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: Old documentation
43 @settitle Installing GCC: GNU Free Documentation License
46 @c Copyright (C) 1988-2021 Free Software Foundation, Inc.
47 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
49 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
50 @c test the generation of HTML documents for the gcc.gnu.org web pages.
52 @c Do not use @footnote{} in this file as it breaks install.texi2html!
54 @c Include everything if we're not making html
58 @set prerequisiteshtml
69 @c Part 2 Summary Description and Copyright
71 Copyright @copyright{} 1988-2021 Free Software Foundation, Inc.
73 Permission is granted to copy, distribute and/or modify this document
74 under the terms of the GNU Free Documentation License, Version 1.3 or
75 any later version published by the Free Software Foundation; with no
76 Invariant Sections, the Front-Cover texts being (a) (see below), and
77 with the Back-Cover Texts being (b) (see below). A copy of the
78 license is included in the section entitled ``@uref{./gfdl.html,,GNU
79 Free Documentation License}''.
81 (a) The FSF's Front-Cover Text is:
85 (b) The FSF's Back-Cover Text is:
87 You have freedom to copy and modify this GNU Manual, like GNU
88 software. Copies published by the Free Software Foundation raise
89 funds for GNU development.
94 @dircategory Software development
96 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
99 @c Part 3 Titlepage and Copyright
101 @title Installing GCC
104 @c The following two commands start the copyright page.
106 @vskip 0pt plus 1filll
110 @c Part 4 Top node, Master Menu, and/or Table of Contents
113 @comment node-name, next, Previous, up
116 * Installing GCC:: This document describes the generic installation
117 procedure for GCC as well as detailing some target
118 specific installation instructions.
120 * Specific:: Host/target specific installation notes for GCC.
121 * Binaries:: Where to get pre-compiled binaries.
123 * Old:: Old installation documentation.
125 * GNU Free Documentation License:: How you can copy and share this manual.
126 * Concept Index:: This index has two entries.
134 @c Part 5 The Body of the Document
135 @c ***Installing GCC**********************************************************
137 @comment node-name, next, previous, up
138 @node Installing GCC, Binaries, , Top
142 @chapter Installing GCC
145 The latest version of this document is always available at
146 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
147 It refers to the current development sources, instructions for
148 specific released versions are included with the sources.
150 This document describes the generic installation procedure for GCC as well
151 as detailing some target specific installation instructions.
153 GCC includes several components that previously were separate distributions
154 with their own installation instructions. This document supersedes all
155 package-specific installation instructions.
157 @emph{Before} starting the build/install procedure please check the
159 @ref{Specific, host/target specific installation notes}.
162 @uref{specific.html,,host/target specific installation notes}.
164 We recommend you browse the entire generic installation instructions before
167 Lists of successful builds for released versions of GCC are
168 available at @uref{http://gcc.gnu.org/buildstat.html}.
169 These lists are updated as new information becomes available.
171 The installation procedure itself is broken into five steps.
176 * Downloading the source::
179 * Testing:: (optional)
186 @uref{prerequisites.html,,Prerequisites}
188 @uref{download.html,,Downloading the source}
190 @uref{configure.html,,Configuration}
192 @uref{build.html,,Building}
194 @uref{test.html,,Testing} (optional)
196 @uref{finalinstall.html,,Final install}
200 Please note that GCC does not support @samp{make uninstall} and probably
201 won't do so in the near future as this would open a can of worms. Instead,
202 we suggest that you install GCC into a directory of its own and simply
203 remove that directory when you do not need that specific version of GCC
204 any longer, and, if shared libraries are installed there as well, no
205 more binaries exist that use them.
208 There are also some @uref{old.html,,old installation instructions},
209 which are mostly obsolete but still contain some information which has
210 not yet been merged into the main part of this manual.
218 @uref{./index.html,,Return to the GCC Installation page}
224 @c ***Prerequisites**************************************************
226 @comment node-name, next, previous, up
227 @node Prerequisites, Downloading the source, , Installing GCC
229 @ifset prerequisiteshtml
231 @chapter Prerequisites
233 @cindex Prerequisites
235 GCC requires that various tools and packages be available for use in the
236 build procedure. Modifying GCC sources requires additional tools
239 @heading Tools/packages necessary for building GCC
241 @item ISO C++11 compiler
242 Necessary to bootstrap GCC.
244 Versions of GCC prior to 11 also allow bootstrapping with an ISO C++98
245 compiler, versions of GCC prior to 4.8 also allow bootstrapping with a
246 ISO C89 compiler, and versions of GCC prior to 3.4 also allow
247 bootstrapping with a traditional (K&R) C compiler.
249 To build all languages in a cross-compiler or other configuration where
250 3-stage bootstrap is not performed, you need to start with an existing
251 GCC binary (version 4.8 or later) because source code for language
252 frontends other than C might use GCC extensions.
254 Note that to bootstrap GCC with versions of GCC earlier than 4.8, you
255 may need to use @option{--disable-stage1-checking}, though
256 bootstrapping the compiler with such earlier compilers is strongly
259 @item C standard library and headers
261 In order to build GCC, the C standard library and headers must be present
262 for all target variants for which target libraries will be built (and not
263 only the variant of the host C++ compiler).
265 This affects the popular @samp{x86_64-pc-linux-gnu} platform (among
266 other multilib targets), for which 64-bit (@samp{x86_64}) and 32-bit
267 (@samp{i386}) libc headers are usually packaged separately. If you do a
268 build of a native compiler on @samp{x86_64-pc-linux-gnu}, make sure you
269 either have the 32-bit libc developer package properly installed (the exact
270 name of the package depends on your distro) or you must build GCC as a
271 64-bit only compiler by configuring with the option
272 @option{--disable-multilib}. Otherwise, you may encounter an error such as
273 @samp{fatal error: gnu/stubs-32.h: No such file}
275 @item @anchor{GNAT-prerequisite}GNAT
277 In order to build GNAT, the Ada compiler, you need a working GNAT
278 compiler (GCC version 4.7 or later).
280 This includes GNAT tools such as @command{gnatmake} and
281 @command{gnatlink}, since the Ada front end is written in Ada and
282 uses some GNAT-specific extensions.
284 In order to build a cross compiler, it is strongly recommended to install
285 the new compiler as native first, and then use it to build the cross
286 compiler. Other native compiler versions may work but this is not guaranteed and
287 will typically fail with hard to understand compilation errors during the
290 Similarly, it is strongly recommended to use an older version of GNAT to build
291 GNAT. More recent versions of GNAT than the version built are not guaranteed
292 to work and will often fail during the build with compilation errors.
294 Note that @command{configure} does not test whether the GNAT installation works
295 and has a sufficiently recent version; if too old a GNAT version is
296 installed and @option{--enable-languages=ada} is used, the build will fail.
298 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
299 must not be set when building the Ada compiler, the Ada tools, or the
300 Ada runtime libraries. You can check that your build environment is clean
301 by verifying that @samp{gnatls -v} lists only one explicit path in each
304 @item A ``working'' POSIX compatible shell, or GNU bash
306 Necessary when running @command{configure} because some
307 @command{/bin/sh} shells have bugs and may crash when configuring the
308 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
309 have disastrous corner-case performance problems. This
310 can cause target @command{configure} runs to literally take days to
311 complete in some cases.
313 So on some platforms @command{/bin/ksh} is sufficient, on others it
314 isn't. See the host/target specific instructions for your platform, or
315 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
316 environment to your ``good'' shell prior to running
317 @command{configure}/@command{make}.
319 @command{zsh} is not a fully compliant POSIX shell and will not
320 work when configuring GCC@.
322 @item A POSIX or SVR4 awk
324 Necessary for creating some of the generated source files for GCC@.
325 If in doubt, use a recent GNU awk version, as some of the older ones
326 are broken. GNU awk version 3.1.5 is known to work.
330 Necessary in some circumstances, optional in others. See the
331 host/target specific instructions for your platform for the exact
334 @item gzip version 1.2.4 (or later) or
335 @itemx bzip2 version 1.0.2 (or later)
337 Necessary to uncompress GCC @command{tar} files when source code is
338 obtained via HTTPS mirror sites.
340 @item GNU make version 3.80 (or later)
342 You must have GNU make installed to build GCC@.
344 @item GNU tar version 1.14 (or later)
346 Necessary (only on some platforms) to untar the source code. Many
347 systems' @command{tar} programs will also work, only try GNU
348 @command{tar} if you have problems.
350 @item Perl version between 5.6.1 and 5.6.24
352 Necessary when targeting Darwin, building @samp{libstdc++},
353 and not using @option{--disable-symvers}.
354 Necessary when targeting Solaris 2 with Solaris @command{ld} and not using
355 @option{--disable-symvers}.
357 Necessary when regenerating @file{Makefile} dependencies in libiberty.
358 Necessary when regenerating @file{libiberty/functions.texi}.
359 Necessary when generating manpages from Texinfo manuals.
360 Used by various scripts to generate some files included in the source
361 repository (mainly Unicode-related and rarely changing) from source
364 Used by @command{automake}.
368 Several support libraries are necessary to build GCC, some are required,
369 others optional. While any sufficiently new version of required tools
370 usually work, library requirements are generally stricter. Newer
371 versions may work in some cases, but it's safer to use the exact
372 versions documented. We appreciate bug reports about problems with
373 newer versions, though. If your OS vendor provides packages for the
374 support libraries then using those packages may be the simplest way to
375 install the libraries.
378 @item GNU Multiple Precision Library (GMP) version 4.3.2 (or later)
380 Necessary to build GCC@. If a GMP source distribution is found in a
381 subdirectory of your GCC sources named @file{gmp}, it will be built
382 together with GCC. Alternatively, if GMP is already installed but it
383 is not in your library search path, you will have to configure with the
384 @option{--with-gmp} configure option. See also @option{--with-gmp-lib}
385 and @option{--with-gmp-include}.
386 The in-tree build is only supported with the GMP version that
387 download_prerequisites installs.
389 @item MPFR Library version 3.1.0 (or later)
391 Necessary to build GCC@. It can be downloaded from
392 @uref{https://www.mpfr.org}. If an MPFR source distribution is found
393 in a subdirectory of your GCC sources named @file{mpfr}, it will be
394 built together with GCC. Alternatively, if MPFR is already installed
395 but it is not in your default library search path, the
396 @option{--with-mpfr} configure option should be used. See also
397 @option{--with-mpfr-lib} and @option{--with-mpfr-include}.
398 The in-tree build is only supported with the MPFR version that
399 download_prerequisites installs.
401 @item MPC Library version 1.0.1 (or later)
403 Necessary to build GCC@. It can be downloaded from
404 @uref{http://www.multiprecision.org/mpc/}. If an MPC source distribution
405 is found in a subdirectory of your GCC sources named @file{mpc}, it
406 will be built together with GCC. Alternatively, if MPC is already
407 installed but it is not in your default library search path, the
408 @option{--with-mpc} configure option should be used. See also
409 @option{--with-mpc-lib} and @option{--with-mpc-include}.
410 The in-tree build is only supported with the MPC version that
411 download_prerequisites installs.
413 @item isl Library version 0.15 or later.
415 Necessary to build GCC with the Graphite loop optimizations.
416 It can be downloaded from @uref{https://gcc.gnu.org/pub/gcc/infrastructure/}.
417 If an isl source distribution is found
418 in a subdirectory of your GCC sources named @file{isl}, it will be
419 built together with GCC. Alternatively, the @option{--with-isl} configure
420 option should be used if isl is not installed in your default library
425 Necessary to build GCC with zstd compression used for LTO bytecode.
426 The library is searched in your default library patch search.
427 Alternatively, the @option{--with-zstd} configure option should be used.
431 @heading Tools/packages necessary for modifying GCC
433 @item autoconf version 2.69
434 @itemx GNU m4 version 1.4.6 (or later)
436 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
437 to regenerate @file{configure} and @file{config.in} files.
439 @item automake version 1.15.1
441 Necessary when modifying a @file{Makefile.am} file to regenerate its
442 associated @file{Makefile.in}.
444 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
445 file. Specifically this applies to the @file{gcc}, @file{intl},
446 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
447 as any of their subdirectories.
449 For directories that use automake, GCC requires the latest release in
450 the 1.15 series, which is currently 1.15.1. When regenerating a directory
451 to a newer version, please update all the directories using an older 1.15
452 to the latest released version.
454 @item gettext version 0.14.5 (or later)
456 Needed to regenerate @file{gcc.pot}.
458 @item gperf version 2.7.2 (or later)
460 Necessary when modifying @command{gperf} input files, e.g.@:
461 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
462 @file{gcc/cp/cfns.h}.
467 @c Once Tcl 8.5 or higher is required, remove any obsolete
468 @c compatibility workarounds:
469 @c git grep 'compatibility with earlier Tcl releases'
471 Necessary to run the GCC testsuite; see the section on testing for
474 @item autogen version 5.5.4 (or later) and
475 @itemx guile version 1.4.1 (or later)
477 Necessary to regenerate @file{fixinc/fixincl.x} from
478 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
480 Necessary to run @samp{make check} for @file{fixinc}.
482 Necessary to regenerate the top level @file{Makefile.in} file from
483 @file{Makefile.tpl} and @file{Makefile.def}.
485 @item Flex version 2.5.4 (or later)
487 Necessary when modifying @file{*.l} files.
489 Necessary to build GCC during development because the generated output
490 files are not included in the version-controlled source repository.
491 They are included in releases.
493 @item Texinfo version 4.7 (or later)
495 Necessary for running @command{makeinfo} when modifying @file{*.texi}
496 files to test your changes.
498 Necessary for running @command{make dvi} or @command{make pdf} to
499 create printable documentation in DVI or PDF format. Texinfo version
500 4.8 or later is required for @command{make pdf}.
502 Necessary to build GCC documentation during development because the
503 generated output files are not included in the repository. They are
504 included in releases.
506 @item @TeX{} (any working version)
508 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
509 are used when running @command{make dvi} or @command{make pdf} to create
510 DVI or PDF files, respectively.
512 @item Sphinx version 1.0 (or later)
514 Necessary to regenerate @file{jit/docs/_build/texinfo} from the @file{.rst}
515 files in the directories below @file{jit/docs}.
517 @item git (any version)
518 @itemx SSH (any version)
520 Necessary to access the source repository. Public releases and weekly
521 snapshots of the development sources are also available via HTTPS@.
523 @item GNU diffutils version 2.7 (or later)
525 Useful when submitting patches for the GCC source code.
527 @item patch version 2.5.4 (or later)
529 Necessary when applying patches, created with @command{diff}, to one's
539 @uref{./index.html,,Return to the GCC Installation page}
543 @c ***Downloading the source**************************************************
545 @comment node-name, next, previous, up
546 @node Downloading the source, Configuration, Prerequisites, Installing GCC
550 @chapter Downloading GCC
552 @cindex Downloading GCC
553 @cindex Downloading the Source
555 GCC is distributed via @uref{http://gcc.gnu.org/git.html,,git} and via
556 HTTPS as tarballs compressed with @command{gzip} or @command{bzip2}.
558 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
559 for information on how to obtain GCC@.
561 The source distribution includes the C, C++, Objective-C, Fortran,
562 and Ada (in the case of GCC 3.1 and later) compilers, as well as
563 runtime libraries for C++, Objective-C, and Fortran.
564 For previous versions these were downloadable as separate components such
565 as the core GCC distribution, which included the C language front end and
566 shared components, and language-specific distributions including the
567 language front end and the language runtime (where appropriate).
569 If you also intend to build binutils (either to upgrade an existing
570 installation or for use in place of the corresponding tools of your
571 OS), unpack the binutils distribution either in the same directory or
572 a separate one. In the latter case, add symbolic links to any
573 components of the binutils you intend to build alongside the compiler
574 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
575 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
577 Likewise the GMP, MPFR and MPC libraries can be automatically built
578 together with GCC. You may simply run the
579 @command{contrib/download_prerequisites} script in the GCC source directory
580 to set up everything.
581 Otherwise unpack the GMP, MPFR and/or MPC source
582 distributions in the directory containing the GCC sources and rename
583 their directories to @file{gmp}, @file{mpfr} and @file{mpc},
584 respectively (or use symbolic links with the same name).
591 @uref{./index.html,,Return to the GCC Installation page}
595 @c ***Configuration***********************************************************
597 @comment node-name, next, previous, up
598 @node Configuration, Building, Downloading the source, Installing GCC
602 @chapter Installing GCC: Configuration
604 @cindex Configuration
605 @cindex Installing GCC: Configuration
607 Like most GNU software, GCC must be configured before it can be built.
608 This document describes the recommended configuration procedure
609 for both native and cross targets.
611 We use @var{srcdir} to refer to the toplevel source directory for
612 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
614 If you obtained the sources by cloning the repository, @var{srcdir}
615 must refer to the top @file{gcc} directory, the one where the
616 @file{MAINTAINERS} file can be found, and not its @file{gcc}
617 subdirectory, otherwise the build will fail.
619 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
620 file system, the shell's built-in @command{pwd} command will return
621 temporary pathnames. Using these can lead to various sorts of build
622 problems. To avoid this issue, set the @env{PWDCMD} environment
623 variable to an automounter-aware @command{pwd} command, e.g.,
624 @command{pawd} or @samp{amq -w}, during the configuration and build
627 First, we @strong{highly} recommend that GCC be built into a
628 separate directory from the sources which does @strong{not} reside
629 within the source tree. This is how we generally build GCC; building
630 where @var{srcdir} == @var{objdir} should still work, but doesn't
631 get extensive testing; building where @var{objdir} is a subdirectory
632 of @var{srcdir} is unsupported.
634 If you have previously built GCC in the same directory for a
635 different target machine, do @samp{make distclean} to delete all files
636 that might be invalid. One of the files this deletes is @file{Makefile};
637 if @samp{make distclean} complains that @file{Makefile} does not exist
638 or issues a message like ``don't know how to make distclean'' it probably
639 means that the directory is already suitably clean. However, with the
640 recommended method of building in a separate @var{objdir}, you should
641 simply use a different @var{objdir} for each target.
643 Second, when configuring a native system, either @command{cc} or
644 @command{gcc} must be in your path or you must set @env{CC} in
645 your environment before running configure. Otherwise the configuration
649 Note that the bootstrap compiler and the resulting GCC must be link
650 compatible, else the bootstrap will fail with linker errors about
651 incompatible object file formats. Several multilibed targets are
652 affected by this requirement, see
654 @ref{Specific, host/target specific installation notes}.
657 @uref{specific.html,,host/target specific installation notes}.
666 % @var{srcdir}/configure [@var{options}] [@var{target}]
669 @heading Distributor options
671 If you will be distributing binary versions of GCC, with modifications
672 to the source code, you should use the options described in this
673 section to make clear that your version contains modifications.
676 @item --with-pkgversion=@var{version}
677 Specify a string that identifies your package. You may wish
678 to include a build number or build date. This version string will be
679 included in the output of @command{gcc --version}. This suffix does
680 not replace the default version string, only the @samp{GCC} part.
682 The default value is @samp{GCC}.
684 @item --with-bugurl=@var{url}
685 Specify the URL that users should visit if they wish to report a bug.
686 You are of course welcome to forward bugs reported to you to the FSF,
687 if you determine that they are not bugs in your modifications.
689 The default value refers to the FSF's GCC bug tracker.
691 @item --with-documentation-root-url=@var{url}
692 Specify the URL root that contains GCC option documentation. The @var{url}
693 should end with a @code{/} character.
695 The default value is @uref{https://gcc.gnu.org/onlinedocs/,,https://gcc.gnu.org/onlinedocs/}.
697 @item --with-changes-root-url=@var{url}
698 Specify the URL root that contains information about changes in GCC
699 releases like @code{gcc-@var{version}/changes.html}.
700 The @var{url} should end with a @code{/} character.
702 The default value is @uref{https://gcc.gnu.org/,,https://gcc.gnu.org/}.
706 @heading Target specification
709 GCC has code to correctly determine the correct value for @var{target}
710 for nearly all native systems. Therefore, we highly recommend you do
711 not provide a configure target when configuring a native compiler.
714 @var{target} must be specified as @option{--target=@var{target}}
715 when configuring a cross compiler; examples of valid targets would be
716 m68k-elf, sh-elf, etc.
719 Specifying just @var{target} instead of @option{--target=@var{target}}
720 implies that the host defaults to @var{target}.
724 @heading Options specification
726 Use @var{options} to override several configure time options for
727 GCC@. A list of supported @var{options} follows; @samp{configure
728 --help} may list other options, but those not listed below may not
729 work and should not normally be used.
731 Note that each @option{--enable} option has a corresponding
732 @option{--disable} option and that each @option{--with} option has a
733 corresponding @option{--without} option.
736 @item --prefix=@var{dirname}
737 Specify the toplevel installation
738 directory. This is the recommended way to install the tools into a directory
739 other than the default. The toplevel installation directory defaults to
742 We @strong{highly} recommend against @var{dirname} being the same or a
743 subdirectory of @var{objdir} or vice versa. If specifying a directory
744 beneath a user's home directory tree, some shells will not expand
745 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
748 The following standard @command{autoconf} options are supported. Normally you
749 should not need to use these options.
751 @item --exec-prefix=@var{dirname}
752 Specify the toplevel installation directory for architecture-dependent
753 files. The default is @file{@var{prefix}}.
755 @item --bindir=@var{dirname}
756 Specify the installation directory for the executables called by users
757 (such as @command{gcc} and @command{g++}). The default is
758 @file{@var{exec-prefix}/bin}.
760 @item --libdir=@var{dirname}
761 Specify the installation directory for object code libraries and
762 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
764 @item --libexecdir=@var{dirname}
765 Specify the installation directory for internal executables of GCC@.
766 The default is @file{@var{exec-prefix}/libexec}.
768 @item --with-slibdir=@var{dirname}
769 Specify the installation directory for the shared libgcc library. The
770 default is @file{@var{libdir}}.
