | Now that GCC is configured, you are ready to build the compiler and | |
2740 | runtime libraries. | |
2741 | ||
2742 | Some commands executed when making the compiler may fail (return a | |
2743 | nonzero status) and be ignored by @command{make}. These failures, which | |
2744 | are often due to files that were not found, are expected, and can safely | |
2745 | be ignored. | |
2746 | ||
2747 | It is normal to have compiler warnings when compiling certain files. | |
2748 | Unless you are a GCC developer, you can generally ignore these warnings | |
2749 | unless they cause compilation to fail. Developers should attempt to fix | |
2750 | any warnings encountered, however they can temporarily continue past | |
2751 | warnings-as-errors by specifying the configure flag | |
2752 | @option{--disable-werror}. | |
2753 | ||
2754 | On certain old systems, defining certain environment variables such as | |
2755 | @env{CC} can interfere with the functioning of @command{make}. | |
2756 | ||
2757 | If you encounter seemingly strange errors when trying to build the | |
2758 | compiler in a directory other than the source directory, it could be | |
2759 | because you have previously configured the compiler in the source | |
2760 | directory. Make sure you have done all the necessary preparations. | |
2761 | ||
2762 | If you build GCC on a BSD system using a directory stored in an old System | |
2763 | V file system, problems may occur in running @command{fixincludes} if the | |
2764 | System V file system doesn't support symbolic links. These problems | |
2765 | result in a failure to fix the declaration of @code{size_t} in | |
2766 | @file{sys/types.h}. If you find that @code{size_t} is a signed type and | |
2767 | that type mismatches occur, this could be the cause. | |
2768 | ||
2769 | The solution is not to use such a directory for building GCC@. | |
2770 | ||
2771 | Similarly, when building from the source repository or snapshots, or if you modify | |
2772 | @file{*.l} files, you need the Flex lexical analyzer generator | |
2773 | installed. If you do not modify @file{*.l} files, releases contain | |
2774 | the Flex-generated files and you do not need Flex installed to build | |
2775 | them. There is still one Flex-based lexical analyzer (part of the | |
2776 | build machinery, not of GCC itself) that is used even if you only | |
2777 | build the C front end. | |
2778 | ||
2779 | When building from the source repository or snapshots, or if you modify Texinfo | |
2780 | documentation, you need version 4.7 or later of Texinfo installed if you | |
2781 | want Info documentation to be regenerated. Releases contain Info | |
2782 | documentation pre-built for the unmodified documentation in the release. | |
2783 | ||
2784 | @section Building a native compiler | |
2785 | ||
2786 | For a native build, the default configuration is to perform | |
2787 | a 3-stage bootstrap of the compiler when @samp{make} is invoked. | |
2788 | This will build the entire GCC system and ensure that it compiles | |
2789 | itself correctly. It can be disabled with the @option{--disable-bootstrap} | |
2790 | parameter to @samp{configure}, but bootstrapping is suggested because | |
2791 | the compiler will be tested more completely and could also have | |
2792 | better performance. | |
2793 | ||
2794 | The bootstrapping process will complete the following steps: | |
2795 | ||
2796 | @itemize @bullet | |
2797 | @item | |
2798 | Build tools necessary to build the compiler. | |
2799 | ||
2800 | @item | |
2801 | Perform a 3-stage bootstrap of the compiler. This includes building | |
2802 | three times the target tools for use by the compiler such as binutils | |
2803 | (bfd, binutils, gas, gprof, ld, and opcodes) if they have been | |
2804 | individually linked or moved into the top level GCC source tree before | |
2805 | configuring. | |
2806 | ||
2807 | @item | |
2808 | Perform a comparison test of the stage2 and stage3 compilers. | |
2809 | ||
2810 | @item | |
2811 | Build runtime libraries using the stage3 compiler from the previous step. | |
2812 | ||
2813 | @end itemize | |
2814 | ||
2815 | If you are short on disk space you might consider @samp{make | |
2816 | bootstrap-lean} instead. The sequence of compilation is the | |
2817 | same described above, but object files from the stage1 and | |
2818 | stage2 of the 3-stage bootstrap of the compiler are deleted as | |
2819 | soon as they are no longer needed. | |
2820 | ||
2821 | If you wish to use non-default GCC flags when compiling the stage2 | |
2822 | and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when | |
2823 | doing @samp{make}. For example, if you want to save additional space | |
2824 | during the bootstrap and in the final installation as well, you can | |
2825 | build the compiler binaries without debugging information as in the | |
2826 | following example. This will save roughly 40% of disk space both for | |
2827 | the bootstrap and the final installation. (Libraries will still contain | |
2828 | debugging information.) | |
2829 | ||
2830 | @smallexample | |
2831 | make BOOT_CFLAGS='-O' bootstrap | |
2832 | @end smallexample | |
2833 | ||
2834 | You can place non-default optimization flags into @code{BOOT_CFLAGS}; they | |
2835 | are less well tested here than the default of @samp{-g -O2}, but should | |
2836 | still work. In a few cases, you may find that you need to specify special | |
2837 | flags such as @option{-msoft-float} here to complete the bootstrap; or, | |
2838 | if the native compiler miscompiles the stage1 compiler, you may need | |
2839 | to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts | |
2840 | of the stage1 compiler that were miscompiled, or by using @samp{make | |
2841 | bootstrap4} to increase the number of stages of bootstrap. | |
2842 | ||
2843 | @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries. | |
2844 | Since these are always compiled with the compiler currently being | |
2845 | bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their | |
2846 | compilation flags, as for non-bootstrapped target libraries. | |
2847 | Again, if the native compiler miscompiles the stage1 compiler, you may | |
2848 | need to work around this by avoiding non-working parts of the stage1 | |
2849 | compiler. Use @code{STAGE1_TFLAGS} to this end. | |
2850 | ||
2851 | If you used the flag @option{--enable-languages=@dots{}} to restrict | |
2852 | the compilers to be built, only those you've actually enabled will be | |
2853 | built. This will of course only build those runtime libraries, for | |
2854 | which the particular compiler has been built. Please note, | |
2855 | that re-defining @env{LANGUAGES} when calling @samp{make} | |
2856 | @strong{does not} work anymore! | |
2857 | ||
2858 | If the comparison of stage2 and stage3 fails, this normally indicates | |
2859 | that the stage2 compiler has compiled GCC incorrectly, and is therefore | |
2860 | a potentially serious bug which you should investigate and report. (On | |
2861 | a few systems, meaningful comparison of object files is impossible; they | |
2862 | always appear ``different''. If you encounter this problem, you will | |
2863 | need to disable comparison in the @file{Makefile}.) | |
2864 | ||
2865 | If you do not want to bootstrap your compiler, you can configure with | |
2866 | @option{--disable-bootstrap}. In particular cases, you may want to | |
2867 | bootstrap your compiler even if the target system is not the same as | |
2868 | the one you are building on: for example, you could build a | |
2869 | @code{powerpc-unknown-linux-gnu} toolchain on a | |
2870 | @code{powerpc64-unknown-linux-gnu} host. In this case, pass | |
2871 | @option{--enable-bootstrap} to the configure script. | |
2872 | ||
2873 | @code{BUILD_CONFIG} can be used to bring in additional customization | |
2874 | to the build. It can be set to a whitespace-separated list of names. | |
2875 | For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will | |
2876 | be included by the top-level @file{Makefile}, bringing in any settings | |
2877 | it contains. The default @code{BUILD_CONFIG} can be set using the | |
2878 | configure option @option{--with-build-config=@code{NAME}...}. Some | |
2879 | examples of supported build configurations are: | |
2880 | ||
2881 | @table @asis | |
2882 | @item @samp{bootstrap-O1} | |
2883 | Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds | |
2884 | @option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to | |
2885 | @samp{BOOT_CFLAGS='-g -O1'}. | |
2886 | ||
2887 | @item @samp{bootstrap-O3} | |
2888 | @itemx @samp{bootstrap-Og} | |
2889 | Analogous to @code{bootstrap-O1}. | |
2890 | ||
2891 | @item @samp{bootstrap-lto} | |
2892 | Enables Link-Time Optimization for host tools during bootstrapping. | |
2893 | @samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding | |
2894 | @option{-flto} to @samp{BOOT_CFLAGS}. This option assumes that the host | |
2895 | supports the linker plugin (e.g.@: GNU ld version 2.21 or later or GNU gold | |
2896 | version 2.21 or later). | |
2897 | ||
2898 | @item @samp{bootstrap-lto-noplugin} | |
2899 | This option is similar to @code{bootstrap-lto}, but is intended for | |
2900 | hosts that do not support the linker plugin. Without the linker plugin | |
2901 | static libraries are not compiled with link-time optimizations. Since | |
2902 | the GCC middle end and back end are in @file{libbackend.a} this means | |
2903 | that only the front end is actually LTO optimized. | |
2904 | ||
2905 | @item @samp{bootstrap-lto-lean} | |
2906 | This option is similar to @code{bootstrap-lto}, but is intended for | |
2907 | faster build by only using LTO in the final bootstrap stage. | |
2908 | With @samp{make profiledbootstrap} the LTO frontend | |
2909 | is trained only on generator files. | |
2910 | ||
2911 | @item @samp{bootstrap-debug} | |
2912 | Verifies that the compiler generates the same executable code, whether | |
2913 | or not it is asked to emit debug information. To this end, this | |
2914 | option builds stage2 host programs without debug information, and uses | |
2915 | @file{contrib/compare-debug} to compare them with the stripped stage3 | |
2916 | object files. If @code{BOOT_CFLAGS} is overridden so as to not enable | |
2917 | debug information, stage2 will have it, and stage3 won't. This option | |
2918 | is enabled by default when GCC bootstrapping is enabled, if | |
2919 | @code{strip} can turn object files compiled with and without debug | |
2920 | info into identical object files. In addition to better test | |
2921 | coverage, this option makes default bootstraps faster and leaner. | |
2922 | ||
2923 | @item @samp{bootstrap-debug-big} | |
2924 | Rather than comparing stripped object files, as in | |
2925 | @code{bootstrap-debug}, this option saves internal compiler dumps | |
2926 | during stage2 and stage3 and compares them as well, which helps catch | |
2927 | additional potential problems, but at a great cost in terms of disk | |
2928 | space. It can be specified in addition to @samp{bootstrap-debug}. | |
2929 | ||
2930 | @item @samp{bootstrap-debug-lean} | |
2931 | This option saves disk space compared with @code{bootstrap-debug-big}, | |
2932 | but at the expense of some recompilation. Instead of saving the dumps | |
2933 | of stage2 and stage3 until the final compare, it uses | |
2934 | @option{-fcompare-debug} to generate, compare and remove the dumps | |
2935 | during stage3, repeating the compilation that already took place in | |
2936 | stage2, whose dumps were not saved. | |
2937 | ||
2938 | @item @samp{bootstrap-debug-lib} | |
2939 | This option tests executable code invariance over debug information | |
2940 | generation on target libraries, just like @code{bootstrap-debug-lean} | |
2941 | tests it on host programs. It builds stage3 libraries with | |
2942 | @option{-fcompare-debug}, and it can be used along with any of the | |
2943 | @code{bootstrap-debug} options above. | |
2944 | ||
2945 | There aren't @code{-lean} or @code{-big} counterparts to this option | |
2946 | because most libraries are only build in stage3, so bootstrap compares | |
2947 | would not get significant coverage. Moreover, the few libraries built | |
2948 | in stage2 are used in stage3 host programs, so we wouldn't want to | |
2949 | compile stage2 libraries with different options for comparison purposes. | |
2950 | ||
2951 | @item @samp{bootstrap-debug-ckovw} | |
2952 | Arranges for error messages to be issued if the compiler built on any | |
2953 | stage is run without the option @option{-fcompare-debug}. This is | |
2954 | useful to verify the full @option{-fcompare-debug} testing coverage. It | |
2955 | must be used along with @code{bootstrap-debug-lean} and | |
2956 | @code{bootstrap-debug-lib}. | |
2957 | ||
2958 | @item @samp{bootstrap-cet} | |
2959 | This option enables Intel CET for host tools during bootstrapping. | |
2960 | @samp{BUILD_CONFIG=bootstrap-cet} is equivalent to adding | |
2961 | @option{-fcf-protection} to @samp{BOOT_CFLAGS}. This option | |
2962 | assumes that the host supports Intel CET (e.g.@: GNU assembler version | |
2963 | 2.30 or later). | |
2964 | ||
2965 | @item @samp{bootstrap-time} | |
2966 | Arranges for the run time of each program started by the GCC driver, | |
2967 | built in any stage, to be logged to @file{time.log}, in the top level of | |
2968 | the build tree. | |
2969 | ||
2970 | @item @samp{bootstrap-asan} | |
2971 | Compiles GCC itself using Address Sanitization in order to catch invalid memory | |
2972 | accesses within the GCC code. | |
2973 | ||
2974 | @item @samp{bootstrap-hwasan} | |
2975 | Compiles GCC itself using HWAddress Sanitization in order to catch invalid | |
2976 | memory accesses within the GCC code. This option is only available on AArch64 | |
2977 | systems that are running Linux kernel version 5.4 or later. | |
2978 | ||
2979 | @end table | |
2980 | ||
2981 | @section Building a cross compiler | |
2982 | ||
2983 | When building a cross compiler, it is not generally possible to do a | |
2984 | 3-stage bootstrap of the compiler. This makes for an interesting problem | |
2985 | as parts of GCC can only be built with GCC@. | |
2986 | ||
2987 | To build a cross compiler, we recommend first building and installing a | |
2988 | native compiler. You can then use the native GCC compiler to build the | |
2989 | cross compiler. The installed native compiler needs to be GCC version | |
2990 | 2.95 or later. | |
2991 | ||
2992 | Assuming you have already installed a native copy of GCC and configured | |
2993 | your cross compiler, issue the command @command{make}, which performs the | |
2994 | following steps: | |
2995 | ||
2996 | @itemize @bullet | |
2997 | @item | |
2998 | Build host tools necessary to build the compiler. | |
2999 | ||
3000 | @item | |
3001 | Build target tools for use by the compiler such as binutils (bfd, | |
3002 | binutils, gas, gprof, ld, and opcodes) | |
