1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename gccinstall.info
9 @include gcc-common.texi
11 @c Specify title for specific html page
13 @settitle Installing GCC
16 @settitle Host/Target specific installation notes for GCC
18 @ifset prerequisiteshtml
19 @settitle Prerequisites for GCC
22 @settitle Downloading GCC
25 @settitle Installing GCC: Configuration
28 @settitle Installing GCC: Building
31 @settitle Installing GCC: Testing
33 @ifset finalinstallhtml
34 @settitle Installing GCC: Final installation
37 @settitle Installing GCC: Binaries
40 @settitle Installing GCC: GNU Free Documentation License
43 @c Copyright (C) 1988-2022 Free Software Foundation, Inc.
44 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
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.
49 @c Do not use @footnote{} in this file as it breaks install.texi2html!
51 @c Include everything if we're not making html
55 @set prerequisiteshtml
65 @c Part 2 Summary Description and Copyright
67 Copyright @copyright{} 1988-2022 Free Software Foundation, Inc.
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}''.
77 (a) The FSF's Front-Cover Text is:
81 (b) The FSF's Back-Cover Text is:
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.
90 @dircategory Software development
92 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
95 @c Part 3 Titlepage and Copyright
100 @c The following two commands start the copyright page.
102 @vskip 0pt plus 1filll
106 @c Part 4 Top node, Master Menu, and/or Table of Contents
109 @comment node-name, next, Previous, up
112 * Installing GCC:: This document describes the generic installation
113 procedure for GCC as well as detailing some target
114 specific installation instructions.
116 * Specific:: Host/target specific installation notes for GCC.
117 * Binaries:: Where to get pre-compiled binaries.
119 * GNU Free Documentation License:: How you can copy and share this manual.
120 * Concept Index:: This index has two entries.
128 @c Part 5 The Body of the Document
129 @c ***Installing GCC**********************************************************
131 @comment node-name, next, previous, up
132 @node Installing GCC, Binaries, , Top
136 @chapter Installing GCC
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.
144 This document describes the generic installation procedure for GCC as well
145 as detailing some target specific installation instructions.
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.
151 @emph{Before} starting the build/install procedure please check the
153 @ref{Specific, host/target specific installation notes}.
156 @uref{specific.html,,host/target specific installation notes}.
158 We recommend you browse the entire generic installation instructions before
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.
165 The installation procedure itself is broken into five steps.
170 * Downloading the source::
173 * Testing:: (optional)
180 @uref{prerequisites.html,,Prerequisites}
182 @uref{download.html,,Downloading the source}
184 @uref{configure.html,,Configuration}
186 @uref{build.html,,Building}
188 @uref{test.html,,Testing} (optional)
190 @uref{finalinstall.html,,Final install}
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.
206 @uref{./index.html,,Return to the GCC Installation page}
212 @c ***Prerequisites**************************************************
214 @comment node-name, next, previous, up
215 @node Prerequisites, Downloading the source, , Installing GCC
217 @ifset prerequisiteshtml
219 @chapter Prerequisites
221 @cindex Prerequisites
223 GCC requires that various tools and packages be available for use in the
224 build procedure. Modifying GCC sources requires additional tools
227 @heading Tools/packages necessary for building GCC
229 @item ISO C++11 compiler
230 Necessary to bootstrap GCC.
232 Versions of GCC prior to 11 also allow bootstrapping with an ISO C++98
233 compiler, versions of GCC prior to 4.8 also allow bootstrapping with a
234 ISO C89 compiler, and versions of GCC prior to 3.4 also allow
235 bootstrapping with a traditional (K&R) C compiler.
237 To build all languages in a cross-compiler or other configuration where
238 3-stage bootstrap is not performed, you need to start with an existing
239 GCC binary (version 4.8 or later) because source code for language
240 frontends other than C might use GCC extensions.
242 @item C standard library and headers
244 In order to build GCC, the C standard library and headers must be present
245 for all target variants for which target libraries will be built (and not
246 only the variant of the host C++ compiler).
248 This affects the popular @samp{x86_64-pc-linux-gnu} platform (among
249 other multilib targets), for which 64-bit (@samp{x86_64}) and 32-bit
250 (@samp{i386}) libc headers are usually packaged separately. If you do a
251 build of a native compiler on @samp{x86_64-pc-linux-gnu}, make sure you
252 either have the 32-bit libc developer package properly installed (the exact
253 name of the package depends on your distro) or you must build GCC as a
254 64-bit only compiler by configuring with the option
255 @option{--disable-multilib}. Otherwise, you may encounter an error such as
256 @samp{fatal error: gnu/stubs-32.h: No such file}
258 @item @anchor{GNAT-prerequisite}GNAT
260 In order to build GNAT, the Ada compiler, you need a working GNAT
261 compiler (GCC version 5.1 or later).
263 This includes GNAT tools such as @command{gnatmake} and
264 @command{gnatlink}, since the Ada front end is written in Ada and
265 uses some GNAT-specific extensions.
267 In order to build a cross compiler, it is strongly recommended to install
268 the new compiler as native first, and then use it to build the cross
269 compiler. Other native compiler versions may work but this is not guaranteed and
270 will typically fail with hard to understand compilation errors during the
273 Similarly, it is strongly recommended to use an older version of GNAT to build
274 GNAT. More recent versions of GNAT than the version built are not guaranteed
275 to work and will often fail during the build with compilation errors.
277 Note that @command{configure} does not test whether the GNAT installation works
278 and has a sufficiently recent version; if too old a GNAT version is
279 installed and @option{--enable-languages=ada} is used, the build will fail.
281 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
282 must not be set when building the Ada compiler, the Ada tools, or the
283 Ada runtime libraries. You can check that your build environment is clean
284 by verifying that @samp{gnatls -v} lists only one explicit path in each
287 @item @anchor{GDC-prerequisite}GDC
289 In order to build GDC, the D compiler, you need a working GDC
290 compiler (GCC version 9.1 or later) and D runtime library,
291 @samp{libphobos}, as the D front end is written in D.
293 Versions of GDC prior to 12 can be built with an ISO C++11 compiler, which can
294 then be installed and used to bootstrap newer versions of the D front end.
296 It is strongly recommended to use an older version of GDC to build GDC. More
297 recent versions of GDC than the version built are not guaranteed to work and
298 will often fail during the build with compilation errors relating to
299 deprecations or removed features.
301 Note that @command{configure} does not test whether the GDC installation works
302 and has a sufficiently recent version. Though the implementation of the D
303 front end does not make use of any GDC-specific extensions, or novel features
304 of the D language, if too old a GDC version is installed and
305 @option{--enable-languages=d} is used, the build will fail.
307 On some targets, @samp{libphobos} isn't enabled by default, but compiles
308 and works if @option{--enable-libphobos} is used. Specifics are
309 documented for affected targets.
311 @item A ``working'' POSIX compatible shell, or GNU bash
313 Necessary when running @command{configure} because some
314 @command{/bin/sh} shells have bugs and may crash when configuring the
315 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
316 have disastrous corner-case performance problems. This
317 can cause target @command{configure} runs to literally take days to
318 complete in some cases.
320 So on some platforms @command{/bin/ksh} is sufficient, on others it
321 isn't. See the host/target specific instructions for your platform, or
322 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
323 environment to your ``good'' shell prior to running
324 @command{configure}/@command{make}.
326 @command{zsh} is not a fully compliant POSIX shell and will not
327 work when configuring GCC@.
329 @item A POSIX or SVR4 awk
331 Necessary for creating some of the generated source files for GCC@.
332 If in doubt, use a recent GNU awk version, as some of the older ones
333 are broken. GNU awk version 3.1.5 is known to work.
337 Necessary in some circumstances, optional in others. See the
338 host/target specific instructions for your platform for the exact
341 Note binutils 2.35 or newer is required for LTO to work correctly
342 with GNU libtool that includes doing a bootstrap with LTO enabled.
344 @item gzip version 1.2.4 (or later) or
345 @itemx bzip2 version 1.0.2 (or later)
347 Necessary to uncompress GCC @command{tar} files when source code is
348 obtained via HTTPS mirror sites.
350 @item GNU make version 3.80 (or later)
352 You must have GNU make installed to build GCC@.
354 @item GNU tar version 1.14 (or later)
356 Necessary (only on some platforms) to untar the source code. Many
357 systems' @command{tar} programs will also work, only try GNU
358 @command{tar} if you have problems.
360 @item Perl version between 5.6.1 and 5.6.24
362 Necessary when targeting Darwin, building @samp{libstdc++},
363 and not using @option{--disable-symvers}.
364 Necessary when targeting Solaris 2 with Solaris @command{ld} and not using
365 @option{--disable-symvers}.
367 Necessary when regenerating @file{Makefile} dependencies in libiberty.
368 Necessary when regenerating @file{libiberty/functions.texi}.
369 Necessary when generating manpages from Texinfo manuals.
370 Used by various scripts to generate some files included in the source
371 repository (mainly Unicode-related and rarely changing) from source
374 Used by @command{automake}.
378 Several support libraries are necessary to build GCC, some are required,
379 others optional. While any sufficiently new version of required tools
380 usually work, library requirements are generally stricter. Newer
381 versions may work in some cases, but it's safer to use the exact
382 versions documented. We appreciate bug reports about problems with
383 newer versions, though. If your OS vendor provides packages for the
384 support libraries then using those packages may be the simplest way to
385 install the libraries.
388 @item GNU Multiple Precision Library (GMP) version 4.3.2 (or later)
390 Necessary to build GCC@. If a GMP source distribution is found in a
391 subdirectory of your GCC sources named @file{gmp}, it will be built
392 together with GCC. Alternatively, if GMP is already installed but it
393 is not in your library search path, you will have to configure with the
394 @option{--with-gmp} configure option. See also @option{--with-gmp-lib}
395 and @option{--with-gmp-include}.
396 The in-tree build is only supported with the GMP version that
397 download_prerequisites installs.
399 @item MPFR Library version 3.1.0 (or later)
401 Necessary to build GCC@. It can be downloaded from
402 @uref{https://www.mpfr.org}. If an MPFR source distribution is found
403 in a subdirectory of your GCC sources named @file{mpfr}, it will be
404 built together with GCC. Alternatively, if MPFR is already installed
405 but it is not in your default library search path, the
406 @option{--with-mpfr} configure option should be used. See also
407 @option{--with-mpfr-lib} and @option{--with-mpfr-include}.
408 The in-tree build is only supported with the MPFR version that
409 download_prerequisites installs.
411 @item MPC Library version 1.0.1 (or later)
413 Necessary to build GCC@. It can be downloaded from
414 @uref{https://www.multiprecision.org/mpc/}. If an MPC source distribution
415 is found in a subdirectory of your GCC sources named @file{mpc}, it
416 will be built together with GCC. Alternatively, if MPC is already
417 installed but it is not in your default library search path, the
418 @option{--with-mpc} configure option should be used. See also
419 @option{--with-mpc-lib} and @option{--with-mpc-include}.
420 The in-tree build is only supported with the MPC version that
421 download_prerequisites installs.
423 @item isl Library version 0.15 or later.
425 Necessary to build GCC with the Graphite loop optimizations.
426 It can be downloaded from @uref{https://gcc.gnu.org/pub/gcc/infrastructure/}.
427 If an isl source distribution is found
428 in a subdirectory of your GCC sources named @file{isl}, it will be
429 built together with GCC. Alternatively, the @option{--with-isl} configure
430 option should be used if isl is not installed in your default library
435 Necessary to build GCC with zstd compression used for LTO bytecode.
436 The library is searched in your default library patch search.
437 Alternatively, the @option{--with-zstd} configure option should be used.
441 @heading Tools/packages necessary for modifying GCC
443 @item autoconf version 2.69
444 @itemx GNU m4 version 1.4.6 (or later)
446 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
447 to regenerate @file{configure} and @file{config.in} files.
449 @item automake version 1.15.1
451 Necessary when modifying a @file{Makefile.am} file to regenerate its
452 associated @file{Makefile.in}.
454 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
455 file. Specifically this applies to the @file{gcc}, @file{intl},
456 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
457 as any of their subdirectories.
459 For directories that use automake, GCC requires the latest release in
460 the 1.15 series, which is currently 1.15.1. When regenerating a directory
461 to a newer version, please update all the directories using an older 1.15
462 to the latest released version.
464 @item gettext version 0.14.5 (or later)
466 Needed to regenerate @file{gcc.pot}.
468 @item gperf version 2.7.2 (or later)
470 Necessary when modifying @command{gperf} input files, e.g.@:
471 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
472 @file{gcc/cp/cfns.h}.
474 @item DejaGnu version 1.5.3 (or later)
477 @c Once Tcl 8.5 or higher is required, remove any obsolete
478 @c compatibility workarounds:
479 @c git grep 'compatibility with earlier Tcl releases'
481 Necessary to run the GCC testsuite; see the section on testing for
484 @item autogen version 5.5.4 (or later) and
485 @itemx guile version 1.4.1 (or later)
487 Necessary to regenerate @file{fixinc/fixincl.x} from
488 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
490 Necessary to run @samp{make check} for @file{fixinc}.
492 Necessary to regenerate the top level @file{Makefile.in} file from
493 @file{Makefile.tpl} and @file{Makefile.def}.
495 @item Flex version 2.5.4 (or later)
497 Necessary when modifying @file{*.l} files.
499 Necessary to build GCC during development because the generated output
500 files are not included in the version-controlled source repository.
501 They are included in releases.
503 @item Texinfo version 4.7 (or later)
505 Necessary for running @command{makeinfo} when modifying @file{*.texi}
506 files to test your changes.
508 Necessary for running @command{make dvi} or @command{make pdf} to
509 create printable documentation in DVI or PDF format. Texinfo version
510 4.8 or later is required for @command{make pdf}.
512 Necessary to build GCC documentation during development because the
513 generated output files are not included in the repository. They are
514 included in releases.
516 @item @TeX{} (any working version)
518 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
519 are used when running @command{make dvi} or @command{make pdf} to create
520 DVI or PDF files, respectively.
522 @item Sphinx version 1.0 (or later)
524 Necessary to regenerate @file{jit/docs/_build/texinfo} from the @file{.rst}
525 files in the directories below @file{jit/docs}.
527 @item git (any version)
528 @itemx SSH (any version)
530 Necessary to access the source repository. Public releases and weekly
531 snapshots of the development sources are also available via HTTPS@.
533 @item GNU diffutils version 2.7 (or later)
535 Useful when submitting patches for the GCC source code.
537 @item patch version 2.5.4 (or later)
539 Necessary when applying patches, created with @command{diff}, to one's
549 @uref{./index.html,,Return to the GCC Installation page}
553 @c ***Downloading the source**************************************************
555 @comment node-name, next, previous, up
556 @node Downloading the source, Configuration, Prerequisites, Installing GCC
560 @chapter Downloading GCC
562 @cindex Downloading GCC
563 @cindex Downloading the Source
565 GCC is distributed via @uref{https://gcc.gnu.org/git.html,,git} and via
566 HTTPS as tarballs compressed with @command{gzip} or @command{bzip2}.
568 Please refer to the @uref{https://gcc.gnu.org/releases.html,,releases web page}
569 for information on how to obtain GCC@.
571 The source distribution includes the C, C++, Objective-C, Fortran,
572 and Ada (in the case of GCC 3.1 and later) compilers, as well as
573 runtime libraries for C++, Objective-C, and Fortran.
574 For previous versions these were downloadable as separate components such
575 as the core GCC distribution, which included the C language front end and
576 shared components, and language-specific distributions including the
577 language front end and the language runtime (where appropriate).
579 If you also intend to build binutils (either to upgrade an existing
580 installation or for use in place of the corresponding tools of your
581 OS), unpack the binutils distribution either in the same directory or
582 a separate one. In the latter case, add symbolic links to any
583 components of the binutils you intend to build alongside the compiler
584 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
585 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
587 Likewise the GMP, MPFR and MPC libraries can be automatically built
588 together with GCC. You may simply run the
589 @command{contrib/download_prerequisites} script in the GCC source directory
590 to set up everything.
591 Otherwise unpack the GMP, MPFR and/or MPC source
592 distributions in the directory containing the GCC sources and rename
593 their directories to @file{gmp}, @file{mpfr} and @file{mpc},
594 respectively (or use symbolic links with the same name).
601 @uref{./index.html,,Return to the GCC Installation page}
605 @c ***Configuration***********************************************************
607 @comment node-name, next, previous, up
608 @node Configuration, Building, Downloading the source, Installing GCC
612 @chapter Installing GCC: Configuration
614 @cindex Configuration
615 @cindex Installing GCC: Configuration
617 Like most GNU software, GCC must be configured before it can be built.
618 This document describes the recommended configuration procedure
619 for both native and cross targets.
621 We use @var{srcdir} to refer to the toplevel source directory for
622 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
624 If you obtained the sources by cloning the repository, @var{srcdir}
625 must refer to the top @file{gcc} directory, the one where the
626 @file{MAINTAINERS} file can be found, and not its @file{gcc}
627 subdirectory, otherwise the build will fail.
629 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
630 file system, the shell's built-in @command{pwd} command will return
631 temporary pathnames. Using these can lead to various sorts of build
632 problems. To avoid this issue, set the @env{PWDCMD} environment
633 variable to an automounter-aware @command{pwd} command, e.g.,
634 @command{pawd} or @samp{amq -w}, during the configuration and build
637 First, we @strong{highly} recommend that GCC be built into a
638 separate directory from the sources which does @strong{not} reside
639 within the source tree. This is how we generally build GCC; building
640 where @var{srcdir} == @var{objdir} should still work, but doesn't
641 get extensive testing; building where @var{objdir} is a subdirectory
642 of @var{srcdir} is unsupported.
644 If you have previously built GCC in the same directory for a
645 different target machine, do @samp{make distclean} to delete all files
646 that might be invalid. One of the files this deletes is @file{Makefile};
647 if @samp{make distclean} complains that @file{Makefile} does not exist
648 or issues a message like ``don't know how to make distclean'' it probably
649 means that the directory is already suitably clean. However, with the
650 recommended method of building in a separate @var{objdir}, you should
651 simply use a different @var{objdir} for each target.
653 Second, when configuring a native system, either @command{cc} or
654 @command{gcc} must be in your path or you must set @env{CC} in
655 your environment before running configure. Otherwise the configuration
659 Note that the bootstrap compiler and the resulting GCC must be link
660 compatible, else the bootstrap will fail with linker errors about
661 incompatible object file formats. Several multilibed targets are
662 affected by this requirement, see
664 @ref{Specific, host/target specific installation notes}.
667 @uref{specific.html,,host/target specific installation notes}.
676 % @var{srcdir}/configure [@var{options}] [@var{target}]
679 @heading Distributor options
681 If you will be distributing binary versions of GCC, with modifications
682 to the source code, you should use the options described in this
683 section to make clear that your version contains modifications.
686 @item --with-pkgversion=@var{version}
687 Specify a string that identifies your package. You may wish
688 to include a build number or build date. This version string will be
689 included in the output of @command{gcc --version}. This suffix does
690 not replace the default version string, only the @samp{GCC} part.
692 The default value is @samp{GCC}.
694 @item --with-bugurl=@var{url}
695 Specify the URL that users should visit if they wish to report a bug.
696 You are of course welcome to forward bugs reported to you to the FSF,
697 if you determine that they are not bugs in your modifications.
699 The default value refers to the FSF's GCC bug tracker.
701 @item --with-documentation-root-url=@var{url}
702 Specify the URL root that contains GCC option documentation. The @var{url}
703 should end with a @code{/} character.
705 The default value is @uref{https://gcc.gnu.org/onlinedocs/,,https://gcc.gnu.org/onlinedocs/}.
707 @item --with-changes-root-url=@var{url}
708 Specify the URL root that contains information about changes in GCC
709 releases like @code{gcc-@var{version}/changes.html}.
710 The @var{url} should end with a @code{/} character.
712 The default value is @uref{https://gcc.gnu.org/,,https://gcc.gnu.org/}.
716 @heading Host, Build and Target specification
718 Specify the host, build and target machine configurations. You do this
719 when you run the @file{configure} script.
721 The @dfn{build} machine is the system which you are using, the
722 @dfn{host} machine is the system where you want to run the resulting
723 compiler (normally the build machine), and the @dfn{target} machine is
724 the system for which you want the compiler to generate code.
726 If you are building a compiler to produce code for the machine it runs
727 on (a native compiler), you normally do not need to specify any operands
728 to @file{configure}; it will try to guess the type of machine you are on
729 and use that as the build, host and target machines. So you don't need
730 to specify a configuration when building a native compiler unless
731 @file{configure} cannot figure out what your configuration is or guesses
734 In those cases, specify the build machine's @dfn{configuration name}
735 with the @option{--host} option; the host and target will default to be
736 the same as the host machine.
741 ./configure --host=x86_64-pc-linux-gnu
744 A configuration name may be canonical or it may be more or less
745 abbreviated (@file{config.sub} script produces canonical versions).
747 A canonical configuration name has three parts, separated by dashes.
748 It looks like this: @samp{@var{cpu}-@var{company}-@var{system}}.
750 Here are the possible CPU types:
753 aarch64, aarch64_be, alpha, alpha64, amdgcn, arc, arceb, arm, armeb, avr, bfin,
754 bpf, cris, csky, epiphany, fido, fr30, frv, ft32, h8300, hppa, hppa2.0,
755 hppa64, i486, i686, ia64, iq2000, lm32, loongarch64, m32c, m32r, m32rle, m68k,
756 mcore, microblaze, microblazeel, mips, mips64, mips64el, mips64octeon,
757 mips64orion, mips64vr, mipsel, mipsisa32, mipsisa32r2, mipsisa64, mipsisa64r2,
758 mipsisa64r2el, mipsisa64sb1, mipsisa64sr71k, mipstx39, mmix, mn10300, moxie,
759 msp430, nds32be, nds32le, nios2, nvptx, or1k, pdp11, powerpc, powerpc64,
760 powerpc64le, powerpcle, pru, riscv32, riscv32be, riscv64, riscv64be, rl78, rx,
761 s390, s390x, sh, shle, sparc, sparc64, tic6x, v850,
762 v850e, v850e1, vax, visium, x86_64, xstormy16, xtensa
765 Here is a list of system types:
768 aix@var{version}, amdhsa, aout, cygwin, darwin@var{version},
769 eabi, eabialtivec, eabisim, eabisimaltivec, elf, elf32,
770 elfbare, elfoabi, freebsd@var{version}, gnu, hpux, hpux@var{version},
771 kfreebsd-gnu, kopensolaris-gnu, linux-androideabi, linux-gnu,
772 linux-gnu_altivec, linux-musl, linux-uclibc, lynxos, mingw32, mingw32crt,
773 mmixware, msdosdjgpp, netbsd, netbsdelf@var{version}, nto-qnx, openbsd,
774 rtems, solaris@var{version}, symbianelf, tpf, uclinux, uclinux_eabi, vms,
775 vxworks, vxworksae, vxworksmils
778 @heading Options specification
780 Use @var{options} to override several configure time options for
781 GCC@. A list of supported @var{options} follows; @samp{configure
782 --help} may list other options, but those not listed below may not
783 work and should not normally be used.
