1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename gccinstall.info
5 @settitle Installing GCC
10 @include gcc-common.texi
12 @c Specify title for specific html page
14 @settitle Installing GCC
17 @settitle Host/Target specific installation notes for GCC
19 @ifset prerequisiteshtml
20 @settitle Prerequisites for GCC
23 @settitle Downloading GCC
26 @settitle Installing GCC: Configuration
29 @settitle Installing GCC: Building
32 @settitle Installing GCC: Testing
34 @ifset finalinstallhtml
35 @settitle Installing GCC: Final installation
38 @settitle Installing GCC: Binaries
41 @settitle Installing GCC: Old documentation
44 @settitle Installing GCC: GNU Free Documentation License
47 @c Copyright (C) 1988-2018 Free Software Foundation, Inc.
48 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
50 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
51 @c test the generation of HTML documents for the gcc.gnu.org web pages.
53 @c Do not use @footnote{} in this file as it breaks install.texi2html!
55 @c Include everything if we're not making html
59 @set prerequisiteshtml
70 @c Part 2 Summary Description and Copyright
72 Copyright @copyright{} 1988-2018 Free Software Foundation, Inc.
74 Permission is granted to copy, distribute and/or modify this document
75 under the terms of the GNU Free Documentation License, Version 1.3 or
76 any later version published by the Free Software Foundation; with no
77 Invariant Sections, the Front-Cover texts being (a) (see below), and
78 with the Back-Cover Texts being (b) (see below). A copy of the
79 license is included in the section entitled ``@uref{./gfdl.html,,GNU
80 Free Documentation License}''.
82 (a) The FSF's Front-Cover Text is:
86 (b) The FSF's Back-Cover Text is:
88 You have freedom to copy and modify this GNU Manual, like GNU
89 software. Copies published by the Free Software Foundation raise
90 funds for GNU development.
95 @dircategory Software development
97 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
100 @c Part 3 Titlepage and Copyright
102 @title Installing GCC
105 @c The following two commands start the copyright page.
107 @vskip 0pt plus 1filll
111 @c Part 4 Top node, Master Menu, and/or Table of Contents
114 @comment node-name, next, Previous, up
117 * Installing GCC:: This document describes the generic installation
118 procedure for GCC as well as detailing some target
119 specific installation instructions.
121 * Specific:: Host/target specific installation notes for GCC.
122 * Binaries:: Where to get pre-compiled binaries.
124 * Old:: Old installation documentation.
126 * GNU Free Documentation License:: How you can copy and share this manual.
127 * Concept Index:: This index has two entries.
135 @c Part 5 The Body of the Document
136 @c ***Installing GCC**********************************************************
138 @comment node-name, next, previous, up
139 @node Installing GCC, Binaries, , Top
143 @chapter Installing GCC
146 The latest version of this document is always available at
147 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
148 It refers to the current development sources, instructions for
149 specific released versions are included with the sources.
151 This document describes the generic installation procedure for GCC as well
152 as detailing some target specific installation instructions.
154 GCC includes several components that previously were separate distributions
155 with their own installation instructions. This document supersedes all
156 package-specific installation instructions.
158 @emph{Before} starting the build/install procedure please check the
160 @ref{Specific, host/target specific installation notes}.
163 @uref{specific.html,,host/target specific installation notes}.
165 We recommend you browse the entire generic installation instructions before
168 Lists of successful builds for released versions of GCC are
169 available at @uref{http://gcc.gnu.org/buildstat.html}.
170 These lists are updated as new information becomes available.
172 The installation procedure itself is broken into five steps.
177 * Downloading the source::
180 * Testing:: (optional)
187 @uref{prerequisites.html,,Prerequisites}
189 @uref{download.html,,Downloading the source}
191 @uref{configure.html,,Configuration}
193 @uref{build.html,,Building}
195 @uref{test.html,,Testing} (optional)
197 @uref{finalinstall.html,,Final install}
201 Please note that GCC does not support @samp{make uninstall} and probably
202 won't do so in the near future as this would open a can of worms. Instead,
203 we suggest that you install GCC into a directory of its own and simply
204 remove that directory when you do not need that specific version of GCC
205 any longer, and, if shared libraries are installed there as well, no
206 more binaries exist that use them.
209 There are also some @uref{old.html,,old installation instructions},
210 which are mostly obsolete but still contain some information which has
211 not yet been merged into the main part of this manual.
219 @uref{./index.html,,Return to the GCC Installation page}
225 @c ***Prerequisites**************************************************
227 @comment node-name, next, previous, up
228 @node Prerequisites, Downloading the source, , Installing GCC
230 @ifset prerequisiteshtml
232 @chapter Prerequisites
234 @cindex Prerequisites
236 GCC requires that various tools and packages be available for use in the
237 build procedure. Modifying GCC sources requires additional tools
240 @heading Tools/packages necessary for building GCC
242 @item ISO C++98 compiler
243 Necessary to bootstrap GCC, although versions of GCC prior
244 to 4.8 also allow bootstrapping with a ISO C89 compiler and versions
245 of GCC prior to 3.4 also allow bootstrapping with a traditional
248 To build all languages in a cross-compiler or other configuration where
249 3-stage bootstrap is not performed, you need to start with an existing
250 GCC binary (version 3.4 or later) because source code for language
251 frontends other than C might use GCC extensions.
253 Note that to bootstrap GCC with versions of GCC earlier than 3.4, you
254 may need to use @option{--disable-stage1-checking}, though
255 bootstrapping the compiler with such earlier compilers is strongly
258 @item C standard library and headers
260 In order to build GCC, the C standard library and headers must be present
261 for all target variants for which target libraries will be built (and not
262 only the variant of the host C++ compiler).
264 This affects the popular @samp{x86_64-unknown-linux-gnu} platform (among
265 other multilib targets), for which 64-bit (@samp{x86_64}) and 32-bit
266 (@samp{i386}) libc headers are usually packaged separately. If you do a
267 build of a native compiler on @samp{x86_64-unknown-linux-gnu}, make sure you
268 either have the 32-bit libc developer package properly installed (the exact
269 name of the package depends on your distro) or you must build GCC as a
270 64-bit only compiler by configuring with the option
271 @option{--disable-multilib}. Otherwise, you may encounter an error such as
272 @samp{fatal error: gnu/stubs-32.h: No such file}
276 In order to build the Ada compiler (GNAT) you must already have GNAT
277 installed because portions of the Ada frontend are written in Ada (with
278 GNAT extensions.) Refer to the Ada installation instructions for more
279 specific information.
281 @item A ``working'' POSIX compatible shell, or GNU bash
283 Necessary when running @command{configure} because some
284 @command{/bin/sh} shells have bugs and may crash when configuring the
285 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
286 have disastrous corner-case performance problems. This
287 can cause target @command{configure} runs to literally take days to
288 complete in some cases.
290 So on some platforms @command{/bin/ksh} is sufficient, on others it
291 isn't. See the host/target specific instructions for your platform, or
292 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
293 environment to your ``good'' shell prior to running
294 @command{configure}/@command{make}.
296 @command{zsh} is not a fully compliant POSIX shell and will not
297 work when configuring GCC@.
299 @item A POSIX or SVR4 awk
301 Necessary for creating some of the generated source files for GCC@.
302 If in doubt, use a recent GNU awk version, as some of the older ones
303 are broken. GNU awk version 3.1.5 is known to work.
307 Necessary in some circumstances, optional in others. See the
308 host/target specific instructions for your platform for the exact
311 @item gzip version 1.2.4 (or later) or
312 @itemx bzip2 version 1.0.2 (or later)
314 Necessary to uncompress GCC @command{tar} files when source code is
315 obtained via FTP mirror sites.
317 @item GNU make version 3.80 (or later)
319 You must have GNU make installed to build GCC@.
321 @item GNU tar version 1.14 (or later)
323 Necessary (only on some platforms) to untar the source code. Many
324 systems' @command{tar} programs will also work, only try GNU
325 @command{tar} if you have problems.
327 @item Perl version between 5.6.1 and 5.6.24
329 Necessary when targeting Darwin, building @samp{libstdc++},
330 and not using @option{--disable-symvers}.