3003 | if they have been individually linked or moved into the top level GCC source | |
3004 | tree before configuring. | |
3005 | ||
3006 | @item | |
3007 | Build the compiler (single stage only). | |
3008 | ||
3009 | @item | |
3010 | Build runtime libraries using the compiler from the previous step. | |
3011 | @end itemize | |
3012 | ||
3013 | Note that if an error occurs in any step the make process will exit. | |
3014 | ||
3015 | If you are not building GNU binutils in the same source tree as GCC, | |
3016 | you will need a cross-assembler and cross-linker installed before | |
3017 | configuring GCC@. Put them in the directory | |
3018 | @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools | |
3019 | you should put in this directory: | |
3020 | ||
3021 | @table @file | |
3022 | @item as | |
3023 | This should be the cross-assembler. | |
3024 | ||
3025 | @item ld | |
3026 | This should be the cross-linker. | |
3027 | ||
3028 | @item ar | |
3029 | This should be the cross-archiver: a program which can manipulate | |
3030 | archive files (linker libraries) in the target machine's format. | |
3031 | ||
3032 | @item ranlib | |
3033 | This should be a program to construct a symbol table in an archive file. | |
3034 | @end table | |
3035 | ||
3036 | The installation of GCC will find these programs in that directory, | |
3037 | and copy or link them to the proper place to for the cross-compiler to | |
3038 | find them when run later. | |
3039 | ||
3040 | The easiest way to provide these files is to build the Binutils package. | |
3041 | Configure it with the same @option{--host} and @option{--target} | |
3042 | options that you use for configuring GCC, then build and install | |
3043 | them. They install their executables automatically into the proper | |
3044 | directory. Alas, they do not support all the targets that GCC | |
3045 | supports. | |
3046 | ||
3047 | If you are not building a C library in the same source tree as GCC, | |
3048 | you should also provide the target libraries and headers before | |
3049 | configuring GCC, specifying the directories with | |
3050 | @option{--with-sysroot} or @option{--with-headers} and | |
3051 | @option{--with-libs}. Many targets also require ``start files'' such | |
3052 | as @file{crt0.o} and | |
3053 | @file{crtn.o} which are linked into each executable. There may be several | |
3054 | alternatives for @file{crt0.o}, for use with profiling or other | |
3055 | compilation options. Check your target's definition of | |
3056 | @code{STARTFILE_SPEC} to find out what start files it uses. | |
3057 | ||
3058 | @section Building in parallel | |
3059 | ||
3060 | GNU Make 3.80 and above, which is necessary to build GCC, support | |
3061 | building in parallel. To activate this, you can use @samp{make -j 2} | |
3062 | instead of @samp{make}. You can also specify a bigger number, and | |
3063 | in most cases using a value greater than the number of processors in | |
3064 | your machine will result in fewer and shorter I/O latency hits, thus | |
3065 | improving overall throughput; this is especially true for slow drives | |
3066 | and network filesystems. | |
3067 | ||
3068 | @section Building the Ada compiler | |
3069 | ||
3070 | @ifnothtml | |
3071 | @ref{GNAT-prerequisite}. | |
3072 | @end ifnothtml | |
3073 | @ifhtml | |
3074 | @uref{prerequisites.html#GNAT-prerequisite,,GNAT prerequisites}. | |
3075 | @end ifhtml | |
3076 | ||
3077 | @section Building the D compiler | |
3078 | ||
3079 | @ifnothtml | |
3080 | @ref{GDC-prerequisite}. | |
3081 | @end ifnothtml | |
3082 | @ifhtml | |
3083 | @uref{prerequisites.html#GDC-prerequisite,,GDC prerequisites}. | |
3084 | @end ifhtml | |
3085 | ||
3086 | @section Building with profile feedback | |
3087 | ||
3088 | It is possible to use profile feedback to optimize the compiler itself. This | |
3089 | should result in a faster compiler binary. Experiments done on x86 using gcc | |
3090 | 3.3 showed approximately 7 percent speedup on compiling C programs. To | |
3091 | bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}. | |
3092 | ||
3093 | When @samp{make profiledbootstrap} is run, it will first build a @code{stage1} | |
3094 | compiler. This compiler is used to build a @code{stageprofile} compiler | |
3095 | instrumented to collect execution counts of instruction and branch | |
3096 | probabilities. Training run is done by building @code{stagetrain} | |
3097 | compiler. Finally a @code{stagefeedback} compiler is built | |
3098 | using the information collected. | |
3099 | ||
3100 | Unlike standard bootstrap, several additional restrictions apply. The | |
3101 | compiler used to build @code{stage1} needs to support a 64-bit integral type. | |
3102 | It is recommended to only use GCC for this. | |
3103 | ||
3104 | On Linux/x86_64 hosts with some restrictions (no virtualization) it is | |
3105 | also possible to do autofdo build with @samp{make | |
3106 | autoprofiledback}. This uses Linux perf to sample branches in the | |
3107 | binary and then rebuild it with feedback derived from the profile. | |
3108 | Linux perf and the @code{autofdo} toolkit needs to be installed for | |
3109 | this. | |
3110 | ||
3111 | Only the profile from the current build is used, so when an error | |
3112 | occurs it is recommended to clean before restarting. Otherwise | |
3113 | the code quality may be much worse. | |
3114 | ||
3115 | @html | |
3116 | <hr /> | |
3117 | <p> | |
3118 | @end html | |
3119 | @ifhtml | |
3120 | @uref{./index.html,,Return to the GCC Installation page} | |
3121 | @end ifhtml | |
3122 | @end ifset | |
3123 | ||
3124 | @c ***Testing***************************************************************** | |
3125 | @ifnothtml | |
3126 | @comment node-name, next, previous, up | |
3127 | @node Testing, Final install, Building, Installing GCC | |
3128 | @end ifnothtml | |
3129 | @ifset testhtml | |
3130 | @ifnothtml | |
3131 | @chapter Installing GCC: Testing | |
3132 | @end ifnothtml | |
3133 | @cindex Testing | |
3134 | @cindex Installing GCC: Testing | |
3135 | @cindex Testsuite | |
3136 | ||
3137 | Before you install GCC, we encourage you to run the testsuites and to | |
3138 | compare your results with results from a similar configuration that have | |
3139 | been submitted to the | |
3140 | @uref{https://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}. | |
3141 | Some of these archived results are linked from the build status lists | |
3142 | at @uref{https://gcc.gnu.org/buildstat.html}, although not everyone who | |
3143 | reports a successful build runs the testsuites and submits the results. | |
3144 | This step is optional and may require you to download additional software, | |
3145 | but it can give you confidence in your new GCC installation or point out | |
3146 | problems before you install and start using your new GCC@. | |
3147 | ||
3148 | First, you must have @uref{download.html,,downloaded the testsuites}. | |
ce51e843 | 3149 | These are included in the source tarball. |
d77de738 ML |
3150 | |
3151 | Second, you must have the testing tools installed. This includes | |
3152 | @uref{https://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect; | |
3153 | the DejaGnu site has links to these. | |
3154 | Some optional tests also require Python3 and pytest module. | |
3155 | ||
3156 | If the directories where @command{runtest} and @command{expect} were | |
3157 | installed are not in the @env{PATH}, you may need to set the following | |
3158 | environment variables appropriately, as in the following example (which | |
3159 | assumes that DejaGnu has been installed under @file{/usr/local}): | |
3160 | ||
3161 | @smallexample | |
3162 | TCL_LIBRARY = /usr/local/share/tcl8.0 | |
3163 | DEJAGNULIBS = /usr/local/share/dejagnu | |
3164 | @end smallexample | |
3165 | ||
3166 | (On systems such as Cygwin, these paths are required to be actual | |
3167 | paths, not mounts or links; presumably this is due to some lack of | |
3168 | portability in the DejaGnu code.) | |
3169 | ||
3170 | ||
3171 | Finally, you can run the testsuite (which may take a long time): | |
3172 | @smallexample | |
3173 | cd @var{objdir}; make -k check | |
3174 | @end smallexample | |
3175 | ||
3176 | This will test various components of GCC, such as compiler | |
3177 | front ends and runtime libraries. While running the testsuite, DejaGnu | |
3178 | might emit some harmless messages resembling | |
3179 | @samp{WARNING: Couldn't find the global config file.} or | |
3180 | @samp{WARNING: Couldn't find tool init file} that can be ignored. | |
3181 | ||
3182 | If you are testing a cross-compiler, you may want to run the testsuite | |
3183 | on a simulator as described at @uref{https://gcc.gnu.org/simtest-howto.html}. | |
3184 | ||
3185 | @section How can you run the testsuite on selected tests? | |
3186 | ||
3187 | In order to run sets of tests selectively, there are targets | |
3188 | @samp{make check-gcc} and language specific @samp{make check-c}, | |
3189 | @samp{make check-c++}, @samp{make check-d} @samp{make check-fortran}, | |
1eee94d3 GM |
3190 | @samp{make check-ada}, @samp{make check-m2}, @samp{make check-objc}, |
3191 | @samp{make check-obj-c++}, @samp{make check-lto} in the @file{gcc} | |
3192 | subdirectory of the object directory. You can also just run | |
3193 | @samp{make check} in a subdirectory of the object directory. | |
d77de738 ML |
3194 | |
3195 | ||
3196 | A more selective way to just run all @command{gcc} execute tests in the | |
3197 | testsuite is to use | |
3198 | ||
3199 | @smallexample | |
3200 | make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}" | |
3201 | @end smallexample | |
3202 | ||
3203 | Likewise, in order to run only the @command{g++} ``old-deja'' tests in | |
3204 | the testsuite with filenames matching @samp{9805*}, you would use | |
3205 | ||
3206 | @smallexample | |
3207 | make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}" | |
3208 | @end smallexample | |
3209 | ||
3210 | The file-matching expression following @var{filename}@command{.exp=} is treated | |
3211 | as a series of whitespace-delimited glob expressions so that multiple patterns | |
3212 | may be passed, although any whitespace must either be escaped or surrounded by | |
3213 | single quotes if multiple expressions are desired. For example, | |
3214 | ||
3215 | @smallexample | |
3216 | make check-g++ RUNTESTFLAGS="old-deja.exp=9805*\ virtual2.c @var{other-options}" | |
3217 | make check-g++ RUNTESTFLAGS="'old-deja.exp=9805* virtual2.c' @var{other-options}" | |
3218 | @end smallexample | |
3219 | ||
3220 | The @file{*.exp} files are located in the testsuite directories of the GCC | |
3221 | source, the most important ones being @file{compile.exp}, | |
3222 | @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}. | |
3223 | To get a list of the possible @file{*.exp} files, pipe the | |
3224 | output of @samp{make check} into a file and look at the | |
3225 | @samp{Running @dots{} .exp} lines. | |
3226 | ||
3227 | @section Passing options and running multiple testsuites | |
3228 | ||
3229 | You can pass multiple options to the testsuite using the | |
3230 | @samp{--target_board} option of DejaGNU, either passed as part of | |
3231 | @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to | |
3232 | work outside the makefiles. For example, | |
3233 | ||
3234 | @smallexample | |
3235 | make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants" | |
3236 | @end smallexample | |
3237 | ||
3238 | will run the standard @command{g++} testsuites (``unix'' is the target name | |
3239 | for a standard native testsuite situation), passing | |
3240 | @samp{-O3 -fmerge-constants} to the compiler on every test, i.e., | |
3241 | slashes separate options. | |
3242 | ||
3243 | You can run the testsuites multiple times using combinations of options | |
3244 | with a syntax similar to the brace expansion of popular shells: | |
3245 | ||
3246 | @smallexample | |
3247 | @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}" | |
3248 | @end smallexample | |
3249 | ||
3250 | (Note the empty option caused by the trailing comma in the final group.) | |
3251 | The following will run each testsuite eight times using the @samp{arm-sim} | |
3252 | target, as if you had specified all possible combinations yourself: | |
3253 | ||
3254 | @smallexample | |
3255 | --target_board='arm-sim/-mhard-float/-O1 \ | |
3256 | arm-sim/-mhard-float/-O2 \ | |
3257 | arm-sim/-mhard-float/-O3 \ | |
3258 | arm-sim/-mhard-float \ | |
3259 | arm-sim/-msoft-float/-O1 \ | |
3260 | arm-sim/-msoft-float/-O2 \ | |
3261 | arm-sim/-msoft-float/-O3 \ | |
3262 | arm-sim/-msoft-float' | |
3263 | @end smallexample | |
3264 | ||
3265 | They can be combined as many times as you wish, in arbitrary ways. This | |
3266 | list: | |
3267 | ||
3268 | @smallexample | |
3269 | @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}" | |
3270 | @end smallexample | |
3271 | ||
3272 | will generate four combinations, all involving @samp{-Wextra}. | |
3273 | ||
3274 | The disadvantage to this method is that the testsuites are run in serial, | |
3275 | which is a waste on multiprocessor systems. For users with GNU Make and | |
3276 | a shell which performs brace expansion, you can run the testsuites in | |
3277 | parallel by having the shell perform the combinations and @command{make} | |
3278 | do the parallel runs. Instead of using @samp{--target_board}, use a | |
3279 | special makefile target: | |
3280 | ||
3281 | @smallexample | |
3282 | make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{} | |
3283 | @end smallexample | |
3284 | ||
3285 | For example, | |
3286 | ||
3287 | @smallexample | |
3288 | make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@} | |
3289 | @end smallexample | |
3290 | ||
3291 | will run three concurrent ``make-gcc'' testsuites, eventually testing all | |
3292 | ten combinations as described above. Note that this is currently only | |
3293 | supported in the @file{gcc} subdirectory. (To see how this works, try | |
3294 | typing @command{echo} before the example given here.) | |
3295 | ||
3296 | ||
3297 | @section How to interpret test results | |
3298 | ||
3299 | The result of running the testsuite are various @file{*.sum} and @file{*.log} | |
3300 | files in the testsuite subdirectories. The @file{*.log} files contain a | |
3301 | detailed log of the compiler invocations and the corresponding | |
3302 | results, the @file{*.sum} files summarize the results. These summaries | |
3303 | contain status codes for all tests: | |
3304 | ||
3305 | @itemize @bullet | |
3306 | @item | |
3307 | PASS: the test passed as expected | |
3308 | @item | |
3309 | XPASS: the test unexpectedly passed | |
3310 | @item | |
3311 | FAIL: the test unexpectedly failed | |
3312 | @item | |
3313 | XFAIL: the test failed as expected | |
3314 | @item | |
3315 | UNSUPPORTED: the test is not supported on this platform | |
3316 | @item | |
3317 | ERROR: the testsuite detected an error | |
3318 | @item | |
