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) 1988-2019 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{} 1988-2019 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++98 compiler
242 Necessary to bootstrap GCC, although versions of GCC prior
243 to 4.8 also allow bootstrapping with a ISO C89 compiler and versions
244 of GCC prior to 3.4 also allow bootstrapping with a traditional
247 To build all languages in a cross-compiler or other configuration where
248 3-stage bootstrap is not performed, you need to start with an existing
249 GCC binary (version 3.4 or later) because source code for language
250 frontends other than C might use GCC extensions.
252 Note that to bootstrap GCC with versions of GCC earlier than 3.4, you
253 may need to use @option{--disable-stage1-checking}, though
254 bootstrapping the compiler with such earlier compilers is strongly
257 @item C standard library and headers
259 In order to build GCC, the C standard library and headers must be present
260 for all target variants for which target libraries will be built (and not
261 only the variant of the host C++ compiler).
263 This affects the popular @samp{x86_64-pc-linux-gnu} platform (among
264 other multilib targets), for which 64-bit (@samp{x86_64}) and 32-bit
265 (@samp{i386}) libc headers are usually packaged separately. If you do a
266 build of a native compiler on @samp{x86_64-pc-linux-gnu}, make sure you
267 either have the 32-bit libc developer package properly installed (the exact
268 name of the package depends on your distro) or you must build GCC as a
269 64-bit only compiler by configuring with the option
270 @option{--disable-multilib}. Otherwise, you may encounter an error such as
271 @samp{fatal error: gnu/stubs-32.h: No such file}
275 In order to build the Ada compiler (GNAT) you must already have GNAT
276 installed because portions of the Ada frontend are written in Ada (with
277 GNAT extensions.) Refer to the Ada installation instructions for more
278 specific information.
280 @item A ``working'' POSIX compatible shell, or GNU bash
282 Necessary when running @command{configure} because some
283 @command{/bin/sh} shells have bugs and may crash when configuring the
284 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
285 have disastrous corner-case performance problems. This
286 can cause target @command{configure} runs to literally take days to
287 complete in some cases.
289 So on some platforms @command{/bin/ksh} is sufficient, on others it
290 isn't. See the host/target specific instructions for your platform, or
291 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
292 environment to your ``good'' shell prior to running
293 @command{configure}/@command{make}.
295 @command{zsh} is not a fully compliant POSIX shell and will not
296 work when configuring GCC@.
298 @item A POSIX or SVR4 awk
300 Necessary for creating some of the generated source files for GCC@.
301 If in doubt, use a recent GNU awk version, as some of the older ones
302 are broken. GNU awk version 3.1.5 is known to work.
306 Necessary in some circumstances, optional in others. See the
307 host/target specific instructions for your platform for the exact
310 @item gzip version 1.2.4 (or later) or
311 @itemx bzip2 version 1.0.2 (or later)
313 Necessary to uncompress GCC @command{tar} files when source code is
314 obtained via FTP mirror sites.
316 @item GNU make version 3.80 (or later)
318 You must have GNU make installed to build GCC@.
320 @item GNU tar version 1.14 (or later)
322 Necessary (only on some platforms) to untar the source code. Many
323 systems' @command{tar} programs will also work, only try GNU
324 @command{tar} if you have problems.
326 @item Perl version between 5.6.1 and 5.6.24
328 Necessary when targeting Darwin, building @samp{libstdc++},
329 and not using @option{--disable-symvers}.
330 Necessary when targeting Solaris 2 with Sun @command{ld} and not using
331 @option{--disable-symvers}. The bundled @command{perl} in Solaris@tie{}8
334 Necessary when regenerating @file{Makefile} dependencies in libiberty.
335 Necessary when regenerating @file{libiberty/functions.texi}.
336 Necessary when generating manpages from Texinfo manuals.
337 Used by various scripts to generate some files included in SVN (mainly
338 Unicode-related and rarely changing) from source tables.
340 Used by @command{automake}.
344 Several support libraries are necessary to build GCC, some are required,
345 others optional. While any sufficiently new version of required tools
346 usually work, library requirements are generally stricter. Newer
347 versions may work in some cases, but it's safer to use the exact
348 versions documented. We appreciate bug reports about problems with
349 newer versions, though. If your OS vendor provides packages for the
350 support libraries then using those packages may be the simplest way to
351 install the libraries.
354 @item GNU Multiple Precision Library (GMP) version 4.3.2 (or later)
356 Necessary to build GCC@. If a GMP source distribution is found in a
357 subdirectory of your GCC sources named @file{gmp}, it will be built
358 together with GCC. Alternatively, if GMP is already installed but it
359 is not in your library search path, you will have to configure with the
360 @option{--with-gmp} configure option. See also @option{--with-gmp-lib}
361 and @option{--with-gmp-include}.
362 The in-tree build is only supported with the GMP version that
363 download_prerequisites installs.
365 @item MPFR Library version 2.4.2 (or later)
367 Necessary to build GCC@. It can be downloaded from
368 @uref{https://www.mpfr.org}. If an MPFR source distribution is found
369 in a subdirectory of your GCC sources named @file{mpfr}, it will be
370 built together with GCC. Alternatively, if MPFR is already installed
371 but it is not in your default library search path, the
372 @option{--with-mpfr} configure option should be used. See also
373 @option{--with-mpfr-lib} and @option{--with-mpfr-include}.
374 The in-tree build is only supported with the MPFR version that
375 download_prerequisites installs.
377 @item MPC Library version 0.8.1 (or later)
379 Necessary to build GCC@. It can be downloaded from
380 @uref{http://www.multiprecision.org/mpc/}. If an MPC source distribution
381 is found in a subdirectory of your GCC sources named @file{mpc}, it
382 will be built together with GCC. Alternatively, if MPC is already
383 installed but it is not in your default library search path, the
384 @option{--with-mpc} configure option should be used. See also
385 @option{--with-mpc-lib} and @option{--with-mpc-include}.
386 The in-tree build is only supported with the MPC version that
387 download_prerequisites installs.
389 @item isl Library version 0.15 or later.
391 Necessary to build GCC with the Graphite loop optimizations.
392 It can be downloaded from @uref{ftp://gcc.gnu.org/pub/gcc/infrastructure/}.
393 If an isl source distribution is found
394 in a subdirectory of your GCC sources named @file{isl}, it will be
395 built together with GCC. Alternatively, the @option{--with-isl} configure
396 option should be used if isl is not installed in your default library
401 Necessary to build GCC with zstd compression used for LTO bytecode.
402 The library is searched in your default library patch search.
403 Alternatively, the @option{--with-zstd} configure option should be used.
407 @heading Tools/packages necessary for modifying GCC
409 @item autoconf version 2.69
410 @itemx GNU m4 version 1.4.6 (or later)
412 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
413 to regenerate @file{configure} and @file{config.in} files.
415 @item automake version 1.15.1
417 Necessary when modifying a @file{Makefile.am} file to regenerate its
418 associated @file{Makefile.in}.
420 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
421 file. Specifically this applies to the @file{gcc}, @file{intl},
422 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
423 as any of their subdirectories.
425 For directories that use automake, GCC requires the latest release in
426 the 1.15 series, which is currently 1.15.1. When regenerating a directory
427 to a newer version, please update all the directories using an older 1.15
428 to the latest released version.
430 @item gettext version 0.14.5 (or later)
432 Needed to regenerate @file{gcc.pot}.
434 @item gperf version 2.7.2 (or later)
436 Necessary when modifying @command{gperf} input files, e.g.@:
437 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
438 @file{gcc/cp/cfns.h}.
444 Necessary to run the GCC testsuite; see the section on testing for
445 details. Tcl 8.6 has a known regression in RE pattern handling that
446 make parts of the testsuite fail. See
447 @uref{http://core.tcl.tk/tcl/tktview/267b7e2334ee2e9de34c4b00d6e72e2f1997085f}
448 for more information. This bug has been fixed in 8.6.1.
450 @item autogen version 5.5.4 (or later) and
451 @itemx guile version 1.4.1 (or later)
453 Necessary to regenerate @file{fixinc/fixincl.x} from
454 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
456 Necessary to run @samp{make check} for @file{fixinc}.
458 Necessary to regenerate the top level @file{Makefile.in} file from
459 @file{Makefile.tpl} and @file{Makefile.def}.
461 @item Flex version 2.5.4 (or later)
463 Necessary when modifying @file{*.l} files.
465 Necessary to build GCC during development because the generated output
466 files are not included in the SVN repository. They are included in
469 @item Texinfo version 4.7 (or later)
471 Necessary for running @command{makeinfo} when modifying @file{*.texi}
472 files to test your changes.
474 Necessary for running @command{make dvi} or @command{make pdf} to
475 create printable documentation in DVI or PDF format. Texinfo version
476 4.8 or later is required for @command{make pdf}.
478 Necessary to build GCC documentation during development because the
479 generated output files are not included in the SVN repository. They are
480 included in releases.
482 @item @TeX{} (any working version)
484 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
485 are used when running @command{make dvi} or @command{make pdf} to create
486 DVI or PDF files, respectively.
488 @item Sphinx version 1.0 (or later)
490 Necessary to regenerate @file{jit/docs/_build/texinfo} from the @file{.rst}
491 files in the directories below @file{jit/docs}.
493 @item SVN (any version)
494 @itemx SSH (any version)
496 Necessary to access the SVN repository. Public releases and weekly
497 snapshots of the development sources are also available via FTP@.
499 @item GNU diffutils version 2.7 (or later)
501 Useful when submitting patches for the GCC source code.
503 @item patch version 2.5.4 (or later)
505 Necessary when applying patches, created with @command{diff}, to one's
515 @uref{./index.html,,Return to the GCC Installation page}
519 @c ***Downloading the source**************************************************
521 @comment node-name, next, previous, up
522 @node Downloading the source, Configuration, Prerequisites, Installing GCC
526 @chapter Downloading GCC
528 @cindex Downloading GCC
529 @cindex Downloading the Source
531 GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP
532 tarballs compressed with @command{gzip} or
535 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
536 for information on how to obtain GCC@.
538 The source distribution includes the C, C++, Objective-C, Fortran,
539 and Ada (in the case of GCC 3.1 and later) compilers, as well as
540 runtime libraries for C++, Objective-C, and Fortran.
541 For previous versions these were downloadable as separate components such
542 as the core GCC distribution, which included the C language front end and
543 shared components, and language-specific distributions including the
544 language front end and the language runtime (where appropriate).
546 If you also intend to build binutils (either to upgrade an existing
547 installation or for use in place of the corresponding tools of your
548 OS), unpack the binutils distribution either in the same directory or
549 a separate one. In the latter case, add symbolic links to any
550 components of the binutils you intend to build alongside the compiler
551 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
552 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
554 Likewise the GMP, MPFR and MPC libraries can be automatically built
555 together with GCC. You may simply run the
556 @command{contrib/download_prerequisites} script in the GCC source directory
557 to set up everything.
558 Otherwise unpack the GMP, MPFR and/or MPC source
559 distributions in the directory containing the GCC sources and rename
560 their directories to @file{gmp}, @file{mpfr} and @file{mpc},
561 respectively (or use symbolic links with the same name).
568 @uref{./index.html,,Return to the GCC Installation page}
572 @c ***Configuration***********************************************************
574 @comment node-name, next, previous, up
575 @node Configuration, Building, Downloading the source, Installing GCC
579 @chapter Installing GCC: Configuration
581 @cindex Configuration
582 @cindex Installing GCC: Configuration
584 Like most GNU software, GCC must be configured before it can be built.
585 This document describes the recommended configuration procedure
586 for both native and cross targets.
588 We use @var{srcdir} to refer to the toplevel source directory for
589 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
591 If you obtained the sources via SVN, @var{srcdir} must refer to the top
592 @file{gcc} directory, the one where the @file{MAINTAINERS} file can be
593 found, and not its @file{gcc} subdirectory, otherwise the build will fail.
595 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
596 file system, the shell's built-in @command{pwd} command will return
597 temporary pathnames. Using these can lead to various sorts of build
598 problems. To avoid this issue, set the @env{PWDCMD} environment
599 variable to an automounter-aware @command{pwd} command, e.g.,
600 @command{pawd} or @samp{amq -w}, during the configuration and build
603 First, we @strong{highly} recommend that GCC be built into a
604 separate directory from the sources which does @strong{not} reside
605 within the source tree. This is how we generally build GCC; building
606 where @var{srcdir} == @var{objdir} should still work, but doesn't
607 get extensive testing; building where @var{objdir} is a subdirectory
608 of @var{srcdir} is unsupported.
610 If you have previously built GCC in the same directory for a
611 different target machine, do @samp{make distclean} to delete all files
612 that might be invalid. One of the files this deletes is @file{Makefile};
613 if @samp{make distclean} complains that @file{Makefile} does not exist
614 or issues a message like ``don't know how to make distclean'' it probably
615 means that the directory is already suitably clean. However, with the
616 recommended method of building in a separate @var{objdir}, you should
617 simply use a different @var{objdir} for each target.
619 Second, when configuring a native system, either @command{cc} or
620 @command{gcc} must be in your path or you must set @env{CC} in
621 your environment before running configure. Otherwise the configuration
625 Note that the bootstrap compiler and the resulting GCC must be link
626 compatible, else the bootstrap will fail with linker errors about
627 incompatible object file formats. Several multilibed targets are
628 affected by this requirement, see
630 @ref{Specific, host/target specific installation notes}.
633 @uref{specific.html,,host/target specific installation notes}.
642 % @var{srcdir}/configure [@var{options}] [@var{target}]
645 @heading Distributor options
647 If you will be distributing binary versions of GCC, with modifications
648 to the source code, you should use the options described in this
649 section to make clear that your version contains modifications.
652 @item --with-pkgversion=@var{version}
653 Specify a string that identifies your package. You may wish
654 to include a build number or build date. This version string will be
655 included in the output of @command{gcc --version}. This suffix does
656 not replace the default version string, only the @samp{GCC} part.
658 The default value is @samp{GCC}.
660 @item --with-bugurl=@var{url}
661 Specify the URL that users should visit if they wish to report a bug.
662 You are of course welcome to forward bugs reported to you to the FSF,
663 if you determine that they are not bugs in your modifications.
665 The default value refers to the FSF's GCC bug tracker.
669 @heading Target specification
672 GCC has code to correctly determine the correct value for @var{target}
673 for nearly all native systems. Therefore, we highly recommend you do
674 not provide a configure target when configuring a native compiler.
677 @var{target} must be specified as @option{--target=@var{target}}
678 when configuring a cross compiler; examples of valid targets would be
679 m68k-elf, sh-elf, etc.
682 Specifying just @var{target} instead of @option{--target=@var{target}}
683 implies that the host defaults to @var{target}.
687 @heading Options specification
689 Use @var{options} to override several configure time options for
690 GCC@. A list of supported @var{options} follows; @samp{configure
691 --help} may list other options, but those not listed below may not
692 work and should not normally be used.
694 Note that each @option{--enable} option has a corresponding
695 @option{--disable} option and that each @option{--with} option has a
696 corresponding @option{--without} option.
699 @item --prefix=@var{dirname}
700 Specify the toplevel installation
701 directory. This is the recommended way to install the tools into a directory
702 other than the default. The toplevel installation directory defaults to
705 We @strong{highly} recommend against @var{dirname} being the same or a
706 subdirectory of @var{objdir} or vice versa. If specifying a directory
707 beneath a user's home directory tree, some shells will not expand
708 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
711 The following standard @command{autoconf} options are supported. Normally you
712 should not need to use these options.
714 @item --exec-prefix=@var{dirname}
715 Specify the toplevel installation directory for architecture-dependent
716 files. The default is @file{@var{prefix}}.
718 @item --bindir=@var{dirname}
719 Specify the installation directory for the executables called by users
720 (such as @command{gcc} and @command{g++}). The default is
721 @file{@var{exec-prefix}/bin}.
723 @item --libdir=@var{dirname}
724 Specify the installation directory for object code libraries and
725 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
727 @item --libexecdir=@var{dirname}
728 Specify the installation directory for internal executables of GCC@.
729 The default is @file{@var{exec-prefix}/libexec}.
731 @item --with-slibdir=@var{dirname}
732 Specify the installation directory for the shared libgcc library. The
733 default is @file{@var{libdir}}.
735 @item --datarootdir=@var{dirname}
736 Specify the root of the directory tree for read-only architecture-independent
737 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
739 @item --infodir=@var{dirname}
740 Specify the installation directory for documentation in info format.
741 The default is @file{@var{datarootdir}/info}.
743 @item --datadir=@var{dirname}
744 Specify the installation directory for some architecture-independent
745 data files referenced by GCC@. The default is @file{@var{datarootdir}}.
747 @item --docdir=@var{dirname}
748 Specify the installation directory for documentation files (other
749 than Info) for GCC@. The default is @file{@var{datarootdir}/doc}.
751 @item --htmldir=@var{dirname}
752 Specify the installation directory for HTML documentation files.
753 The default is @file{@var{docdir}}.
755 @item --pdfdir=@var{dirname}
756 Specify the installation directory for PDF documentation files.
757 The default is @file{@var{docdir}}.
759 @item --mandir=@var{dirname}
760 Specify the installation directory for manual pages. The default is
761 @file{@var{datarootdir}/man}. (Note that the manual pages are only extracts
762 from the full GCC manuals, which are provided in Texinfo format. The manpages
763 are derived by an automatic conversion process from parts of the full
766 @item --with-gxx-include-dir=@var{dirname}
768 the installation directory for G++ header files. The default depends
769 on other configuration options, and differs between cross and native
772 @item --with-specs=@var{specs}
773 Specify additional command line driver SPECS.
774 This can be useful if you need to turn on a non-standard feature by
775 default without modifying the compiler's source code, for instance
776 @option{--with-specs=%@{!fcommon:%@{!fno-common:-fno-common@}@}}.
778 @xref{Spec Files,, Specifying subprocesses and the switches to pass to them,
779 gcc, Using the GNU Compiler Collection (GCC)},
782 See ``Spec Files'' in the main manual
787 @item --program-prefix=@var{prefix}
788 GCC supports some transformations of the names of its programs when
789 installing them. This option prepends @var{prefix} to the names of
790 programs to install in @var{bindir} (see above). For example, specifying
791 @option{--program-prefix=foo-} would result in @samp{gcc}
792 being installed as @file{/usr/local/bin/foo-gcc}.
794 @item --program-suffix=@var{suffix}
795 Appends @var{suffix} to the names of programs to install in @var{bindir}
796 (see above). For example, specifying @option{--program-suffix=-3.1}
797 would result in @samp{gcc} being installed as
798 @file{/usr/local/bin/gcc-3.1}.
800 @item --program-transform-name=@var{pattern}
801 Applies the @samp{sed} script @var{pattern} to be applied to the names
802 of programs to install in @var{bindir} (see above). @var{pattern} has to
803 consist of one or more basic @samp{sed} editing commands, separated by
804 semicolons. For example, if you want the @samp{gcc} program name to be
805 transformed to the installed program @file{/usr/local/bin/myowngcc} and
806 the @samp{g++} program name to be transformed to
807 @file{/usr/local/bin/gspecial++} without changing other program names,
808 you could use the pattern
809 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
810 to achieve this effect.
812 All three options can be combined and used together, resulting in more
813 complex conversion patterns. As a basic rule, @var{prefix} (and
814 @var{suffix}) are prepended (appended) before further transformations
815 can happen with a special transformation script @var{pattern}.
817 As currently implemented, this option only takes effect for native
818 builds; cross compiler binaries' names are not transformed even when a
819 transformation is explicitly asked for by one of these options.
821 For native builds, some of the installed programs are also installed
822 with the target alias in front of their name, as in
823 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
824 before the target alias is prepended to the name---so, specifying
825 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
826 resulting binary would be installed as
827 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
829 As a last shortcoming, none of the installed Ada programs are
830 transformed yet, which will be fixed in some time.
832 @item --with-local-prefix=@var{dirname}
834 installation directory for local include files. The default is
835 @file{/usr/local}. Specify this option if you want the compiler to
836 search directory @file{@var{dirname}/include} for locally installed
837 header files @emph{instead} of @file{/usr/local/include}.