331 Necessary when targeting Solaris 2 with Sun @command{ld} and not using
332 @option{--disable-symvers}. The bundled @command{perl} in Solaris@tie{}8
335 Necessary when regenerating @file{Makefile} dependencies in libiberty.
336 Necessary when regenerating @file{libiberty/functions.texi}.
337 Necessary when generating manpages from Texinfo manuals.
338 Used by various scripts to generate some files included in SVN (mainly
339 Unicode-related and rarely changing) from source tables.
341 Used by @command{automake}.
345 Several support libraries are necessary to build GCC, some are required,
346 others optional. While any sufficiently new version of required tools
347 usually work, library requirements are generally stricter. Newer
348 versions may work in some cases, but it's safer to use the exact
349 versions documented. We appreciate bug reports about problems with
350 newer versions, though. If your OS vendor provides packages for the
351 support libraries then using those packages may be the simplest way to
352 install the libraries.
355 @item GNU Multiple Precision Library (GMP) version 4.3.2 (or later)
357 Necessary to build GCC@. If a GMP source distribution is found in a
358 subdirectory of your GCC sources named @file{gmp}, it will be built
359 together with GCC. Alternatively, if GMP is already installed but it
360 is not in your library search path, you will have to configure with the
361 @option{--with-gmp} configure option. See also @option{--with-gmp-lib}
362 and @option{--with-gmp-include}.
363 The in-tree build is only supported with the GMP version that
364 download_prerequisites installs.
366 @item MPFR Library version 2.4.2 (or later)
368 Necessary to build GCC@. It can be downloaded from
369 @uref{https://www.mpfr.org}. If an MPFR source distribution is found
370 in a subdirectory of your GCC sources named @file{mpfr}, it will be
371 built together with GCC. Alternatively, if MPFR is already installed
372 but it is not in your default library search path, the
373 @option{--with-mpfr} configure option should be used. See also
374 @option{--with-mpfr-lib} and @option{--with-mpfr-include}.
375 The in-tree build is only supported with the MPFR version that
376 download_prerequisites installs.
378 @item MPC Library version 0.8.1 (or later)
380 Necessary to build GCC@. It can be downloaded from
381 @uref{http://www.multiprecision.org/mpc/}. If an MPC source distribution
382 is found in a subdirectory of your GCC sources named @file{mpc}, it
383 will be built together with GCC. Alternatively, if MPC is already
384 installed but it is not in your default library search path, the
385 @option{--with-mpc} configure option should be used. See also
386 @option{--with-mpc-lib} and @option{--with-mpc-include}.
387 The in-tree build is only supported with the MPC version that
388 download_prerequisites installs.
390 @item isl Library version 0.15 or later.
392 Necessary to build GCC with the Graphite loop optimizations.
393 It can be downloaded from @uref{ftp://gcc.gnu.org/pub/gcc/infrastructure/}.
394 If an isl source distribution is found
395 in a subdirectory of your GCC sources named @file{isl}, it will be
396 built together with GCC. Alternatively, the @option{--with-isl} configure
397 option should be used if isl is not installed in your default library
402 @heading Tools/packages necessary for modifying GCC
404 @item autoconf version 2.64
405 @itemx GNU m4 version 1.4.6 (or later)
407 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
408 to regenerate @file{configure} and @file{config.in} files.
410 @item automake version 1.11.6
412 Necessary when modifying a @file{Makefile.am} file to regenerate its
413 associated @file{Makefile.in}.
415 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
416 file. Specifically this applies to the @file{gcc}, @file{intl},
417 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
418 as any of their subdirectories.
420 For directories that use automake, GCC requires the latest release in
421 the 1.11 series, which is currently 1.11.6. When regenerating a directory
422 to a newer version, please update all the directories using an older 1.11
423 to the latest released version.
425 Note that @command{automake} 1.11.6 is incompatible with
426 @command{perl} version 5.6.26.
428 @item gettext version 0.14.5 (or later)
430 Needed to regenerate @file{gcc.pot}.
432 @item gperf version 2.7.2 (or later)
434 Necessary when modifying @command{gperf} input files, e.g.@:
435 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
436 @file{gcc/cp/cfns.h}.
442 Necessary to run the GCC testsuite; see the section on testing for
443 details. Tcl 8.6 has a known regression in RE pattern handling that
444 make parts of the testsuite fail. See
445 @uref{http://core.tcl.tk/tcl/tktview/267b7e2334ee2e9de34c4b00d6e72e2f1997085f}
446 for more information. This bug has been fixed in 8.6.1.
448 @item autogen version 5.5.4 (or later) and
449 @itemx guile version 1.4.1 (or later)
451 Necessary to regenerate @file{fixinc/fixincl.x} from
452 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
454 Necessary to run @samp{make check} for @file{fixinc}.
456 Necessary to regenerate the top level @file{Makefile.in} file from
457 @file{Makefile.tpl} and @file{Makefile.def}.
459 @item Flex version 2.5.4 (or later)
461 Necessary when modifying @file{*.l} files.
463 Necessary to build GCC during development because the generated output
464 files are not included in the SVN repository. They are included in
467 @item Texinfo version 4.7 (or later)
469 Necessary for running @command{makeinfo} when modifying @file{*.texi}
470 files to test your changes.
472 Necessary for running @command{make dvi} or @command{make pdf} to
473 create printable documentation in DVI or PDF format. Texinfo version
474 4.8 or later is required for @command{make pdf}.
476 Necessary to build GCC documentation during development because the
477 generated output files are not included in the SVN repository. They are
478 included in releases.
480 @item @TeX{} (any working version)
482 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
483 are used when running @command{make dvi} or @command{make pdf} to create
484 DVI or PDF files, respectively.
486 @item Sphinx version 1.0 (or later)
488 Necessary to regenerate @file{jit/docs/_build/texinfo} from the @file{.rst}
489 files in the directories below @file{jit/docs}.
491 @item SVN (any version)
492 @itemx SSH (any version)
494 Necessary to access the SVN repository. Public releases and weekly
495 snapshots of the development sources are also available via FTP@.
497 @item GNU diffutils version 2.7 (or later)
499 Useful when submitting patches for the GCC source code.
501 @item patch version 2.5.4 (or later)
503 Necessary when applying patches, created with @command{diff}, to one's
513 @uref{./index.html,,Return to the GCC Installation page}
517 @c ***Downloading the source**************************************************
519 @comment node-name, next, previous, up
520 @node Downloading the source, Configuration, Prerequisites, Installing GCC
524 @chapter Downloading GCC
526 @cindex Downloading GCC
527 @cindex Downloading the Source
529 GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP
530 tarballs compressed with @command{gzip} or
533 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
534 for information on how to obtain GCC@.
536 The source distribution includes the C, C++, Objective-C, Fortran,
537 and Ada (in the case of GCC 3.1 and later) compilers, as well as
538 runtime libraries for C++, Objective-C, and Fortran.
539 For previous versions these were downloadable as separate components such
540 as the core GCC distribution, which included the C language front end and
541 shared components, and language-specific distributions including the
542 language front end and the language runtime (where appropriate).
544 If you also intend to build binutils (either to upgrade an existing
545 installation or for use in place of the corresponding tools of your
546 OS), unpack the binutils distribution either in the same directory or
547 a separate one. In the latter case, add symbolic links to any
548 components of the binutils you intend to build alongside the compiler
549 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
550 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
552 Likewise the GMP, MPFR and MPC libraries can be automatically built
553 together with GCC. You may simply run the
554 @command{contrib/download_prerequisites} script in the GCC source directory
555 to set up everything.
556 Otherwise unpack the GMP, MPFR and/or MPC source
557 distributions in the directory containing the GCC sources and rename
558 their directories to @file{gmp}, @file{mpfr} and @file{mpc},
559 respectively (or use symbolic links with the same name).
566 @uref{./index.html,,Return to the GCC Installation page}
570 @c ***Configuration***********************************************************
572 @comment node-name, next, previous, up
573 @node Configuration, Building, Downloading the source, Installing GCC
577 @chapter Installing GCC: Configuration
579 @cindex Configuration
580 @cindex Installing GCC: Configuration
582 Like most GNU software, GCC must be configured before it can be built.
583 This document describes the recommended configuration procedure
584 for both native and cross targets.
586 We use @var{srcdir} to refer to the toplevel source directory for
587 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
589 If you obtained the sources via SVN, @var{srcdir} must refer to the top
590 @file{gcc} directory, the one where the @file{MAINTAINERS} file can be
591 found, and not its @file{gcc} subdirectory, otherwise the build will fail.