3319 | WARNING: the testsuite detected a possible problem | |
3320 | @end itemize | |
3321 | ||
3322 | It is normal for some tests to report unexpected failures. At the | |
3323 | current time the testing harness does not allow fine grained control | |
3324 | over whether or not a test is expected to fail. This problem should | |
3325 | be fixed in future releases. | |
3326 | ||
3327 | ||
3328 | @section Submitting test results | |
3329 | ||
3330 | If you want to report the results to the GCC project, use the | |
3331 | @file{contrib/test_summary} shell script. Start it in the @var{objdir} with | |
3332 | ||
3333 | @smallexample | |
3334 | @var{srcdir}/contrib/test_summary -p your_commentary.txt \ | |
3335 | -m gcc-testresults@@gcc.gnu.org |sh | |
3336 | @end smallexample | |
3337 | ||
3338 | This script uses the @command{Mail} program to send the results, so | |
3339 | make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is | |
3340 | prepended to the testsuite summary and should contain any special | |
3341 | remarks you have on your results or your build environment. Please | |
3342 | do not edit the testsuite result block or the subject line, as these | |
3343 | messages may be automatically processed. | |
3344 | ||
3345 | @html | |
3346 | <hr /> | |
3347 | <p> | |
3348 | @end html | |
3349 | @ifhtml | |
3350 | @uref{./index.html,,Return to the GCC Installation page} | |
3351 | @end ifhtml | |
3352 | @end ifset | |
3353 | ||
3354 | @c ***Final install*********************************************************** | |
3355 | @ifnothtml | |
3356 | @comment node-name, next, previous, up | |
3357 | @node Final install, , Testing, Installing GCC | |
3358 | @end ifnothtml | |
3359 | @ifset finalinstallhtml | |
3360 | @ifnothtml | |
3361 | @chapter Installing GCC: Final installation | |
3362 | @end ifnothtml | |
3363 | ||
3364 | Now that GCC has been built (and optionally tested), you can install it with | |
3365 | @smallexample | |
3366 | cd @var{objdir} && make install | |
3367 | @end smallexample | |
3368 | ||
3369 | We strongly recommend to install into a target directory where there is | |
3370 | no previous version of GCC present. Also, the GNAT runtime should not | |
3371 | be stripped, as this would break certain features of the debugger that | |
3372 | depend on this debugging information (catching Ada exceptions for | |
3373 | instance). | |
3374 | ||
3375 | That step completes the installation of GCC; user level binaries can | |
3376 | be found in @file{@var{prefix}/bin} where @var{prefix} is the value | |
3377 | you specified with the @option{--prefix} to configure (or | |
3378 | @file{/usr/local} by default). (If you specified @option{--bindir}, | |
3379 | that directory will be used instead; otherwise, if you specified | |
3380 | @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.) | |
3381 | Headers for the C++ library are installed in | |
3382 | @file{@var{prefix}/include}; libraries in @file{@var{libdir}} | |
3383 | (normally @file{@var{prefix}/lib}); internal parts of the compiler in | |
3384 | @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation | |
3385 | in info format in @file{@var{infodir}} (normally | |
3386 | @file{@var{prefix}/info}). | |
3387 | ||
3388 | When installing cross-compilers, GCC's executables | |
3389 | are not only installed into @file{@var{bindir}}, that | |
3390 | is, @file{@var{exec-prefix}/bin}, but additionally into | |
3391 | @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory | |
3392 | exists. Typically, such @dfn{tooldirs} hold target-specific | |
3393 | binutils, including assembler and linker. | |
3394 | ||
3395 | Installation into a temporary staging area or into a @command{chroot} | |
3396 | jail can be achieved with the command | |
3397 | ||
3398 | @smallexample | |
3399 | make DESTDIR=@var{path-to-rootdir} install | |
3400 | @end smallexample | |
3401 | ||
3402 | @noindent | |
3403 | where @var{path-to-rootdir} is the absolute path of | |
3404 | a directory relative to which all installation paths will be | |
3405 | interpreted. Note that the directory specified by @code{DESTDIR} | |
3406 | need not exist yet; it will be created if necessary. | |
3407 | ||
3408 | There is a subtle point with tooldirs and @code{DESTDIR}: | |
3409 | If you relocate a cross-compiler installation with | |
3410 | e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory | |
3411 | @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will | |
3412 | be filled with duplicated GCC executables only if it already exists, | |
3413 | it will not be created otherwise. This is regarded as a feature, | |
3414 | not as a bug, because it gives slightly more control to the packagers | |
3415 | using the @code{DESTDIR} feature. | |
3416 | ||
3417 | You can install stripped programs and libraries with | |
3418 | ||
3419 | @smallexample | |
3420 | make install-strip | |
3421 | @end smallexample | |
3422 | ||
3423 | If you are bootstrapping a released version of GCC then please | |
3424 | quickly review the build status page for your release, available from | |
3425 | @uref{https://gcc.gnu.org/buildstat.html}. | |
3426 | If your system is not listed for the version of GCC that you built, | |
3427 | send a note to | |
3428 | @email{gcc@@gcc.gnu.org} indicating | |
3429 | that you successfully built and installed GCC@. | |
3430 | Include the following information: | |
3431 | ||
3432 | @itemize @bullet | |
3433 | @item | |
3434 | Output from running @file{@var{srcdir}/config.guess}. Do not send | |
3435 | that file itself, just the one-line output from running it. | |
3436 | ||
3437 | @item | |
3438 | The output of @samp{gcc -v} for your newly installed @command{gcc}. | |
3439 | This tells us which version of GCC you built and the options you passed to | |
3440 | configure. | |
3441 | ||
3442 | @item | |
3443 | Whether you enabled all languages or a subset of them. If you used a | |
3444 | full distribution then this information is part of the configure | |
3445 | options in the output of @samp{gcc -v}, but if you downloaded the | |
3446 | ``core'' compiler plus additional front ends then it isn't apparent | |
3447 | which ones you built unless you tell us about it. | |
3448 | ||
3449 | @item | |
3450 | If the build was for GNU/Linux, also include: | |
3451 | @itemize @bullet | |
3452 | @item | |
3453 | The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3); | |
3454 | this information should be available from @file{/etc/issue}. | |
3455 | ||
3456 | @item | |
3457 | The version of the Linux kernel, available from @samp{uname --version} | |
3458 | or @samp{uname -a}. | |
3459 | ||
3460 | @item | |
3461 | The version of glibc you used; for RPM-based systems like Red Hat, | |
3462 | Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version, | |
3463 | and on systems like Debian and Progeny use @samp{dpkg -l libc6}. | |
3464 | @end itemize | |
3465 | For other systems, you can include similar information if you think it is | |
3466 | relevant. | |
3467 | ||
3468 | @item | |
3469 | Any other information that you think would be useful to people building | |
3470 | GCC on the same configuration. The new entry in the build status list | |
3471 | will include a link to the archived copy of your message. | |
3472 | @end itemize | |
3473 | ||
3474 | We'd also like to know if the | |
3475 | @ifnothtml | |
3476 | @ref{Specific, host/target specific installation notes} | |
3477 | @end ifnothtml | |
3478 | @ifhtml | |
3479 | @uref{specific.html,,host/target specific installation notes} | |
3480 | @end ifhtml | |
3481 | didn't include your host/target information or if that information is | |
3482 | incomplete or out of date. Send a note to | |
3483 | @email{gcc@@gcc.gnu.org} detailing how the information should be changed. | |
3484 | ||
3485 | If you find a bug, please report it following the | |
3486 | @uref{../bugs/,,bug reporting guidelines}. | |
3487 | ||
3488 | If you want to print the GCC manuals, do @samp{cd @var{objdir}; make | |
3489 | dvi}. You will need to have @command{texi2dvi} (version at least 4.7) | |
3490 | and @TeX{} installed. This creates a number of @file{.dvi} files in | |
3491 | subdirectories of @file{@var{objdir}}; these may be converted for | |
3492 | printing with programs such as @command{dvips}. Alternately, by using | |
3493 | @samp{make pdf} in place of @samp{make dvi}, you can create documentation | |
3494 | in the form of @file{.pdf} files; this requires @command{texi2pdf}, which | |
3495 | is included with Texinfo version 4.8 and later. You can also | |
3496 | @uref{https://shop.fsf.org/,,buy printed manuals from the | |
3497 | Free Software Foundation}, though such manuals may not be for the most | |
3498 | recent version of GCC@. | |
3499 | ||
3500 | If you would like to generate online HTML documentation, do @samp{cd | |
3501 | @var{objdir}; make html} and HTML will be generated for the gcc manuals in | |
3502 | @file{@var{objdir}/gcc/HTML}. | |
3503 | ||
3504 | @html | |
3505 | <hr /> | |
3506 | <p> | |
3507 | @end html | |
3508 | @ifhtml | |
3509 | @uref{./index.html,,Return to the GCC Installation page} | |
3510 | @end ifhtml | |
3511 | @end ifset | |
3512 | ||
3513 | @c ***Binaries**************************************************************** | |
3514 | @ifnothtml | |
3515 | @comment node-name, next, previous, up | |
3516 | @node Binaries, Specific, Installing GCC, Top | |
3517 | @end ifnothtml | |
3518 | @ifset binarieshtml | |
3519 | @ifnothtml | |
3520 | @chapter Installing GCC: Binaries | |
3521 | @end ifnothtml | |
3522 | @cindex Binaries | |
3523 | @cindex Installing GCC: Binaries | |
3524 | ||
3525 | We are often asked about pre-compiled versions of GCC@. While we cannot | |
3526 | provide these for all platforms, below you'll find links to binaries for | |
3527 | various platforms where creating them by yourself is not easy due to various | |
3528 | reasons. | |
3529 | ||
3530 | Please note that we did not create these binaries, nor do we | |
3531 | support them. If you have any problems installing them, please | |
3532 | contact their makers. | |
3533 | ||
3534 | @itemize | |
3535 | @item | |
3536 | AIX: | |
3537 | @itemize | |
3538 | @item | |
3539 | @uref{http://www.perzl.org/aix/,,AIX Open Source Packages (AIX5L AIX 6.1 | |
3540 | AIX 7.1)}. | |
3541 | @end itemize | |
3542 | ||
3543 | @item | |
3544 | DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}. | |
3545 | ||
3546 | @item | |
3547 | HP-UX: | |
3548 | @itemize | |
3549 | @item | |
3550 | @uref{http://hpux.connect.org.uk/,,HP-UX Porting Center}; | |
3551 | @end itemize | |
3552 | ||
3553 | @item | |
3554 | Solaris 2 (SPARC, Intel): | |
3555 | @itemize | |
3556 | @item | |
3557 | @uref{https://www.opencsw.org/,,OpenCSW} | |
3558 | @end itemize | |
3559 | ||
3560 | @item | |
3561 | macOS: | |
3562 | @itemize | |
3563 | @item | |
3564 | The @uref{https://brew.sh,,Homebrew} package manager; | |
3565 | @item | |
3566 | @uref{https://www.macports.org,,MacPorts}. | |
3567 | @end itemize | |
3568 | ||
3569 | @item | |
3570 | Microsoft Windows: | |
3571 | @itemize | |
3572 | @item | |
3573 | The @uref{https://sourceware.org/cygwin/,,Cygwin} project; | |
3574 | @item | |
3575 | The @uref{https://osdn.net/projects/mingw/,,MinGW} and | |
3576 | @uref{https://www.mingw-w64.org/,,mingw-w64} projects. | |
3577 | @end itemize | |
3578 | ||
3579 | @item | |
3580 | @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a | |
3581 | number of platforms. | |
3582 | ||
3583 | @item | |
3584 | The @uref{https://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has | |
3585 | links to GNU Fortran binaries for several platforms. | |
3586 | @end itemize | |
3587 | ||
3588 | @html | |
3589 | <hr /> | |
3590 | <p> | |
3591 | @end html | |
3592 | @ifhtml | |
3593 | @uref{./index.html,,Return to the GCC Installation page} | |
3594 | @end ifhtml | |
3595 | @end ifset | |
3596 | ||
3597 | @c ***Specific**************************************************************** | |
3598 | @ifnothtml | |
3599 | @comment node-name, next, previous, up | |
3600 | @node Specific, GNU Free Documentation License, Binaries, Top | |
3601 | @end ifnothtml | |
3602 | @ifset specifichtml | |
3603 | @ifnothtml | |
3604 | @chapter Host/target specific installation notes for GCC | |
3605 | @end ifnothtml | |
3606 | @cindex Specific | |
3607 | @cindex Specific installation notes | |
3608 | @cindex Target specific installation | |
3609 | @cindex Host specific installation | |
3610 | @cindex Target specific installation notes | |
3611 | ||
3612 | Please read this document carefully @emph{before} installing the | |
3613 | GNU Compiler Collection on your machine. | |
3614 | ||
3615 | Note that this list of install notes is @emph{not} a list of supported | |
3616 | hosts or targets. Not all supported hosts and targets are listed | |
3617 | here, only the ones that require host-specific or target-specific | |
3618 | information have to. | |
3619 | ||
3620 | @ifhtml | |
3621 | @itemize | |
3622 | @item | |
3623 | @uref{#aarch64-x-x,,aarch64*-*-*} | |
3624 | @item | |
3625 | @uref{#alpha-x-x,,alpha*-*-*} | |
3626 | @item | |
3627 | @uref{#amdgcn-x-amdhsa,,amdgcn-*-amdhsa} | |
3628 | @item | |
3629 | @uref{#amd64-x-solaris2,,amd64-*-solaris2*} | |
3630 | @item | |
3631 | @uref{#arc-x-elf32,,arc-*-elf32} | |
3632 | @item | |
3633 | @uref{#arc-linux-uclibc,,arc-linux-uclibc} | |
3634 | @item | |
3635 | @uref{#arm-x-eabi,,arm-*-eabi} | |
3636 | @item | |
3637 | @uref{#avr,,avr} | |
3638 | @item | |
3639 | @uref{#bfin,,Blackfin} | |
3640 | @item | |
3641 | @uref{#cris,,cris} | |
3642 | @item | |
3643 | @uref{#dos,,DOS} | |
3644 | @item | |
3645 | @uref{#epiphany-x-elf,,epiphany-*-elf} | |
3646 | @item | |
3647 | @uref{#ft32-x-elf,,ft32-*-elf} | |
3648 | @item | |
3649 | @uref{#x-x-freebsd,,*-*-freebsd*} | |
3650 | @item | |
3651 | @uref{#h8300-hms,,h8300-hms} | |
3652 | @item | |
3653 | @uref{#hppa-hp-hpux,,hppa*-hp-hpux*} | |
3654 | @item | |
3655 | @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10} | |
3656 | @item | |
3657 | @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11} | |
3658 | @item | |
3659 | @uref{#x-x-linux-gnu,,*-*-linux-gnu} | |
3660 | @item | |
3661 | @uref{#ix86-x-linux,,i?86-*-linux*} | |
3662 | @item | |
3663 | @uref{#ix86-x-solaris2,,i?86-*-solaris2*} | |
3664 | @item | |
3665 | @uref{#ia64-x-linux,,ia64-*-linux} | |
3666 | @item | |
3667 | @uref{#ia64-x-hpux,,ia64-*-hpux*} | |
3668 | @item | |
3669 | @uref{#x-ibm-aix,,*-ibm-aix*} | |
3670 | @item | |
3671 | @uref{#iq2000-x-elf,,iq2000-*-elf} | |
3672 | @item | |
3673 | @uref{#loongarch,,loongarch} | |
3674 | @item | |
3675 | @uref{#lm32-x-elf,,lm32-*-elf} | |
3676 | @item | |
3677 | @uref{#lm32-x-uclinux,,lm32-*-uclinux} | |
3678 | @item | |
3679 | @uref{#m32c-x-elf,,m32c-*-elf} | |
3680 | @item | |
3681 | @uref{#m32r-x-elf,,m32r-*-elf} | |
3682 | @item | |
3683 | @uref{#m68k-x-x,,m68k-*-*} | |
3684 | @item | |
3685 | @uref{#m68k-x-uclinux,,m68k-*-uclinux} | |
3686 | @item | |