839 You should specify @option{--with-local-prefix} @strong{only} if your
840 site has a different convention (not @file{/usr/local}) for where to put
843 The default value for @option{--with-local-prefix} is @file{/usr/local}
844 regardless of the value of @option{--prefix}. Specifying
845 @option{--prefix} has no effect on which directory GCC searches for
846 local header files. This may seem counterintuitive, but actually it is
849 The purpose of @option{--prefix} is to specify where to @emph{install
850 GCC}. The local header files in @file{/usr/local/include}---if you put
851 any in that directory---are not part of GCC@. They are part of other
852 programs---perhaps many others. (GCC installs its own header files in
853 another directory which is based on the @option{--prefix} value.)
855 Both the local-prefix include directory and the GCC-prefix include
856 directory are part of GCC's ``system include'' directories. Although these
857 two directories are not fixed, they need to be searched in the proper
858 order for the correct processing of the include_next directive. The
859 local-prefix include directory is searched before the GCC-prefix
860 include directory. Another characteristic of system include directories
861 is that pedantic warnings are turned off for headers in these directories.
863 Some autoconf macros add @option{-I @var{directory}} options to the
864 compiler command line, to ensure that directories containing installed
865 packages' headers are searched. When @var{directory} is one of GCC's
866 system include directories, GCC will ignore the option so that system
867 directories continue to be processed in the correct order. This
868 may result in a search order different from what was specified but the
869 directory will still be searched.
871 GCC automatically searches for ordinary libraries using
872 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
873 used for both GCC and packages, GCC will automatically search for
874 both headers and libraries. This provides a configuration that is
875 easy to use. GCC behaves in a manner similar to that when it is
876 installed as a system compiler in @file{/usr}.
878 Sites that need to install multiple versions of GCC may not want to
879 use the above simple configuration. It is possible to use the
880 @option{--program-prefix}, @option{--program-suffix} and
881 @option{--program-transform-name} options to install multiple versions
882 into a single directory, but it may be simpler to use different prefixes
883 and the @option{--with-local-prefix} option to specify the location of the
884 site-specific files for each version. It will then be necessary for
885 users to specify explicitly the location of local site libraries
886 (e.g., with @env{LIBRARY_PATH}).
888 The same value can be used for both @option{--with-local-prefix} and
889 @option{--prefix} provided it is not @file{/usr}. This can be used
890 to avoid the default search of @file{/usr/local/include}.
892 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
893 The directory you use for @option{--with-local-prefix} @strong{must not}
894 contain any of the system's standard header files. If it did contain
895 them, certain programs would be miscompiled (including GNU Emacs, on
896 certain targets), because this would override and nullify the header
897 file corrections made by the @command{fixincludes} script.
899 Indications are that people who use this option use it based on mistaken
900 ideas of what it is for. People use it as if it specified where to
901 install part of GCC@. Perhaps they make this assumption because
902 installing GCC creates the directory.
904 @item --with-gcc-major-version-only
905 Specifies that GCC should use only the major number rather than
906 @var{major}.@var{minor}.@var{patchlevel} in filesystem paths.
908 @item --with-native-system-header-dir=@var{dirname}
909 Specifies that @var{dirname} is the directory that contains native system
910 header files, rather than @file{/usr/include}. This option is most useful
911 if you are creating a compiler that should be isolated from the system
912 as much as possible. It is most commonly used with the
913 @option{--with-sysroot} option and will cause GCC to search
914 @var{dirname} inside the system root specified by that option.
916 @item --enable-shared[=@var{package}[,@dots{}]]
917 Build shared versions of libraries, if shared libraries are supported on
918 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
919 are enabled by default on all platforms that support shared libraries.
921 If a list of packages is given as an argument, build shared libraries
922 only for the listed packages. For other packages, only static libraries
923 will be built. Package names currently recognized in the GCC tree are
924 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
925 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
926 @samp{ada}, @samp{libada}, @samp{libgo}, @samp{libobjc}, and @samp{libphobos}.
927 Note @samp{libiberty} does not support shared libraries at all.
929 Use @option{--disable-shared} to build only static libraries. Note that
930 @option{--disable-shared} does not accept a list of package names as
931 argument, only @option{--enable-shared} does.
933 Contrast with @option{--enable-host-shared}, which affects @emph{host}
936 @item --enable-host-shared
937 Specify that the @emph{host} code should be built into position-independent
938 machine code (with -fPIC), allowing it to be used within shared libraries,
939 but yielding a slightly slower compiler.
941 This option is required when building the libgccjit.so library.
943 Contrast with @option{--enable-shared}, which affects @emph{target}
946 @item @anchor{with-gnu-as}--with-gnu-as
947 Specify that the compiler should assume that the
948 assembler it finds is the GNU assembler. However, this does not modify
949 the rules to find an assembler and will result in confusion if the
950 assembler found is not actually the GNU assembler. (Confusion may also
951 result if the compiler finds the GNU assembler but has not been
952 configured with @option{--with-gnu-as}.) If you have more than one
953 assembler installed on your system, you may want to use this option in
954 connection with @option{--with-as=@var{pathname}} or
955 @option{--with-build-time-tools=@var{pathname}}.
957 The following systems are the only ones where it makes a difference
958 whether you use the GNU assembler. On any other system,
959 @option{--with-gnu-as} has no effect.
962 @item @samp{hppa1.0-@var{any}-@var{any}}
963 @item @samp{hppa1.1-@var{any}-@var{any}}
964 @item @samp{sparc-sun-solaris2.@var{any}}
965 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
968 @item @anchor{with-as}--with-as=@var{pathname}
969 Specify that the compiler should use the assembler pointed to by
970 @var{pathname}, rather than the one found by the standard rules to find
971 an assembler, which are:
974 Unless GCC is being built with a cross compiler, check the
975 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
976 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
977 @var{exec-prefix} defaults to @var{prefix}, which
978 defaults to @file{/usr/local} unless overridden by the
979 @option{--prefix=@var{pathname}} switch described above. @var{target}
980 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
981 @var{version} denotes the GCC version, such as 3.0.
984 If the target system is the same that you are building on, check
985 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
989 Check in the @env{PATH} for a tool whose name is prefixed by the
990 target system triple.
993 Check in the @env{PATH} for a tool whose name is not prefixed by the
994 target system triple, if the host and target system triple are
995 the same (in other words, we use a host tool if it can be used for
999 You may want to use @option{--with-as} if no assembler
1000 is installed in the directories listed above, or if you have multiple
1001 assemblers installed and want to choose one that is not found by the
1004 @item @anchor{with-gnu-ld}--with-gnu-ld
1005 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
1008 @item --with-ld=@var{pathname}
1009 Same as @uref{#with-as,,@option{--with-as}}
1013 Specify that stabs debugging
1014 information should be used instead of whatever format the host normally
1015 uses. Normally GCC uses the same debug format as the host system.
1017 @item --with-tls=@var{dialect}
1018 Specify the default TLS dialect, for systems were there is a choice.
1019 For ARM targets, possible values for @var{dialect} are @code{gnu} or
1020 @code{gnu2}, which select between the original GNU dialect and the GNU TLS
1021 descriptor-based dialect.
1023 @item --enable-multiarch
1024 Specify whether to enable or disable multiarch support. The default is
1025 to check for glibc start files in a multiarch location, and enable it
1026 if the files are found. The auto detection is enabled for native builds,
1027 and for cross builds configured with @option{--with-sysroot}, and without
1028 @option{--with-native-system-header-dir}.
1029 More documentation about multiarch can be found at
1030 @uref{https://wiki.debian.org/Multiarch}.
1032 @item --enable-sjlj-exceptions
1033 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1034 @samp{configure} ordinarily picks the correct value based on the platform.
1035 Only use this option if you are sure you need a different setting.
1037 @item --enable-vtable-verify
1038 Specify whether to enable or disable the vtable verification feature.
1039 Enabling this feature causes libstdc++ to be built with its virtual calls
1040 in verifiable mode. This means that, when linked with libvtv, every
1041 virtual call in libstdc++ will verify the vtable pointer through which the
1042 call will be made before actually making the call. If not linked with libvtv,
1043 the verifier will call stub functions (in libstdc++ itself) and do nothing.
1044 If vtable verification is disabled, then libstdc++ is not built with its
1045 virtual calls in verifiable mode at all. However the libvtv library will
1046 still be built (see @option{--disable-libvtv} to turn off building libvtv).
1047 @option{--disable-vtable-verify} is the default.
1049 @item --disable-gcov
1050 Specify that the run-time library used for coverage analysis
1051 and associated host tools should not be built.
1053 @item --disable-multilib
1054 Specify that multiple target
1055 libraries to support different target variants, calling
1056 conventions, etc.@: should not be built. The default is to build a
1057 predefined set of them.
1059 Some targets provide finer-grained control over which multilibs are built
1060 (e.g., @option{--disable-softfloat}):
1063 fpu, 26bit, underscore, interwork, biendian, nofmult.
1066 softfloat, m68881, m68000, m68020.
1069 single-float, biendian, softfloat.
1071 @item powerpc*-*-*, rs6000*-*-*
1072 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
1077 @item --with-multilib-list=@var{list}
1078 @itemx --without-multilib-list
1079 Specify what multilibs to build. @var{list} is a comma separated list of
1080 values, possibly consisting of a single value. Currently only implemented
1081 for aarch64*-*-*, arm*-*-*, riscv*-*-*, sh*-*-* and x86-64-*-linux*. The
1082 accepted values and meaning for each target is given below.
1086 @var{list} is a comma separated list of @code{ilp32}, and @code{lp64}
1087 to enable ILP32 and LP64 run-time libraries, respectively. If
1088 @var{list} is empty, then there will be no multilibs and only the
1089 default run-time library will be built. If @var{list} is
1090 @code{default} or --with-multilib-list= is not specified, then the
1091 default set of libraries is selected based on the value of
1095 @var{list} is a comma separated list of @code{aprofile} and
1096 @code{rmprofile} to build multilibs for A or R and M architecture
1097 profiles respectively. Note that, due to some limitation of the current
1098 multilib framework, using the combined @code{aprofile,rmprofile}
1099 multilibs selects in some cases a less optimal multilib than when using
1100 the multilib profile for the architecture targetted. The special value
1101 @code{default} is also accepted and is equivalent to omitting the
1102 option, i.e., only the default run-time library will be enabled.
1104 @var{list} may instead contain @code{@@name}, to use the multilib
1105 configuration Makefile fragment @file{name} in @file{gcc/config/arm} in
1106 the source tree (it is part of the corresponding sources, after all).
1107 It is recommended, but not required, that files used for this purpose to
1108 be named starting with @file{t-ml-}, to make their intended purpose
1109 self-evident, in line with GCC conventions. Such files enable custom,
1110 user-chosen multilib lists to be configured. Whether multiple such
1111 files can be used together depends on the contents of the supplied
1112 files. See @file{gcc/config/arm/t-multilib} and its supplementary
1113 @file{gcc/config/arm/t-*profile} files for an example of what such
1114 Makefile fragments might look like for this version of GCC. The macros
1115 expected to be defined in these fragments are not stable across GCC
1116 releases, so make sure they define the @code{MULTILIB}-related macros
1117 expected by the version of GCC you are building.
1119 @xref{Target Fragment,, Target Makefile Fragments, gccint, GNU Compiler
1120 Collection (GCC) Internals}.
1123 See ``Target Makefile Fragments'' in the internals manual.
1126 The table below gives the combination of ISAs, architectures, FPUs and
1127 floating-point ABIs for which multilibs are built for each predefined
1128 profile. The union of these options is considered when specifying both
1129 @code{aprofile} and @code{rmprofile}.
1131 @multitable @columnfractions .15 .28 .30
1132 @item Option @tab aprofile @tab rmprofile
1134 @tab @code{-marm} and @code{-mthumb}
1136 @item Architectures@*@*@*@*@*@*
1137 @tab default architecture@*
1138 @code{-march=armv7-a}@*
1139 @code{-march=armv7ve}@*
1140 @code{-march=armv8-a}@*@*@*
1141 @tab default architecture@*
1142 @code{-march=armv6s-m}@*
1143 @code{-march=armv7-m}@*
1144 @code{-march=armv7e-m}@*
1145 @code{-march=armv8-m.base}@*
1146 @code{-march=armv8-m.main}@*
1148 @item FPUs@*@*@*@*@*
1150 @code{-mfpu=vfpv3-d16}@*
1152 @code{-mfpu=vfpv4-d16}@*
1153 @code{-mfpu=neon-vfpv4}@*
1154 @code{-mfpu=neon-fp-armv8}
1156 @code{-mfpu=vfpv3-d16}@*
1157 @code{-mfpu=fpv4-sp-d16}@*
1158 @code{-mfpu=fpv5-sp-d16}@*
1159 @code{-mfpu=fpv5-d16}@*
1160 @item floating-point@/ ABIs@*@*
1161 @tab @code{-mfloat-abi=soft}@*
1162 @code{-mfloat-abi=softfp}@*
1163 @code{-mfloat-abi=hard}
1164 @tab @code{-mfloat-abi=soft}@*
1165 @code{-mfloat-abi=softfp}@*
1166 @code{-mfloat-abi=hard}
1170 @var{list} is a single ABI name. The target architecture must be either
1171 @code{rv32gc} or @code{rv64gc}. This will build a single multilib for the
1172 specified architecture and ABI pair. If @code{--with-multilib-list} is not
1173 given, then a default set of multilibs is selected based on the value of
1174 @option{--target}. This is usually a large set of multilibs.
1177 @var{list} is a comma separated list of CPU names. These must be of the
1178 form @code{sh*} or @code{m*} (in which case they match the compiler option
1179 for that processor). The list should not contain any endian options -
1180 these are handled by @option{--with-endian}.
1182 If @var{list} is empty, then there will be no multilibs for extra
1183 processors. The multilib for the secondary endian remains enabled.
1185 As a special case, if an entry in the list starts with a @code{!}
1186 (exclamation point), then it is added to the list of excluded multilibs.
1187 Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES}
1188 (once the leading @code{!} has been stripped).
1190 If @option{--with-multilib-list} is not given, then a default set of
1191 multilibs is selected based on the value of @option{--target}. This is
1192 usually the complete set of libraries, but some targets imply a more
1195 Example 1: to configure a compiler for SH4A only, but supporting both
1196 endians, with little endian being the default:
1198 --with-cpu=sh4a --with-endian=little,big --with-multilib-list=
1201 Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with
1202 only little endian SH4AL:
1204 --with-cpu=sh4a --with-endian=little,big \
1205 --with-multilib-list=sh4al,!mb/m4al
1208 @item x86-64-*-linux*
1209 @var{list} is a comma separated list of @code{m32}, @code{m64} and
1210 @code{mx32} to enable 32-bit, 64-bit and x32 run-time libraries,
1211 respectively. If @var{list} is empty, then there will be no multilibs
1212 and only the default run-time library will be enabled.
1214 If @option{--with-multilib-list} is not given, then only 32-bit and
1215 64-bit run-time libraries will be enabled.
1218 @item --with-endian=@var{endians}
1219 Specify what endians to use.
1220 Currently only implemented for sh*-*-*.
1222 @var{endians} may be one of the following:
1225 Use big endian exclusively.
1227 Use little endian exclusively.
1229 Use big endian by default. Provide a multilib for little endian.
1231 Use little endian by default. Provide a multilib for big endian.
1234 @item --enable-threads
1235 Specify that the target
1236 supports threads. This affects the Objective-C compiler and runtime
1237 library, and exception handling for other languages like C++.
1238 On some systems, this is the default.
1240 In general, the best (and, in many cases, the only known) threading
1241 model available will be configured for use. Beware that on some
1242 systems, GCC has not been taught what threading models are generally
1243 available for the system. In this case, @option{--enable-threads} is an
1244 alias for @option{--enable-threads=single}.
1246 @item --disable-threads
1247 Specify that threading support should be disabled for the system.
1248 This is an alias for @option{--enable-threads=single}.
1250 @item --enable-threads=@var{lib}
1252 @var{lib} is the thread support library. This affects the Objective-C
1253 compiler and runtime library, and exception handling for other languages
1254 like C++. The possibilities for @var{lib} are:
1262 LynxOS thread support.
1264 MIPS SDE thread support.
1266 This is an alias for @samp{single}.
1268 Generic POSIX/Unix98 thread support.
1270 RTEMS thread support.
1272 Disable thread support, should work for all platforms.
1276 VxWorks thread support.
1278 Microsoft Win32 API thread support.
1282 Specify that the target supports TLS (Thread Local Storage). Usually
1283 configure can correctly determine if TLS is supported. In cases where
1284 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1285 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1286 the assembler supports TLS but the C library does not, or if the
1287 assumptions made by the configure test are incorrect.
1290 Specify that the target does not support TLS.
1291 This is an alias for @option{--enable-tls=no}.
1293 @item --disable-tm-clone-registry
1294 Disable TM clone registry in libgcc. It is enabled in libgcc by default.
1295 This option helps to reduce code size for embedded targets which do
1296 not use transactional memory.
1298 @item --with-cpu=@var{cpu}
1299 @itemx --with-cpu-32=@var{cpu}
1300 @itemx --with-cpu-64=@var{cpu}
1301 Specify which cpu variant the compiler should generate code for by default.
1302 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1303 This option is only supported on some targets, including ARC, ARM, i386, M68k,
1304 PowerPC, and SPARC@. It is mandatory for ARC@. The @option{--with-cpu-32} and
1305 @option{--with-cpu-64} options specify separate default CPUs for
1306 32-bit and 64-bit modes; these options are only supported for i386,
1307 x86-64, PowerPC, and SPARC@.
1309 @item --with-schedule=@var{cpu}
1310 @itemx --with-arch=@var{cpu}
1311 @itemx --with-arch-32=@var{cpu}
1312 @itemx --with-arch-64=@var{cpu}
1313 @itemx --with-tune=@var{cpu}
1314 @itemx --with-tune-32=@var{cpu}
1315 @itemx --with-tune-64=@var{cpu}
1316 @itemx --with-abi=@var{abi}
1317 @itemx --with-fpu=@var{type}
1318 @itemx --with-float=@var{type}
1319 These configure options provide default values for the @option{-mschedule=},
1320 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1321 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1322 @option{--with-cpu}, which switches will be accepted and acceptable values
1323 of the arguments depend on the target.
1325 @item --with-mode=@var{mode}
1326 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1327 This option is only supported on ARM targets.
1329 @item --with-stack-offset=@var{num}
1330 This option sets the default for the -mstack-offset=@var{num} option,
1331 and will thus generally also control the setting of this option for
1332 libraries. This option is only supported on Epiphany targets.
1334 @item --with-fpmath=@var{isa}
1335 This options sets @option{-mfpmath=sse} by default and specifies the default
1336 ISA for floating-point arithmetics. You can select either @samp{sse} which
1337 enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default.
1338 This option is only supported on i386 and x86-64 targets.
1340 @item --with-fp-32=@var{mode}
1341 On MIPS targets, set the default value for the @option{-mfp} option when using
1342 the o32 ABI. The possibilities for @var{mode} are:
1345 Use the o32 FP32 ABI extension, as with the @option{-mfp32} command-line
1348 Use the o32 FPXX ABI extension, as with the @option{-mfpxx} command-line
1351 Use the o32 FP64 ABI extension, as with the @option{-mfp64} command-line
1354 In the absence of this configuration option the default is to use the o32
1357 @item --with-odd-spreg-32
1358 On MIPS targets, set the @option{-modd-spreg} option by default when using
1361 @item --without-odd-spreg-32
1362 On MIPS targets, set the @option{-mno-odd-spreg} option by default when using
1363 the o32 ABI. This is normally used in conjunction with
1364 @option{--with-fp-32=64} in order to target the o32 FP64A ABI extension.
1366 @item --with-nan=@var{encoding}
1367 On MIPS targets, set the default encoding convention to use for the
1368 special not-a-number (NaN) IEEE 754 floating-point data. The
1369 possibilities for @var{encoding} are:
1372 Use the legacy encoding, as with the @option{-mnan=legacy} command-line
1375 Use the 754-2008 encoding, as with the @option{-mnan=2008} command-line
1378 To use this configuration option you must have an assembler version
1379 installed that supports the @option{-mnan=} command-line option too.