593 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
594 file system, the shell's built-in @command{pwd} command will return
595 temporary pathnames. Using these can lead to various sorts of build
596 problems. To avoid this issue, set the @env{PWDCMD} environment
597 variable to an automounter-aware @command{pwd} command, e.g.,
598 @command{pawd} or @samp{amq -w}, during the configuration and build
601 First, we @strong{highly} recommend that GCC be built into a
602 separate directory from the sources which does @strong{not} reside
603 within the source tree. This is how we generally build GCC; building
604 where @var{srcdir} == @var{objdir} should still work, but doesn't
605 get extensive testing; building where @var{objdir} is a subdirectory
606 of @var{srcdir} is unsupported.
608 If you have previously built GCC in the same directory for a
609 different target machine, do @samp{make distclean} to delete all files
610 that might be invalid. One of the files this deletes is @file{Makefile};
611 if @samp{make distclean} complains that @file{Makefile} does not exist
612 or issues a message like ``don't know how to make distclean'' it probably
613 means that the directory is already suitably clean. However, with the
614 recommended method of building in a separate @var{objdir}, you should
615 simply use a different @var{objdir} for each target.
617 Second, when configuring a native system, either @command{cc} or
618 @command{gcc} must be in your path or you must set @env{CC} in
619 your environment before running configure. Otherwise the configuration
623 Note that the bootstrap compiler and the resulting GCC must be link
624 compatible, else the bootstrap will fail with linker errors about
625 incompatible object file formats. Several multilibed targets are
626 affected by this requirement, see
628 @ref{Specific, host/target specific installation notes}.
631 @uref{specific.html,,host/target specific installation notes}.
640 % @var{srcdir}/configure [@var{options}] [@var{target}]
643 @heading Distributor options
645 If you will be distributing binary versions of GCC, with modifications
646 to the source code, you should use the options described in this
647 section to make clear that your version contains modifications.
650 @item --with-pkgversion=@var{version}
651 Specify a string that identifies your package. You may wish
652 to include a build number or build date. This version string will be
653 included in the output of @command{gcc --version}. This suffix does
654 not replace the default version string, only the @samp{GCC} part.
656 The default value is @samp{GCC}.
658 @item --with-bugurl=@var{url}
659 Specify the URL that users should visit if they wish to report a bug.
660 You are of course welcome to forward bugs reported to you to the FSF,
661 if you determine that they are not bugs in your modifications.
663 The default value refers to the FSF's GCC bug tracker.
667 @heading Target specification
670 GCC has code to correctly determine the correct value for @var{target}
671 for nearly all native systems. Therefore, we highly recommend you do
672 not provide a configure target when configuring a native compiler.
675 @var{target} must be specified as @option{--target=@var{target}}
676 when configuring a cross compiler; examples of valid targets would be
677 m68k-elf, sh-elf, etc.
680 Specifying just @var{target} instead of @option{--target=@var{target}}
681 implies that the host defaults to @var{target}.
685 @heading Options specification
687 Use @var{options} to override several configure time options for
688 GCC@. A list of supported @var{options} follows; @samp{configure
689 --help} may list other options, but those not listed below may not
690 work and should not normally be used.
692 Note that each @option{--enable} option has a corresponding
693 @option{--disable} option and that each @option{--with} option has a
694 corresponding @option{--without} option.
697 @item --prefix=@var{dirname}
698 Specify the toplevel installation
699 directory. This is the recommended way to install the tools into a directory
700 other than the default. The toplevel installation directory defaults to
703 We @strong{highly} recommend against @var{dirname} being the same or a
704 subdirectory of @var{objdir} or vice versa. If specifying a directory
705 beneath a user's home directory tree, some shells will not expand
706 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
709 The following standard @command{autoconf} options are supported. Normally you
710 should not need to use these options.
712 @item --exec-prefix=@var{dirname}
713 Specify the toplevel installation directory for architecture-dependent
714 files. The default is @file{@var{prefix}}.
716 @item --bindir=@var{dirname}
717 Specify the installation directory for the executables called by users
718 (such as @command{gcc} and @command{g++}). The default is
719 @file{@var{exec-prefix}/bin}.
721 @item --libdir=@var{dirname}
722 Specify the installation directory for object code libraries and
723 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
725 @item --libexecdir=@var{dirname}
726 Specify the installation directory for internal executables of GCC@.
727 The default is @file{@var{exec-prefix}/libexec}.
729 @item --with-slibdir=@var{dirname}
730 Specify the installation directory for the shared libgcc library. The
731 default is @file{@var{libdir}}.
733 @item --datarootdir=@var{dirname}
734 Specify the root of the directory tree for read-only architecture-independent
735 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
737 @item --infodir=@var{dirname}
738 Specify the installation directory for documentation in info format.
739 The default is @file{@var{datarootdir}/info}.
741 @item --datadir=@var{dirname}
742 Specify the installation directory for some architecture-independent
743 data files referenced by GCC@. The default is @file{@var{datarootdir}}.
745 @item --docdir=@var{dirname}
746 Specify the installation directory for documentation files (other
747 than Info) for GCC@. The default is @file{@var{datarootdir}/doc}.
749 @item --htmldir=@var{dirname}
750 Specify the installation directory for HTML documentation files.
751 The default is @file{@var{docdir}}.
753 @item --pdfdir=@var{dirname}
754 Specify the installation directory for PDF documentation files.
755 The default is @file{@var{docdir}}.
757 @item --mandir=@var{dirname}
758 Specify the installation directory for manual pages. The default is
759 @file{@var{datarootdir}/man}. (Note that the manual pages are only extracts
760 from the full GCC manuals, which are provided in Texinfo format. The manpages
761 are derived by an automatic conversion process from parts of the full
764 @item --with-gxx-include-dir=@var{dirname}
766 the installation directory for G++ header files. The default depends
767 on other configuration options, and differs between cross and native
770 @item --with-specs=@var{specs}
771 Specify additional command line driver SPECS.
772 This can be useful if you need to turn on a non-standard feature by
773 default without modifying the compiler's source code, for instance
774 @option{--with-specs=%@{!fcommon:%@{!fno-common:-fno-common@}@}}.
776 @xref{Spec Files,, Specifying subprocesses and the switches to pass to them,
777 gcc, Using the GNU Compiler Collection (GCC)},
780 See ``Spec Files'' in the main manual
785 @item --program-prefix=@var{prefix}
786 GCC supports some transformations of the names of its programs when
787 installing them. This option prepends @var{prefix} to the names of
788 programs to install in @var{bindir} (see above). For example, specifying
789 @option{--program-prefix=foo-} would result in @samp{gcc}
790 being installed as @file{/usr/local/bin/foo-gcc}.
792 @item --program-suffix=@var{suffix}
793 Appends @var{suffix} to the names of programs to install in @var{bindir}
794 (see above). For example, specifying @option{--program-suffix=-3.1}
795 would result in @samp{gcc} being installed as
796 @file{/usr/local/bin/gcc-3.1}.
798 @item --program-transform-name=@var{pattern}
799 Applies the @samp{sed} script @var{pattern} to be applied to the names
800 of programs to install in @var{bindir} (see above). @var{pattern} has to
801 consist of one or more basic @samp{sed} editing commands, separated by
802 semicolons. For example, if you want the @samp{gcc} program name to be
803 transformed to the installed program @file{/usr/local/bin/myowngcc} and
804 the @samp{g++} program name to be transformed to
805 @file{/usr/local/bin/gspecial++} without changing other program names,
806 you could use the pattern
807 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
808 to achieve this effect.
810 All three options can be combined and used together, resulting in more
811 complex conversion patterns. As a basic rule, @var{prefix} (and
812 @var{suffix}) are prepended (appended) before further transformations
813 can happen with a special transformation script @var{pattern}.
815 As currently implemented, this option only takes effect for native
816 builds; cross compiler binaries' names are not transformed even when a
817 transformation is explicitly asked for by one of these options.
819 For native builds, some of the installed programs are also installed
820 with the target alias in front of their name, as in
821 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
822 before the target alias is prepended to the name---so, specifying
823 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
824 resulting binary would be installed as
825 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
827 As a last shortcoming, none of the installed Ada programs are
828 transformed yet, which will be fixed in some time.
830 @item --with-local-prefix=@var{dirname}
832 installation directory for local include files. The default is
833 @file{/usr/local}. Specify this option if you want the compiler to
834 search directory @file{@var{dirname}/include} for locally installed
835 header files @emph{instead} of @file{/usr/local/include}.