3687 | @uref{#microblaze-x-elf,,microblaze-*-elf} | |
3688 | @item | |
3689 | @uref{#mips-x-x,,mips-*-*} | |
3690 | @item | |
3691 | @uref{#moxie-x-elf,,moxie-*-elf} | |
3692 | @item | |
3693 | @uref{#msp430-x-elf,,msp430-*-elf} | |
3694 | @item | |
3695 | @uref{#nds32le-x-elf,,nds32le-*-elf} | |
3696 | @item | |
3697 | @uref{#nds32be-x-elf,,nds32be-*-elf} | |
3698 | @item | |
3699 | @uref{#nvptx-x-none,,nvptx-*-none} | |
3700 | @item | |
3701 | @uref{#or1k-x-elf,,or1k-*-elf} | |
3702 | @item | |
3703 | @uref{#or1k-x-linux,,or1k-*-linux} | |
3704 | @item | |
3705 | @uref{#powerpc-x-x,,powerpc*-*-*} | |
3706 | @item | |
3707 | @uref{#powerpc-x-darwin,,powerpc-*-darwin*} | |
3708 | @item | |
3709 | @uref{#powerpc-x-elf,,powerpc-*-elf} | |
3710 | @item | |
3711 | @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*} | |
3712 | @item | |
3713 | @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*} | |
3714 | @item | |
3715 | @uref{#powerpc-x-eabisim,,powerpc-*-eabisim} | |
3716 | @item | |
3717 | @uref{#powerpc-x-eabi,,powerpc-*-eabi} | |
3718 | @item | |
3719 | @uref{#powerpcle-x-elf,,powerpcle-*-elf} | |
3720 | @item | |
3721 | @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim} | |
3722 | @item | |
3723 | @uref{#powerpcle-x-eabi,,powerpcle-*-eabi} | |
3724 | @item | |
3725 | @uref{#riscv32-x-elf,,riscv32-*-elf} | |
3726 | @item | |
3727 | @uref{#riscv32-x-linux,,riscv32-*-linux} | |
3728 | @item | |
3729 | @uref{#riscv64-x-elf,,riscv64-*-elf} | |
3730 | @item | |
3731 | @uref{#riscv64-x-linux,,riscv64-*-linux} | |
3732 | @item | |
3733 | @uref{#rl78-x-elf,,rl78-*-elf} | |
3734 | @item | |
3735 | @uref{#rx-x-elf,,rx-*-elf} | |
3736 | @item | |
3737 | @uref{#s390-x-linux,,s390-*-linux*} | |
3738 | @item | |
3739 | @uref{#s390x-x-linux,,s390x-*-linux*} | |
3740 | @item | |
3741 | @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*} | |
3742 | @item | |
3743 | @uref{#x-x-solaris2,,*-*-solaris2*} | |
3744 | @item | |
3745 | @uref{#sparc-x-x,,sparc*-*-*} | |
3746 | @item | |
3747 | @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*} | |
3748 | @item | |
3749 | @uref{#sparc-x-linux,,sparc-*-linux*} | |
3750 | @item | |
3751 | @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*} | |
3752 | @item | |
3753 | @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*} | |
3754 | @item | |
3755 | @uref{#c6x-x-x,,c6x-*-*} | |
3756 | @item | |
3757 | @uref{#visium-x-elf, visium-*-elf} | |
3758 | @item | |
3759 | @uref{#x-x-vxworks,,*-*-vxworks*} | |
3760 | @item | |
3761 | @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*} | |
3762 | @item | |
3763 | @uref{#x86-64-x-solaris2,,x86_64-*-solaris2*} | |
3764 | @item | |
3765 | @uref{#xtensa-x-elf,,xtensa*-*-elf} | |
3766 | @item | |
3767 | @uref{#xtensa-x-linux,,xtensa*-*-linux*} | |
3768 | @item | |
3769 | @uref{#windows,,Microsoft Windows} | |
3770 | @item | |
3771 | @uref{#x-x-cygwin,,*-*-cygwin} | |
3772 | @item | |
3773 | @uref{#x-x-mingw32,,*-*-mingw32} | |
3774 | @item | |
3775 | @uref{#os2,,OS/2} | |
3776 | @item | |
3777 | @uref{#older,,Older systems} | |
3778 | @end itemize | |
3779 | ||
3780 | @itemize | |
3781 | @item | |
3782 | @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.) | |
3783 | @end itemize | |
3784 | @end ifhtml | |
3785 | ||
3786 | ||
3787 | @html | |
3788 | <!-- -------- host/target specific issues start here ---------------- --> | |
3789 | <hr /> | |
3790 | @end html | |
3791 | @anchor{aarch64-x-x} | |
3792 | @heading aarch64*-*-* | |
3793 | Binutils pre 2.24 does not have support for selecting @option{-mabi} and | |
3794 | does not support ILP32. If it is used to build GCC 4.9 or later, GCC will | |
3795 | not support option @option{-mabi=ilp32}. | |
3796 | ||
3797 | To enable a workaround for the Cortex-A53 erratum number 835769 by default | |
3798 | (for all CPUs regardless of -mcpu option given) at configure time use the | |
3799 | @option{--enable-fix-cortex-a53-835769} option. This will enable the fix by | |
3800 | default and can be explicitly disabled during compilation by passing the | |
3801 | @option{-mno-fix-cortex-a53-835769} option. Conversely, | |
3802 | @option{--disable-fix-cortex-a53-835769} will disable the workaround by | |
3803 | default. The workaround is disabled by default if neither of | |
3804 | @option{--enable-fix-cortex-a53-835769} or | |
3805 | @option{--disable-fix-cortex-a53-835769} is given at configure time. | |
3806 | ||
3807 | To enable a workaround for the Cortex-A53 erratum number 843419 by default | |
3808 | (for all CPUs regardless of -mcpu option given) at configure time use the | |
3809 | @option{--enable-fix-cortex-a53-843419} option. This workaround is applied at | |
3810 | link time. Enabling the workaround will cause GCC to pass the relevant option | |
3811 | to the linker. It can be explicitly disabled during compilation by passing the | |
3812 | @option{-mno-fix-cortex-a53-843419} option. Conversely, | |
3813 | @option{--disable-fix-cortex-a53-843419} will disable the workaround by default. | |
3814 | The workaround is disabled by default if neither of | |
3815 | @option{--enable-fix-cortex-a53-843419} or | |
3816 | @option{--disable-fix-cortex-a53-843419} is given at configure time. | |
3817 | ||
3818 | To enable Branch Target Identification Mechanism and Return Address Signing by | |
3819 | default at configure time use the @option{--enable-standard-branch-protection} | |
3820 | option. This is equivalent to having @option{-mbranch-protection=standard} | |
3821 | during compilation. This can be explicitly disabled during compilation by | |
3822 | passing the @option{-mbranch-protection=none} option which turns off all | |
3823 | types of branch protections. Conversely, | |
3824 | @option{--disable-standard-branch-protection} will disable both the | |
3825 | protections by default. This mechanism is turned off by default if neither | |
3826 | of the options are given at configure time. | |
3827 | ||
3828 | @html | |
3829 | <hr /> | |
3830 | @end html | |
3831 | @anchor{alpha-x-x} | |
3832 | @heading alpha*-*-* | |
3833 | This section contains general configuration information for all | |
3834 | Alpha-based platforms using ELF@. In addition to reading this | |
3835 | section, please read all other sections that match your target. | |
3836 | ||
3837 | @html | |
3838 | <hr /> | |
3839 | @end html | |
3840 | @anchor{amd64-x-solaris2} | |
3841 | @heading amd64-*-solaris2* | |
3842 | This is a synonym for @samp{x86_64-*-solaris2*}. | |
3843 | ||
3844 | @html | |
3845 | <hr /> | |
3846 | @end html | |
3847 | @anchor{amdgcn-x-amdhsa} | |
3848 | @heading amdgcn-*-amdhsa | |
3849 | AMD GCN GPU target. | |
3850 | ||
3851 | Instead of GNU Binutils, you will need to install LLVM 13.0.1, or later, and copy | |
3852 | @file{bin/llvm-mc} to @file{amdgcn-amdhsa/bin/as}, | |
3853 | @file{bin/lld} to @file{amdgcn-amdhsa/bin/ld}, | |
3854 | @file{bin/llvm-nm} to @file{amdgcn-amdhsa/bin/nm}, and | |
3855 | @file{bin/llvm-ar} to both @file{bin/amdgcn-amdhsa-ar} and | |
3856 | @file{bin/amdgcn-amdhsa-ranlib}. | |
3857 | ||
3858 | Use Newlib (3.2.0, or newer). | |
3859 | ||
3860 | To run the binaries, install the HSA Runtime from the | |
3861 | @uref{https://rocm.github.io,,ROCm Platform}, and use | |
3862 | @file{libexec/gcc/amdhsa-amdhsa/@var{version}/gcn-run} to launch them | |
3863 | on the GPU. | |
3864 | ||
3865 | @html | |
3866 | <hr /> | |
3867 | @end html | |
3868 | @anchor{arc-x-elf32} | |
3869 | @heading arc-*-elf32 | |
3870 | ||
3871 | Use @samp{configure --target=arc-elf32 --with-cpu=@var{cpu} --enable-languages="c,c++"} | |
3872 | to configure GCC, with @var{cpu} being one of @samp{arc600}, @samp{arc601}, | |
3873 | or @samp{arc700}@. | |
3874 | ||
3875 | @html | |
3876 | <hr /> | |
3877 | @end html | |
3878 | @anchor{arc-linux-uclibc} | |
3879 | @heading arc-linux-uclibc | |
3880 | ||
3881 | Use @samp{configure --target=arc-linux-uclibc --with-cpu=arc700 --enable-languages="c,c++"} to configure GCC@. | |
3882 | ||
3883 | @html | |
3884 | <hr /> | |
3885 | @end html | |
3886 | @anchor{arm-x-eabi} | |
3887 | @heading arm-*-eabi | |
3888 | ARM-family processors. | |
3889 | ||
3890 | Building the Ada frontend commonly fails (an infinite loop executing | |
3891 | @code{xsinfo}) if the host compiler is GNAT 4.8. Host compilers built from the | |
3892 | GNAT 4.6, 4.9 or 5 release branches are known to succeed. | |
3893 | ||
3894 | @html | |
3895 | <hr /> | |
3896 | @end html | |
3897 | @anchor{avr} | |
3898 | @heading avr | |
3899 | ATMEL AVR-family micro controllers. These are used in embedded | |
3900 | applications. There are no standard Unix configurations. | |
3901 | @ifnothtml | |
3902 | @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler | |
3903 | Collection (GCC)}, | |
3904 | @end ifnothtml | |
3905 | @ifhtml | |
3906 | See ``AVR Options'' in the main manual | |
3907 | @end ifhtml | |
3908 | for the list of supported MCU types. | |
3909 | ||
3910 | Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@. | |
3911 | ||
3912 | Further installation notes and other useful information about AVR tools | |
3913 | can also be obtained from: | |
3914 | ||
3915 | @itemize @bullet | |
3916 | @item | |
3917 | @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/} | |
3918 | @item | |
3919 | @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/} | |
3920 | @end itemize | |
3921 | ||
3922 | The following error: | |
3923 | @smallexample | |
3924 | Error: register required | |
3925 | @end smallexample | |
3926 | ||
3927 | indicates that you should upgrade to a newer version of the binutils. | |
3928 | ||
3929 | @html | |
3930 | <hr /> | |
3931 | @end html | |
3932 | @anchor{bfin} | |
3933 | @heading Blackfin | |
3934 | The Blackfin processor, an Analog Devices DSP. | |
3935 | @ifnothtml | |
3936 | @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler | |
3937 | Collection (GCC)}, | |
3938 | @end ifnothtml | |
3939 | @ifhtml | |
3940 | See ``Blackfin Options'' in the main manual | |
3941 | @end ifhtml | |
3942 | ||
3943 | More information, and a version of binutils with support for this processor, | |
3944 | are available at @uref{https://sourceforge.net/projects/adi-toolchain/}. | |
3945 | ||
3946 | @html | |
3947 | <hr /> | |
3948 | @end html | |
3949 | @anchor{cris} | |
3950 | @heading CRIS | |
3951 | CRIS is a CPU architecture in Axis Communications systems-on-a-chip, for | |
3952 | example the ETRAX series. These are used in embedded applications. | |
3953 | ||
3954 | @ifnothtml | |
3955 | @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler | |
3956 | Collection (GCC)}, | |
3957 | @end ifnothtml | |
3958 | @ifhtml | |
3959 | See ``CRIS Options'' in the main manual | |
3960 | @end ifhtml | |
3961 | for a list of CRIS-specific options. | |
3962 | ||
3963 | Use @samp{configure --target=cris-elf} to configure GCC@ for building | |
3964 | a cross-compiler for CRIS. | |
3965 | @html | |
3966 | <hr /> | |
3967 | @end html | |
3968 | @anchor{dos} | |
3969 | @heading DOS | |
3970 | Please have a look at the @uref{binaries.html,,binaries page}. | |
3971 | ||
3972 | You cannot install GCC by itself on MSDOS; it will not compile under | |
3973 | any MSDOS compiler except itself. You need to get the complete | |
3974 | compilation package DJGPP, which includes binaries as well as sources, | |
3975 | and includes all the necessary compilation tools and libraries. | |
3976 | ||
3977 | @html | |
3978 | <hr /> | |
3979 | @end html | |
3980 | @anchor{epiphany-x-elf} | |
3981 | @heading epiphany-*-elf | |
3982 | Adapteva Epiphany. | |
3983 | This configuration is intended for embedded systems. | |
3984 | ||
3985 | @html | |
3986 | <hr /> | |
3987 | @end html | |
3988 | @anchor{x-x-freebsd} | |
3989 | @heading *-*-freebsd* | |
3990 | In order to better utilize FreeBSD base system functionality and match | |
3991 | the configuration of the system compiler, GCC 4.5 and above as well as | |
3992 | GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present | |
3993 | on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default | |
3994 | (on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside | |
3995 | @file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled | |
3996 | by GCC 4.5 and above. | |
3997 | ||
3998 | We support FreeBSD using the ELF file format with DWARF 2 debugging | |
3999 | for all CPU architectures. There are | |
4000 | no known issues with mixing object files and libraries with different | |
4001 | debugging formats. Otherwise, this release of GCC should now match | |
4002 | more of the configuration used in the stock FreeBSD configuration of | |
4003 | GCC@. In particular, @option{--enable-threads} is now configured by | |
4004 | default. However, as a general user, do not attempt to replace the | |
4005 | system compiler with this release. Known to bootstrap and check with | |
4006 | good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap | |
4007 | and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4, | |
4008 | 4.5, 4.8, 4.9 and 5-CURRENT@. | |
4009 | ||
4010 | The version of binutils installed in @file{/usr/bin} probably works | |
4011 | with this release of GCC@. Bootstrapping against the latest GNU | |
4012 | binutils and/or the version found in @file{/usr/ports/devel/binutils} has | |
4013 | been known to enable additional features and improve overall testsuite | |
4014 | results. However, it is currently known that boehm-gc may not configure | |
4015 | properly on FreeBSD prior to the FreeBSD 7.0 release with GNU binutils | |
4016 | after 2.16.1. | |
4017 | ||
4018 | @html | |
4019 | <hr /> | |
4020 | @end html | |
4021 | @anchor{ft32-x-elf} | |
4022 | @heading ft32-*-elf | |
4023 | The FT32 processor. | |
4024 | This configuration is intended for embedded systems. | |
4025 | ||
4026 | @html | |
4027 | <hr /> | |
4028 | @end html | |
4029 | @anchor{h8300-hms} | |
4030 | @heading h8300-hms | |
4031 | Renesas H8/300 series of processors. | |
4032 | ||
4033 | Please have a look at the @uref{binaries.html,,binaries page}. | |
4034 | ||
4035 | The calling convention and structure layout has changed in release 2.6. | |
4036 | All code must be recompiled. The calling convention now passes the | |
4037 | first three arguments in function calls in registers. Structures are no | |
4038 | longer a multiple of 2 bytes. | |
4039 | ||
4040 | @html | |
4041 | <hr /> | |
4042 | @end html | |
4043 | @anchor{hppa-hp-hpux} | |
4044 | @heading hppa*-hp-hpux* | |
4045 | Support for HP-UX version 9 and older was discontinued in GCC 3.4. | |
4046 | ||
4047 | We require using gas/binutils on all hppa platforms. Version 2.19 or | |
4048 | later is recommended. | |
4049 | ||
4050 | It may be helpful to configure GCC with the | |
4051 | @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and | |
4052 | @option{--with-as=@dots{}} options to ensure that GCC can find GAS@. | |
4053 | ||
4054 | The HP assembler should not be used with GCC. It is rarely tested and may | |