1380 In the absence of this configuration option the default convention is
1381 the legacy encoding, as when neither of the @option{-mnan=2008} and
1382 @option{-mnan=legacy} command-line options has been used.
1384 @item --with-divide=@var{type}
1385 Specify how the compiler should generate code for checking for
1386 division by zero. This option is only supported on the MIPS target.
1387 The possibilities for @var{type} are:
1390 Division by zero checks use conditional traps (this is the default on
1391 systems that support conditional traps).
1393 Division by zero checks use the break instruction.
1396 @c If you make --with-llsc the default for additional targets,
1397 @c update the --with-llsc description in the MIPS section below.
1400 On MIPS targets, make @option{-mllsc} the default when no
1401 @option{-mno-llsc} option is passed. This is the default for
1402 Linux-based targets, as the kernel will emulate them if the ISA does
1405 @item --without-llsc
1406 On MIPS targets, make @option{-mno-llsc} the default when no
1407 @option{-mllsc} option is passed.
1410 On MIPS targets, make @option{-msynci} the default when no
1411 @option{-mno-synci} option is passed.
1413 @item --without-synci
1414 On MIPS targets, make @option{-mno-synci} the default when no
1415 @option{-msynci} option is passed. This is the default.
1417 @item --with-lxc1-sxc1
1418 On MIPS targets, make @option{-mlxc1-sxc1} the default when no
1419 @option{-mno-lxc1-sxc1} option is passed. This is the default.
1421 @item --without-lxc1-sxc1
1422 On MIPS targets, make @option{-mno-lxc1-sxc1} the default when no
1423 @option{-mlxc1-sxc1} option is passed. The indexed load/store
1424 instructions are not directly a problem but can lead to unexpected
1425 behaviour when deployed in an application intended for a 32-bit address
1426 space but run on a 64-bit processor. The issue is seen because all
1427 known MIPS 64-bit Linux kernels execute o32 and n32 applications
1428 with 64-bit addressing enabled which affects the overflow behaviour
1429 of the indexed addressing mode. GCC will assume that ordinary
1430 32-bit arithmetic overflow behaviour is the same whether performed
1431 as an @code{addu} instruction or as part of the address calculation
1432 in @code{lwxc1} type instructions. This assumption holds true in a
1433 pure 32-bit environment and can hold true in a 64-bit environment if
1434 the address space is accurately set to be 32-bit for o32 and n32.
1437 On MIPS targets, make @option{-mmadd4} the default when no
1438 @option{-mno-madd4} option is passed. This is the default.
1440 @item --without-madd4
1441 On MIPS targets, make @option{-mno-madd4} the default when no
1442 @option{-mmadd4} option is passed. The @code{madd4} instruction
1443 family can be problematic when targeting a combination of cores that
1444 implement these instructions differently. There are two known cores
1445 that implement these as fused operations instead of unfused (where
1446 unfused is normally expected). Disabling these instructions is the
1447 only way to ensure compatible code is generated; this will incur
1448 a performance penalty.
1450 @item --with-mips-plt
1451 On MIPS targets, make use of copy relocations and PLTs.
1452 These features are extensions to the traditional
1453 SVR4-based MIPS ABIs and require support from GNU binutils
1454 and the runtime C library.
1456 @item --with-stack-clash-protection-guard-size=@var{size}
1457 On certain targets this option sets the default stack clash protection guard
1458 size as a power of two in bytes. On AArch64 @var{size} is required to be either
1459 12 (4KB) or 16 (64KB).
1461 @item --enable-__cxa_atexit
1462 Define if you want to use __cxa_atexit, rather than atexit, to
1463 register C++ destructors for local statics and global objects.
1464 This is essential for fully standards-compliant handling of
1465 destructors, but requires __cxa_atexit in libc. This option is currently
1466 only available on systems with GNU libc. When enabled, this will cause
1467 @option{-fuse-cxa-atexit} to be passed by default.
1469 @item --enable-gnu-indirect-function
1470 Define if you want to enable the @code{ifunc} attribute. This option is
1471 currently only available on systems with GNU libc on certain targets.
1473 @item --enable-target-optspace
1475 libraries should be optimized for code space instead of code speed.
1476 This is the default for the m32r platform.
1478 @item --with-cpp-install-dir=@var{dirname}
1479 Specify that the user visible @command{cpp} program should be installed
1480 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1482 @item --enable-comdat
1483 Enable COMDAT group support. This is primarily used to override the
1484 automatically detected value.
1486 @item --enable-initfini-array
1487 Force the use of sections @code{.init_array} and @code{.fini_array}
1488 (instead of @code{.init} and @code{.fini}) for constructors and
1489 destructors. Option @option{--disable-initfini-array} has the
1490 opposite effect. If neither option is specified, the configure script
1491 will try to guess whether the @code{.init_array} and
1492 @code{.fini_array} sections are supported and, if they are, use them.
1494 @item --enable-link-mutex
1495 When building GCC, use a mutex to avoid linking the compilers for
1496 multiple languages at the same time, to avoid thrashing on build
1497 systems with limited free memory. The default is not to use such a mutex.
1499 @item --enable-maintainer-mode
1500 The build rules that regenerate the Autoconf and Automake output files as
1501 well as the GCC master message catalog @file{gcc.pot} are normally
1502 disabled. This is because it can only be rebuilt if the complete source
1503 tree is present. If you have changed the sources and want to rebuild the
1504 catalog, configuring with @option{--enable-maintainer-mode} will enable
1505 this. Note that you need a recent version of the @code{gettext} tools
1508 @item --disable-bootstrap
1509 For a native build, the default configuration is to perform
1510 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1511 testing that GCC can compile itself correctly. If you want to disable
1512 this process, you can configure with @option{--disable-bootstrap}.
1514 @item --enable-bootstrap
1515 In special cases, you may want to perform a 3-stage build
1516 even if the target and host triplets are different.
1517 This is possible when the host can run code compiled for
1518 the target (e.g.@: host is i686-linux, target is i486-linux).
1519 Starting from GCC 4.2, to do this you have to configure explicitly
1520 with @option{--enable-bootstrap}.
1522 @item --enable-generated-files-in-srcdir
1523 Neither the .c and .h files that are generated from Bison and flex nor the
1524 info manuals and man pages that are built from the .texi files are present
1525 in the SVN development tree. When building GCC from that development tree,
1526 or from one of our snapshots, those generated files are placed in your
1527 build directory, which allows for the source to be in a readonly
1530 If you configure with @option{--enable-generated-files-in-srcdir} then those
1531 generated files will go into the source directory. This is mainly intended
1532 for generating release or prerelease tarballs of the GCC sources, since it
1533 is not a requirement that the users of source releases to have flex, Bison,
1536 @item --enable-version-specific-runtime-libs
1538 that runtime libraries should be installed in the compiler specific
1539 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1540 addition, @samp{libstdc++}'s include files will be installed into
1541 @file{@var{libdir}} unless you overruled it by using
1542 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1543 particularly useful if you intend to use several versions of GCC in
1544 parallel. This is currently supported by @samp{libgfortran},
1545 @samp{libstdc++}, and @samp{libobjc}.
1547 @item @anchor{WithAixSoname}--with-aix-soname=@samp{aix}, @samp{svr4} or @samp{both}
1548 Traditional AIX shared library versioning (versioned @code{Shared Object}
1549 files as members of unversioned @code{Archive Library} files named
1550 @samp{lib.a}) causes numerous headaches for package managers. However,
1551 @code{Import Files} as members of @code{Archive Library} files allow for
1552 @strong{filename-based versioning} of shared libraries as seen on Linux/SVR4,
1553 where this is called the "SONAME". But as they prevent static linking,
1554 @code{Import Files} may be used with @code{Runtime Linking} only, where the
1555 linker does search for @samp{libNAME.so} before @samp{libNAME.a} library
1556 filenames with the @samp{-lNAME} linker flag.
1558 @anchor{AixLdCommand}For detailed information please refer to the AIX
1559 @uref{https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22,,ld
1562 As long as shared library creation is enabled, upon:
1564 @item --with-aix-soname=aix
1565 @item --with-aix-soname=both
1566 A (traditional AIX) @code{Shared Archive Library} file is created:
1568 @item using the @samp{libNAME.a} filename scheme
1569 @item with the @code{Shared Object} file as archive member named
1570 @samp{libNAME.so.V} (except for @samp{libgcc_s}, where the @code{Shared
1571 Object} file is named @samp{shr.o} for backwards compatibility), which
1573 @item is used for runtime loading from inside the @samp{libNAME.a} file
1574 @item is used for dynamic loading via
1575 @code{dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)}
1576 @item is used for shared linking
1577 @item is used for static linking, so no separate @code{Static Archive
1578 Library} file is needed
1581 @item --with-aix-soname=both
1582 @item --with-aix-soname=svr4
1583 A (second) @code{Shared Archive Library} file is created:
1585 @item using the @samp{libNAME.so.V} filename scheme
1586 @item with the @code{Shared Object} file as archive member named
1589 @item is created with the @code{-G linker flag}
1590 @item has the @code{F_LOADONLY} flag set
1591 @item is used for runtime loading from inside the @samp{libNAME.so.V} file
1592 @item is used for dynamic loading via @code{dlopen("libNAME.so.V(shr.o)",
1595 @item with the @code{Import File} as archive member named @samp{shr.imp},
1598 @item refers to @samp{libNAME.so.V(shr.o)} as the "SONAME", to be recorded
1599 in the @code{Loader Section} of subsequent binaries
1600 @item indicates whether @samp{libNAME.so.V(shr.o)} is 32 or 64 bit
1601 @item lists all the public symbols exported by @samp{lib.so.V(shr.o)},
1602 eventually decorated with the @code{@samp{weak} Keyword}
1603 @item is necessary for shared linking against @samp{lib.so.V(shr.o)}
1606 A symbolic link using the @samp{libNAME.so} filename scheme is created:
1608 @item pointing to the @samp{libNAME.so.V} @code{Shared Archive Library} file
1609 @item to permit the @code{ld Command} to find @samp{lib.so.V(shr.imp)} via
1610 the @samp{-lNAME} argument (requires @code{Runtime Linking} to be enabled)
1611 @item to permit dynamic loading of @samp{lib.so.V(shr.o)} without the need
1612 to specify the version number via @code{dlopen("libNAME.so(shr.o)",
1617 As long as static library creation is enabled, upon:
1619 @item --with-aix-soname=svr4
1620 A @code{Static Archive Library} is created:
1622 @item using the @samp{libNAME.a} filename scheme
1623 @item with all the @code{Static Object} files as archive members, which
1625 @item are used for static linking
1630 While the aix-soname=@samp{svr4} option does not create @code{Shared Object}
1631 files as members of unversioned @code{Archive Library} files any more, package
1632 managers still are responsible to
1633 @uref{./specific.html#TransferAixShobj,,transfer} @code{Shared Object} files
1634 found as member of a previously installed unversioned @code{Archive Library}
1635 file into the newly installed @code{Archive Library} file with the same
1638 @emph{WARNING:} Creating @code{Shared Object} files with @code{Runtime Linking}
1639 enabled may bloat the TOC, eventually leading to @code{TOC overflow} errors,
1640 requiring the use of either the @option{-Wl,-bbigtoc} linker flag (seen to
1641 break with the @code{GDB} debugger) or some of the TOC-related compiler flags,
1643 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1644 Using the GNU Compiler Collection (GCC)}.
1647 see ``RS/6000 and PowerPC Options'' in the main manual.
1650 @option{--with-aix-soname} is currently supported by @samp{libgcc_s} only, so
1651 this option is still experimental and not for normal use yet.
1653 Default is the traditional behavior @option{--with-aix-soname=@samp{aix}}.
1655 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1656 Specify that only a particular subset of compilers and
1657 their runtime libraries should be built. For a list of valid values for
1658 @var{langN} you can issue the following command in the
1659 @file{gcc} directory of your GCC source tree:@*
1661 grep ^language= */config-lang.in
1663 Currently, you can use any of the following:
1664 @code{all}, @code{default}, @code{ada}, @code{c}, @code{c++}, @code{d},
1665 @code{fortran}, @code{go}, @code{jit}, @code{lto}, @code{objc}, @code{obj-c++}.
1666 Building the Ada compiler has special requirements, see below.
1667 If you do not pass this flag, or specify the option @code{default}, then the
1668 default languages available in the @file{gcc} sub-tree will be configured.
1669 Ada, D, Go, Jit, and Objective-C++ are not default languages. LTO is not a
1670 default language, but is built by default because @option{--enable-lto} is
1671 enabled by default. The other languages are default languages. If
1672 @code{all} is specified, then all available languages are built. An
1673 exception is @code{jit} language, which requires
1674 @option{--enable-host-shared} to be included with @code{all}.
1676 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1677 Specify that a particular subset of compilers and their runtime
1678 libraries should be built with the system C compiler during stage 1 of
1679 the bootstrap process, rather than only in later stages with the
1680 bootstrapped C compiler. The list of valid values is the same as for
1681 @option{--enable-languages}, and the option @code{all} will select all
1682 of the languages enabled by @option{--enable-languages}. This option is
1683 primarily useful for GCC development; for instance, when a development
1684 version of the compiler cannot bootstrap due to compiler bugs, or when
1685 one is debugging front ends other than the C front end. When this
1686 option is used, one can then build the target libraries for the
1687 specified languages with the stage-1 compiler by using @command{make
1688 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1689 for the specified languages using @command{make stage1-start check-gcc}.
1691 @item --disable-libada
1692 Specify that the run-time libraries and tools used by GNAT should not
1693 be built. This can be useful for debugging, or for compatibility with
1694 previous Ada build procedures, when it was required to explicitly
1695 do a @samp{make -C gcc gnatlib_and_tools}.
1697 @item --disable-libsanitizer
1698 Specify that the run-time libraries for the various sanitizers should
1701 @item --disable-libssp
1702 Specify that the run-time libraries for stack smashing protection
1703 should not be built or linked against. On many targets library support
1704 is provided by the C library instead.
1706 @item --disable-libquadmath
1707 Specify that the GCC quad-precision math library should not be built.
1708 On some systems, the library is required to be linkable when building
1709 the Fortran front end, unless @option{--disable-libquadmath-support}
1712 @item --disable-libquadmath-support
1713 Specify that the Fortran front end and @code{libgfortran} do not add
1714 support for @code{libquadmath} on systems supporting it.
1716 @item --disable-libgomp
1717 Specify that the GNU Offloading and Multi Processing Runtime Library
1718 should not be built.
1720 @item --disable-libvtv
1721 Specify that the run-time libraries used by vtable verification
1722 should not be built.
1725 Specify that the compiler should
1726 use DWARF 2 debugging information as the default.
1728 @item --with-advance-toolchain=@var{at}
1729 On 64-bit PowerPC Linux systems, configure the compiler to use the
1730 header files, library files, and the dynamic linker from the Advance
1731 Toolchain release @var{at} instead of the default versions that are
1732 provided by the Linux distribution. In general, this option is
1733 intended for the developers of GCC, and it is not intended for general
1736 @item --enable-targets=all
1737 @itemx --enable-targets=@var{target_list}
1738 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1739 These are compilers that are able to generate either 64-bit or 32-bit
1740 code. Typically, the corresponding 32-bit target, e.g.@:
1741 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1742 option enables the 32-bit target to be a bi-arch compiler, which is
1743 useful when you want a bi-arch compiler that defaults to 32-bit, and
1744 you are building a bi-arch or multi-arch binutils in a combined tree.
1745 On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64),
1747 Currently, this option only affects sparc-linux, powerpc-linux, x86-linux,
1748 mips-linux and s390-linux.
1750 @item --enable-default-pie
1751 Turn on @option{-fPIE} and @option{-pie} by default.
1753 @item --enable-secureplt
1754 This option enables @option{-msecure-plt} by default for powerpc-linux.
1756 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1757 Using the GNU Compiler Collection (GCC)},
1760 See ``RS/6000 and PowerPC Options'' in the main manual
1763 @item --enable-default-ssp
1764 Turn on @option{-fstack-protector-strong} by default.
1767 This option enables @option{-mcld} by default for 32-bit x86 targets.
1769 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1770 Using the GNU Compiler Collection (GCC)},
1773 See ``i386 and x86-64 Options'' in the main manual
1776 @item --enable-large-address-aware
1777 The @option{--enable-large-address-aware} option arranges for MinGW
1778 executables to be linked using the @option{--large-address-aware}
1779 option, that enables the use of more than 2GB of memory. If GCC is
1780 configured with this option, its effects can be reversed by passing the
1781 @option{-Wl,--disable-large-address-aware} option to the so-configured
1784 @item --enable-win32-registry
1785 @itemx --enable-win32-registry=@var{key}
1786 @itemx --disable-win32-registry
1787 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1788 to look up installations paths in the registry using the following key:
1791 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1794 @var{key} defaults to GCC version number, and can be overridden by the
1795 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1796 who use custom installers are encouraged to provide a different key,
1797 perhaps one comprised of vendor name and GCC version number, to
1798 avoid conflict with existing installations. This feature is enabled
1799 by default, and can be disabled by @option{--disable-win32-registry}
1800 option. This option has no effect on the other hosts.
1803 Specify that the machine does not have a floating point unit. This
1804 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1805 system, @option{--nfp} has no effect.
1807 @item --enable-werror
1808 @itemx --disable-werror
1809 @itemx --enable-werror=yes
1810 @itemx --enable-werror=no
1811 When you specify this option, it controls whether certain files in the
1812 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1813 If you don't specify it, @option{-Werror} is turned on for the main
1814 development trunk. However it defaults to off for release branches and
1815 final releases. The specific files which get @option{-Werror} are
1816 controlled by the Makefiles.
1818 @item --enable-checking
1819 @itemx --enable-checking=@var{list}
1820 When you specify this option, the compiler is built to perform internal
1821 consistency checks of the requested complexity. This does not change the
1822 generated code, but adds error checking within the compiler. This will
1823 slow down the compiler and may only work properly if you are building
1824 the compiler with GCC@. This is @samp{yes,extra} by default when building
1825 from SVN or snapshots, but @samp{release} for releases. The default
1826 for building the stage1 compiler is @samp{yes}. More control
1827 over the checks may be had by specifying @var{list}. The categories of
1828 checks available are @samp{yes} (most common checks
1829 @samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at
1830 all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest
1831 checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1832 Individual checks can be enabled with these flags @samp{assert},
1833 @samp{df}, @samp{fold}, @samp{gc}, @samp{gcac}, @samp{misc}, @samp{rtl},
1834 @samp{rtlflag}, @samp{runtime}, @samp{tree}, @samp{extra} and @samp{valgrind}.
1835 @samp{extra} adds for @samp{misc} checking extra checks that might affect
1836 code generation and should therefore not differ between stage1 and later
1839 The @samp{valgrind} check requires the external @command{valgrind}
1840 simulator, available from @uref{http://valgrind.org/}. The
1841 @samp{df}, @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive.
1842 To disable all checking, @samp{--disable-checking} or
1843 @samp{--enable-checking=none} must be explicitly requested. Disabling
1844 assertions will make the compiler and runtime slightly faster but
1845 increase the risk of undetected internal errors causing wrong code to be
1848 @item --disable-stage1-checking
1849 @itemx --enable-stage1-checking
1850 @itemx --enable-stage1-checking=@var{list}
1851 If no @option{--enable-checking} option is specified the stage1
1852 compiler will be built with @samp{yes} checking enabled, otherwise
1853 the stage1 checking flags are the same as specified by
1854 @option{--enable-checking}. To build the stage1 compiler with
1855 different checking options use @option{--enable-stage1-checking}.
1856 The list of checking options is the same as for @option{--enable-checking}.
1857 If your system is too slow or too small to bootstrap a released compiler
1858 with checking for stage1 enabled, you can use @samp{--disable-stage1-checking}
1859 to disable checking for the stage1 compiler.
1861 @item --enable-coverage
1862 @itemx --enable-coverage=@var{level}
1863 With this option, the compiler is built to collect self coverage
1864 information, every time it is run. This is for internal development
1865 purposes, and only works when the compiler is being built with gcc. The
1866 @var{level} argument controls whether the compiler is built optimized or
1867 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1868 want to disable optimization, for performance analysis you want to
1869 enable optimization. When coverage is enabled, the default level is
1870 without optimization.