837 You should specify @option{--with-local-prefix} @strong{only} if your
838 site has a different convention (not @file{/usr/local}) for where to put
841 The default value for @option{--with-local-prefix} is @file{/usr/local}
842 regardless of the value of @option{--prefix}. Specifying
843 @option{--prefix} has no effect on which directory GCC searches for
844 local header files. This may seem counterintuitive, but actually it is
847 The purpose of @option{--prefix} is to specify where to @emph{install
848 GCC}. The local header files in @file{/usr/local/include}---if you put
849 any in that directory---are not part of GCC@. They are part of other
850 programs---perhaps many others. (GCC installs its own header files in
851 another directory which is based on the @option{--prefix} value.)
853 Both the local-prefix include directory and the GCC-prefix include
854 directory are part of GCC's ``system include'' directories. Although these
855 two directories are not fixed, they need to be searched in the proper
856 order for the correct processing of the include_next directive. The
857 local-prefix include directory is searched before the GCC-prefix
858 include directory. Another characteristic of system include directories
859 is that pedantic warnings are turned off for headers in these directories.
861 Some autoconf macros add @option{-I @var{directory}} options to the
862 compiler command line, to ensure that directories containing installed
863 packages' headers are searched. When @var{directory} is one of GCC's
864 system include directories, GCC will ignore the option so that system
865 directories continue to be processed in the correct order. This
866 may result in a search order different from what was specified but the
867 directory will still be searched.
869 GCC automatically searches for ordinary libraries using
870 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
871 used for both GCC and packages, GCC will automatically search for
872 both headers and libraries. This provides a configuration that is
873 easy to use. GCC behaves in a manner similar to that when it is
874 installed as a system compiler in @file{/usr}.
876 Sites that need to install multiple versions of GCC may not want to
877 use the above simple configuration. It is possible to use the
878 @option{--program-prefix}, @option{--program-suffix} and
879 @option{--program-transform-name} options to install multiple versions
880 into a single directory, but it may be simpler to use different prefixes
881 and the @option{--with-local-prefix} option to specify the location of the
882 site-specific files for each version. It will then be necessary for
883 users to specify explicitly the location of local site libraries
884 (e.g., with @env{LIBRARY_PATH}).
886 The same value can be used for both @option{--with-local-prefix} and
887 @option{--prefix} provided it is not @file{/usr}. This can be used
888 to avoid the default search of @file{/usr/local/include}.
890 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
891 The directory you use for @option{--with-local-prefix} @strong{must not}
892 contain any of the system's standard header files. If it did contain
893 them, certain programs would be miscompiled (including GNU Emacs, on
894 certain targets), because this would override and nullify the header
895 file corrections made by the @command{fixincludes} script.
897 Indications are that people who use this option use it based on mistaken
898 ideas of what it is for. People use it as if it specified where to
899 install part of GCC@. Perhaps they make this assumption because
900 installing GCC creates the directory.
902 @item --with-gcc-major-version-only
903 Specifies that GCC should use only the major number rather than
904 @var{major}.@var{minor}.@var{patchlevel} in filesystem paths.
906 @item --with-native-system-header-dir=@var{dirname}
907 Specifies that @var{dirname} is the directory that contains native system
908 header files, rather than @file{/usr/include}. This option is most useful
909 if you are creating a compiler that should be isolated from the system
910 as much as possible. It is most commonly used with the
911 @option{--with-sysroot} option and will cause GCC to search
912 @var{dirname} inside the system root specified by that option.
914 @item --enable-shared[=@var{package}[,@dots{}]]
915 Build shared versions of libraries, if shared libraries are supported on
916 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
917 are enabled by default on all platforms that support shared libraries.
919 If a list of packages is given as an argument, build shared libraries
920 only for the listed packages. For other packages, only static libraries
921 will be built. Package names currently recognized in the GCC tree are
922 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
923 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
924 @samp{ada}, @samp{libada}, @samp{libgo}, @samp{libobjc}, and @samp{libphobos}.
925 Note @samp{libiberty} does not support shared libraries at all.
927 Use @option{--disable-shared} to build only static libraries. Note that
928 @option{--disable-shared} does not accept a list of package names as
929 argument, only @option{--enable-shared} does.
931 Contrast with @option{--enable-host-shared}, which affects @emph{host}
934 @item --enable-host-shared
935 Specify that the @emph{host} code should be built into position-independent
936 machine code (with -fPIC), allowing it to be used within shared libraries,
937 but yielding a slightly slower compiler.
939 This option is required when building the libgccjit.so library.
941 Contrast with @option{--enable-shared}, which affects @emph{target}
944 @item @anchor{with-gnu-as}--with-gnu-as
945 Specify that the compiler should assume that the
946 assembler it finds is the GNU assembler. However, this does not modify
947 the rules to find an assembler and will result in confusion if the
948 assembler found is not actually the GNU assembler. (Confusion may also
949 result if the compiler finds the GNU assembler but has not been
950 configured with @option{--with-gnu-as}.) If you have more than one
951 assembler installed on your system, you may want to use this option in
952 connection with @option{--with-as=@var{pathname}} or
953 @option{--with-build-time-tools=@var{pathname}}.
955 The following systems are the only ones where it makes a difference
956 whether you use the GNU assembler. On any other system,
957 @option{--with-gnu-as} has no effect.
960 @item @samp{hppa1.0-@var{any}-@var{any}}
961 @item @samp{hppa1.1-@var{any}-@var{any}}
962 @item @samp{sparc-sun-solaris2.@var{any}}
963 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
966 @item @anchor{with-as}--with-as=@var{pathname}
967 Specify that the compiler should use the assembler pointed to by
968 @var{pathname}, rather than the one found by the standard rules to find
969 an assembler, which are:
972 Unless GCC is being built with a cross compiler, check the
973 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
974 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
975 @var{exec-prefix} defaults to @var{prefix}, which
976 defaults to @file{/usr/local} unless overridden by the
977 @option{--prefix=@var{pathname}} switch described above. @var{target}
978 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
979 @var{version} denotes the GCC version, such as 3.0.
982 If the target system is the same that you are building on, check
983 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
987 Check in the @env{PATH} for a tool whose name is prefixed by the
988 target system triple.
991 Check in the @env{PATH} for a tool whose name is not prefixed by the
992 target system triple, if the host and target system triple are
993 the same (in other words, we use a host tool if it can be used for
997 You may want to use @option{--with-as} if no assembler
998 is installed in the directories listed above, or if you have multiple
999 assemblers installed and want to choose one that is not found by the
1002 @item @anchor{with-gnu-ld}--with-gnu-ld
1003 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
1006 @item --with-ld=@var{pathname}
1007 Same as @uref{#with-as,,@option{--with-as}}
1011 Specify that stabs debugging
1012 information should be used instead of whatever format the host normally
1013 uses. Normally GCC uses the same debug format as the host system.
1015 @item --with-tls=@var{dialect}
1016 Specify the default TLS dialect, for systems were there is a choice.
1017 For ARM targets, possible values for @var{dialect} are @code{gnu} or
1018 @code{gnu2}, which select between the original GNU dialect and the GNU TLS
1019 descriptor-based dialect.
1021 @item --enable-multiarch
1022 Specify whether to enable or disable multiarch support. The default is
1023 to check for glibc start files in a multiarch location, and enable it
1024 if the files are found. The auto detection is enabled for native builds,
1025 and for cross builds configured with @option{--with-sysroot}, and without
1026 @option{--with-native-system-header-dir}.
1027 More documentation about multiarch can be found at
1028 @uref{https://wiki.debian.org/Multiarch}.
1030 @item --enable-sjlj-exceptions
1031 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1032 @samp{configure} ordinarily picks the correct value based on the platform.
1033 Only use this option if you are sure you need a different setting.
1035 @item --enable-vtable-verify
1036 Specify whether to enable or disable the vtable verification feature.
1037 Enabling this feature causes libstdc++ to be built with its virtual calls
1038 in verifiable mode. This means that, when linked with libvtv, every
1039 virtual call in libstdc++ will verify the vtable pointer through which the
1040 call will be made before actually making the call. If not linked with libvtv,
1041 the verifier will call stub functions (in libstdc++ itself) and do nothing.