4055 | not work. It shouldn't be used with any languages other than C due to its | |
4056 | many limitations. | |
4057 | ||
4058 | Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging | |
4059 | format which GCC does not know about). It also inserts timestamps | |
4060 | into each object file it creates, causing the 3-stage comparison test to | |
4061 | fail during a bootstrap. You should be able to continue by saying | |
4062 | @samp{make all-host all-target} after getting the failure from @samp{make}. | |
4063 | ||
4064 | Various GCC features are not supported. For example, it does not support weak | |
4065 | symbols or alias definitions. As a result, explicit template instantiations | |
4066 | are required when using C++. This makes it difficult if not impossible to | |
4067 | build many C++ applications. | |
4068 | ||
4069 | There are two default scheduling models for instructions. These are | |
4070 | PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc | |
4071 | architecture specified for the target machine when configuring. | |
4072 | PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when | |
4073 | the target is a @samp{hppa1*} machine. | |
4074 | ||
4075 | The PROCESSOR_8000 model is not well suited to older processors. Thus, | |
4076 | it is important to completely specify the machine architecture when | |
4077 | configuring if you want a model other than PROCESSOR_8000. The macro | |
4078 | TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different | |
4079 | default scheduling model is desired. | |
4080 | ||
4081 | As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10 | |
4082 | through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later. | |
4083 | This namespace change might cause problems when bootstrapping with | |
4084 | an earlier version of GCC or the HP compiler as essentially the same | |
4085 | namespace is required for an entire build. This problem can be avoided | |
4086 | in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95} | |
4087 | or @samp{98}. Another way is to add an appropriate set of predefines | |
4088 | to @env{CC}. The description for the @option{munix=} option contains | |
4089 | a list of the predefines used with each standard. | |
4090 | ||
4091 | More specific information to @samp{hppa*-hp-hpux*} targets follows. | |
4092 | ||
4093 | @html | |
4094 | <hr /> | |
4095 | @end html | |
4096 | @anchor{hppa-hp-hpux10} | |
4097 | @heading hppa*-hp-hpux10 | |
4098 | For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch | |
4099 | @code{PHCO_19798} from HP@. | |
4100 | ||
4101 | The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are | |
4102 | used for one-only code and data. This resolves many of the previous | |
4103 | problems in using C++ on this target. However, the ABI is not compatible | |
4104 | with the one implemented under HP-UX 11 using secondary definitions. | |
4105 | ||
4106 | @html | |
4107 | <hr /> | |
4108 | @end html | |
4109 | @anchor{hppa-hp-hpux11} | |
4110 | @heading hppa*-hp-hpux11 | |
4111 | GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot | |
4112 | be used to compile GCC 3.0 and up. | |
4113 | ||
4114 | The libffi library haven't been ported to 64-bit HP-UX@ and doesn't build. | |
4115 | ||
4116 | Refer to @uref{binaries.html,,binaries} for information about obtaining | |
4117 | precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained | |
4118 | to build the Ada language as it cannot be bootstrapped using C@. Ada is | |
4119 | only available for the 32-bit PA-RISC runtime. | |
4120 | ||
4121 | Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The | |
4122 | bundled compiler supports only traditional C; you will need either HP's | |
4123 | unbundled compiler, or a binary distribution of GCC@. | |
4124 | ||
4125 | It is possible to build GCC 3.3 starting with the bundled HP compiler, | |
4126 | but the process requires several steps. GCC 3.3 can then be used to | |
4127 | build later versions. | |
4128 | ||
4129 | There are several possible approaches to building the distribution. | |
4130 | Binutils can be built first using the HP tools. Then, the GCC | |
4131 | distribution can be built. The second approach is to build GCC | |
4132 | first using the HP tools, then build binutils, then rebuild GCC@. | |
4133 | There have been problems with various binary distributions, so it | |
4134 | is best not to start from a binary distribution. | |
4135 | ||
4136 | On 64-bit capable systems, there are two distinct targets. Different | |
4137 | installation prefixes must be used if both are to be installed on | |
4138 | the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code | |
4139 | for the 32-bit PA-RISC runtime architecture and uses the HP linker. | |
4140 | The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the | |
4141 | PA-RISC 2.0 architecture. | |
4142 | ||
4143 | The script config.guess now selects the target type based on the compiler | |
4144 | detected during configuration. You must define @env{PATH} or @env{CC} so | |
4145 | that configure finds an appropriate compiler for the initial bootstrap. | |
4146 | When @env{CC} is used, the definition should contain the options that are | |
4147 | needed whenever @env{CC} is used. | |
4148 | ||
4149 | Specifically, options that determine the runtime architecture must be | |
4150 | in @env{CC} to correctly select the target for the build. It is also | |
4151 | convenient to place many other compiler options in @env{CC}. For example, | |
4152 | @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"} | |
4153 | can be used to bootstrap the GCC 3.3 branch with the HP compiler in | |
4154 | 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in | |
4155 | the automatic selection of the @samp{hppa64-hp-hpux11*} target. The | |
4156 | macro definition table of cpp needs to be increased for a successful | |
4157 | build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to | |
4158 | be defined when building with the bundled compiler, or when using the | |
4159 | @option{-Ac} option. These defines aren't necessary with @option{-Ae}. | |
4160 | ||
4161 | It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target | |
4162 | with the @option{--with-ld=@dots{}} option. This overrides the standard | |
4163 | search for ld. The two linkers supported on this target require different | |
4164 | commands. The default linker is determined during configuration. As a | |
4165 | result, it's not possible to switch linkers in the middle of a GCC build. | |
4166 | This has been reported to sometimes occur in unified builds of binutils | |
4167 | and GCC@. | |
4168 | ||
4169 | A recent linker patch must be installed for the correct operation of | |
4170 | GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the | |
4171 | oldest linker patches that are known to work. They are for HP-UX | |
4172 | 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to | |
4173 | @code{PHSS_24304}, might be usable but it hasn't been tested. These | |
4174 | patches have been superseded. Consult the HP patch database to obtain | |
4175 | the currently recommended linker patch for your system. | |
4176 | ||
4177 | The patches are necessary for the support of weak symbols on the | |
4178 | 32-bit port, and for the running of initializers and finalizers. Weak | |
4179 | symbols are implemented using SOM secondary definition symbols. Prior | |
4180 | to HP-UX 11, there are bugs in the linker support for secondary symbols. | |
4181 | The patches correct a problem of linker core dumps creating shared | |
4182 | libraries containing secondary symbols, as well as various other | |
4183 | linking issues involving secondary symbols. | |
4184 | ||
4185 | GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to | |
4186 | run initializers and finalizers on the 64-bit port. The 32-bit port | |
4187 | uses the linker @option{+init} and @option{+fini} options for the same | |
4188 | purpose. The patches correct various problems with the +init/+fini | |
4189 | options, including program core dumps. Binutils 2.14 corrects a | |
4190 | problem on the 64-bit port resulting from HP's non-standard use of | |
4191 | the .init and .fini sections for array initializers and finalizers. | |
4192 | ||
4193 | Although the HP and GNU linkers are both supported for the | |
4194 | @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the | |
4195 | HP linker be used for link editing on this target. | |
4196 | ||
4197 | At this time, the GNU linker does not support the creation of long | |
4198 | branch stubs. As a result, it cannot successfully link binaries | |
4199 | containing branch offsets larger than 8 megabytes. In addition, | |
4200 | there are problems linking shared libraries, linking executables | |
4201 | with @option{-static}, and with dwarf2 unwind and exception support. | |
4202 | It also doesn't provide stubs for internal calls to global functions | |
4203 | in shared libraries, so these calls cannot be overloaded. | |
4204 | ||
4205 | The HP dynamic loader does not support GNU symbol versioning, so symbol | |
4206 | versioning is not supported. It may be necessary to disable symbol | |
4207 | versioning with @option{--disable-symvers} when using GNU ld. | |
4208 | ||
4209 | POSIX threads are the default. The optional DCE thread library is not | |
4210 | supported, so @option{--enable-threads=dce} does not work. | |
4211 | ||
4212 | @html | |
4213 | <hr /> | |
4214 | @end html | |
4215 | @anchor{x-x-linux-gnu} | |
4216 | @heading *-*-linux-gnu | |
4217 | The @code{.init_array} and @code{.fini_array} sections are enabled | |
4218 | unconditionally which requires at least glibc 2.1 and binutils 2.12. | |
4219 | ||
4220 | Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present | |
4221 | in glibc 2.2.5 and later. More information is available in the | |
4222 | libstdc++-v3 documentation. | |
4223 | ||
4224 | @html | |
4225 | <hr /> | |
4226 | @end html | |
4227 | @anchor{ix86-x-linux} | |
4228 | @heading i?86-*-linux* | |
4229 | As of GCC 3.3, binutils 2.13.1 or later is required for this platform. | |
4230 | See @uref{https://gcc.gnu.org/PR10877,,bug 10877} for more information. | |
4231 | ||
4232 | If you receive Signal 11 errors when building on GNU/Linux, then it is | |
4233 | possible you have a hardware problem. Further information on this can be | |
4234 | found on @uref{https://www.bitwizard.nl/sig11/,,www.bitwizard.nl}. | |
4235 | ||
4236 | @html | |
4237 | <hr /> | |
4238 | @end html | |
4239 | @anchor{ix86-x-solaris2} | |
4240 | @heading i?86-*-solaris2* | |
4241 | Use this for Solaris 11.3 or later on x86 and x86-64 systems. Starting | |
4242 | with GCC 4.7, there is also a 64-bit @samp{amd64-*-solaris2*} or | |
4243 | @samp{x86_64-*-solaris2*} configuration that corresponds to | |
4244 | @samp{sparcv9-sun-solaris2*}. | |
4245 | ||
4246 | It is recommended that you configure GCC to use the GNU assembler. The | |
4247 | versions included in Solaris 11.3, from GNU binutils 2.23.1 or | |
4248 | newer (available as @file{/usr/bin/gas} and | |
4249 | @file{/usr/gnu/bin/as}), work fine. The current version, from GNU | |
4250 | binutils 2.34, is known to work. Recent versions of the Solaris assembler in | |
4251 | @file{/usr/bin/as} work almost as well, though. | |
4252 | ||
4253 | For linking, the Solaris linker is preferred. If you want to use the GNU | |
4254 | linker instead, the version in Solaris 11.3, from GNU binutils 2.23.1 or | |
4255 | newer (in @file{/usr/gnu/bin/ld} and @file{/usr/bin/gld}), works, | |
4256 | as does the latest version, from GNU binutils 2.34. | |
4257 | ||
4258 | To use GNU @command{as}, configure with the options | |
4259 | @option{--with-gnu-as --with-as=@//usr/@/gnu/@/bin/@/as}. It may be necessary | |
4260 | to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to | |
4261 | guarantee use of Solaris @command{ld}. | |
4262 | @c FIXME: why --without-gnu-ld --with-ld? | |
4263 | ||
4264 | @html | |
4265 | <hr /> | |
4266 | @end html | |
4267 | @anchor{ia64-x-linux} | |
4268 | @heading ia64-*-linux | |
4269 | IA-64 processor (also known as IPF, or Itanium Processor Family) | |
4270 | running GNU/Linux. | |
4271 | ||
4272 | If you are using the installed system libunwind library with | |
4273 | @option{--with-system-libunwind}, then you must use libunwind 0.98 or | |
4274 | later. | |
4275 | ||
4276 | @html | |
4277 | <hr /> | |
4278 | @end html | |
4279 | @anchor{ia64-x-hpux} | |
4280 | @heading ia64-*-hpux* | |
4281 | Building GCC on this target requires the GNU Assembler. The bundled HP | |
4282 | assembler will not work. To prevent GCC from using the wrong assembler, | |
4283 | the option @option{--with-gnu-as} may be necessary. | |
4284 | ||
4285 | The GCC libunwind library has not been ported to HPUX@. This means that for | |
4286 | GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions} | |
4287 | is required to build GCC@. For GCC 3.3 and later, this is the default. | |
4288 | For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is | |
4289 | removed and the system libunwind library will always be used. | |
4290 | ||
4291 | @html | |
4292 | <hr /> | |
4293 | <!-- rs6000-ibm-aix*, powerpc-ibm-aix* --> | |
4294 | @end html | |
4295 | @anchor{x-ibm-aix} | |
4296 | @heading *-ibm-aix* | |
4297 | Support for AIX version 3 and older was discontinued in GCC 3.4. | |
4298 | Support for AIX version 4.2 and older was discontinued in GCC 4.5. | |
4299 | ||
4300 | ``out of memory'' bootstrap failures may indicate a problem with | |
4301 | process resource limits (ulimit). Hard limits are configured in the | |
4302 | @file{/etc/security/limits} system configuration file. | |
4303 | ||
4304 | GCC 4.9 and above require a C++ compiler for bootstrap. IBM VAC++ / xlC | |
4305 | cannot bootstrap GCC. xlc can bootstrap an older version of GCC and | |
4306 | G++ can bootstrap recent releases of GCC. | |
4307 | ||
4308 | GCC can bootstrap with recent versions of IBM XLC, but bootstrapping | |
4309 | with an earlier release of GCC is recommended. Bootstrapping with XLC | |
4310 | requires a larger data segment, which can be enabled through the | |
4311 | @var{LDR_CNTRL} environment variable, e.g., | |
4312 | ||
4313 | @smallexample | |
4314 | % LDR_CNTRL=MAXDATA=0x50000000 | |
4315 | % export LDR_CNTRL | |
4316 | @end smallexample | |
4317 | ||
4318 | One can start with a pre-compiled version of GCC to build from | |
4319 | sources. One may delete GCC's ``fixed'' header files when starting | |
4320 | with a version of GCC built for an earlier release of AIX. | |
4321 | ||
4322 | To speed up the configuration phases of bootstrapping and installing GCC, | |
4323 | one may use GNU Bash instead of AIX @command{/bin/sh}, e.g., | |
4324 | ||
4325 | @smallexample | |
4326 | % CONFIG_SHELL=/opt/freeware/bin/bash | |
4327 | % export CONFIG_SHELL | |
4328 | @end smallexample | |
4329 | ||
4330 | and then proceed as described in @uref{build.html,,the build | |