1872 @item --enable-gather-detailed-mem-stats
1873 When this option is specified more detailed information on memory
1874 allocation is gathered. This information is printed when using
1875 @option{-fmem-report}.
1877 @item --enable-valgrind-annotations
1878 Mark selected memory related operations in the compiler when run under
1879 valgrind to suppress false positives.
1882 @itemx --disable-nls
1883 The @option{--enable-nls} option enables Native Language Support (NLS),
1884 which lets GCC output diagnostics in languages other than American
1885 English. Native Language Support is enabled by default if not doing a
1886 canadian cross build. The @option{--disable-nls} option disables NLS@.
1888 @item --with-included-gettext
1889 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1890 procedure to prefer its copy of GNU @command{gettext}.
1892 @item --with-catgets
1893 If NLS is enabled, and if the host lacks @code{gettext} but has the
1894 inferior @code{catgets} interface, the GCC build procedure normally
1895 ignores @code{catgets} and instead uses GCC's copy of the GNU
1896 @code{gettext} library. The @option{--with-catgets} option causes the
1897 build procedure to use the host's @code{catgets} in this situation.
1899 @item --with-libiconv-prefix=@var{dir}
1900 Search for libiconv header files in @file{@var{dir}/include} and
1901 libiconv library files in @file{@var{dir}/lib}.
1903 @item --enable-obsolete
1904 Enable configuration for an obsoleted system. If you attempt to
1905 configure GCC for a system (build, host, or target) which has been
1906 obsoleted, and you do not specify this flag, configure will halt with an
1909 All support for systems which have been obsoleted in one release of GCC
1910 is removed entirely in the next major release, unless someone steps
1911 forward to maintain the port.
1913 @item --enable-decimal-float
1914 @itemx --enable-decimal-float=yes
1915 @itemx --enable-decimal-float=no
1916 @itemx --enable-decimal-float=bid
1917 @itemx --enable-decimal-float=dpd
1918 @itemx --disable-decimal-float
1919 Enable (or disable) support for the C decimal floating point extension
1920 that is in the IEEE 754-2008 standard. This is enabled by default only
1921 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
1922 support it, but require the user to specifically enable it. You can
1923 optionally control which decimal floating point format is used (either
1924 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
1925 format is default on i386 and x86_64 systems, and the @samp{dpd}
1926 (densely packed decimal) format is default on PowerPC systems.
1928 @item --enable-fixed-point
1929 @itemx --disable-fixed-point
1930 Enable (or disable) support for C fixed-point arithmetic.
1931 This option is enabled by default for some targets (such as MIPS) which
1932 have hardware-support for fixed-point operations. On other targets, you
1933 may enable this option manually.
1935 @item --with-long-double-128
1936 Specify if @code{long double} type should be 128-bit by default on selected
1937 GNU/Linux architectures. If using @code{--without-long-double-128},
1938 @code{long double} will be by default 64-bit, the same as @code{double} type.
1939 When neither of these configure options are used, the default will be
1940 128-bit @code{long double} when built against GNU C Library 2.4 and later,
1941 64-bit @code{long double} otherwise.
1943 @item --with-long-double-format=ibm
1944 @itemx --with-long-double-format=ieee
1945 Specify whether @code{long double} uses the IBM extended double format
1946 or the IEEE 128-bit floating point format on PowerPC Linux systems.
1947 This configuration switch will only work on little endian PowerPC
1948 Linux systems and on big endian 64-bit systems where the default cpu
1949 is at least power7 (i.e.@: @option{--with-cpu=power7},
1950 @option{--with-cpu=power8}, or @option{--with-cpu=power9} is used).
1952 If you use the @option{--with-long-double-64} configuration option,
1953 the @option{--with-long-double-format=ibm} and
1954 @option{--with-long-double-format=ieee} options are ignored.
1956 The default @code{long double} format is to use IBM extended double.
1957 Until all of the libraries are converted to use IEEE 128-bit floating
1958 point, it is not recommended to use
1959 @option{--with-long-double-format=ieee}.
1961 On little endian PowerPC Linux systems, if you explicitly set the
1962 @code{long double} type, it will build multilibs to allow you to
1963 select either @code{long double} format, unless you disable multilibs
1964 with the @code{--disable-multilib} option. At present,
1965 @code{long double} multilibs are not built on big endian PowerPC Linux
1966 systems. If you are building multilibs, you will need to configure
1967 the compiler using the @option{--with-system-zlib} option.
1969 If you do not set the @code{long double} type explicitly, no multilibs
1972 @item --enable-fdpic
1973 On SH Linux systems, generate ELF FDPIC code.
1975 @item --with-gmp=@var{pathname}
1976 @itemx --with-gmp-include=@var{pathname}
1977 @itemx --with-gmp-lib=@var{pathname}
1978 @itemx --with-mpfr=@var{pathname}
1979 @itemx --with-mpfr-include=@var{pathname}
1980 @itemx --with-mpfr-lib=@var{pathname}
1981 @itemx --with-mpc=@var{pathname}
1982 @itemx --with-mpc-include=@var{pathname}
1983 @itemx --with-mpc-lib=@var{pathname}
1984 If you want to build GCC but do not have the GMP library, the MPFR
1985 library and/or the MPC library installed in a standard location and
1986 do not have their sources present in the GCC source tree then you
1987 can explicitly specify the directory where they are installed
1988 (@samp{--with-gmp=@var{gmpinstalldir}},
1989 @samp{--with-mpfr=@/@var{mpfrinstalldir}},
1990 @samp{--with-mpc=@/@var{mpcinstalldir}}). The
1991 @option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for
1992 @option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and
1993 @option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the
1994 @option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for
1995 @option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and
1996 @option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the
1997 @option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for
1998 @option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and
1999 @option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these
2000 shorthand assumptions are not correct, you can use the explicit
2001 include and lib options directly. You might also need to ensure the
2002 shared libraries can be found by the dynamic linker when building and
2003 using GCC, for example by setting the runtime shared library path
2004 variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems).
2006 These flags are applicable to the host platform only. When building
2007 a cross compiler, they will not be used to configure target libraries.
2009 @item --with-isl=@var{pathname}
2010 @itemx --with-isl-include=@var{pathname}
2011 @itemx --with-isl-lib=@var{pathname}
2012 If you do not have the isl library installed in a standard location and you
2013 want to build GCC, you can explicitly specify the directory where it is
2014 installed (@samp{--with-isl=@/@var{islinstalldir}}). The
2015 @option{--with-isl=@/@var{islinstalldir}} option is shorthand for
2016 @option{--with-isl-lib=@/@var{islinstalldir}/lib} and
2017 @option{--with-isl-include=@/@var{islinstalldir}/include}. If this
2018 shorthand assumption is not correct, you can use the explicit
2019 include and lib options directly.
2021 These flags are applicable to the host platform only. When building
2022 a cross compiler, they will not be used to configure target libraries.
2024 @item --with-stage1-ldflags=@var{flags}
2025 This option may be used to set linker flags to be used when linking
2026 stage 1 of GCC. These are also used when linking GCC if configured with
2027 @option{--disable-bootstrap}. If @option{--with-stage1-libs} is not set to a
2028 value, then the default is @samp{-static-libstdc++ -static-libgcc}, if
2031 @item --with-stage1-libs=@var{libs}
2032 This option may be used to set libraries to be used when linking stage 1
2033 of GCC. These are also used when linking GCC if configured with
2034 @option{--disable-bootstrap}.
2036 @item --with-boot-ldflags=@var{flags}
2037 This option may be used to set linker flags to be used when linking
2038 stage 2 and later when bootstrapping GCC. If --with-boot-libs
2039 is not is set to a value, then the default is
2040 @samp{-static-libstdc++ -static-libgcc}.
2042 @item --with-boot-libs=@var{libs}
2043 This option may be used to set libraries to be used when linking stage 2
2044 and later when bootstrapping GCC.
2046 @item --with-debug-prefix-map=@var{map}
2047 Convert source directory names using @option{-fdebug-prefix-map} when
2048 building runtime libraries. @samp{@var{map}} is a space-separated
2049 list of maps of the form @samp{@var{old}=@var{new}}.
2051 @item --enable-linker-build-id
2052 Tells GCC to pass @option{--build-id} option to the linker for all final
2053 links (links performed without the @option{-r} or @option{--relocatable}
2054 option), if the linker supports it. If you specify
2055 @option{--enable-linker-build-id}, but your linker does not
2056 support @option{--build-id} option, a warning is issued and the
2057 @option{--enable-linker-build-id} option is ignored. The default is off.
2059 @item --with-linker-hash-style=@var{choice}
2060 Tells GCC to pass @option{--hash-style=@var{choice}} option to the
2061 linker for all final links. @var{choice} can be one of
2062 @samp{sysv}, @samp{gnu}, and @samp{both} where @samp{sysv} is the default.
2064 @item --enable-gnu-unique-object
2065 @itemx --disable-gnu-unique-object
2066 Tells GCC to use the gnu_unique_object relocation for C++ template
2067 static data members and inline function local statics. Enabled by
2068 default for a toolchain with an assembler that accepts it and
2069 GLIBC 2.11 or above, otherwise disabled.
2071 @item --with-diagnostics-color=@var{choice}
2072 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-color=}
2073 option (if not used explicitly on the command line). @var{choice}
2074 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2075 where @samp{auto} is the default. @samp{auto-if-env} means that
2076 @option{-fdiagnostics-color=auto} will be the default if @code{GCC_COLORS}
2077 is present and non-empty in the environment, and
2078 @option{-fdiagnostics-color=never} otherwise.
2081 @itemx --disable-lto
2082 Enable support for link-time optimization (LTO). This is enabled by
2083 default, and may be disabled using @option{--disable-lto}.
2085 @item --enable-linker-plugin-configure-flags=FLAGS
2086 @itemx --enable-linker-plugin-flags=FLAGS
2087 By default, linker plugins (such as the LTO plugin) are built for the
2088 host system architecture. For the case that the linker has a
2089 different (but run-time compatible) architecture, these flags can be
2090 specified to build plugins that are compatible to the linker. For
2091 example, if you are building GCC for a 64-bit x86_64
2092 (@samp{x86_64-pc-linux-gnu}) host system, but have a 32-bit x86
2093 GNU/Linux (@samp{i686-pc-linux-gnu}) linker executable (which is
2094 executable on the former system), you can configure GCC as follows for
2095 getting compatible linker plugins:
2098 % @var{srcdir}/configure \
2099 --host=x86_64-pc-linux-gnu \
2100 --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \
2101 --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib'
2104 @item --with-plugin-ld=@var{pathname}
2105 Enable an alternate linker to be used at link-time optimization (LTO)
2106 link time when @option{-fuse-linker-plugin} is enabled.
2107 This linker should have plugin support such as gold starting with
2108 version 2.20 or GNU ld starting with version 2.21.
2109 See @option{-fuse-linker-plugin} for details.
2111 @item --enable-canonical-system-headers
2112 @itemx --disable-canonical-system-headers
2113 Enable system header path canonicalization for @file{libcpp}. This can
2114 produce shorter header file paths in diagnostics and dependency output
2115 files, but these changed header paths may conflict with some compilation
2116 environments. Enabled by default, and may be disabled using
2117 @option{--disable-canonical-system-headers}.
2119 @item --with-glibc-version=@var{major}.@var{minor}
2120 Tell GCC that when the GNU C Library (glibc) is used on the target it
2121 will be version @var{major}.@var{minor} or later. Normally this can
2122 be detected from the C library's header files, but this option may be
2123 needed when bootstrapping a cross toolchain without the header files
2124 available for building the initial bootstrap compiler.
2126 If GCC is configured with some multilibs that use glibc and some that
2127 do not, this option applies only to the multilibs that use glibc.
2128 However, such configurations may not work well as not all the relevant
2129 configuration in GCC is on a per-multilib basis.
2131 @item --enable-as-accelerator-for=@var{target}
2132 Build as offload target compiler. Specify offload host triple by @var{target}.
2134 @item --enable-offload-targets=@var{target1}[=@var{path1}],@dots{},@var{targetN}[=@var{pathN}]
2135 Enable offloading to targets @var{target1}, @dots{}, @var{targetN}.
2136 Offload compilers are expected to be already installed. Default search
2137 path for them is @file{@var{exec-prefix}}, but it can be changed by
2138 specifying paths @var{path1}, @dots{}, @var{pathN}.
2141 % @var{srcdir}/configure \
2142 --enable-offload-target=i686-unknown-linux-gnu=/path/to/i686/compiler,x86_64-pc-linux-gnu
2145 If @samp{hsa} is specified as one of the targets, the compiler will be
2146 built with support for HSA GPU accelerators. Because the same
2147 compiler will emit the accelerator code, no path should be specified.
2149 @item --with-hsa-runtime=@var{pathname}
2150 @itemx --with-hsa-runtime-include=@var{pathname}
2151 @itemx --with-hsa-runtime-lib=@var{pathname}
2153 If you configure GCC with HSA offloading but do not have the HSA
2154 run-time library installed in a standard location then you can
2155 explicitly specify the directory where they are installed. The
2156 @option{--with-hsa-runtime=@/@var{hsainstalldir}} option is a
2158 @option{--with-hsa-runtime-lib=@/@var{hsainstalldir}/lib} and
2159 @option{--with-hsa-runtime-include=@/@var{hsainstalldir}/include}.
2162 @itemx --disable-cet
2163 Enable building target run-time libraries with control-flow
2164 instrumentation, see @option{-fcf-protection} option. When
2165 @code{--enable-cet} is specified target libraries are configured
2166 to add @option{-fcf-protection} and, if needed, other target
2167 specific options to a set of building options.
2169 The option is disabled by default. When @code{--enable-cet=auto}
2170 is used, it is enabled on Linux/x86 if target binutils
2171 supports @code{Intel CET} instructions and disabled otherwise.
2172 In this case the target libraries are configured to get additional
2173 @option{-fcf-protection} option.
2175 @item --with-riscv-attribute=@samp{yes}, @samp{no} or @samp{default}
2176 Generate RISC-V attribute by default, in order to record extra build
2177 information in object.
2179 The option is disabled by default. It is enabled on RISC-V/ELF (bare-metal)
2180 target if target binutils supported.
2183 @subheading Cross-Compiler-Specific Options
2184 The following options only apply to building cross compilers.
2187 @item --with-sysroot
2188 @itemx --with-sysroot=@var{dir}
2189 Tells GCC to consider @var{dir} as the root of a tree that contains
2190 (a subset of) the root filesystem of the target operating system.
2191 Target system headers, libraries and run-time object files will be
2192 searched for in there. More specifically, this acts as if
2193 @option{--sysroot=@var{dir}} was added to the default options of the built
2194 compiler. The specified directory is not copied into the
2195 install tree, unlike the options @option{--with-headers} and
2196 @option{--with-libs} that this option obsoletes. The default value,
2197 in case @option{--with-sysroot} is not given an argument, is
2198 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
2199 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
2200 the GCC binaries if the installation tree is moved.
2202 This option affects the system root for the compiler used to build
2203 target libraries (which runs on the build system) and the compiler newly
2204 installed with @code{make install}; it does not affect the compiler which is
2205 used to build GCC itself.
2207 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2208 option then the compiler will search that directory within @var{dirname} for
2209 native system headers rather than the default @file{/usr/include}.
2211 @item --with-build-sysroot
2212 @itemx --with-build-sysroot=@var{dir}
2213 Tells GCC to consider @var{dir} as the system root (see
2214 @option{--with-sysroot}) while building target libraries, instead of
2215 the directory specified with @option{--with-sysroot}. This option is
2216 only useful when you are already using @option{--with-sysroot}. You
2217 can use @option{--with-build-sysroot} when you are configuring with
2218 @option{--prefix} set to a directory that is different from the one in
2219 which you are installing GCC and your target libraries.
2221 This option affects the system root for the compiler used to build
2222 target libraries (which runs on the build system); it does not affect
2223 the compiler which is used to build GCC itself.
2225 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2226 option then the compiler will search that directory within @var{dirname} for
2227 native system headers rather than the default @file{/usr/include}.
2229 @item --with-headers
2230 @itemx --with-headers=@var{dir}
2231 Deprecated in favor of @option{--with-sysroot}.
2232 Specifies that target headers are available when building a cross compiler.
2233 The @var{dir} argument specifies a directory which has the target include
2234 files. These include files will be copied into the @file{gcc} install
2235 directory. @emph{This option with the @var{dir} argument is required} when
2236 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
2237 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
2238 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
2239 will be run on these files to make them compatible with GCC@.
2241 @item --without-headers
2242 Tells GCC not use any target headers from a libc when building a cross
2243 compiler. When crossing to GNU/Linux, you need the headers so GCC
2244 can build the exception handling for libgcc.
2247 @itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}"
2248 Deprecated in favor of @option{--with-sysroot}.
2249 Specifies a list of directories which contain the target runtime
2250 libraries. These libraries will be copied into the @file{gcc} install
2251 directory. If the directory list is omitted, this option has no
2255 Specifies that @samp{newlib} is
2256 being used as the target C library. This causes @code{__eprintf} to be
2257 omitted from @file{libgcc.a} on the assumption that it will be provided by
2260 @item --with-avrlibc
2261 Specifies that @samp{AVR-Libc} is
2262 being used as the target C library. This causes float support
2263 functions like @code{__addsf3} to be omitted from @file{libgcc.a} on
2264 the assumption that it will be provided by @file{libm.a}. For more
2265 technical details, cf. @uref{http://gcc.gnu.org/PR54461,,PR54461}.
2266 This option is only supported for the AVR target. It is not supported for
2267 RTEMS configurations, which currently use newlib. The option is
2268 supported since version 4.7.2 and is the default in 4.8.0 and newer.
2270 @item --with-nds32-lib=@var{library}
2271 Specifies that @var{library} setting is used for building @file{libgcc.a}.
2272 Currently, the valid @var{library} is @samp{newlib} or @samp{mculib}.
2273 This option is only supported for the NDS32 target.
2275 @item --with-build-time-tools=@var{dir}
2276 Specifies where to find the set of target tools (assembler, linker, etc.)
2277 that will be used while building GCC itself. This option can be useful
2278 if the directory layouts are different between the system you are building
2279 GCC on, and the system where you will deploy it.
2281 For example, on an @samp{ia64-hp-hpux} system, you may have the GNU
2282 assembler and linker in @file{/usr/bin}, and the native tools in a
2283 different path, and build a toolchain that expects to find the
2284 native tools in @file{/usr/bin}.
2286 When you use this option, you should ensure that @var{dir} includes
2287 @command{ar}, @command{as}, @command{ld}, @command{nm},
2288 @command{ranlib} and @command{strip} if necessary, and possibly
2289 @command{objdump}. Otherwise, GCC may use an inconsistent set of
2293 @subsubheading Overriding @command{configure} test results
2295 Sometimes, it might be necessary to override the result of some
2296 @command{configure} test, for example in order to ease porting to a new
2297 system or work around a bug in a test. The toplevel @command{configure}
2298 script provides three variables for this:
2302 @item build_configargs
2303 @cindex @code{build_configargs}
2304 The contents of this variable is passed to all build @command{configure}
2307 @item host_configargs
2308 @cindex @code{host_configargs}
2309 The contents of this variable is passed to all host @command{configure}
2312 @item target_configargs
2313 @cindex @code{target_configargs}
2314 The contents of this variable is passed to all target @command{configure}
2319 In order to avoid shell and @command{make} quoting issues for complex
2320 overrides, you can pass a setting for @env{CONFIG_SITE} and set
2321 variables in the site file.
2323 @subheading Objective-C-Specific Options
2325 The following options apply to the build of the Objective-C runtime library.
2328 @item --enable-objc-gc
2329 Specify that an additional variant of the GNU Objective-C runtime library
2330 is built, using an external build of the Boehm-Demers-Weiser garbage
2331 collector (@uref{http://www.hboehm.info/gc/}). This library needs to be
2332 available for each multilib variant, unless configured with
2333 @option{--enable-objc-gc=@samp{auto}} in which case the build of the
2334 additional runtime library is skipped when not available and the build
2337 @item --with-target-bdw-gc=@var{list}
2338 @itemx --with-target-bdw-gc-include=@var{list}
2339 @itemx --with-target-bdw-gc-lib=@var{list}
2340 Specify search directories for the garbage collector header files and
2341 libraries. @var{list} is a comma separated list of key value pairs of the
2342 form @samp{@var{multilibdir}=@var{path}}, where the default multilib key
2343 is named as @samp{.} (dot), or is omitted (e.g.@:
2344 @samp{--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32}).