772 @item --datarootdir=@var{dirname}
773 Specify the root of the directory tree for read-only architecture-independent
774 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
776 @item --infodir=@var{dirname}
777 Specify the installation directory for documentation in info format.
778 The default is @file{@var{datarootdir}/info}.
780 @item --datadir=@var{dirname}
781 Specify the installation directory for some architecture-independent
782 data files referenced by GCC@. The default is @file{@var{datarootdir}}.
784 @item --docdir=@var{dirname}
785 Specify the installation directory for documentation files (other
786 than Info) for GCC@. The default is @file{@var{datarootdir}/doc}.
788 @item --htmldir=@var{dirname}
789 Specify the installation directory for HTML documentation files.
790 The default is @file{@var{docdir}}.
792 @item --pdfdir=@var{dirname}
793 Specify the installation directory for PDF documentation files.
794 The default is @file{@var{docdir}}.
796 @item --mandir=@var{dirname}
797 Specify the installation directory for manual pages. The default is
798 @file{@var{datarootdir}/man}. (Note that the manual pages are only extracts
799 from the full GCC manuals, which are provided in Texinfo format. The manpages
800 are derived by an automatic conversion process from parts of the full
803 @item --with-gxx-include-dir=@var{dirname}
805 the installation directory for G++ header files. The default depends
806 on other configuration options, and differs between cross and native
809 @item --with-specs=@var{specs}
810 Specify additional command line driver SPECS.
811 This can be useful if you need to turn on a non-standard feature by
812 default without modifying the compiler's source code, for instance
813 @option{--with-specs=%@{!fcommon:%@{!fno-common:-fno-common@}@}}.
815 @xref{Spec Files,, Specifying subprocesses and the switches to pass to them,
816 gcc, Using the GNU Compiler Collection (GCC)},
819 See ``Spec Files'' in the main manual
824 @item --program-prefix=@var{prefix}
825 GCC supports some transformations of the names of its programs when
826 installing them. This option prepends @var{prefix} to the names of
827 programs to install in @var{bindir} (see above). For example, specifying
828 @option{--program-prefix=foo-} would result in @samp{gcc}
829 being installed as @file{/usr/local/bin/foo-gcc}.
831 @item --program-suffix=@var{suffix}
832 Appends @var{suffix} to the names of programs to install in @var{bindir}
833 (see above). For example, specifying @option{--program-suffix=-3.1}
834 would result in @samp{gcc} being installed as
835 @file{/usr/local/bin/gcc-3.1}.
837 @item --program-transform-name=@var{pattern}
838 Applies the @samp{sed} script @var{pattern} to be applied to the names
839 of programs to install in @var{bindir} (see above). @var{pattern} has to
840 consist of one or more basic @samp{sed} editing commands, separated by
841 semicolons. For example, if you want the @samp{gcc} program name to be
842 transformed to the installed program @file{/usr/local/bin/myowngcc} and
843 the @samp{g++} program name to be transformed to
844 @file{/usr/local/bin/gspecial++} without changing other program names,
845 you could use the pattern
846 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
847 to achieve this effect.
849 All three options can be combined and used together, resulting in more
850 complex conversion patterns. As a basic rule, @var{prefix} (and
851 @var{suffix}) are prepended (appended) before further transformations
852 can happen with a special transformation script @var{pattern}.
854 As currently implemented, this option only takes effect for native
855 builds; cross compiler binaries' names are not transformed even when a
856 transformation is explicitly asked for by one of these options.
858 For native builds, some of the installed programs are also installed
859 with the target alias in front of their name, as in
860 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
861 before the target alias is prepended to the name---so, specifying
862 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
863 resulting binary would be installed as
864 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
866 As a last shortcoming, none of the installed Ada programs are
867 transformed yet, which will be fixed in some time.
869 @item --with-local-prefix=@var{dirname}
871 installation directory for local include files. The default is
872 @file{/usr/local}. Specify this option if you want the compiler to
873 search directory @file{@var{dirname}/include} for locally installed
874 header files @emph{instead} of @file{/usr/local/include}.
876 You should specify @option{--with-local-prefix} @strong{only} if your
877 site has a different convention (not @file{/usr/local}) for where to put
880 The default value for @option{--with-local-prefix} is @file{/usr/local}
881 regardless of the value of @option{--prefix}. Specifying
882 @option{--prefix} has no effect on which directory GCC searches for
883 local header files. This may seem counterintuitive, but actually it is
886 The purpose of @option{--prefix} is to specify where to @emph{install
887 GCC}. The local header files in @file{/usr/local/include}---if you put
888 any in that directory---are not part of GCC@. They are part of other
889 programs---perhaps many others. (GCC installs its own header files in
890 another directory which is based on the @option{--prefix} value.)
892 Both the local-prefix include directory and the GCC-prefix include
893 directory are part of GCC's ``system include'' directories. Although these
894 two directories are not fixed, they need to be searched in the proper
895 order for the correct processing of the include_next directive. The
896 local-prefix include directory is searched before the GCC-prefix
897 include directory. Another characteristic of system include directories
898 is that pedantic warnings are turned off for headers in these directories.
900 Some autoconf macros add @option{-I @var{directory}} options to the
901 compiler command line, to ensure that directories containing installed
902 packages' headers are searched. When @var{directory} is one of GCC's
903 system include directories, GCC will ignore the option so that system
904 directories continue to be processed in the correct order. This
905 may result in a search order different from what was specified but the
906 directory will still be searched.
908 GCC automatically searches for ordinary libraries using
909 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
910 used for both GCC and packages, GCC will automatically search for
911 both headers and libraries. This provides a configuration that is
912 easy to use. GCC behaves in a manner similar to that when it is
913 installed as a system compiler in @file{/usr}.
915 Sites that need to install multiple versions of GCC may not want to
916 use the above simple configuration. It is possible to use the
917 @option{--program-prefix}, @option{--program-suffix} and
918 @option{--program-transform-name} options to install multiple versions
919 into a single directory, but it may be simpler to use different prefixes
920 and the @option{--with-local-prefix} option to specify the location of the
921 site-specific files for each version. It will then be necessary for
922 users to specify explicitly the location of local site libraries
923 (e.g., with @env{LIBRARY_PATH}).
925 The same value can be used for both @option{--with-local-prefix} and
926 @option{--prefix} provided it is not @file{/usr}. This can be used
927 to avoid the default search of @file{/usr/local/include}.
929 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
930 The directory you use for @option{--with-local-prefix} @strong{must not}
931 contain any of the system's standard header files. If it did contain
932 them, certain programs would be miscompiled (including GNU Emacs, on
933 certain targets), because this would override and nullify the header
934 file corrections made by the @command{fixincludes} script.
936 Indications are that people who use this option use it based on mistaken
937 ideas of what it is for. People use it as if it specified where to
938 install part of GCC@. Perhaps they make this assumption because
939 installing GCC creates the directory.
941 @item --with-gcc-major-version-only
942 Specifies that GCC should use only the major number rather than
943 @var{major}.@var{minor}.@var{patchlevel} in filesystem paths.
945 @item --with-native-system-header-dir=@var{dirname}
946 Specifies that @var{dirname} is the directory that contains native system
947 header files, rather than @file{/usr/include}. This option is most useful
948 if you are creating a compiler that should be isolated from the system
949 as much as possible. It is most commonly used with the
950 @option{--with-sysroot} option and will cause GCC to search
951 @var{dirname} inside the system root specified by that option.
953 @item --enable-shared[=@var{package}[,@dots{}]]
954 Build shared versions of libraries, if shared libraries are supported on
955 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
956 are enabled by default on all platforms that support shared libraries.
958 If a list of packages is given as an argument, build shared libraries
959 only for the listed packages. For other packages, only static libraries
960 will be built. Package names currently recognized in the GCC tree are
961 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
962 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
963 @samp{ada}, @samp{libada}, @samp{libgo}, @samp{libobjc}, and @samp{libphobos}.
964 Note @samp{libiberty} does not support shared libraries at all.
966 Use @option{--disable-shared} to build only static libraries. Note that
967 @option{--disable-shared} does not accept a list of package names as
968 argument, only @option{--enable-shared} does.
970 Contrast with @option{--enable-host-shared}, which affects @emph{host}
973 @item --enable-host-shared
974 Specify that the @emph{host} code should be built into position-independent
975 machine code (with -fPIC), allowing it to be used within shared libraries,
976 but yielding a slightly slower compiler.
978 This option is required when building the libgccjit.so library.
980 Contrast with @option{--enable-shared}, which affects @emph{target}
983 @item @anchor{with-gnu-as}--with-gnu-as
984 Specify that the compiler should assume that the
985 assembler it finds is the GNU assembler. However, this does not modify
986 the rules to find an assembler and will result in confusion if the
987 assembler found is not actually the GNU assembler. (Confusion may also
988 result if the compiler finds the GNU assembler but has not been
989 configured with @option{--with-gnu-as}.) If you have more than one
990 assembler installed on your system, you may want to use this option in
991 connection with @option{--with-as=@var{pathname}} or
992 @option{--with-build-time-tools=@var{pathname}}.
994 The following systems are the only ones where it makes a difference
995 whether you use the GNU assembler. On any other system,
996 @option{--with-gnu-as} has no effect.
999 @item @samp{hppa1.0-@var{any}-@var{any}}
1000 @item @samp{hppa1.1-@var{any}-@var{any}}
1001 @item @samp{sparc-sun-solaris2.@var{any}}
1002 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
1005 @item @anchor{with-as}--with-as=@var{pathname}
1006 Specify that the compiler should use the assembler pointed to by
1007 @var{pathname}, rather than the one found by the standard rules to find
1008 an assembler, which are:
1011 Unless GCC is being built with a cross compiler, check the
1012 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
1013 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
1014 @var{exec-prefix} defaults to @var{prefix}, which
1015 defaults to @file{/usr/local} unless overridden by the
1016 @option{--prefix=@var{pathname}} switch described above. @var{target}
1017 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
1018 @var{version} denotes the GCC version, such as 3.0.
1021 If the target system is the same that you are building on, check
1022 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
1026 Check in the @env{PATH} for a tool whose name is prefixed by the
1027 target system triple.
1030 Check in the @env{PATH} for a tool whose name is not prefixed by the
1031 target system triple, if the host and target system triple are
1032 the same (in other words, we use a host tool if it can be used for
1033 the target as well).
1036 You may want to use @option{--with-as} if no assembler
1037 is installed in the directories listed above, or if you have multiple
1038 assemblers installed and want to choose one that is not found by the
1041 @item @anchor{with-gnu-ld}--with-gnu-ld
1042 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
1045 @item --with-ld=@var{pathname}
1046 Same as @uref{#with-as,,@option{--with-as}}
1050 Specify that stabs debugging
1051 information should be used instead of whatever format the host normally
1052 uses. Normally GCC uses the same debug format as the host system.
1054 @item --with-tls=@var{dialect}
1055 Specify the default TLS dialect, for systems were there is a choice.
1056 For ARM targets, possible values for @var{dialect} are @code{gnu} or
1057 @code{gnu2}, which select between the original GNU dialect and the GNU TLS
1058 descriptor-based dialect.
1060 @item --enable-multiarch
1061 Specify whether to enable or disable multiarch support. The default is
1062 to check for glibc start files in a multiarch location, and enable it
1063 if the files are found. The auto detection is enabled for native builds,
1064 and for cross builds configured with @option{--with-sysroot}, and without
1065 @option{--with-native-system-header-dir}.
1066 More documentation about multiarch can be found at
1067 @uref{https://wiki.debian.org/Multiarch}.
1069 @item --enable-sjlj-exceptions
1070 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1071 @samp{configure} ordinarily picks the correct value based on the platform.
1072 Only use this option if you are sure you need a different setting.
1074 @item --enable-vtable-verify
1075 Specify whether to enable or disable the vtable verification feature.
1076 Enabling this feature causes libstdc++ to be built with its virtual calls
1077 in verifiable mode. This means that, when linked with libvtv, every
1078 virtual call in libstdc++ will verify the vtable pointer through which the
1079 call will be made before actually making the call. If not linked with libvtv,
1080 the verifier will call stub functions (in libstdc++ itself) and do nothing.
1081 If vtable verification is disabled, then libstdc++ is not built with its
1082 virtual calls in verifiable mode at all. However the libvtv library will
1083 still be built (see @option{--disable-libvtv} to turn off building libvtv).
1084 @option{--disable-vtable-verify} is the default.
1086 @item --disable-gcov
1087 Specify that the run-time library used for coverage analysis
1088 and associated host tools should not be built.
1090 @item --disable-multilib
1091 Specify that multiple target
1092 libraries to support different target variants, calling
1093 conventions, etc.@: should not be built. The default is to build a
1094 predefined set of them.
1096 Some targets provide finer-grained control over which multilibs are built
1097 (e.g., @option{--disable-softfloat}):
1100 fpu, 26bit, underscore, interwork, biendian, nofmult.
1103 softfloat, m68881, m68000, m68020.
1106 single-float, biendian, softfloat.
1111 @item powerpc*-*-*, rs6000*-*-*
1112 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
1117 @item --with-multilib-list=@var{list}
1118 @itemx --without-multilib-list
1119 Specify what multilibs to build. @var{list} is a comma separated list of
1120 values, possibly consisting of a single value. Currently only implemented
1121 for aarch64*-*-*, arm*-*-*, riscv*-*-*, sh*-*-* and x86-64-*-linux*. The
1122 accepted values and meaning for each target is given below.
1126 @var{list} is a comma separated list of @code{ilp32}, and @code{lp64}
1127 to enable ILP32 and LP64 run-time libraries, respectively. If
1128 @var{list} is empty, then there will be no multilibs and only the
1129 default run-time library will be built. If @var{list} is
1130 @code{default} or --with-multilib-list= is not specified, then the
1131 default set of libraries is selected based on the value of
1135 @var{list} is a comma separated list of @code{aprofile} and
1136 @code{rmprofile} to build multilibs for A or R and M architecture
1137 profiles respectively. Note that, due to some limitation of the current
1138 multilib framework, using the combined @code{aprofile,rmprofile}
1139 multilibs selects in some cases a less optimal multilib than when using
1140 the multilib profile for the architecture targetted. The special value
1141 @code{default} is also accepted and is equivalent to omitting the
1142 option, i.e., only the default run-time library will be enabled.
1144 @var{list} may instead contain @code{@@name}, to use the multilib
1145 configuration Makefile fragment @file{name} in @file{gcc/config/arm} in
1146 the source tree (it is part of the corresponding sources, after all).
1147 It is recommended, but not required, that files used for this purpose to
1148 be named starting with @file{t-ml-}, to make their intended purpose
1149 self-evident, in line with GCC conventions. Such files enable custom,
1150 user-chosen multilib lists to be configured. Whether multiple such
1151 files can be used together depends on the contents of the supplied
1152 files. See @file{gcc/config/arm/t-multilib} and its supplementary
1153 @file{gcc/config/arm/t-*profile} files for an example of what such
1154 Makefile fragments might look like for this version of GCC. The macros
1155 expected to be defined in these fragments are not stable across GCC
1156 releases, so make sure they define the @code{MULTILIB}-related macros
1157 expected by the version of GCC you are building.
1159 @xref{Target Fragment,, Target Makefile Fragments, gccint, GNU Compiler
1160 Collection (GCC) Internals}.
1163 See ``Target Makefile Fragments'' in the internals manual.
1166 The table below gives the combination of ISAs, architectures, FPUs and
1167 floating-point ABIs for which multilibs are built for each predefined
1168 profile. The union of these options is considered when specifying both
1169 @code{aprofile} and @code{rmprofile}.
1171 @multitable @columnfractions .15 .28 .30
1172 @item Option @tab aprofile @tab rmprofile
1174 @tab @code{-marm} and @code{-mthumb}
1176 @item Architectures@*@*@*@*@*@*
1177 @tab default architecture@*
1178 @code{-march=armv7-a}@*
1179 @code{-march=armv7ve}@*
1180 @code{-march=armv8-a}@*@*@*
1181 @tab default architecture@*
1182 @code{-march=armv6s-m}@*
1183 @code{-march=armv7-m}@*
1184 @code{-march=armv7e-m}@*
1185 @code{-march=armv8-m.base}@*
1186 @code{-march=armv8-m.main}@*
1188 @item FPUs@*@*@*@*@*
1190 @code{-mfpu=vfpv3-d16}@*
1192 @code{-mfpu=vfpv4-d16}@*
1193 @code{-mfpu=neon-vfpv4}@*
1194 @code{-mfpu=neon-fp-armv8}
1196 @code{-mfpu=vfpv3-d16}@*
1197 @code{-mfpu=fpv4-sp-d16}@*
1198 @code{-mfpu=fpv5-sp-d16}@*
1199 @code{-mfpu=fpv5-d16}@*
1200 @item floating-point@/ ABIs@*@*
1201 @tab @code{-mfloat-abi=soft}@*
1202 @code{-mfloat-abi=softfp}@*
1203 @code{-mfloat-abi=hard}
1204 @tab @code{-mfloat-abi=soft}@*
1205 @code{-mfloat-abi=softfp}@*
1206 @code{-mfloat-abi=hard}
1210 @var{list} is a single ABI name. The target architecture must be either
1211 @code{rv32gc} or @code{rv64gc}. This will build a single multilib for the
1212 specified architecture and ABI pair. If @code{--with-multilib-list} is not
1213 given, then a default set of multilibs is selected based on the value of
1214 @option{--target}. This is usually a large set of multilibs.
1217 @var{list} is a comma separated list of CPU names. These must be of the
1218 form @code{sh*} or @code{m*} (in which case they match the compiler option
1219 for that processor). The list should not contain any endian options -
1220 these are handled by @option{--with-endian}.
1222 If @var{list} is empty, then there will be no multilibs for extra
1223 processors. The multilib for the secondary endian remains enabled.
1225 As a special case, if an entry in the list starts with a @code{!}
1226 (exclamation point), then it is added to the list of excluded multilibs.
1227 Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES}
1228 (once the leading @code{!} has been stripped).
1230 If @option{--with-multilib-list} is not given, then a default set of
1231 multilibs is selected based on the value of @option{--target}. This is
1232 usually the complete set of libraries, but some targets imply a more
1235 Example 1: to configure a compiler for SH4A only, but supporting both
1236 endians, with little endian being the default:
1238 --with-cpu=sh4a --with-endian=little,big --with-multilib-list=
1241 Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with
1242 only little endian SH4AL:
1244 --with-cpu=sh4a --with-endian=little,big \
1245 --with-multilib-list=sh4al,!mb/m4al
1248 @item x86-64-*-linux*
1249 @var{list} is a comma separated list of @code{m32}, @code{m64} and
1250 @code{mx32} to enable 32-bit, 64-bit and x32 run-time libraries,
1251 respectively. If @var{list} is empty, then there will be no multilibs
1252 and only the default run-time library will be enabled.
1254 If @option{--with-multilib-list} is not given, then only 32-bit and
1255 64-bit run-time libraries will be enabled.
1258 @item --with-multilib-generator=@var{config}
1259 Specify what multilibs to build. @var{config} is a semicolon separated list of
1260 values, possibly consisting of a single value. Currently only implemented
1261 for riscv*-*-elf*. The accepted values and meanings are given below.
1264 Every config is constructed with four components: architecture string, ABI,
1265 reuse rule with architecture string and reuse rule with sub-extension.
1267 Example 1: Add multi-lib suppport for rv32i with ilp32.
1272 Example 2: Add multi-lib suppport for rv32i with ilp32 and rv32imafd with ilp32.
1274 rv32i-ilp32--;rv32imafd-ilp32--
1277 Example 3: Add multi-lib suppport for rv32i with ilp32; rv32im with ilp32 and
1278 rv32ic with ilp32 will reuse this multi-lib set.
1280 rv32i-ilp32-rv32im-c
1283 Example 4: Add multi-lib suppport for rv64ima with lp64; rv64imaf with lp64,
1284 rv64imac with lp64 and rv64imafc with lp64 will reuse this multi-lib set.
1286 rv64ima-lp64--f,c,fc
1289 @item --with-endian=@var{endians}
1290 Specify what endians to use.
1291 Currently only implemented for sh*-*-*.
1293 @var{endians} may be one of the following:
1296 Use big endian exclusively.
1298 Use little endian exclusively.
1300 Use big endian by default. Provide a multilib for little endian.
1302 Use little endian by default. Provide a multilib for big endian.
1305 @item --enable-threads
1306 Specify that the target
1307 supports threads. This affects the Objective-C compiler and runtime
1308 library, and exception handling for other languages like C++.
1309 On some systems, this is the default.
1311 In general, the best (and, in many cases, the only known) threading
1312 model available will be configured for use. Beware that on some
1313 systems, GCC has not been taught what threading models are generally
1314 available for the system. In this case, @option{--enable-threads} is an
1315 alias for @option{--enable-threads=single}.
1317 @item --disable-threads
1318 Specify that threading support should be disabled for the system.
1319 This is an alias for @option{--enable-threads=single}.
1321 @item --enable-threads=@var{lib}
1323 @var{lib} is the thread support library. This affects the Objective-C
1324 compiler and runtime library, and exception handling for other languages
1325 like C++. The possibilities for @var{lib} are:
1333 LynxOS thread support.
1335 MIPS SDE thread support.
1337 This is an alias for @samp{single}.
1339 Generic POSIX/Unix98 thread support.
1341 RTEMS thread support.
1343 Disable thread support, should work for all platforms.
1347 VxWorks thread support.
1349 Microsoft Win32 API thread support.
1353 Specify that the target supports TLS (Thread Local Storage). Usually
1354 configure can correctly determine if TLS is supported. In cases where
1355 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1356 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1357 the assembler supports TLS but the C library does not, or if the
1358 assumptions made by the configure test are incorrect.