785 Note that each @option{--enable} option has a corresponding
786 @option{--disable} option and that each @option{--with} option has a
787 corresponding @option{--without} option.
790 @item --prefix=@var{dirname}
791 Specify the toplevel installation
792 directory. This is the recommended way to install the tools into a directory
793 other than the default. The toplevel installation directory defaults to
796 We @strong{highly} recommend against @var{dirname} being the same or a
797 subdirectory of @var{objdir} or vice versa. If specifying a directory
798 beneath a user's home directory tree, some shells will not expand
799 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
802 The following standard @command{autoconf} options are supported. Normally you
803 should not need to use these options.
805 @item --exec-prefix=@var{dirname}
806 Specify the toplevel installation directory for architecture-dependent
807 files. The default is @file{@var{prefix}}.
809 @item --bindir=@var{dirname}
810 Specify the installation directory for the executables called by users
811 (such as @command{gcc} and @command{g++}). The default is
812 @file{@var{exec-prefix}/bin}.
814 @item --libdir=@var{dirname}
815 Specify the installation directory for object code libraries and
816 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
818 @item --libexecdir=@var{dirname}
819 Specify the installation directory for internal executables of GCC@.
820 The default is @file{@var{exec-prefix}/libexec}.
822 @item --with-slibdir=@var{dirname}
823 Specify the installation directory for the shared libgcc library. The
824 default is @file{@var{libdir}}.
826 @item --datarootdir=@var{dirname}
827 Specify the root of the directory tree for read-only architecture-independent
828 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
830 @item --infodir=@var{dirname}
831 Specify the installation directory for documentation in info format.
832 The default is @file{@var{datarootdir}/info}.
834 @item --datadir=@var{dirname}
835 Specify the installation directory for some architecture-independent
836 data files referenced by GCC@. The default is @file{@var{datarootdir}}.
838 @item --docdir=@var{dirname}
839 Specify the installation directory for documentation files (other
840 than Info) for GCC@. The default is @file{@var{datarootdir}/doc}.
842 @item --htmldir=@var{dirname}
843 Specify the installation directory for HTML documentation files.
844 The default is @file{@var{docdir}}.
846 @item --pdfdir=@var{dirname}
847 Specify the installation directory for PDF documentation files.
848 The default is @file{@var{docdir}}.
850 @item --mandir=@var{dirname}
851 Specify the installation directory for manual pages. The default is
852 @file{@var{datarootdir}/man}. (Note that the manual pages are only extracts
853 from the full GCC manuals, which are provided in Texinfo format. The manpages
854 are derived by an automatic conversion process from parts of the full
857 @item --with-gxx-include-dir=@var{dirname}
859 the installation directory for G++ header files. The default depends
860 on other configuration options, and differs between cross and native
863 @item --with-specs=@var{specs}
864 Specify additional command line driver SPECS.
865 This can be useful if you need to turn on a non-standard feature by
866 default without modifying the compiler's source code, for instance
867 @option{--with-specs=%@{!fcommon:%@{!fno-common:-fno-common@}@}}.
869 @xref{Spec Files,, Specifying subprocesses and the switches to pass to them,
870 gcc, Using the GNU Compiler Collection (GCC)},
873 See ``Spec Files'' in the main manual
878 @item --program-prefix=@var{prefix}
879 GCC supports some transformations of the names of its programs when
880 installing them. This option prepends @var{prefix} to the names of
881 programs to install in @var{bindir} (see above). For example, specifying
882 @option{--program-prefix=foo-} would result in @samp{gcc}
883 being installed as @file{/usr/local/bin/foo-gcc}.
885 @item --program-suffix=@var{suffix}
886 Appends @var{suffix} to the names of programs to install in @var{bindir}
887 (see above). For example, specifying @option{--program-suffix=-3.1}
888 would result in @samp{gcc} being installed as
889 @file{/usr/local/bin/gcc-3.1}.
891 @item --program-transform-name=@var{pattern}
892 Applies the @samp{sed} script @var{pattern} to be applied to the names
893 of programs to install in @var{bindir} (see above). @var{pattern} has to
894 consist of one or more basic @samp{sed} editing commands, separated by
895 semicolons. For example, if you want the @samp{gcc} program name to be
896 transformed to the installed program @file{/usr/local/bin/myowngcc} and
897 the @samp{g++} program name to be transformed to
898 @file{/usr/local/bin/gspecial++} without changing other program names,
899 you could use the pattern
900 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
901 to achieve this effect.
903 All three options can be combined and used together, resulting in more
904 complex conversion patterns. As a basic rule, @var{prefix} (and
905 @var{suffix}) are prepended (appended) before further transformations
906 can happen with a special transformation script @var{pattern}.
908 As currently implemented, this option only takes effect for native
909 builds; cross compiler binaries' names are not transformed even when a
910 transformation is explicitly asked for by one of these options.
912 For native builds, some of the installed programs are also installed
913 with the target alias in front of their name, as in
914 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
915 before the target alias is prepended to the name---so, specifying
916 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
917 resulting binary would be installed as
918 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
920 As a last shortcoming, none of the installed Ada programs are
921 transformed yet, which will be fixed in some time.
923 @item --with-local-prefix=@var{dirname}
925 installation directory for local include files. The default is
926 @file{/usr/local}. Specify this option if you want the compiler to
927 search directory @file{@var{dirname}/include} for locally installed
928 header files @emph{instead} of @file{/usr/local/include}.
930 You should specify @option{--with-local-prefix} @strong{only} if your
931 site has a different convention (not @file{/usr/local}) for where to put
934 The default value for @option{--with-local-prefix} is @file{/usr/local}
935 regardless of the value of @option{--prefix}. Specifying
936 @option{--prefix} has no effect on which directory GCC searches for
937 local header files. This may seem counterintuitive, but actually it is
940 The purpose of @option{--prefix} is to specify where to @emph{install
941 GCC}. The local header files in @file{/usr/local/include}---if you put
942 any in that directory---are not part of GCC@. They are part of other
943 programs---perhaps many others. (GCC installs its own header files in
944 another directory which is based on the @option{--prefix} value.)
946 Both the local-prefix include directory and the GCC-prefix include
947 directory are part of GCC's ``system include'' directories. Although these
948 two directories are not fixed, they need to be searched in the proper
949 order for the correct processing of the include_next directive. The
950 local-prefix include directory is searched before the GCC-prefix
951 include directory. Another characteristic of system include directories
952 is that pedantic warnings are turned off for headers in these directories.
954 Some autoconf macros add @option{-I @var{directory}} options to the
955 compiler command line, to ensure that directories containing installed
956 packages' headers are searched. When @var{directory} is one of GCC's
957 system include directories, GCC will ignore the option so that system
958 directories continue to be processed in the correct order. This
959 may result in a search order different from what was specified but the
960 directory will still be searched.
962 GCC automatically searches for ordinary libraries using
963 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
964 used for both GCC and packages, GCC will automatically search for
965 both headers and libraries. This provides a configuration that is
966 easy to use. GCC behaves in a manner similar to that when it is
967 installed as a system compiler in @file{/usr}.
969 Sites that need to install multiple versions of GCC may not want to
970 use the above simple configuration. It is possible to use the
971 @option{--program-prefix}, @option{--program-suffix} and
972 @option{--program-transform-name} options to install multiple versions
973 into a single directory, but it may be simpler to use different prefixes
974 and the @option{--with-local-prefix} option to specify the location of the
975 site-specific files for each version. It will then be necessary for
976 users to specify explicitly the location of local site libraries
977 (e.g., with @env{LIBRARY_PATH}).
979 The same value can be used for both @option{--with-local-prefix} and
980 @option{--prefix} provided it is not @file{/usr}. This can be used
981 to avoid the default search of @file{/usr/local/include}.
983 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
984 The directory you use for @option{--with-local-prefix} @strong{must not}
985 contain any of the system's standard header files. If it did contain
986 them, certain programs would be miscompiled (including GNU Emacs, on
987 certain targets), because this would override and nullify the header
988 file corrections made by the @command{fixincludes} script.
990 Indications are that people who use this option use it based on mistaken
991 ideas of what it is for. People use it as if it specified where to
992 install part of GCC@. Perhaps they make this assumption because
993 installing GCC creates the directory.
995 @item --with-gcc-major-version-only
996 Specifies that GCC should use only the major number rather than
997 @var{major}.@var{minor}.@var{patchlevel} in filesystem paths.
999 @item --with-native-system-header-dir=@var{dirname}
1000 Specifies that @var{dirname} is the directory that contains native system
1001 header files, rather than @file{/usr/include}. This option is most useful
1002 if you are creating a compiler that should be isolated from the system
1003 as much as possible. It is most commonly used with the
1004 @option{--with-sysroot} option and will cause GCC to search
1005 @var{dirname} inside the system root specified by that option.
1007 @item --enable-shared[=@var{package}[,@dots{}]]
1008 Build shared versions of libraries, if shared libraries are supported on
1009 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
1010 are enabled by default on all platforms that support shared libraries.
1012 If a list of packages is given as an argument, build shared libraries
1013 only for the listed packages. For other packages, only static libraries
1014 will be built. Package names currently recognized in the GCC tree are
1015 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
1016 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
1017 @samp{ada}, @samp{libada}, @samp{libgo}, @samp{libobjc}, and @samp{libphobos}.
1018 Note @samp{libiberty} does not support shared libraries at all.
1020 Use @option{--disable-shared} to build only static libraries. Note that
1021 @option{--disable-shared} does not accept a list of package names as
1022 argument, only @option{--enable-shared} does.
1024 Contrast with @option{--enable-host-shared}, which affects @emph{host}
1027 @item --enable-host-shared
1028 Specify that the @emph{host} code should be built into position-independent
1029 machine code (with @option{-fPIC}), allowing it to be used within shared
1030 libraries, but yielding a slightly slower compiler.
1032 This option is required when building the libgccjit.so library.
1034 Contrast with @option{--enable-shared}, which affects @emph{target}
1037 @item --enable-host-pie
1038 Specify that the @emph{host} executables should be built into
1039 position-independent executables (with @option{-fPIE} and @option{-pie}),
1040 yielding a slightly slower compiler (but faster than
1041 @option{--enable-host-shared}). Position-independent executables are loaded
1042 at random addresses each time they are executed, therefore provide additional
1043 protection against Return Oriented Programming (ROP) attacks.
1045 @option{--enable-host-pie}) may be used with @option{--enable-host-shared}),
1046 in which case @option{-fPIC} is used when compiling, and @option{-pie} when
1049 @item @anchor{with-gnu-as}--with-gnu-as
1050 Specify that the compiler should assume that the
1051 assembler it finds is the GNU assembler. However, this does not modify
1052 the rules to find an assembler and will result in confusion if the
1053 assembler found is not actually the GNU assembler. (Confusion may also
1054 result if the compiler finds the GNU assembler but has not been
1055 configured with @option{--with-gnu-as}.) If you have more than one
1056 assembler installed on your system, you may want to use this option in
1057 connection with @option{--with-as=@var{pathname}} or
1058 @option{--with-build-time-tools=@var{pathname}}.
1060 The following systems are the only ones where it makes a difference
1061 whether you use the GNU assembler. On any other system,
1062 @option{--with-gnu-as} has no effect.
1065 @item @samp{hppa1.0-@var{any}-@var{any}}
1066 @item @samp{hppa1.1-@var{any}-@var{any}}
1067 @item @samp{sparc-sun-solaris2.@var{any}}
1068 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
1071 @item @anchor{with-as}--with-as=@var{pathname}
1072 Specify that the compiler should use the assembler pointed to by
1073 @var{pathname}, rather than the one found by the standard rules to find
1074 an assembler, which are:
1077 Unless GCC is being built with a cross compiler, check the
1078 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
1079 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
1080 @var{exec-prefix} defaults to @var{prefix}, which
1081 defaults to @file{/usr/local} unless overridden by the
1082 @option{--prefix=@var{pathname}} switch described above. @var{target}
1083 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
1084 @var{version} denotes the GCC version, such as 3.0.
1087 If the target system is the same that you are building on, check
1088 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
1092 Check in the @env{PATH} for a tool whose name is prefixed by the
1093 target system triple.
1096 Check in the @env{PATH} for a tool whose name is not prefixed by the
1097 target system triple, if the host and target system triple are
1098 the same (in other words, we use a host tool if it can be used for
1099 the target as well).
1102 You may want to use @option{--with-as} if no assembler
1103 is installed in the directories listed above, or if you have multiple
1104 assemblers installed and want to choose one that is not found by the
1107 @item @anchor{with-gnu-ld}--with-gnu-ld
1108 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
1111 @item --with-ld=@var{pathname}
1112 Same as @uref{#with-as,,@option{--with-as}}
1115 @item --with-dsymutil=@var{pathname}
1116 Same as @uref{#with-as,,@option{--with-as}}
1117 but for the debug linker (only used on Darwin platforms so far).
1119 @item --with-tls=@var{dialect}
1120 Specify the default TLS dialect, for systems were there is a choice.
1121 For ARM targets, possible values for @var{dialect} are @code{gnu} or
1122 @code{gnu2}, which select between the original GNU dialect and the GNU TLS
1123 descriptor-based dialect.
1125 @item --enable-multiarch
1126 Specify whether to enable or disable multiarch support. The default is
1127 to check for glibc start files in a multiarch location, and enable it
1128 if the files are found. The auto detection is enabled for native builds,
1129 and for cross builds configured with @option{--with-sysroot}, and without
1130 @option{--with-native-system-header-dir}.
1131 More documentation about multiarch can be found at
1132 @uref{https://wiki.debian.org/Multiarch}.
1134 @item --enable-sjlj-exceptions
1135 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1136 @samp{configure} ordinarily picks the correct value based on the platform.
1137 Only use this option if you are sure you need a different setting.
1139 @item --enable-vtable-verify
1140 Specify whether to enable or disable the vtable verification feature.
1141 Enabling this feature causes libstdc++ to be built with its virtual calls
1142 in verifiable mode. This means that, when linked with libvtv, every
1143 virtual call in libstdc++ will verify the vtable pointer through which the
1144 call will be made before actually making the call. If not linked with libvtv,
1145 the verifier will call stub functions (in libstdc++ itself) and do nothing.
1146 If vtable verification is disabled, then libstdc++ is not built with its
1147 virtual calls in verifiable mode at all. However the libvtv library will
1148 still be built (see @option{--disable-libvtv} to turn off building libvtv).
1149 @option{--disable-vtable-verify} is the default.
1151 @item --disable-gcov
1152 Specify that the run-time library used for coverage analysis
1153 and associated host tools should not be built.
1155 @item --disable-multilib
1156 Specify that multiple target
1157 libraries to support different target variants, calling
1158 conventions, etc.@: should not be built. The default is to build a
1159 predefined set of them.
1161 Some targets provide finer-grained control over which multilibs are built
1162 (e.g., @option{--disable-softfloat}):
1165 fpu, 26bit, underscore, interwork, biendian, nofmult.
1168 softfloat, m68881, m68000, m68020.
1171 single-float, biendian, softfloat.
1176 @item powerpc*-*-*, rs6000*-*-*
1177 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
1182 @item --with-multilib-list=@var{list}
1183 @itemx --without-multilib-list
1184 Specify what multilibs to build. @var{list} is a comma separated list of
1185 values, possibly consisting of a single value. Currently only implemented
1186 for aarch64*-*-*, arm*-*-*, loongarch64-*-*, riscv*-*-*, sh*-*-* and
1187 x86-64-*-linux*. The accepted values and meaning for each target is given
1192 @var{list} is a comma separated list of @code{ilp32}, and @code{lp64}
1193 to enable ILP32 and LP64 run-time libraries, respectively. If
1194 @var{list} is empty, then there will be no multilibs and only the
1195 default run-time library will be built. If @var{list} is
1196 @code{default} or --with-multilib-list= is not specified, then the
1197 default set of libraries is selected based on the value of
1201 @var{list} is a comma separated list of @code{aprofile} and
1202 @code{rmprofile} to build multilibs for A or R and M architecture
1203 profiles respectively. Note that, due to some limitation of the current
1204 multilib framework, using the combined @code{aprofile,rmprofile}
1205 multilibs selects in some cases a less optimal multilib than when using
1206 the multilib profile for the architecture targetted. The special value
1207 @code{default} is also accepted and is equivalent to omitting the
1208 option, i.e., only the default run-time library will be enabled.
1210 @var{list} may instead contain @code{@@name}, to use the multilib
1211 configuration Makefile fragment @file{name} in @file{gcc/config/arm} in
1212 the source tree (it is part of the corresponding sources, after all).
1213 It is recommended, but not required, that files used for this purpose to
1214 be named starting with @file{t-ml-}, to make their intended purpose
1215 self-evident, in line with GCC conventions. Such files enable custom,
1216 user-chosen multilib lists to be configured. Whether multiple such
1217 files can be used together depends on the contents of the supplied
1218 files. See @file{gcc/config/arm/t-multilib} and its supplementary
1219 @file{gcc/config/arm/t-*profile} files for an example of what such
1220 Makefile fragments might look like for this version of GCC. The macros
1221 expected to be defined in these fragments are not stable across GCC
1222 releases, so make sure they define the @code{MULTILIB}-related macros
1223 expected by the version of GCC you are building.
1225 @xref{Target Fragment,, Target Makefile Fragments, gccint, GNU Compiler
1226 Collection (GCC) Internals}.
1229 See ``Target Makefile Fragments'' in the internals manual.
1232 The table below gives the combination of ISAs, architectures, FPUs and
1233 floating-point ABIs for which multilibs are built for each predefined
1234 profile. The union of these options is considered when specifying both
1235 @code{aprofile} and @code{rmprofile}.
1237 @multitable @columnfractions .15 .28 .30
1238 @item Option @tab aprofile @tab rmprofile
1240 @tab @code{-marm} and @code{-mthumb}
1242 @item Architectures@*@*@*@*@*@*
1243 @tab default architecture@*
1244 @code{-march=armv7-a}@*
1245 @code{-march=armv7ve}@*
1246 @code{-march=armv8-a}@*@*@*
1247 @tab default architecture@*
1248 @code{-march=armv6s-m}@*
1249 @code{-march=armv7-m}@*
1250 @code{-march=armv7e-m}@*
1251 @code{-march=armv8-m.base}@*
1252 @code{-march=armv8-m.main}@*
1254 @item FPUs@*@*@*@*@*
1256 @code{-mfpu=vfpv3-d16}@*
1258 @code{-mfpu=vfpv4-d16}@*
1259 @code{-mfpu=neon-vfpv4}@*
1260 @code{-mfpu=neon-fp-armv8}
1262 @code{-mfpu=vfpv3-d16}@*
1263 @code{-mfpu=fpv4-sp-d16}@*
1264 @code{-mfpu=fpv5-sp-d16}@*
1265 @code{-mfpu=fpv5-d16}@*
1266 @item floating-point@/ ABIs@*@*
1267 @tab @code{-mfloat-abi=soft}@*
1268 @code{-mfloat-abi=softfp}@*
1269 @code{-mfloat-abi=hard}
1270 @tab @code{-mfloat-abi=soft}@*
1271 @code{-mfloat-abi=softfp}@*
1272 @code{-mfloat-abi=hard}
1275 @item loongarch*-*-*
1276 @var{list} is a comma-separated list of the following ABI identifiers:
1277 @code{lp64d[/base]} @code{lp64f[/base]} @code{lp64d[/base]}, where the
1278 @code{/base} suffix may be omitted, to enable their respective run-time
1279 libraries. If @var{list} is empty or @code{default},
1280 or if @option{--with-multilib-list} is not specified, then the default ABI
1281 as specified by @option{--with-abi} or implied by @option{--target} is selected.
1284 @var{list} is a single ABI name. The target architecture must be either
1285 @code{rv32gc} or @code{rv64gc}. This will build a single multilib for the
1286 specified architecture and ABI pair. If @code{--with-multilib-list} is not
1287 given, then a default set of multilibs is selected based on the value of
1288 @option{--target}. This is usually a large set of multilibs.
1291 @var{list} is a comma separated list of CPU names. These must be of the
1292 form @code{sh*} or @code{m*} (in which case they match the compiler option
1293 for that processor). The list should not contain any endian options -
1294 these are handled by @option{--with-endian}.
1296 If @var{list} is empty, then there will be no multilibs for extra
1297 processors. The multilib for the secondary endian remains enabled.
1299 As a special case, if an entry in the list starts with a @code{!}
1300 (exclamation point), then it is added to the list of excluded multilibs.
1301 Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES}
1302 (once the leading @code{!} has been stripped).
1304 If @option{--with-multilib-list} is not given, then a default set of
1305 multilibs is selected based on the value of @option{--target}. This is
1306 usually the complete set of libraries, but some targets imply a more
1309 Example 1: to configure a compiler for SH4A only, but supporting both
1310 endians, with little endian being the default:
1312 --with-cpu=sh4a --with-endian=little,big --with-multilib-list=
1315 Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with
1316 only little endian SH4AL:
1318 --with-cpu=sh4a --with-endian=little,big \
1319 --with-multilib-list=sh4al,!mb/m4al
1322 @item x86-64-*-linux*
1323 @var{list} is a comma separated list of @code{m32}, @code{m64} and
1324 @code{mx32} to enable 32-bit, 64-bit and x32 run-time libraries,
1325 respectively. If @var{list} is empty, then there will be no multilibs
1326 and only the default run-time library will be enabled.
1328 If @option{--with-multilib-list} is not given, then only 32-bit and
1329 64-bit run-time libraries will be enabled.
1332 @item --with-multilib-generator=@var{config}
1333 Specify what multilibs to build. @var{config} is a semicolon separated list of
1334 values, possibly consisting of a single value. Currently only implemented
1335 for riscv*-*-elf*. The accepted values and meanings are given below.
1338 Every config is constructed with four components: architecture string, ABI,
1339 reuse rule with architecture string and reuse rule with sub-extension.