2346 The options @option{--with-target-bdw-gc-include} and
2347 @option{--with-target-bdw-gc-lib} must always be specified together
2348 for each multilib variant and they take precedence over
2349 @option{--with-target-bdw-gc}. If @option{--with-target-bdw-gc-include}
2350 is missing values for a multilib, then the value for the default
2351 multilib is used (e.g.@: @samp{--with-target-bdw-gc-include=/opt/bdw-gc/include}
2352 @samp{--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32}).
2353 If none of these options are specified, the library is assumed in
2357 @subheading D-Specific Options
2359 The following options apply to the build of the D runtime library.
2362 @item --with-target-system-zlib
2363 Use installed @samp{zlib} rather than that included with GCC@. This needs
2364 to be available for each multilib variant, unless configured with
2365 @option{--with-target-system-zlib=@samp{auto}} in which case the GCC@ included
2366 @samp{zlib} is only used when the system installed library is not available.
2374 @uref{./index.html,,Return to the GCC Installation page}
2378 @c ***Building****************************************************************
2380 @comment node-name, next, previous, up
2381 @node Building, Testing, Configuration, Installing GCC
2387 @cindex Installing GCC: Building
2389 Now that GCC is configured, you are ready to build the compiler and
2392 Some commands executed when making the compiler may fail (return a
2393 nonzero status) and be ignored by @command{make}. These failures, which
2394 are often due to files that were not found, are expected, and can safely
2397 It is normal to have compiler warnings when compiling certain files.
2398 Unless you are a GCC developer, you can generally ignore these warnings
2399 unless they cause compilation to fail. Developers should attempt to fix
2400 any warnings encountered, however they can temporarily continue past
2401 warnings-as-errors by specifying the configure flag
2402 @option{--disable-werror}.
2404 On certain old systems, defining certain environment variables such as
2405 @env{CC} can interfere with the functioning of @command{make}.
2407 If you encounter seemingly strange errors when trying to build the
2408 compiler in a directory other than the source directory, it could be
2409 because you have previously configured the compiler in the source
2410 directory. Make sure you have done all the necessary preparations.
2412 If you build GCC on a BSD system using a directory stored in an old System
2413 V file system, problems may occur in running @command{fixincludes} if the
2414 System V file system doesn't support symbolic links. These problems
2415 result in a failure to fix the declaration of @code{size_t} in
2416 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
2417 that type mismatches occur, this could be the cause.
2419 The solution is not to use such a directory for building GCC@.
2421 Similarly, when building from SVN or snapshots, or if you modify
2422 @file{*.l} files, you need the Flex lexical analyzer generator
2423 installed. If you do not modify @file{*.l} files, releases contain
2424 the Flex-generated files and you do not need Flex installed to build
2425 them. There is still one Flex-based lexical analyzer (part of the
2426 build machinery, not of GCC itself) that is used even if you only
2427 build the C front end.
2429 When building from SVN or snapshots, or if you modify Texinfo
2430 documentation, you need version 4.7 or later of Texinfo installed if you
2431 want Info documentation to be regenerated. Releases contain Info
2432 documentation pre-built for the unmodified documentation in the release.
2434 @section Building a native compiler
2436 For a native build, the default configuration is to perform
2437 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
2438 This will build the entire GCC system and ensure that it compiles
2439 itself correctly. It can be disabled with the @option{--disable-bootstrap}
2440 parameter to @samp{configure}, but bootstrapping is suggested because
2441 the compiler will be tested more completely and could also have
2444 The bootstrapping process will complete the following steps:
2448 Build tools necessary to build the compiler.
2451 Perform a 3-stage bootstrap of the compiler. This includes building
2452 three times the target tools for use by the compiler such as binutils
2453 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
2454 individually linked or moved into the top level GCC source tree before
2458 Perform a comparison test of the stage2 and stage3 compilers.
2461 Build runtime libraries using the stage3 compiler from the previous step.
2465 If you are short on disk space you might consider @samp{make
2466 bootstrap-lean} instead. The sequence of compilation is the
2467 same described above, but object files from the stage1 and
2468 stage2 of the 3-stage bootstrap of the compiler are deleted as
2469 soon as they are no longer needed.
2471 If you wish to use non-default GCC flags when compiling the stage2
2472 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
2473 doing @samp{make}. For example, if you want to save additional space
2474 during the bootstrap and in the final installation as well, you can
2475 build the compiler binaries without debugging information as in the
2476 following example. This will save roughly 40% of disk space both for
2477 the bootstrap and the final installation. (Libraries will still contain
2478 debugging information.)
2481 make BOOT_CFLAGS='-O' bootstrap
2484 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
2485 are less well tested here than the default of @samp{-g -O2}, but should
2486 still work. In a few cases, you may find that you need to specify special
2487 flags such as @option{-msoft-float} here to complete the bootstrap; or,
2488 if the native compiler miscompiles the stage1 compiler, you may need
2489 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
2490 of the stage1 compiler that were miscompiled, or by using @samp{make
2491 bootstrap4} to increase the number of stages of bootstrap.
2493 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
2494 Since these are always compiled with the compiler currently being
2495 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
2496 compilation flags, as for non-bootstrapped target libraries.
2497 Again, if the native compiler miscompiles the stage1 compiler, you may
2498 need to work around this by avoiding non-working parts of the stage1
2499 compiler. Use @code{STAGE1_TFLAGS} to this end.
2501 If you used the flag @option{--enable-languages=@dots{}} to restrict
2502 the compilers to be built, only those you've actually enabled will be
2503 built. This will of course only build those runtime libraries, for
2504 which the particular compiler has been built. Please note,
2505 that re-defining @env{LANGUAGES} when calling @samp{make}
2506 @strong{does not} work anymore!
2508 If the comparison of stage2 and stage3 fails, this normally indicates
2509 that the stage2 compiler has compiled GCC incorrectly, and is therefore
2510 a potentially serious bug which you should investigate and report. (On
2511 a few systems, meaningful comparison of object files is impossible; they
2512 always appear ``different''. If you encounter this problem, you will
2513 need to disable comparison in the @file{Makefile}.)
2515 If you do not want to bootstrap your compiler, you can configure with
2516 @option{--disable-bootstrap}. In particular cases, you may want to
2517 bootstrap your compiler even if the target system is not the same as
2518 the one you are building on: for example, you could build a
2519 @code{powerpc-unknown-linux-gnu} toolchain on a
2520 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
2521 @option{--enable-bootstrap} to the configure script.
2523 @code{BUILD_CONFIG} can be used to bring in additional customization
2524 to the build. It can be set to a whitespace-separated list of names.
2525 For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will
2526 be included by the top-level @file{Makefile}, bringing in any settings
2527 it contains. The default @code{BUILD_CONFIG} can be set using the
2528 configure option @option{--with-build-config=@code{NAME}...}. Some
2529 examples of supported build configurations are:
2532 @item @samp{bootstrap-O1}
2533 Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds
2534 @option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to
2535 @samp{BOOT_CFLAGS='-g -O1'}.
2537 @item @samp{bootstrap-O3}
2538 Analogous to @code{bootstrap-O1}.
2540 @item @samp{bootstrap-lto}
2541 Enables Link-Time Optimization for host tools during bootstrapping.
2542 @samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding
2543 @option{-flto} to @samp{BOOT_CFLAGS}. This option assumes that the host
2544 supports the linker plugin (e.g.@: GNU ld version 2.21 or later or GNU gold
2545 version 2.21 or later).
2547 @item @samp{bootstrap-lto-noplugin}
2548 This option is similar to @code{bootstrap-lto}, but is intended for
2549 hosts that do not support the linker plugin. Without the linker plugin
2550 static libraries are not compiled with link-time optimizations. Since
2551 the GCC middle end and back end are in @file{libbackend.a} this means
2552 that only the front end is actually LTO optimized.
2554 @item @samp{bootstrap-lto-lean}
2555 This option is similar to @code{bootstrap-lto}, but is intended for
2556 faster build by only using LTO in the final bootstrap stage.
2557 With @samp{make profiledbootstrap} the LTO frontend
2558 is trained only on generator files.
2560 @item @samp{bootstrap-debug}
2561 Verifies that the compiler generates the same executable code, whether
2562 or not it is asked to emit debug information. To this end, this
2563 option builds stage2 host programs without debug information, and uses
2564 @file{contrib/compare-debug} to compare them with the stripped stage3
2565 object files. If @code{BOOT_CFLAGS} is overridden so as to not enable
2566 debug information, stage2 will have it, and stage3 won't. This option
2567 is enabled by default when GCC bootstrapping is enabled, if
2568 @code{strip} can turn object files compiled with and without debug
2569 info into identical object files. In addition to better test
2570 coverage, this option makes default bootstraps faster and leaner.
2572 @item @samp{bootstrap-debug-big}
2573 Rather than comparing stripped object files, as in
2574 @code{bootstrap-debug}, this option saves internal compiler dumps
2575 during stage2 and stage3 and compares them as well, which helps catch
2576 additional potential problems, but at a great cost in terms of disk
2577 space. It can be specified in addition to @samp{bootstrap-debug}.
2579 @item @samp{bootstrap-debug-lean}
2580 This option saves disk space compared with @code{bootstrap-debug-big},
2581 but at the expense of some recompilation. Instead of saving the dumps
2582 of stage2 and stage3 until the final compare, it uses
2583 @option{-fcompare-debug} to generate, compare and remove the dumps
2584 during stage3, repeating the compilation that already took place in
2585 stage2, whose dumps were not saved.
2587 @item @samp{bootstrap-debug-lib}
2588 This option tests executable code invariance over debug information
2589 generation on target libraries, just like @code{bootstrap-debug-lean}
2590 tests it on host programs. It builds stage3 libraries with
2591 @option{-fcompare-debug}, and it can be used along with any of the
2592 @code{bootstrap-debug} options above.
2594 There aren't @code{-lean} or @code{-big} counterparts to this option
2595 because most libraries are only build in stage3, so bootstrap compares
2596 would not get significant coverage. Moreover, the few libraries built
2597 in stage2 are used in stage3 host programs, so we wouldn't want to
2598 compile stage2 libraries with different options for comparison purposes.
2600 @item @samp{bootstrap-debug-ckovw}
2601 Arranges for error messages to be issued if the compiler built on any
2602 stage is run without the option @option{-fcompare-debug}. This is
2603 useful to verify the full @option{-fcompare-debug} testing coverage. It
2604 must be used along with @code{bootstrap-debug-lean} and
2605 @code{bootstrap-debug-lib}.
2607 @item @samp{bootstrap-cet}
2608 This option enables Intel CET for host tools during bootstrapping.
2609 @samp{BUILD_CONFIG=bootstrap-cet} is equivalent to adding
2610 @option{-fcf-protection} to @samp{BOOT_CFLAGS}. This option
2611 assumes that the host supports Intel CET (e.g.@: GNU assembler version
2614 @item @samp{bootstrap-time}
2615 Arranges for the run time of each program started by the GCC driver,
2616 built in any stage, to be logged to @file{time.log}, in the top level of
2621 @section Building a cross compiler
2623 When building a cross compiler, it is not generally possible to do a
2624 3-stage bootstrap of the compiler. This makes for an interesting problem
2625 as parts of GCC can only be built with GCC@.
2627 To build a cross compiler, we recommend first building and installing a
2628 native compiler. You can then use the native GCC compiler to build the
2629 cross compiler. The installed native compiler needs to be GCC version
2632 Assuming you have already installed a native copy of GCC and configured
2633 your cross compiler, issue the command @command{make}, which performs the
2638 Build host tools necessary to build the compiler.
2641 Build target tools for use by the compiler such as binutils (bfd,
2642 binutils, gas, gprof, ld, and opcodes)
2643 if they have been individually linked or moved into the top level GCC source
2644 tree before configuring.
2647 Build the compiler (single stage only).
2650 Build runtime libraries using the compiler from the previous step.
2653 Note that if an error occurs in any step the make process will exit.
2655 If you are not building GNU binutils in the same source tree as GCC,
2656 you will need a cross-assembler and cross-linker installed before
2657 configuring GCC@. Put them in the directory
2658 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
2659 you should put in this directory:
2663 This should be the cross-assembler.
2666 This should be the cross-linker.
2669 This should be the cross-archiver: a program which can manipulate
2670 archive files (linker libraries) in the target machine's format.
2673 This should be a program to construct a symbol table in an archive file.
2676 The installation of GCC will find these programs in that directory,
2677 and copy or link them to the proper place to for the cross-compiler to
2678 find them when run later.
2680 The easiest way to provide these files is to build the Binutils package.
2681 Configure it with the same @option{--host} and @option{--target}
2682 options that you use for configuring GCC, then build and install
2683 them. They install their executables automatically into the proper
2684 directory. Alas, they do not support all the targets that GCC
2687 If you are not building a C library in the same source tree as GCC,
2688 you should also provide the target libraries and headers before
2689 configuring GCC, specifying the directories with
2690 @option{--with-sysroot} or @option{--with-headers} and
2691 @option{--with-libs}. Many targets also require ``start files'' such
2692 as @file{crt0.o} and
2693 @file{crtn.o} which are linked into each executable. There may be several
2694 alternatives for @file{crt0.o}, for use with profiling or other
2695 compilation options. Check your target's definition of
2696 @code{STARTFILE_SPEC} to find out what start files it uses.
2698 @section Building in parallel
2700 GNU Make 3.80 and above, which is necessary to build GCC, support
2701 building in parallel. To activate this, you can use @samp{make -j 2}
2702 instead of @samp{make}. You can also specify a bigger number, and
2703 in most cases using a value greater than the number of processors in
2704 your machine will result in fewer and shorter I/O latency hits, thus
2705 improving overall throughput; this is especially true for slow drives
2706 and network filesystems.
2708 @section Building the Ada compiler
2710 In order to build GNAT, the Ada compiler, you need a working GNAT
2711 compiler (GCC version 4.0 or later).
2712 This includes GNAT tools such as @command{gnatmake} and
2713 @command{gnatlink}, since the Ada front end is written in Ada and
2714 uses some GNAT-specific extensions.
2716 In order to build a cross compiler, it is suggested to install
2717 the new compiler as native first, and then use it to build the cross
2720 @command{configure} does not test whether the GNAT installation works
2721 and has a sufficiently recent version; if too old a GNAT version is
2722 installed, the build will fail unless @option{--enable-languages} is
2723 used to disable building the Ada front end.
2725 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
2726 must not be set when building the Ada compiler, the Ada tools, or the
2727 Ada runtime libraries. You can check that your build environment is clean
2728 by verifying that @samp{gnatls -v} lists only one explicit path in each
2731 @section Building with profile feedback
2733 It is possible to use profile feedback to optimize the compiler itself. This
2734 should result in a faster compiler binary. Experiments done on x86 using gcc
2735 3.3 showed approximately 7 percent speedup on compiling C programs. To
2736 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
2738 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
2739 compiler. This compiler is used to build a @code{stageprofile} compiler
2740 instrumented to collect execution counts of instruction and branch
2741 probabilities. Training run is done by building @code{stagetrain}
2742 compiler. Finally a @code{stagefeedback} compiler is built
2743 using the information collected.
2745 Unlike standard bootstrap, several additional restrictions apply. The
2746 compiler used to build @code{stage1} needs to support a 64-bit integral type.
2747 It is recommended to only use GCC for this.
2749 On Linux/x86_64 hosts with some restrictions (no virtualization) it is
2750 also possible to do autofdo build with @samp{make
2751 autoprofiledback}. This uses Linux perf to sample branches in the
2752 binary and then rebuild it with feedback derived from the profile.
2753 Linux perf and the @code{autofdo} toolkit needs to be installed for
2756 Only the profile from the current build is used, so when an error
2757 occurs it is recommended to clean before restarting. Otherwise
2758 the code quality may be much worse.
2765 @uref{./index.html,,Return to the GCC Installation page}
2769 @c ***Testing*****************************************************************
2771 @comment node-name, next, previous, up
2772 @node Testing, Final install, Building, Installing GCC
2776 @chapter Installing GCC: Testing
2779 @cindex Installing GCC: Testing
2782 Before you install GCC, we encourage you to run the testsuites and to
2783 compare your results with results from a similar configuration that have
2784 been submitted to the
2785 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
2786 Some of these archived results are linked from the build status lists
2787 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
2788 reports a successful build runs the testsuites and submits the results.
2789 This step is optional and may require you to download additional software,
2790 but it can give you confidence in your new GCC installation or point out
2791 problems before you install and start using your new GCC@.
2793 First, you must have @uref{download.html,,downloaded the testsuites}.
2794 These are part of the full distribution, but if you downloaded the
2795 ``core'' compiler plus any front ends, you must download the testsuites
2798 Second, you must have the testing tools installed. This includes
2799 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
2800 the DejaGnu site has links to these. For running the BRIG frontend
2801 tests, a tool to assemble the binary BRIGs from HSAIL text,
2802 @uref{https://github.com/HSAFoundation/HSAIL-Tools/,,HSAILasm} must
2805 If the directories where @command{runtest} and @command{expect} were
2806 installed are not in the @env{PATH}, you may need to set the following
2807 environment variables appropriately, as in the following example (which
2808 assumes that DejaGnu has been installed under @file{/usr/local}):
2811 TCL_LIBRARY = /usr/local/share/tcl8.0
2812 DEJAGNULIBS = /usr/local/share/dejagnu
2815 (On systems such as Cygwin, these paths are required to be actual
2816 paths, not mounts or links; presumably this is due to some lack of
2817 portability in the DejaGnu code.)
2820 Finally, you can run the testsuite (which may take a long time):
2822 cd @var{objdir}; make -k check
2825 This will test various components of GCC, such as compiler
2826 front ends and runtime libraries. While running the testsuite, DejaGnu
2827 might emit some harmless messages resembling
2828 @samp{WARNING: Couldn't find the global config file.} or
2829 @samp{WARNING: Couldn't find tool init file} that can be ignored.
2831 If you are testing a cross-compiler, you may want to run the testsuite
2832 on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}.
2834 @section How can you run the testsuite on selected tests?
2836 In order to run sets of tests selectively, there are targets
2837 @samp{make check-gcc} and language specific @samp{make check-c},
2838 @samp{make check-c++}, @samp{make check-d} @samp{make check-fortran},
2839 @samp{make check-ada}, @samp{make check-objc}, @samp{make check-obj-c++},
2840 @samp{make check-lto}
2841 in the @file{gcc} subdirectory of the object directory. You can also
2842 just run @samp{make check} in a subdirectory of the object directory.
2845 A more selective way to just run all @command{gcc} execute tests in the
2849 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
2852 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
2853 the testsuite with filenames matching @samp{9805*}, you would use
2856 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
2859 The file-matching expression following @var{filename}@command{.exp=} is treated
2860 as a series of whitespace-delimited glob expressions so that multiple patterns
2861 may be passed, although any whitespace must either be escaped or surrounded by
2862 single quotes if multiple expressions are desired. For example,
2865 make check-g++ RUNTESTFLAGS="old-deja.exp=9805*\ virtual2.c @var{other-options}"
2866 make check-g++ RUNTESTFLAGS="'old-deja.exp=9805* virtual2.c' @var{other-options}"
2869 The @file{*.exp} files are located in the testsuite directories of the GCC
2870 source, the most important ones being @file{compile.exp},
2871 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
2872 To get a list of the possible @file{*.exp} files, pipe the
2873 output of @samp{make check} into a file and look at the
2874 @samp{Running @dots{} .exp} lines.
2876 @section Passing options and running multiple testsuites
2878 You can pass multiple options to the testsuite using the
2879 @samp{--target_board} option of DejaGNU, either passed as part of
2880 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
2881 work outside the makefiles. For example,
2884 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
2887 will run the standard @command{g++} testsuites (``unix'' is the target name
2888 for a standard native testsuite situation), passing
2889 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
2890 slashes separate options.