1361 Specify that the target does not support TLS.
1362 This is an alias for @option{--enable-tls=no}.
1364 @item --disable-tm-clone-registry
1365 Disable TM clone registry in libgcc. It is enabled in libgcc by default.
1366 This option helps to reduce code size for embedded targets which do
1367 not use transactional memory.
1369 @item --with-cpu=@var{cpu}
1370 @itemx --with-cpu-32=@var{cpu}
1371 @itemx --with-cpu-64=@var{cpu}
1372 Specify which cpu variant the compiler should generate code for by default.
1373 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1374 This option is only supported on some targets, including ARC, ARM, i386, M68k,
1375 PowerPC, and SPARC@. It is mandatory for ARC@. The @option{--with-cpu-32} and
1376 @option{--with-cpu-64} options specify separate default CPUs for
1377 32-bit and 64-bit modes; these options are only supported for aarch64, i386,
1378 x86-64, PowerPC, and SPARC@.
1380 @item --with-schedule=@var{cpu}
1381 @itemx --with-arch=@var{cpu}
1382 @itemx --with-arch-32=@var{cpu}
1383 @itemx --with-arch-64=@var{cpu}
1384 @itemx --with-tune=@var{cpu}
1385 @itemx --with-tune-32=@var{cpu}
1386 @itemx --with-tune-64=@var{cpu}
1387 @itemx --with-abi=@var{abi}
1388 @itemx --with-fpu=@var{type}
1389 @itemx --with-float=@var{type}
1390 These configure options provide default values for the @option{-mschedule=},
1391 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1392 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1393 @option{--with-cpu}, which switches will be accepted and acceptable values
1394 of the arguments depend on the target.
1396 @item --with-mode=@var{mode}
1397 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1398 This option is only supported on ARM targets.
1400 @item --with-stack-offset=@var{num}
1401 This option sets the default for the -mstack-offset=@var{num} option,
1402 and will thus generally also control the setting of this option for
1403 libraries. This option is only supported on Epiphany targets.
1405 @item --with-fpmath=@var{isa}
1406 This options sets @option{-mfpmath=sse} by default and specifies the default
1407 ISA for floating-point arithmetics. You can select either @samp{sse} which
1408 enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default.
1409 This option is only supported on i386 and x86-64 targets.
1411 @item --with-fp-32=@var{mode}
1412 On MIPS targets, set the default value for the @option{-mfp} option when using
1413 the o32 ABI. The possibilities for @var{mode} are:
1416 Use the o32 FP32 ABI extension, as with the @option{-mfp32} command-line
1419 Use the o32 FPXX ABI extension, as with the @option{-mfpxx} command-line
1422 Use the o32 FP64 ABI extension, as with the @option{-mfp64} command-line
1425 In the absence of this configuration option the default is to use the o32
1428 @item --with-odd-spreg-32
1429 On MIPS targets, set the @option{-modd-spreg} option by default when using
1432 @item --without-odd-spreg-32
1433 On MIPS targets, set the @option{-mno-odd-spreg} option by default when using
1434 the o32 ABI. This is normally used in conjunction with
1435 @option{--with-fp-32=64} in order to target the o32 FP64A ABI extension.
1437 @item --with-nan=@var{encoding}
1438 On MIPS targets, set the default encoding convention to use for the
1439 special not-a-number (NaN) IEEE 754 floating-point data. The
1440 possibilities for @var{encoding} are:
1443 Use the legacy encoding, as with the @option{-mnan=legacy} command-line
1446 Use the 754-2008 encoding, as with the @option{-mnan=2008} command-line
1449 To use this configuration option you must have an assembler version
1450 installed that supports the @option{-mnan=} command-line option too.
1451 In the absence of this configuration option the default convention is
1452 the legacy encoding, as when neither of the @option{-mnan=2008} and
1453 @option{-mnan=legacy} command-line options has been used.
1455 @item --with-divide=@var{type}
1456 Specify how the compiler should generate code for checking for
1457 division by zero. This option is only supported on the MIPS target.
1458 The possibilities for @var{type} are:
1461 Division by zero checks use conditional traps (this is the default on
1462 systems that support conditional traps).
1464 Division by zero checks use the break instruction.
1467 @c If you make --with-llsc the default for additional targets,
1468 @c update the --with-llsc description in the MIPS section below.
1471 On MIPS targets, make @option{-mllsc} the default when no
1472 @option{-mno-llsc} option is passed. This is the default for
1473 Linux-based targets, as the kernel will emulate them if the ISA does
1476 @item --without-llsc
1477 On MIPS targets, make @option{-mno-llsc} the default when no
1478 @option{-mllsc} option is passed.
1481 On MIPS targets, make @option{-msynci} the default when no
1482 @option{-mno-synci} option is passed.
1484 @item --without-synci
1485 On MIPS targets, make @option{-mno-synci} the default when no
1486 @option{-msynci} option is passed. This is the default.
1488 @item --with-lxc1-sxc1
1489 On MIPS targets, make @option{-mlxc1-sxc1} the default when no
1490 @option{-mno-lxc1-sxc1} option is passed. This is the default.
1492 @item --without-lxc1-sxc1
1493 On MIPS targets, make @option{-mno-lxc1-sxc1} the default when no
1494 @option{-mlxc1-sxc1} option is passed. The indexed load/store
1495 instructions are not directly a problem but can lead to unexpected
1496 behaviour when deployed in an application intended for a 32-bit address
1497 space but run on a 64-bit processor. The issue is seen because all
1498 known MIPS 64-bit Linux kernels execute o32 and n32 applications
1499 with 64-bit addressing enabled which affects the overflow behaviour
1500 of the indexed addressing mode. GCC will assume that ordinary
1501 32-bit arithmetic overflow behaviour is the same whether performed
1502 as an @code{addu} instruction or as part of the address calculation
1503 in @code{lwxc1} type instructions. This assumption holds true in a
1504 pure 32-bit environment and can hold true in a 64-bit environment if
1505 the address space is accurately set to be 32-bit for o32 and n32.
1508 On MIPS targets, make @option{-mmadd4} the default when no
1509 @option{-mno-madd4} option is passed. This is the default.
1511 @item --without-madd4
1512 On MIPS targets, make @option{-mno-madd4} the default when no
1513 @option{-mmadd4} option is passed. The @code{madd4} instruction
1514 family can be problematic when targeting a combination of cores that
1515 implement these instructions differently. There are two known cores
1516 that implement these as fused operations instead of unfused (where
1517 unfused is normally expected). Disabling these instructions is the
1518 only way to ensure compatible code is generated; this will incur
1519 a performance penalty.
1521 @item --with-mips-plt
1522 On MIPS targets, make use of copy relocations and PLTs.
1523 These features are extensions to the traditional
1524 SVR4-based MIPS ABIs and require support from GNU binutils
1525 and the runtime C library.
1527 @item --with-stack-clash-protection-guard-size=@var{size}
1528 On certain targets this option sets the default stack clash protection guard
1529 size as a power of two in bytes. On AArch64 @var{size} is required to be either
1530 12 (4KB) or 16 (64KB).
1532 @item --enable-__cxa_atexit
1533 Define if you want to use __cxa_atexit, rather than atexit, to
1534 register C++ destructors for local statics and global objects.
1535 This is essential for fully standards-compliant handling of
1536 destructors, but requires __cxa_atexit in libc. This option is currently
1537 only available on systems with GNU libc. When enabled, this will cause
1538 @option{-fuse-cxa-atexit} to be passed by default.
1540 @item --enable-gnu-indirect-function
1541 Define if you want to enable the @code{ifunc} attribute. This option is
1542 currently only available on systems with GNU libc on certain targets.
1544 @item --enable-target-optspace
1546 libraries should be optimized for code space instead of code speed.
1547 This is the default for the m32r platform.
1549 @item --with-cpp-install-dir=@var{dirname}
1550 Specify that the user visible @command{cpp} program should be installed
1551 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1553 @item --enable-comdat
1554 Enable COMDAT group support. This is primarily used to override the
1555 automatically detected value.
1557 @item --enable-initfini-array
1558 Force the use of sections @code{.init_array} and @code{.fini_array}
1559 (instead of @code{.init} and @code{.fini}) for constructors and
1560 destructors. Option @option{--disable-initfini-array} has the
1561 opposite effect. If neither option is specified, the configure script
1562 will try to guess whether the @code{.init_array} and
1563 @code{.fini_array} sections are supported and, if they are, use them.
1565 @item --enable-link-mutex
1566 When building GCC, use a mutex to avoid linking the compilers for
1567 multiple languages at the same time, to avoid thrashing on build
1568 systems with limited free memory. The default is not to use such a mutex.
1570 @item --enable-link-serialization
1571 When building GCC, use make dependencies to serialize linking the compilers for
1572 multiple languages, to avoid thrashing on build
1573 systems with limited free memory. The default is not to add such
1574 dependencies and thus with parallel make potentially link different
1575 compilers concurrently. If the argument is a positive integer, allow
1576 that number of concurrent link processes for the large binaries.
1578 @item --enable-maintainer-mode
1579 The build rules that regenerate the Autoconf and Automake output files as
1580 well as the GCC master message catalog @file{gcc.pot} are normally
1581 disabled. This is because it can only be rebuilt if the complete source
1582 tree is present. If you have changed the sources and want to rebuild the
1583 catalog, configuring with @option{--enable-maintainer-mode} will enable
1584 this. Note that you need a recent version of the @code{gettext} tools
1587 @item --disable-bootstrap
1588 For a native build, the default configuration is to perform
1589 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1590 testing that GCC can compile itself correctly. If you want to disable
1591 this process, you can configure with @option{--disable-bootstrap}.
1593 @item --enable-bootstrap
1594 In special cases, you may want to perform a 3-stage build
1595 even if the target and host triplets are different.
1596 This is possible when the host can run code compiled for
1597 the target (e.g.@: host is i686-linux, target is i486-linux).
1598 Starting from GCC 4.2, to do this you have to configure explicitly
1599 with @option{--enable-bootstrap}.
1601 @item --enable-generated-files-in-srcdir
1602 Neither the .c and .h files that are generated from Bison and flex nor the
1603 info manuals and man pages that are built from the .texi files are present
1604 in the repository development tree. When building GCC from that development tree,
1605 or from one of our snapshots, those generated files are placed in your
1606 build directory, which allows for the source to be in a readonly
1609 If you configure with @option{--enable-generated-files-in-srcdir} then those
1610 generated files will go into the source directory. This is mainly intended
1611 for generating release or prerelease tarballs of the GCC sources, since it
1612 is not a requirement that the users of source releases to have flex, Bison,
1615 @item --enable-version-specific-runtime-libs
1617 that runtime libraries should be installed in the compiler specific
1618 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1619 addition, @samp{libstdc++}'s include files will be installed into
1620 @file{@var{libdir}} unless you overruled it by using
1621 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1622 particularly useful if you intend to use several versions of GCC in
1623 parallel. The default is @samp{yes} for @samp{libada}, and @samp{no} for
1624 the remaining libraries.
1626 @item @anchor{WithAixSoname}--with-aix-soname=@samp{aix}, @samp{svr4} or @samp{both}
1627 Traditional AIX shared library versioning (versioned @code{Shared Object}
1628 files as members of unversioned @code{Archive Library} files named
1629 @samp{lib.a}) causes numerous headaches for package managers. However,
1630 @code{Import Files} as members of @code{Archive Library} files allow for
1631 @strong{filename-based versioning} of shared libraries as seen on Linux/SVR4,
1632 where this is called the "SONAME". But as they prevent static linking,
1633 @code{Import Files} may be used with @code{Runtime Linking} only, where the
1634 linker does search for @samp{libNAME.so} before @samp{libNAME.a} library
1635 filenames with the @samp{-lNAME} linker flag.
1637 @anchor{AixLdCommand}For detailed information please refer to the AIX
1638 @uref{https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22,,ld
1641 As long as shared library creation is enabled, upon:
1643 @item --with-aix-soname=aix
1644 @item --with-aix-soname=both
1645 A (traditional AIX) @code{Shared Archive Library} file is created:
1647 @item using the @samp{libNAME.a} filename scheme
1648 @item with the @code{Shared Object} file as archive member named
1649 @samp{libNAME.so.V} (except for @samp{libgcc_s}, where the @code{Shared
1650 Object} file is named @samp{shr.o} for backwards compatibility), which
1652 @item is used for runtime loading from inside the @samp{libNAME.a} file
1653 @item is used for dynamic loading via
1654 @code{dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)}
1655 @item is used for shared linking
1656 @item is used for static linking, so no separate @code{Static Archive
1657 Library} file is needed
1660 @item --with-aix-soname=both
1661 @item --with-aix-soname=svr4
1662 A (second) @code{Shared Archive Library} file is created:
1664 @item using the @samp{libNAME.so.V} filename scheme
1665 @item with the @code{Shared Object} file as archive member named
1668 @item is created with the @code{-G linker flag}
1669 @item has the @code{F_LOADONLY} flag set
1670 @item is used for runtime loading from inside the @samp{libNAME.so.V} file
1671 @item is used for dynamic loading via @code{dlopen("libNAME.so.V(shr.o)",
1674 @item with the @code{Import File} as archive member named @samp{shr.imp},
1677 @item refers to @samp{libNAME.so.V(shr.o)} as the "SONAME", to be recorded
1678 in the @code{Loader Section} of subsequent binaries
1679 @item indicates whether @samp{libNAME.so.V(shr.o)} is 32 or 64 bit
1680 @item lists all the public symbols exported by @samp{lib.so.V(shr.o)},
1681 eventually decorated with the @code{@samp{weak} Keyword}
1682 @item is necessary for shared linking against @samp{lib.so.V(shr.o)}
1685 A symbolic link using the @samp{libNAME.so} filename scheme is created:
1687 @item pointing to the @samp{libNAME.so.V} @code{Shared Archive Library} file
1688 @item to permit the @code{ld Command} to find @samp{lib.so.V(shr.imp)} via
1689 the @samp{-lNAME} argument (requires @code{Runtime Linking} to be enabled)
1690 @item to permit dynamic loading of @samp{lib.so.V(shr.o)} without the need
1691 to specify the version number via @code{dlopen("libNAME.so(shr.o)",
1696 As long as static library creation is enabled, upon:
1698 @item --with-aix-soname=svr4
1699 A @code{Static Archive Library} is created:
1701 @item using the @samp{libNAME.a} filename scheme
1702 @item with all the @code{Static Object} files as archive members, which
1704 @item are used for static linking
1709 While the aix-soname=@samp{svr4} option does not create @code{Shared Object}
1710 files as members of unversioned @code{Archive Library} files any more, package
1711 managers still are responsible to
1712 @uref{./specific.html#TransferAixShobj,,transfer} @code{Shared Object} files
1713 found as member of a previously installed unversioned @code{Archive Library}
1714 file into the newly installed @code{Archive Library} file with the same
1717 @emph{WARNING:} Creating @code{Shared Object} files with @code{Runtime Linking}
1718 enabled may bloat the TOC, eventually leading to @code{TOC overflow} errors,
1719 requiring the use of either the @option{-Wl,-bbigtoc} linker flag (seen to
1720 break with the @code{GDB} debugger) or some of the TOC-related compiler flags,
1722 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1723 Using the GNU Compiler Collection (GCC)}.
1726 see ``RS/6000 and PowerPC Options'' in the main manual.
1729 @option{--with-aix-soname} is currently supported by @samp{libgcc_s} only, so
1730 this option is still experimental and not for normal use yet.
1732 Default is the traditional behavior @option{--with-aix-soname=@samp{aix}}.
1734 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1735 Specify that only a particular subset of compilers and
1736 their runtime libraries should be built. For a list of valid values for
1737 @var{langN} you can issue the following command in the
1738 @file{gcc} directory of your GCC source tree:@*
1740 grep ^language= */config-lang.in
1742 Currently, you can use any of the following:
1743 @code{all}, @code{default}, @code{ada}, @code{c}, @code{c++}, @code{d},
1744 @code{fortran}, @code{go}, @code{jit}, @code{lto}, @code{objc}, @code{obj-c++}.
1745 Building the Ada compiler has special requirements, see below.
1746 If you do not pass this flag, or specify the option @code{default}, then the
1747 default languages available in the @file{gcc} sub-tree will be configured.
1748 Ada, D, Go, Jit, and Objective-C++ are not default languages. LTO is not a
1749 default language, but is built by default because @option{--enable-lto} is
1750 enabled by default. The other languages are default languages. If
1751 @code{all} is specified, then all available languages are built. An
1752 exception is @code{jit} language, which requires
1753 @option{--enable-host-shared} to be included with @code{all}.
1755 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1756 Specify that a particular subset of compilers and their runtime
1757 libraries should be built with the system C compiler during stage 1 of
1758 the bootstrap process, rather than only in later stages with the
1759 bootstrapped C compiler. The list of valid values is the same as for
1760 @option{--enable-languages}, and the option @code{all} will select all
1761 of the languages enabled by @option{--enable-languages}. This option is
1762 primarily useful for GCC development; for instance, when a development
1763 version of the compiler cannot bootstrap due to compiler bugs, or when
1764 one is debugging front ends other than the C front end. When this
1765 option is used, one can then build the target libraries for the
1766 specified languages with the stage-1 compiler by using @command{make
1767 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1768 for the specified languages using @command{make stage1-start check-gcc}.
1770 @item --disable-libada
1771 Specify that the run-time libraries and tools used by GNAT should not
1772 be built. This can be useful for debugging, or for compatibility with
1773 previous Ada build procedures, when it was required to explicitly
1774 do a @samp{make -C gcc gnatlib_and_tools}.
1776 @item --disable-libsanitizer
1777 Specify that the run-time libraries for the various sanitizers should
1780 @item --disable-libssp
1781 Specify that the run-time libraries for stack smashing protection
1782 should not be built or linked against. On many targets library support
1783 is provided by the C library instead.
1785 @item --disable-libquadmath
1786 Specify that the GCC quad-precision math library should not be built.
1787 On some systems, the library is required to be linkable when building
1788 the Fortran front end, unless @option{--disable-libquadmath-support}
1791 @item --disable-libquadmath-support
1792 Specify that the Fortran front end and @code{libgfortran} do not add
1793 support for @code{libquadmath} on systems supporting it.
1795 @item --disable-libgomp
1796 Specify that the GNU Offloading and Multi Processing Runtime Library
1797 should not be built.
1799 @item --disable-libvtv
1800 Specify that the run-time libraries used by vtable verification
1801 should not be built.
1804 Specify that the compiler should
1805 use DWARF 2 debugging information as the default.
1807 @item --with-advance-toolchain=@var{at}
1808 On 64-bit PowerPC Linux systems, configure the compiler to use the
1809 header files, library files, and the dynamic linker from the Advance
1810 Toolchain release @var{at} instead of the default versions that are
1811 provided by the Linux distribution. In general, this option is
1812 intended for the developers of GCC, and it is not intended for general
1815 @item --enable-targets=all
1816 @itemx --enable-targets=@var{target_list}
1817 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1818 These are compilers that are able to generate either 64-bit or 32-bit
1819 code. Typically, the corresponding 32-bit target, e.g.@:
1820 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1821 option enables the 32-bit target to be a bi-arch compiler, which is
1822 useful when you want a bi-arch compiler that defaults to 32-bit, and
1823 you are building a bi-arch or multi-arch binutils in a combined tree.
1824 On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64),
1826 Currently, this option only affects sparc-linux, powerpc-linux, x86-linux,
1827 mips-linux and s390-linux.
1829 @item --enable-default-pie
1830 Turn on @option{-fPIE} and @option{-pie} by default.
1832 @item --enable-secureplt
1833 This option enables @option{-msecure-plt} by default for powerpc-linux.
1835 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1836 Using the GNU Compiler Collection (GCC)},
1839 See ``RS/6000 and PowerPC Options'' in the main manual
1842 @item --enable-default-ssp
1843 Turn on @option{-fstack-protector-strong} by default.
1846 This option enables @option{-mcld} by default for 32-bit x86 targets.
1848 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1849 Using the GNU Compiler Collection (GCC)},
1852 See ``i386 and x86-64 Options'' in the main manual
1855 @item --enable-large-address-aware
1856 The @option{--enable-large-address-aware} option arranges for MinGW
1857 executables to be linked using the @option{--large-address-aware}
1858 option, that enables the use of more than 2GB of memory. If GCC is
1859 configured with this option, its effects can be reversed by passing the
1860 @option{-Wl,--disable-large-address-aware} option to the so-configured
1863 @item --enable-win32-registry
1864 @itemx --enable-win32-registry=@var{key}
1865 @itemx --disable-win32-registry
1866 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1867 to look up installations paths in the registry using the following key:
1870 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1873 @var{key} defaults to GCC version number, and can be overridden by the
1874 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1875 who use custom installers are encouraged to provide a different key,
1876 perhaps one comprised of vendor name and GCC version number, to
1877 avoid conflict with existing installations. This feature is enabled
1878 by default, and can be disabled by @option{--disable-win32-registry}
1879 option. This option has no effect on the other hosts.
1882 Specify that the machine does not have a floating point unit. This
1883 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1884 system, @option{--nfp} has no effect.
1886 @item --enable-werror
1887 @itemx --disable-werror
1888 @itemx --enable-werror=yes
1889 @itemx --enable-werror=no
1890 When you specify this option, it controls whether certain files in the
1891 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1892 If you don't specify it, @option{-Werror} is turned on for the main
1893 development trunk. However it defaults to off for release branches and
1894 final releases. The specific files which get @option{-Werror} are
1895 controlled by the Makefiles.
1897 @item --enable-checking
1898 @itemx --disable-checking
1899 @itemx --enable-checking=@var{list}
1900 This option controls performing internal consistency checks in the compiler.
1901 It does not change the generated code, but adds error checking of the
1902 requested complexity. This slows down the compiler and may only work
1903 properly if you are building the compiler with GCC@.