4331 | instructions}, where we strongly recommend specifying an absolute path | |
4332 | to invoke @var{srcdir}/configure. | |
4333 | ||
4334 | Because GCC on AIX is built as a 32-bit executable by default, | |
4335 | (although it can generate 64-bit programs) the GMP and MPFR libraries | |
4336 | required by gfortran must be 32-bit libraries. Building GMP and MPFR | |
4337 | as static archive libraries works better than shared libraries. | |
4338 | ||
4339 | Errors involving @code{alloca} when building GCC generally are due | |
4340 | to an incorrect definition of @code{CC} in the Makefile or mixing files | |
4341 | compiled with the native C compiler and GCC@. During the stage1 phase of | |
4342 | the build, the native AIX compiler @strong{must} be invoked as @command{cc} | |
4343 | (not @command{xlc}). Once @command{configure} has been informed of | |
4344 | @command{xlc}, one needs to use @samp{make distclean} to remove the | |
4345 | configure cache files and ensure that @env{CC} environment variable | |
4346 | does not provide a definition that will confuse @command{configure}. | |
4347 | If this error occurs during stage2 or later, then the problem most likely | |
4348 | is the version of Make (see above). | |
4349 | ||
4350 | The native @command{as} and @command{ld} are recommended for | |
4351 | bootstrapping on AIX@. The GNU Assembler, GNU Linker, and GNU | |
4352 | Binutils version 2.20 is the minimum level that supports bootstrap on | |
4353 | AIX 5@. The GNU Assembler has not been updated to support AIX 6@ or | |
4354 | AIX 7. The native AIX tools do interoperate with GCC@. | |
4355 | ||
4356 | AIX 7.1 added partial support for DWARF debugging, but full support | |
4357 | requires AIX 7.1 TL03 SP7 that supports additional DWARF sections and | |
4358 | fixes a bug in the assembler. AIX 7.1 TL03 SP5 distributed a version | |
4359 | of libm.a missing important symbols; a fix for IV77796 will be | |
4360 | included in SP6. | |
4361 | ||
4362 | AIX 5.3 TL10, AIX 6.1 TL05 and AIX 7.1 TL00 introduced an AIX | |
4363 | assembler change that sometimes produces corrupt assembly files | |
4364 | causing AIX linker errors. The bug breaks GCC bootstrap on AIX and | |
4365 | can cause compilation failures with existing GCC installations. An | |
4366 | AIX iFix for AIX 5.3 is available (APAR IZ98385 for AIX 5.3 TL10, APAR | |
4367 | IZ98477 for AIX 5.3 TL11 and IZ98134 for AIX 5.3 TL12). AIX 5.3 TL11 SP8, | |
4368 | AIX 5.3 TL12 SP5, AIX 6.1 TL04 SP11, AIX 6.1 TL05 SP7, AIX 6.1 TL06 SP6, | |
4369 | AIX 6.1 TL07 and AIX 7.1 TL01 should include the fix. | |
4370 | ||
4371 | Building @file{libstdc++.a} requires a fix for an AIX Assembler bug | |
4372 | APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a | |
4373 | fix for another AIX Assembler bug and a co-dependent AIX Archiver fix | |
4374 | referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1) | |
4375 | ||
4376 | @anchor{TransferAixShobj} | |
4377 | @samp{libstdc++} in GCC 3.4 increments the major version number of the | |
4378 | shared object and GCC installation places the @file{libstdc++.a} | |
4379 | shared library in a common location which will overwrite the and GCC | |
4380 | 3.3 version of the shared library. Applications either need to be | |
4381 | re-linked against the new shared library or the GCC 3.1 and GCC 3.3 | |
4382 | versions of the @samp{libstdc++} shared object needs to be available | |
4383 | to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if | |
4384 | present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be | |
4385 | installed for runtime dynamic loading using the following steps to set | |
4386 | the @samp{F_LOADONLY} flag in the shared object for @emph{each} | |
4387 | multilib @file{libstdc++.a} installed: | |
4388 | ||
4389 | Extract the shared objects from the currently installed | |
4390 | @file{libstdc++.a} archive: | |
4391 | @smallexample | |
4392 | % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5 | |
4393 | @end smallexample | |
4394 | ||
4395 | Enable the @samp{F_LOADONLY} flag so that the shared object will be | |
4396 | available for runtime dynamic loading, but not linking: | |
4397 | @smallexample | |
4398 | % strip -e libstdc++.so.4 libstdc++.so.5 | |
4399 | @end smallexample | |
4400 | ||
4401 | Archive the runtime-only shared object in the GCC 3.4 | |
4402 | @file{libstdc++.a} archive: | |
4403 | @smallexample | |
4404 | % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5 | |
4405 | @end smallexample | |
4406 | ||
4407 | Eventually, the | |
4408 | @uref{./configure.html#WithAixSoname,,@option{--with-aix-soname=svr4}} | |
4409 | configure option may drop the need for this procedure for libraries that | |
4410 | support it. | |
4411 | ||
4412 | Linking executables and shared libraries may produce warnings of | |
4413 | duplicate symbols. The assembly files generated by GCC for AIX always | |
4414 | have included multiple symbol definitions for certain global variable | |
4415 | and function declarations in the original program. The warnings should | |
4416 | not prevent the linker from producing a correct library or runnable | |
4417 | executable. | |
4418 | ||
4419 | AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and | |
4420 | 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1 | |
4421 | to parse archive libraries did not handle the new format correctly. | |
4422 | These routines are used by GCC and result in error messages during | |
4423 | linking such as ``not a COFF file''. The version of the routines shipped | |
4424 | with AIX 4.3.1 should work for a 32-bit environment. The @option{-g} | |
4425 | option of the archive command may be used to create archives of 32-bit | |
4426 | objects using the original ``small format''. A correct version of the | |
4427 | routines is shipped with AIX 4.3.2 and above. | |
4428 | ||
4429 | Some versions of the AIX binder (linker) can fail with a relocation | |
4430 | overflow severe error when the @option{-bbigtoc} option is used to link | |
4431 | GCC-produced object files into an executable that overflows the TOC@. A fix | |
4432 | for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is | |
4433 | available from IBM Customer Support and from its | |
4434 | @uref{https://techsupport.services.ibm.com/,,techsupport.services.ibm.com} | |
4435 | website as PTF U455193. | |
4436 | ||
4437 | The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core | |
4438 | with a segmentation fault when invoked by any version of GCC@. A fix for | |
4439 | APAR IX87327 is available from IBM Customer Support and from its | |
4440 | @uref{https://techsupport.services.ibm.com/,,techsupport.services.ibm.com} | |
4441 | website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above. | |
4442 | ||
4443 | The initial assembler shipped with AIX 4.3.0 generates incorrect object | |
4444 | files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS | |
4445 | TO ASSEMBLE/BIND) is available from IBM Customer Support and from its | |
4446 | @uref{https://techsupport.services.ibm.com/,,techsupport.services.ibm.com} | |
4447 | website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above. | |
4448 | ||
4449 | AIX provides National Language Support (NLS)@. Compilers and assemblers | |
4450 | use NLS to support locale-specific representations of various data | |
4451 | formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for | |
4452 | separating decimal fractions). There have been problems reported where | |
4453 | GCC does not produce the same floating-point formats that the assembler | |
4454 | expects. If one encounters this problem, set the @env{LANG} | |
4455 | environment variable to @samp{C} or @samp{En_US}. | |
4456 | ||
4457 | A default can be specified with the @option{-mcpu=@var{cpu_type}} | |
4458 | switch and using the configure option @option{--with-cpu-@var{cpu_type}}. | |
4459 | ||
4460 | @html | |
4461 | <hr /> | |
4462 | @end html | |
4463 | @anchor{iq2000-x-elf} | |
4464 | @heading iq2000-*-elf | |
4465 | Vitesse IQ2000 processors. These are used in embedded | |
4466 | applications. There are no standard Unix configurations. | |
4467 | ||
4468 | @html | |
4469 | <hr /> | |
4470 | @end html | |
4471 | @anchor{lm32-x-elf} | |
4472 | @heading lm32-*-elf | |
4473 | Lattice Mico32 processor. | |
4474 | This configuration is intended for embedded systems. | |
4475 | ||
4476 | @html | |
4477 | <hr /> | |
4478 | @end html | |
4479 | @anchor{lm32-x-uclinux} | |
4480 | @heading lm32-*-uclinux | |
4481 | Lattice Mico32 processor. | |
4482 | This configuration is intended for embedded systems running uClinux. | |
4483 | ||
4484 | @html | |
4485 | <hr /> | |
4486 | @end html | |
4487 | @anchor{loongarch} | |
4488 | @heading LoongArch | |
4489 | LoongArch processor. | |
4490 | The following LoongArch targets are available: | |
4491 | @table @code | |
4492 | @item loongarch64-linux-gnu* | |
4493 | LoongArch processor running GNU/Linux. This target triplet may be coupled | |
4494 | with a small set of possible suffixes to identify their default ABI type: | |
4495 | @table @code | |
4496 | @item f64 | |
4497 | Uses @code{lp64d/base} ABI by default. | |
4498 | @item f32 | |
4499 | Uses @code{lp64f/base} ABI by default. | |
4500 | @item sf | |
4501 | Uses @code{lp64s/base} ABI by default. | |
4502 | @end table | |
4503 | ||
4504 | @item loongarch64-linux-gnu | |
4505 | Same as @code{loongarch64-linux-gnuf64}, but may be used with | |
4506 | @option{--with-abi=*} to configure the default ABI type. | |
4507 | @end table | |
4508 | ||
4509 | More information about LoongArch can be found at | |
4510 | @uref{https://github.com/loongson/LoongArch-Documentation}. | |
4511 | ||
4512 | @html | |
4513 | <hr /> | |
4514 | @end html | |
4515 | @anchor{m32c-x-elf} | |
4516 | @heading m32c-*-elf | |
4517 | Renesas M32C processor. | |
4518 | This configuration is intended for embedded systems. | |
4519 | ||
4520 | @html | |
4521 | <hr /> | |
4522 | @end html | |
4523 | @anchor{m32r-x-elf} | |
4524 | @heading m32r-*-elf | |
4525 | Renesas M32R processor. | |
4526 | This configuration is intended for embedded systems. | |
4527 | ||
4528 | @html | |
4529 | <hr /> | |
4530 | @end html | |
4531 | @anchor{m68k-x-x} | |
4532 | @heading m68k-*-* | |
4533 | By default, | |
4534 | @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and | |
4535 | @samp{m68k-*-linux} | |
4536 | build libraries for both M680x0 and ColdFire processors. If you only | |
4537 | need the M680x0 libraries, you can omit the ColdFire ones by passing | |
4538 | @option{--with-arch=m68k} to @command{configure}. Alternatively, you | |
4539 | can omit the M680x0 libraries by passing @option{--with-arch=cf} to | |
4540 | @command{configure}. These targets default to 5206 or 5475 code as | |
4541 | appropriate for the target system when | |
4542 | configured with @option{--with-arch=cf} and 68020 code otherwise. | |
4543 | ||
4544 | The @samp{m68k-*-netbsd} and | |
4545 | @samp{m68k-*-openbsd} targets also support the @option{--with-arch} | |
4546 | option. They will generate ColdFire CFV4e code when configured with | |
4547 | @option{--with-arch=cf} and 68020 code otherwise. | |
4548 | ||
4549 | You can override the default processors listed above by configuring | |
4550 | with @option{--with-cpu=@var{target}}. This @var{target} can either | |
4551 | be a @option{-mcpu} argument or one of the following values: | |
4552 | @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030}, | |
4553 | @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}. | |
4554 | ||
4555 | GCC requires at least binutils version 2.17 on these targets. | |
4556 | ||
4557 | @html | |
4558 | <hr /> | |
4559 | @end html | |
4560 | @anchor{m68k-x-uclinux} | |
4561 | @heading m68k-*-uclinux | |
4562 | GCC 4.3 changed the uClinux configuration so that it uses the | |
4563 | @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI. | |
4564 | It also added improved support for C++ and flat shared libraries, | |
4565 | both of which were ABI changes. | |
4566 | ||
4567 | @html | |
4568 | <hr /> | |
4569 | @end html | |
4570 | @anchor{microblaze-x-elf} | |
4571 | @heading microblaze-*-elf | |
4572 | Xilinx MicroBlaze processor. | |
4573 | This configuration is intended for embedded systems. | |
4574 | ||
4575 | @html | |
4576 | <hr /> | |
4577 | @end html | |
4578 | @anchor{mips-x-x} | |
4579 | @heading mips-*-* | |
4580 | If on a MIPS system you get an error message saying ``does not have gp | |
4581 | sections for all it's [sic] sectons [sic]'', don't worry about it. This | |
4582 | happens whenever you use GAS with the MIPS linker, but there is not | |
4583 | really anything wrong, and it is okay to use the output file. You can | |
4584 | stop such warnings by installing the GNU linker. | |
4585 | ||
4586 | It would be nice to extend GAS to produce the gp tables, but they are | |
4587 | optional, and there should not be a warning about their absence. | |
4588 | ||
4589 | The libstdc++ atomic locking routines for MIPS targets requires MIPS II | |
4590 | and later. A patch went in just after the GCC 3.3 release to | |
4591 | make @samp{mips*-*-*} use the generic implementation instead. You can also | |
4592 | configure for @samp{mipsel-elf} as a workaround. The | |
4593 | @samp{mips*-*-linux*} target continues to use the MIPS II routines. More | |
4594 | work on this is expected in future releases. | |
4595 | ||
4596 | @c If you make --with-llsc the default for another target, please also | |
4597 | @c update the description of the --with-llsc option. | |
4598 | ||
4599 | The built-in @code{__sync_*} functions are available on MIPS II and | |
4600 | later systems and others that support the @samp{ll}, @samp{sc} and | |
4601 | @samp{sync} instructions. This can be overridden by passing | |
4602 | @option{--with-llsc} or @option{--without-llsc} when configuring GCC. | |
4603 | Since the Linux kernel emulates these instructions if they are | |
4604 | missing, the default for @samp{mips*-*-linux*} targets is | |
4605 | @option{--with-llsc}. The @option{--with-llsc} and | |
4606 | @option{--without-llsc} configure options may be overridden at compile | |
4607 | time by passing the @option{-mllsc} or @option{-mno-llsc} options to | |
4608 | the compiler. | |
4609 | ||
4610 | MIPS systems check for division by zero (unless | |
4611 | @option{-mno-check-zero-division} is passed to the compiler) by | |
4612 | generating either a conditional trap or a break instruction. Using | |
4613 | trap results in smaller code, but is only supported on MIPS II and | |
4614 | later. Also, some versions of the Linux kernel have a bug that | |
4615 | prevents trap from generating the proper signal (@code{SIGFPE}). To enable | |
4616 | the use of break, use the @option{--with-divide=breaks} | |
4617 | @command{configure} option when configuring GCC@. The default is to | |
4618 | use traps on systems that support them. | |
4619 | ||
4620 | @html | |
4621 | <hr /> | |