2892 You can run the testsuites multiple times using combinations of options
2893 with a syntax similar to the brace expansion of popular shells:
2896 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
2899 (Note the empty option caused by the trailing comma in the final group.)
2900 The following will run each testsuite eight times using the @samp{arm-sim}
2901 target, as if you had specified all possible combinations yourself:
2904 --target_board='arm-sim/-mhard-float/-O1 \
2905 arm-sim/-mhard-float/-O2 \
2906 arm-sim/-mhard-float/-O3 \
2907 arm-sim/-mhard-float \
2908 arm-sim/-msoft-float/-O1 \
2909 arm-sim/-msoft-float/-O2 \
2910 arm-sim/-msoft-float/-O3 \
2911 arm-sim/-msoft-float'
2914 They can be combined as many times as you wish, in arbitrary ways. This
2918 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
2921 will generate four combinations, all involving @samp{-Wextra}.
2923 The disadvantage to this method is that the testsuites are run in serial,
2924 which is a waste on multiprocessor systems. For users with GNU Make and
2925 a shell which performs brace expansion, you can run the testsuites in
2926 parallel by having the shell perform the combinations and @command{make}
2927 do the parallel runs. Instead of using @samp{--target_board}, use a
2928 special makefile target:
2931 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
2937 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
2940 will run three concurrent ``make-gcc'' testsuites, eventually testing all
2941 ten combinations as described above. Note that this is currently only
2942 supported in the @file{gcc} subdirectory. (To see how this works, try
2943 typing @command{echo} before the example given here.)
2946 @section How to interpret test results
2948 The result of running the testsuite are various @file{*.sum} and @file{*.log}
2949 files in the testsuite subdirectories. The @file{*.log} files contain a
2950 detailed log of the compiler invocations and the corresponding
2951 results, the @file{*.sum} files summarize the results. These summaries
2952 contain status codes for all tests:
2956 PASS: the test passed as expected
2958 XPASS: the test unexpectedly passed
2960 FAIL: the test unexpectedly failed
2962 XFAIL: the test failed as expected
2964 UNSUPPORTED: the test is not supported on this platform
2966 ERROR: the testsuite detected an error
2968 WARNING: the testsuite detected a possible problem
2971 It is normal for some tests to report unexpected failures. At the
2972 current time the testing harness does not allow fine grained control
2973 over whether or not a test is expected to fail. This problem should
2974 be fixed in future releases.
2977 @section Submitting test results
2979 If you want to report the results to the GCC project, use the
2980 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
2983 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
2984 -m gcc-testresults@@gcc.gnu.org |sh
2987 This script uses the @command{Mail} program to send the results, so
2988 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
2989 prepended to the testsuite summary and should contain any special
2990 remarks you have on your results or your build environment. Please
2991 do not edit the testsuite result block or the subject line, as these
2992 messages may be automatically processed.
2999 @uref{./index.html,,Return to the GCC Installation page}
3003 @c ***Final install***********************************************************
3005 @comment node-name, next, previous, up
3006 @node Final install, , Testing, Installing GCC
3008 @ifset finalinstallhtml
3010 @chapter Installing GCC: Final installation
3013 Now that GCC has been built (and optionally tested), you can install it with
3015 cd @var{objdir} && make install
3018 We strongly recommend to install into a target directory where there is
3019 no previous version of GCC present. Also, the GNAT runtime should not
3020 be stripped, as this would break certain features of the debugger that
3021 depend on this debugging information (catching Ada exceptions for
3024 That step completes the installation of GCC; user level binaries can
3025 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
3026 you specified with the @option{--prefix} to configure (or
3027 @file{/usr/local} by default). (If you specified @option{--bindir},
3028 that directory will be used instead; otherwise, if you specified
3029 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
3030 Headers for the C++ library are installed in
3031 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
3032 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
3033 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
3034 in info format in @file{@var{infodir}} (normally
3035 @file{@var{prefix}/info}).
3037 When installing cross-compilers, GCC's executables
3038 are not only installed into @file{@var{bindir}}, that
3039 is, @file{@var{exec-prefix}/bin}, but additionally into
3040 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
3041 exists. Typically, such @dfn{tooldirs} hold target-specific
3042 binutils, including assembler and linker.
3044 Installation into a temporary staging area or into a @command{chroot}
3045 jail can be achieved with the command
3048 make DESTDIR=@var{path-to-rootdir} install
3052 where @var{path-to-rootdir} is the absolute path of
3053 a directory relative to which all installation paths will be
3054 interpreted. Note that the directory specified by @code{DESTDIR}
3055 need not exist yet; it will be created if necessary.
3057 There is a subtle point with tooldirs and @code{DESTDIR}:
3058 If you relocate a cross-compiler installation with
3059 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
3060 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
3061 be filled with duplicated GCC executables only if it already exists,
3062 it will not be created otherwise. This is regarded as a feature,
3063 not as a bug, because it gives slightly more control to the packagers
3064 using the @code{DESTDIR} feature.
3066 You can install stripped programs and libraries with
3072 If you are bootstrapping a released version of GCC then please
3073 quickly review the build status page for your release, available from
3074 @uref{http://gcc.gnu.org/buildstat.html}.
3075 If your system is not listed for the version of GCC that you built,
3077 @email{gcc@@gcc.gnu.org} indicating
3078 that you successfully built and installed GCC@.
3079 Include the following information:
3083 Output from running @file{@var{srcdir}/config.guess}. Do not send
3084 that file itself, just the one-line output from running it.
3087 The output of @samp{gcc -v} for your newly installed @command{gcc}.
3088 This tells us which version of GCC you built and the options you passed to
3092 Whether you enabled all languages or a subset of them. If you used a
3093 full distribution then this information is part of the configure
3094 options in the output of @samp{gcc -v}, but if you downloaded the
3095 ``core'' compiler plus additional front ends then it isn't apparent
3096 which ones you built unless you tell us about it.
3099 If the build was for GNU/Linux, also include:
3102 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
3103 this information should be available from @file{/etc/issue}.
3106 The version of the Linux kernel, available from @samp{uname --version}
3110 The version of glibc you used; for RPM-based systems like Red Hat,
3111 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
3112 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
3114 For other systems, you can include similar information if you think it is
3118 Any other information that you think would be useful to people building
3119 GCC on the same configuration. The new entry in the build status list
3120 will include a link to the archived copy of your message.
3123 We'd also like to know if the
3125 @ref{Specific, host/target specific installation notes}
3128 @uref{specific.html,,host/target specific installation notes}
3130 didn't include your host/target information or if that information is
3131 incomplete or out of date. Send a note to
3132 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
3134 If you find a bug, please report it following the
3135 @uref{../bugs/,,bug reporting guidelines}.
3137 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
3138 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
3139 and @TeX{} installed. This creates a number of @file{.dvi} files in
3140 subdirectories of @file{@var{objdir}}; these may be converted for
3141 printing with programs such as @command{dvips}. Alternately, by using
3142 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
3143 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
3144 is included with Texinfo version 4.8 and later. You can also
3145 @uref{https://shop.fsf.org/,,buy printed manuals from the
3146 Free Software Foundation}, though such manuals may not be for the most
3147 recent version of GCC@.
3149 If you would like to generate online HTML documentation, do @samp{cd
3150 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
3151 @file{@var{objdir}/gcc/HTML}.
3158 @uref{./index.html,,Return to the GCC Installation page}
3162 @c ***Binaries****************************************************************
3164 @comment node-name, next, previous, up
3165 @node Binaries, Specific, Installing GCC, Top
3169 @chapter Installing GCC: Binaries
3172 @cindex Installing GCC: Binaries
3174 We are often asked about pre-compiled versions of GCC@. While we cannot
3175 provide these for all platforms, below you'll find links to binaries for
3176 various platforms where creating them by yourself is not easy due to various
3179 Please note that we did not create these binaries, nor do we
3180 support them. If you have any problems installing them, please
3181 contact their makers.
3188 @uref{http://www.bullfreeware.com,,Bull's Open Source Software Archive for
3189 for AIX 5L and AIX 6};
3192 @uref{http://www.perzl.org/aix/,,AIX Open Source Packages (AIX5L AIX 6.1
3197 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
3203 @uref{http://hpux.connect.org.uk/,,HP-UX Porting Center};
3207 Solaris 2 (SPARC, Intel):
3210 @uref{https://www.opencsw.org/,,OpenCSW}
3213 @uref{http://jupiterrise.com/tgcware/,,TGCware}
3220 The @uref{https://brew.sh,,Homebrew} package manager;
3222 @uref{https://www.macports.org,,MacPorts}.
3229 The @uref{https://sourceware.org/cygwin/,,Cygwin} project;
3231 The @uref{http://www.mingw.org/,,MinGW} and
3232 @uref{http://mingw-w64.org/doku.php,,mingw-w64} projects.
3236 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
3237 number of platforms.
3240 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
3241 links to GNU Fortran binaries for several platforms.
3249 @uref{./index.html,,Return to the GCC Installation page}
3253 @c ***Specific****************************************************************
3255 @comment node-name, next, previous, up
3256 @node Specific, Old, Binaries, Top
3260 @chapter Host/target specific installation notes for GCC
3263 @cindex Specific installation notes
3264 @cindex Target specific installation
3265 @cindex Host specific installation
3266 @cindex Target specific installation notes
3268 Please read this document carefully @emph{before} installing the
3269 GNU Compiler Collection on your machine.
3271 Note that this list of install notes is @emph{not} a list of supported
3272 hosts or targets. Not all supported hosts and targets are listed
3273 here, only the ones that require host-specific or target-specific
3274 information have to.
3279 @uref{#aarch64-x-x,,aarch64*-*-*}
3281 @uref{#alpha-x-x,,alpha*-*-*}
3283 @uref{#amd64-x-solaris2,,amd64-*-solaris2*}
3285 @uref{#arm-x-eabi,,arm-*-eabi}
3289 @uref{#bfin,,Blackfin}
3293 @uref{#x-x-freebsd,,*-*-freebsd*}
3295 @uref{#h8300-hms,,h8300-hms}
3297 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
3299 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
3301 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
3303 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
3305 @uref{#ix86-x-linux,,i?86-*-linux*}
3307 @uref{#ix86-x-solaris2,,i?86-*-solaris2*}
3309 @uref{#ia64-x-linux,,ia64-*-linux}
3311 @uref{#ia64-x-hpux,,ia64-*-hpux*}
3313 @uref{#x-ibm-aix,,*-ibm-aix*}
3315 @uref{#iq2000-x-elf,,iq2000-*-elf}
3317 @uref{#lm32-x-elf,,lm32-*-elf}
3319 @uref{#lm32-x-uclinux,,lm32-*-uclinux}
3321 @uref{#m32c-x-elf,,m32c-*-elf}
3323 @uref{#m32r-x-elf,,m32r-*-elf}
3325 @uref{#m68k-x-x,,m68k-*-*}
3327 @uref{#m68k-uclinux,,m68k-uclinux}
3329 @uref{#microblaze-x-elf,,microblaze-*-elf}
3331 @uref{#mips-x-x,,mips-*-*}
3333 @uref{#nds32le-x-elf,,nds32le-*-elf}
3335 @uref{#nds32be-x-elf,,nds32be-*-elf}
3337 @uref{#nvptx-x-none,,nvptx-*-none}
3339 @uref{#or1k-x-elf,,or1k-*-elf}
3341 @uref{#or1k-x-linux,,or1k-*-linux}
3343 @uref{#powerpc-x-x,,powerpc*-*-*}
3345 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
3347 @uref{#powerpc-x-elf,,powerpc-*-elf}
3349 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
3351 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
3353 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
3355 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
3357 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
3359 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
3361 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
3363 @uref{#riscv32-x-elf,,riscv32-*-elf}
3365 @uref{#riscv32-x-linux,,riscv32-*-linux}
3367 @uref{#riscv64-x-elf,,riscv64-*-elf}
3369 @uref{#riscv64-x-linux,,riscv64-*-linux}
3371 @uref{#s390-x-linux,,s390-*-linux*}
3373 @uref{#s390x-x-linux,,s390x-*-linux*}
3375 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
3377 @uref{#x-x-solaris2,,*-*-solaris2*}
3379 @uref{#sparc-x-x,,sparc*-*-*}
3381 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
3383 @uref{#sparc-x-linux,,sparc-*-linux*}
3385 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
3387 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
3389 @uref{#c6x-x-x,,c6x-*-*}
3391 @uref{#tilegx-x-linux,,tilegx-*-linux*}
3393 @uref{#tilegxbe-x-linux,,tilegxbe-*-linux*}
3395 @uref{#tilepro-x-linux,,tilepro-*-linux*}
3397 @uref{#visium-x-elf, visium-*-elf}
3399 @uref{#x-x-vxworks,,*-*-vxworks*}
3401 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
3403 @uref{#x86-64-x-solaris2,,x86_64-*-solaris2*}
3405 @uref{#xtensa-x-elf,,xtensa*-*-elf}
3407 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
3409 @uref{#windows,,Microsoft Windows}
3411 @uref{#x-x-cygwin,,*-*-cygwin}
3413 @uref{#x-x-mingw32,,*-*-mingw32}
3417 @uref{#older,,Older systems}
3422 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
3428 <!-- -------- host/target specific issues start here ---------------- -->
3431 @anchor{aarch64-x-x}
3432 @heading aarch64*-*-*
3433 Binutils pre 2.24 does not have support for selecting @option{-mabi} and
3434 does not support ILP32. If it is used to build GCC 4.9 or later, GCC will
3435 not support option @option{-mabi=ilp32}.
3437 To enable a workaround for the Cortex-A53 erratum number 835769 by default
3438 (for all CPUs regardless of -mcpu option given) at configure time use the
3439 @option{--enable-fix-cortex-a53-835769} option. This will enable the fix by
3440 default and can be explicitly disabled during compilation by passing the
3441 @option{-mno-fix-cortex-a53-835769} option. Conversely,
3442 @option{--disable-fix-cortex-a53-835769} will disable the workaround by
3443 default. The workaround is disabled by default if neither of
3444 @option{--enable-fix-cortex-a53-835769} or
3445 @option{--disable-fix-cortex-a53-835769} is given at configure time.
3447 To enable a workaround for the Cortex-A53 erratum number 843419 by default
3448 (for all CPUs regardless of -mcpu option given) at configure time use the
3449 @option{--enable-fix-cortex-a53-843419} option. This workaround is applied at
3450 link time. Enabling the workaround will cause GCC to pass the relevant option
3451 to the linker. It can be explicitly disabled during compilation by passing the
3452 @option{-mno-fix-cortex-a53-843419} option. Conversely,
3453 @option{--disable-fix-cortex-a53-843419} will disable the workaround by default.
3454 The workaround is disabled by default if neither of
3455 @option{--enable-fix-cortex-a53-843419} or
3456 @option{--disable-fix-cortex-a53-843419} is given at configure time.
3458 To enable Branch Target Identification Mechanism and Return Address Signing by
3459 default at configure time use the @option{--enable-standard-branch-protection}
3460 option. This is equivalent to having @option{-mbranch-protection=standard}
3461 during compilation. This can be explicitly disabled during compilation by
3462 passing the @option{-mbranch-protection=none} option which turns off all
3463 types of branch protections. Conversely,
3464 @option{--disable-standard-branch-protection} will disable both the
3465 protections by default. This mechanism is turned off by default if neither
3466 of the options are given at configure time.
3473 This section contains general configuration information for all
3474 Alpha-based platforms using ELF@. In addition to reading this
3475 section, please read all other sections that match your target.
3477 We require binutils 2.11.2 or newer.
3478 Previous binutils releases had a number of problems with DWARF 2
3479 debugging information, not the least of which is incorrect linking of
3485 @anchor{amd64-x-solaris2}
3486 @heading amd64-*-solaris2*
3487 This is a synonym for @samp{x86_64-*-solaris2*}.
3492 @anchor{amdgcn-unknown-amdhsa}
3493 @heading amdgcn-unknown-amdhsa
3496 Instead of GNU Binutils, you will need to install LLVM 6, or later, and copy
3497 @file{bin/llvm-mc} to @file{amdgcn-unknown-amdhsa/bin/as},
3498 @file{bin/lld} to @file{amdgcn-unknown-amdhsa/bin/ld},
3499 @file{bin/llvm-nm} to @file{amdgcn-unknown-amdhsa/bin/nm}, and
3500 @file{bin/llvm-ar} to both @file{bin/amdgcn-unknown-amdhsa-ar} and
3501 @file{bin/amdgcn-unknown-amdhsa-ranlib}.
3503 Use Newlib (2019-01-16, or newer).
3505 To run the binaries, install the HSA Runtime from the
3506 @uref{https://rocm.github.io,,ROCm Platform}, and use
3507 @file{libexec/gcc/amdhsa-unknown-amdhsa/@var{version}/gcn-run} to launch them
3513 @anchor{arc-x-elf32}
3514 @heading arc-*-elf32
3516 Use @samp{configure --target=arc-elf32 --with-cpu=@var{cpu} --enable-languages="c,c++"}
3517 to configure GCC, with @var{cpu} being one of @samp{arc600}, @samp{arc601},
3523 @anchor{arc-linux-uclibc}
3524 @heading arc-linux-uclibc
3526 Use @samp{configure --target=arc-linux-uclibc --with-cpu=arc700 --enable-languages="c,c++"} to configure GCC@.
3533 ARM-family processors.
3535 Building the Ada frontend commonly fails (an infinite loop executing
3536 @code{xsinfo}) if the host compiler is GNAT 4.8. Host compilers built from the
3537 GNAT 4.6, 4.9 or 5 release branches are known to succeed.
3544 ATMEL AVR-family micro controllers. These are used in embedded
3545 applications. There are no standard Unix configurations.
3547 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
3551 See ``AVR Options'' in the main manual
3553 for the list of supported MCU types.
3555 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
3557 Further installation notes and other useful information about AVR tools
3558 can also be obtained from:
3562 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
3564 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
3567 The following error:
3569 Error: register required
3572 indicates that you should upgrade to a newer version of the binutils.
3579 The Blackfin processor, an Analog Devices DSP.
3581 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
3585 See ``Blackfin Options'' in the main manual
3588 More information, and a version of binutils with support for this processor,
3589 is available at @uref{https://blackfin.uclinux.org}
3596 The CR16 CompactRISC architecture is a 16-bit architecture. This
3597 architecture is used in embedded applications.
3600 @xref{CR16 Options,, CR16 Options, gcc, Using and Porting the GNU Compiler
3605 See ``CR16 Options'' in the main manual for a list of CR16-specific options.
3608 Use @samp{configure --target=cr16-elf --enable-languages=c,c++} to configure
3609 GCC@ for building a CR16 elf cross-compiler.
3611 Use @samp{configure --target=cr16-uclinux --enable-languages=c,c++} to
3612 configure GCC@ for building a CR16 uclinux cross-compiler.
3619 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
3620 series. These are used in embedded applications.
3623 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
3627 See ``CRIS Options'' in the main manual
3629 for a list of CRIS-specific options.
3631 There are a few different CRIS targets:
3634 Mainly for monolithic embedded systems. Includes a multilib for the
3635 @samp{v10} core used in @samp{ETRAX 100 LX}.
3636 @item cris-axis-linux-gnu
3637 A GNU/Linux port for the CRIS architecture, currently targeting
3638 @samp{ETRAX 100 LX} by default.
3641 Pre-packaged tools can be obtained from
3642 @uref{ftp://ftp.axis.com/@/pub/@/axis/@/tools/@/cris/@/compiler-kit/}. More
3643 information about this platform is available at
3644 @uref{http://developer.axis.com/}.
3651 Please have a look at the @uref{binaries.html,,binaries page}.
3653 You cannot install GCC by itself on MSDOS; it will not compile under
3654 any MSDOS compiler except itself. You need to get the complete
3655 compilation package DJGPP, which includes binaries as well as sources,
3656 and includes all the necessary compilation tools and libraries.