1042 If vtable verification is disabled, then libstdc++ is not built with its
1043 virtual calls in verifiable mode at all. However the libvtv library will
1044 still be built (see @option{--disable-libvtv} to turn off building libvtv).
1045 @option{--disable-vtable-verify} is the default.
1047 @item --disable-gcov
1048 Specify that the run-time library used for coverage analysis
1049 and associated host tools should not be built.
1051 @item --disable-multilib
1052 Specify that multiple target
1053 libraries to support different target variants, calling
1054 conventions, etc.@: should not be built. The default is to build a
1055 predefined set of them.
1057 Some targets provide finer-grained control over which multilibs are built
1058 (e.g., @option{--disable-softfloat}):
1061 fpu, 26bit, underscore, interwork, biendian, nofmult.
1064 softfloat, m68881, m68000, m68020.
1067 single-float, biendian, softfloat.
1069 @item powerpc*-*-*, rs6000*-*-*
1070 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
1075 @item --with-multilib-list=@var{list}
1076 @itemx --without-multilib-list
1077 Specify what multilibs to build. @var{list} is a comma separated list of
1078 values, possibly consisting of a single value. Currently only implemented
1079 for arm*-*-*, riscv*-*-*, sh*-*-* and x86-64-*-linux*. The accepted
1080 values and meaning for each target is given below.
1084 @var{list} is a comma separated list of @code{aprofile} and @code{rmprofile}
1085 to build multilibs for A or R and M architecture profiles respectively. Note
1086 that, due to some limitation of the current multilib framework, using the
1087 combined @code{aprofile,rmprofile} multilibs selects in some cases a less
1088 optimal multilib than when using the multilib profile for the architecture
1089 targetted. The special value @code{default} is also accepted and is equivalent
1090 to omitting the option, ie. only the default run-time library will be enabled.
1092 The table below gives the combination of ISAs, architectures, FPUs and
1093 floating-point ABIs for which multilibs are built for each accepted value.
1094 The union of these options is considered when specifying both @code{aprofile}
1095 and @code{rmprofile}.
1097 @multitable @columnfractions .15 .28 .30
1098 @item Option @tab aprofile @tab rmprofile
1100 @tab @code{-marm} and @code{-mthumb}
1102 @item Architectures@*@*@*@*@*@*
1103 @tab default architecture@*
1104 @code{-march=armv7-a}@*
1105 @code{-march=armv7ve}@*
1106 @code{-march=armv8-a}@*@*@*
1107 @tab default architecture@*
1108 @code{-march=armv6s-m}@*
1109 @code{-march=armv7-m}@*
1110 @code{-march=armv7e-m}@*
1111 @code{-march=armv8-m.base}@*
1112 @code{-march=armv8-m.main}@*
1114 @item FPUs@*@*@*@*@*
1116 @code{-mfpu=vfpv3-d16}@*
1118 @code{-mfpu=vfpv4-d16}@*
1119 @code{-mfpu=neon-vfpv4}@*
1120 @code{-mfpu=neon-fp-armv8}
1122 @code{-mfpu=vfpv3-d16}@*
1123 @code{-mfpu=fpv4-sp-d16}@*
1124 @code{-mfpu=fpv5-sp-d16}@*
1125 @code{-mfpu=fpv5-d16}@*
1126 @item floating-point@/ ABIs@*@*
1127 @tab @code{-mfloat-abi=soft}@*
1128 @code{-mfloat-abi=softfp}@*
1129 @code{-mfloat-abi=hard}
1130 @tab @code{-mfloat-abi=soft}@*
1131 @code{-mfloat-abi=softfp}@*
1132 @code{-mfloat-abi=hard}
1136 @var{list} is a single ABI name. The target architecture must be either
1137 @code{rv32gc} or @code{rv64gc}. This will build a single multilib for the
1138 specified architecture and ABI pair. If @code{--with-multilib-list} is not
1139 given, then a default set of multilibs is selected based on the value of
1140 @option{--target}. This is usually a large set of multilibs.
1143 @var{list} is a comma separated list of CPU names. These must be of the
1144 form @code{sh*} or @code{m*} (in which case they match the compiler option
1145 for that processor). The list should not contain any endian options -
1146 these are handled by @option{--with-endian}.
1148 If @var{list} is empty, then there will be no multilibs for extra
1149 processors. The multilib for the secondary endian remains enabled.
1151 As a special case, if an entry in the list starts with a @code{!}
1152 (exclamation point), then it is added to the list of excluded multilibs.
1153 Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES}
1154 (once the leading @code{!} has been stripped).
1156 If @option{--with-multilib-list} is not given, then a default set of
1157 multilibs is selected based on the value of @option{--target}. This is
1158 usually the complete set of libraries, but some targets imply a more
1161 Example 1: to configure a compiler for SH4A only, but supporting both
1162 endians, with little endian being the default:
1164 --with-cpu=sh4a --with-endian=little,big --with-multilib-list=
1167 Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with
1168 only little endian SH4AL:
1170 --with-cpu=sh4a --with-endian=little,big \
1171 --with-multilib-list=sh4al,!mb/m4al
1174 @item x86-64-*-linux*
1175 @var{list} is a comma separated list of @code{m32}, @code{m64} and
1176 @code{mx32} to enable 32-bit, 64-bit and x32 run-time libraries,
1177 respectively. If @var{list} is empty, then there will be no multilibs
1178 and only the default run-time library will be enabled.
1180 If @option{--with-multilib-list} is not given, then only 32-bit and
1181 64-bit run-time libraries will be enabled.
1184 @item --with-endian=@var{endians}
1185 Specify what endians to use.
1186 Currently only implemented for sh*-*-*.
1188 @var{endians} may be one of the following:
1191 Use big endian exclusively.
1193 Use little endian exclusively.
1195 Use big endian by default. Provide a multilib for little endian.
1197 Use little endian by default. Provide a multilib for big endian.
1200 @item --enable-threads
1201 Specify that the target
1202 supports threads. This affects the Objective-C compiler and runtime
1203 library, and exception handling for other languages like C++.
1204 On some systems, this is the default.
1206 In general, the best (and, in many cases, the only known) threading
1207 model available will be configured for use. Beware that on some
1208 systems, GCC has not been taught what threading models are generally
1209 available for the system. In this case, @option{--enable-threads} is an
1210 alias for @option{--enable-threads=single}.
1212 @item --disable-threads
1213 Specify that threading support should be disabled for the system.
1214 This is an alias for @option{--enable-threads=single}.
1216 @item --enable-threads=@var{lib}
1218 @var{lib} is the thread support library. This affects the Objective-C
1219 compiler and runtime library, and exception handling for other languages
1220 like C++. The possibilities for @var{lib} are:
1228 LynxOS thread support.
1230 MIPS SDE thread support.
1232 This is an alias for @samp{single}.
1234 Generic POSIX/Unix98 thread support.
1236 RTEMS thread support.
1238 Disable thread support, should work for all platforms.
1242 VxWorks thread support.
1244 Microsoft Win32 API thread support.
1248 Specify that the target supports TLS (Thread Local Storage). Usually
1249 configure can correctly determine if TLS is supported. In cases where
1250 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1251 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1252 the assembler supports TLS but the C library does not, or if the
1253 assumptions made by the configure test are incorrect.
1256 Specify that the target does not support TLS.
1257 This is an alias for @option{--enable-tls=no}.
1259 @item --with-cpu=@var{cpu}
1260 @itemx --with-cpu-32=@var{cpu}
1261 @itemx --with-cpu-64=@var{cpu}
1262 Specify which cpu variant the compiler should generate code for by default.
1263 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1264 This option is only supported on some targets, including ARC, ARM, i386, M68k,
1265 PowerPC, and SPARC@. It is mandatory for ARC@. The @option{--with-cpu-32} and
1266 @option{--with-cpu-64} options specify separate default CPUs for
1267 32-bit and 64-bit modes; these options are only supported for i386,
1268 x86-64, PowerPC, and SPARC@.
1270 @item --with-schedule=@var{cpu}
1271 @itemx --with-arch=@var{cpu}
1272 @itemx --with-arch-32=@var{cpu}
1273 @itemx --with-arch-64=@var{cpu}
1274 @itemx --with-tune=@var{cpu}
1275 @itemx --with-tune-32=@var{cpu}
1276 @itemx --with-tune-64=@var{cpu}
1277 @itemx --with-abi=@var{abi}
1278 @itemx --with-fpu=@var{type}
1279 @itemx --with-float=@var{type}
1280 These configure options provide default values for the @option{-mschedule=},
1281 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1282 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1283 @option{--with-cpu}, which switches will be accepted and acceptable values
1284 of the arguments depend on the target.