1341 Example 1: Add multi-lib suppport for rv32i with ilp32.
1346 Example 2: Add multi-lib suppport for rv32i with ilp32 and rv32imafd with ilp32.
1348 rv32i-ilp32--;rv32imafd-ilp32--
1351 Example 3: Add multi-lib suppport for rv32i with ilp32; rv32im with ilp32 and
1352 rv32ic with ilp32 will reuse this multi-lib set.
1354 rv32i-ilp32-rv32im-c
1357 Example 4: Add multi-lib suppport for rv64ima with lp64; rv64imaf with lp64,
1358 rv64imac with lp64 and rv64imafc with lp64 will reuse this multi-lib set.
1360 rv64ima-lp64--f,c,fc
1363 @option{--with-multilib-generator} have an optional configuration argument
1364 @option{--cmodel=val} for code model, this option will expand with other
1365 config options, @var{val} is a comma separated list of possible code model,
1366 currently we support medlow and medany.
1368 Example 5: Add multi-lib suppport for rv64ima with lp64; rv64ima with lp64 and
1371 rv64ima-lp64--;--cmodel=medlow
1374 Example 6: Add multi-lib suppport for rv64ima with lp64; rv64ima with lp64 and
1375 medlow code model; rv64ima with lp64 and medany code model
1377 rv64ima-lp64--;--cmodel=medlow,medany
1380 @item --with-endian=@var{endians}
1381 Specify what endians to use.
1382 Currently only implemented for sh*-*-*.
1384 @var{endians} may be one of the following:
1387 Use big endian exclusively.
1389 Use little endian exclusively.
1391 Use big endian by default. Provide a multilib for little endian.
1393 Use little endian by default. Provide a multilib for big endian.
1396 @item --enable-threads
1397 Specify that the target
1398 supports threads. This affects the Objective-C compiler and runtime
1399 library, and exception handling for other languages like C++.
1400 On some systems, this is the default.
1402 In general, the best (and, in many cases, the only known) threading
1403 model available will be configured for use. Beware that on some
1404 systems, GCC has not been taught what threading models are generally
1405 available for the system. In this case, @option{--enable-threads} is an
1406 alias for @option{--enable-threads=single}.
1408 @item --disable-threads
1409 Specify that threading support should be disabled for the system.
1410 This is an alias for @option{--enable-threads=single}.
1412 @item --enable-threads=@var{lib}
1414 @var{lib} is the thread support library. This affects the Objective-C
1415 compiler and runtime library, and exception handling for other languages
1416 like C++. The possibilities for @var{lib} are:
1424 LynxOS thread support.
1426 MIPS SDE thread support.
1428 This is an alias for @samp{single}.
1430 Generic POSIX/Unix98 thread support.
1432 RTEMS thread support.
1434 Disable thread support, should work for all platforms.
1438 VxWorks thread support.
1440 Microsoft Win32 API thread support.
1444 Specify that the target supports TLS (Thread Local Storage). Usually
1445 configure can correctly determine if TLS is supported. In cases where
1446 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1447 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1448 the assembler supports TLS but the C library does not, or if the
1449 assumptions made by the configure test are incorrect.
1452 Specify that the target does not support TLS.
1453 This is an alias for @option{--enable-tls=no}.
1455 @item --disable-tm-clone-registry
1456 Disable TM clone registry in libgcc. It is enabled in libgcc by default.
1457 This option helps to reduce code size for embedded targets which do
1458 not use transactional memory.
1460 @item --with-cpu=@var{cpu}
1461 @itemx --with-cpu-32=@var{cpu}
1462 @itemx --with-cpu-64=@var{cpu}
1463 Specify which cpu variant the compiler should generate code for by default.
1464 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1465 This option is only supported on some targets, including ARC, ARM, i386, M68k,
1466 PowerPC, and SPARC@. It is mandatory for ARC@. The @option{--with-cpu-32} and
1467 @option{--with-cpu-64} options specify separate default CPUs for
1468 32-bit and 64-bit modes; these options are only supported for aarch64, i386,
1469 x86-64, PowerPC, and SPARC@.
1471 @item --with-schedule=@var{cpu}
1472 @itemx --with-arch=@var{cpu}
1473 @itemx --with-arch-32=@var{cpu}
1474 @itemx --with-arch-64=@var{cpu}
1475 @itemx --with-tune=@var{cpu}
1476 @itemx --with-tune-32=@var{cpu}
1477 @itemx --with-tune-64=@var{cpu}
1478 @itemx --with-abi=@var{abi}
1479 @itemx --with-fpu=@var{type}
1480 @itemx --with-float=@var{type}
1481 These configure options provide default values for the @option{-mschedule=},
1482 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1483 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1484 @option{--with-cpu}, which switches will be accepted and acceptable values
1485 of the arguments depend on the target.
1487 @item --with-mode=@var{mode}
1488 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1489 This option is only supported on ARM targets.
1491 @item --with-stack-offset=@var{num}
1492 This option sets the default for the -mstack-offset=@var{num} option,
1493 and will thus generally also control the setting of this option for
1494 libraries. This option is only supported on Epiphany targets.
1496 @item --with-fpmath=@var{isa}
1497 This options sets @option{-mfpmath=sse} by default and specifies the default
1498 ISA for floating-point arithmetics. You can select either @samp{sse} which
1499 enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default.
1500 This option is only supported on i386 and x86-64 targets.
1502 @item --with-fp-32=@var{mode}
1503 On MIPS targets, set the default value for the @option{-mfp} option when using
1504 the o32 ABI. The possibilities for @var{mode} are:
1507 Use the o32 FP32 ABI extension, as with the @option{-mfp32} command-line
1510 Use the o32 FPXX ABI extension, as with the @option{-mfpxx} command-line
1513 Use the o32 FP64 ABI extension, as with the @option{-mfp64} command-line
1516 In the absence of this configuration option the default is to use the o32
1519 @item --with-odd-spreg-32
1520 On MIPS targets, set the @option{-modd-spreg} option by default when using
1523 @item --without-odd-spreg-32
1524 On MIPS targets, set the @option{-mno-odd-spreg} option by default when using
1525 the o32 ABI. This is normally used in conjunction with
1526 @option{--with-fp-32=64} in order to target the o32 FP64A ABI extension.
1528 @item --with-nan=@var{encoding}
1529 On MIPS targets, set the default encoding convention to use for the
1530 special not-a-number (NaN) IEEE 754 floating-point data. The
1531 possibilities for @var{encoding} are:
1534 Use the legacy encoding, as with the @option{-mnan=legacy} command-line
1537 Use the 754-2008 encoding, as with the @option{-mnan=2008} command-line
1540 To use this configuration option you must have an assembler version
1541 installed that supports the @option{-mnan=} command-line option too.
1542 In the absence of this configuration option the default convention is
1543 the legacy encoding, as when neither of the @option{-mnan=2008} and
1544 @option{-mnan=legacy} command-line options has been used.
1546 @item --with-divide=@var{type}
1547 Specify how the compiler should generate code for checking for
1548 division by zero. This option is only supported on the MIPS target.
1549 The possibilities for @var{type} are:
1552 Division by zero checks use conditional traps (this is the default on
1553 systems that support conditional traps).
1555 Division by zero checks use the break instruction.
1558 @item --with-compact-branches=@var{policy}
1559 Specify how the compiler should generate branch instructions.
1560 This option is only supported on the MIPS target.
1561 The possibilities for @var{type} are:
1564 Cause a delay slot branch to be used if one is available in the
1565 current ISA and the delay slot is successfully filled. If the delay slot
1566 is not filled, a compact branch will be chosen if one is available.
1568 Ensures that compact branch instructions will never be generated.
1570 Ensures that a compact branch instruction will be generated if available.
1571 If a compact branch instruction is not available,
1572 a delay slot form of the branch will be used instead.
1573 This option is supported from MIPS Release 6 onwards.
1574 For pre-R6/microMIPS/MIPS16, this option is just same as never/optimal.
1577 @c If you make --with-llsc the default for additional targets,
1578 @c update the --with-llsc description in the MIPS section below.
1581 On MIPS targets, make @option{-mllsc} the default when no
1582 @option{-mno-llsc} option is passed. This is the default for
1583 Linux-based targets, as the kernel will emulate them if the ISA does
1586 @item --without-llsc
1587 On MIPS targets, make @option{-mno-llsc} the default when no
1588 @option{-mllsc} option is passed.
1591 On MIPS targets, make @option{-msynci} the default when no
1592 @option{-mno-synci} option is passed.
1594 @item --without-synci
1595 On MIPS targets, make @option{-mno-synci} the default when no
1596 @option{-msynci} option is passed. This is the default.
1598 @item --with-lxc1-sxc1
1599 On MIPS targets, make @option{-mlxc1-sxc1} the default when no
1600 @option{-mno-lxc1-sxc1} option is passed. This is the default.
1602 @item --without-lxc1-sxc1
1603 On MIPS targets, make @option{-mno-lxc1-sxc1} the default when no
1604 @option{-mlxc1-sxc1} option is passed. The indexed load/store
1605 instructions are not directly a problem but can lead to unexpected
1606 behaviour when deployed in an application intended for a 32-bit address
1607 space but run on a 64-bit processor. The issue is seen because all
1608 known MIPS 64-bit Linux kernels execute o32 and n32 applications
1609 with 64-bit addressing enabled which affects the overflow behaviour
1610 of the indexed addressing mode. GCC will assume that ordinary
1611 32-bit arithmetic overflow behaviour is the same whether performed
1612 as an @code{addu} instruction or as part of the address calculation
1613 in @code{lwxc1} type instructions. This assumption holds true in a
1614 pure 32-bit environment and can hold true in a 64-bit environment if
1615 the address space is accurately set to be 32-bit for o32 and n32.
1618 On MIPS targets, make @option{-mmadd4} the default when no
1619 @option{-mno-madd4} option is passed. This is the default.
1621 @item --without-madd4
1622 On MIPS targets, make @option{-mno-madd4} the default when no
1623 @option{-mmadd4} option is passed. The @code{madd4} instruction
1624 family can be problematic when targeting a combination of cores that
1625 implement these instructions differently. There are two known cores
1626 that implement these as fused operations instead of unfused (where
1627 unfused is normally expected). Disabling these instructions is the
1628 only way to ensure compatible code is generated; this will incur
1629 a performance penalty.
1631 @item --with-mips-plt
1632 On MIPS targets, make use of copy relocations and PLTs.
1633 These features are extensions to the traditional
1634 SVR4-based MIPS ABIs and require support from GNU binutils
1635 and the runtime C library.
1637 @item --with-stack-clash-protection-guard-size=@var{size}
1638 On certain targets this option sets the default stack clash protection guard
1639 size as a power of two in bytes. On AArch64 @var{size} is required to be either
1640 12 (4KB) or 16 (64KB).
1642 @item --with-isa-spec=@var{ISA-spec-string}
1643 On RISC-V targets specify the default version of the RISC-V Unprivileged
1644 (formerly User-Level) ISA specification to produce code conforming to.
1645 The possibilities for @var{ISA-spec-string} are:
1648 Produce code conforming to version 2.2.
1650 Produce code conforming to version 20190608.
1652 Produce code conforming to version 20191213.
1654 In the absence of this configuration option the default version is 20191213.
1656 @item --enable-__cxa_atexit
1657 Define if you want to use __cxa_atexit, rather than atexit, to
1658 register C++ destructors for local statics and global objects.
1659 This is essential for fully standards-compliant handling of
1660 destructors, but requires __cxa_atexit in libc. This option is currently
1661 only available on systems with GNU libc. When enabled, this will cause
1662 @option{-fuse-cxa-atexit} to be passed by default.
1664 @item --enable-gnu-indirect-function
1665 Define if you want to enable the @code{ifunc} attribute. This option is
1666 currently only available on systems with GNU libc on certain targets.
1668 @item --enable-target-optspace
1670 libraries should be optimized for code space instead of code speed.
1671 This is the default for the m32r platform.
1673 @item --with-cpp-install-dir=@var{dirname}
1674 Specify that the user visible @command{cpp} program should be installed
1675 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1677 @item --enable-comdat
1678 Enable COMDAT group support. This is primarily used to override the
1679 automatically detected value.
1681 @item --enable-initfini-array
1682 Force the use of sections @code{.init_array} and @code{.fini_array}
1683 (instead of @code{.init} and @code{.fini}) for constructors and
1684 destructors. Option @option{--disable-initfini-array} has the
1685 opposite effect. If neither option is specified, the configure script
1686 will try to guess whether the @code{.init_array} and
1687 @code{.fini_array} sections are supported and, if they are, use them.
1689 @item --enable-link-mutex
1690 When building GCC, use a mutex to avoid linking the compilers for
1691 multiple languages at the same time, to avoid thrashing on build
1692 systems with limited free memory. The default is not to use such a mutex.
1694 @item --enable-link-serialization
1695 When building GCC, use make dependencies to serialize linking the compilers for
1696 multiple languages, to avoid thrashing on build
1697 systems with limited free memory. The default is not to add such
1698 dependencies and thus with parallel make potentially link different
1699 compilers concurrently. If the argument is a positive integer, allow
1700 that number of concurrent link processes for the large binaries.
1702 @item --enable-maintainer-mode
1703 The build rules that regenerate the Autoconf and Automake output files as
1704 well as the GCC master message catalog @file{gcc.pot} are normally
1705 disabled. This is because it can only be rebuilt if the complete source
1706 tree is present. If you have changed the sources and want to rebuild the
1707 catalog, configuring with @option{--enable-maintainer-mode} will enable
1708 this. Note that you need a recent version of the @code{gettext} tools
1711 @item --disable-bootstrap
1712 For a native build, the default configuration is to perform
1713 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1714 testing that GCC can compile itself correctly. If you want to disable
1715 this process, you can configure with @option{--disable-bootstrap}.
1717 @item --enable-bootstrap
1718 In special cases, you may want to perform a 3-stage build
1719 even if the target and host triplets are different.
1720 This is possible when the host can run code compiled for
1721 the target (e.g.@: host is i686-linux, target is i486-linux).
1722 Starting from GCC 4.2, to do this you have to configure explicitly
1723 with @option{--enable-bootstrap}.
1725 @item --enable-generated-files-in-srcdir
1726 Neither the .c and .h files that are generated from Bison and flex nor the
1727 info manuals and man pages that are built from the .texi files are present
1728 in the repository development tree. When building GCC from that development tree,
1729 or from one of our snapshots, those generated files are placed in your
1730 build directory, which allows for the source to be in a readonly
1733 If you configure with @option{--enable-generated-files-in-srcdir} then those
1734 generated files will go into the source directory. This is mainly intended
1735 for generating release or prerelease tarballs of the GCC sources, since it
1736 is not a requirement that the users of source releases to have flex, Bison,
1739 @item --enable-version-specific-runtime-libs
1741 that runtime libraries should be installed in the compiler specific
1742 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1743 addition, @samp{libstdc++}'s include files will be installed into
1744 @file{@var{libdir}} unless you overruled it by using
1745 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1746 particularly useful if you intend to use several versions of GCC in
1747 parallel. The default is @samp{yes} for @samp{libada}, and @samp{no} for
1748 the remaining libraries.
1750 @item @anchor{WithAixSoname}--with-aix-soname=@samp{aix}, @samp{svr4} or @samp{both}
1751 Traditional AIX shared library versioning (versioned @code{Shared Object}
1752 files as members of unversioned @code{Archive Library} files named
1753 @samp{lib.a}) causes numerous headaches for package managers. However,
1754 @code{Import Files} as members of @code{Archive Library} files allow for
1755 @strong{filename-based versioning} of shared libraries as seen on Linux/SVR4,
1756 where this is called the "SONAME". But as they prevent static linking,
1757 @code{Import Files} may be used with @code{Runtime Linking} only, where the
1758 linker does search for @samp{libNAME.so} before @samp{libNAME.a} library
1759 filenames with the @samp{-lNAME} linker flag.
1761 @anchor{AixLdCommand}For detailed information please refer to the AIX
1762 @uref{https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22,,ld
1765 As long as shared library creation is enabled, upon:
1767 @item --with-aix-soname=aix
1768 @item --with-aix-soname=both
1769 A (traditional AIX) @code{Shared Archive Library} file is created:
1771 @item using the @samp{libNAME.a} filename scheme
1772 @item with the @code{Shared Object} file as archive member named
1773 @samp{libNAME.so.V} (except for @samp{libgcc_s}, where the @code{Shared
1774 Object} file is named @samp{shr.o} for backwards compatibility), which
1776 @item is used for runtime loading from inside the @samp{libNAME.a} file
1777 @item is used for dynamic loading via
1778 @code{dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)}
1779 @item is used for shared linking
1780 @item is used for static linking, so no separate @code{Static Archive
1781 Library} file is needed
1784 @item --with-aix-soname=both
1785 @item --with-aix-soname=svr4
1786 A (second) @code{Shared Archive Library} file is created:
1788 @item using the @samp{libNAME.so.V} filename scheme
1789 @item with the @code{Shared Object} file as archive member named
1792 @item is created with the @code{-G linker flag}
1793 @item has the @code{F_LOADONLY} flag set
1794 @item is used for runtime loading from inside the @samp{libNAME.so.V} file
1795 @item is used for dynamic loading via @code{dlopen("libNAME.so.V(shr.o)",
1798 @item with the @code{Import File} as archive member named @samp{shr.imp},
1801 @item refers to @samp{libNAME.so.V(shr.o)} as the "SONAME", to be recorded
1802 in the @code{Loader Section} of subsequent binaries
1803 @item indicates whether @samp{libNAME.so.V(shr.o)} is 32 or 64 bit
1804 @item lists all the public symbols exported by @samp{lib.so.V(shr.o)},
1805 eventually decorated with the @code{@samp{weak} Keyword}
1806 @item is necessary for shared linking against @samp{lib.so.V(shr.o)}
1809 A symbolic link using the @samp{libNAME.so} filename scheme is created:
1811 @item pointing to the @samp{libNAME.so.V} @code{Shared Archive Library} file
1812 @item to permit the @code{ld Command} to find @samp{lib.so.V(shr.imp)} via
1813 the @samp{-lNAME} argument (requires @code{Runtime Linking} to be enabled)
1814 @item to permit dynamic loading of @samp{lib.so.V(shr.o)} without the need
1815 to specify the version number via @code{dlopen("libNAME.so(shr.o)",
1820 As long as static library creation is enabled, upon:
1822 @item --with-aix-soname=svr4
1823 A @code{Static Archive Library} is created:
1825 @item using the @samp{libNAME.a} filename scheme
1826 @item with all the @code{Static Object} files as archive members, which
1828 @item are used for static linking
1833 While the aix-soname=@samp{svr4} option does not create @code{Shared Object}
1834 files as members of unversioned @code{Archive Library} files any more, package
1835 managers still are responsible to
1836 @uref{./specific.html#TransferAixShobj,,transfer} @code{Shared Object} files
1837 found as member of a previously installed unversioned @code{Archive Library}
1838 file into the newly installed @code{Archive Library} file with the same
1841 @emph{WARNING:} Creating @code{Shared Object} files with @code{Runtime Linking}
1842 enabled may bloat the TOC, eventually leading to @code{TOC overflow} errors,
1843 requiring the use of either the @option{-Wl,-bbigtoc} linker flag (seen to
1844 break with the @code{GDB} debugger) or some of the TOC-related compiler flags,
1846 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1847 Using the GNU Compiler Collection (GCC)}.
1850 see ``RS/6000 and PowerPC Options'' in the main manual.
1853 @option{--with-aix-soname} is currently supported by @samp{libgcc_s} only, so
1854 this option is still experimental and not for normal use yet.
1856 Default is the traditional behavior @option{--with-aix-soname=@samp{aix}}.
1858 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1859 Specify that only a particular subset of compilers and
1860 their runtime libraries should be built. For a list of valid values for
1861 @var{langN} you can issue the following command in the
1862 @file{gcc} directory of your GCC source tree:@*
1864 grep ^language= */config-lang.in
1866 Currently, you can use any of the following:
1867 @code{all}, @code{default}, @code{ada}, @code{c}, @code{c++}, @code{d},
1868 @code{fortran}, @code{go}, @code{jit}, @code{lto}, @code{objc}, @code{obj-c++}.
1869 Building the Ada compiler has special requirements, see below.
1870 If you do not pass this flag, or specify the option @code{default}, then the
1871 default languages available in the @file{gcc} sub-tree will be configured.
1872 Ada, D, Go, Jit, and Objective-C++ are not default languages. LTO is not a
1873 default language, but is built by default because @option{--enable-lto} is
1874 enabled by default. The other languages are default languages. If
1875 @code{all} is specified, then all available languages are built. An
1876 exception is @code{jit} language, which requires
1877 @option{--enable-host-shared} to be included with @code{all}.
1879 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1880 Specify that a particular subset of compilers and their runtime
1881 libraries should be built with the system C compiler during stage 1 of
1882 the bootstrap process, rather than only in later stages with the
1883 bootstrapped C compiler. The list of valid values is the same as for
1884 @option{--enable-languages}, and the option @code{all} will select all
1885 of the languages enabled by @option{--enable-languages}. This option is
1886 primarily useful for GCC development; for instance, when a development
1887 version of the compiler cannot bootstrap due to compiler bugs, or when
1888 one is debugging front ends other than the C front end. When this
1889 option is used, one can then build the target libraries for the
1890 specified languages with the stage-1 compiler by using @command{make
1891 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1892 for the specified languages using @command{make stage1-start check-gcc}.
1894 @item --disable-libada
1895 Specify that the run-time libraries and tools used by GNAT should not
1896 be built. This can be useful for debugging, or for compatibility with
1897 previous Ada build procedures, when it was required to explicitly
1898 do a @samp{make -C gcc gnatlib_and_tools}.
1900 @item --disable-libsanitizer
1901 Specify that the run-time libraries for the various sanitizers should
1904 @item --disable-libssp
1905 Specify that the run-time libraries for stack smashing protection
1906 should not be built or linked against. On many targets library support
1907 is provided by the C library instead.
1909 @item --disable-libquadmath
1910 Specify that the GCC quad-precision math library should not be built.
1911 On some systems, the library is required to be linkable when building
1912 the Fortran front end, unless @option{--disable-libquadmath-support}
1915 @item --disable-libquadmath-support
1916 Specify that the Fortran front end and @code{libgfortran} do not add
1917 support for @code{libquadmath} on systems supporting it.
1919 @item --disable-libgomp
1920 Specify that the GNU Offloading and Multi Processing Runtime Library
1921 should not be built.
1923 @item --disable-libvtv
1924 Specify that the run-time libraries used by vtable verification
1925 should not be built.
1928 Specify that the compiler should
1929 use DWARF 2 debugging information as the default.