4622 | @end html | |
4623 | @anchor{moxie-x-elf} | |
4624 | @heading moxie-*-elf | |
4625 | The moxie processor. | |
4626 | ||
4627 | @html | |
4628 | <hr /> | |
4629 | @end html | |
4630 | @anchor{msp430-x-elf} | |
4631 | @heading msp430-*-elf* | |
4632 | TI MSP430 processor. | |
4633 | This configuration is intended for embedded systems. | |
4634 | ||
4635 | @samp{msp430-*-elf} is the standard configuration with most GCC | |
4636 | features enabled by default. | |
4637 | ||
4638 | @samp{msp430-*-elfbare} is tuned for a bare-metal environment, and disables | |
4639 | features related to shared libraries and other functionality not used for | |
4640 | this device. This reduces code and data usage of the GCC libraries, resulting | |
4641 | in a minimal run-time environment by default. | |
4642 | ||
4643 | Features disabled by default include: | |
4644 | @itemize | |
4645 | @item transactional memory | |
4646 | @item __cxa_atexit | |
4647 | @end itemize | |
4648 | ||
4649 | @html | |
4650 | <hr /> | |
4651 | @end html | |
4652 | @anchor{nds32le-x-elf} | |
4653 | @heading nds32le-*-elf | |
4654 | Andes NDS32 target in little endian mode. | |
4655 | ||
4656 | @html | |
4657 | <hr /> | |
4658 | @end html | |
4659 | @anchor{nds32be-x-elf} | |
4660 | @heading nds32be-*-elf | |
4661 | Andes NDS32 target in big endian mode. | |
4662 | ||
4663 | @html | |
4664 | <hr /> | |
4665 | @end html | |
4666 | @anchor{nvptx-x-none} | |
4667 | @heading nvptx-*-none | |
4668 | Nvidia PTX target. | |
4669 | ||
4670 | Instead of GNU binutils, you will need to install | |
4671 | @uref{https://github.com/MentorEmbedded/nvptx-tools/,,nvptx-tools}. | |
4672 | Tell GCC where to find it: | |
4673 | @option{--with-build-time-tools=[install-nvptx-tools]/nvptx-none/bin}. | |
4674 | ||
4675 | You will need newlib 3.1.0 or later. It can be | |
4676 | automatically built together with GCC@. For this, add a symbolic link | |
4677 | to nvptx-newlib's @file{newlib} directory to the directory containing | |
4678 | the GCC sources. | |
4679 | ||
4680 | Use the @option{--disable-sjlj-exceptions} and | |
4681 | @option{--enable-newlib-io-long-long} options when configuring. | |
4682 | ||
4683 | The @option{--with-arch} option may be specified to override the | |
4684 | default value for the @option{-march} option, and to also build | |
4685 | corresponding target libraries. | |
4686 | The default is @option{--with-arch=sm_30}. | |
4687 | ||
4688 | For example, if @option{--with-arch=sm_70} is specified, | |
4689 | @option{-march=sm_30} and @option{-march=sm_70} target libraries are | |
4690 | built, and code generation defaults to @option{-march=sm_70}. | |
4691 | ||
4692 | @html | |
4693 | <hr /> | |
4694 | @end html | |
4695 | @anchor{or1k-x-elf} | |
4696 | @heading or1k-*-elf | |
4697 | The OpenRISC 1000 32-bit processor with delay slots. | |
4698 | This configuration is intended for embedded systems. | |
4699 | ||
4700 | @html | |
4701 | <hr /> | |
4702 | @end html | |
4703 | @anchor{or1k-x-linux} | |
4704 | @heading or1k-*-linux | |
4705 | The OpenRISC 1000 32-bit processor with delay slots. | |
4706 | ||
4707 | @html | |
4708 | <hr /> | |
4709 | @end html | |
4710 | @anchor{powerpc-x-x} | |
4711 | @heading powerpc-*-* | |
4712 | You can specify a default version for the @option{-mcpu=@var{cpu_type}} | |
4713 | switch by using the configure option @option{--with-cpu-@var{cpu_type}}. | |
4714 | ||
4715 | You will need GNU binutils 2.20 or newer. | |
4716 | ||
4717 | @html | |
4718 | <hr /> | |
4719 | @end html | |
4720 | @anchor{powerpc-x-darwin} | |
4721 | @heading powerpc-*-darwin* | |
4722 | PowerPC running Darwin (Mac OS X kernel). | |
4723 | ||
4724 | Pre-installed versions of Mac OS X may not include any developer tools, | |
4725 | meaning that you will not be able to build GCC from source. Tool | |
4726 | binaries are available at | |
4727 | @uref{https://opensource.apple.com}. | |
4728 | ||
4729 | This version of GCC requires at least cctools-590.36. The | |
4730 | cctools-590.36 package referenced from | |
4731 | @uref{https://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work | |
4732 | on systems older than 10.3.9 (aka darwin7.9.0). | |
4733 | ||
4734 | @html | |
4735 | <hr /> | |
4736 | @end html | |
4737 | @anchor{powerpc-x-elf} | |
4738 | @heading powerpc-*-elf | |
4739 | PowerPC system in big endian mode, running System V.4. | |
4740 | ||
4741 | @html | |
4742 | <hr /> | |
4743 | @end html | |
4744 | @anchor{powerpc-x-linux-gnu} | |
4745 | @heading powerpc*-*-linux-gnu* | |
4746 | PowerPC system in big endian mode running Linux. | |
4747 | ||
4748 | @html | |
4749 | <hr /> | |
4750 | @end html | |
4751 | @anchor{powerpc-x-netbsd} | |
4752 | @heading powerpc-*-netbsd* | |
4753 | PowerPC system in big endian mode running NetBSD@. | |
4754 | ||
4755 | @html | |
4756 | <hr /> | |
4757 | @end html | |
4758 | @anchor{powerpc-x-eabisim} | |
4759 | @heading powerpc-*-eabisim | |
4760 | Embedded PowerPC system in big endian mode for use in running under the | |
4761 | PSIM simulator. | |
4762 | ||
4763 | @html | |
4764 | <hr /> | |
4765 | @end html | |
4766 | @anchor{powerpc-x-eabi} | |
4767 | @heading powerpc-*-eabi | |
4768 | Embedded PowerPC system in big endian mode. | |
4769 | ||
4770 | @html | |
4771 | <hr /> | |
4772 | @end html | |
4773 | @anchor{powerpcle-x-elf} | |
4774 | @heading powerpcle-*-elf | |
4775 | PowerPC system in little endian mode, running System V.4. | |
4776 | ||
4777 | @html | |
4778 | <hr /> | |
4779 | @end html | |
4780 | @anchor{powerpcle-x-eabisim} | |
4781 | @heading powerpcle-*-eabisim | |
4782 | Embedded PowerPC system in little endian mode for use in running under | |
4783 | the PSIM simulator. | |
4784 | ||
4785 | @html | |
4786 | <hr /> | |
4787 | @end html | |
4788 | @anchor{powerpcle-x-eabi} | |
4789 | @heading powerpcle-*-eabi | |
4790 | Embedded PowerPC system in little endian mode. | |
4791 | ||
4792 | @html | |
4793 | <hr /> | |
4794 | @end html | |
4795 | @anchor{rl78-x-elf} | |
4796 | @heading rl78-*-elf | |
4797 | The Renesas RL78 processor. | |
4798 | This configuration is intended for embedded systems. | |
4799 | ||
4800 | @html | |
4801 | <hr /> | |
4802 | @end html | |
4803 | @anchor{riscv32-x-elf} | |
4804 | @heading riscv32-*-elf | |
4805 | The RISC-V RV32 instruction set. | |
4806 | This configuration is intended for embedded systems. | |
4807 | This (and all other RISC-V) targets require the binutils 2.30 release. | |
4808 | ||
4809 | @html | |
4810 | <hr /> | |
4811 | @end html | |
4812 | @anchor{riscv32-x-linux} | |
4813 | @heading riscv32-*-linux | |
4814 | The RISC-V RV32 instruction set running GNU/Linux. | |
4815 | This (and all other RISC-V) targets require the binutils 2.30 release. | |
4816 | ||
4817 | @html | |
4818 | <hr /> | |
4819 | @end html | |
4820 | @anchor{riscv64-x-elf} | |
4821 | @heading riscv64-*-elf | |
4822 | The RISC-V RV64 instruction set. | |
4823 | This configuration is intended for embedded systems. | |
4824 | This (and all other RISC-V) targets require the binutils 2.30 release. | |
4825 | ||
4826 | @html | |
4827 | <hr /> | |
4828 | @end html | |
4829 | @anchor{riscv64-x-linux} | |
4830 | @heading riscv64-*-linux | |
4831 | The RISC-V RV64 instruction set running GNU/Linux. | |
4832 | This (and all other RISC-V) targets require the binutils 2.30 release. | |
4833 | ||
4834 | @html | |
4835 | <hr /> | |
4836 | @end html | |
4837 | @anchor{rx-x-elf} | |
4838 | @heading rx-*-elf | |
4839 | The Renesas RX processor. | |
4840 | ||
4841 | @html | |
4842 | <hr /> | |
4843 | @end html | |
4844 | @anchor{s390-x-linux} | |
4845 | @heading s390-*-linux* | |
4846 | S/390 system running GNU/Linux for S/390@. | |
4847 | ||
4848 | @html | |
4849 | <hr /> | |
4850 | @end html | |
4851 | @anchor{s390x-x-linux} | |
4852 | @heading s390x-*-linux* | |
4853 | zSeries system (64-bit) running GNU/Linux for zSeries@. | |
4854 | ||
4855 | @html | |
4856 | <hr /> | |
4857 | @end html | |
4858 | @anchor{s390x-ibm-tpf} | |
4859 | @heading s390x-ibm-tpf* | |
4860 | zSeries system (64-bit) running TPF@. This platform is | |
4861 | supported as cross-compilation target only. | |
4862 | ||
4863 | @html | |
4864 | <hr /> | |
4865 | @end html | |
4866 | @c Please use Solaris 2 to refer to all release of Solaris, starting | |
4867 | @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for | |
4868 | @c SunOS 4 releases which we don't use to avoid confusion. Solaris | |
4869 | @c alone is too unspecific and must be avoided. | |
4870 | @anchor{x-x-solaris2} | |
4871 | @heading *-*-solaris2* | |
fd69977f RO |
4872 | Support for Solaris 11.3 and earlier has been obsoleted in GCC 13, but |
4873 | can still be enabled by configuring with @option{--enable-obsolete}. | |
d77de738 ML |
4874 | Support for Solaris 10 has been removed in GCC 10. Support for Solaris |
4875 | 9 has been removed in GCC 5. Support for Solaris 8 has been removed in | |
4876 | GCC 4.8. Support for Solaris 7 has been removed in GCC 4.6. | |
4877 | ||
4878 | Solaris 11.3 provides GCC 4.5.2, 4.7.3, and 4.8.2 as | |
4879 | @command{/usr/gcc/4.5/bin/gcc} or similar. Newer Solaris versions | |
4880 | provide one or more of GCC 5, 7, and 9. Alternatively, | |
4881 | you can install a pre-built GCC to bootstrap and install GCC. See the | |
4882 | @uref{binaries.html,,binaries page} for details. | |
4883 | ||
4884 | The Solaris 2 @command{/bin/sh} will often fail to configure | |
4885 | @samp{libstdc++-v3}. We therefore recommend using the | |
4886 | following initial sequence of commands | |
4887 | ||
4888 | @smallexample | |
4889 | % CONFIG_SHELL=/bin/ksh | |
4890 | % export CONFIG_SHELL | |
4891 | @end smallexample | |
4892 | ||
4893 | @noindent | |
4894 | and proceed as described in @uref{configure.html,,the configure instructions}. | |
4895 | In addition we strongly recommend specifying an absolute path to invoke | |
4896 | @command{@var{srcdir}/configure}. | |
4897 | ||
4898 | In Solaris 11, you need to check for @code{system/header}, | |
4899 | @code{system/linker}, and @code{developer/assembler} packages. | |
4900 | ||
4901 | Trying to use the linker and other tools in | |
4902 | @file{/usr/ucb} to install GCC has been observed to cause trouble. | |
4903 | For example, the linker may hang indefinitely. The fix is to remove | |
4904 | @file{/usr/ucb} from your @env{PATH}. | |
4905 | ||
4906 | The build process works more smoothly with the legacy Solaris tools so, if you | |
4907 | have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place | |
4908 | @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build. | |
4909 | ||
4910 | We recommend the use of the Solaris assembler or the GNU assembler, in | |
4911 | conjunction with the Solaris linker. The GNU @command{as} | |
4912 | versions included in Solaris 11.3, | |
4913 | from GNU binutils 2.23.1 or newer (in @file{/usr/bin/gas} and | |
4914 | @file{/usr/gnu/bin/as}), are known to work. | |
4915 | The current version, from GNU binutils 2.34, | |
4916 | is known to work as well. Note that your mileage may vary | |
4917 | if you use a combination of the GNU tools and the Solaris tools: while the | |
4918 | combination GNU @command{as} + Solaris @command{ld} should reasonably work, | |
4919 | the reverse combination Solaris @command{as} + GNU @command{ld} may fail to | |
4920 | build or cause memory corruption at runtime in some cases for C++ programs. | |
4921 | @c FIXME: still? | |
4922 | GNU @command{ld} usually works as well. Again, the current | |
4923 | version (2.34) is known to work, but generally lacks platform specific | |
4924 | features, so better stay with Solaris @command{ld}. To use the LTO linker | |
4925 | plugin (@option{-fuse-linker-plugin}) with GNU @command{ld}, GNU | |
4926 | binutils @emph{must} be configured with @option{--enable-largefile}. | |
4927 | ||
4928 | To enable symbol versioning in @samp{libstdc++} with the Solaris linker, | |
4929 | you need to have any version of GNU @command{c++filt}, which is part of | |
4930 | GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no | |
4931 | appropriate version is found. Solaris @command{c++filt} from the Solaris | |
4932 | Studio compilers does @emph{not} work. | |
4933 | ||
4934 | In order to build the GNU D compiler, GDC, a working @samp{libphobos} is | |
4935 | needed. That library wasn't built by default in GCC 9--11 on SPARC, or | |
4936 | on x86 when the Solaris assembler is used, but can be enabled by | |
4937 | configuring with @option{--enable-libphobos}. Also, GDC 9.4.0 is | |
4938 | required on x86, while GDC 9.3.0 is known to work on SPARC. | |
4939 | ||
4940 | The versions of the GNU Multiple Precision Library (GMP), the MPFR | |
4941 | library and the MPC library bundled with Solaris 11.3 and later are | |
4942 | usually recent enough to match GCC's requirements. There are two | |
4943 | caveats: | |
4944 | ||
4945 | @itemize @bullet | |
4946 | @item | |
4947 | While the version of the GMP library in Solaris 11.3 works with GCC, you | |
4948 | need to configure with @option{--with-gmp-include=/usr/include/gmp}. | |
4949 | ||
4950 | @item | |
4951 | The version of the MPFR libary included in Solaris 11.3 is too old; you | |
4952 | need to provide a more recent one. | |
4953 | ||
4954 | @end itemize | |
4955 | ||
4956 | @html | |
4957 | <hr /> | |
4958 | @end html | |
4959 | @anchor{sparc-x-x} | |
4960 | @heading sparc*-*-* | |
4961 | This section contains general configuration information for all | |
4962 | SPARC-based platforms. In addition to reading this section, please | |
4963 | read all other sections that match your target. | |
4964 | ||
4965 | Newer versions of the GNU Multiple Precision Library (GMP), the MPFR | |
4966 | library and the MPC library are known to be miscompiled by earlier | |
4967 | versions of GCC on these platforms. We therefore recommend the use | |
4968 | of the exact versions of these libraries listed as minimal versions | |
4969 | in @uref{prerequisites.html,,the prerequisites}. | |
4970 | ||
4971 | @html | |
4972 | <hr /> | |
4973 | @end html | |
4974 | @anchor{sparc-sun-solaris2} | |
4975 | @heading sparc-sun-solaris2* | |
4976 | When GCC is configured to use GNU binutils 2.14 or later, the binaries | |
4977 | produced are smaller than the ones produced using Solaris native tools; | |
4978 | this difference is quite significant for binaries containing debugging | |
4979 | information. | |
4980 | ||
4981 | Starting with Solaris 7, the operating system is capable of executing | |
4982 | 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports | |
4983 | this; the @option{-m64} option enables 64-bit code generation. | |
4984 | However, if all you want is code tuned for the UltraSPARC CPU, you | |
4985 | should try the @option{-mtune=ultrasparc} option instead, which produces | |
4986 | code that, unlike full 64-bit code, can still run on non-UltraSPARC | |
4987 | machines. | |
4988 | ||
4989 | When configuring the GNU Multiple Precision Library (GMP), the MPFR | |