1286 @item --with-mode=@var{mode}
1287 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1288 This option is only supported on ARM targets.
1290 @item --with-stack-offset=@var{num}
1291 This option sets the default for the -mstack-offset=@var{num} option,
1292 and will thus generally also control the setting of this option for
1293 libraries. This option is only supported on Epiphany targets.
1295 @item --with-fpmath=@var{isa}
1296 This options sets @option{-mfpmath=sse} by default and specifies the default
1297 ISA for floating-point arithmetics. You can select either @samp{sse} which
1298 enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default.
1299 This option is only supported on i386 and x86-64 targets.
1301 @item --with-fp-32=@var{mode}
1302 On MIPS targets, set the default value for the @option{-mfp} option when using
1303 the o32 ABI. The possibilities for @var{mode} are:
1306 Use the o32 FP32 ABI extension, as with the @option{-mfp32} command-line
1309 Use the o32 FPXX ABI extension, as with the @option{-mfpxx} command-line
1312 Use the o32 FP64 ABI extension, as with the @option{-mfp64} command-line
1315 In the absence of this configuration option the default is to use the o32
1318 @item --with-odd-spreg-32
1319 On MIPS targets, set the @option{-modd-spreg} option by default when using
1322 @item --without-odd-spreg-32
1323 On MIPS targets, set the @option{-mno-odd-spreg} option by default when using
1324 the o32 ABI. This is normally used in conjunction with
1325 @option{--with-fp-32=64} in order to target the o32 FP64A ABI extension.
1327 @item --with-nan=@var{encoding}
1328 On MIPS targets, set the default encoding convention to use for the
1329 special not-a-number (NaN) IEEE 754 floating-point data. The
1330 possibilities for @var{encoding} are:
1333 Use the legacy encoding, as with the @option{-mnan=legacy} command-line
1336 Use the 754-2008 encoding, as with the @option{-mnan=2008} command-line
1339 To use this configuration option you must have an assembler version
1340 installed that supports the @option{-mnan=} command-line option too.
1341 In the absence of this configuration option the default convention is
1342 the legacy encoding, as when neither of the @option{-mnan=2008} and
1343 @option{-mnan=legacy} command-line options has been used.
1345 @item --with-divide=@var{type}
1346 Specify how the compiler should generate code for checking for
1347 division by zero. This option is only supported on the MIPS target.
1348 The possibilities for @var{type} are:
1351 Division by zero checks use conditional traps (this is the default on
1352 systems that support conditional traps).
1354 Division by zero checks use the break instruction.
1357 @c If you make --with-llsc the default for additional targets,
1358 @c update the --with-llsc description in the MIPS section below.
1361 On MIPS targets, make @option{-mllsc} the default when no
1362 @option{-mno-llsc} option is passed. This is the default for
1363 Linux-based targets, as the kernel will emulate them if the ISA does
1366 @item --without-llsc
1367 On MIPS targets, make @option{-mno-llsc} the default when no
1368 @option{-mllsc} option is passed.
1371 On MIPS targets, make @option{-msynci} the default when no
1372 @option{-mno-synci} option is passed.
1374 @item --without-synci
1375 On MIPS targets, make @option{-mno-synci} the default when no
1376 @option{-msynci} option is passed. This is the default.
1378 @item --with-lxc1-sxc1
1379 On MIPS targets, make @option{-mlxc1-sxc1} the default when no
1380 @option{-mno-lxc1-sxc1} option is passed. This is the default.
1382 @item --without-lxc1-sxc1
1383 On MIPS targets, make @option{-mno-lxc1-sxc1} the default when no
1384 @option{-mlxc1-sxc1} option is passed. The indexed load/store
1385 instructions are not directly a problem but can lead to unexpected
1386 behaviour when deployed in an application intended for a 32-bit address
1387 space but run on a 64-bit processor. The issue is seen because all
1388 known MIPS 64-bit Linux kernels execute o32 and n32 applications
1389 with 64-bit addressing enabled which affects the overflow behaviour
1390 of the indexed addressing mode. GCC will assume that ordinary
1391 32-bit arithmetic overflow behaviour is the same whether performed
1392 as an @code{addu} instruction or as part of the address calculation
1393 in @code{lwxc1} type instructions. This assumption holds true in a
1394 pure 32-bit environment and can hold true in a 64-bit environment if
1395 the address space is accurately set to be 32-bit for o32 and n32.
1398 On MIPS targets, make @option{-mmadd4} the default when no
1399 @option{-mno-madd4} option is passed. This is the default.
1401 @item --without-madd4
1402 On MIPS targets, make @option{-mno-madd4} the default when no
1403 @option{-mmadd4} option is passed. The @code{madd4} instruction
1404 family can be problematic when targeting a combination of cores that
1405 implement these instructions differently. There are two known cores
1406 that implement these as fused operations instead of unfused (where
1407 unfused is normally expected). Disabling these instructions is the
1408 only way to ensure compatible code is generated; this will incur
1409 a performance penalty.
1411 @item --with-mips-plt
1412 On MIPS targets, make use of copy relocations and PLTs.
1413 These features are extensions to the traditional
1414 SVR4-based MIPS ABIs and require support from GNU binutils
1415 and the runtime C library.
1417 @item --with-stack-clash-protection-guard-size=@var{size}
1418 On certain targets this option sets the default stack clash protection guard
1419 size as a power of two in bytes. On AArch64 @var{size} is required to be either
1420 12 (4KB) or 16 (64KB).
1422 @item --enable-__cxa_atexit
1423 Define if you want to use __cxa_atexit, rather than atexit, to
1424 register C++ destructors for local statics and global objects.
1425 This is essential for fully standards-compliant handling of
1426 destructors, but requires __cxa_atexit in libc. This option is currently
1427 only available on systems with GNU libc. When enabled, this will cause
1428 @option{-fuse-cxa-atexit} to be passed by default.
1430 @item --enable-gnu-indirect-function
1431 Define if you want to enable the @code{ifunc} attribute. This option is
1432 currently only available on systems with GNU libc on certain targets.
1434 @item --enable-target-optspace
1436 libraries should be optimized for code space instead of code speed.
1437 This is the default for the m32r platform.
1439 @item --with-cpp-install-dir=@var{dirname}
1440 Specify that the user visible @command{cpp} program should be installed
1441 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1443 @item --enable-comdat
1444 Enable COMDAT group support. This is primarily used to override the
1445 automatically detected value.
1447 @item --enable-initfini-array
1448 Force the use of sections @code{.init_array} and @code{.fini_array}
1449 (instead of @code{.init} and @code{.fini}) for constructors and
1450 destructors. Option @option{--disable-initfini-array} has the
1451 opposite effect. If neither option is specified, the configure script
1452 will try to guess whether the @code{.init_array} and
1453 @code{.fini_array} sections are supported and, if they are, use them.
1455 @item --enable-link-mutex
1456 When building GCC, use a mutex to avoid linking the compilers for
1457 multiple languages at the same time, to avoid thrashing on build
1458 systems with limited free memory. The default is not to use such a mutex.
1460 @item --enable-maintainer-mode
1461 The build rules that regenerate the Autoconf and Automake output files as
1462 well as the GCC master message catalog @file{gcc.pot} are normally
1463 disabled. This is because it can only be rebuilt if the complete source
1464 tree is present. If you have changed the sources and want to rebuild the
1465 catalog, configuring with @option{--enable-maintainer-mode} will enable
1466 this. Note that you need a recent version of the @code{gettext} tools
1469 @item --disable-bootstrap
1470 For a native build, the default configuration is to perform
1471 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1472 testing that GCC can compile itself correctly. If you want to disable
1473 this process, you can configure with @option{--disable-bootstrap}.
1475 @item --enable-bootstrap
1476 In special cases, you may want to perform a 3-stage build
1477 even if the target and host triplets are different.
1478 This is possible when the host can run code compiled for
1479 the target (e.g.@: host is i686-linux, target is i486-linux).
1480 Starting from GCC 4.2, to do this you have to configure explicitly
1481 with @option{--enable-bootstrap}.