1931 @item --with-advance-toolchain=@var{at}
1932 On 64-bit PowerPC Linux systems, configure the compiler to use the
1933 header files, library files, and the dynamic linker from the Advance
1934 Toolchain release @var{at} instead of the default versions that are
1935 provided by the Linux distribution. In general, this option is
1936 intended for the developers of GCC, and it is not intended for general
1939 @item --enable-targets=all
1940 @itemx --enable-targets=@var{target_list}
1941 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1942 These are compilers that are able to generate either 64-bit or 32-bit
1943 code. Typically, the corresponding 32-bit target, e.g.@:
1944 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1945 option enables the 32-bit target to be a bi-arch compiler, which is
1946 useful when you want a bi-arch compiler that defaults to 32-bit, and
1947 you are building a bi-arch or multi-arch binutils in a combined tree.
1948 On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64),
1950 Currently, this option only affects sparc-linux, powerpc-linux, x86-linux,
1951 mips-linux and s390-linux.
1953 @item --enable-default-pie
1954 Turn on @option{-fPIE} and @option{-pie} by default.
1956 @item --enable-secureplt
1957 This option enables @option{-msecure-plt} by default for powerpc-linux.
1959 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1960 Using the GNU Compiler Collection (GCC)},
1963 See ``RS/6000 and PowerPC Options'' in the main manual
1966 @item --enable-default-ssp
1967 Turn on @option{-fstack-protector-strong} by default.
1970 This option enables @option{-mcld} by default for 32-bit x86 targets.
1972 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1973 Using the GNU Compiler Collection (GCC)},
1976 See ``i386 and x86-64 Options'' in the main manual
1979 @item --enable-large-address-aware
1980 The @option{--enable-large-address-aware} option arranges for MinGW
1981 executables to be linked using the @option{--large-address-aware}
1982 option, that enables the use of more than 2GB of memory. If GCC is
1983 configured with this option, its effects can be reversed by passing the
1984 @option{-Wl,--disable-large-address-aware} option to the so-configured
1987 @item --enable-win32-registry
1988 @itemx --enable-win32-registry=@var{key}
1989 @itemx --disable-win32-registry
1990 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1991 to look up installations paths in the registry using the following key:
1994 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1997 @var{key} defaults to GCC version number, and can be overridden by the
1998 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1999 who use custom installers are encouraged to provide a different key,
2000 perhaps one comprised of vendor name and GCC version number, to
2001 avoid conflict with existing installations. This feature is enabled
2002 by default, and can be disabled by @option{--disable-win32-registry}
2003 option. This option has no effect on the other hosts.
2006 Specify that the machine does not have a floating point unit. This
2007 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
2008 system, @option{--nfp} has no effect.
2010 @item --enable-werror
2011 @itemx --disable-werror
2012 @itemx --enable-werror=yes
2013 @itemx --enable-werror=no
2014 When you specify this option, it controls whether certain files in the
2015 compiler are built with @option{-Werror} in bootstrap stage2 and later.
2016 If you don't specify it, @option{-Werror} is turned on for the main
2017 development trunk. However it defaults to off for release branches and
2018 final releases. The specific files which get @option{-Werror} are
2019 controlled by the Makefiles.
2021 @item --enable-checking
2022 @itemx --disable-checking
2023 @itemx --enable-checking=@var{list}
2024 This option controls performing internal consistency checks in the compiler.
2025 It does not change the generated code, but adds error checking of the
2026 requested complexity. This slows down the compiler and may only work
2027 properly if you are building the compiler with GCC@.
2029 When the option is not specified, the active set of checks depends on context.
2030 Namely, bootstrap stage 1 defaults to @samp{--enable-checking=yes}, builds
2031 from release branches or release archives default to
2032 @samp{--enable-checking=release}, and otherwise
2033 @samp{--enable-checking=yes,extra} is used. When the option is
2034 specified without a @var{list}, the result is the same as
2035 @samp{--enable-checking=yes}. Likewise, @samp{--disable-checking} is
2036 equivalent to @samp{--enable-checking=no}.
2038 The categories of checks available in @var{list} are @samp{yes} (most common
2039 checks @samp{assert,misc,gc,gimple,rtlflag,runtime,tree,types}), @samp{no}
2040 (no checks at all), @samp{all} (all but @samp{valgrind}), @samp{release}
2041 (cheapest checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
2042 @samp{release} checks are always on and to disable them
2043 @samp{--disable-checking} or @samp{--enable-checking=no[,<other checks>]}
2044 must be explicitly requested. Disabling assertions makes the compiler and
2045 runtime slightly faster but increases the risk of undetected internal errors
2046 causing wrong code to be generated.
2048 Individual checks can be enabled with these flags: @samp{assert}, @samp{df},
2049 @samp{extra}, @samp{fold}, @samp{gc}, @samp{gcac}, @samp{gimple},
2050 @samp{misc}, @samp{rtl}, @samp{rtlflag}, @samp{runtime}, @samp{tree},
2051 @samp{types} and @samp{valgrind}. @samp{extra} extends @samp{misc}
2052 checking with extra checks that might affect code generation and should
2053 therefore not differ between stage1 and later stages in bootstrap.
2055 The @samp{valgrind} check requires the external @command{valgrind} simulator,
2056 available from @uref{https://valgrind.org}. The @samp{rtl} checks are
2057 expensive and the @samp{df}, @samp{gcac} and @samp{valgrind} checks are very
2060 @item --disable-stage1-checking
2061 @itemx --enable-stage1-checking
2062 @itemx --enable-stage1-checking=@var{list}
2063 This option affects only bootstrap build. If no @option{--enable-checking}
2064 option is specified the stage1 compiler is built with @samp{yes} checking
2065 enabled, otherwise the stage1 checking flags are the same as specified by
2066 @option{--enable-checking}. To build the stage1 compiler with
2067 different checking options use @option{--enable-stage1-checking}.
2068 The list of checking options is the same as for @option{--enable-checking}.
2069 If your system is too slow or too small to bootstrap a released compiler
2070 with checking for stage1 enabled, you can use @samp{--disable-stage1-checking}
2071 to disable checking for the stage1 compiler.
2073 @item --enable-coverage
2074 @itemx --enable-coverage=@var{level}
2075 With this option, the compiler is built to collect self coverage
2076 information, every time it is run. This is for internal development
2077 purposes, and only works when the compiler is being built with gcc. The
2078 @var{level} argument controls whether the compiler is built optimized or
2079 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
2080 want to disable optimization, for performance analysis you want to
2081 enable optimization. When coverage is enabled, the default level is
2082 without optimization.
2084 @item --enable-gather-detailed-mem-stats
2085 When this option is specified more detailed information on memory
2086 allocation is gathered. This information is printed when using
2087 @option{-fmem-report}.
2089 @item --enable-valgrind-annotations
2090 Mark selected memory related operations in the compiler when run under
2091 valgrind to suppress false positives.
2094 @itemx --disable-nls
2095 The @option{--enable-nls} option enables Native Language Support (NLS),
2096 which lets GCC output diagnostics in languages other than American
2097 English. Native Language Support is enabled by default if not doing a
2098 canadian cross build. The @option{--disable-nls} option disables NLS@.
2100 @item --with-included-gettext
2101 If NLS is enabled, the @option{--with-included-gettext} option causes the build
2102 procedure to prefer its copy of GNU @command{gettext}.
2104 @item --with-catgets
2105 If NLS is enabled, and if the host lacks @code{gettext} but has the
2106 inferior @code{catgets} interface, the GCC build procedure normally
2107 ignores @code{catgets} and instead uses GCC's copy of the GNU
2108 @code{gettext} library. The @option{--with-catgets} option causes the
2109 build procedure to use the host's @code{catgets} in this situation.
2111 @item --with-libiconv-prefix=@var{dir}
2112 Search for libiconv header files in @file{@var{dir}/include} and
2113 libiconv library files in @file{@var{dir}/lib}.
2115 @item --enable-obsolete
2116 Enable configuration for an obsoleted system. If you attempt to
2117 configure GCC for a system (build, host, or target) which has been
2118 obsoleted, and you do not specify this flag, configure will halt with an
2121 All support for systems which have been obsoleted in one release of GCC
2122 is removed entirely in the next major release, unless someone steps
2123 forward to maintain the port.
2125 @item --enable-decimal-float
2126 @itemx --enable-decimal-float=yes
2127 @itemx --enable-decimal-float=no
2128 @itemx --enable-decimal-float=bid
2129 @itemx --enable-decimal-float=dpd
2130 @itemx --disable-decimal-float
2131 Enable (or disable) support for the C decimal floating point extension
2132 that is in the IEEE 754-2008 standard. This is enabled by default only
2133 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
2134 support it, but require the user to specifically enable it. You can
2135 optionally control which decimal floating point format is used (either
2136 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
2137 format is default on i386 and x86_64 systems, and the @samp{dpd}
2138 (densely packed decimal) format is default on PowerPC systems.
2140 @item --enable-fixed-point
2141 @itemx --disable-fixed-point
2142 Enable (or disable) support for C fixed-point arithmetic.
2143 This option is enabled by default for some targets (such as MIPS) which
2144 have hardware-support for fixed-point operations. On other targets, you
2145 may enable this option manually.
2147 @item --with-long-double-128
2148 Specify if @code{long double} type should be 128-bit by default on selected
2149 GNU/Linux architectures. If using @code{--without-long-double-128},
2150 @code{long double} will be by default 64-bit, the same as @code{double} type.
2151 When neither of these configure options are used, the default will be
2152 128-bit @code{long double} when built against GNU C Library 2.4 and later,
2153 64-bit @code{long double} otherwise.
2155 @item --with-long-double-format=ibm
2156 @itemx --with-long-double-format=ieee
2157 Specify whether @code{long double} uses the IBM extended double format
2158 or the IEEE 128-bit floating point format on PowerPC Linux systems.
2159 This configuration switch will only work on little endian PowerPC
2160 Linux systems and on big endian 64-bit systems where the default cpu
2161 is at least power7 (i.e.@: @option{--with-cpu=power7},
2162 @option{--with-cpu=power8}, or @option{--with-cpu=power9} is used).
2164 If you use the @option{--with-long-double-64} configuration option,
2165 the @option{--with-long-double-format=ibm} and
2166 @option{--with-long-double-format=ieee} options are ignored.
2168 The default @code{long double} format is to use IBM extended double.
2169 Until all of the libraries are converted to use IEEE 128-bit floating
2170 point, it is not recommended to use
2171 @option{--with-long-double-format=ieee}.
2173 @item --enable-fdpic
2174 On SH Linux systems, generate ELF FDPIC code.
2176 @item --with-gmp=@var{pathname}
2177 @itemx --with-gmp-include=@var{pathname}
2178 @itemx --with-gmp-lib=@var{pathname}
2179 @itemx --with-mpfr=@var{pathname}
2180 @itemx --with-mpfr-include=@var{pathname}
2181 @itemx --with-mpfr-lib=@var{pathname}
2182 @itemx --with-mpc=@var{pathname}
2183 @itemx --with-mpc-include=@var{pathname}
2184 @itemx --with-mpc-lib=@var{pathname}
2185 If you want to build GCC but do not have the GMP library, the MPFR
2186 library and/or the MPC library installed in a standard location and
2187 do not have their sources present in the GCC source tree then you
2188 can explicitly specify the directory where they are installed
2189 (@samp{--with-gmp=@var{gmpinstalldir}},
2190 @samp{--with-mpfr=@/@var{mpfrinstalldir}},
2191 @samp{--with-mpc=@/@var{mpcinstalldir}}). The
2192 @option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for
2193 @option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and
2194 @option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the
2195 @option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for
2196 @option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and
2197 @option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the
2198 @option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for
2199 @option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and
2200 @option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these
2201 shorthand assumptions are not correct, you can use the explicit
2202 include and lib options directly. You might also need to ensure the
2203 shared libraries can be found by the dynamic linker when building and
2204 using GCC, for example by setting the runtime shared library path
2205 variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems).
2207 These flags are applicable to the host platform only. When building
2208 a cross compiler, they will not be used to configure target libraries.
2210 @item --with-isl=@var{pathname}
2211 @itemx --with-isl-include=@var{pathname}
2212 @itemx --with-isl-lib=@var{pathname}
2213 If you do not have the isl library installed in a standard location and you
2214 want to build GCC, you can explicitly specify the directory where it is
2215 installed (@samp{--with-isl=@/@var{islinstalldir}}). The
2216 @option{--with-isl=@/@var{islinstalldir}} option is shorthand for
2217 @option{--with-isl-lib=@/@var{islinstalldir}/lib} and
2218 @option{--with-isl-include=@/@var{islinstalldir}/include}. If this
2219 shorthand assumption is not correct, you can use the explicit
2220 include and lib options directly.
2222 These flags are applicable to the host platform only. When building
2223 a cross compiler, they will not be used to configure target libraries.
2225 @item --with-stage1-ldflags=@var{flags}
2226 This option may be used to set linker flags to be used when linking
2227 stage 1 of GCC. These are also used when linking GCC if configured with
2228 @option{--disable-bootstrap}. If @option{--with-stage1-libs} is not set to a
2229 value, then the default is @samp{-static-libstdc++ -static-libgcc}, if
2232 @item --with-stage1-libs=@var{libs}
2233 This option may be used to set libraries to be used when linking stage 1
2234 of GCC. These are also used when linking GCC if configured with
2235 @option{--disable-bootstrap}.
2237 @item --with-boot-ldflags=@var{flags}
2238 This option may be used to set linker flags to be used when linking
2239 stage 2 and later when bootstrapping GCC. If --with-boot-libs
2240 is not is set to a value, then the default is
2241 @samp{-static-libstdc++ -static-libgcc}.
2243 @item --with-boot-libs=@var{libs}
2244 This option may be used to set libraries to be used when linking stage 2
2245 and later when bootstrapping GCC.
2247 @item --with-debug-prefix-map=@var{map}
2248 Convert source directory names using @option{-fdebug-prefix-map} when
2249 building runtime libraries. @samp{@var{map}} is a space-separated
2250 list of maps of the form @samp{@var{old}=@var{new}}.
2252 @item --enable-linker-build-id
2253 Tells GCC to pass @option{--build-id} option to the linker for all final
2254 links (links performed without the @option{-r} or @option{--relocatable}
2255 option), if the linker supports it. If you specify
2256 @option{--enable-linker-build-id}, but your linker does not
2257 support @option{--build-id} option, a warning is issued and the
2258 @option{--enable-linker-build-id} option is ignored. The default is off.
2260 @item --with-linker-hash-style=@var{choice}
2261 Tells GCC to pass @option{--hash-style=@var{choice}} option to the
2262 linker for all final links. @var{choice} can be one of
2263 @samp{sysv}, @samp{gnu}, and @samp{both} where @samp{sysv} is the default.
2265 @item --enable-gnu-unique-object
2266 @itemx --disable-gnu-unique-object
2267 Tells GCC to use the gnu_unique_object relocation for C++ template
2268 static data members and inline function local statics. Enabled by
2269 default for a toolchain with an assembler that accepts it and
2270 GLIBC 2.11 or above, otherwise disabled.
2272 @item --with-diagnostics-color=@var{choice}
2273 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-color=}
2274 option (if not used explicitly on the command line). @var{choice}
2275 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2276 where @samp{auto} is the default. @samp{auto-if-env} makes
2277 @option{-fdiagnostics-color=auto} the default if @env{GCC_COLORS}
2278 is present and non-empty in the environment of the compiler, and
2279 @option{-fdiagnostics-color=never} otherwise.
2281 @item --with-diagnostics-urls=@var{choice}
2282 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-urls=}
2283 option (if not used explicitly on the command line). @var{choice}
2284 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2285 where @samp{auto} is the default. @samp{auto-if-env} makes
2286 @option{-fdiagnostics-urls=auto} the default if @env{GCC_URLS}
2287 or @env{TERM_URLS} is present and non-empty in the environment of the
2288 compiler, and @option{-fdiagnostics-urls=never} otherwise.
2291 @itemx --disable-lto
2292 Enable support for link-time optimization (LTO). This is enabled by
2293 default, and may be disabled using @option{--disable-lto}.
2295 @item --enable-linker-plugin-configure-flags=FLAGS
2296 @itemx --enable-linker-plugin-flags=FLAGS
2297 By default, linker plugins (such as the LTO plugin) are built for the
2298 host system architecture. For the case that the linker has a
2299 different (but run-time compatible) architecture, these flags can be
2300 specified to build plugins that are compatible to the linker. For
2301 example, if you are building GCC for a 64-bit x86_64
2302 (@samp{x86_64-pc-linux-gnu}) host system, but have a 32-bit x86
2303 GNU/Linux (@samp{i686-pc-linux-gnu}) linker executable (which is
2304 executable on the former system), you can configure GCC as follows for
2305 getting compatible linker plugins:
2308 % @var{srcdir}/configure \
2309 --host=x86_64-pc-linux-gnu \
2310 --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \
2311 --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib'
2314 @item --with-plugin-ld=@var{pathname}
2315 Enable an alternate linker to be used at link-time optimization (LTO)
2316 link time when @option{-fuse-linker-plugin} is enabled.
2317 This linker should have plugin support such as gold starting with
2318 version 2.20 or GNU ld starting with version 2.21.
2319 See @option{-fuse-linker-plugin} for details.
2321 @item --enable-canonical-system-headers
2322 @itemx --disable-canonical-system-headers
2323 Enable system header path canonicalization for @file{libcpp}. This can
2324 produce shorter header file paths in diagnostics and dependency output
2325 files, but these changed header paths may conflict with some compilation
2326 environments. Enabled by default, and may be disabled using
2327 @option{--disable-canonical-system-headers}.
2329 @item --with-glibc-version=@var{major}.@var{minor}
2330 Tell GCC that when the GNU C Library (glibc) is used on the target it
2331 will be version @var{major}.@var{minor} or later. Normally this can
2332 be detected from the C library's header files, but this option may be
2333 needed when bootstrapping a cross toolchain without the header files
2334 available for building the initial bootstrap compiler.
2336 If GCC is configured with some multilibs that use glibc and some that
2337 do not, this option applies only to the multilibs that use glibc.
2338 However, such configurations may not work well as not all the relevant
2339 configuration in GCC is on a per-multilib basis.
2341 @item --enable-as-accelerator-for=@var{target}
2342 Build as offload target compiler. Specify offload host triple by @var{target}.
2344 @item --enable-offload-targets=@var{target1}[=@var{path1}],@dots{},@var{targetN}[=@var{pathN}]
2345 Enable offloading to targets @var{target1}, @dots{}, @var{targetN}.
2346 Offload compilers are expected to be already installed. Default search
2347 path for them is @file{@var{exec-prefix}}, but it can be changed by
2348 specifying paths @var{path1}, @dots{}, @var{pathN}.
2351 % @var{srcdir}/configure \
2352 --enable-offload-targets=amdgcn-amdhsa,nvptx-none
2355 @item --enable-offload-defaulted
2357 Tell GCC that configured but not installed offload compilers and libgomp
2358 plugins are silently ignored. Useful for distribution compilers where
2359 those are in separate optional packages and where the presence or absence
2360 of those optional packages should determine the actual supported offloading
2361 target set rather than the GCC configure-time selection.
2364 @itemx --disable-cet
2365 Enable building target run-time libraries with control-flow
2366 instrumentation, see @option{-fcf-protection} option. When
2367 @code{--enable-cet} is specified target libraries are configured
2368 to add @option{-fcf-protection} and, if needed, other target
2369 specific options to a set of building options.
2371 @code{--enable-cet=auto} is default. CET is enabled on Linux/x86 if
2372 target binutils supports @code{Intel CET} instructions and disabled
2373 otherwise. In this case, the target libraries are configured to get
2374 additional @option{-fcf-protection} option.
2376 @item --with-riscv-attribute=@samp{yes}, @samp{no} or @samp{default}
2377 Generate RISC-V attribute by default, in order to record extra build
2378 information in object.
2380 The option is disabled by default. It is enabled on RISC-V/ELF (bare-metal)
2381 target if target binutils supported.
2383 @item --enable-s390-excess-float-precision
2384 @itemx --disable-s390-excess-float-precision
2385 On s390(x) targets, enable treatment of float expressions with double precision
2386 when in standards-compliant mode (e.g., when @code{--std=c99} or
2387 @code{-fexcess-precision=standard} are given).
2389 For a native build and cross compiles that have target headers, the option's
2390 default is derived from glibc's behavior. When glibc clamps float_t to double,
2391 GCC follows and enables the option. For other cross compiles, the default is
2394 @item --with-zstd=@var{pathname}
2395 @itemx --with-zstd-include=@var{pathname}
2396 @itemx --with-zstd-lib=@var{pathname}
2397 If you do not have the @code{zstd} library installed in a standard
2398 location and you want to build GCC, you can explicitly specify the
2399 directory where it is installed (@samp{--with-zstd=@/@var{zstdinstalldir}}).
2400 The @option{--with-zstd=@/@var{zstdinstalldir}} option is shorthand for
2401 @option{--with-zstd-lib=@/@var{zstdinstalldir}/lib} and
2402 @option{--with-zstd-include=@/@var{zstdinstalldir}/include}. If this
2403 shorthand assumption is not correct, you can use the explicit
2404 include and lib options directly.
2406 These flags are applicable to the host platform only. When building
2407 a cross compiler, they will not be used to configure target libraries.
2410 @subheading Cross-Compiler-Specific Options
2411 The following options only apply to building cross compilers.
2414 @item --with-toolexeclibdir=@var{dir}
2415 Specify the installation directory for libraries built with a cross compiler.
2416 The default is @option{$@{gcc_tooldir@}/lib}.
2418 @item --with-sysroot
2419 @itemx --with-sysroot=@var{dir}
2420 Tells GCC to consider @var{dir} as the root of a tree that contains
2421 (a subset of) the root filesystem of the target operating system.
2422 Target system headers, libraries and run-time object files will be
2423 searched for in there. More specifically, this acts as if
2424 @option{--sysroot=@var{dir}} was added to the default options of the built
2425 compiler. The specified directory is not copied into the
2426 install tree, unlike the options @option{--with-headers} and
2427 @option{--with-libs} that this option obsoletes. The default value,
2428 in case @option{--with-sysroot} is not given an argument, is
2429 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
2430 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
2431 the GCC binaries if the installation tree is moved.
2433 This option affects the system root for the compiler used to build
2434 target libraries (which runs on the build system) and the compiler newly
2435 installed with @code{make install}; it does not affect the compiler which is
2436 used to build GCC itself.
2438 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2439 option then the compiler will search that directory within @var{dirname} for
2440 native system headers rather than the default @file{/usr/include}.