1905 When the option is not specified, the active set of checks depends on context.
1906 Namely, bootstrap stage 1 defaults to @samp{--enable-checking=yes}, builds
1907 from release branches or release archives default to
1908 @samp{--enable-checking=release}, and otherwise
1909 @samp{--enable-checking=yes,extra} is used. When the option is
1910 specified without a @var{list}, the result is the same as
1911 @samp{--enable-checking=yes}. Likewise, @samp{--disable-checking} is
1912 equivalent to @samp{--enable-checking=no}.
1914 The categories of checks available in @var{list} are @samp{yes} (most common
1915 checks @samp{assert,misc,gc,gimple,rtlflag,runtime,tree,types}), @samp{no}
1916 (no checks at all), @samp{all} (all but @samp{valgrind}), @samp{release}
1917 (cheapest checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1918 @samp{release} checks are always on and to disable them
1919 @samp{--disable-checking} or @samp{--enable-checking=no[,<other checks>]}
1920 must be explicitly requested. Disabling assertions makes the compiler and
1921 runtime slightly faster but increases the risk of undetected internal errors
1922 causing wrong code to be generated.
1924 Individual checks can be enabled with these flags: @samp{assert}, @samp{df},
1925 @samp{extra}, @samp{fold}, @samp{gc}, @samp{gcac}, @samp{gimple},
1926 @samp{misc}, @samp{rtl}, @samp{rtlflag}, @samp{runtime}, @samp{tree},
1927 @samp{types} and @samp{valgrind}. @samp{extra} extends @samp{misc}
1928 checking with extra checks that might affect code generation and should
1929 therefore not differ between stage1 and later stages in bootstrap.
1931 The @samp{valgrind} check requires the external @command{valgrind} simulator,
1932 available from @uref{https://valgrind.org}. The @samp{rtl} checks are
1933 expensive and the @samp{df}, @samp{gcac} and @samp{valgrind} checks are very
1936 @item --disable-stage1-checking
1937 @itemx --enable-stage1-checking
1938 @itemx --enable-stage1-checking=@var{list}
1939 This option affects only bootstrap build. If no @option{--enable-checking}
1940 option is specified the stage1 compiler is built with @samp{yes} checking
1941 enabled, otherwise the stage1 checking flags are the same as specified by
1942 @option{--enable-checking}. To build the stage1 compiler with
1943 different checking options use @option{--enable-stage1-checking}.
1944 The list of checking options is the same as for @option{--enable-checking}.
1945 If your system is too slow or too small to bootstrap a released compiler
1946 with checking for stage1 enabled, you can use @samp{--disable-stage1-checking}
1947 to disable checking for the stage1 compiler.
1949 @item --enable-coverage
1950 @itemx --enable-coverage=@var{level}
1951 With this option, the compiler is built to collect self coverage
1952 information, every time it is run. This is for internal development
1953 purposes, and only works when the compiler is being built with gcc. The
1954 @var{level} argument controls whether the compiler is built optimized or
1955 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1956 want to disable optimization, for performance analysis you want to
1957 enable optimization. When coverage is enabled, the default level is
1958 without optimization.
1960 @item --enable-gather-detailed-mem-stats
1961 When this option is specified more detailed information on memory
1962 allocation is gathered. This information is printed when using
1963 @option{-fmem-report}.
1965 @item --enable-valgrind-annotations
1966 Mark selected memory related operations in the compiler when run under
1967 valgrind to suppress false positives.
1970 @itemx --disable-nls
1971 The @option{--enable-nls} option enables Native Language Support (NLS),
1972 which lets GCC output diagnostics in languages other than American
1973 English. Native Language Support is enabled by default if not doing a
1974 canadian cross build. The @option{--disable-nls} option disables NLS@.
1976 @item --with-included-gettext
1977 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1978 procedure to prefer its copy of GNU @command{gettext}.
1980 @item --with-catgets
1981 If NLS is enabled, and if the host lacks @code{gettext} but has the
1982 inferior @code{catgets} interface, the GCC build procedure normally
1983 ignores @code{catgets} and instead uses GCC's copy of the GNU
1984 @code{gettext} library. The @option{--with-catgets} option causes the
1985 build procedure to use the host's @code{catgets} in this situation.
1987 @item --with-libiconv-prefix=@var{dir}
1988 Search for libiconv header files in @file{@var{dir}/include} and
1989 libiconv library files in @file{@var{dir}/lib}.
1991 @item --enable-obsolete
1992 Enable configuration for an obsoleted system. If you attempt to
1993 configure GCC for a system (build, host, or target) which has been
1994 obsoleted, and you do not specify this flag, configure will halt with an
1997 All support for systems which have been obsoleted in one release of GCC
1998 is removed entirely in the next major release, unless someone steps
1999 forward to maintain the port.
2001 @item --enable-decimal-float
2002 @itemx --enable-decimal-float=yes
2003 @itemx --enable-decimal-float=no
2004 @itemx --enable-decimal-float=bid
2005 @itemx --enable-decimal-float=dpd
2006 @itemx --disable-decimal-float
2007 Enable (or disable) support for the C decimal floating point extension
2008 that is in the IEEE 754-2008 standard. This is enabled by default only
2009 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
2010 support it, but require the user to specifically enable it. You can
2011 optionally control which decimal floating point format is used (either
2012 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
2013 format is default on i386 and x86_64 systems, and the @samp{dpd}
2014 (densely packed decimal) format is default on PowerPC systems.
2016 @item --enable-fixed-point
2017 @itemx --disable-fixed-point
2018 Enable (or disable) support for C fixed-point arithmetic.
2019 This option is enabled by default for some targets (such as MIPS) which
2020 have hardware-support for fixed-point operations. On other targets, you
2021 may enable this option manually.
2023 @item --with-long-double-128
2024 Specify if @code{long double} type should be 128-bit by default on selected
2025 GNU/Linux architectures. If using @code{--without-long-double-128},
2026 @code{long double} will be by default 64-bit, the same as @code{double} type.
2027 When neither of these configure options are used, the default will be
2028 128-bit @code{long double} when built against GNU C Library 2.4 and later,
2029 64-bit @code{long double} otherwise.
2031 @item --with-long-double-format=ibm
2032 @itemx --with-long-double-format=ieee
2033 Specify whether @code{long double} uses the IBM extended double format
2034 or the IEEE 128-bit floating point format on PowerPC Linux systems.
2035 This configuration switch will only work on little endian PowerPC
2036 Linux systems and on big endian 64-bit systems where the default cpu
2037 is at least power7 (i.e.@: @option{--with-cpu=power7},
2038 @option{--with-cpu=power8}, or @option{--with-cpu=power9} is used).
2040 If you use the @option{--with-long-double-64} configuration option,
2041 the @option{--with-long-double-format=ibm} and
2042 @option{--with-long-double-format=ieee} options are ignored.
2044 The default @code{long double} format is to use IBM extended double.
2045 Until all of the libraries are converted to use IEEE 128-bit floating
2046 point, it is not recommended to use
2047 @option{--with-long-double-format=ieee}.
2049 On little endian PowerPC Linux systems, if you explicitly set the
2050 @code{long double} type, it will build multilibs to allow you to
2051 select either @code{long double} format, unless you disable multilibs
2052 with the @code{--disable-multilib} option. At present,
2053 @code{long double} multilibs are not built on big endian PowerPC Linux
2054 systems. If you are building multilibs, you will need to configure
2055 the compiler using the @option{--with-system-zlib} option.
2057 If you do not set the @code{long double} type explicitly, no multilibs
2060 @item --enable-fdpic
2061 On SH Linux systems, generate ELF FDPIC code.
2063 @item --with-gmp=@var{pathname}
2064 @itemx --with-gmp-include=@var{pathname}
2065 @itemx --with-gmp-lib=@var{pathname}
2066 @itemx --with-mpfr=@var{pathname}
2067 @itemx --with-mpfr-include=@var{pathname}
2068 @itemx --with-mpfr-lib=@var{pathname}
2069 @itemx --with-mpc=@var{pathname}
2070 @itemx --with-mpc-include=@var{pathname}
2071 @itemx --with-mpc-lib=@var{pathname}
2072 If you want to build GCC but do not have the GMP library, the MPFR
2073 library and/or the MPC library installed in a standard location and
2074 do not have their sources present in the GCC source tree then you
2075 can explicitly specify the directory where they are installed
2076 (@samp{--with-gmp=@var{gmpinstalldir}},
2077 @samp{--with-mpfr=@/@var{mpfrinstalldir}},
2078 @samp{--with-mpc=@/@var{mpcinstalldir}}). The
2079 @option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for
2080 @option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and
2081 @option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the
2082 @option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for
2083 @option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and
2084 @option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the
2085 @option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for
2086 @option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and
2087 @option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these
2088 shorthand assumptions are not correct, you can use the explicit
2089 include and lib options directly. You might also need to ensure the
2090 shared libraries can be found by the dynamic linker when building and
2091 using GCC, for example by setting the runtime shared library path
2092 variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems).
2094 These flags are applicable to the host platform only. When building
2095 a cross compiler, they will not be used to configure target libraries.
2097 @item --with-isl=@var{pathname}
2098 @itemx --with-isl-include=@var{pathname}
2099 @itemx --with-isl-lib=@var{pathname}
2100 If you do not have the isl library installed in a standard location and you
2101 want to build GCC, you can explicitly specify the directory where it is
2102 installed (@samp{--with-isl=@/@var{islinstalldir}}). The
2103 @option{--with-isl=@/@var{islinstalldir}} option is shorthand for
2104 @option{--with-isl-lib=@/@var{islinstalldir}/lib} and
2105 @option{--with-isl-include=@/@var{islinstalldir}/include}. If this
2106 shorthand assumption is not correct, you can use the explicit
2107 include and lib options directly.
2109 These flags are applicable to the host platform only. When building
2110 a cross compiler, they will not be used to configure target libraries.
2112 @item --with-stage1-ldflags=@var{flags}
2113 This option may be used to set linker flags to be used when linking
2114 stage 1 of GCC. These are also used when linking GCC if configured with
2115 @option{--disable-bootstrap}. If @option{--with-stage1-libs} is not set to a
2116 value, then the default is @samp{-static-libstdc++ -static-libgcc}, if
2119 @item --with-stage1-libs=@var{libs}
2120 This option may be used to set libraries to be used when linking stage 1
2121 of GCC. These are also used when linking GCC if configured with
2122 @option{--disable-bootstrap}.
2124 @item --with-boot-ldflags=@var{flags}
2125 This option may be used to set linker flags to be used when linking
2126 stage 2 and later when bootstrapping GCC. If --with-boot-libs
2127 is not is set to a value, then the default is
2128 @samp{-static-libstdc++ -static-libgcc}.
2130 @item --with-boot-libs=@var{libs}
2131 This option may be used to set libraries to be used when linking stage 2
2132 and later when bootstrapping GCC.
2134 @item --with-debug-prefix-map=@var{map}
2135 Convert source directory names using @option{-fdebug-prefix-map} when
2136 building runtime libraries. @samp{@var{map}} is a space-separated
2137 list of maps of the form @samp{@var{old}=@var{new}}.
2139 @item --enable-linker-build-id
2140 Tells GCC to pass @option{--build-id} option to the linker for all final
2141 links (links performed without the @option{-r} or @option{--relocatable}
2142 option), if the linker supports it. If you specify
2143 @option{--enable-linker-build-id}, but your linker does not
2144 support @option{--build-id} option, a warning is issued and the
2145 @option{--enable-linker-build-id} option is ignored. The default is off.
2147 @item --with-linker-hash-style=@var{choice}
2148 Tells GCC to pass @option{--hash-style=@var{choice}} option to the
2149 linker for all final links. @var{choice} can be one of
2150 @samp{sysv}, @samp{gnu}, and @samp{both} where @samp{sysv} is the default.
2152 @item --enable-gnu-unique-object
2153 @itemx --disable-gnu-unique-object
2154 Tells GCC to use the gnu_unique_object relocation for C++ template
2155 static data members and inline function local statics. Enabled by
2156 default for a toolchain with an assembler that accepts it and
2157 GLIBC 2.11 or above, otherwise disabled.
2159 @item --with-diagnostics-color=@var{choice}
2160 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-color=}
2161 option (if not used explicitly on the command line). @var{choice}
2162 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2163 where @samp{auto} is the default. @samp{auto-if-env} makes
2164 @option{-fdiagnostics-color=auto} the default if @env{GCC_COLORS}
2165 is present and non-empty in the environment of the compiler, and
2166 @option{-fdiagnostics-color=never} otherwise.
2168 @item --with-diagnostics-urls=@var{choice}
2169 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-urls=}
2170 option (if not used explicitly on the command line). @var{choice}
2171 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2172 where @samp{auto} is the default. @samp{auto-if-env} makes
2173 @option{-fdiagnostics-urls=auto} the default if @env{GCC_URLS}
2174 or @env{TERM_URLS} is present and non-empty in the environment of the
2175 compiler, and @option{-fdiagnostics-urls=never} otherwise.
2178 @itemx --disable-lto
2179 Enable support for link-time optimization (LTO). This is enabled by
2180 default, and may be disabled using @option{--disable-lto}.
2182 @item --enable-linker-plugin-configure-flags=FLAGS
2183 @itemx --enable-linker-plugin-flags=FLAGS
2184 By default, linker plugins (such as the LTO plugin) are built for the
2185 host system architecture. For the case that the linker has a
2186 different (but run-time compatible) architecture, these flags can be
2187 specified to build plugins that are compatible to the linker. For
2188 example, if you are building GCC for a 64-bit x86_64
2189 (@samp{x86_64-pc-linux-gnu}) host system, but have a 32-bit x86
2190 GNU/Linux (@samp{i686-pc-linux-gnu}) linker executable (which is
2191 executable on the former system), you can configure GCC as follows for
2192 getting compatible linker plugins:
2195 % @var{srcdir}/configure \
2196 --host=x86_64-pc-linux-gnu \
2197 --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \
2198 --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib'
2201 @item --with-plugin-ld=@var{pathname}
2202 Enable an alternate linker to be used at link-time optimization (LTO)
2203 link time when @option{-fuse-linker-plugin} is enabled.
2204 This linker should have plugin support such as gold starting with
2205 version 2.20 or GNU ld starting with version 2.21.
2206 See @option{-fuse-linker-plugin} for details.
2208 @item --enable-canonical-system-headers
2209 @itemx --disable-canonical-system-headers
2210 Enable system header path canonicalization for @file{libcpp}. This can
2211 produce shorter header file paths in diagnostics and dependency output
2212 files, but these changed header paths may conflict with some compilation
2213 environments. Enabled by default, and may be disabled using
2214 @option{--disable-canonical-system-headers}.
2216 @item --with-glibc-version=@var{major}.@var{minor}
2217 Tell GCC that when the GNU C Library (glibc) is used on the target it
2218 will be version @var{major}.@var{minor} or later. Normally this can
2219 be detected from the C library's header files, but this option may be
2220 needed when bootstrapping a cross toolchain without the header files
2221 available for building the initial bootstrap compiler.
2223 If GCC is configured with some multilibs that use glibc and some that
2224 do not, this option applies only to the multilibs that use glibc.
2225 However, such configurations may not work well as not all the relevant
2226 configuration in GCC is on a per-multilib basis.
2228 @item --enable-as-accelerator-for=@var{target}
2229 Build as offload target compiler. Specify offload host triple by @var{target}.
2231 @item --enable-offload-targets=@var{target1}[=@var{path1}],@dots{},@var{targetN}[=@var{pathN}]
2232 Enable offloading to targets @var{target1}, @dots{}, @var{targetN}.
2233 Offload compilers are expected to be already installed. Default search
2234 path for them is @file{@var{exec-prefix}}, but it can be changed by
2235 specifying paths @var{path1}, @dots{}, @var{pathN}.
2238 % @var{srcdir}/configure \
2239 --enable-offload-targets=x86_64-intelmicemul-linux-gnu=/path/to/x86_64/compiler,nvptx-none
2242 @item --with-hsa-runtime=@var{pathname}
2243 @itemx --with-hsa-runtime-include=@var{pathname}
2244 @itemx --with-hsa-runtime-lib=@var{pathname}
2246 If you configure GCC with offloading which uses an HSA run-time such as
2247 AMDGCN but do not have the HSA run-time library installed in a standard
2248 location then you can explicitly specify the directory where they are
2249 installed. The @option{--with-hsa-runtime=@/@var{hsainstalldir}} option
2251 @option{--with-hsa-runtime-lib=@/@var{hsainstalldir}/lib} and
2252 @option{--with-hsa-runtime-include=@/@var{hsainstalldir}/include}.
2255 @itemx --disable-cet
2256 Enable building target run-time libraries with control-flow
2257 instrumentation, see @option{-fcf-protection} option. When
2258 @code{--enable-cet} is specified target libraries are configured
2259 to add @option{-fcf-protection} and, if needed, other target
2260 specific options to a set of building options.
2262 @code{--enable-cet=auto} is default. CET is enabled on Linux/x86 if
2263 target binutils supports @code{Intel CET} instructions and disabled
2264 otherwise. In this case, the target libraries are configured to get
2265 additional @option{-fcf-protection} option.
2267 @item --with-riscv-attribute=@samp{yes}, @samp{no} or @samp{default}
2268 Generate RISC-V attribute by default, in order to record extra build
2269 information in object.
2271 The option is disabled by default. It is enabled on RISC-V/ELF (bare-metal)
2272 target if target binutils supported.
2274 @item --enable-s390-excess-float-precision
2275 @itemx --disable-s390-excess-float-precision
2276 On s390(x) targets, enable treatment of float expressions with double precision
2277 when in standards-compliant mode (e.g., when @code{--std=c99} or
2278 @code{-fexcess-precision=standard} are given).
2280 For a native build and cross compiles that have target headers, the option's
2281 default is derived from glibc's behavior. When glibc clamps float_t to double,
2282 GCC follows and enables the option. For other cross compiles, the default is
2286 @subheading Cross-Compiler-Specific Options
2287 The following options only apply to building cross compilers.
2290 @item --with-toolexeclibdir=@var{dir}
2291 Specify the installation directory for libraries built with a cross compiler.
2292 The default is @option{$@{gcc_tooldir@}/lib}.
2294 @item --with-sysroot
2295 @itemx --with-sysroot=@var{dir}
2296 Tells GCC to consider @var{dir} as the root of a tree that contains
2297 (a subset of) the root filesystem of the target operating system.
2298 Target system headers, libraries and run-time object files will be
2299 searched for in there. More specifically, this acts as if
2300 @option{--sysroot=@var{dir}} was added to the default options of the built
2301 compiler. The specified directory is not copied into the
2302 install tree, unlike the options @option{--with-headers} and
2303 @option{--with-libs} that this option obsoletes. The default value,
2304 in case @option{--with-sysroot} is not given an argument, is
2305 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
2306 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
2307 the GCC binaries if the installation tree is moved.
2309 This option affects the system root for the compiler used to build
2310 target libraries (which runs on the build system) and the compiler newly
2311 installed with @code{make install}; it does not affect the compiler which is
2312 used to build GCC itself.
2314 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2315 option then the compiler will search that directory within @var{dirname} for
2316 native system headers rather than the default @file{/usr/include}.
2318 @item --with-build-sysroot
2319 @itemx --with-build-sysroot=@var{dir}
2320 Tells GCC to consider @var{dir} as the system root (see
2321 @option{--with-sysroot}) while building target libraries, instead of
2322 the directory specified with @option{--with-sysroot}. This option is
2323 only useful when you are already using @option{--with-sysroot}. You
2324 can use @option{--with-build-sysroot} when you are configuring with
2325 @option{--prefix} set to a directory that is different from the one in
2326 which you are installing GCC and your target libraries.
2328 This option affects the system root for the compiler used to build
2329 target libraries (which runs on the build system); it does not affect
2330 the compiler which is used to build GCC itself.
2332 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2333 option then the compiler will search that directory within @var{dirname} for
2334 native system headers rather than the default @file{/usr/include}.
2336 @item --with-headers
2337 @itemx --with-headers=@var{dir}
2338 Deprecated in favor of @option{--with-sysroot}.
2339 Specifies that target headers are available when building a cross compiler.
2340 The @var{dir} argument specifies a directory which has the target include
2341 files. These include files will be copied into the @file{gcc} install
2342 directory. @emph{This option with the @var{dir} argument is required} when
2343 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
2344 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
2345 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
2346 will be run on these files to make them compatible with GCC@.
2348 @item --without-headers
2349 Tells GCC not use any target headers from a libc when building a cross
2350 compiler. When crossing to GNU/Linux, you need the headers so GCC
2351 can build the exception handling for libgcc.
2354 @itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}"
2355 Deprecated in favor of @option{--with-sysroot}.
2356 Specifies a list of directories which contain the target runtime
2357 libraries. These libraries will be copied into the @file{gcc} install
2358 directory. If the directory list is omitted, this option has no
2362 Specifies that @samp{newlib} is
2363 being used as the target C library. This causes @code{__eprintf} to be
2364 omitted from @file{libgcc.a} on the assumption that it will be provided by
2370 @item --with-avrlibc
2371 Only supported for the AVR target. Specifies that @samp{AVR-Libc} is
2372 being used as the target C@tie{} library. This causes float support
2373 functions like @code{__addsf3} to be omitted from @file{libgcc.a} on
2374 the assumption that it will be provided by @file{libm.a}. For more
2375 technical details, cf. @uref{http://gcc.gnu.org/PR54461,,PR54461}.
2376 It is not supported for
2377 RTEMS configurations, which currently use newlib. The option is
2378 supported since version 4.7.2 and is the default in 4.8.0 and newer.
2380 @item --with-double=@{32|64|32,64|64,32@}
2381 @itemx --with-long-double=@{32|64|32,64|64,32|double@}
2382 Only supported for the AVR target since version@tie{}10.
2383 Specify the default layout available for the C/C++ @samp{double}
2384 and @samp{long double} type, respectively. The following rules apply:
2387 The first value after the @samp{=} specifies the default layout (in bits)
2388 of the type and also the default for the @option{-mdouble=} resp.