1483 @item --enable-generated-files-in-srcdir
1484 Neither the .c and .h files that are generated from Bison and flex nor the
1485 info manuals and man pages that are built from the .texi files are present
1486 in the SVN development tree. When building GCC from that development tree,
1487 or from one of our snapshots, those generated files are placed in your
1488 build directory, which allows for the source to be in a readonly
1491 If you configure with @option{--enable-generated-files-in-srcdir} then those
1492 generated files will go into the source directory. This is mainly intended
1493 for generating release or prerelease tarballs of the GCC sources, since it
1494 is not a requirement that the users of source releases to have flex, Bison,
1497 @item --enable-version-specific-runtime-libs
1499 that runtime libraries should be installed in the compiler specific
1500 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1501 addition, @samp{libstdc++}'s include files will be installed into
1502 @file{@var{libdir}} unless you overruled it by using
1503 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1504 particularly useful if you intend to use several versions of GCC in
1505 parallel. This is currently supported by @samp{libgfortran},
1506 @samp{libstdc++}, and @samp{libobjc}.
1508 @item @anchor{WithAixSoname}--with-aix-soname=@samp{aix}, @samp{svr4} or @samp{both}
1509 Traditional AIX shared library versioning (versioned @code{Shared Object}
1510 files as members of unversioned @code{Archive Library} files named
1511 @samp{lib.a}) causes numerous headaches for package managers. However,
1512 @code{Import Files} as members of @code{Archive Library} files allow for
1513 @strong{filename-based versioning} of shared libraries as seen on Linux/SVR4,
1514 where this is called the "SONAME". But as they prevent static linking,
1515 @code{Import Files} may be used with @code{Runtime Linking} only, where the
1516 linker does search for @samp{libNAME.so} before @samp{libNAME.a} library
1517 filenames with the @samp{-lNAME} linker flag.
1519 @anchor{AixLdCommand}For detailed information please refer to the AIX
1520 @uref{https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22,,ld
1523 As long as shared library creation is enabled, upon:
1525 @item --with-aix-soname=aix
1526 @item --with-aix-soname=both
1527 A (traditional AIX) @code{Shared Archive Library} file is created:
1529 @item using the @samp{libNAME.a} filename scheme
1530 @item with the @code{Shared Object} file as archive member named
1531 @samp{libNAME.so.V} (except for @samp{libgcc_s}, where the @code{Shared
1532 Object} file is named @samp{shr.o} for backwards compatibility), which
1534 @item is used for runtime loading from inside the @samp{libNAME.a} file
1535 @item is used for dynamic loading via
1536 @code{dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)}
1537 @item is used for shared linking
1538 @item is used for static linking, so no separate @code{Static Archive
1539 Library} file is needed
1542 @item --with-aix-soname=both
1543 @item --with-aix-soname=svr4
1544 A (second) @code{Shared Archive Library} file is created:
1546 @item using the @samp{libNAME.so.V} filename scheme
1547 @item with the @code{Shared Object} file as archive member named
1550 @item is created with the @code{-G linker flag}
1551 @item has the @code{F_LOADONLY} flag set
1552 @item is used for runtime loading from inside the @samp{libNAME.so.V} file
1553 @item is used for dynamic loading via @code{dlopen("libNAME.so.V(shr.o)",
1556 @item with the @code{Import File} as archive member named @samp{shr.imp},
1559 @item refers to @samp{libNAME.so.V(shr.o)} as the "SONAME", to be recorded
1560 in the @code{Loader Section} of subsequent binaries
1561 @item indicates whether @samp{libNAME.so.V(shr.o)} is 32 or 64 bit
1562 @item lists all the public symbols exported by @samp{lib.so.V(shr.o)},
1563 eventually decorated with the @code{@samp{weak} Keyword}
1564 @item is necessary for shared linking against @samp{lib.so.V(shr.o)}
1567 A symbolic link using the @samp{libNAME.so} filename scheme is created:
1569 @item pointing to the @samp{libNAME.so.V} @code{Shared Archive Library} file
1570 @item to permit the @code{ld Command} to find @samp{lib.so.V(shr.imp)} via
1571 the @samp{-lNAME} argument (requires @code{Runtime Linking} to be enabled)
1572 @item to permit dynamic loading of @samp{lib.so.V(shr.o)} without the need
1573 to specify the version number via @code{dlopen("libNAME.so(shr.o)",
1578 As long as static library creation is enabled, upon:
1580 @item --with-aix-soname=svr4
1581 A @code{Static Archive Library} is created:
1583 @item using the @samp{libNAME.a} filename scheme
1584 @item with all the @code{Static Object} files as archive members, which
1586 @item are used for static linking
1591 While the aix-soname=@samp{svr4} option does not create @code{Shared Object}
1592 files as members of unversioned @code{Archive Library} files any more, package
1593 managers still are responsible to
1594 @uref{./specific.html#TransferAixShobj,,transfer} @code{Shared Object} files
1595 found as member of a previously installed unversioned @code{Archive Library}
1596 file into the newly installed @code{Archive Library} file with the same
1599 @emph{WARNING:} Creating @code{Shared Object} files with @code{Runtime Linking}
1600 enabled may bloat the TOC, eventually leading to @code{TOC overflow} errors,
1601 requiring the use of either the @option{-Wl,-bbigtoc} linker flag (seen to
1602 break with the @code{GDB} debugger) or some of the TOC-related compiler flags,
1604 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1605 Using the GNU Compiler Collection (GCC)}.
1608 see ``RS/6000 and PowerPC Options'' in the main manual.
1611 @option{--with-aix-soname} is currently supported by @samp{libgcc_s} only, so
1612 this option is still experimental and not for normal use yet.
1614 Default is the traditional behavior @option{--with-aix-soname=@samp{aix}}.
1616 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1617 Specify that only a particular subset of compilers and
1618 their runtime libraries should be built. For a list of valid values for
1619 @var{langN} you can issue the following command in the
1620 @file{gcc} directory of your GCC source tree:@*
1622 grep ^language= */config-lang.in
1624 Currently, you can use any of the following:
1625 @code{all}, @code{default}, @code{ada}, @code{c}, @code{c++}, @code{d},
1626 @code{fortran}, @code{go}, @code{jit}, @code{lto}, @code{objc}, @code{obj-c++}.
1627 Building the Ada compiler has special requirements, see below.
1628 If you do not pass this flag, or specify the option @code{default}, then the
1629 default languages available in the @file{gcc} sub-tree will be configured.
1630 Ada, D, Go, Jit, and Objective-C++ are not default languages. LTO is not a
1631 default language, but is built by default because @option{--enable-lto} is
1632 enabled by default. The other languages are default languages. If
1633 @code{all} is specified, then all available languages are built. An
1634 exception is @code{jit} language, which requires
1635 @option{--enable-host-shared} to be included with @code{all}.
1637 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1638 Specify that a particular subset of compilers and their runtime
1639 libraries should be built with the system C compiler during stage 1 of
1640 the bootstrap process, rather than only in later stages with the
1641 bootstrapped C compiler. The list of valid values is the same as for
1642 @option{--enable-languages}, and the option @code{all} will select all
1643 of the languages enabled by @option{--enable-languages}. This option is
1644 primarily useful for GCC development; for instance, when a development
1645 version of the compiler cannot bootstrap due to compiler bugs, or when
1646 one is debugging front ends other than the C front end. When this
1647 option is used, one can then build the target libraries for the
1648 specified languages with the stage-1 compiler by using @command{make
1649 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1650 for the specified languages using @command{make stage1-start check-gcc}.
1652 @item --disable-libada
1653 Specify that the run-time libraries and tools used by GNAT should not
1654 be built. This can be useful for debugging, or for compatibility with
1655 previous Ada build procedures, when it was required to explicitly
1656 do a @samp{make -C gcc gnatlib_and_tools}.
1658 @item --disable-libsanitizer
1659 Specify that the run-time libraries for the various sanitizers should
1662 @item --disable-libssp
1663 Specify that the run-time libraries for stack smashing protection
1664 should not be built or linked against. On many targets library support
1665 is provided by the C library instead.
1667 @item --disable-libquadmath
1668 Specify that the GCC quad-precision math library should not be built.
1669 On some systems, the library is required to be linkable when building
1670 the Fortran front end, unless @option{--disable-libquadmath-support}
1673 @item --disable-libquadmath-support
1674 Specify that the Fortran front end and @code{libgfortran} do not add
1675 support for @code{libquadmath} on systems supporting it.