2442 @item --with-build-sysroot
2443 @itemx --with-build-sysroot=@var{dir}
2444 Tells GCC to consider @var{dir} as the system root (see
2445 @option{--with-sysroot}) while building target libraries, instead of
2446 the directory specified with @option{--with-sysroot}. This option is
2447 only useful when you are already using @option{--with-sysroot}. You
2448 can use @option{--with-build-sysroot} when you are configuring with
2449 @option{--prefix} set to a directory that is different from the one in
2450 which you are installing GCC and your target libraries.
2452 This option affects the system root for the compiler used to build
2453 target libraries (which runs on the build system); it does not affect
2454 the compiler which is used to build GCC itself.
2456 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2457 option then the compiler will search that directory within @var{dirname} for
2458 native system headers rather than the default @file{/usr/include}.
2460 @item --with-headers
2461 @itemx --with-headers=@var{dir}
2462 Deprecated in favor of @option{--with-sysroot}.
2463 Specifies that target headers are available when building a cross compiler.
2464 The @var{dir} argument specifies a directory which has the target include
2465 files. These include files will be copied into the @file{gcc} install
2466 directory. @emph{This option with the @var{dir} argument is required} when
2467 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
2468 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
2469 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
2470 will be run on these files to make them compatible with GCC@.
2472 @item --without-headers
2473 Tells GCC not use any target headers from a libc when building a cross
2474 compiler. When crossing to GNU/Linux, you need the headers so GCC
2475 can build the exception handling for libgcc.
2478 @itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}"
2479 Deprecated in favor of @option{--with-sysroot}.
2480 Specifies a list of directories which contain the target runtime
2481 libraries. These libraries will be copied into the @file{gcc} install
2482 directory. If the directory list is omitted, this option has no
2486 Specifies that @samp{newlib} is
2487 being used as the target C library. This causes @code{__eprintf} to be
2488 omitted from @file{libgcc.a} on the assumption that it will be provided by
2494 @item --with-avrlibc
2495 Only supported for the AVR target. Specifies that @samp{AVR-Libc} is
2496 being used as the target C@tie{} library. This causes float support
2497 functions like @code{__addsf3} to be omitted from @file{libgcc.a} on
2498 the assumption that it will be provided by @file{libm.a}. For more
2499 technical details, cf. @uref{https://gcc.gnu.org/PR54461,,PR54461}.
2500 It is not supported for
2501 RTEMS configurations, which currently use newlib. The option is
2502 supported since version 4.7.2 and is the default in 4.8.0 and newer.
2504 @item --with-double=@{32|64|32,64|64,32@}
2505 @itemx --with-long-double=@{32|64|32,64|64,32|double@}
2506 Only supported for the AVR target since version@tie{}10.
2507 Specify the default layout available for the C/C++ @samp{double}
2508 and @samp{long double} type, respectively. The following rules apply:
2511 The first value after the @samp{=} specifies the default layout (in bits)
2512 of the type and also the default for the @option{-mdouble=} resp.
2513 @option{-mlong-double=} compiler option.
2515 If more than one value is specified, respective multilib variants are
2516 available, and @option{-mdouble=} resp. @option{-mlong-double=} acts
2517 as a multilib option.
2519 If @option{--with-long-double=double} is specified, @samp{double} and
2520 @samp{long double} will have the same layout.
2522 The defaults are @option{--with-long-double=64,32} and
2523 @option{--with-double=32,64}. The default @samp{double} layout imposed by
2524 the latter is compatible with older versions of the compiler that implement
2525 @samp{double} as a 32-bit type, which does not comply to the language standard.
2527 Not all combinations of @option{--with-double=} and
2528 @option{--with-long-double=} are valid. For example, the combination
2529 @option{--with-double=32,64} @option{--with-long-double=32} will be
2530 rejected because the first option specifies the availability of
2531 multilibs for @samp{double}, whereas the second option implies
2532 that @samp{long double} --- and hence also @samp{double} --- is always
2535 @item --with-double-comparison=@{tristate|bool|libf7@}
2536 Only supported for the AVR target since version@tie{}10.
2537 Specify what result format is returned by library functions that
2538 compare 64-bit floating point values (@code{DFmode}).
2539 The GCC default is @samp{tristate}. If the floating point
2540 implementation returns a boolean instead, set it to @samp{bool}.
2542 @item --with-libf7=@{libgcc|math|math-symbols|no@}
2543 Only supported for the AVR target since version@tie{}10.
2544 Specify to which degree code from LibF7 is included in libgcc.
2545 LibF7 is an ad-hoc, AVR-specific, 64-bit floating point emulation
2546 written in C and (inline) assembly. @samp{libgcc} adds support
2547 for functions that one would usually expect in libgcc like double addition,
2548 double comparisons and double conversions. @samp{math} also adds routines
2549 that one would expect in @file{libm.a}, but with @code{__} (two underscores)
2550 prepended to the symbol names as specified by @file{math.h}.
2551 @samp{math-symbols} also defines weak aliases for the functions
2552 declared in @file{math.h}. However, @code{--with-libf7} won't
2553 install no @file{math.h} header file whatsoever, this file must come
2554 from elsewhere. This option sets @option{--with-double-comparison}
2557 @item --with-nds32-lib=@var{library}
2558 Specifies that @var{library} setting is used for building @file{libgcc.a}.
2559 Currently, the valid @var{library} is @samp{newlib} or @samp{mculib}.
2560 This option is only supported for the NDS32 target.
2562 @item --with-build-time-tools=@var{dir}
2563 Specifies where to find the set of target tools (assembler, linker, etc.)
2564 that will be used while building GCC itself. This option can be useful
2565 if the directory layouts are different between the system you are building
2566 GCC on, and the system where you will deploy it.
2568 For example, on an @samp{ia64-hp-hpux} system, you may have the GNU
2569 assembler and linker in @file{/usr/bin}, and the native tools in a
2570 different path, and build a toolchain that expects to find the
2571 native tools in @file{/usr/bin}.
2573 When you use this option, you should ensure that @var{dir} includes
2574 @command{ar}, @command{as}, @command{ld}, @command{nm},
2575 @command{ranlib} and @command{strip} if necessary, and possibly
2576 @command{objdump}. Otherwise, GCC may use an inconsistent set of
2580 @subsubheading Overriding @command{configure} test results
2582 Sometimes, it might be necessary to override the result of some
2583 @command{configure} test, for example in order to ease porting to a new
2584 system or work around a bug in a test. The toplevel @command{configure}
2585 script provides three variables for this:
2589 @item build_configargs
2590 @cindex @code{build_configargs}
2591 The contents of this variable is passed to all build @command{configure}
2594 @item host_configargs
2595 @cindex @code{host_configargs}
2596 The contents of this variable is passed to all host @command{configure}
2599 @item target_configargs
2600 @cindex @code{target_configargs}
2601 The contents of this variable is passed to all target @command{configure}
2606 In order to avoid shell and @command{make} quoting issues for complex
2607 overrides, you can pass a setting for @env{CONFIG_SITE} and set
2608 variables in the site file.
2610 @subheading Objective-C-Specific Options
2612 The following options apply to the build of the Objective-C runtime library.
2615 @item --enable-objc-gc
2616 Specify that an additional variant of the GNU Objective-C runtime library
2617 is built, using an external build of the Boehm-Demers-Weiser garbage
2618 collector (@uref{https://www.hboehm.info/gc/}). This library needs to be
2619 available for each multilib variant, unless configured with
2620 @option{--enable-objc-gc=@samp{auto}} in which case the build of the
2621 additional runtime library is skipped when not available and the build
2624 @item --with-target-bdw-gc=@var{list}
2625 @itemx --with-target-bdw-gc-include=@var{list}
2626 @itemx --with-target-bdw-gc-lib=@var{list}
2627 Specify search directories for the garbage collector header files and
2628 libraries. @var{list} is a comma separated list of key value pairs of the
2629 form @samp{@var{multilibdir}=@var{path}}, where the default multilib key
2630 is named as @samp{.} (dot), or is omitted (e.g.@:
2631 @samp{--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32}).
2633 The options @option{--with-target-bdw-gc-include} and
2634 @option{--with-target-bdw-gc-lib} must always be specified together
2635 for each multilib variant and they take precedence over
2636 @option{--with-target-bdw-gc}. If @option{--with-target-bdw-gc-include}
2637 is missing values for a multilib, then the value for the default
2638 multilib is used (e.g.@: @samp{--with-target-bdw-gc-include=/opt/bdw-gc/include}
2639 @samp{--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32}).
2640 If none of these options are specified, the library is assumed in
2644 @subheading D-Specific Options
2646 The following options apply to the build of the D runtime library.
2649 @item --enable-libphobos-checking
2650 @itemx --disable-libphobos-checking
2651 @itemx --enable-libphobos-checking=@var{list}
2652 This option controls whether run-time checks and contracts are compiled into
2653 the D runtime library. When the option is not specified, the library is built
2654 with @samp{release} checking. When the option is specified without a
2655 @var{list}, the result is the same as @samp{--enable-libphobos-checking=yes}.
2656 Likewise, @samp{--disable-libphobos-checking} is equivalent to
2657 @samp{--enable-libphobos-checking=no}.
2659 The categories of checks available in @var{list} are @samp{yes} (compiles
2660 libphobos with @option{-fno-release}), @samp{no} (compiles libphobos with
2661 @option{-frelease}), @samp{all} (same as @samp{yes}), @samp{none} or
2662 @samp{release} (same as @samp{no}).
2664 Individual checks available in @var{list} are @samp{assert} (compiles libphobos
2665 with an extra option @option{-fassert}).
2667 @item --with-libphobos-druntime-only
2668 @itemx --with-libphobos-druntime-only=@var{choice}
2669 Specify whether to build only the core D runtime library (druntime), or both
2670 the core and standard library (phobos) into libphobos. This is useful for
2671 targets that have full support in druntime, but no or incomplete support
2672 in phobos. @var{choice} can be one of @samp{auto}, @samp{yes}, and @samp{no}
2673 where @samp{auto} is the default.
2675 When the option is not specified, the default choice @samp{auto} means that it
2676 is inferred whether the target has support for the phobos standard library.
2677 When the option is specified without a @var{choice}, the result is the same as
2678 @samp{--with-libphobos-druntime-only=yes}.
2680 @item --with-target-system-zlib
2681 Use installed @samp{zlib} rather than that included with GCC@. This needs
2682 to be available for each multilib variant, unless configured with
2683 @option{--with-target-system-zlib=@samp{auto}} in which case the GCC@ included
2684 @samp{zlib} is only used when the system installed library is not available.
2692 @uref{./index.html,,Return to the GCC Installation page}
2696 @c ***Building****************************************************************
2698 @comment node-name, next, previous, up
2699 @node Building, Testing, Configuration, Installing GCC
2705 @cindex Installing GCC: Building
2707 Now that GCC is configured, you are ready to build the compiler and
2710 Some commands executed when making the compiler may fail (return a
2711 nonzero status) and be ignored by @command{make}. These failures, which
2712 are often due to files that were not found, are expected, and can safely
2715 It is normal to have compiler warnings when compiling certain files.
2716 Unless you are a GCC developer, you can generally ignore these warnings
2717 unless they cause compilation to fail. Developers should attempt to fix
2718 any warnings encountered, however they can temporarily continue past
2719 warnings-as-errors by specifying the configure flag
2720 @option{--disable-werror}.
2722 On certain old systems, defining certain environment variables such as
2723 @env{CC} can interfere with the functioning of @command{make}.
2725 If you encounter seemingly strange errors when trying to build the
2726 compiler in a directory other than the source directory, it could be
2727 because you have previously configured the compiler in the source
2728 directory. Make sure you have done all the necessary preparations.
2730 If you build GCC on a BSD system using a directory stored in an old System
2731 V file system, problems may occur in running @command{fixincludes} if the
2732 System V file system doesn't support symbolic links. These problems
2733 result in a failure to fix the declaration of @code{size_t} in
2734 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
2735 that type mismatches occur, this could be the cause.
2737 The solution is not to use such a directory for building GCC@.
2739 Similarly, when building from the source repository or snapshots, or if you modify
2740 @file{*.l} files, you need the Flex lexical analyzer generator
2741 installed. If you do not modify @file{*.l} files, releases contain
2742 the Flex-generated files and you do not need Flex installed to build
2743 them. There is still one Flex-based lexical analyzer (part of the
2744 build machinery, not of GCC itself) that is used even if you only
2745 build the C front end.
2747 When building from the source repository or snapshots, or if you modify Texinfo
2748 documentation, you need version 4.7 or later of Texinfo installed if you
2749 want Info documentation to be regenerated. Releases contain Info
2750 documentation pre-built for the unmodified documentation in the release.
2752 @section Building a native compiler
2754 For a native build, the default configuration is to perform
2755 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
2756 This will build the entire GCC system and ensure that it compiles
2757 itself correctly. It can be disabled with the @option{--disable-bootstrap}
2758 parameter to @samp{configure}, but bootstrapping is suggested because
2759 the compiler will be tested more completely and could also have
2762 The bootstrapping process will complete the following steps:
2766 Build tools necessary to build the compiler.
2769 Perform a 3-stage bootstrap of the compiler. This includes building
2770 three times the target tools for use by the compiler such as binutils
2771 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
2772 individually linked or moved into the top level GCC source tree before
2776 Perform a comparison test of the stage2 and stage3 compilers.
2779 Build runtime libraries using the stage3 compiler from the previous step.
2783 If you are short on disk space you might consider @samp{make
2784 bootstrap-lean} instead. The sequence of compilation is the
2785 same described above, but object files from the stage1 and
2786 stage2 of the 3-stage bootstrap of the compiler are deleted as
2787 soon as they are no longer needed.
2789 If you wish to use non-default GCC flags when compiling the stage2
2790 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
2791 doing @samp{make}. For example, if you want to save additional space
2792 during the bootstrap and in the final installation as well, you can
2793 build the compiler binaries without debugging information as in the
2794 following example. This will save roughly 40% of disk space both for
2795 the bootstrap and the final installation. (Libraries will still contain
2796 debugging information.)
2799 make BOOT_CFLAGS='-O' bootstrap
2802 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
2803 are less well tested here than the default of @samp{-g -O2}, but should
2804 still work. In a few cases, you may find that you need to specify special
2805 flags such as @option{-msoft-float} here to complete the bootstrap; or,
2806 if the native compiler miscompiles the stage1 compiler, you may need
2807 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
2808 of the stage1 compiler that were miscompiled, or by using @samp{make
2809 bootstrap4} to increase the number of stages of bootstrap.
2811 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
2812 Since these are always compiled with the compiler currently being
2813 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
2814 compilation flags, as for non-bootstrapped target libraries.
2815 Again, if the native compiler miscompiles the stage1 compiler, you may
2816 need to work around this by avoiding non-working parts of the stage1
2817 compiler. Use @code{STAGE1_TFLAGS} to this end.
2819 If you used the flag @option{--enable-languages=@dots{}} to restrict
2820 the compilers to be built, only those you've actually enabled will be
2821 built. This will of course only build those runtime libraries, for
2822 which the particular compiler has been built. Please note,
2823 that re-defining @env{LANGUAGES} when calling @samp{make}
2824 @strong{does not} work anymore!
2826 If the comparison of stage2 and stage3 fails, this normally indicates
2827 that the stage2 compiler has compiled GCC incorrectly, and is therefore
2828 a potentially serious bug which you should investigate and report. (On
2829 a few systems, meaningful comparison of object files is impossible; they
2830 always appear ``different''. If you encounter this problem, you will
2831 need to disable comparison in the @file{Makefile}.)
2833 If you do not want to bootstrap your compiler, you can configure with
2834 @option{--disable-bootstrap}. In particular cases, you may want to
2835 bootstrap your compiler even if the target system is not the same as
2836 the one you are building on: for example, you could build a
2837 @code{powerpc-unknown-linux-gnu} toolchain on a
2838 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
2839 @option{--enable-bootstrap} to the configure script.
2841 @code{BUILD_CONFIG} can be used to bring in additional customization
2842 to the build. It can be set to a whitespace-separated list of names.
2843 For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will
2844 be included by the top-level @file{Makefile}, bringing in any settings
2845 it contains. The default @code{BUILD_CONFIG} can be set using the
2846 configure option @option{--with-build-config=@code{NAME}...}. Some
2847 examples of supported build configurations are:
2850 @item @samp{bootstrap-O1}
2851 Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds
2852 @option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to
2853 @samp{BOOT_CFLAGS='-g -O1'}.
2855 @item @samp{bootstrap-O3}
2856 @itemx @samp{bootstrap-Og}
2857 Analogous to @code{bootstrap-O1}.
2859 @item @samp{bootstrap-lto}
2860 Enables Link-Time Optimization for host tools during bootstrapping.
2861 @samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding
2862 @option{-flto} to @samp{BOOT_CFLAGS}. This option assumes that the host
2863 supports the linker plugin (e.g.@: GNU ld version 2.21 or later or GNU gold
2864 version 2.21 or later).
2866 @item @samp{bootstrap-lto-noplugin}
2867 This option is similar to @code{bootstrap-lto}, but is intended for
2868 hosts that do not support the linker plugin. Without the linker plugin
2869 static libraries are not compiled with link-time optimizations. Since
2870 the GCC middle end and back end are in @file{libbackend.a} this means
2871 that only the front end is actually LTO optimized.
2873 @item @samp{bootstrap-lto-lean}
2874 This option is similar to @code{bootstrap-lto}, but is intended for
2875 faster build by only using LTO in the final bootstrap stage.
2876 With @samp{make profiledbootstrap} the LTO frontend
2877 is trained only on generator files.
2879 @item @samp{bootstrap-debug}
2880 Verifies that the compiler generates the same executable code, whether
2881 or not it is asked to emit debug information. To this end, this
2882 option builds stage2 host programs without debug information, and uses
2883 @file{contrib/compare-debug} to compare them with the stripped stage3
2884 object files. If @code{BOOT_CFLAGS} is overridden so as to not enable
2885 debug information, stage2 will have it, and stage3 won't. This option
2886 is enabled by default when GCC bootstrapping is enabled, if
2887 @code{strip} can turn object files compiled with and without debug
2888 info into identical object files. In addition to better test
2889 coverage, this option makes default bootstraps faster and leaner.
2891 @item @samp{bootstrap-debug-big}
2892 Rather than comparing stripped object files, as in
2893 @code{bootstrap-debug}, this option saves internal compiler dumps
2894 during stage2 and stage3 and compares them as well, which helps catch
2895 additional potential problems, but at a great cost in terms of disk
2896 space. It can be specified in addition to @samp{bootstrap-debug}.
2898 @item @samp{bootstrap-debug-lean}
2899 This option saves disk space compared with @code{bootstrap-debug-big},
2900 but at the expense of some recompilation. Instead of saving the dumps
2901 of stage2 and stage3 until the final compare, it uses
2902 @option{-fcompare-debug} to generate, compare and remove the dumps
2903 during stage3, repeating the compilation that already took place in
2904 stage2, whose dumps were not saved.
2906 @item @samp{bootstrap-debug-lib}
2907 This option tests executable code invariance over debug information
2908 generation on target libraries, just like @code{bootstrap-debug-lean}
2909 tests it on host programs. It builds stage3 libraries with
2910 @option{-fcompare-debug}, and it can be used along with any of the
2911 @code{bootstrap-debug} options above.
2913 There aren't @code{-lean} or @code{-big} counterparts to this option
2914 because most libraries are only build in stage3, so bootstrap compares
2915 would not get significant coverage. Moreover, the few libraries built
2916 in stage2 are used in stage3 host programs, so we wouldn't want to
2917 compile stage2 libraries with different options for comparison purposes.
2919 @item @samp{bootstrap-debug-ckovw}
2920 Arranges for error messages to be issued if the compiler built on any
2921 stage is run without the option @option{-fcompare-debug}. This is
2922 useful to verify the full @option{-fcompare-debug} testing coverage. It
2923 must be used along with @code{bootstrap-debug-lean} and
2924 @code{bootstrap-debug-lib}.
2926 @item @samp{bootstrap-cet}
2927 This option enables Intel CET for host tools during bootstrapping.
2928 @samp{BUILD_CONFIG=bootstrap-cet} is equivalent to adding
2929 @option{-fcf-protection} to @samp{BOOT_CFLAGS}. This option
2930 assumes that the host supports Intel CET (e.g.@: GNU assembler version
2933 @item @samp{bootstrap-time}
2934 Arranges for the run time of each program started by the GCC driver,
2935 built in any stage, to be logged to @file{time.log}, in the top level of
2938 @item @samp{bootstrap-asan}
2939 Compiles GCC itself using Address Sanitization in order to catch invalid memory
2940 accesses within the GCC code.
2942 @item @samp{bootstrap-hwasan}
2943 Compiles GCC itself using HWAddress Sanitization in order to catch invalid
2944 memory accesses within the GCC code. This option is only available on AArch64
2945 systems that are running Linux kernel version 5.4 or later.
2949 @section Building a cross compiler
2951 When building a cross compiler, it is not generally possible to do a
2952 3-stage bootstrap of the compiler. This makes for an interesting problem
2953 as parts of GCC can only be built with GCC@.
2955 To build a cross compiler, we recommend first building and installing a
2956 native compiler. You can then use the native GCC compiler to build the
2957 cross compiler. The installed native compiler needs to be GCC version
2960 Assuming you have already installed a native copy of GCC and configured
2961 your cross compiler, issue the command @command{make}, which performs the
2966 Build host tools necessary to build the compiler.
2969 Build target tools for use by the compiler such as binutils (bfd,
2970 binutils, gas, gprof, ld, and opcodes)
2971 if they have been individually linked or moved into the top level GCC source
2972 tree before configuring.
2975 Build the compiler (single stage only).
2978 Build runtime libraries using the compiler from the previous step.
2981 Note that if an error occurs in any step the make process will exit.
2983 If you are not building GNU binutils in the same source tree as GCC,
2984 you will need a cross-assembler and cross-linker installed before
2985 configuring GCC@. Put them in the directory
2986 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
2987 you should put in this directory:
2991 This should be the cross-assembler.
2994 This should be the cross-linker.
2997 This should be the cross-archiver: a program which can manipulate
2998 archive files (linker libraries) in the target machine's format.
3001 This should be a program to construct a symbol table in an archive file.
3004 The installation of GCC will find these programs in that directory,
3005 and copy or link them to the proper place to for the cross-compiler to
3006 find them when run later.