3661 @anchor{epiphany-x-elf}
3662 @heading epiphany-*-elf
3664 This configuration is intended for embedded systems.
3669 @anchor{x-x-freebsd}
3670 @heading *-*-freebsd*
3671 Support for FreeBSD 1 was discontinued in GCC 3.2. Support for
3672 FreeBSD 2 (and any mutant a.out variants of FreeBSD 3) was
3673 discontinued in GCC 4.0.
3675 In order to better utilize FreeBSD base system functionality and match
3676 the configuration of the system compiler, GCC 4.5 and above as well as
3677 GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present
3678 on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default
3679 (on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside
3680 @file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled
3681 by GCC 4.5 and above.
3683 We support FreeBSD using the ELF file format with DWARF 2 debugging
3684 for all CPU architectures. You may use @option{-gstabs} instead of
3685 @option{-g}, if you really want the old debugging format. There are
3686 no known issues with mixing object files and libraries with different
3687 debugging formats. Otherwise, this release of GCC should now match
3688 more of the configuration used in the stock FreeBSD configuration of
3689 GCC@. In particular, @option{--enable-threads} is now configured by
3690 default. However, as a general user, do not attempt to replace the
3691 system compiler with this release. Known to bootstrap and check with
3692 good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap
3693 and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4,
3694 4.5, 4.8, 4.9 and 5-CURRENT@.
3696 The version of binutils installed in @file{/usr/bin} probably works
3697 with this release of GCC@. Bootstrapping against the latest GNU
3698 binutils and/or the version found in @file{/usr/ports/devel/binutils} has
3699 been known to enable additional features and improve overall testsuite
3700 results. However, it is currently known that boehm-gc may not configure
3701 properly on FreeBSD prior to the FreeBSD 7.0 release with GNU binutils
3710 This configuration is intended for embedded systems.
3717 Renesas H8/300 series of processors.
3719 Please have a look at the @uref{binaries.html,,binaries page}.
3721 The calling convention and structure layout has changed in release 2.6.
3722 All code must be recompiled. The calling convention now passes the
3723 first three arguments in function calls in registers. Structures are no
3724 longer a multiple of 2 bytes.
3729 @anchor{hppa-hp-hpux}
3730 @heading hppa*-hp-hpux*
3731 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
3733 We require using gas/binutils on all hppa platforms. Version 2.19 or
3734 later is recommended.
3736 It may be helpful to configure GCC with the
3737 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
3738 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
3740 The HP assembler should not be used with GCC. It is rarely tested and may
3741 not work. It shouldn't be used with any languages other than C due to its
3744 Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging
3745 format which GCC does not know about). It also inserts timestamps
3746 into each object file it creates, causing the 3-stage comparison test to
3747 fail during a bootstrap. You should be able to continue by saying
3748 @samp{make all-host all-target} after getting the failure from @samp{make}.
3750 Various GCC features are not supported. For example, it does not support weak
3751 symbols or alias definitions. As a result, explicit template instantiations
3752 are required when using C++. This makes it difficult if not impossible to
3753 build many C++ applications.
3755 There are two default scheduling models for instructions. These are
3756 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
3757 architecture specified for the target machine when configuring.
3758 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
3759 the target is a @samp{hppa1*} machine.
3761 The PROCESSOR_8000 model is not well suited to older processors. Thus,
3762 it is important to completely specify the machine architecture when
3763 configuring if you want a model other than PROCESSOR_8000. The macro
3764 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
3765 default scheduling model is desired.
3767 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
3768 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
3769 This namespace change might cause problems when bootstrapping with
3770 an earlier version of GCC or the HP compiler as essentially the same
3771 namespace is required for an entire build. This problem can be avoided
3772 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
3773 or @samp{98}. Another way is to add an appropriate set of predefines
3774 to @env{CC}. The description for the @option{munix=} option contains
3775 a list of the predefines used with each standard.
3777 More specific information to @samp{hppa*-hp-hpux*} targets follows.
3782 @anchor{hppa-hp-hpux10}
3783 @heading hppa*-hp-hpux10
3784 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
3785 @code{PHCO_19798} from HP@.
3787 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
3788 used for one-only code and data. This resolves many of the previous
3789 problems in using C++ on this target. However, the ABI is not compatible
3790 with the one implemented under HP-UX 11 using secondary definitions.
3795 @anchor{hppa-hp-hpux11}
3796 @heading hppa*-hp-hpux11
3797 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
3798 be used to compile GCC 3.0 and up.
3800 The libffi library haven't been ported to 64-bit HP-UX@ and doesn't build.
3802 Refer to @uref{binaries.html,,binaries} for information about obtaining
3803 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
3804 to build the Ada language as it cannot be bootstrapped using C@. Ada is
3805 only available for the 32-bit PA-RISC runtime.
3807 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
3808 bundled compiler supports only traditional C; you will need either HP's
3809 unbundled compiler, or a binary distribution of GCC@.
3811 It is possible to build GCC 3.3 starting with the bundled HP compiler,
3812 but the process requires several steps. GCC 3.3 can then be used to
3813 build later versions.
3815 There are several possible approaches to building the distribution.
3816 Binutils can be built first using the HP tools. Then, the GCC
3817 distribution can be built. The second approach is to build GCC
3818 first using the HP tools, then build binutils, then rebuild GCC@.
3819 There have been problems with various binary distributions, so it
3820 is best not to start from a binary distribution.
3822 On 64-bit capable systems, there are two distinct targets. Different
3823 installation prefixes must be used if both are to be installed on
3824 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
3825 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
3826 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
3827 PA-RISC 2.0 architecture.
3829 The script config.guess now selects the target type based on the compiler
3830 detected during configuration. You must define @env{PATH} or @env{CC} so
3831 that configure finds an appropriate compiler for the initial bootstrap.
3832 When @env{CC} is used, the definition should contain the options that are
3833 needed whenever @env{CC} is used.
3835 Specifically, options that determine the runtime architecture must be
3836 in @env{CC} to correctly select the target for the build. It is also
3837 convenient to place many other compiler options in @env{CC}. For example,
3838 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
3839 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
3840 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
3841 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
3842 macro definition table of cpp needs to be increased for a successful
3843 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
3844 be defined when building with the bundled compiler, or when using the
3845 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
3847 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
3848 with the @option{--with-ld=@dots{}} option. This overrides the standard
3849 search for ld. The two linkers supported on this target require different
3850 commands. The default linker is determined during configuration. As a
3851 result, it's not possible to switch linkers in the middle of a GCC build.
3852 This has been reported to sometimes occur in unified builds of binutils
3855 A recent linker patch must be installed for the correct operation of
3856 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
3857 oldest linker patches that are known to work. They are for HP-UX
3858 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
3859 @code{PHSS_24304}, might be usable but it hasn't been tested. These
3860 patches have been superseded. Consult the HP patch database to obtain
3861 the currently recommended linker patch for your system.
3863 The patches are necessary for the support of weak symbols on the
3864 32-bit port, and for the running of initializers and finalizers. Weak
3865 symbols are implemented using SOM secondary definition symbols. Prior
3866 to HP-UX 11, there are bugs in the linker support for secondary symbols.
3867 The patches correct a problem of linker core dumps creating shared
3868 libraries containing secondary symbols, as well as various other
3869 linking issues involving secondary symbols.
3871 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
3872 run initializers and finalizers on the 64-bit port. The 32-bit port
3873 uses the linker @option{+init} and @option{+fini} options for the same
3874 purpose. The patches correct various problems with the +init/+fini
3875 options, including program core dumps. Binutils 2.14 corrects a
3876 problem on the 64-bit port resulting from HP's non-standard use of
3877 the .init and .fini sections for array initializers and finalizers.
3879 Although the HP and GNU linkers are both supported for the
3880 @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the
3881 HP linker be used for link editing on this target.
3883 At this time, the GNU linker does not support the creation of long
3884 branch stubs. As a result, it cannot successfully link binaries
3885 containing branch offsets larger than 8 megabytes. In addition,
3886 there are problems linking shared libraries, linking executables
3887 with @option{-static}, and with dwarf2 unwind and exception support.
3888 It also doesn't provide stubs for internal calls to global functions
3889 in shared libraries, so these calls cannot be overloaded.
3891 The HP dynamic loader does not support GNU symbol versioning, so symbol
3892 versioning is not supported. It may be necessary to disable symbol
3893 versioning with @option{--disable-symvers} when using GNU ld.
3895 POSIX threads are the default. The optional DCE thread library is not
3896 supported, so @option{--enable-threads=dce} does not work.
3901 @anchor{x-x-linux-gnu}
3902 @heading *-*-linux-gnu
3903 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
3904 in glibc 2.2.5 and later. More information is available in the
3905 libstdc++-v3 documentation.
3910 @anchor{ix86-x-linux}
3911 @heading i?86-*-linux*
3912 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
3913 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
3915 If you receive Signal 11 errors when building on GNU/Linux, then it is
3916 possible you have a hardware problem. Further information on this can be
3917 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
3922 @anchor{ix86-x-solaris2}
3923 @heading i?86-*-solaris2*
3924 Use this for Solaris 11 or later on x86 and x86-64 systems. Starting
3925 with GCC 4.7, there is also a 64-bit @samp{amd64-*-solaris2*} or
3926 @samp{x86_64-*-solaris2*} configuration that corresponds to
3927 @samp{sparcv9-sun-solaris2*}.
3929 It is recommended that you configure GCC to use the GNU assembler. The
3930 versions included in Solaris 11, from GNU binutils 2.19 or
3931 newer (also available as @file{/usr/bin/gas} and
3932 @file{/usr/gnu/bin/as}), work fine. The current version, from GNU
3933 binutils 2.32, is known to work, but the version from GNU binutils 2.26
3934 must be avoided. Recent versions of the Solaris assembler in
3935 @file{/usr/ccs/bin/as} work almost as well, though.
3936 @c FIXME: as patch requirements?
3938 For linking, the Solaris linker, is preferred. If you want to use the GNU
3939 linker instead, the version in Solaris 11, from GNU binutils 2.19 or
3940 newer (also in @file{/usr/gnu/bin/ld} and @file{/usr/bin/gld}), works,
3941 as does the latest version, from GNU binutils 2.32.
3943 To use GNU @command{as}, configure with the options
3944 @option{--with-gnu-as --with-as=@//usr/@/gnu/@/bin/@/as}. It may be necessary
3945 to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to
3946 guarantee use of Sun @command{ld}.
3947 @c FIXME: why --without-gnu-ld --with-ld?
3952 @anchor{ia64-x-linux}
3953 @heading ia64-*-linux
3954 IA-64 processor (also known as IPF, or Itanium Processor Family)
3957 If you are using the installed system libunwind library with
3958 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
3961 None of the following versions of GCC has an ABI that is compatible
3962 with any of the other versions in this list, with the exception that
3963 Red Hat 2.96 and Trillian 000171 are compatible with each other:
3964 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
3965 This primarily affects C++ programs and programs that create shared libraries.
3966 GCC 3.1 or later is recommended for compiling linux, the kernel.
3967 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
3968 more major ABI changes are expected.
3973 @anchor{ia64-x-hpux}
3974 @heading ia64-*-hpux*
3975 Building GCC on this target requires the GNU Assembler. The bundled HP
3976 assembler will not work. To prevent GCC from using the wrong assembler,
3977 the option @option{--with-gnu-as} may be necessary.
3979 The GCC libunwind library has not been ported to HPUX@. This means that for
3980 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
3981 is required to build GCC@. For GCC 3.3 and later, this is the default.
3982 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
3983 removed and the system libunwind library will always be used.
3987 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
3991 Support for AIX version 3 and older was discontinued in GCC 3.4.
3992 Support for AIX version 4.2 and older was discontinued in GCC 4.5.
3994 ``out of memory'' bootstrap failures may indicate a problem with
3995 process resource limits (ulimit). Hard limits are configured in the
3996 @file{/etc/security/limits} system configuration file.
3998 GCC 4.9 and above require a C++ compiler for bootstrap. IBM VAC++ / xlC
3999 cannot bootstrap GCC. xlc can bootstrap an older version of GCC and
4000 G++ can bootstrap recent releases of GCC.
4002 GCC can bootstrap with recent versions of IBM XLC, but bootstrapping
4003 with an earlier release of GCC is recommended. Bootstrapping with XLC
4004 requires a larger data segment, which can be enabled through the
4005 @var{LDR_CNTRL} environment variable, e.g.,
4008 % LDR_CNTRL=MAXDATA=0x50000000
4012 One can start with a pre-compiled version of GCC to build from
4013 sources. One may delete GCC's ``fixed'' header files when starting
4014 with a version of GCC built for an earlier release of AIX.
4016 To speed up the configuration phases of bootstrapping and installing GCC,
4017 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
4020 % CONFIG_SHELL=/opt/freeware/bin/bash
4021 % export CONFIG_SHELL
4024 and then proceed as described in @uref{build.html,,the build
4025 instructions}, where we strongly recommend specifying an absolute path
4026 to invoke @var{srcdir}/configure.
4028 Because GCC on AIX is built as a 32-bit executable by default,
4029 (although it can generate 64-bit programs) the GMP and MPFR libraries
4030 required by gfortran must be 32-bit libraries. Building GMP and MPFR
4031 as static archive libraries works better than shared libraries.
4033 Errors involving @code{alloca} when building GCC generally are due
4034 to an incorrect definition of @code{CC} in the Makefile or mixing files
4035 compiled with the native C compiler and GCC@. During the stage1 phase of
4036 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
4037 (not @command{xlc}). Once @command{configure} has been informed of
4038 @command{xlc}, one needs to use @samp{make distclean} to remove the
4039 configure cache files and ensure that @env{CC} environment variable
4040 does not provide a definition that will confuse @command{configure}.
4041 If this error occurs during stage2 or later, then the problem most likely
4042 is the version of Make (see above).
4044 The native @command{as} and @command{ld} are recommended for
4045 bootstrapping on AIX@. The GNU Assembler, GNU Linker, and GNU
4046 Binutils version 2.20 is the minimum level that supports bootstrap on
4047 AIX 5@. The GNU Assembler has not been updated to support AIX 6@ or
4048 AIX 7. The native AIX tools do interoperate with GCC@.
4050 AIX 7.1 added partial support for DWARF debugging, but full support
4051 requires AIX 7.1 TL03 SP7 that supports additional DWARF sections and
4052 fixes a bug in the assembler. AIX 7.1 TL03 SP5 distributed a version
4053 of libm.a missing important symbols; a fix for IV77796 will be
4056 AIX 5.3 TL10, AIX 6.1 TL05 and AIX 7.1 TL00 introduced an AIX
4057 assembler change that sometimes produces corrupt assembly files
4058 causing AIX linker errors. The bug breaks GCC bootstrap on AIX and
4059 can cause compilation failures with existing GCC installations. An
4060 AIX iFix for AIX 5.3 is available (APAR IZ98385 for AIX 5.3 TL10, APAR
4061 IZ98477 for AIX 5.3 TL11 and IZ98134 for AIX 5.3 TL12). AIX 5.3 TL11 SP8,
4062 AIX 5.3 TL12 SP5, AIX 6.1 TL04 SP11, AIX 6.1 TL05 SP7, AIX 6.1 TL06 SP6,
4063 AIX 6.1 TL07 and AIX 7.1 TL01 should include the fix.
4065 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
4066 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
4067 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
4068 referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1)
4070 @anchor{TransferAixShobj}
4071 @samp{libstdc++} in GCC 3.4 increments the major version number of the
4072 shared object and GCC installation places the @file{libstdc++.a}
4073 shared library in a common location which will overwrite the and GCC
4074 3.3 version of the shared library. Applications either need to be
4075 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
4076 versions of the @samp{libstdc++} shared object needs to be available
4077 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
4078 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
4079 installed for runtime dynamic loading using the following steps to set
4080 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
4081 multilib @file{libstdc++.a} installed:
4083 Extract the shared objects from the currently installed
4084 @file{libstdc++.a} archive:
4086 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
4089 Enable the @samp{F_LOADONLY} flag so that the shared object will be
4090 available for runtime dynamic loading, but not linking:
4092 % strip -e libstdc++.so.4 libstdc++.so.5
4095 Archive the runtime-only shared object in the GCC 3.4
4096 @file{libstdc++.a} archive:
4098 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
4102 @uref{./configure.html#WithAixSoname,,@option{--with-aix-soname=svr4}}
4103 configure option may drop the need for this procedure for libraries that
4106 Linking executables and shared libraries may produce warnings of
4107 duplicate symbols. The assembly files generated by GCC for AIX always
4108 have included multiple symbol definitions for certain global variable
4109 and function declarations in the original program. The warnings should
4110 not prevent the linker from producing a correct library or runnable
4113 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
4114 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
4115 to parse archive libraries did not handle the new format correctly.
4116 These routines are used by GCC and result in error messages during
4117 linking such as ``not a COFF file''. The version of the routines shipped
4118 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
4119 option of the archive command may be used to create archives of 32-bit
4120 objects using the original ``small format''. A correct version of the
4121 routines is shipped with AIX 4.3.2 and above.
4123 Some versions of the AIX binder (linker) can fail with a relocation
4124 overflow severe error when the @option{-bbigtoc} option is used to link
4125 GCC-produced object files into an executable that overflows the TOC@. A fix
4126 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
4127 available from IBM Customer Support and from its
4128 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4129 website as PTF U455193.
4131 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
4132 with a segmentation fault when invoked by any version of GCC@. A fix for
4133 APAR IX87327 is available from IBM Customer Support and from its
4134 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4135 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
4137 The initial assembler shipped with AIX 4.3.0 generates incorrect object
4138 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
4139 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
4140 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4141 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
4143 AIX provides National Language Support (NLS)@. Compilers and assemblers
4144 use NLS to support locale-specific representations of various data
4145 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
4146 separating decimal fractions). There have been problems reported where
4147 GCC does not produce the same floating-point formats that the assembler
4148 expects. If one encounters this problem, set the @env{LANG}
4149 environment variable to @samp{C} or @samp{En_US}.
4151 A default can be specified with the @option{-mcpu=@var{cpu_type}}
4152 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
4157 @anchor{iq2000-x-elf}
4158 @heading iq2000-*-elf
4159 Vitesse IQ2000 processors. These are used in embedded
4160 applications. There are no standard Unix configurations.
4167 Lattice Mico32 processor.
4168 This configuration is intended for embedded systems.
4173 @anchor{lm32-x-uclinux}
4174 @heading lm32-*-uclinux
4175 Lattice Mico32 processor.
4176 This configuration is intended for embedded systems running uClinux.
4183 Renesas M32C processor.
4184 This configuration is intended for embedded systems.
4191 Renesas M32R processor.
4192 This configuration is intended for embedded systems.
4200 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
4202 build libraries for both M680x0 and ColdFire processors. If you only
4203 need the M680x0 libraries, you can omit the ColdFire ones by passing
4204 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
4205 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
4206 @command{configure}. These targets default to 5206 or 5475 code as
4207 appropriate for the target system when
4208 configured with @option{--with-arch=cf} and 68020 code otherwise.
4210 The @samp{m68k-*-netbsd} and
4211 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
4212 option. They will generate ColdFire CFV4e code when configured with
4213 @option{--with-arch=cf} and 68020 code otherwise.
4215 You can override the default processors listed above by configuring
4216 with @option{--with-cpu=@var{target}}. This @var{target} can either
4217 be a @option{-mcpu} argument or one of the following values:
4218 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
4219 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
4221 GCC requires at least binutils version 2.17 on these targets.
4226 @anchor{m68k-x-uclinux}
4227 @heading m68k-*-uclinux
4228 GCC 4.3 changed the uClinux configuration so that it uses the
4229 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
4230 It also added improved support for C++ and flat shared libraries,
4231 both of which were ABI changes.
4236 @anchor{microblaze-x-elf}
4237 @heading microblaze-*-elf
4238 Xilinx MicroBlaze processor.
4239 This configuration is intended for embedded systems.
4246 If on a MIPS system you get an error message saying ``does not have gp
4247 sections for all it's [sic] sectons [sic]'', don't worry about it. This
4248 happens whenever you use GAS with the MIPS linker, but there is not
4249 really anything wrong, and it is okay to use the output file. You can
4250 stop such warnings by installing the GNU linker.
4252 It would be nice to extend GAS to produce the gp tables, but they are
4253 optional, and there should not be a warning about their absence.