2389 @option{-mlong-double=} compiler option.
2391 If more than one value is specified, respective multilib variants are
2392 available, and @option{-mdouble=} resp. @option{-mlong-double=} acts
2393 as a multilib option.
2395 If @option{--with-long-double=double} is specified, @samp{double} and
2396 @samp{long double} will have the same layout.
2398 The defaults are @option{--with-long-double=64,32} and
2399 @option{--with-double=32,64}. The default @samp{double} layout imposed by
2400 the latter is compatible with older versions of the compiler that implement
2401 @samp{double} as a 32-bit type, which does not comply to the language standard.
2403 Not all combinations of @option{--with-double=} and
2404 @option{--with-long-double=} are valid. For example, the combination
2405 @option{--with-double=32,64} @option{--with-long-double=32} will be
2406 rejected because the first option specifies the availability of
2407 multilibs for @samp{double}, whereas the second option implies
2408 that @samp{long double} --- and hence also @samp{double} --- is always
2411 @item --with-double-comparison=@{tristate|bool|libf7@}
2412 Only supported for the AVR target since version@tie{}10.
2413 Specify what result format is returned by library functions that
2414 compare 64-bit floating point values (@code{DFmode}).
2415 The GCC default is @samp{tristate}. If the floating point
2416 implementation returns a boolean instead, set it to @samp{bool}.
2418 @item --with-libf7=@{libgcc|math|math-symbols|no@}
2419 Only supported for the AVR target since version@tie{}10.
2420 Specify to which degree code from LibF7 is included in libgcc.
2421 LibF7 is an ad-hoc, AVR-specific, 64-bit floating point emulation
2422 written in C and (inline) assembly. @samp{libgcc} adds support
2423 for functions that one would usually expect in libgcc like double addition,
2424 double comparisons and double conversions. @samp{math} also adds routines
2425 that one would expect in @file{libm.a}, but with @code{__} (two underscores)
2426 prepended to the symbol names as specified by @file{math.h}.
2427 @samp{math-symbols} also defines weak aliases for the functions
2428 declared in @file{math.h}. However, @code{--with-libf7} won't
2429 install no @file{math.h} header file whatsoever, this file must come
2430 from elsewhere. This option sets @option{--with-double-comparison}
2433 @item --with-nds32-lib=@var{library}
2434 Specifies that @var{library} setting is used for building @file{libgcc.a}.
2435 Currently, the valid @var{library} is @samp{newlib} or @samp{mculib}.
2436 This option is only supported for the NDS32 target.
2438 @item --with-build-time-tools=@var{dir}
2439 Specifies where to find the set of target tools (assembler, linker, etc.)
2440 that will be used while building GCC itself. This option can be useful
2441 if the directory layouts are different between the system you are building
2442 GCC on, and the system where you will deploy it.
2444 For example, on an @samp{ia64-hp-hpux} system, you may have the GNU
2445 assembler and linker in @file{/usr/bin}, and the native tools in a
2446 different path, and build a toolchain that expects to find the
2447 native tools in @file{/usr/bin}.
2449 When you use this option, you should ensure that @var{dir} includes
2450 @command{ar}, @command{as}, @command{ld}, @command{nm},
2451 @command{ranlib} and @command{strip} if necessary, and possibly
2452 @command{objdump}. Otherwise, GCC may use an inconsistent set of
2456 @subsubheading Overriding @command{configure} test results
2458 Sometimes, it might be necessary to override the result of some
2459 @command{configure} test, for example in order to ease porting to a new
2460 system or work around a bug in a test. The toplevel @command{configure}
2461 script provides three variables for this:
2465 @item build_configargs
2466 @cindex @code{build_configargs}
2467 The contents of this variable is passed to all build @command{configure}
2470 @item host_configargs
2471 @cindex @code{host_configargs}
2472 The contents of this variable is passed to all host @command{configure}
2475 @item target_configargs
2476 @cindex @code{target_configargs}
2477 The contents of this variable is passed to all target @command{configure}
2482 In order to avoid shell and @command{make} quoting issues for complex
2483 overrides, you can pass a setting for @env{CONFIG_SITE} and set
2484 variables in the site file.
2486 @subheading Objective-C-Specific Options
2488 The following options apply to the build of the Objective-C runtime library.
2491 @item --enable-objc-gc
2492 Specify that an additional variant of the GNU Objective-C runtime library
2493 is built, using an external build of the Boehm-Demers-Weiser garbage
2494 collector (@uref{https://www.hboehm.info/gc/}). This library needs to be
2495 available for each multilib variant, unless configured with
2496 @option{--enable-objc-gc=@samp{auto}} in which case the build of the
2497 additional runtime library is skipped when not available and the build
2500 @item --with-target-bdw-gc=@var{list}
2501 @itemx --with-target-bdw-gc-include=@var{list}
2502 @itemx --with-target-bdw-gc-lib=@var{list}
2503 Specify search directories for the garbage collector header files and
2504 libraries. @var{list} is a comma separated list of key value pairs of the
2505 form @samp{@var{multilibdir}=@var{path}}, where the default multilib key
2506 is named as @samp{.} (dot), or is omitted (e.g.@:
2507 @samp{--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32}).
2509 The options @option{--with-target-bdw-gc-include} and
2510 @option{--with-target-bdw-gc-lib} must always be specified together
2511 for each multilib variant and they take precedence over
2512 @option{--with-target-bdw-gc}. If @option{--with-target-bdw-gc-include}
2513 is missing values for a multilib, then the value for the default
2514 multilib is used (e.g.@: @samp{--with-target-bdw-gc-include=/opt/bdw-gc/include}
2515 @samp{--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32}).
2516 If none of these options are specified, the library is assumed in
2520 @subheading D-Specific Options
2522 The following options apply to the build of the D runtime library.
2525 @item --enable-libphobos-checking
2526 @itemx --disable-libphobos-checking
2527 @itemx --enable-libphobos-checking=@var{list}
2528 This option controls whether run-time checks and contracts are compiled into
2529 the D runtime library. When the option is not specified, the library is built
2530 with @samp{release} checking. When the option is specified without a
2531 @var{list}, the result is the same as @samp{--enable-libphobos-checking=yes}.
2532 Likewise, @samp{--disable-libphobos-checking} is equivalent to
2533 @samp{--enable-libphobos-checking=no}.
2535 The categories of checks available in @var{list} are @samp{yes} (compiles
2536 libphobos with @option{-fno-release}), @samp{no} (compiles libphobos with
2537 @option{-frelease}), @samp{all} (same as @samp{yes}), @samp{none} or
2538 @samp{release} (same as @samp{no}).
2540 Individual checks available in @var{list} are @samp{assert} (compiles libphobos
2541 with an extra option @option{-fassert}).
2543 @item --with-libphobos-druntime-only
2544 @itemx --with-libphobos-druntime-only=@var{choice}
2545 Specify whether to build only the core D runtime library (druntime), or both
2546 the core and standard library (phobos) into libphobos. This is useful for
2547 targets that have full support in druntime, but no or incomplete support
2548 in phobos. @var{choice} can be one of @samp{auto}, @samp{yes}, and @samp{no}
2549 where @samp{auto} is the default.
2551 When the option is not specified, the default choice @samp{auto} means that it
2552 is inferred whether the target has support for the phobos standard library.
2553 When the option is specified without a @var{choice}, the result is the same as
2554 @samp{--with-libphobos-druntime-only=yes}.
2556 @item --with-target-system-zlib
2557 Use installed @samp{zlib} rather than that included with GCC@. This needs
2558 to be available for each multilib variant, unless configured with
2559 @option{--with-target-system-zlib=@samp{auto}} in which case the GCC@ included
2560 @samp{zlib} is only used when the system installed library is not available.
2568 @uref{./index.html,,Return to the GCC Installation page}
2572 @c ***Building****************************************************************
2574 @comment node-name, next, previous, up
2575 @node Building, Testing, Configuration, Installing GCC
2581 @cindex Installing GCC: Building
2583 Now that GCC is configured, you are ready to build the compiler and
2586 Some commands executed when making the compiler may fail (return a
2587 nonzero status) and be ignored by @command{make}. These failures, which
2588 are often due to files that were not found, are expected, and can safely
2591 It is normal to have compiler warnings when compiling certain files.
2592 Unless you are a GCC developer, you can generally ignore these warnings
2593 unless they cause compilation to fail. Developers should attempt to fix
2594 any warnings encountered, however they can temporarily continue past
2595 warnings-as-errors by specifying the configure flag
2596 @option{--disable-werror}.
2598 On certain old systems, defining certain environment variables such as
2599 @env{CC} can interfere with the functioning of @command{make}.
2601 If you encounter seemingly strange errors when trying to build the
2602 compiler in a directory other than the source directory, it could be
2603 because you have previously configured the compiler in the source
2604 directory. Make sure you have done all the necessary preparations.
2606 If you build GCC on a BSD system using a directory stored in an old System
2607 V file system, problems may occur in running @command{fixincludes} if the
2608 System V file system doesn't support symbolic links. These problems
2609 result in a failure to fix the declaration of @code{size_t} in
2610 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
2611 that type mismatches occur, this could be the cause.
2613 The solution is not to use such a directory for building GCC@.
2615 Similarly, when building from the source repository or snapshots, or if you modify
2616 @file{*.l} files, you need the Flex lexical analyzer generator
2617 installed. If you do not modify @file{*.l} files, releases contain
2618 the Flex-generated files and you do not need Flex installed to build
2619 them. There is still one Flex-based lexical analyzer (part of the
2620 build machinery, not of GCC itself) that is used even if you only
2621 build the C front end.
2623 When building from the source repository or snapshots, or if you modify Texinfo
2624 documentation, you need version 4.7 or later of Texinfo installed if you
2625 want Info documentation to be regenerated. Releases contain Info
2626 documentation pre-built for the unmodified documentation in the release.
2628 @section Building a native compiler
2630 For a native build, the default configuration is to perform
2631 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
2632 This will build the entire GCC system and ensure that it compiles
2633 itself correctly. It can be disabled with the @option{--disable-bootstrap}
2634 parameter to @samp{configure}, but bootstrapping is suggested because
2635 the compiler will be tested more completely and could also have
2638 The bootstrapping process will complete the following steps:
2642 Build tools necessary to build the compiler.
2645 Perform a 3-stage bootstrap of the compiler. This includes building
2646 three times the target tools for use by the compiler such as binutils
2647 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
2648 individually linked or moved into the top level GCC source tree before
2652 Perform a comparison test of the stage2 and stage3 compilers.
2655 Build runtime libraries using the stage3 compiler from the previous step.
2659 If you are short on disk space you might consider @samp{make
2660 bootstrap-lean} instead. The sequence of compilation is the
2661 same described above, but object files from the stage1 and
2662 stage2 of the 3-stage bootstrap of the compiler are deleted as
2663 soon as they are no longer needed.
2665 If you wish to use non-default GCC flags when compiling the stage2
2666 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
2667 doing @samp{make}. For example, if you want to save additional space
2668 during the bootstrap and in the final installation as well, you can
2669 build the compiler binaries without debugging information as in the
2670 following example. This will save roughly 40% of disk space both for
2671 the bootstrap and the final installation. (Libraries will still contain
2672 debugging information.)
2675 make BOOT_CFLAGS='-O' bootstrap
2678 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
2679 are less well tested here than the default of @samp{-g -O2}, but should
2680 still work. In a few cases, you may find that you need to specify special
2681 flags such as @option{-msoft-float} here to complete the bootstrap; or,
2682 if the native compiler miscompiles the stage1 compiler, you may need
2683 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
2684 of the stage1 compiler that were miscompiled, or by using @samp{make
2685 bootstrap4} to increase the number of stages of bootstrap.
2687 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
2688 Since these are always compiled with the compiler currently being
2689 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
2690 compilation flags, as for non-bootstrapped target libraries.
2691 Again, if the native compiler miscompiles the stage1 compiler, you may
2692 need to work around this by avoiding non-working parts of the stage1
2693 compiler. Use @code{STAGE1_TFLAGS} to this end.
2695 If you used the flag @option{--enable-languages=@dots{}} to restrict
2696 the compilers to be built, only those you've actually enabled will be
2697 built. This will of course only build those runtime libraries, for
2698 which the particular compiler has been built. Please note,
2699 that re-defining @env{LANGUAGES} when calling @samp{make}
2700 @strong{does not} work anymore!
2702 If the comparison of stage2 and stage3 fails, this normally indicates
2703 that the stage2 compiler has compiled GCC incorrectly, and is therefore
2704 a potentially serious bug which you should investigate and report. (On
2705 a few systems, meaningful comparison of object files is impossible; they
2706 always appear ``different''. If you encounter this problem, you will
2707 need to disable comparison in the @file{Makefile}.)
2709 If you do not want to bootstrap your compiler, you can configure with
2710 @option{--disable-bootstrap}. In particular cases, you may want to
2711 bootstrap your compiler even if the target system is not the same as
2712 the one you are building on: for example, you could build a
2713 @code{powerpc-unknown-linux-gnu} toolchain on a
2714 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
2715 @option{--enable-bootstrap} to the configure script.
2717 @code{BUILD_CONFIG} can be used to bring in additional customization
2718 to the build. It can be set to a whitespace-separated list of names.
2719 For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will
2720 be included by the top-level @file{Makefile}, bringing in any settings
2721 it contains. The default @code{BUILD_CONFIG} can be set using the
2722 configure option @option{--with-build-config=@code{NAME}...}. Some
2723 examples of supported build configurations are:
2726 @item @samp{bootstrap-O1}
2727 Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds
2728 @option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to
2729 @samp{BOOT_CFLAGS='-g -O1'}.
2731 @item @samp{bootstrap-O3}
2732 @itemx @samp{bootstrap-Og}
2733 Analogous to @code{bootstrap-O1}.
2735 @item @samp{bootstrap-lto}
2736 Enables Link-Time Optimization for host tools during bootstrapping.
2737 @samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding
2738 @option{-flto} to @samp{BOOT_CFLAGS}. This option assumes that the host
2739 supports the linker plugin (e.g.@: GNU ld version 2.21 or later or GNU gold
2740 version 2.21 or later).
2742 @item @samp{bootstrap-lto-noplugin}
2743 This option is similar to @code{bootstrap-lto}, but is intended for
2744 hosts that do not support the linker plugin. Without the linker plugin
2745 static libraries are not compiled with link-time optimizations. Since
2746 the GCC middle end and back end are in @file{libbackend.a} this means
2747 that only the front end is actually LTO optimized.
2749 @item @samp{bootstrap-lto-lean}
2750 This option is similar to @code{bootstrap-lto}, but is intended for
2751 faster build by only using LTO in the final bootstrap stage.
2752 With @samp{make profiledbootstrap} the LTO frontend
2753 is trained only on generator files.
2755 @item @samp{bootstrap-debug}
2756 Verifies that the compiler generates the same executable code, whether
2757 or not it is asked to emit debug information. To this end, this
2758 option builds stage2 host programs without debug information, and uses
2759 @file{contrib/compare-debug} to compare them with the stripped stage3
2760 object files. If @code{BOOT_CFLAGS} is overridden so as to not enable
2761 debug information, stage2 will have it, and stage3 won't. This option
2762 is enabled by default when GCC bootstrapping is enabled, if
2763 @code{strip} can turn object files compiled with and without debug
2764 info into identical object files. In addition to better test
2765 coverage, this option makes default bootstraps faster and leaner.
2767 @item @samp{bootstrap-debug-big}
2768 Rather than comparing stripped object files, as in
2769 @code{bootstrap-debug}, this option saves internal compiler dumps
2770 during stage2 and stage3 and compares them as well, which helps catch
2771 additional potential problems, but at a great cost in terms of disk
2772 space. It can be specified in addition to @samp{bootstrap-debug}.
2774 @item @samp{bootstrap-debug-lean}
2775 This option saves disk space compared with @code{bootstrap-debug-big},
2776 but at the expense of some recompilation. Instead of saving the dumps
2777 of stage2 and stage3 until the final compare, it uses
2778 @option{-fcompare-debug} to generate, compare and remove the dumps
2779 during stage3, repeating the compilation that already took place in
2780 stage2, whose dumps were not saved.
2782 @item @samp{bootstrap-debug-lib}
2783 This option tests executable code invariance over debug information
2784 generation on target libraries, just like @code{bootstrap-debug-lean}
2785 tests it on host programs. It builds stage3 libraries with
2786 @option{-fcompare-debug}, and it can be used along with any of the
2787 @code{bootstrap-debug} options above.
2789 There aren't @code{-lean} or @code{-big} counterparts to this option
2790 because most libraries are only build in stage3, so bootstrap compares
2791 would not get significant coverage. Moreover, the few libraries built
2792 in stage2 are used in stage3 host programs, so we wouldn't want to
2793 compile stage2 libraries with different options for comparison purposes.
2795 @item @samp{bootstrap-debug-ckovw}
2796 Arranges for error messages to be issued if the compiler built on any
2797 stage is run without the option @option{-fcompare-debug}. This is
2798 useful to verify the full @option{-fcompare-debug} testing coverage. It
2799 must be used along with @code{bootstrap-debug-lean} and
2800 @code{bootstrap-debug-lib}.
2802 @item @samp{bootstrap-cet}
2803 This option enables Intel CET for host tools during bootstrapping.
2804 @samp{BUILD_CONFIG=bootstrap-cet} is equivalent to adding
2805 @option{-fcf-protection} to @samp{BOOT_CFLAGS}. This option
2806 assumes that the host supports Intel CET (e.g.@: GNU assembler version
2809 @item @samp{bootstrap-time}
2810 Arranges for the run time of each program started by the GCC driver,
2811 built in any stage, to be logged to @file{time.log}, in the top level of
2814 @item @samp{bootstrap-asan}
2815 Compiles GCC itself using Address Sanitization in order to catch invalid memory
2816 accesses within the GCC code.
2818 @item @samp{bootstrap-hwasan}
2819 Compiles GCC itself using HWAddress Sanitization in order to catch invalid
2820 memory accesses within the GCC code. This option is only available on AArch64
2821 systems that are running Linux kernel version 5.4 or later.
2825 @section Building a cross compiler
2827 When building a cross compiler, it is not generally possible to do a
2828 3-stage bootstrap of the compiler. This makes for an interesting problem
2829 as parts of GCC can only be built with GCC@.
2831 To build a cross compiler, we recommend first building and installing a
2832 native compiler. You can then use the native GCC compiler to build the
2833 cross compiler. The installed native compiler needs to be GCC version
2836 Assuming you have already installed a native copy of GCC and configured
2837 your cross compiler, issue the command @command{make}, which performs the
2842 Build host tools necessary to build the compiler.
2845 Build target tools for use by the compiler such as binutils (bfd,
2846 binutils, gas, gprof, ld, and opcodes)
2847 if they have been individually linked or moved into the top level GCC source
2848 tree before configuring.
2851 Build the compiler (single stage only).
2854 Build runtime libraries using the compiler from the previous step.
2857 Note that if an error occurs in any step the make process will exit.
2859 If you are not building GNU binutils in the same source tree as GCC,
2860 you will need a cross-assembler and cross-linker installed before
2861 configuring GCC@. Put them in the directory
2862 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
2863 you should put in this directory:
2867 This should be the cross-assembler.
2870 This should be the cross-linker.
2873 This should be the cross-archiver: a program which can manipulate
2874 archive files (linker libraries) in the target machine's format.
2877 This should be a program to construct a symbol table in an archive file.
2880 The installation of GCC will find these programs in that directory,
2881 and copy or link them to the proper place to for the cross-compiler to
2882 find them when run later.
2884 The easiest way to provide these files is to build the Binutils package.
2885 Configure it with the same @option{--host} and @option{--target}
2886 options that you use for configuring GCC, then build and install
2887 them. They install their executables automatically into the proper
2888 directory. Alas, they do not support all the targets that GCC
2891 If you are not building a C library in the same source tree as GCC,
2892 you should also provide the target libraries and headers before
2893 configuring GCC, specifying the directories with
2894 @option{--with-sysroot} or @option{--with-headers} and
2895 @option{--with-libs}. Many targets also require ``start files'' such
2896 as @file{crt0.o} and
2897 @file{crtn.o} which are linked into each executable. There may be several
2898 alternatives for @file{crt0.o}, for use with profiling or other
2899 compilation options. Check your target's definition of
2900 @code{STARTFILE_SPEC} to find out what start files it uses.
2902 @section Building in parallel
2904 GNU Make 3.80 and above, which is necessary to build GCC, support
2905 building in parallel. To activate this, you can use @samp{make -j 2}
2906 instead of @samp{make}. You can also specify a bigger number, and
2907 in most cases using a value greater than the number of processors in
2908 your machine will result in fewer and shorter I/O latency hits, thus
2909 improving overall throughput; this is especially true for slow drives
2910 and network filesystems.
2912 @section Building the Ada compiler
2915 @ref{GNAT-prerequisite}.
2918 @uref{prerequisites.html#GNAT-prerequisite,,GNAT prerequisites}.
2921 @section Building with profile feedback
2923 It is possible to use profile feedback to optimize the compiler itself. This
2924 should result in a faster compiler binary. Experiments done on x86 using gcc
2925 3.3 showed approximately 7 percent speedup on compiling C programs. To
2926 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
2928 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
2929 compiler. This compiler is used to build a @code{stageprofile} compiler
2930 instrumented to collect execution counts of instruction and branch
2931 probabilities. Training run is done by building @code{stagetrain}
2932 compiler. Finally a @code{stagefeedback} compiler is built
2933 using the information collected.
2935 Unlike standard bootstrap, several additional restrictions apply. The
2936 compiler used to build @code{stage1} needs to support a 64-bit integral type.
2937 It is recommended to only use GCC for this.
2939 On Linux/x86_64 hosts with some restrictions (no virtualization) it is
2940 also possible to do autofdo build with @samp{make
2941 autoprofiledback}. This uses Linux perf to sample branches in the
2942 binary and then rebuild it with feedback derived from the profile.