3008 The easiest way to provide these files is to build the Binutils package.
3009 Configure it with the same @option{--host} and @option{--target}
3010 options that you use for configuring GCC, then build and install
3011 them. They install their executables automatically into the proper
3012 directory. Alas, they do not support all the targets that GCC
3015 If you are not building a C library in the same source tree as GCC,
3016 you should also provide the target libraries and headers before
3017 configuring GCC, specifying the directories with
3018 @option{--with-sysroot} or @option{--with-headers} and
3019 @option{--with-libs}. Many targets also require ``start files'' such
3020 as @file{crt0.o} and
3021 @file{crtn.o} which are linked into each executable. There may be several
3022 alternatives for @file{crt0.o}, for use with profiling or other
3023 compilation options. Check your target's definition of
3024 @code{STARTFILE_SPEC} to find out what start files it uses.
3026 @section Building in parallel
3028 GNU Make 3.80 and above, which is necessary to build GCC, support
3029 building in parallel. To activate this, you can use @samp{make -j 2}
3030 instead of @samp{make}. You can also specify a bigger number, and
3031 in most cases using a value greater than the number of processors in
3032 your machine will result in fewer and shorter I/O latency hits, thus
3033 improving overall throughput; this is especially true for slow drives
3034 and network filesystems.
3036 @section Building the Ada compiler
3039 @ref{GNAT-prerequisite}.
3042 @uref{prerequisites.html#GNAT-prerequisite,,GNAT prerequisites}.
3045 @section Building the D compiler
3048 @ref{GDC-prerequisite}.
3051 @uref{prerequisites.html#GDC-prerequisite,,GDC prerequisites}.
3054 @section Building with profile feedback
3056 It is possible to use profile feedback to optimize the compiler itself. This
3057 should result in a faster compiler binary. Experiments done on x86 using gcc
3058 3.3 showed approximately 7 percent speedup on compiling C programs. To
3059 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
3061 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
3062 compiler. This compiler is used to build a @code{stageprofile} compiler
3063 instrumented to collect execution counts of instruction and branch
3064 probabilities. Training run is done by building @code{stagetrain}
3065 compiler. Finally a @code{stagefeedback} compiler is built
3066 using the information collected.
3068 Unlike standard bootstrap, several additional restrictions apply. The
3069 compiler used to build @code{stage1} needs to support a 64-bit integral type.
3070 It is recommended to only use GCC for this.
3072 On Linux/x86_64 hosts with some restrictions (no virtualization) it is
3073 also possible to do autofdo build with @samp{make
3074 autoprofiledback}. This uses Linux perf to sample branches in the
3075 binary and then rebuild it with feedback derived from the profile.
3076 Linux perf and the @code{autofdo} toolkit needs to be installed for
3079 Only the profile from the current build is used, so when an error
3080 occurs it is recommended to clean before restarting. Otherwise
3081 the code quality may be much worse.
3088 @uref{./index.html,,Return to the GCC Installation page}
3092 @c ***Testing*****************************************************************
3094 @comment node-name, next, previous, up
3095 @node Testing, Final install, Building, Installing GCC
3099 @chapter Installing GCC: Testing
3102 @cindex Installing GCC: Testing
3105 Before you install GCC, we encourage you to run the testsuites and to
3106 compare your results with results from a similar configuration that have
3107 been submitted to the
3108 @uref{https://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
3109 Some of these archived results are linked from the build status lists
3110 at @uref{https://gcc.gnu.org/buildstat.html}, although not everyone who
3111 reports a successful build runs the testsuites and submits the results.
3112 This step is optional and may require you to download additional software,
3113 but it can give you confidence in your new GCC installation or point out
3114 problems before you install and start using your new GCC@.
3116 First, you must have @uref{download.html,,downloaded the testsuites}.
3117 These are included in the source tarball.
3119 Second, you must have the testing tools installed. This includes
3120 @uref{https://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
3121 the DejaGnu site has links to these.
3122 Some optional tests also require Python3 and pytest module.
3124 If the directories where @command{runtest} and @command{expect} were
3125 installed are not in the @env{PATH}, you may need to set the following
3126 environment variables appropriately, as in the following example (which
3127 assumes that DejaGnu has been installed under @file{/usr/local}):
3130 TCL_LIBRARY = /usr/local/share/tcl8.0
3131 DEJAGNULIBS = /usr/local/share/dejagnu
3134 (On systems such as Cygwin, these paths are required to be actual
3135 paths, not mounts or links; presumably this is due to some lack of
3136 portability in the DejaGnu code.)
3139 Finally, you can run the testsuite (which may take a long time):
3141 cd @var{objdir}; make -k check
3144 This will test various components of GCC, such as compiler
3145 front ends and runtime libraries. While running the testsuite, DejaGnu
3146 might emit some harmless messages resembling
3147 @samp{WARNING: Couldn't find the global config file.} or
3148 @samp{WARNING: Couldn't find tool init file} that can be ignored.
3150 If you are testing a cross-compiler, you may want to run the testsuite
3151 on a simulator as described at @uref{https://gcc.gnu.org/simtest-howto.html}.
3153 @section How can you run the testsuite on selected tests?
3155 In order to run sets of tests selectively, there are targets
3156 @samp{make check-gcc} and language specific @samp{make check-c},
3157 @samp{make check-c++}, @samp{make check-d} @samp{make check-fortran},
3158 @samp{make check-ada}, @samp{make check-objc}, @samp{make check-obj-c++},
3159 @samp{make check-lto}
3160 in the @file{gcc} subdirectory of the object directory. You can also
3161 just run @samp{make check} in a subdirectory of the object directory.
3164 A more selective way to just run all @command{gcc} execute tests in the
3168 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
3171 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
3172 the testsuite with filenames matching @samp{9805*}, you would use
3175 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
3178 The file-matching expression following @var{filename}@command{.exp=} is treated
3179 as a series of whitespace-delimited glob expressions so that multiple patterns
3180 may be passed, although any whitespace must either be escaped or surrounded by
3181 single quotes if multiple expressions are desired. For example,
3184 make check-g++ RUNTESTFLAGS="old-deja.exp=9805*\ virtual2.c @var{other-options}"
3185 make check-g++ RUNTESTFLAGS="'old-deja.exp=9805* virtual2.c' @var{other-options}"
3188 The @file{*.exp} files are located in the testsuite directories of the GCC
3189 source, the most important ones being @file{compile.exp},
3190 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
3191 To get a list of the possible @file{*.exp} files, pipe the
3192 output of @samp{make check} into a file and look at the
3193 @samp{Running @dots{} .exp} lines.
3195 @section Passing options and running multiple testsuites
3197 You can pass multiple options to the testsuite using the
3198 @samp{--target_board} option of DejaGNU, either passed as part of
3199 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
3200 work outside the makefiles. For example,
3203 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
3206 will run the standard @command{g++} testsuites (``unix'' is the target name
3207 for a standard native testsuite situation), passing
3208 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
3209 slashes separate options.
3211 You can run the testsuites multiple times using combinations of options
3212 with a syntax similar to the brace expansion of popular shells:
3215 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
3218 (Note the empty option caused by the trailing comma in the final group.)
3219 The following will run each testsuite eight times using the @samp{arm-sim}
3220 target, as if you had specified all possible combinations yourself:
3223 --target_board='arm-sim/-mhard-float/-O1 \
3224 arm-sim/-mhard-float/-O2 \
3225 arm-sim/-mhard-float/-O3 \
3226 arm-sim/-mhard-float \
3227 arm-sim/-msoft-float/-O1 \
3228 arm-sim/-msoft-float/-O2 \
3229 arm-sim/-msoft-float/-O3 \
3230 arm-sim/-msoft-float'
3233 They can be combined as many times as you wish, in arbitrary ways. This
3237 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
3240 will generate four combinations, all involving @samp{-Wextra}.
3242 The disadvantage to this method is that the testsuites are run in serial,
3243 which is a waste on multiprocessor systems. For users with GNU Make and
3244 a shell which performs brace expansion, you can run the testsuites in
3245 parallel by having the shell perform the combinations and @command{make}
3246 do the parallel runs. Instead of using @samp{--target_board}, use a
3247 special makefile target:
3250 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
3256 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
3259 will run three concurrent ``make-gcc'' testsuites, eventually testing all
3260 ten combinations as described above. Note that this is currently only
3261 supported in the @file{gcc} subdirectory. (To see how this works, try
3262 typing @command{echo} before the example given here.)
3265 @section How to interpret test results
3267 The result of running the testsuite are various @file{*.sum} and @file{*.log}
3268 files in the testsuite subdirectories. The @file{*.log} files contain a
3269 detailed log of the compiler invocations and the corresponding
3270 results, the @file{*.sum} files summarize the results. These summaries
3271 contain status codes for all tests:
3275 PASS: the test passed as expected
3277 XPASS: the test unexpectedly passed
3279 FAIL: the test unexpectedly failed
3281 XFAIL: the test failed as expected
3283 UNSUPPORTED: the test is not supported on this platform
3285 ERROR: the testsuite detected an error
3287 WARNING: the testsuite detected a possible problem
3290 It is normal for some tests to report unexpected failures. At the
3291 current time the testing harness does not allow fine grained control
3292 over whether or not a test is expected to fail. This problem should
3293 be fixed in future releases.
3296 @section Submitting test results
3298 If you want to report the results to the GCC project, use the
3299 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
3302 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
3303 -m gcc-testresults@@gcc.gnu.org |sh
3306 This script uses the @command{Mail} program to send the results, so
3307 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
3308 prepended to the testsuite summary and should contain any special
3309 remarks you have on your results or your build environment. Please
3310 do not edit the testsuite result block or the subject line, as these
3311 messages may be automatically processed.
3318 @uref{./index.html,,Return to the GCC Installation page}
3322 @c ***Final install***********************************************************
3324 @comment node-name, next, previous, up
3325 @node Final install, , Testing, Installing GCC
3327 @ifset finalinstallhtml
3329 @chapter Installing GCC: Final installation
3332 Now that GCC has been built (and optionally tested), you can install it with
3334 cd @var{objdir} && make install
3337 We strongly recommend to install into a target directory where there is
3338 no previous version of GCC present. Also, the GNAT runtime should not
3339 be stripped, as this would break certain features of the debugger that
3340 depend on this debugging information (catching Ada exceptions for
3343 That step completes the installation of GCC; user level binaries can
3344 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
3345 you specified with the @option{--prefix} to configure (or
3346 @file{/usr/local} by default). (If you specified @option{--bindir},
3347 that directory will be used instead; otherwise, if you specified
3348 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
3349 Headers for the C++ library are installed in
3350 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
3351 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
3352 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
3353 in info format in @file{@var{infodir}} (normally
3354 @file{@var{prefix}/info}).
3356 When installing cross-compilers, GCC's executables
3357 are not only installed into @file{@var{bindir}}, that
3358 is, @file{@var{exec-prefix}/bin}, but additionally into
3359 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
3360 exists. Typically, such @dfn{tooldirs} hold target-specific
3361 binutils, including assembler and linker.
3363 Installation into a temporary staging area or into a @command{chroot}
3364 jail can be achieved with the command
3367 make DESTDIR=@var{path-to-rootdir} install
3371 where @var{path-to-rootdir} is the absolute path of
3372 a directory relative to which all installation paths will be
3373 interpreted. Note that the directory specified by @code{DESTDIR}
3374 need not exist yet; it will be created if necessary.
3376 There is a subtle point with tooldirs and @code{DESTDIR}:
3377 If you relocate a cross-compiler installation with
3378 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
3379 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
3380 be filled with duplicated GCC executables only if it already exists,
3381 it will not be created otherwise. This is regarded as a feature,
3382 not as a bug, because it gives slightly more control to the packagers
3383 using the @code{DESTDIR} feature.
3385 You can install stripped programs and libraries with
3391 If you are bootstrapping a released version of GCC then please
3392 quickly review the build status page for your release, available from
3393 @uref{https://gcc.gnu.org/buildstat.html}.
3394 If your system is not listed for the version of GCC that you built,
3396 @email{gcc@@gcc.gnu.org} indicating
3397 that you successfully built and installed GCC@.
3398 Include the following information:
3402 Output from running @file{@var{srcdir}/config.guess}. Do not send
3403 that file itself, just the one-line output from running it.
3406 The output of @samp{gcc -v} for your newly installed @command{gcc}.
3407 This tells us which version of GCC you built and the options you passed to
3411 Whether you enabled all languages or a subset of them. If you used a
3412 full distribution then this information is part of the configure
3413 options in the output of @samp{gcc -v}, but if you downloaded the
3414 ``core'' compiler plus additional front ends then it isn't apparent
3415 which ones you built unless you tell us about it.
3418 If the build was for GNU/Linux, also include:
3421 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
3422 this information should be available from @file{/etc/issue}.
3425 The version of the Linux kernel, available from @samp{uname --version}
3429 The version of glibc you used; for RPM-based systems like Red Hat,
3430 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
3431 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
3433 For other systems, you can include similar information if you think it is
3437 Any other information that you think would be useful to people building
3438 GCC on the same configuration. The new entry in the build status list
3439 will include a link to the archived copy of your message.
3442 We'd also like to know if the
3444 @ref{Specific, host/target specific installation notes}
3447 @uref{specific.html,,host/target specific installation notes}
3449 didn't include your host/target information or if that information is
3450 incomplete or out of date. Send a note to
3451 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
3453 If you find a bug, please report it following the
3454 @uref{../bugs/,,bug reporting guidelines}.
3456 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
3457 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
3458 and @TeX{} installed. This creates a number of @file{.dvi} files in
3459 subdirectories of @file{@var{objdir}}; these may be converted for
3460 printing with programs such as @command{dvips}. Alternately, by using
3461 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
3462 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
3463 is included with Texinfo version 4.8 and later. You can also
3464 @uref{https://shop.fsf.org/,,buy printed manuals from the
3465 Free Software Foundation}, though such manuals may not be for the most
3466 recent version of GCC@.
3468 If you would like to generate online HTML documentation, do @samp{cd
3469 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
3470 @file{@var{objdir}/gcc/HTML}.
3477 @uref{./index.html,,Return to the GCC Installation page}
3481 @c ***Binaries****************************************************************
3483 @comment node-name, next, previous, up
3484 @node Binaries, Specific, Installing GCC, Top
3488 @chapter Installing GCC: Binaries
3491 @cindex Installing GCC: Binaries
3493 We are often asked about pre-compiled versions of GCC@. While we cannot
3494 provide these for all platforms, below you'll find links to binaries for
3495 various platforms where creating them by yourself is not easy due to various
3498 Please note that we did not create these binaries, nor do we
3499 support them. If you have any problems installing them, please
3500 contact their makers.
3507 @uref{http://www.perzl.org/aix/,,AIX Open Source Packages (AIX5L AIX 6.1
3512 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
3518 @uref{http://hpux.connect.org.uk/,,HP-UX Porting Center};
3522 Solaris 2 (SPARC, Intel):
3525 @uref{https://www.opencsw.org/,,OpenCSW}
3532 The @uref{https://brew.sh,,Homebrew} package manager;
3534 @uref{https://www.macports.org,,MacPorts}.
3541 The @uref{https://sourceware.org/cygwin/,,Cygwin} project;
3543 The @uref{https://osdn.net/projects/mingw/,,MinGW} and
3544 @uref{https://www.mingw-w64.org/,,mingw-w64} projects.
3548 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
3549 number of platforms.
3552 The @uref{https://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
3553 links to GNU Fortran binaries for several platforms.
3561 @uref{./index.html,,Return to the GCC Installation page}
3565 @c ***Specific****************************************************************
3567 @comment node-name, next, previous, up
3568 @node Specific, GNU Free Documentation License, Binaries, Top
3572 @chapter Host/target specific installation notes for GCC
3575 @cindex Specific installation notes
3576 @cindex Target specific installation
3577 @cindex Host specific installation
3578 @cindex Target specific installation notes
3580 Please read this document carefully @emph{before} installing the
3581 GNU Compiler Collection on your machine.
3583 Note that this list of install notes is @emph{not} a list of supported
3584 hosts or targets. Not all supported hosts and targets are listed
3585 here, only the ones that require host-specific or target-specific
3586 information have to.
3591 @uref{#aarch64-x-x,,aarch64*-*-*}
3593 @uref{#alpha-x-x,,alpha*-*-*}
3595 @uref{#amdgcn-x-amdhsa,,amdgcn-*-amdhsa}
3597 @uref{#amd64-x-solaris2,,amd64-*-solaris2*}
3599 @uref{#arc-x-elf32,,arc-*-elf32}
3601 @uref{#arc-linux-uclibc,,arc-linux-uclibc}
3603 @uref{#arm-x-eabi,,arm-*-eabi}
3607 @uref{#bfin,,Blackfin}
3613 @uref{#epiphany-x-elf,,epiphany-*-elf}
3615 @uref{#ft32-x-elf,,ft32-*-elf}
3617 @uref{#x-x-freebsd,,*-*-freebsd*}
3619 @uref{#h8300-hms,,h8300-hms}
3621 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
3623 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
3625 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
3627 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
3629 @uref{#ix86-x-linux,,i?86-*-linux*}
3631 @uref{#ix86-x-solaris2,,i?86-*-solaris2*}
3633 @uref{#ia64-x-linux,,ia64-*-linux}
3635 @uref{#ia64-x-hpux,,ia64-*-hpux*}
3637 @uref{#x-ibm-aix,,*-ibm-aix*}
3639 @uref{#iq2000-x-elf,,iq2000-*-elf}
3641 @uref{#loongarch,,loongarch}
3643 @uref{#lm32-x-elf,,lm32-*-elf}
3645 @uref{#lm32-x-uclinux,,lm32-*-uclinux}
3647 @uref{#m32c-x-elf,,m32c-*-elf}
3649 @uref{#m32r-x-elf,,m32r-*-elf}
3651 @uref{#m68k-x-x,,m68k-*-*}
3653 @uref{#m68k-x-uclinux,,m68k-*-uclinux}
3655 @uref{#microblaze-x-elf,,microblaze-*-elf}
3657 @uref{#mips-x-x,,mips-*-*}
3659 @uref{#moxie-x-elf,,moxie-*-elf}
3661 @uref{#msp430-x-elf,,msp430-*-elf}
3663 @uref{#nds32le-x-elf,,nds32le-*-elf}
3665 @uref{#nds32be-x-elf,,nds32be-*-elf}
3667 @uref{#nvptx-x-none,,nvptx-*-none}
3669 @uref{#or1k-x-elf,,or1k-*-elf}
3671 @uref{#or1k-x-linux,,or1k-*-linux}
3673 @uref{#powerpc-x-x,,powerpc*-*-*}
3675 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
3677 @uref{#powerpc-x-elf,,powerpc-*-elf}
3679 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
3681 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
3683 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
3685 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
3687 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
3689 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
3691 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
3693 @uref{#riscv32-x-elf,,riscv32-*-elf}
3695 @uref{#riscv32-x-linux,,riscv32-*-linux}
3697 @uref{#riscv64-x-elf,,riscv64-*-elf}
3699 @uref{#riscv64-x-linux,,riscv64-*-linux}
3701 @uref{#rl78-x-elf,,rl78-*-elf}
3703 @uref{#rx-x-elf,,rx-*-elf}
3705 @uref{#s390-x-linux,,s390-*-linux*}
3707 @uref{#s390x-x-linux,,s390x-*-linux*}
3709 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
3711 @uref{#x-x-solaris2,,*-*-solaris2*}
3713 @uref{#sparc-x-x,,sparc*-*-*}
3715 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
3717 @uref{#sparc-x-linux,,sparc-*-linux*}
3719 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
3721 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
3723 @uref{#c6x-x-x,,c6x-*-*}
3725 @uref{#visium-x-elf, visium-*-elf}
3727 @uref{#x-x-vxworks,,*-*-vxworks*}
3729 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
3731 @uref{#x86-64-x-solaris2,,x86_64-*-solaris2*}
3733 @uref{#xtensa-x-elf,,xtensa*-*-elf}
3735 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
3737 @uref{#windows,,Microsoft Windows}
3739 @uref{#x-x-cygwin,,*-*-cygwin}
3741 @uref{#x-x-mingw32,,*-*-mingw32}
3745 @uref{#older,,Older systems}
3750 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
3756 <!-- -------- host/target specific issues start here ---------------- -->
3759 @anchor{aarch64-x-x}
3760 @heading aarch64*-*-*
3761 Binutils pre 2.24 does not have support for selecting @option{-mabi} and
3762 does not support ILP32. If it is used to build GCC 4.9 or later, GCC will
3763 not support option @option{-mabi=ilp32}.
3765 To enable a workaround for the Cortex-A53 erratum number 835769 by default
3766 (for all CPUs regardless of -mcpu option given) at configure time use the
3767 @option{--enable-fix-cortex-a53-835769} option. This will enable the fix by
3768 default and can be explicitly disabled during compilation by passing the
3769 @option{-mno-fix-cortex-a53-835769} option. Conversely,
3770 @option{--disable-fix-cortex-a53-835769} will disable the workaround by
3771 default. The workaround is disabled by default if neither of
3772 @option{--enable-fix-cortex-a53-835769} or
3773 @option{--disable-fix-cortex-a53-835769} is given at configure time.
3775 To enable a workaround for the Cortex-A53 erratum number 843419 by default
3776 (for all CPUs regardless of -mcpu option given) at configure time use the
3777 @option{--enable-fix-cortex-a53-843419} option. This workaround is applied at
3778 link time. Enabling the workaround will cause GCC to pass the relevant option
3779 to the linker. It can be explicitly disabled during compilation by passing the
3780 @option{-mno-fix-cortex-a53-843419} option. Conversely,
3781 @option{--disable-fix-cortex-a53-843419} will disable the workaround by default.
3782 The workaround is disabled by default if neither of
3783 @option{--enable-fix-cortex-a53-843419} or
3784 @option{--disable-fix-cortex-a53-843419} is given at configure time.
3786 To enable Branch Target Identification Mechanism and Return Address Signing by
3787 default at configure time use the @option{--enable-standard-branch-protection}
3788 option. This is equivalent to having @option{-mbranch-protection=standard}
3789 during compilation. This can be explicitly disabled during compilation by
3790 passing the @option{-mbranch-protection=none} option which turns off all
3791 types of branch protections. Conversely,
3792 @option{--disable-standard-branch-protection} will disable both the
3793 protections by default. This mechanism is turned off by default if neither
3794 of the options are given at configure time.
3801 This section contains general configuration information for all
3802 Alpha-based platforms using ELF@. In addition to reading this
3803 section, please read all other sections that match your target.