4255 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
4256 and later. A patch went in just after the GCC 3.3 release to
4257 make @samp{mips*-*-*} use the generic implementation instead. You can also
4258 configure for @samp{mipsel-elf} as a workaround. The
4259 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
4260 work on this is expected in future releases.
4262 @c If you make --with-llsc the default for another target, please also
4263 @c update the description of the --with-llsc option.
4265 The built-in @code{__sync_*} functions are available on MIPS II and
4266 later systems and others that support the @samp{ll}, @samp{sc} and
4267 @samp{sync} instructions. This can be overridden by passing
4268 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
4269 Since the Linux kernel emulates these instructions if they are
4270 missing, the default for @samp{mips*-*-linux*} targets is
4271 @option{--with-llsc}. The @option{--with-llsc} and
4272 @option{--without-llsc} configure options may be overridden at compile
4273 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
4276 MIPS systems check for division by zero (unless
4277 @option{-mno-check-zero-division} is passed to the compiler) by
4278 generating either a conditional trap or a break instruction. Using
4279 trap results in smaller code, but is only supported on MIPS II and
4280 later. Also, some versions of the Linux kernel have a bug that
4281 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
4282 the use of break, use the @option{--with-divide=breaks}
4283 @command{configure} option when configuring GCC@. The default is to
4284 use traps on systems that support them.
4289 @anchor{moxie-x-elf}
4290 @heading moxie-*-elf
4291 The moxie processor.
4296 @anchor{msp430-x-elf}
4297 @heading msp430-*-elf
4298 TI MSP430 processor.
4299 This configuration is intended for embedded systems.
4304 @anchor{nds32le-x-elf}
4305 @heading nds32le-*-elf
4306 Andes NDS32 target in little endian mode.
4311 @anchor{nds32be-x-elf}
4312 @heading nds32be-*-elf
4313 Andes NDS32 target in big endian mode.
4318 @anchor{nvptx-x-none}
4319 @heading nvptx-*-none
4322 Instead of GNU binutils, you will need to install
4323 @uref{https://github.com/MentorEmbedded/nvptx-tools/,,nvptx-tools}.
4324 Tell GCC where to find it:
4325 @option{--with-build-time-tools=[install-nvptx-tools]/nvptx-none/bin}.
4327 You will need newlib 3.0 git revision
4328 cd31fbb2aea25f94d7ecedc9db16dfc87ab0c316 or later. It can be
4329 automatically built together with GCC@. For this, add a symbolic link
4330 to nvptx-newlib's @file{newlib} directory to the directory containing
4333 Use the @option{--disable-sjlj-exceptions} and
4334 @option{--enable-newlib-io-long-long} options when configuring.
4341 The OpenRISC 1000 32-bit processor with delay slots.
4342 This configuration is intended for embedded systems.
4347 @anchor{or1k-x-linux}
4348 @heading or1k-*-linux
4349 The OpenRISC 1000 32-bit processor with delay slots.
4354 @anchor{powerpc-x-x}
4355 @heading powerpc-*-*
4356 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
4357 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
4359 You will need GNU binutils 2.15 or newer.
4364 @anchor{powerpc-x-darwin}
4365 @heading powerpc-*-darwin*
4366 PowerPC running Darwin (Mac OS X kernel).
4368 Pre-installed versions of Mac OS X may not include any developer tools,
4369 meaning that you will not be able to build GCC from source. Tool
4370 binaries are available at
4371 @uref{https://opensource.apple.com}.
4373 This version of GCC requires at least cctools-590.36. The
4374 cctools-590.36 package referenced from
4375 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
4376 on systems older than 10.3.9 (aka darwin7.9.0).
4381 @anchor{powerpc-x-elf}
4382 @heading powerpc-*-elf
4383 PowerPC system in big endian mode, running System V.4.
4388 @anchor{powerpc-x-linux-gnu}
4389 @heading powerpc*-*-linux-gnu*
4390 PowerPC system in big endian mode running Linux.
4395 @anchor{powerpc-x-netbsd}
4396 @heading powerpc-*-netbsd*
4397 PowerPC system in big endian mode running NetBSD@.
4402 @anchor{powerpc-x-eabisim}
4403 @heading powerpc-*-eabisim
4404 Embedded PowerPC system in big endian mode for use in running under the
4410 @anchor{powerpc-x-eabi}
4411 @heading powerpc-*-eabi
4412 Embedded PowerPC system in big endian mode.
4417 @anchor{powerpcle-x-elf}
4418 @heading powerpcle-*-elf
4419 PowerPC system in little endian mode, running System V.4.
4424 @anchor{powerpcle-x-eabisim}
4425 @heading powerpcle-*-eabisim
4426 Embedded PowerPC system in little endian mode for use in running under
4432 @anchor{powerpcle-x-eabi}
4433 @heading powerpcle-*-eabi
4434 Embedded PowerPC system in little endian mode.
4441 The Renesas RL78 processor.
4442 This configuration is intended for embedded systems.
4447 @anchor{riscv32-x-elf}
4448 @heading riscv32-*-elf
4449 The RISC-V RV32 instruction set.
4450 This configuration is intended for embedded systems.
4451 This (and all other RISC-V) targets are supported upstream as of the
4452 binutils 2.28 release.
4457 @anchor{riscv32-x-linux}
4458 @heading riscv32-*-linux
4459 The RISC-V RV32 instruction set running GNU/Linux.
4460 This (and all other RISC-V) targets are supported upstream as of the
4461 binutils 2.28 release.
4466 @anchor{riscv64-x-elf}
4467 @heading riscv64-*-elf
4468 The RISC-V RV64 instruction set.
4469 This configuration is intended for embedded systems.
4470 This (and all other RISC-V) targets are supported upstream as of the
4471 binutils 2.28 release.
4476 @anchor{riscv64-x-linux}
4477 @heading riscv64-*-linux
4478 The RISC-V RV64 instruction set running GNU/Linux.
4479 This (and all other RISC-V) targets are supported upstream as of the
4480 binutils 2.28 release.
4487 The Renesas RX processor.
4492 @anchor{s390-x-linux}
4493 @heading s390-*-linux*
4494 S/390 system running GNU/Linux for S/390@.
4499 @anchor{s390x-x-linux}
4500 @heading s390x-*-linux*
4501 zSeries system (64-bit) running GNU/Linux for zSeries@.
4506 @anchor{s390x-ibm-tpf}
4507 @heading s390x-ibm-tpf*
4508 zSeries system (64-bit) running TPF@. This platform is
4509 supported as cross-compilation target only.
4514 @c Please use Solaris 2 to refer to all release of Solaris, starting
4515 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
4516 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
4517 @c alone is too unspecific and must be avoided.
4518 @anchor{x-x-solaris2}
4519 @heading *-*-solaris2*
4520 Support for Solaris 10 has been removed in GCC 10. Support for Solaris
4521 9 has been removed in GCC 5. Support for Solaris 8 has been removed in
4522 GCC 4.8. Support for Solaris 7 has been removed in GCC 4.6.
4524 Solaris 11 provides GCC 4.5.2, 4.7.3, and 4.8.2 as
4525 @command{/usr/gcc/4.5/bin/gcc} or similar. Alternatively,
4526 you can install a pre-built GCC to bootstrap and install GCC. See the
4527 @uref{binaries.html,,binaries page} for details.
4529 The Solaris 2 @command{/bin/sh} will often fail to configure
4530 @samp{libstdc++-v3}or @samp{boehm-gc}. We therefore recommend using the
4531 following initial sequence of commands
4534 % CONFIG_SHELL=/bin/ksh
4535 % export CONFIG_SHELL
4539 and proceed as described in @uref{configure.html,,the configure instructions}.
4540 In addition we strongly recommend specifying an absolute path to invoke
4541 @command{@var{srcdir}/configure}.
4543 In Solaris 11, you need to check for @code{system/header},
4544 @code{system/linker}, and @code{developer/assembler} packages.
4546 Trying to use the linker and other tools in
4547 @file{/usr/ucb} to install GCC has been observed to cause trouble.
4548 For example, the linker may hang indefinitely. The fix is to remove
4549 @file{/usr/ucb} from your @env{PATH}.
4551 The build process works more smoothly with the legacy Sun tools so, if you
4552 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
4553 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
4555 We recommend the use of the Solaris assembler or the GNU assembler, in
4556 conjunction with the Solaris linker. The GNU @command{as}
4557 versions included in Solaris 11,
4558 from GNU binutils 2.19 or newer (also in @file{/usr/bin/gas} and
4559 @file{/usr/gnu/bin/as}), are known to work.
4560 The current version, from GNU binutils 2.32,
4561 is known to work as well. Note that your mileage may vary
4562 if you use a combination of the GNU tools and the Solaris tools: while the
4563 combination GNU @command{as} + Sun @command{ld} should reasonably work,
4564 the reverse combination Sun @command{as} + GNU @command{ld} may fail to
4565 build or cause memory corruption at runtime in some cases for C++ programs.
4567 GNU @command{ld} usually works as well. Again, the current
4568 version (2.32) is known to work, but generally lacks platform specific
4569 features, so better stay with Solaris @command{ld}. To use the LTO linker
4570 plugin (@option{-fuse-linker-plugin}) with GNU @command{ld}, GNU
4571 binutils @emph{must} be configured with @option{--enable-largefile}.
4573 To enable symbol versioning in @samp{libstdc++} with the Solaris linker,
4574 you need to have any version of GNU @command{c++filt}, which is part of
4575 GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no
4576 appropriate version is found. Solaris @command{c++filt} from the Solaris
4577 Studio compilers does @emph{not} work.
4584 This section contains general configuration information for all
4585 SPARC-based platforms. In addition to reading this section, please
4586 read all other sections that match your target.
4588 Newer versions of the GNU Multiple Precision Library (GMP), the MPFR
4589 library and the MPC library are known to be miscompiled by earlier
4590 versions of GCC on these platforms. We therefore recommend the use
4591 of the exact versions of these libraries listed as minimal versions
4592 in @uref{prerequisites.html,,the prerequisites}.
4597 @anchor{sparc-sun-solaris2}
4598 @heading sparc-sun-solaris2*
4599 When GCC is configured to use GNU binutils 2.14 or later, the binaries
4600 produced are smaller than the ones produced using Sun's native tools;
4601 this difference is quite significant for binaries containing debugging
4604 Starting with Solaris 7, the operating system is capable of executing
4605 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
4606 this; the @option{-m64} option enables 64-bit code generation.
4607 However, if all you want is code tuned for the UltraSPARC CPU, you
4608 should try the @option{-mtune=ultrasparc} option instead, which produces
4609 code that, unlike full 64-bit code, can still run on non-UltraSPARC
4612 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4613 library or the MPC library on a Solaris 7 or later system, the canonical
4614 target triplet must be specified as the @command{build} parameter on the
4615 configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and
4616 not that of GMP or MPFR or MPC). For example on a Solaris 9 system:
4619 % ./configure --build=sparc-sun-solaris2.9 --prefix=xxx
4625 @anchor{sparc-x-linux}
4626 @heading sparc-*-linux*
4631 @anchor{sparc64-x-solaris2}
4632 @heading sparc64-*-solaris2*
4633 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4634 library or the MPC library, the canonical target triplet must be specified
4635 as the @command{build} parameter on the configure line. For example
4636 on a Solaris 9 system:
4639 % ./configure --build=sparc64-sun-solaris2.9 --prefix=xxx
4645 @anchor{sparcv9-x-solaris2}
4646 @heading sparcv9-*-solaris2*
4647 This is a synonym for @samp{sparc64-*-solaris2*}.
4654 The C6X family of processors. This port requires binutils-2.22 or newer.
4659 @anchor{tilegx-*-linux}
4660 @heading tilegx-*-linux*
4661 The TILE-Gx processor in little endian mode, running GNU/Linux. This
4662 port requires binutils-2.22 or newer.
4667 @anchor{tilegxbe-*-linux}
4668 @heading tilegxbe-*-linux*
4669 The TILE-Gx processor in big endian mode, running GNU/Linux. This
4670 port requires binutils-2.23 or newer.
4675 @anchor{tilepro-*-linux}
4676 @heading tilepro-*-linux*
4677 The TILEPro processor running GNU/Linux. This port requires
4678 binutils-2.22 or newer.
4683 @anchor{visium-x-elf}
4684 @heading visium-*-elf
4685 CDS VISIUMcore processor.
4686 This configuration is intended for embedded systems.
4691 @anchor{x-x-vxworks}
4692 @heading *-*-vxworks*
4693 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4694 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4695 We welcome patches for other architectures supported by VxWorks 5.5.
4696 Support for VxWorks AE would also be welcome; we believe this is merely
4697 a matter of writing an appropriate ``configlette'' (see below). We are
4698 not interested in supporting older, a.out or COFF-based, versions of
4701 VxWorks comes with an older version of GCC installed in
4702 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4703 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4704 Before running @command{configure}, create the directories @file{@var{prefix}}
4705 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4706 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4707 include that directory while running both @command{configure} and
4710 You must give @command{configure} the
4711 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4712 find the VxWorks system headers. Since VxWorks is a cross compilation
4713 target only, you must also specify @option{--target=@var{target}}.
4714 @command{configure} will attempt to create the directory
4715 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4716 make sure the user running @command{configure} has sufficient privilege
4719 GCC's exception handling runtime requires a special ``configlette''
4720 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4721 that file to add the module to your kernel build. (Future versions of
4722 VxWorks will incorporate this module.)
4728 @heading x86_64-*-*, amd64-*-*
4729 GCC supports the x86-64 architecture implemented by the AMD64 processor
4730 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4731 On GNU/Linux the default is a bi-arch compiler which is able to generate
4732 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4737 @anchor{x86-64-x-solaris2}
4738 @heading x86_64-*-solaris2*
4739 GCC also supports the x86-64 architecture implemented by the AMD64
4740 processor (@samp{amd64-*-*} is an alias for @samp{x86_64-*-*}) on
4741 Solaris 10 or later. Unlike other systems, without special options a
4742 bi-arch compiler is built which generates 32-bit code by default, but
4743 can generate 64-bit x86-64 code with the @option{-m64} switch. Since
4744 GCC 4.7, there is also a configuration that defaults to 64-bit code, but
4745 can generate 32-bit code with @option{-m32}. To configure and build
4746 this way, you have to provide all support libraries like @file{libgmp}
4747 as 64-bit code, configure with @option{--target=x86_64-pc-solaris2.1x}
4748 and @samp{CC=gcc -m64}.
4753 @anchor{xtensa-x-elf}
4754 @heading xtensa*-*-elf
4755 This target is intended for embedded Xtensa systems using the
4756 @samp{newlib} C library. It uses ELF but does not support shared
4757 objects. Designed-defined instructions specified via the
4758 Tensilica Instruction Extension (TIE) language are only supported
4759 through inline assembly.
4761 The Xtensa configuration information must be specified prior to
4762 building GCC@. The @file{include/xtensa-config.h} header
4763 file contains the configuration information. If you created your
4764 own Xtensa configuration with the Xtensa Processor Generator, the
4765 downloaded files include a customized copy of this header file,
4766 which you can use to replace the default header file.
4771 @anchor{xtensa-x-linux}
4772 @heading xtensa*-*-linux*
4773 This target is for Xtensa systems running GNU/Linux. It supports ELF
4774 shared objects and the GNU C library (glibc). It also generates
4775 position-independent code (PIC) regardless of whether the
4776 @option{-fpic} or @option{-fPIC} options are used. In other
4777 respects, this target is the same as the
4778 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
4784 @heading Microsoft Windows
4786 @subheading Intel 16-bit versions
4787 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
4790 However, the 32-bit port has limited support for Microsoft
4791 Windows 3.11 in the Win32s environment, as a target only. See below.
4793 @subheading Intel 32-bit versions
4794 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
4795 XP, and Windows Vista, are supported by several different target
4796 platforms. These targets differ in which Windows subsystem they target
4797 and which C libraries are used.
4800 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
4801 Linux API emulation layer in the Win32 subsystem.
4802 @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for
4803 the Win32 subsystem that provides a subset of POSIX.
4804 @item MKS i386-pc-mks: NuTCracker from MKS. See
4805 @uref{https://www.mkssoftware.com} for more information.
4808 @subheading Intel 64-bit versions
4809 GCC contains support for x86-64 using the mingw-w64
4810 runtime library, available from @uref{http://mingw-w64.org/doku.php}.
4811 This library should be used with the target triple x86_64-pc-mingw32.
4813 Presently Windows for Itanium is not supported.
4815 @subheading Windows CE
4816 Windows CE is supported as a target only on Hitachi
4817 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
4819 @subheading Other Windows Platforms
4820 GCC no longer supports Windows NT on the Alpha or PowerPC.
4822 GCC no longer supports the Windows POSIX subsystem. However, it does
4823 support the Interix subsystem. See above.
4825 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
4827 PW32 (i386-pc-pw32) support was never completed, and the project seems to
4828 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
4830 UWIN support has been removed due to a lack of maintenance.
4837 Ports of GCC are included with the
4838 @uref{http://www.cygwin.com/,,Cygwin environment}.
4840 GCC will build under Cygwin without modification; it does not build
4841 with Microsoft's C++ compiler and there are no plans to make it do so.
4843 The Cygwin native compiler can be configured to target any 32-bit x86
4844 cpu architecture desired; the default is i686-pc-cygwin. It should be
4845 used with as up-to-date a version of binutils as possible; use either
4846 the latest official GNU binutils release in the Cygwin distribution,
4847 or version 2.20 or above if building your own.
4852 @anchor{x-x-mingw32}
4853 @heading *-*-mingw32
4854 GCC will build with and support only MinGW runtime 3.12 and later.
4855 Earlier versions of headers are incompatible with the new default semantics
4856 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
4862 @heading Older systems
4863 GCC contains support files for many older (1980s and early
4864 1990s) Unix variants. For the most part, support for these systems
4865 has not been deliberately removed, but it has not been maintained for
4866 several years and may suffer from bitrot.
4868 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
4869 Support for these systems is still present in that release, but
4870 @command{configure} will fail unless the @option{--enable-obsolete}
4871 option is given. Unless a maintainer steps forward, support for these
4872 systems will be removed from the next release of GCC@.
4874 Support for old systems as hosts for GCC can cause problems if the
4875 workarounds for compiler, library and operating system bugs affect the
4876 cleanliness or maintainability of the rest of GCC@. In some cases, to
4877 bring GCC up on such a system, if still possible with current GCC, may
4878 require first installing an old version of GCC which did work on that
4879 system, and using it to compile a more recent GCC, to avoid bugs in the
4880 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
4881 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
4882 sites}. Header bugs may generally be avoided using
4883 @command{fixincludes}, but bugs or deficiencies in libraries and the
4884 operating system may still cause problems.
4886 Support for older systems as targets for cross-compilation is less
4887 problematic than support for them as hosts for GCC; if an enthusiast
4888 wishes to make such a target work again (including resurrecting any of
4889 the targets that never worked with GCC 2, starting from the last
4890 version before they were removed), patches
4891 @uref{../contribute.html,,following the usual requirements} would be
4892 likely to be accepted, since they should not affect the support for more
4895 For some systems, old versions of GNU binutils may also be useful,
4896 and are available from @file{pub/binutils/old-releases} on
4897 @uref{https://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
4899 Some of the information on specific systems above relates to
4900 such older systems, but much of the information
4901 about GCC on such systems (which may no longer be applicable to
4902 current GCC) is to be found in the GCC texinfo manual.
4908 @heading all ELF targets (SVR4, Solaris 2, etc.)
4909 C++ support is significantly better on ELF targets if you use the
4910 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
4911 inlines, vtables and template instantiations will be discarded
4920 @uref{./index.html,,Return to the GCC Installation page}
4924 @c ***Old documentation******************************************************
4926 @include install-old.texi
4932 @uref{./index.html,,Return to the GCC Installation page}
4936 @c ***GFDL********************************************************************
4944 @uref{./index.html,,Return to the GCC Installation page}
4948 @c ***************************************************************************
4949 @c Part 6 The End of the Document
4951 @comment node-name, next, previous, up
4952 @node Concept Index, , GNU Free Documentation License, Top
4956 @unnumbered Concept Index