2943 Linux perf and the @code{autofdo} toolkit needs to be installed for
2946 Only the profile from the current build is used, so when an error
2947 occurs it is recommended to clean before restarting. Otherwise
2948 the code quality may be much worse.
2955 @uref{./index.html,,Return to the GCC Installation page}
2959 @c ***Testing*****************************************************************
2961 @comment node-name, next, previous, up
2962 @node Testing, Final install, Building, Installing GCC
2966 @chapter Installing GCC: Testing
2969 @cindex Installing GCC: Testing
2972 Before you install GCC, we encourage you to run the testsuites and to
2973 compare your results with results from a similar configuration that have
2974 been submitted to the
2975 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
2976 Some of these archived results are linked from the build status lists
2977 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
2978 reports a successful build runs the testsuites and submits the results.
2979 This step is optional and may require you to download additional software,
2980 but it can give you confidence in your new GCC installation or point out
2981 problems before you install and start using your new GCC@.
2983 First, you must have @uref{download.html,,downloaded the testsuites}.
2984 These are part of the full distribution, but if you downloaded the
2985 ``core'' compiler plus any front ends, you must download the testsuites
2988 Second, you must have the testing tools installed. This includes
2989 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
2990 the DejaGnu site has links to these. For running the BRIG frontend
2991 tests, a tool to assemble the binary BRIGs from HSAIL text,
2992 @uref{https://github.com/HSAFoundation/HSAIL-Tools/,,HSAILasm} must
2995 If the directories where @command{runtest} and @command{expect} were
2996 installed are not in the @env{PATH}, you may need to set the following
2997 environment variables appropriately, as in the following example (which
2998 assumes that DejaGnu has been installed under @file{/usr/local}):
3001 TCL_LIBRARY = /usr/local/share/tcl8.0
3002 DEJAGNULIBS = /usr/local/share/dejagnu
3005 (On systems such as Cygwin, these paths are required to be actual
3006 paths, not mounts or links; presumably this is due to some lack of
3007 portability in the DejaGnu code.)
3010 Finally, you can run the testsuite (which may take a long time):
3012 cd @var{objdir}; make -k check
3015 This will test various components of GCC, such as compiler
3016 front ends and runtime libraries. While running the testsuite, DejaGnu
3017 might emit some harmless messages resembling
3018 @samp{WARNING: Couldn't find the global config file.} or
3019 @samp{WARNING: Couldn't find tool init file} that can be ignored.
3021 If you are testing a cross-compiler, you may want to run the testsuite
3022 on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}.
3024 @section How can you run the testsuite on selected tests?
3026 In order to run sets of tests selectively, there are targets
3027 @samp{make check-gcc} and language specific @samp{make check-c},
3028 @samp{make check-c++}, @samp{make check-d} @samp{make check-fortran},
3029 @samp{make check-ada}, @samp{make check-objc}, @samp{make check-obj-c++},
3030 @samp{make check-lto}
3031 in the @file{gcc} subdirectory of the object directory. You can also
3032 just run @samp{make check} in a subdirectory of the object directory.
3035 A more selective way to just run all @command{gcc} execute tests in the
3039 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
3042 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
3043 the testsuite with filenames matching @samp{9805*}, you would use
3046 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
3049 The file-matching expression following @var{filename}@command{.exp=} is treated
3050 as a series of whitespace-delimited glob expressions so that multiple patterns
3051 may be passed, although any whitespace must either be escaped or surrounded by
3052 single quotes if multiple expressions are desired. For example,
3055 make check-g++ RUNTESTFLAGS="old-deja.exp=9805*\ virtual2.c @var{other-options}"
3056 make check-g++ RUNTESTFLAGS="'old-deja.exp=9805* virtual2.c' @var{other-options}"
3059 The @file{*.exp} files are located in the testsuite directories of the GCC
3060 source, the most important ones being @file{compile.exp},
3061 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
3062 To get a list of the possible @file{*.exp} files, pipe the
3063 output of @samp{make check} into a file and look at the
3064 @samp{Running @dots{} .exp} lines.
3066 @section Passing options and running multiple testsuites
3068 You can pass multiple options to the testsuite using the
3069 @samp{--target_board} option of DejaGNU, either passed as part of
3070 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
3071 work outside the makefiles. For example,
3074 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
3077 will run the standard @command{g++} testsuites (``unix'' is the target name
3078 for a standard native testsuite situation), passing
3079 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
3080 slashes separate options.
3082 You can run the testsuites multiple times using combinations of options
3083 with a syntax similar to the brace expansion of popular shells:
3086 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
3089 (Note the empty option caused by the trailing comma in the final group.)
3090 The following will run each testsuite eight times using the @samp{arm-sim}
3091 target, as if you had specified all possible combinations yourself:
3094 --target_board='arm-sim/-mhard-float/-O1 \
3095 arm-sim/-mhard-float/-O2 \
3096 arm-sim/-mhard-float/-O3 \
3097 arm-sim/-mhard-float \
3098 arm-sim/-msoft-float/-O1 \
3099 arm-sim/-msoft-float/-O2 \
3100 arm-sim/-msoft-float/-O3 \
3101 arm-sim/-msoft-float'
3104 They can be combined as many times as you wish, in arbitrary ways. This
3108 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
3111 will generate four combinations, all involving @samp{-Wextra}.
3113 The disadvantage to this method is that the testsuites are run in serial,
3114 which is a waste on multiprocessor systems. For users with GNU Make and
3115 a shell which performs brace expansion, you can run the testsuites in
3116 parallel by having the shell perform the combinations and @command{make}
3117 do the parallel runs. Instead of using @samp{--target_board}, use a
3118 special makefile target:
3121 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
3127 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
3130 will run three concurrent ``make-gcc'' testsuites, eventually testing all
3131 ten combinations as described above. Note that this is currently only
3132 supported in the @file{gcc} subdirectory. (To see how this works, try
3133 typing @command{echo} before the example given here.)
3136 @section How to interpret test results
3138 The result of running the testsuite are various @file{*.sum} and @file{*.log}
3139 files in the testsuite subdirectories. The @file{*.log} files contain a
3140 detailed log of the compiler invocations and the corresponding
3141 results, the @file{*.sum} files summarize the results. These summaries
3142 contain status codes for all tests:
3146 PASS: the test passed as expected
3148 XPASS: the test unexpectedly passed
3150 FAIL: the test unexpectedly failed
3152 XFAIL: the test failed as expected
3154 UNSUPPORTED: the test is not supported on this platform
3156 ERROR: the testsuite detected an error
3158 WARNING: the testsuite detected a possible problem
3161 It is normal for some tests to report unexpected failures. At the
3162 current time the testing harness does not allow fine grained control
3163 over whether or not a test is expected to fail. This problem should
3164 be fixed in future releases.
3167 @section Submitting test results
3169 If you want to report the results to the GCC project, use the
3170 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
3173 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
3174 -m gcc-testresults@@gcc.gnu.org |sh
3177 This script uses the @command{Mail} program to send the results, so
3178 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
3179 prepended to the testsuite summary and should contain any special
3180 remarks you have on your results or your build environment. Please
3181 do not edit the testsuite result block or the subject line, as these
3182 messages may be automatically processed.
3189 @uref{./index.html,,Return to the GCC Installation page}
3193 @c ***Final install***********************************************************
3195 @comment node-name, next, previous, up
3196 @node Final install, , Testing, Installing GCC
3198 @ifset finalinstallhtml
3200 @chapter Installing GCC: Final installation
3203 Now that GCC has been built (and optionally tested), you can install it with
3205 cd @var{objdir} && make install
3208 We strongly recommend to install into a target directory where there is
3209 no previous version of GCC present. Also, the GNAT runtime should not
3210 be stripped, as this would break certain features of the debugger that
3211 depend on this debugging information (catching Ada exceptions for
3214 That step completes the installation of GCC; user level binaries can
3215 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
3216 you specified with the @option{--prefix} to configure (or
3217 @file{/usr/local} by default). (If you specified @option{--bindir},
3218 that directory will be used instead; otherwise, if you specified
3219 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
3220 Headers for the C++ library are installed in
3221 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
3222 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
3223 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
3224 in info format in @file{@var{infodir}} (normally
3225 @file{@var{prefix}/info}).
3227 When installing cross-compilers, GCC's executables
3228 are not only installed into @file{@var{bindir}}, that
3229 is, @file{@var{exec-prefix}/bin}, but additionally into
3230 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
3231 exists. Typically, such @dfn{tooldirs} hold target-specific
3232 binutils, including assembler and linker.
3234 Installation into a temporary staging area or into a @command{chroot}
3235 jail can be achieved with the command
3238 make DESTDIR=@var{path-to-rootdir} install
3242 where @var{path-to-rootdir} is the absolute path of
3243 a directory relative to which all installation paths will be
3244 interpreted. Note that the directory specified by @code{DESTDIR}
3245 need not exist yet; it will be created if necessary.
3247 There is a subtle point with tooldirs and @code{DESTDIR}:
3248 If you relocate a cross-compiler installation with
3249 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
3250 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
3251 be filled with duplicated GCC executables only if it already exists,
3252 it will not be created otherwise. This is regarded as a feature,
3253 not as a bug, because it gives slightly more control to the packagers
3254 using the @code{DESTDIR} feature.
3256 You can install stripped programs and libraries with
3262 If you are bootstrapping a released version of GCC then please
3263 quickly review the build status page for your release, available from
3264 @uref{http://gcc.gnu.org/buildstat.html}.
3265 If your system is not listed for the version of GCC that you built,
3267 @email{gcc@@gcc.gnu.org} indicating
3268 that you successfully built and installed GCC@.
3269 Include the following information:
3273 Output from running @file{@var{srcdir}/config.guess}. Do not send
3274 that file itself, just the one-line output from running it.
3277 The output of @samp{gcc -v} for your newly installed @command{gcc}.
3278 This tells us which version of GCC you built and the options you passed to
3282 Whether you enabled all languages or a subset of them. If you used a
3283 full distribution then this information is part of the configure
3284 options in the output of @samp{gcc -v}, but if you downloaded the
3285 ``core'' compiler plus additional front ends then it isn't apparent
3286 which ones you built unless you tell us about it.
3289 If the build was for GNU/Linux, also include:
3292 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
3293 this information should be available from @file{/etc/issue}.
3296 The version of the Linux kernel, available from @samp{uname --version}
3300 The version of glibc you used; for RPM-based systems like Red Hat,
3301 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
3302 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
3304 For other systems, you can include similar information if you think it is
3308 Any other information that you think would be useful to people building
3309 GCC on the same configuration. The new entry in the build status list
3310 will include a link to the archived copy of your message.
3313 We'd also like to know if the
3315 @ref{Specific, host/target specific installation notes}
3318 @uref{specific.html,,host/target specific installation notes}
3320 didn't include your host/target information or if that information is
3321 incomplete or out of date. Send a note to
3322 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
3324 If you find a bug, please report it following the
3325 @uref{../bugs/,,bug reporting guidelines}.
3327 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
3328 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
3329 and @TeX{} installed. This creates a number of @file{.dvi} files in
3330 subdirectories of @file{@var{objdir}}; these may be converted for
3331 printing with programs such as @command{dvips}. Alternately, by using
3332 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
3333 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
3334 is included with Texinfo version 4.8 and later. You can also
3335 @uref{https://shop.fsf.org/,,buy printed manuals from the
3336 Free Software Foundation}, though such manuals may not be for the most
3337 recent version of GCC@.
3339 If you would like to generate online HTML documentation, do @samp{cd
3340 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
3341 @file{@var{objdir}/gcc/HTML}.
3348 @uref{./index.html,,Return to the GCC Installation page}
3352 @c ***Binaries****************************************************************
3354 @comment node-name, next, previous, up
3355 @node Binaries, Specific, Installing GCC, Top
3359 @chapter Installing GCC: Binaries
3362 @cindex Installing GCC: Binaries
3364 We are often asked about pre-compiled versions of GCC@. While we cannot
3365 provide these for all platforms, below you'll find links to binaries for
3366 various platforms where creating them by yourself is not easy due to various
3369 Please note that we did not create these binaries, nor do we
3370 support them. If you have any problems installing them, please
3371 contact their makers.
3378 @uref{http://www.bullfreeware.com,,Bull's Open Source Software Archive for
3379 for AIX 6 and AIX 7};
3382 @uref{http://www.perzl.org/aix/,,AIX Open Source Packages (AIX5L AIX 6.1
3387 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
3393 @uref{http://hpux.connect.org.uk/,,HP-UX Porting Center};
3397 Solaris 2 (SPARC, Intel):
3400 @uref{https://www.opencsw.org/,,OpenCSW}
3407 The @uref{https://brew.sh,,Homebrew} package manager;
3409 @uref{https://www.macports.org,,MacPorts}.
3416 The @uref{https://sourceware.org/cygwin/,,Cygwin} project;
3418 The @uref{http://www.mingw.org/,,MinGW} and
3419 @uref{http://mingw-w64.org/doku.php,,mingw-w64} projects.
3423 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
3424 number of platforms.
3427 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
3428 links to GNU Fortran binaries for several platforms.
3436 @uref{./index.html,,Return to the GCC Installation page}
3440 @c ***Specific****************************************************************
3442 @comment node-name, next, previous, up
3443 @node Specific, Old, Binaries, Top
3447 @chapter Host/target specific installation notes for GCC
3450 @cindex Specific installation notes
3451 @cindex Target specific installation
3452 @cindex Host specific installation
3453 @cindex Target specific installation notes
3455 Please read this document carefully @emph{before} installing the
3456 GNU Compiler Collection on your machine.
3458 Note that this list of install notes is @emph{not} a list of supported
3459 hosts or targets. Not all supported hosts and targets are listed
3460 here, only the ones that require host-specific or target-specific
3461 information have to.
3466 @uref{#aarch64-x-x,,aarch64*-*-*}
3468 @uref{#alpha-x-x,,alpha*-*-*}
3470 @uref{#amd64-x-solaris2,,amd64-*-solaris2*}
3472 @uref{#arm-x-eabi,,arm-*-eabi}
3476 @uref{#bfin,,Blackfin}
3480 @uref{#x-x-freebsd,,*-*-freebsd*}
3482 @uref{#h8300-hms,,h8300-hms}
3484 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
3486 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
3488 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
3490 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
3492 @uref{#ix86-x-linux,,i?86-*-linux*}
3494 @uref{#ix86-x-solaris2,,i?86-*-solaris2*}
3496 @uref{#ia64-x-linux,,ia64-*-linux}
3498 @uref{#ia64-x-hpux,,ia64-*-hpux*}
3500 @uref{#x-ibm-aix,,*-ibm-aix*}
3502 @uref{#iq2000-x-elf,,iq2000-*-elf}
3504 @uref{#lm32-x-elf,,lm32-*-elf}
3506 @uref{#lm32-x-uclinux,,lm32-*-uclinux}
3508 @uref{#m32c-x-elf,,m32c-*-elf}
3510 @uref{#m32r-x-elf,,m32r-*-elf}
3512 @uref{#m68k-x-x,,m68k-*-*}
3514 @uref{#m68k-uclinux,,m68k-uclinux}
3516 @uref{#microblaze-x-elf,,microblaze-*-elf}
3518 @uref{#mips-x-x,,mips-*-*}
3520 @uref{#nds32le-x-elf,,nds32le-*-elf}
3522 @uref{#nds32be-x-elf,,nds32be-*-elf}
3524 @uref{#nvptx-x-none,,nvptx-*-none}
3526 @uref{#or1k-x-elf,,or1k-*-elf}
3528 @uref{#or1k-x-linux,,or1k-*-linux}
3530 @uref{#powerpc-x-x,,powerpc*-*-*}
3532 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
3534 @uref{#powerpc-x-elf,,powerpc-*-elf}
3536 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
3538 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
3540 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
3542 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
3544 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
3546 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
3548 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
3550 @uref{#riscv32-x-elf,,riscv32-*-elf}
3552 @uref{#riscv32-x-linux,,riscv32-*-linux}
3554 @uref{#riscv64-x-elf,,riscv64-*-elf}
3556 @uref{#riscv64-x-linux,,riscv64-*-linux}
3558 @uref{#s390-x-linux,,s390-*-linux*}
3560 @uref{#s390x-x-linux,,s390x-*-linux*}
3562 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
3564 @uref{#x-x-solaris2,,*-*-solaris2*}
3566 @uref{#sparc-x-x,,sparc*-*-*}
3568 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
3570 @uref{#sparc-x-linux,,sparc-*-linux*}
3572 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
3574 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
3576 @uref{#c6x-x-x,,c6x-*-*}
3578 @uref{#tilegx-x-linux,,tilegx-*-linux*}
3580 @uref{#tilegxbe-x-linux,,tilegxbe-*-linux*}
3582 @uref{#tilepro-x-linux,,tilepro-*-linux*}
3584 @uref{#visium-x-elf, visium-*-elf}
3586 @uref{#x-x-vxworks,,*-*-vxworks*}
3588 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
3590 @uref{#x86-64-x-solaris2,,x86_64-*-solaris2*}
3592 @uref{#xtensa-x-elf,,xtensa*-*-elf}
3594 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
3596 @uref{#windows,,Microsoft Windows}
3598 @uref{#x-x-cygwin,,*-*-cygwin}
3600 @uref{#x-x-mingw32,,*-*-mingw32}
3604 @uref{#older,,Older systems}
3609 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
3615 <!-- -------- host/target specific issues start here ---------------- -->
3618 @anchor{aarch64-x-x}
3619 @heading aarch64*-*-*
3620 Binutils pre 2.24 does not have support for selecting @option{-mabi} and
3621 does not support ILP32. If it is used to build GCC 4.9 or later, GCC will
3622 not support option @option{-mabi=ilp32}.
3624 To enable a workaround for the Cortex-A53 erratum number 835769 by default
3625 (for all CPUs regardless of -mcpu option given) at configure time use the
3626 @option{--enable-fix-cortex-a53-835769} option. This will enable the fix by
3627 default and can be explicitly disabled during compilation by passing the
3628 @option{-mno-fix-cortex-a53-835769} option. Conversely,
3629 @option{--disable-fix-cortex-a53-835769} will disable the workaround by
3630 default. The workaround is disabled by default if neither of
3631 @option{--enable-fix-cortex-a53-835769} or
3632 @option{--disable-fix-cortex-a53-835769} is given at configure time.
3634 To enable a workaround for the Cortex-A53 erratum number 843419 by default
3635 (for all CPUs regardless of -mcpu option given) at configure time use the
3636 @option{--enable-fix-cortex-a53-843419} option. This workaround is applied at
3637 link time. Enabling the workaround will cause GCC to pass the relevant option
3638 to the linker. It can be explicitly disabled during compilation by passing the
3639 @option{-mno-fix-cortex-a53-843419} option. Conversely,
3640 @option{--disable-fix-cortex-a53-843419} will disable the workaround by default.
3641 The workaround is disabled by default if neither of
3642 @option{--enable-fix-cortex-a53-843419} or
3643 @option{--disable-fix-cortex-a53-843419} is given at configure time.
3645 To enable Branch Target Identification Mechanism and Return Address Signing by
3646 default at configure time use the @option{--enable-standard-branch-protection}
3647 option. This is equivalent to having @option{-mbranch-protection=standard}
3648 during compilation. This can be explicitly disabled during compilation by
3649 passing the @option{-mbranch-protection=none} option which turns off all
3650 types of branch protections. Conversely,
3651 @option{--disable-standard-branch-protection} will disable both the
3652 protections by default. This mechanism is turned off by default if neither
3653 of the options are given at configure time.
3660 This section contains general configuration information for all
3661 Alpha-based platforms using ELF@. In addition to reading this
3662 section, please read all other sections that match your target.
3667 @anchor{amd64-x-solaris2}
3668 @heading amd64-*-solaris2*
3669 This is a synonym for @samp{x86_64-*-solaris2*}.
3674 @anchor{amdgcn-x-amdhsa}
3675 @heading amdgcn-*-amdhsa
3678 Instead of GNU Binutils, you will need to install LLVM 6, or later, and copy
3679 @file{bin/llvm-mc} to @file{amdgcn-amdhsa/bin/as},
3680 @file{bin/lld} to @file{amdgcn-amdhsa/bin/ld},
3681 @file{bin/llvm-nm} to @file{amdgcn-amdhsa/bin/nm}, and
3682 @file{bin/llvm-ar} to both @file{bin/amdgcn-amdhsa-ar} and
3683 @file{bin/amdgcn-amdhsa-ranlib}.
3685 Use Newlib (2019-01-16, or newer).
3687 To run the binaries, install the HSA Runtime from the
3688 @uref{https://rocm.github.io,,ROCm Platform}, and use
3689 @file{libexec/gcc/amdhsa-amdhsa/@var{version}/gcn-run} to launch them
3695 @anchor{arc-x-elf32}
3696 @heading arc-*-elf32
3698 Use @samp{configure --target=arc-elf32 --with-cpu=@var{cpu} --enable-languages="c,c++"}
3699 to configure GCC, with @var{cpu} being one of @samp{arc600}, @samp{arc601},
3705 @anchor{arc-linux-uclibc}
3706 @heading arc-linux-uclibc
3708 Use @samp{configure --target=arc-linux-uclibc --with-cpu=arc700 --enable-languages="c,c++"} to configure GCC@.
3715 ARM-family processors.
3717 Building the Ada frontend commonly fails (an infinite loop executing
3718 @code{xsinfo}) if the host compiler is GNAT 4.8. Host compilers built from the
3719 GNAT 4.6, 4.9 or 5 release branches are known to succeed.
3726 ATMEL AVR-family micro controllers. These are used in embedded
3727 applications. There are no standard Unix configurations.
3729 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
3733 See ``AVR Options'' in the main manual
3735 for the list of supported MCU types.