3808 @anchor{amd64-x-solaris2}
3809 @heading amd64-*-solaris2*
3810 This is a synonym for @samp{x86_64-*-solaris2*}.
3815 @anchor{amdgcn-x-amdhsa}
3816 @heading amdgcn-*-amdhsa
3819 Instead of GNU Binutils, you will need to install LLVM 13.0.1, or later, and copy
3820 @file{bin/llvm-mc} to @file{amdgcn-amdhsa/bin/as},
3821 @file{bin/lld} to @file{amdgcn-amdhsa/bin/ld},
3822 @file{bin/llvm-nm} to @file{amdgcn-amdhsa/bin/nm}, and
3823 @file{bin/llvm-ar} to both @file{bin/amdgcn-amdhsa-ar} and
3824 @file{bin/amdgcn-amdhsa-ranlib}.
3826 Use Newlib (3.2.0, or newer).
3828 To run the binaries, install the HSA Runtime from the
3829 @uref{https://rocm.github.io,,ROCm Platform}, and use
3830 @file{libexec/gcc/amdhsa-amdhsa/@var{version}/gcn-run} to launch them
3836 @anchor{arc-x-elf32}
3837 @heading arc-*-elf32
3839 Use @samp{configure --target=arc-elf32 --with-cpu=@var{cpu} --enable-languages="c,c++"}
3840 to configure GCC, with @var{cpu} being one of @samp{arc600}, @samp{arc601},
3846 @anchor{arc-linux-uclibc}
3847 @heading arc-linux-uclibc
3849 Use @samp{configure --target=arc-linux-uclibc --with-cpu=arc700 --enable-languages="c,c++"} to configure GCC@.
3856 ARM-family processors.
3858 Building the Ada frontend commonly fails (an infinite loop executing
3859 @code{xsinfo}) if the host compiler is GNAT 4.8. Host compilers built from the
3860 GNAT 4.6, 4.9 or 5 release branches are known to succeed.
3867 ATMEL AVR-family micro controllers. These are used in embedded
3868 applications. There are no standard Unix configurations.
3870 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
3874 See ``AVR Options'' in the main manual
3876 for the list of supported MCU types.
3878 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
3880 Further installation notes and other useful information about AVR tools
3881 can also be obtained from:
3885 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
3887 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
3890 The following error:
3892 Error: register required
3895 indicates that you should upgrade to a newer version of the binutils.
3902 The Blackfin processor, an Analog Devices DSP.
3904 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
3908 See ``Blackfin Options'' in the main manual
3911 More information, and a version of binutils with support for this processor,
3912 are available at @uref{https://sourceforge.net/projects/adi-toolchain/}.
3919 CRIS is a CPU architecture in Axis Communications systems-on-a-chip, for
3920 example the ETRAX series. These are used in embedded applications.
3923 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
3927 See ``CRIS Options'' in the main manual
3929 for a list of CRIS-specific options.
3931 Use @samp{configure --target=cris-elf} to configure GCC@ for building
3932 a cross-compiler for CRIS.
3938 Please have a look at the @uref{binaries.html,,binaries page}.
3940 You cannot install GCC by itself on MSDOS; it will not compile under
3941 any MSDOS compiler except itself. You need to get the complete
3942 compilation package DJGPP, which includes binaries as well as sources,
3943 and includes all the necessary compilation tools and libraries.
3948 @anchor{epiphany-x-elf}
3949 @heading epiphany-*-elf
3951 This configuration is intended for embedded systems.
3956 @anchor{x-x-freebsd}
3957 @heading *-*-freebsd*
3958 In order to better utilize FreeBSD base system functionality and match
3959 the configuration of the system compiler, GCC 4.5 and above as well as
3960 GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present
3961 on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default
3962 (on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside
3963 @file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled
3964 by GCC 4.5 and above.
3966 We support FreeBSD using the ELF file format with DWARF 2 debugging
3967 for all CPU architectures. There are
3968 no known issues with mixing object files and libraries with different
3969 debugging formats. Otherwise, this release of GCC should now match
3970 more of the configuration used in the stock FreeBSD configuration of
3971 GCC@. In particular, @option{--enable-threads} is now configured by
3972 default. However, as a general user, do not attempt to replace the
3973 system compiler with this release. Known to bootstrap and check with
3974 good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap
3975 and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4,
3976 4.5, 4.8, 4.9 and 5-CURRENT@.
3978 The version of binutils installed in @file{/usr/bin} probably works
3979 with this release of GCC@. Bootstrapping against the latest GNU
3980 binutils and/or the version found in @file{/usr/ports/devel/binutils} has
3981 been known to enable additional features and improve overall testsuite
3982 results. However, it is currently known that boehm-gc may not configure
3983 properly on FreeBSD prior to the FreeBSD 7.0 release with GNU binutils
3992 This configuration is intended for embedded systems.
3999 Renesas H8/300 series of processors.
4001 Please have a look at the @uref{binaries.html,,binaries page}.
4003 The calling convention and structure layout has changed in release 2.6.
4004 All code must be recompiled. The calling convention now passes the
4005 first three arguments in function calls in registers. Structures are no
4006 longer a multiple of 2 bytes.
4011 @anchor{hppa-hp-hpux}
4012 @heading hppa*-hp-hpux*
4013 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
4015 We require using gas/binutils on all hppa platforms. Version 2.19 or
4016 later is recommended.
4018 It may be helpful to configure GCC with the
4019 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
4020 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
4022 The HP assembler should not be used with GCC. It is rarely tested and may
4023 not work. It shouldn't be used with any languages other than C due to its
4026 Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging
4027 format which GCC does not know about). It also inserts timestamps
4028 into each object file it creates, causing the 3-stage comparison test to
4029 fail during a bootstrap. You should be able to continue by saying
4030 @samp{make all-host all-target} after getting the failure from @samp{make}.
4032 Various GCC features are not supported. For example, it does not support weak
4033 symbols or alias definitions. As a result, explicit template instantiations
4034 are required when using C++. This makes it difficult if not impossible to
4035 build many C++ applications.
4037 There are two default scheduling models for instructions. These are
4038 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
4039 architecture specified for the target machine when configuring.
4040 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
4041 the target is a @samp{hppa1*} machine.
4043 The PROCESSOR_8000 model is not well suited to older processors. Thus,
4044 it is important to completely specify the machine architecture when
4045 configuring if you want a model other than PROCESSOR_8000. The macro
4046 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
4047 default scheduling model is desired.
4049 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
4050 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
4051 This namespace change might cause problems when bootstrapping with
4052 an earlier version of GCC or the HP compiler as essentially the same
4053 namespace is required for an entire build. This problem can be avoided
4054 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
4055 or @samp{98}. Another way is to add an appropriate set of predefines
4056 to @env{CC}. The description for the @option{munix=} option contains
4057 a list of the predefines used with each standard.
4059 More specific information to @samp{hppa*-hp-hpux*} targets follows.
4064 @anchor{hppa-hp-hpux10}
4065 @heading hppa*-hp-hpux10
4066 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
4067 @code{PHCO_19798} from HP@.
4069 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
4070 used for one-only code and data. This resolves many of the previous
4071 problems in using C++ on this target. However, the ABI is not compatible
4072 with the one implemented under HP-UX 11 using secondary definitions.
4077 @anchor{hppa-hp-hpux11}
4078 @heading hppa*-hp-hpux11
4079 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
4080 be used to compile GCC 3.0 and up.
4082 The libffi library haven't been ported to 64-bit HP-UX@ and doesn't build.
4084 Refer to @uref{binaries.html,,binaries} for information about obtaining
4085 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
4086 to build the Ada language as it cannot be bootstrapped using C@. Ada is
4087 only available for the 32-bit PA-RISC runtime.
4089 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
4090 bundled compiler supports only traditional C; you will need either HP's
4091 unbundled compiler, or a binary distribution of GCC@.
4093 It is possible to build GCC 3.3 starting with the bundled HP compiler,
4094 but the process requires several steps. GCC 3.3 can then be used to
4095 build later versions.
4097 There are several possible approaches to building the distribution.
4098 Binutils can be built first using the HP tools. Then, the GCC
4099 distribution can be built. The second approach is to build GCC
4100 first using the HP tools, then build binutils, then rebuild GCC@.
4101 There have been problems with various binary distributions, so it
4102 is best not to start from a binary distribution.
4104 On 64-bit capable systems, there are two distinct targets. Different
4105 installation prefixes must be used if both are to be installed on
4106 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
4107 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
4108 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
4109 PA-RISC 2.0 architecture.
4111 The script config.guess now selects the target type based on the compiler
4112 detected during configuration. You must define @env{PATH} or @env{CC} so
4113 that configure finds an appropriate compiler for the initial bootstrap.
4114 When @env{CC} is used, the definition should contain the options that are
4115 needed whenever @env{CC} is used.
4117 Specifically, options that determine the runtime architecture must be
4118 in @env{CC} to correctly select the target for the build. It is also
4119 convenient to place many other compiler options in @env{CC}. For example,
4120 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
4121 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
4122 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
4123 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
4124 macro definition table of cpp needs to be increased for a successful
4125 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
4126 be defined when building with the bundled compiler, or when using the
4127 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
4129 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
4130 with the @option{--with-ld=@dots{}} option. This overrides the standard
4131 search for ld. The two linkers supported on this target require different
4132 commands. The default linker is determined during configuration. As a
4133 result, it's not possible to switch linkers in the middle of a GCC build.
4134 This has been reported to sometimes occur in unified builds of binutils
4137 A recent linker patch must be installed for the correct operation of
4138 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
4139 oldest linker patches that are known to work. They are for HP-UX
4140 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
4141 @code{PHSS_24304}, might be usable but it hasn't been tested. These
4142 patches have been superseded. Consult the HP patch database to obtain
4143 the currently recommended linker patch for your system.
4145 The patches are necessary for the support of weak symbols on the
4146 32-bit port, and for the running of initializers and finalizers. Weak
4147 symbols are implemented using SOM secondary definition symbols. Prior
4148 to HP-UX 11, there are bugs in the linker support for secondary symbols.
4149 The patches correct a problem of linker core dumps creating shared
4150 libraries containing secondary symbols, as well as various other
4151 linking issues involving secondary symbols.
4153 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
4154 run initializers and finalizers on the 64-bit port. The 32-bit port
4155 uses the linker @option{+init} and @option{+fini} options for the same
4156 purpose. The patches correct various problems with the +init/+fini
4157 options, including program core dumps. Binutils 2.14 corrects a
4158 problem on the 64-bit port resulting from HP's non-standard use of
4159 the .init and .fini sections for array initializers and finalizers.
4161 Although the HP and GNU linkers are both supported for the
4162 @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the
4163 HP linker be used for link editing on this target.
4165 At this time, the GNU linker does not support the creation of long
4166 branch stubs. As a result, it cannot successfully link binaries
4167 containing branch offsets larger than 8 megabytes. In addition,
4168 there are problems linking shared libraries, linking executables
4169 with @option{-static}, and with dwarf2 unwind and exception support.
4170 It also doesn't provide stubs for internal calls to global functions
4171 in shared libraries, so these calls cannot be overloaded.
4173 The HP dynamic loader does not support GNU symbol versioning, so symbol
4174 versioning is not supported. It may be necessary to disable symbol
4175 versioning with @option{--disable-symvers} when using GNU ld.
4177 POSIX threads are the default. The optional DCE thread library is not
4178 supported, so @option{--enable-threads=dce} does not work.
4183 @anchor{x-x-linux-gnu}
4184 @heading *-*-linux-gnu
4185 The @code{.init_array} and @code{.fini_array} sections are enabled
4186 unconditionally which requires at least glibc 2.1 and binutils 2.12.
4188 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
4189 in glibc 2.2.5 and later. More information is available in the
4190 libstdc++-v3 documentation.
4195 @anchor{ix86-x-linux}
4196 @heading i?86-*-linux*
4197 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
4198 See @uref{https://gcc.gnu.org/PR10877,,bug 10877} for more information.
4200 If you receive Signal 11 errors when building on GNU/Linux, then it is
4201 possible you have a hardware problem. Further information on this can be
4202 found on @uref{https://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
4207 @anchor{ix86-x-solaris2}
4208 @heading i?86-*-solaris2*
4209 Use this for Solaris 11.3 or later on x86 and x86-64 systems. Starting
4210 with GCC 4.7, there is also a 64-bit @samp{amd64-*-solaris2*} or
4211 @samp{x86_64-*-solaris2*} configuration that corresponds to
4212 @samp{sparcv9-sun-solaris2*}.
4214 It is recommended that you configure GCC to use the GNU assembler. The
4215 versions included in Solaris 11.3, from GNU binutils 2.23.1 or
4216 newer (available as @file{/usr/bin/gas} and
4217 @file{/usr/gnu/bin/as}), work fine. The current version, from GNU
4218 binutils 2.34, is known to work. Recent versions of the Solaris assembler in
4219 @file{/usr/bin/as} work almost as well, though.
4221 For linking, the Solaris linker is preferred. If you want to use the GNU
4222 linker instead, the version in Solaris 11.3, from GNU binutils 2.23.1 or
4223 newer (in @file{/usr/gnu/bin/ld} and @file{/usr/bin/gld}), works,
4224 as does the latest version, from GNU binutils 2.34.
4226 To use GNU @command{as}, configure with the options
4227 @option{--with-gnu-as --with-as=@//usr/@/gnu/@/bin/@/as}. It may be necessary
4228 to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to
4229 guarantee use of Solaris @command{ld}.
4230 @c FIXME: why --without-gnu-ld --with-ld?
4235 @anchor{ia64-x-linux}
4236 @heading ia64-*-linux
4237 IA-64 processor (also known as IPF, or Itanium Processor Family)
4240 If you are using the installed system libunwind library with
4241 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
4247 @anchor{ia64-x-hpux}
4248 @heading ia64-*-hpux*
4249 Building GCC on this target requires the GNU Assembler. The bundled HP
4250 assembler will not work. To prevent GCC from using the wrong assembler,
4251 the option @option{--with-gnu-as} may be necessary.
4253 The GCC libunwind library has not been ported to HPUX@. This means that for
4254 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
4255 is required to build GCC@. For GCC 3.3 and later, this is the default.
4256 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
4257 removed and the system libunwind library will always be used.
4261 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
4265 Support for AIX version 3 and older was discontinued in GCC 3.4.
4266 Support for AIX version 4.2 and older was discontinued in GCC 4.5.
4268 ``out of memory'' bootstrap failures may indicate a problem with
4269 process resource limits (ulimit). Hard limits are configured in the
4270 @file{/etc/security/limits} system configuration file.
4272 GCC 4.9 and above require a C++ compiler for bootstrap. IBM VAC++ / xlC
4273 cannot bootstrap GCC. xlc can bootstrap an older version of GCC and
4274 G++ can bootstrap recent releases of GCC.
4276 GCC can bootstrap with recent versions of IBM XLC, but bootstrapping
4277 with an earlier release of GCC is recommended. Bootstrapping with XLC
4278 requires a larger data segment, which can be enabled through the
4279 @var{LDR_CNTRL} environment variable, e.g.,
4282 % LDR_CNTRL=MAXDATA=0x50000000
4286 One can start with a pre-compiled version of GCC to build from
4287 sources. One may delete GCC's ``fixed'' header files when starting
4288 with a version of GCC built for an earlier release of AIX.
4290 To speed up the configuration phases of bootstrapping and installing GCC,
4291 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
4294 % CONFIG_SHELL=/opt/freeware/bin/bash
4295 % export CONFIG_SHELL
4298 and then proceed as described in @uref{build.html,,the build
4299 instructions}, where we strongly recommend specifying an absolute path
4300 to invoke @var{srcdir}/configure.
4302 Because GCC on AIX is built as a 32-bit executable by default,
4303 (although it can generate 64-bit programs) the GMP and MPFR libraries
4304 required by gfortran must be 32-bit libraries. Building GMP and MPFR
4305 as static archive libraries works better than shared libraries.
4307 Errors involving @code{alloca} when building GCC generally are due
4308 to an incorrect definition of @code{CC} in the Makefile or mixing files
4309 compiled with the native C compiler and GCC@. During the stage1 phase of
4310 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
4311 (not @command{xlc}). Once @command{configure} has been informed of
4312 @command{xlc}, one needs to use @samp{make distclean} to remove the
4313 configure cache files and ensure that @env{CC} environment variable
4314 does not provide a definition that will confuse @command{configure}.
4315 If this error occurs during stage2 or later, then the problem most likely
4316 is the version of Make (see above).
4318 The native @command{as} and @command{ld} are recommended for
4319 bootstrapping on AIX@. The GNU Assembler, GNU Linker, and GNU
4320 Binutils version 2.20 is the minimum level that supports bootstrap on
4321 AIX 5@. The GNU Assembler has not been updated to support AIX 6@ or
4322 AIX 7. The native AIX tools do interoperate with GCC@.
4324 AIX 7.1 added partial support for DWARF debugging, but full support
4325 requires AIX 7.1 TL03 SP7 that supports additional DWARF sections and
4326 fixes a bug in the assembler. AIX 7.1 TL03 SP5 distributed a version
4327 of libm.a missing important symbols; a fix for IV77796 will be
4330 AIX 5.3 TL10, AIX 6.1 TL05 and AIX 7.1 TL00 introduced an AIX
4331 assembler change that sometimes produces corrupt assembly files
4332 causing AIX linker errors. The bug breaks GCC bootstrap on AIX and
4333 can cause compilation failures with existing GCC installations. An
4334 AIX iFix for AIX 5.3 is available (APAR IZ98385 for AIX 5.3 TL10, APAR
4335 IZ98477 for AIX 5.3 TL11 and IZ98134 for AIX 5.3 TL12). AIX 5.3 TL11 SP8,
4336 AIX 5.3 TL12 SP5, AIX 6.1 TL04 SP11, AIX 6.1 TL05 SP7, AIX 6.1 TL06 SP6,
4337 AIX 6.1 TL07 and AIX 7.1 TL01 should include the fix.
4339 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
4340 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
4341 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
4342 referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1)
4344 @anchor{TransferAixShobj}
4345 @samp{libstdc++} in GCC 3.4 increments the major version number of the
4346 shared object and GCC installation places the @file{libstdc++.a}
4347 shared library in a common location which will overwrite the and GCC
4348 3.3 version of the shared library. Applications either need to be
4349 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
4350 versions of the @samp{libstdc++} shared object needs to be available
4351 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
4352 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
4353 installed for runtime dynamic loading using the following steps to set
4354 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
4355 multilib @file{libstdc++.a} installed:
4357 Extract the shared objects from the currently installed
4358 @file{libstdc++.a} archive:
4360 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
4363 Enable the @samp{F_LOADONLY} flag so that the shared object will be
4364 available for runtime dynamic loading, but not linking:
4366 % strip -e libstdc++.so.4 libstdc++.so.5
4369 Archive the runtime-only shared object in the GCC 3.4
4370 @file{libstdc++.a} archive:
4372 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
4376 @uref{./configure.html#WithAixSoname,,@option{--with-aix-soname=svr4}}
4377 configure option may drop the need for this procedure for libraries that
4380 Linking executables and shared libraries may produce warnings of
4381 duplicate symbols. The assembly files generated by GCC for AIX always
4382 have included multiple symbol definitions for certain global variable
4383 and function declarations in the original program. The warnings should
4384 not prevent the linker from producing a correct library or runnable
4387 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
4388 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
4389 to parse archive libraries did not handle the new format correctly.
4390 These routines are used by GCC and result in error messages during
4391 linking such as ``not a COFF file''. The version of the routines shipped
4392 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
4393 option of the archive command may be used to create archives of 32-bit
4394 objects using the original ``small format''. A correct version of the
4395 routines is shipped with AIX 4.3.2 and above.
4397 Some versions of the AIX binder (linker) can fail with a relocation
4398 overflow severe error when the @option{-bbigtoc} option is used to link
4399 GCC-produced object files into an executable that overflows the TOC@. A fix
4400 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
4401 available from IBM Customer Support and from its
4402 @uref{https://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4403 website as PTF U455193.
4405 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
4406 with a segmentation fault when invoked by any version of GCC@. A fix for
4407 APAR IX87327 is available from IBM Customer Support and from its
4408 @uref{https://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4409 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
4411 The initial assembler shipped with AIX 4.3.0 generates incorrect object
4412 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
4413 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
4414 @uref{https://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4415 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
4417 AIX provides National Language Support (NLS)@. Compilers and assemblers
4418 use NLS to support locale-specific representations of various data
4419 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
4420 separating decimal fractions). There have been problems reported where
4421 GCC does not produce the same floating-point formats that the assembler
4422 expects. If one encounters this problem, set the @env{LANG}
4423 environment variable to @samp{C} or @samp{En_US}.
4425 A default can be specified with the @option{-mcpu=@var{cpu_type}}
4426 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
4431 @anchor{iq2000-x-elf}
4432 @heading iq2000-*-elf
4433 Vitesse IQ2000 processors. These are used in embedded
4434 applications. There are no standard Unix configurations.
4441 Lattice Mico32 processor.
4442 This configuration is intended for embedded systems.
4447 @anchor{lm32-x-uclinux}
4448 @heading lm32-*-uclinux
4449 Lattice Mico32 processor.
4450 This configuration is intended for embedded systems running uClinux.
4457 LoongArch processor.
4458 The following LoongArch targets are available:
4460 @item loongarch64-linux-gnu*
4461 LoongArch processor running GNU/Linux. This target triplet may be coupled
4462 with a small set of possible suffixes to identify their default ABI type:
4465 Uses @code{lp64d/base} ABI by default.
4467 Uses @code{lp64f/base} ABI by default.
4469 Uses @code{lp64s/base} ABI by default.
4472 @item loongarch64-linux-gnu
4473 Same as @code{loongarch64-linux-gnuf64}, but may be used with
4474 @option{--with-abi=*} to configure the default ABI type.
4477 More information about LoongArch can be found at
4478 @uref{https://github.com/loongson/LoongArch-Documentation}.
4485 Renesas M32C processor.
4486 This configuration is intended for embedded systems.
4493 Renesas M32R processor.
4494 This configuration is intended for embedded systems.
4502 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
4504 build libraries for both M680x0 and ColdFire processors. If you only
4505 need the M680x0 libraries, you can omit the ColdFire ones by passing
4506 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
4507 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
4508 @command{configure}. These targets default to 5206 or 5475 code as
4509 appropriate for the target system when
4510 configured with @option{--with-arch=cf} and 68020 code otherwise.
4512 The @samp{m68k-*-netbsd} and
4513 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
4514 option. They will generate ColdFire CFV4e code when configured with
4515 @option{--with-arch=cf} and 68020 code otherwise.