4990 | library or the MPC library on a Solaris 7 or later system, the canonical | |
4991 | target triplet must be specified as the @command{build} parameter on the | |
4992 | configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and | |
4993 | not that of GMP or MPFR or MPC). For example on a Solaris 11 system: | |
4994 | ||
4995 | @smallexample | |
4996 | % ./configure --build=sparc-sun-solaris2.11 --prefix=xxx | |
4997 | @end smallexample | |
4998 | ||
4999 | @html | |
5000 | <hr /> | |
5001 | @end html | |
5002 | @anchor{sparc-x-linux} | |
5003 | @heading sparc-*-linux* | |
5004 | ||
5005 | @html | |
5006 | <hr /> | |
5007 | @end html | |
5008 | @anchor{sparc64-x-solaris2} | |
5009 | @heading sparc64-*-solaris2* | |
5010 | When configuring a 64-bit-default GCC on Solaris/SPARC, you must use a | |
5011 | build compiler that generates 64-bit code, either by default or by | |
5012 | specifying @samp{CC='gcc -m64' CXX='gcc-m64'} to @command{configure}. | |
5013 | Additionally, you @emph{must} pass @option{--build=sparc64-sun-solaris2.11} | |
5014 | or @option{--build=sparcv9-sun-solaris2.11} because @file{config.guess} | |
5015 | misdetects this situation, which can cause build failures. | |
5016 | ||
5017 | When configuring the GNU Multiple Precision Library (GMP), the MPFR | |
5018 | library or the MPC library, the canonical target triplet must be specified | |
5019 | as the @command{build} parameter on the configure line. For example | |
5020 | on a Solaris 11 system: | |
5021 | ||
5022 | @smallexample | |
5023 | % ./configure --build=sparc64-sun-solaris2.11 --prefix=xxx | |
5024 | @end smallexample | |
5025 | ||
5026 | @html | |
5027 | <hr /> | |
5028 | @end html | |
5029 | @anchor{sparcv9-x-solaris2} | |
5030 | @heading sparcv9-*-solaris2* | |
5031 | This is a synonym for @samp{sparc64-*-solaris2*}. | |
5032 | ||
5033 | @html | |
5034 | <hr /> | |
5035 | @end html | |
5036 | @anchor{c6x-x-x} | |
5037 | @heading c6x-*-* | |
5038 | The C6X family of processors. This port requires binutils-2.22 or newer. | |
5039 | ||
5040 | @html | |
5041 | <hr /> | |
5042 | @end html | |
5043 | @anchor{visium-x-elf} | |
5044 | @heading visium-*-elf | |
5045 | CDS VISIUMcore processor. | |
5046 | This configuration is intended for embedded systems. | |
5047 | ||
5048 | @html | |
5049 | <hr /> | |
5050 | @end html | |
5051 | @anchor{x-x-vxworks} | |
5052 | @heading *-*-vxworks* | |
5053 | Support for VxWorks is in flux. At present GCC supports @emph{only} the | |
5054 | very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@. | |
5055 | We welcome patches for other architectures supported by VxWorks 5.5. | |
5056 | Support for VxWorks AE would also be welcome; we believe this is merely | |
5057 | a matter of writing an appropriate ``configlette'' (see below). We are | |
5058 | not interested in supporting older, a.out or COFF-based, versions of | |
5059 | VxWorks in GCC 3. | |
5060 | ||
5061 | VxWorks comes with an older version of GCC installed in | |
5062 | @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it. | |
5063 | Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}. | |
5064 | Before running @command{configure}, create the directories @file{@var{prefix}} | |
5065 | and @file{@var{prefix}/bin}. Link or copy the appropriate assembler, | |
5066 | linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to | |
5067 | include that directory while running both @command{configure} and | |
5068 | @command{make}. | |
5069 | ||
5070 | You must give @command{configure} the | |
5071 | @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can | |
5072 | find the VxWorks system headers. Since VxWorks is a cross compilation | |
5073 | target only, you must also specify @option{--target=@var{target}}. | |
5074 | @command{configure} will attempt to create the directory | |
5075 | @file{@var{prefix}/@var{target}/sys-include} and copy files into it; | |
5076 | make sure the user running @command{configure} has sufficient privilege | |
5077 | to do so. | |
5078 | ||
5079 | GCC's exception handling runtime requires a special ``configlette'' | |
5080 | module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in | |
5081 | that file to add the module to your kernel build. (Future versions of | |
5082 | VxWorks will incorporate this module.) | |
5083 | ||
5084 | @html | |
5085 | <hr /> | |
5086 | @end html | |
5087 | @anchor{x86-64-x-x} | |
5088 | @heading x86_64-*-*, amd64-*-* | |
5089 | GCC supports the x86-64 architecture implemented by the AMD64 processor | |
5090 | (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@. | |
5091 | On GNU/Linux the default is a bi-arch compiler which is able to generate | |
5092 | both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch). | |
5093 | ||
5094 | @html | |
5095 | <hr /> | |
5096 | @end html | |
5097 | @anchor{x86-64-x-solaris2} | |
5098 | @heading x86_64-*-solaris2* | |
5099 | GCC also supports the x86-64 architecture implemented by the AMD64 | |
5100 | processor (@samp{amd64-*-*} is an alias for @samp{x86_64-*-*}) on | |
5101 | Solaris 10 or later. Unlike other systems, without special options a | |
5102 | bi-arch compiler is built which generates 32-bit code by default, but | |
5103 | can generate 64-bit x86-64 code with the @option{-m64} switch. Since | |
5104 | GCC 4.7, there is also a configuration that defaults to 64-bit code, but | |
5105 | can generate 32-bit code with @option{-m32}. To configure and build | |
5106 | this way, you have to provide all support libraries like @file{libgmp} | |
5107 | as 64-bit code, configure with @option{--target=x86_64-pc-solaris2.11} | |
5108 | and @samp{CC=gcc -m64}. | |
5109 | ||
5110 | @html | |
5111 | <hr /> | |
5112 | @end html | |
5113 | @anchor{xtensa-x-elf} | |
5114 | @heading xtensa*-*-elf | |
5115 | This target is intended for embedded Xtensa systems using the | |
5116 | @samp{newlib} C library. It uses ELF but does not support shared | |
5117 | objects. Designed-defined instructions specified via the | |
5118 | Tensilica Instruction Extension (TIE) language are only supported | |
5119 | through inline assembly. | |
5120 | ||
5121 | The Xtensa configuration information must be specified prior to | |
5122 | building GCC@. The @file{include/xtensa-config.h} header | |
5123 | file contains the configuration information. If you created your | |
5124 | own Xtensa configuration with the Xtensa Processor Generator, the | |
5125 | downloaded files include a customized copy of this header file, | |
5126 | which you can use to replace the default header file. | |
5127 | ||
5128 | @html | |
5129 | <hr /> | |
5130 | @end html | |
5131 | @anchor{xtensa-x-linux} | |
5132 | @heading xtensa*-*-linux* | |
5133 | This target is for Xtensa systems running GNU/Linux. It supports ELF | |
5134 | shared objects and the GNU C library (glibc). It also generates | |
5135 | position-independent code (PIC) regardless of whether the | |
5136 | @option{-fpic} or @option{-fPIC} options are used. In other | |
5137 | respects, this target is the same as the | |
5138 | @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target. | |
5139 | ||
5140 | @html | |
5141 | <hr /> | |
5142 | @end html | |
5143 | @anchor{windows} | |
5144 | @heading Microsoft Windows | |
5145 | ||
5146 | @subheading Intel 16-bit versions | |
5147 | The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not | |
5148 | supported. | |
5149 | ||
5150 | However, the 32-bit port has limited support for Microsoft | |
5151 | Windows 3.11 in the Win32s environment, as a target only. See below. | |
5152 | ||
5153 | @subheading Intel 32-bit versions | |
5154 | The 32-bit versions of Windows, including Windows 95, Windows NT, Windows | |
5155 | XP, and Windows Vista, are supported by several different target | |
5156 | platforms. These targets differ in which Windows subsystem they target | |
5157 | and which C libraries are used. | |
5158 | ||
5159 | @itemize | |
5160 | @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space | |
5161 | Linux API emulation layer in the Win32 subsystem. | |
5162 | @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for | |
5163 | the Win32 subsystem that provides a subset of POSIX. | |
5164 | @item MKS i386-pc-mks: NuTCracker from MKS. See | |
5165 | @uref{https://www.mkssoftware.com} for more information. | |
5166 | @end itemize | |
5167 | ||
5168 | @subheading Intel 64-bit versions | |
5169 | GCC contains support for x86-64 using the mingw-w64 | |
5170 | runtime library, available from @uref{https://www.mingw-w64.org/downloads/}. | |
5171 | This library should be used with the target triple x86_64-pc-mingw32. | |
5172 | ||
5173 | @subheading Windows CE | |
5174 | Windows CE is supported as a target only on Hitachi | |
5175 | SuperH (sh-wince-pe), and MIPS (mips-wince-pe). | |
5176 | ||
5177 | @subheading Other Windows Platforms | |
5178 | GCC no longer supports Windows NT on the Alpha or PowerPC. | |
5179 | ||
5180 | GCC no longer supports the Windows POSIX subsystem. However, it does | |
5181 | support the Interix subsystem. See above. | |
5182 | ||
5183 | Old target names including *-*-winnt and *-*-windowsnt are no longer used. | |
5184 | ||
5185 | PW32 (i386-pc-pw32) support was never completed, and the project seems to | |
5186 | be inactive. See @uref{http://pw32.sourceforge.net/} for more information. | |
5187 | ||
5188 | UWIN support has been removed due to a lack of maintenance. | |
5189 | ||
5190 | @html | |
5191 | <hr /> | |
5192 | @end html | |
5193 | @anchor{x-x-cygwin} | |
5194 | @heading *-*-cygwin | |
5195 | Ports of GCC are included with the | |
5196 | @uref{http://www.cygwin.com/,,Cygwin environment}. | |
5197 | ||
5198 | GCC will build under Cygwin without modification; it does not build | |
5199 | with Microsoft's C++ compiler and there are no plans to make it do so. | |
5200 | ||
5201 | The Cygwin native compiler can be configured to target any 32-bit x86 | |
5202 | cpu architecture desired; the default is i686-pc-cygwin. It should be | |
5203 | used with as up-to-date a version of binutils as possible; use either | |
5204 | the latest official GNU binutils release in the Cygwin distribution, | |
5205 | or version 2.20 or above if building your own. | |
5206 | ||
5207 | @html | |
5208 | <hr /> | |
5209 | @end html | |
5210 | @anchor{x-x-mingw32} | |
5211 | @heading *-*-mingw32 | |
5212 | GCC will build with and support only MinGW runtime 3.12 and later. | |
5213 | Earlier versions of headers are incompatible with the new default semantics | |
5214 | of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes. | |
5215 | ||
5216 | To support emitting DWARF debugging info you need to use GNU binutils | |
5217 | version 2.16 or above containing support for the @code{.secrel32} | |
5218 | assembler pseudo-op. | |
5219 | ||
5220 | @html | |
5221 | <hr /> | |
5222 | @end html | |
5223 | @anchor{older} | |
5224 | @heading Older systems | |
5225 | GCC contains support files for many older (1980s and early | |
5226 | 1990s) Unix variants. For the most part, support for these systems | |
5227 | has not been deliberately removed, but it has not been maintained for | |
5228 | several years and may suffer from bitrot. | |
5229 | ||
5230 | Starting with GCC 3.1, each release has a list of ``obsoleted'' systems. | |
5231 | Support for these systems is still present in that release, but | |
5232 | @command{configure} will fail unless the @option{--enable-obsolete} | |
5233 | option is given. Unless a maintainer steps forward, support for these | |
5234 | systems will be removed from the next release of GCC@. | |
5235 | ||
5236 | Support for old systems as hosts for GCC can cause problems if the | |
5237 | workarounds for compiler, library and operating system bugs affect the | |
5238 | cleanliness or maintainability of the rest of GCC@. In some cases, to | |
5239 | bring GCC up on such a system, if still possible with current GCC, may | |
5240 | require first installing an old version of GCC which did work on that | |
5241 | system, and using it to compile a more recent GCC, to avoid bugs in the | |
5242 | vendor compiler. Old releases of GCC 1 and GCC 2 are available in the | |
5243 | @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror | |
5244 | sites}. Header bugs may generally be avoided using | |
5245 | @command{fixincludes}, but bugs or deficiencies in libraries and the | |
5246 | operating system may still cause problems. | |
5247 | ||
5248 | Support for older systems as targets for cross-compilation is less | |
5249 | problematic than support for them as hosts for GCC; if an enthusiast | |
5250 | wishes to make such a target work again (including resurrecting any of | |
5251 | the targets that never worked with GCC 2, starting from the last | |
5252 | version before they were removed), patches | |
5253 | @uref{../contribute.html,,following the usual requirements} would be | |
5254 | likely to be accepted, since they should not affect the support for more | |
5255 | modern targets. | |
5256 | ||
5257 | For some systems, old versions of GNU binutils may also be useful, | |
5258 | and are available from @file{pub/binutils/old-releases} on | |
5259 | @uref{https://sourceware.org/mirrors.html,,sourceware.org mirror sites}. | |
5260 | ||
5261 | Some of the information on specific systems above relates to | |
5262 | such older systems, but much of the information | |
5263 | about GCC on such systems (which may no longer be applicable to | |
5264 | current GCC) is to be found in the GCC texinfo manual. | |
5265 | ||
5266 | @html | |
5267 | <hr /> | |
5268 | @end html | |
5269 | @anchor{elf} | |
5270 | @heading all ELF targets (SVR4, Solaris 2, etc.) | |
5271 | C++ support is significantly better on ELF targets if you use the | |
5272 | @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of | |
5273 | inlines, vtables and template instantiations will be discarded | |
5274 | automatically. | |
5275 | ||
5276 | ||
5277 | @html | |
5278 | <hr /> | |
5279 | <p> | |
5280 | @end html | |
5281 | @ifhtml | |
5282 | @uref{./index.html,,Return to the GCC Installation page} | |
5283 | @end ifhtml | |
5284 | @end ifset | |
5285 | ||
5286 | @c ***GFDL******************************************************************** | |
5287 | @ifset gfdlhtml | |
5288 | @include fdl.texi | |
5289 | @html | |
5290 | <hr /> | |
5291 | <p> | |
5292 | @end html | |
5293 | @ifhtml | |
5294 | @uref{./index.html,,Return to the GCC Installation page} | |
5295 | @end ifhtml | |
5296 | @end ifset | |
5297 | ||
5298 | @c *************************************************************************** | |
5299 | @c Part 6 The End of the Document | |
5300 | @ifinfo | |
5301 | @comment node-name, next, previous, up | |
5302 | @node Concept Index, , GNU Free Documentation License, Top | |
5303 | @end ifinfo | |
5304 | ||
5305 | @ifinfo | |
5306 | @unnumbered Concept Index | |
5307 | ||
5308 | @printindex cp | |
5309 | ||
5310 | @contents | |
5311 | @end ifinfo | |
5312 | @bye |