3737 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
3739 Further installation notes and other useful information about AVR tools
3740 can also be obtained from:
3744 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
3746 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
3749 The following error:
3751 Error: register required
3754 indicates that you should upgrade to a newer version of the binutils.
3761 The Blackfin processor, an Analog Devices DSP.
3763 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
3767 See ``Blackfin Options'' in the main manual
3770 More information, and a version of binutils with support for this processor,
3771 are available at @uref{https://sourceforge.net/projects/adi-toolchain/}.
3778 The CR16 CompactRISC architecture is a 16-bit architecture. This
3779 architecture is used in embedded applications.
3782 @xref{CR16 Options,, CR16 Options, gcc, Using and Porting the GNU Compiler
3787 See ``CR16 Options'' in the main manual for a list of CR16-specific options.
3790 Use @samp{configure --target=cr16-elf --enable-languages=c,c++} to configure
3791 GCC@ for building a CR16 elf cross-compiler.
3793 Use @samp{configure --target=cr16-uclinux --enable-languages=c,c++} to
3794 configure GCC@ for building a CR16 uclinux cross-compiler.
3801 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
3802 series. These are used in embedded applications.
3805 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
3809 See ``CRIS Options'' in the main manual
3811 for a list of CRIS-specific options.
3813 There are a few different CRIS targets:
3816 Mainly for monolithic embedded systems. Includes a multilib for the
3817 @samp{v10} core used in @samp{ETRAX 100 LX}.
3818 @item cris-axis-linux-gnu
3819 A GNU/Linux port for the CRIS architecture, currently targeting
3820 @samp{ETRAX 100 LX} by default.
3823 Pre-packaged tools can be obtained from
3824 @uref{ftp://ftp.axis.com/@/pub/@/axis/@/tools/@/cris/@/compiler-kit/}. More
3825 information about this platform is available at
3826 @uref{http://developer.axis.com/}.
3833 Please have a look at the @uref{binaries.html,,binaries page}.
3835 You cannot install GCC by itself on MSDOS; it will not compile under
3836 any MSDOS compiler except itself. You need to get the complete
3837 compilation package DJGPP, which includes binaries as well as sources,
3838 and includes all the necessary compilation tools and libraries.
3843 @anchor{epiphany-x-elf}
3844 @heading epiphany-*-elf
3846 This configuration is intended for embedded systems.
3851 @anchor{x-x-freebsd}
3852 @heading *-*-freebsd*
3853 Support for FreeBSD 1 was discontinued in GCC 3.2. Support for
3854 FreeBSD 2 (and any mutant a.out variants of FreeBSD 3) was
3855 discontinued in GCC 4.0.
3857 In order to better utilize FreeBSD base system functionality and match
3858 the configuration of the system compiler, GCC 4.5 and above as well as
3859 GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present
3860 on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default
3861 (on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside
3862 @file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled
3863 by GCC 4.5 and above.
3865 We support FreeBSD using the ELF file format with DWARF 2 debugging
3866 for all CPU architectures. You may use @option{-gstabs} instead of
3867 @option{-g}, if you really want the old debugging format. There are
3868 no known issues with mixing object files and libraries with different
3869 debugging formats. Otherwise, this release of GCC should now match
3870 more of the configuration used in the stock FreeBSD configuration of
3871 GCC@. In particular, @option{--enable-threads} is now configured by
3872 default. However, as a general user, do not attempt to replace the
3873 system compiler with this release. Known to bootstrap and check with
3874 good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap
3875 and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4,
3876 4.5, 4.8, 4.9 and 5-CURRENT@.
3878 The version of binutils installed in @file{/usr/bin} probably works
3879 with this release of GCC@. Bootstrapping against the latest GNU
3880 binutils and/or the version found in @file{/usr/ports/devel/binutils} has
3881 been known to enable additional features and improve overall testsuite
3882 results. However, it is currently known that boehm-gc may not configure
3883 properly on FreeBSD prior to the FreeBSD 7.0 release with GNU binutils
3892 This configuration is intended for embedded systems.
3899 Renesas H8/300 series of processors.
3901 Please have a look at the @uref{binaries.html,,binaries page}.
3903 The calling convention and structure layout has changed in release 2.6.
3904 All code must be recompiled. The calling convention now passes the
3905 first three arguments in function calls in registers. Structures are no
3906 longer a multiple of 2 bytes.
3911 @anchor{hppa-hp-hpux}
3912 @heading hppa*-hp-hpux*
3913 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
3915 We require using gas/binutils on all hppa platforms. Version 2.19 or
3916 later is recommended.
3918 It may be helpful to configure GCC with the
3919 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
3920 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
3922 The HP assembler should not be used with GCC. It is rarely tested and may
3923 not work. It shouldn't be used with any languages other than C due to its
3926 Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging
3927 format which GCC does not know about). It also inserts timestamps
3928 into each object file it creates, causing the 3-stage comparison test to
3929 fail during a bootstrap. You should be able to continue by saying
3930 @samp{make all-host all-target} after getting the failure from @samp{make}.
3932 Various GCC features are not supported. For example, it does not support weak
3933 symbols or alias definitions. As a result, explicit template instantiations
3934 are required when using C++. This makes it difficult if not impossible to
3935 build many C++ applications.
3937 There are two default scheduling models for instructions. These are
3938 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
3939 architecture specified for the target machine when configuring.
3940 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
3941 the target is a @samp{hppa1*} machine.
3943 The PROCESSOR_8000 model is not well suited to older processors. Thus,
3944 it is important to completely specify the machine architecture when
3945 configuring if you want a model other than PROCESSOR_8000. The macro
3946 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
3947 default scheduling model is desired.
3949 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
3950 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
3951 This namespace change might cause problems when bootstrapping with
3952 an earlier version of GCC or the HP compiler as essentially the same
3953 namespace is required for an entire build. This problem can be avoided
3954 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
3955 or @samp{98}. Another way is to add an appropriate set of predefines
3956 to @env{CC}. The description for the @option{munix=} option contains
3957 a list of the predefines used with each standard.
3959 More specific information to @samp{hppa*-hp-hpux*} targets follows.
3964 @anchor{hppa-hp-hpux10}
3965 @heading hppa*-hp-hpux10
3966 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
3967 @code{PHCO_19798} from HP@.
3969 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
3970 used for one-only code and data. This resolves many of the previous
3971 problems in using C++ on this target. However, the ABI is not compatible
3972 with the one implemented under HP-UX 11 using secondary definitions.
3977 @anchor{hppa-hp-hpux11}
3978 @heading hppa*-hp-hpux11
3979 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
3980 be used to compile GCC 3.0 and up.
3982 The libffi library haven't been ported to 64-bit HP-UX@ and doesn't build.
3984 Refer to @uref{binaries.html,,binaries} for information about obtaining
3985 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
3986 to build the Ada language as it cannot be bootstrapped using C@. Ada is
3987 only available for the 32-bit PA-RISC runtime.
3989 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
3990 bundled compiler supports only traditional C; you will need either HP's
3991 unbundled compiler, or a binary distribution of GCC@.
3993 It is possible to build GCC 3.3 starting with the bundled HP compiler,
3994 but the process requires several steps. GCC 3.3 can then be used to
3995 build later versions.
3997 There are several possible approaches to building the distribution.
3998 Binutils can be built first using the HP tools. Then, the GCC
3999 distribution can be built. The second approach is to build GCC
4000 first using the HP tools, then build binutils, then rebuild GCC@.
4001 There have been problems with various binary distributions, so it
4002 is best not to start from a binary distribution.
4004 On 64-bit capable systems, there are two distinct targets. Different
4005 installation prefixes must be used if both are to be installed on
4006 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
4007 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
4008 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
4009 PA-RISC 2.0 architecture.
4011 The script config.guess now selects the target type based on the compiler
4012 detected during configuration. You must define @env{PATH} or @env{CC} so
4013 that configure finds an appropriate compiler for the initial bootstrap.
4014 When @env{CC} is used, the definition should contain the options that are
4015 needed whenever @env{CC} is used.
4017 Specifically, options that determine the runtime architecture must be
4018 in @env{CC} to correctly select the target for the build. It is also
4019 convenient to place many other compiler options in @env{CC}. For example,
4020 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
4021 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
4022 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
4023 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
4024 macro definition table of cpp needs to be increased for a successful
4025 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
4026 be defined when building with the bundled compiler, or when using the
4027 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
4029 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
4030 with the @option{--with-ld=@dots{}} option. This overrides the standard
4031 search for ld. The two linkers supported on this target require different
4032 commands. The default linker is determined during configuration. As a
4033 result, it's not possible to switch linkers in the middle of a GCC build.
4034 This has been reported to sometimes occur in unified builds of binutils
4037 A recent linker patch must be installed for the correct operation of
4038 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
4039 oldest linker patches that are known to work. They are for HP-UX
4040 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
4041 @code{PHSS_24304}, might be usable but it hasn't been tested. These
4042 patches have been superseded. Consult the HP patch database to obtain
4043 the currently recommended linker patch for your system.
4045 The patches are necessary for the support of weak symbols on the
4046 32-bit port, and for the running of initializers and finalizers. Weak
4047 symbols are implemented using SOM secondary definition symbols. Prior
4048 to HP-UX 11, there are bugs in the linker support for secondary symbols.
4049 The patches correct a problem of linker core dumps creating shared
4050 libraries containing secondary symbols, as well as various other
4051 linking issues involving secondary symbols.
4053 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
4054 run initializers and finalizers on the 64-bit port. The 32-bit port
4055 uses the linker @option{+init} and @option{+fini} options for the same
4056 purpose. The patches correct various problems with the +init/+fini
4057 options, including program core dumps. Binutils 2.14 corrects a
4058 problem on the 64-bit port resulting from HP's non-standard use of
4059 the .init and .fini sections for array initializers and finalizers.
4061 Although the HP and GNU linkers are both supported for the
4062 @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the
4063 HP linker be used for link editing on this target.
4065 At this time, the GNU linker does not support the creation of long
4066 branch stubs. As a result, it cannot successfully link binaries
4067 containing branch offsets larger than 8 megabytes. In addition,
4068 there are problems linking shared libraries, linking executables
4069 with @option{-static}, and with dwarf2 unwind and exception support.
4070 It also doesn't provide stubs for internal calls to global functions
4071 in shared libraries, so these calls cannot be overloaded.
4073 The HP dynamic loader does not support GNU symbol versioning, so symbol
4074 versioning is not supported. It may be necessary to disable symbol
4075 versioning with @option{--disable-symvers} when using GNU ld.
4077 POSIX threads are the default. The optional DCE thread library is not
4078 supported, so @option{--enable-threads=dce} does not work.
4083 @anchor{x-x-linux-gnu}
4084 @heading *-*-linux-gnu
4085 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
4086 in glibc 2.2.5 and later. More information is available in the
4087 libstdc++-v3 documentation.
4092 @anchor{ix86-x-linux}
4093 @heading i?86-*-linux*
4094 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
4095 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
4097 If you receive Signal 11 errors when building on GNU/Linux, then it is
4098 possible you have a hardware problem. Further information on this can be
4099 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
4104 @anchor{ix86-x-solaris2}
4105 @heading i?86-*-solaris2*
4106 Use this for Solaris 11.3 or later on x86 and x86-64 systems. Starting
4107 with GCC 4.7, there is also a 64-bit @samp{amd64-*-solaris2*} or
4108 @samp{x86_64-*-solaris2*} configuration that corresponds to
4109 @samp{sparcv9-sun-solaris2*}.
4111 It is recommended that you configure GCC to use the GNU assembler. The
4112 versions included in Solaris 11.3, from GNU binutils 2.23.1 or
4113 newer (available as @file{/usr/bin/gas} and
4114 @file{/usr/gnu/bin/as}), work fine. The current version, from GNU
4115 binutils 2.34, is known to work. Recent versions of the Solaris assembler in
4116 @file{/usr/bin/as} work almost as well, though.
4118 For linking, the Solaris linker is preferred. If you want to use the GNU
4119 linker instead, the version in Solaris 11.3, from GNU binutils 2.23.1 or
4120 newer (in @file{/usr/gnu/bin/ld} and @file{/usr/bin/gld}), works,
4121 as does the latest version, from GNU binutils 2.34.
4123 To use GNU @command{as}, configure with the options
4124 @option{--with-gnu-as --with-as=@//usr/@/gnu/@/bin/@/as}. It may be necessary
4125 to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to
4126 guarantee use of Solaris @command{ld}.
4127 @c FIXME: why --without-gnu-ld --with-ld?
4132 @anchor{ia64-x-linux}
4133 @heading ia64-*-linux
4134 IA-64 processor (also known as IPF, or Itanium Processor Family)
4137 If you are using the installed system libunwind library with
4138 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
4141 None of the following versions of GCC has an ABI that is compatible
4142 with any of the other versions in this list, with the exception that
4143 Red Hat 2.96 and Trillian 000171 are compatible with each other:
4144 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
4145 This primarily affects C++ programs and programs that create shared libraries.
4146 GCC 3.1 or later is recommended for compiling linux, the kernel.
4147 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
4148 more major ABI changes are expected.
4153 @anchor{ia64-x-hpux}
4154 @heading ia64-*-hpux*
4155 Building GCC on this target requires the GNU Assembler. The bundled HP
4156 assembler will not work. To prevent GCC from using the wrong assembler,
4157 the option @option{--with-gnu-as} may be necessary.
4159 The GCC libunwind library has not been ported to HPUX@. This means that for
4160 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
4161 is required to build GCC@. For GCC 3.3 and later, this is the default.
4162 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
4163 removed and the system libunwind library will always be used.
4167 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
4171 Support for AIX version 3 and older was discontinued in GCC 3.4.
4172 Support for AIX version 4.2 and older was discontinued in GCC 4.5.
4174 ``out of memory'' bootstrap failures may indicate a problem with
4175 process resource limits (ulimit). Hard limits are configured in the
4176 @file{/etc/security/limits} system configuration file.
4178 GCC 4.9 and above require a C++ compiler for bootstrap. IBM VAC++ / xlC
4179 cannot bootstrap GCC. xlc can bootstrap an older version of GCC and
4180 G++ can bootstrap recent releases of GCC.
4182 GCC can bootstrap with recent versions of IBM XLC, but bootstrapping
4183 with an earlier release of GCC is recommended. Bootstrapping with XLC
4184 requires a larger data segment, which can be enabled through the
4185 @var{LDR_CNTRL} environment variable, e.g.,
4188 % LDR_CNTRL=MAXDATA=0x50000000
4192 One can start with a pre-compiled version of GCC to build from
4193 sources. One may delete GCC's ``fixed'' header files when starting
4194 with a version of GCC built for an earlier release of AIX.
4196 To speed up the configuration phases of bootstrapping and installing GCC,
4197 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
4200 % CONFIG_SHELL=/opt/freeware/bin/bash
4201 % export CONFIG_SHELL
4204 and then proceed as described in @uref{build.html,,the build
4205 instructions}, where we strongly recommend specifying an absolute path
4206 to invoke @var{srcdir}/configure.
4208 Because GCC on AIX is built as a 32-bit executable by default,
4209 (although it can generate 64-bit programs) the GMP and MPFR libraries
4210 required by gfortran must be 32-bit libraries. Building GMP and MPFR
4211 as static archive libraries works better than shared libraries.
4213 Errors involving @code{alloca} when building GCC generally are due
4214 to an incorrect definition of @code{CC} in the Makefile or mixing files
4215 compiled with the native C compiler and GCC@. During the stage1 phase of
4216 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
4217 (not @command{xlc}). Once @command{configure} has been informed of
4218 @command{xlc}, one needs to use @samp{make distclean} to remove the
4219 configure cache files and ensure that @env{CC} environment variable
4220 does not provide a definition that will confuse @command{configure}.
4221 If this error occurs during stage2 or later, then the problem most likely
4222 is the version of Make (see above).
4224 The native @command{as} and @command{ld} are recommended for
4225 bootstrapping on AIX@. The GNU Assembler, GNU Linker, and GNU
4226 Binutils version 2.20 is the minimum level that supports bootstrap on
4227 AIX 5@. The GNU Assembler has not been updated to support AIX 6@ or
4228 AIX 7. The native AIX tools do interoperate with GCC@.
4230 AIX 7.1 added partial support for DWARF debugging, but full support
4231 requires AIX 7.1 TL03 SP7 that supports additional DWARF sections and
4232 fixes a bug in the assembler. AIX 7.1 TL03 SP5 distributed a version
4233 of libm.a missing important symbols; a fix for IV77796 will be
4236 AIX 5.3 TL10, AIX 6.1 TL05 and AIX 7.1 TL00 introduced an AIX
4237 assembler change that sometimes produces corrupt assembly files
4238 causing AIX linker errors. The bug breaks GCC bootstrap on AIX and
4239 can cause compilation failures with existing GCC installations. An
4240 AIX iFix for AIX 5.3 is available (APAR IZ98385 for AIX 5.3 TL10, APAR
4241 IZ98477 for AIX 5.3 TL11 and IZ98134 for AIX 5.3 TL12). AIX 5.3 TL11 SP8,
4242 AIX 5.3 TL12 SP5, AIX 6.1 TL04 SP11, AIX 6.1 TL05 SP7, AIX 6.1 TL06 SP6,
4243 AIX 6.1 TL07 and AIX 7.1 TL01 should include the fix.
4245 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
4246 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
4247 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
4248 referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1)
4250 @anchor{TransferAixShobj}
4251 @samp{libstdc++} in GCC 3.4 increments the major version number of the
4252 shared object and GCC installation places the @file{libstdc++.a}
4253 shared library in a common location which will overwrite the and GCC
4254 3.3 version of the shared library. Applications either need to be
4255 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
4256 versions of the @samp{libstdc++} shared object needs to be available
4257 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
4258 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
4259 installed for runtime dynamic loading using the following steps to set
4260 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
4261 multilib @file{libstdc++.a} installed:
4263 Extract the shared objects from the currently installed
4264 @file{libstdc++.a} archive:
4266 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
4269 Enable the @samp{F_LOADONLY} flag so that the shared object will be
4270 available for runtime dynamic loading, but not linking:
4272 % strip -e libstdc++.so.4 libstdc++.so.5
4275 Archive the runtime-only shared object in the GCC 3.4
4276 @file{libstdc++.a} archive:
4278 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
4282 @uref{./configure.html#WithAixSoname,,@option{--with-aix-soname=svr4}}
4283 configure option may drop the need for this procedure for libraries that
4286 Linking executables and shared libraries may produce warnings of
4287 duplicate symbols. The assembly files generated by GCC for AIX always
4288 have included multiple symbol definitions for certain global variable
4289 and function declarations in the original program. The warnings should
4290 not prevent the linker from producing a correct library or runnable
4293 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
4294 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
4295 to parse archive libraries did not handle the new format correctly.
4296 These routines are used by GCC and result in error messages during
4297 linking such as ``not a COFF file''. The version of the routines shipped
4298 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
4299 option of the archive command may be used to create archives of 32-bit
4300 objects using the original ``small format''. A correct version of the
4301 routines is shipped with AIX 4.3.2 and above.
4303 Some versions of the AIX binder (linker) can fail with a relocation
4304 overflow severe error when the @option{-bbigtoc} option is used to link
4305 GCC-produced object files into an executable that overflows the TOC@. A fix
4306 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
4307 available from IBM Customer Support and from its
4308 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4309 website as PTF U455193.
4311 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
4312 with a segmentation fault when invoked by any version of GCC@. A fix for
4313 APAR IX87327 is available from IBM Customer Support and from its
4314 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4315 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
4317 The initial assembler shipped with AIX 4.3.0 generates incorrect object
4318 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
4319 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
4320 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4321 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
4323 AIX provides National Language Support (NLS)@. Compilers and assemblers
4324 use NLS to support locale-specific representations of various data
4325 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
4326 separating decimal fractions). There have been problems reported where
4327 GCC does not produce the same floating-point formats that the assembler
4328 expects. If one encounters this problem, set the @env{LANG}
4329 environment variable to @samp{C} or @samp{En_US}.
4331 A default can be specified with the @option{-mcpu=@var{cpu_type}}
4332 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
4337 @anchor{iq2000-x-elf}
4338 @heading iq2000-*-elf
4339 Vitesse IQ2000 processors. These are used in embedded
4340 applications. There are no standard Unix configurations.
4347 Lattice Mico32 processor.
4348 This configuration is intended for embedded systems.
4353 @anchor{lm32-x-uclinux}
4354 @heading lm32-*-uclinux
4355 Lattice Mico32 processor.
4356 This configuration is intended for embedded systems running uClinux.
4363 Renesas M32C processor.
4364 This configuration is intended for embedded systems.
4371 Renesas M32R processor.
4372 This configuration is intended for embedded systems.
4380 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
4382 build libraries for both M680x0 and ColdFire processors. If you only
4383 need the M680x0 libraries, you can omit the ColdFire ones by passing
4384 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
4385 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
4386 @command{configure}. These targets default to 5206 or 5475 code as
4387 appropriate for the target system when
4388 configured with @option{--with-arch=cf} and 68020 code otherwise.
4390 The @samp{m68k-*-netbsd} and
4391 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
4392 option. They will generate ColdFire CFV4e code when configured with
4393 @option{--with-arch=cf} and 68020 code otherwise.
4395 You can override the default processors listed above by configuring
4396 with @option{--with-cpu=@var{target}}. This @var{target} can either
4397 be a @option{-mcpu} argument or one of the following values:
4398 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
4399 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
4401 GCC requires at least binutils version 2.17 on these targets.
4406 @anchor{m68k-x-uclinux}
4407 @heading m68k-*-uclinux
4408 GCC 4.3 changed the uClinux configuration so that it uses the
4409 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
4410 It also added improved support for C++ and flat shared libraries,
4411 both of which were ABI changes.