4517 You can override the default processors listed above by configuring
4518 with @option{--with-cpu=@var{target}}. This @var{target} can either
4519 be a @option{-mcpu} argument or one of the following values:
4520 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
4521 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
4523 GCC requires at least binutils version 2.17 on these targets.
4528 @anchor{m68k-x-uclinux}
4529 @heading m68k-*-uclinux
4530 GCC 4.3 changed the uClinux configuration so that it uses the
4531 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
4532 It also added improved support for C++ and flat shared libraries,
4533 both of which were ABI changes.
4538 @anchor{microblaze-x-elf}
4539 @heading microblaze-*-elf
4540 Xilinx MicroBlaze processor.
4541 This configuration is intended for embedded systems.
4548 If on a MIPS system you get an error message saying ``does not have gp
4549 sections for all it's [sic] sectons [sic]'', don't worry about it. This
4550 happens whenever you use GAS with the MIPS linker, but there is not
4551 really anything wrong, and it is okay to use the output file. You can
4552 stop such warnings by installing the GNU linker.
4554 It would be nice to extend GAS to produce the gp tables, but they are
4555 optional, and there should not be a warning about their absence.
4557 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
4558 and later. A patch went in just after the GCC 3.3 release to
4559 make @samp{mips*-*-*} use the generic implementation instead. You can also
4560 configure for @samp{mipsel-elf} as a workaround. The
4561 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
4562 work on this is expected in future releases.
4564 @c If you make --with-llsc the default for another target, please also
4565 @c update the description of the --with-llsc option.
4567 The built-in @code{__sync_*} functions are available on MIPS II and
4568 later systems and others that support the @samp{ll}, @samp{sc} and
4569 @samp{sync} instructions. This can be overridden by passing
4570 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
4571 Since the Linux kernel emulates these instructions if they are
4572 missing, the default for @samp{mips*-*-linux*} targets is
4573 @option{--with-llsc}. The @option{--with-llsc} and
4574 @option{--without-llsc} configure options may be overridden at compile
4575 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
4578 MIPS systems check for division by zero (unless
4579 @option{-mno-check-zero-division} is passed to the compiler) by
4580 generating either a conditional trap or a break instruction. Using
4581 trap results in smaller code, but is only supported on MIPS II and
4582 later. Also, some versions of the Linux kernel have a bug that
4583 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
4584 the use of break, use the @option{--with-divide=breaks}
4585 @command{configure} option when configuring GCC@. The default is to
4586 use traps on systems that support them.
4591 @anchor{moxie-x-elf}
4592 @heading moxie-*-elf
4593 The moxie processor.
4598 @anchor{msp430-x-elf}
4599 @heading msp430-*-elf*
4600 TI MSP430 processor.
4601 This configuration is intended for embedded systems.
4603 @samp{msp430-*-elf} is the standard configuration with most GCC
4604 features enabled by default.
4606 @samp{msp430-*-elfbare} is tuned for a bare-metal environment, and disables
4607 features related to shared libraries and other functionality not used for
4608 this device. This reduces code and data usage of the GCC libraries, resulting
4609 in a minimal run-time environment by default.
4611 Features disabled by default include:
4613 @item transactional memory
4620 @anchor{nds32le-x-elf}
4621 @heading nds32le-*-elf
4622 Andes NDS32 target in little endian mode.
4627 @anchor{nds32be-x-elf}
4628 @heading nds32be-*-elf
4629 Andes NDS32 target in big endian mode.
4634 @anchor{nvptx-x-none}
4635 @heading nvptx-*-none
4638 Instead of GNU binutils, you will need to install
4639 @uref{https://github.com/MentorEmbedded/nvptx-tools/,,nvptx-tools}.
4640 Tell GCC where to find it:
4641 @option{--with-build-time-tools=[install-nvptx-tools]/nvptx-none/bin}.
4643 You will need newlib 3.1.0 or later. It can be
4644 automatically built together with GCC@. For this, add a symbolic link
4645 to nvptx-newlib's @file{newlib} directory to the directory containing
4648 Use the @option{--disable-sjlj-exceptions} and
4649 @option{--enable-newlib-io-long-long} options when configuring.
4651 The @option{--with-arch} option may be specified to override the
4652 default value for the @option{-march} option, and to also build
4653 corresponding target libraries.
4654 The default is @option{--with-arch=sm_30}.
4656 For example, if @option{--with-arch=sm_70} is specified,
4657 @option{-march=sm_30} and @option{-march=sm_70} target libraries are
4658 built, and code generation defaults to @option{-march=sm_70}.
4665 The OpenRISC 1000 32-bit processor with delay slots.
4666 This configuration is intended for embedded systems.
4671 @anchor{or1k-x-linux}
4672 @heading or1k-*-linux
4673 The OpenRISC 1000 32-bit processor with delay slots.
4678 @anchor{powerpc-x-x}
4679 @heading powerpc-*-*
4680 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
4681 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
4683 You will need GNU binutils 2.20 or newer.
4688 @anchor{powerpc-x-darwin}
4689 @heading powerpc-*-darwin*
4690 PowerPC running Darwin (Mac OS X kernel).
4692 Pre-installed versions of Mac OS X may not include any developer tools,
4693 meaning that you will not be able to build GCC from source. Tool
4694 binaries are available at
4695 @uref{https://opensource.apple.com}.
4697 This version of GCC requires at least cctools-590.36. The
4698 cctools-590.36 package referenced from
4699 @uref{https://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
4700 on systems older than 10.3.9 (aka darwin7.9.0).
4705 @anchor{powerpc-x-elf}
4706 @heading powerpc-*-elf
4707 PowerPC system in big endian mode, running System V.4.
4712 @anchor{powerpc-x-linux-gnu}
4713 @heading powerpc*-*-linux-gnu*
4714 PowerPC system in big endian mode running Linux.
4719 @anchor{powerpc-x-netbsd}
4720 @heading powerpc-*-netbsd*
4721 PowerPC system in big endian mode running NetBSD@.
4726 @anchor{powerpc-x-eabisim}
4727 @heading powerpc-*-eabisim
4728 Embedded PowerPC system in big endian mode for use in running under the
4734 @anchor{powerpc-x-eabi}
4735 @heading powerpc-*-eabi
4736 Embedded PowerPC system in big endian mode.
4741 @anchor{powerpcle-x-elf}
4742 @heading powerpcle-*-elf
4743 PowerPC system in little endian mode, running System V.4.
4748 @anchor{powerpcle-x-eabisim}
4749 @heading powerpcle-*-eabisim
4750 Embedded PowerPC system in little endian mode for use in running under
4756 @anchor{powerpcle-x-eabi}
4757 @heading powerpcle-*-eabi
4758 Embedded PowerPC system in little endian mode.
4765 The Renesas RL78 processor.
4766 This configuration is intended for embedded systems.
4771 @anchor{riscv32-x-elf}
4772 @heading riscv32-*-elf
4773 The RISC-V RV32 instruction set.
4774 This configuration is intended for embedded systems.
4775 This (and all other RISC-V) targets require the binutils 2.30 release.
4780 @anchor{riscv32-x-linux}
4781 @heading riscv32-*-linux
4782 The RISC-V RV32 instruction set running GNU/Linux.
4783 This (and all other RISC-V) targets require the binutils 2.30 release.
4788 @anchor{riscv64-x-elf}
4789 @heading riscv64-*-elf
4790 The RISC-V RV64 instruction set.
4791 This configuration is intended for embedded systems.
4792 This (and all other RISC-V) targets require the binutils 2.30 release.
4797 @anchor{riscv64-x-linux}
4798 @heading riscv64-*-linux
4799 The RISC-V RV64 instruction set running GNU/Linux.
4800 This (and all other RISC-V) targets require the binutils 2.30 release.
4807 The Renesas RX processor.
4812 @anchor{s390-x-linux}
4813 @heading s390-*-linux*
4814 S/390 system running GNU/Linux for S/390@.
4819 @anchor{s390x-x-linux}
4820 @heading s390x-*-linux*
4821 zSeries system (64-bit) running GNU/Linux for zSeries@.
4826 @anchor{s390x-ibm-tpf}
4827 @heading s390x-ibm-tpf*
4828 zSeries system (64-bit) running TPF@. This platform is
4829 supported as cross-compilation target only.
4834 @c Please use Solaris 2 to refer to all release of Solaris, starting
4835 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
4836 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
4837 @c alone is too unspecific and must be avoided.
4838 @anchor{x-x-solaris2}
4839 @heading *-*-solaris2*
4840 Support for Solaris 10 has been removed in GCC 10. Support for Solaris
4841 9 has been removed in GCC 5. Support for Solaris 8 has been removed in
4842 GCC 4.8. Support for Solaris 7 has been removed in GCC 4.6.
4844 Solaris 11.3 provides GCC 4.5.2, 4.7.3, and 4.8.2 as
4845 @command{/usr/gcc/4.5/bin/gcc} or similar. Newer Solaris versions
4846 provide one or more of GCC 5, 7, and 9. Alternatively,
4847 you can install a pre-built GCC to bootstrap and install GCC. See the
4848 @uref{binaries.html,,binaries page} for details.
4850 The Solaris 2 @command{/bin/sh} will often fail to configure
4851 @samp{libstdc++-v3}. We therefore recommend using the
4852 following initial sequence of commands
4855 % CONFIG_SHELL=/bin/ksh
4856 % export CONFIG_SHELL
4860 and proceed as described in @uref{configure.html,,the configure instructions}.
4861 In addition we strongly recommend specifying an absolute path to invoke
4862 @command{@var{srcdir}/configure}.
4864 In Solaris 11, you need to check for @code{system/header},
4865 @code{system/linker}, and @code{developer/assembler} packages.
4867 Trying to use the linker and other tools in
4868 @file{/usr/ucb} to install GCC has been observed to cause trouble.
4869 For example, the linker may hang indefinitely. The fix is to remove
4870 @file{/usr/ucb} from your @env{PATH}.
4872 The build process works more smoothly with the legacy Solaris tools so, if you
4873 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
4874 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
4876 We recommend the use of the Solaris assembler or the GNU assembler, in
4877 conjunction with the Solaris linker. The GNU @command{as}
4878 versions included in Solaris 11.3,
4879 from GNU binutils 2.23.1 or newer (in @file{/usr/bin/gas} and
4880 @file{/usr/gnu/bin/as}), are known to work.
4881 The current version, from GNU binutils 2.34,
4882 is known to work as well. Note that your mileage may vary
4883 if you use a combination of the GNU tools and the Solaris tools: while the
4884 combination GNU @command{as} + Solaris @command{ld} should reasonably work,
4885 the reverse combination Solaris @command{as} + GNU @command{ld} may fail to
4886 build or cause memory corruption at runtime in some cases for C++ programs.
4888 GNU @command{ld} usually works as well. Again, the current
4889 version (2.34) is known to work, but generally lacks platform specific
4890 features, so better stay with Solaris @command{ld}. To use the LTO linker
4891 plugin (@option{-fuse-linker-plugin}) with GNU @command{ld}, GNU
4892 binutils @emph{must} be configured with @option{--enable-largefile}.
4894 To enable symbol versioning in @samp{libstdc++} with the Solaris linker,
4895 you need to have any version of GNU @command{c++filt}, which is part of
4896 GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no
4897 appropriate version is found. Solaris @command{c++filt} from the Solaris
4898 Studio compilers does @emph{not} work.
4900 In order to build the GNU D compiler, GDC, a working @samp{libphobos} is
4901 needed. That library wasn't built by default in GCC 9--11 on SPARC, or
4902 on x86 when the Solaris assembler is used, but can be enabled by
4903 configuring with @option{--enable-libphobos}. Also, GDC 9.4.0 is
4904 required on x86, while GDC 9.3.0 is known to work on SPARC.
4906 The versions of the GNU Multiple Precision Library (GMP), the MPFR
4907 library and the MPC library bundled with Solaris 11.3 and later are
4908 usually recent enough to match GCC's requirements. There are two
4913 While the version of the GMP library in Solaris 11.3 works with GCC, you
4914 need to configure with @option{--with-gmp-include=/usr/include/gmp}.
4917 The version of the MPFR libary included in Solaris 11.3 is too old; you
4918 need to provide a more recent one.
4927 This section contains general configuration information for all
4928 SPARC-based platforms. In addition to reading this section, please
4929 read all other sections that match your target.
4931 Newer versions of the GNU Multiple Precision Library (GMP), the MPFR
4932 library and the MPC library are known to be miscompiled by earlier
4933 versions of GCC on these platforms. We therefore recommend the use
4934 of the exact versions of these libraries listed as minimal versions
4935 in @uref{prerequisites.html,,the prerequisites}.
4940 @anchor{sparc-sun-solaris2}
4941 @heading sparc-sun-solaris2*
4942 When GCC is configured to use GNU binutils 2.14 or later, the binaries
4943 produced are smaller than the ones produced using Solaris native tools;
4944 this difference is quite significant for binaries containing debugging
4947 Starting with Solaris 7, the operating system is capable of executing
4948 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
4949 this; the @option{-m64} option enables 64-bit code generation.
4950 However, if all you want is code tuned for the UltraSPARC CPU, you
4951 should try the @option{-mtune=ultrasparc} option instead, which produces
4952 code that, unlike full 64-bit code, can still run on non-UltraSPARC
4955 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4956 library or the MPC library on a Solaris 7 or later system, the canonical
4957 target triplet must be specified as the @command{build} parameter on the
4958 configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and
4959 not that of GMP or MPFR or MPC). For example on a Solaris 11 system:
4962 % ./configure --build=sparc-sun-solaris2.11 --prefix=xxx
4968 @anchor{sparc-x-linux}
4969 @heading sparc-*-linux*
4974 @anchor{sparc64-x-solaris2}
4975 @heading sparc64-*-solaris2*
4976 When configuring a 64-bit-default GCC on Solaris/SPARC, you must use a
4977 build compiler that generates 64-bit code, either by default or by
4978 specifying @samp{CC='gcc -m64' CXX='gcc-m64'} to @command{configure}.
4979 Additionally, you @emph{must} pass @option{--build=sparc64-sun-solaris2.11}
4980 or @option{--build=sparcv9-sun-solaris2.11} because @file{config.guess}
4981 misdetects this situation, which can cause build failures.
4983 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4984 library or the MPC library, the canonical target triplet must be specified
4985 as the @command{build} parameter on the configure line. For example
4986 on a Solaris 11 system:
4989 % ./configure --build=sparc64-sun-solaris2.11 --prefix=xxx
4995 @anchor{sparcv9-x-solaris2}
4996 @heading sparcv9-*-solaris2*
4997 This is a synonym for @samp{sparc64-*-solaris2*}.
5004 The C6X family of processors. This port requires binutils-2.22 or newer.
5009 @anchor{visium-x-elf}
5010 @heading visium-*-elf
5011 CDS VISIUMcore processor.
5012 This configuration is intended for embedded systems.
5017 @anchor{x-x-vxworks}
5018 @heading *-*-vxworks*
5019 Support for VxWorks is in flux. At present GCC supports @emph{only} the
5020 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
5021 We welcome patches for other architectures supported by VxWorks 5.5.
5022 Support for VxWorks AE would also be welcome; we believe this is merely
5023 a matter of writing an appropriate ``configlette'' (see below). We are
5024 not interested in supporting older, a.out or COFF-based, versions of
5027 VxWorks comes with an older version of GCC installed in
5028 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
5029 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
5030 Before running @command{configure}, create the directories @file{@var{prefix}}
5031 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
5032 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
5033 include that directory while running both @command{configure} and
5036 You must give @command{configure} the
5037 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
5038 find the VxWorks system headers. Since VxWorks is a cross compilation
5039 target only, you must also specify @option{--target=@var{target}}.
5040 @command{configure} will attempt to create the directory
5041 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
5042 make sure the user running @command{configure} has sufficient privilege
5045 GCC's exception handling runtime requires a special ``configlette''
5046 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
5047 that file to add the module to your kernel build. (Future versions of
5048 VxWorks will incorporate this module.)
5054 @heading x86_64-*-*, amd64-*-*
5055 GCC supports the x86-64 architecture implemented by the AMD64 processor
5056 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
5057 On GNU/Linux the default is a bi-arch compiler which is able to generate
5058 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
5063 @anchor{x86-64-x-solaris2}
5064 @heading x86_64-*-solaris2*
5065 GCC also supports the x86-64 architecture implemented by the AMD64
5066 processor (@samp{amd64-*-*} is an alias for @samp{x86_64-*-*}) on
5067 Solaris 10 or later. Unlike other systems, without special options a
5068 bi-arch compiler is built which generates 32-bit code by default, but
5069 can generate 64-bit x86-64 code with the @option{-m64} switch. Since
5070 GCC 4.7, there is also a configuration that defaults to 64-bit code, but
5071 can generate 32-bit code with @option{-m32}. To configure and build
5072 this way, you have to provide all support libraries like @file{libgmp}
5073 as 64-bit code, configure with @option{--target=x86_64-pc-solaris2.11}
5074 and @samp{CC=gcc -m64}.
5079 @anchor{xtensa-x-elf}
5080 @heading xtensa*-*-elf
5081 This target is intended for embedded Xtensa systems using the
5082 @samp{newlib} C library. It uses ELF but does not support shared
5083 objects. Designed-defined instructions specified via the
5084 Tensilica Instruction Extension (TIE) language are only supported
5085 through inline assembly.
5087 The Xtensa configuration information must be specified prior to
5088 building GCC@. The @file{include/xtensa-config.h} header
5089 file contains the configuration information. If you created your
5090 own Xtensa configuration with the Xtensa Processor Generator, the
5091 downloaded files include a customized copy of this header file,
5092 which you can use to replace the default header file.
5097 @anchor{xtensa-x-linux}
5098 @heading xtensa*-*-linux*
5099 This target is for Xtensa systems running GNU/Linux. It supports ELF
5100 shared objects and the GNU C library (glibc). It also generates
5101 position-independent code (PIC) regardless of whether the
5102 @option{-fpic} or @option{-fPIC} options are used. In other
5103 respects, this target is the same as the
5104 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
5110 @heading Microsoft Windows
5112 @subheading Intel 16-bit versions
5113 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
5116 However, the 32-bit port has limited support for Microsoft
5117 Windows 3.11 in the Win32s environment, as a target only. See below.
5119 @subheading Intel 32-bit versions
5120 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
5121 XP, and Windows Vista, are supported by several different target
5122 platforms. These targets differ in which Windows subsystem they target
5123 and which C libraries are used.
5126 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
5127 Linux API emulation layer in the Win32 subsystem.
5128 @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for
5129 the Win32 subsystem that provides a subset of POSIX.
5130 @item MKS i386-pc-mks: NuTCracker from MKS. See
5131 @uref{https://www.mkssoftware.com} for more information.
5134 @subheading Intel 64-bit versions
5135 GCC contains support for x86-64 using the mingw-w64
5136 runtime library, available from @uref{https://www.mingw-w64.org/downloads/}.
5137 This library should be used with the target triple x86_64-pc-mingw32.
5139 @subheading Windows CE
5140 Windows CE is supported as a target only on Hitachi
5141 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
5143 @subheading Other Windows Platforms
5144 GCC no longer supports Windows NT on the Alpha or PowerPC.
5146 GCC no longer supports the Windows POSIX subsystem. However, it does
5147 support the Interix subsystem. See above.
5149 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
5151 PW32 (i386-pc-pw32) support was never completed, and the project seems to
5152 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
5154 UWIN support has been removed due to a lack of maintenance.
5161 Ports of GCC are included with the
5162 @uref{http://www.cygwin.com/,,Cygwin environment}.
5164 GCC will build under Cygwin without modification; it does not build
5165 with Microsoft's C++ compiler and there are no plans to make it do so.
5167 The Cygwin native compiler can be configured to target any 32-bit x86
5168 cpu architecture desired; the default is i686-pc-cygwin. It should be
5169 used with as up-to-date a version of binutils as possible; use either
5170 the latest official GNU binutils release in the Cygwin distribution,
5171 or version 2.20 or above if building your own.
5176 @anchor{x-x-mingw32}
5177 @heading *-*-mingw32
5178 GCC will build with and support only MinGW runtime 3.12 and later.
5179 Earlier versions of headers are incompatible with the new default semantics
5180 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
5182 To support emitting DWARF debugging info you need to use GNU binutils
5183 version 2.16 or above containing support for the @code{.secrel32}
5184 assembler pseudo-op.
5190 @heading Older systems
5191 GCC contains support files for many older (1980s and early
5192 1990s) Unix variants. For the most part, support for these systems
5193 has not been deliberately removed, but it has not been maintained for
5194 several years and may suffer from bitrot.
5196 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
5197 Support for these systems is still present in that release, but
5198 @command{configure} will fail unless the @option{--enable-obsolete}
5199 option is given. Unless a maintainer steps forward, support for these
5200 systems will be removed from the next release of GCC@.
5202 Support for old systems as hosts for GCC can cause problems if the
5203 workarounds for compiler, library and operating system bugs affect the
5204 cleanliness or maintainability of the rest of GCC@. In some cases, to
5205 bring GCC up on such a system, if still possible with current GCC, may
5206 require first installing an old version of GCC which did work on that
5207 system, and using it to compile a more recent GCC, to avoid bugs in the
5208 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
5209 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
5210 sites}. Header bugs may generally be avoided using
5211 @command{fixincludes}, but bugs or deficiencies in libraries and the
5212 operating system may still cause problems.
5214 Support for older systems as targets for cross-compilation is less
5215 problematic than support for them as hosts for GCC; if an enthusiast
5216 wishes to make such a target work again (including resurrecting any of
5217 the targets that never worked with GCC 2, starting from the last
5218 version before they were removed), patches
5219 @uref{../contribute.html,,following the usual requirements} would be
5220 likely to be accepted, since they should not affect the support for more
5223 For some systems, old versions of GNU binutils may also be useful,
5224 and are available from @file{pub/binutils/old-releases} on
5225 @uref{https://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
5227 Some of the information on specific systems above relates to
5228 such older systems, but much of the information
5229 about GCC on such systems (which may no longer be applicable to
5230 current GCC) is to be found in the GCC texinfo manual.
5236 @heading all ELF targets (SVR4, Solaris 2, etc.)
5237 C++ support is significantly better on ELF targets if you use the
5238 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
5239 inlines, vtables and template instantiations will be discarded
5248 @uref{./index.html,,Return to the GCC Installation page}
5252 @c ***GFDL********************************************************************
5260 @uref{./index.html,,Return to the GCC Installation page}
5264 @c ***************************************************************************
5265 @c Part 6 The End of the Document
5267 @comment node-name, next, previous, up
5268 @node Concept Index, , GNU Free Documentation License, Top
5272 @unnumbered Concept Index