4416 @anchor{microblaze-x-elf}
4417 @heading microblaze-*-elf
4418 Xilinx MicroBlaze processor.
4419 This configuration is intended for embedded systems.
4426 If on a MIPS system you get an error message saying ``does not have gp
4427 sections for all it's [sic] sectons [sic]'', don't worry about it. This
4428 happens whenever you use GAS with the MIPS linker, but there is not
4429 really anything wrong, and it is okay to use the output file. You can
4430 stop such warnings by installing the GNU linker.
4432 It would be nice to extend GAS to produce the gp tables, but they are
4433 optional, and there should not be a warning about their absence.
4435 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
4436 and later. A patch went in just after the GCC 3.3 release to
4437 make @samp{mips*-*-*} use the generic implementation instead. You can also
4438 configure for @samp{mipsel-elf} as a workaround. The
4439 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
4440 work on this is expected in future releases.
4442 @c If you make --with-llsc the default for another target, please also
4443 @c update the description of the --with-llsc option.
4445 The built-in @code{__sync_*} functions are available on MIPS II and
4446 later systems and others that support the @samp{ll}, @samp{sc} and
4447 @samp{sync} instructions. This can be overridden by passing
4448 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
4449 Since the Linux kernel emulates these instructions if they are
4450 missing, the default for @samp{mips*-*-linux*} targets is
4451 @option{--with-llsc}. The @option{--with-llsc} and
4452 @option{--without-llsc} configure options may be overridden at compile
4453 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
4456 MIPS systems check for division by zero (unless
4457 @option{-mno-check-zero-division} is passed to the compiler) by
4458 generating either a conditional trap or a break instruction. Using
4459 trap results in smaller code, but is only supported on MIPS II and
4460 later. Also, some versions of the Linux kernel have a bug that
4461 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
4462 the use of break, use the @option{--with-divide=breaks}
4463 @command{configure} option when configuring GCC@. The default is to
4464 use traps on systems that support them.
4469 @anchor{moxie-x-elf}
4470 @heading moxie-*-elf
4471 The moxie processor.
4476 @anchor{msp430-x-elf}
4477 @heading msp430-*-elf*
4478 TI MSP430 processor.
4479 This configuration is intended for embedded systems.
4481 @samp{msp430-*-elf} is the standard configuration with most GCC
4482 features enabled by default.
4484 @samp{msp430-*-elfbare} is tuned for a bare-metal environment, and disables
4485 features related to shared libraries and other functionality not used for
4486 this device. This reduces code and data usage of the GCC libraries, resulting
4487 in a minimal run-time environment by default.
4489 Features disabled by default include:
4491 @item transactional memory
4498 @anchor{nds32le-x-elf}
4499 @heading nds32le-*-elf
4500 Andes NDS32 target in little endian mode.
4505 @anchor{nds32be-x-elf}
4506 @heading nds32be-*-elf
4507 Andes NDS32 target in big endian mode.
4512 @anchor{nvptx-x-none}
4513 @heading nvptx-*-none
4516 Instead of GNU binutils, you will need to install
4517 @uref{https://github.com/MentorEmbedded/nvptx-tools/,,nvptx-tools}.
4518 Tell GCC where to find it:
4519 @option{--with-build-time-tools=[install-nvptx-tools]/nvptx-none/bin}.
4521 You will need newlib 3.0 git revision
4522 cd31fbb2aea25f94d7ecedc9db16dfc87ab0c316 or later. It can be
4523 automatically built together with GCC@. For this, add a symbolic link
4524 to nvptx-newlib's @file{newlib} directory to the directory containing
4527 Use the @option{--disable-sjlj-exceptions} and
4528 @option{--enable-newlib-io-long-long} options when configuring.
4535 The OpenRISC 1000 32-bit processor with delay slots.
4536 This configuration is intended for embedded systems.
4541 @anchor{or1k-x-linux}
4542 @heading or1k-*-linux
4543 The OpenRISC 1000 32-bit processor with delay slots.
4548 @anchor{powerpc-x-x}
4549 @heading powerpc-*-*
4550 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
4551 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
4553 You will need GNU binutils 2.20 or newer.
4558 @anchor{powerpc-x-darwin}
4559 @heading powerpc-*-darwin*
4560 PowerPC running Darwin (Mac OS X kernel).
4562 Pre-installed versions of Mac OS X may not include any developer tools,
4563 meaning that you will not be able to build GCC from source. Tool
4564 binaries are available at
4565 @uref{https://opensource.apple.com}.
4567 This version of GCC requires at least cctools-590.36. The
4568 cctools-590.36 package referenced from
4569 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
4570 on systems older than 10.3.9 (aka darwin7.9.0).
4575 @anchor{powerpc-x-elf}
4576 @heading powerpc-*-elf
4577 PowerPC system in big endian mode, running System V.4.
4582 @anchor{powerpc-x-linux-gnu}
4583 @heading powerpc*-*-linux-gnu*
4584 PowerPC system in big endian mode running Linux.
4589 @anchor{powerpc-x-netbsd}
4590 @heading powerpc-*-netbsd*
4591 PowerPC system in big endian mode running NetBSD@.
4596 @anchor{powerpc-x-eabisim}
4597 @heading powerpc-*-eabisim
4598 Embedded PowerPC system in big endian mode for use in running under the
4604 @anchor{powerpc-x-eabi}
4605 @heading powerpc-*-eabi
4606 Embedded PowerPC system in big endian mode.
4611 @anchor{powerpcle-x-elf}
4612 @heading powerpcle-*-elf
4613 PowerPC system in little endian mode, running System V.4.
4618 @anchor{powerpcle-x-eabisim}
4619 @heading powerpcle-*-eabisim
4620 Embedded PowerPC system in little endian mode for use in running under
4626 @anchor{powerpcle-x-eabi}
4627 @heading powerpcle-*-eabi
4628 Embedded PowerPC system in little endian mode.
4635 The Renesas RL78 processor.
4636 This configuration is intended for embedded systems.
4641 @anchor{riscv32-x-elf}
4642 @heading riscv32-*-elf
4643 The RISC-V RV32 instruction set.
4644 This configuration is intended for embedded systems.
4645 This (and all other RISC-V) targets require the binutils 2.30 release.
4650 @anchor{riscv32-x-linux}
4651 @heading riscv32-*-linux
4652 The RISC-V RV32 instruction set running GNU/Linux.
4653 This (and all other RISC-V) targets require the binutils 2.30 release.
4658 @anchor{riscv64-x-elf}
4659 @heading riscv64-*-elf
4660 The RISC-V RV64 instruction set.
4661 This configuration is intended for embedded systems.
4662 This (and all other RISC-V) targets require the binutils 2.30 release.
4667 @anchor{riscv64-x-linux}
4668 @heading riscv64-*-linux
4669 The RISC-V RV64 instruction set running GNU/Linux.
4670 This (and all other RISC-V) targets require the binutils 2.30 release.
4677 The Renesas RX processor.
4682 @anchor{s390-x-linux}
4683 @heading s390-*-linux*
4684 S/390 system running GNU/Linux for S/390@.
4689 @anchor{s390x-x-linux}
4690 @heading s390x-*-linux*
4691 zSeries system (64-bit) running GNU/Linux for zSeries@.
4696 @anchor{s390x-ibm-tpf}
4697 @heading s390x-ibm-tpf*
4698 zSeries system (64-bit) running TPF@. This platform is
4699 supported as cross-compilation target only.
4704 @c Please use Solaris 2 to refer to all release of Solaris, starting
4705 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
4706 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
4707 @c alone is too unspecific and must be avoided.
4708 @anchor{x-x-solaris2}
4709 @heading *-*-solaris2*
4710 Support for Solaris 10 has been removed in GCC 10. Support for Solaris
4711 9 has been removed in GCC 5. Support for Solaris 8 has been removed in
4712 GCC 4.8. Support for Solaris 7 has been removed in GCC 4.6.
4714 Solaris 11.3 provides GCC 4.5.2, 4.7.3, and 4.8.2 as
4715 @command{/usr/gcc/4.5/bin/gcc} or similar. Newer Solaris versions
4716 provide one or more of GCC 5, 7, and 9. Alternatively,
4717 you can install a pre-built GCC to bootstrap and install GCC. See the
4718 @uref{binaries.html,,binaries page} for details.
4720 The Solaris 2 @command{/bin/sh} will often fail to configure
4721 @samp{libstdc++-v3}. We therefore recommend using the
4722 following initial sequence of commands
4725 % CONFIG_SHELL=/bin/ksh
4726 % export CONFIG_SHELL
4730 and proceed as described in @uref{configure.html,,the configure instructions}.
4731 In addition we strongly recommend specifying an absolute path to invoke
4732 @command{@var{srcdir}/configure}.
4734 In Solaris 11, you need to check for @code{system/header},
4735 @code{system/linker}, and @code{developer/assembler} packages.
4737 Trying to use the linker and other tools in
4738 @file{/usr/ucb} to install GCC has been observed to cause trouble.
4739 For example, the linker may hang indefinitely. The fix is to remove
4740 @file{/usr/ucb} from your @env{PATH}.
4742 The build process works more smoothly with the legacy Solaris tools so, if you
4743 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
4744 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
4746 We recommend the use of the Solaris assembler or the GNU assembler, in
4747 conjunction with the Solaris linker. The GNU @command{as}
4748 versions included in Solaris 11.3,
4749 from GNU binutils 2.23.1 or newer (in @file{/usr/bin/gas} and
4750 @file{/usr/gnu/bin/as}), are known to work.
4751 The current version, from GNU binutils 2.34,
4752 is known to work as well. Note that your mileage may vary
4753 if you use a combination of the GNU tools and the Solaris tools: while the
4754 combination GNU @command{as} + Solaris @command{ld} should reasonably work,
4755 the reverse combination Solaris @command{as} + GNU @command{ld} may fail to
4756 build or cause memory corruption at runtime in some cases for C++ programs.
4758 GNU @command{ld} usually works as well. Again, the current
4759 version (2.34) is known to work, but generally lacks platform specific
4760 features, so better stay with Solaris @command{ld}. To use the LTO linker
4761 plugin (@option{-fuse-linker-plugin}) with GNU @command{ld}, GNU
4762 binutils @emph{must} be configured with @option{--enable-largefile}.
4764 To enable symbol versioning in @samp{libstdc++} with the Solaris linker,
4765 you need to have any version of GNU @command{c++filt}, which is part of
4766 GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no
4767 appropriate version is found. Solaris @command{c++filt} from the Solaris
4768 Studio compilers does @emph{not} work.
4770 The versions of the GNU Multiple Precision Library (GMP), the MPFR
4771 library and the MPC library bundled with Solaris 11.3 and later are
4772 usually recent enough to match GCC's requirements. There are two
4777 While the version of the GMP library in Solaris 11.3 works with GCC, you
4778 need to configure with @option{--with-gmp-include=/usr/include/gmp}.
4781 The version of the MPFR libary included in Solaris 11.3 is too old; you
4782 need to provide a more recent one.
4790 This section contains general configuration information for all
4791 SPARC-based platforms. In addition to reading this section, please
4792 read all other sections that match your target.
4794 Newer versions of the GNU Multiple Precision Library (GMP), the MPFR
4795 library and the MPC library are known to be miscompiled by earlier
4796 versions of GCC on these platforms. We therefore recommend the use
4797 of the exact versions of these libraries listed as minimal versions
4798 in @uref{prerequisites.html,,the prerequisites}.
4803 @anchor{sparc-sun-solaris2}
4804 @heading sparc-sun-solaris2*
4805 When GCC is configured to use GNU binutils 2.14 or later, the binaries
4806 produced are smaller than the ones produced using Solaris native tools;
4807 this difference is quite significant for binaries containing debugging
4810 Starting with Solaris 7, the operating system is capable of executing
4811 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
4812 this; the @option{-m64} option enables 64-bit code generation.
4813 However, if all you want is code tuned for the UltraSPARC CPU, you
4814 should try the @option{-mtune=ultrasparc} option instead, which produces
4815 code that, unlike full 64-bit code, can still run on non-UltraSPARC
4818 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4819 library or the MPC library on a Solaris 7 or later system, the canonical
4820 target triplet must be specified as the @command{build} parameter on the
4821 configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and
4822 not that of GMP or MPFR or MPC). For example on a Solaris 11 system:
4825 % ./configure --build=sparc-sun-solaris2.11 --prefix=xxx
4831 @anchor{sparc-x-linux}
4832 @heading sparc-*-linux*
4837 @anchor{sparc64-x-solaris2}
4838 @heading sparc64-*-solaris2*
4839 When configuring a 64-bit-default GCC on Solaris/SPARC, you must use a
4840 build compiler that generates 64-bit code, either by default or by
4841 specifying @samp{CC='gcc -m64' CXX='gcc-m64'} to @command{configure}.
4842 Additionally, you @emph{must} pass @option{--build=sparc64-sun-solaris2.11}
4843 or @option{--build=sparcv9-sun-solaris2.11} because @file{config.guess}
4844 misdetects this situation, which can cause build failures.
4846 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4847 library or the MPC library, the canonical target triplet must be specified
4848 as the @command{build} parameter on the configure line. For example
4849 on a Solaris 11 system:
4852 % ./configure --build=sparc64-sun-solaris2.11 --prefix=xxx
4858 @anchor{sparcv9-x-solaris2}
4859 @heading sparcv9-*-solaris2*
4860 This is a synonym for @samp{sparc64-*-solaris2*}.
4867 The C6X family of processors. This port requires binutils-2.22 or newer.
4872 @anchor{tilegx-*-linux}
4873 @heading tilegx-*-linux*
4874 The TILE-Gx processor in little endian mode, running GNU/Linux. This
4875 port requires binutils-2.22 or newer.
4880 @anchor{tilegxbe-*-linux}
4881 @heading tilegxbe-*-linux*
4882 The TILE-Gx processor in big endian mode, running GNU/Linux. This
4883 port requires binutils-2.23 or newer.
4888 @anchor{tilepro-*-linux}
4889 @heading tilepro-*-linux*
4890 The TILEPro processor running GNU/Linux. This port requires
4891 binutils-2.22 or newer.
4896 @anchor{visium-x-elf}
4897 @heading visium-*-elf
4898 CDS VISIUMcore processor.
4899 This configuration is intended for embedded systems.
4904 @anchor{x-x-vxworks}
4905 @heading *-*-vxworks*
4906 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4907 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4908 We welcome patches for other architectures supported by VxWorks 5.5.
4909 Support for VxWorks AE would also be welcome; we believe this is merely
4910 a matter of writing an appropriate ``configlette'' (see below). We are
4911 not interested in supporting older, a.out or COFF-based, versions of
4914 VxWorks comes with an older version of GCC installed in
4915 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4916 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4917 Before running @command{configure}, create the directories @file{@var{prefix}}
4918 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4919 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4920 include that directory while running both @command{configure} and
4923 You must give @command{configure} the
4924 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4925 find the VxWorks system headers. Since VxWorks is a cross compilation
4926 target only, you must also specify @option{--target=@var{target}}.
4927 @command{configure} will attempt to create the directory
4928 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4929 make sure the user running @command{configure} has sufficient privilege
4932 GCC's exception handling runtime requires a special ``configlette''
4933 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4934 that file to add the module to your kernel build. (Future versions of
4935 VxWorks will incorporate this module.)
4941 @heading x86_64-*-*, amd64-*-*
4942 GCC supports the x86-64 architecture implemented by the AMD64 processor
4943 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4944 On GNU/Linux the default is a bi-arch compiler which is able to generate
4945 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4950 @anchor{x86-64-x-solaris2}
4951 @heading x86_64-*-solaris2*
4952 GCC also supports the x86-64 architecture implemented by the AMD64
4953 processor (@samp{amd64-*-*} is an alias for @samp{x86_64-*-*}) on
4954 Solaris 10 or later. Unlike other systems, without special options a
4955 bi-arch compiler is built which generates 32-bit code by default, but
4956 can generate 64-bit x86-64 code with the @option{-m64} switch. Since
4957 GCC 4.7, there is also a configuration that defaults to 64-bit code, but
4958 can generate 32-bit code with @option{-m32}. To configure and build
4959 this way, you have to provide all support libraries like @file{libgmp}
4960 as 64-bit code, configure with @option{--target=x86_64-pc-solaris2.11}
4961 and @samp{CC=gcc -m64}.
4966 @anchor{xtensa-x-elf}
4967 @heading xtensa*-*-elf
4968 This target is intended for embedded Xtensa systems using the
4969 @samp{newlib} C library. It uses ELF but does not support shared
4970 objects. Designed-defined instructions specified via the
4971 Tensilica Instruction Extension (TIE) language are only supported
4972 through inline assembly.
4974 The Xtensa configuration information must be specified prior to
4975 building GCC@. The @file{include/xtensa-config.h} header
4976 file contains the configuration information. If you created your
4977 own Xtensa configuration with the Xtensa Processor Generator, the
4978 downloaded files include a customized copy of this header file,
4979 which you can use to replace the default header file.
4984 @anchor{xtensa-x-linux}
4985 @heading xtensa*-*-linux*
4986 This target is for Xtensa systems running GNU/Linux. It supports ELF
4987 shared objects and the GNU C library (glibc). It also generates
4988 position-independent code (PIC) regardless of whether the
4989 @option{-fpic} or @option{-fPIC} options are used. In other
4990 respects, this target is the same as the
4991 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
4997 @heading Microsoft Windows
4999 @subheading Intel 16-bit versions
5000 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
5003 However, the 32-bit port has limited support for Microsoft
5004 Windows 3.11 in the Win32s environment, as a target only. See below.
5006 @subheading Intel 32-bit versions
5007 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
5008 XP, and Windows Vista, are supported by several different target
5009 platforms. These targets differ in which Windows subsystem they target
5010 and which C libraries are used.
5013 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
5014 Linux API emulation layer in the Win32 subsystem.
5015 @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for
5016 the Win32 subsystem that provides a subset of POSIX.
5017 @item MKS i386-pc-mks: NuTCracker from MKS. See
5018 @uref{https://www.mkssoftware.com} for more information.
5021 @subheading Intel 64-bit versions
5022 GCC contains support for x86-64 using the mingw-w64
5023 runtime library, available from @uref{http://mingw-w64.org/doku.php}.
5024 This library should be used with the target triple x86_64-pc-mingw32.
5026 Presently Windows for Itanium is not supported.
5028 @subheading Windows CE
5029 Windows CE is supported as a target only on Hitachi
5030 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
5032 @subheading Other Windows Platforms
5033 GCC no longer supports Windows NT on the Alpha or PowerPC.
5035 GCC no longer supports the Windows POSIX subsystem. However, it does
5036 support the Interix subsystem. See above.
5038 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
5040 PW32 (i386-pc-pw32) support was never completed, and the project seems to
5041 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
5043 UWIN support has been removed due to a lack of maintenance.
5050 Ports of GCC are included with the
5051 @uref{http://www.cygwin.com/,,Cygwin environment}.
5053 GCC will build under Cygwin without modification; it does not build
5054 with Microsoft's C++ compiler and there are no plans to make it do so.
5056 The Cygwin native compiler can be configured to target any 32-bit x86
5057 cpu architecture desired; the default is i686-pc-cygwin. It should be
5058 used with as up-to-date a version of binutils as possible; use either
5059 the latest official GNU binutils release in the Cygwin distribution,
5060 or version 2.20 or above if building your own.
5065 @anchor{x-x-mingw32}
5066 @heading *-*-mingw32
5067 GCC will build with and support only MinGW runtime 3.12 and later.
5068 Earlier versions of headers are incompatible with the new default semantics
5069 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
5075 @heading Older systems
5076 GCC contains support files for many older (1980s and early
5077 1990s) Unix variants. For the most part, support for these systems
5078 has not been deliberately removed, but it has not been maintained for
5079 several years and may suffer from bitrot.
5081 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
5082 Support for these systems is still present in that release, but
5083 @command{configure} will fail unless the @option{--enable-obsolete}
5084 option is given. Unless a maintainer steps forward, support for these
5085 systems will be removed from the next release of GCC@.
5087 Support for old systems as hosts for GCC can cause problems if the
5088 workarounds for compiler, library and operating system bugs affect the
5089 cleanliness or maintainability of the rest of GCC@. In some cases, to
5090 bring GCC up on such a system, if still possible with current GCC, may
5091 require first installing an old version of GCC which did work on that
5092 system, and using it to compile a more recent GCC, to avoid bugs in the
5093 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
5094 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
5095 sites}. Header bugs may generally be avoided using
5096 @command{fixincludes}, but bugs or deficiencies in libraries and the
5097 operating system may still cause problems.
5099 Support for older systems as targets for cross-compilation is less
5100 problematic than support for them as hosts for GCC; if an enthusiast
5101 wishes to make such a target work again (including resurrecting any of
5102 the targets that never worked with GCC 2, starting from the last
5103 version before they were removed), patches
5104 @uref{../contribute.html,,following the usual requirements} would be
5105 likely to be accepted, since they should not affect the support for more
5108 For some systems, old versions of GNU binutils may also be useful,
5109 and are available from @file{pub/binutils/old-releases} on
5110 @uref{https://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
5112 Some of the information on specific systems above relates to
5113 such older systems, but much of the information
5114 about GCC on such systems (which may no longer be applicable to
5115 current GCC) is to be found in the GCC texinfo manual.
5121 @heading all ELF targets (SVR4, Solaris 2, etc.)
5122 C++ support is significantly better on ELF targets if you use the
5123 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
5124 inlines, vtables and template instantiations will be discarded
5133 @uref{./index.html,,Return to the GCC Installation page}
5137 @c ***Old documentation******************************************************
5139 @include install-old.texi
5145 @uref{./index.html,,Return to the GCC Installation page}
5149 @c ***GFDL********************************************************************
5157 @uref{./index.html,,Return to the GCC Installation page}
5161 @c ***************************************************************************
5162 @c Part 6 The End of the Document
5164 @comment node-name, next, previous, up
5165 @node Concept Index, , GNU Free Documentation License, Top
5169 @unnumbered Concept Index