1677 @item --disable-libgomp
1678 Specify that the GNU Offloading and Multi Processing Runtime Library
1679 should not be built.
1681 @item --disable-libvtv
1682 Specify that the run-time libraries used by vtable verification
1683 should not be built.
1686 Specify that the compiler should
1687 use DWARF 2 debugging information as the default.
1689 @item --with-advance-toolchain=@var{at}
1690 On 64-bit PowerPC Linux systems, configure the compiler to use the
1691 header files, library files, and the dynamic linker from the Advance
1692 Toolchain release @var{at} instead of the default versions that are
1693 provided by the Linux distribution. In general, this option is
1694 intended for the developers of GCC, and it is not intended for general
1697 @item --enable-targets=all
1698 @itemx --enable-targets=@var{target_list}
1699 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1700 These are compilers that are able to generate either 64-bit or 32-bit
1701 code. Typically, the corresponding 32-bit target, e.g.@:
1702 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1703 option enables the 32-bit target to be a bi-arch compiler, which is
1704 useful when you want a bi-arch compiler that defaults to 32-bit, and
1705 you are building a bi-arch or multi-arch binutils in a combined tree.
1706 On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64),
1708 Currently, this option only affects sparc-linux, powerpc-linux, x86-linux,
1709 mips-linux and s390-linux.
1711 @item --enable-default-pie
1712 Turn on @option{-fPIE} and @option{-pie} by default.
1714 @item --enable-secureplt
1715 This option enables @option{-msecure-plt} by default for powerpc-linux.
1717 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1718 Using the GNU Compiler Collection (GCC)},
1721 See ``RS/6000 and PowerPC Options'' in the main manual
1724 @item --enable-default-ssp
1725 Turn on @option{-fstack-protector-strong} by default.
1728 This option enables @option{-mcld} by default for 32-bit x86 targets.
1730 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1731 Using the GNU Compiler Collection (GCC)},
1734 See ``i386 and x86-64 Options'' in the main manual
1737 @item --enable-large-address-aware
1738 The @option{--enable-large-address-aware} option arranges for MinGW
1739 executables to be linked using the @option{--large-address-aware}
1740 option, that enables the use of more than 2GB of memory. If GCC is
1741 configured with this option, its effects can be reversed by passing the
1742 @option{-Wl,--disable-large-address-aware} option to the so-configured
1745 @item --enable-win32-registry
1746 @itemx --enable-win32-registry=@var{key}
1747 @itemx --disable-win32-registry
1748 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1749 to look up installations paths in the registry using the following key:
1752 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1755 @var{key} defaults to GCC version number, and can be overridden by the
1756 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1757 who use custom installers are encouraged to provide a different key,
1758 perhaps one comprised of vendor name and GCC version number, to
1759 avoid conflict with existing installations. This feature is enabled
1760 by default, and can be disabled by @option{--disable-win32-registry}
1761 option. This option has no effect on the other hosts.
1764 Specify that the machine does not have a floating point unit. This
1765 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1766 system, @option{--nfp} has no effect.
1768 @item --enable-werror
1769 @itemx --disable-werror
1770 @itemx --enable-werror=yes
1771 @itemx --enable-werror=no
1772 When you specify this option, it controls whether certain files in the
1773 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1774 If you don't specify it, @option{-Werror} is turned on for the main
1775 development trunk. However it defaults to off for release branches and
1776 final releases. The specific files which get @option{-Werror} are
1777 controlled by the Makefiles.
1779 @item --enable-checking
1780 @itemx --enable-checking=@var{list}
1781 When you specify this option, the compiler is built to perform internal
1782 consistency checks of the requested complexity. This does not change the
1783 generated code, but adds error checking within the compiler. This will
1784 slow down the compiler and may only work properly if you are building
1785 the compiler with GCC@. This is @samp{yes,extra} by default when building
1786 from SVN or snapshots, but @samp{release} for releases. The default
1787 for building the stage1 compiler is @samp{yes}. More control
1788 over the checks may be had by specifying @var{list}. The categories of
1789 checks available are @samp{yes} (most common checks
1790 @samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at
1791 all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest
1792 checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1793 Individual checks can be enabled with these flags @samp{assert},
1794 @samp{df}, @samp{fold}, @samp{gc}, @samp{gcac}, @samp{misc}, @samp{rtl},
1795 @samp{rtlflag}, @samp{runtime}, @samp{tree}, @samp{extra} and @samp{valgrind}.
1796 @samp{extra} adds for @samp{misc} checking extra checks that might affect
1797 code generation and should therefore not differ between stage1 and later
1800 The @samp{valgrind} check requires the external @command{valgrind}
1801 simulator, available from @uref{http://valgrind.org/}. The
1802 @samp{df}, @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive.
1803 To disable all checking, @samp{--disable-checking} or
1804 @samp{--enable-checking=none} must be explicitly requested. Disabling
1805 assertions will make the compiler and runtime slightly faster but
1806 increase the risk of undetected internal errors causing wrong code to be
1809 @item --disable-stage1-checking
1810 @itemx --enable-stage1-checking
1811 @itemx --enable-stage1-checking=@var{list}
1812 If no @option{--enable-checking} option is specified the stage1
1813 compiler will be built with @samp{yes} checking enabled, otherwise
1814 the stage1 checking flags are the same as specified by
1815 @option{--enable-checking}. To build the stage1 compiler with
1816 different checking options use @option{--enable-stage1-checking}.
1817 The list of checking options is the same as for @option{--enable-checking}.
1818 If your system is too slow or too small to bootstrap a released compiler
1819 with checking for stage1 enabled, you can use @samp{--disable-stage1-checking}
1820 to disable checking for the stage1 compiler.
1822 @item --enable-coverage
1823 @itemx --enable-coverage=@var{level}
1824 With this option, the compiler is built to collect self coverage
1825 information, every time it is run. This is for internal development
1826 purposes, and only works when the compiler is being built with gcc. The
1827 @var{level} argument controls whether the compiler is built optimized or
1828 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1829 want to disable optimization, for performance analysis you want to
1830 enable optimization. When coverage is enabled, the default level is
1831 without optimization.
1833 @item --enable-gather-detailed-mem-stats
1834 When this option is specified more detailed information on memory
1835 allocation is gathered. This information is printed when using
1836 @option{-fmem-report}.
1838 @item --enable-valgrind-annotations
1839 Mark selected memory related operations in the compiler when run under
1840 valgrind to suppress false positives.
1843 @itemx --disable-nls
1844 The @option{--enable-nls} option enables Native Language Support (NLS),
1845 which lets GCC output diagnostics in languages other than American
1846 English. Native Language Support is enabled by default if not doing a
1847 canadian cross build. The @option{--disable-nls} option disables NLS@.
1849 @item --with-included-gettext
1850 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1851 procedure to prefer its copy of GNU @command{gettext}.
1853 @item --with-catgets
1854 If NLS is enabled, and if the host lacks @code{gettext} but has the
1855 inferior @code{catgets} interface, the GCC build procedure normally
1856 ignores @code{catgets} and instead uses GCC's copy of the GNU
1857 @code{gettext} library. The @option{--with-catgets} option causes the
1858 build procedure to use the host's @code{catgets} in this situation.
1860 @item --with-libiconv-prefix=@var{dir}
1861 Search for libiconv header files in @file{@var{dir}/include} and
1862 libiconv library files in @file{@var{dir}/lib}.
1864 @item --enable-obsolete
1865 Enable configuration for an obsoleted system. If you attempt to
1866 configure GCC for a system (build, host, or target) which has been
1867 obsoleted, and you do not specify this flag, configure will halt with an
1870 All support for systems which have been obsoleted in one release of GCC
1871 is removed entirely in the next major release, unless someone steps
1872 forward to maintain the port.
1874 @item --enable-decimal-float
1875 @itemx --enable-decimal-float=yes
1876 @itemx --enable-decimal-float=no
1877 @itemx --enable-decimal-float=bid
1878 @itemx --enable-decimal-float=dpd
1879 @itemx --disable-decimal-float
1880 Enable (or disable) support for the C decimal floating point extension
1881 that is in the IEEE 754-2008 standard. This is enabled by default only
1882 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
1883 support it, but require the user to specifically enable it. You can
1884 optionally control which decimal floating point format is used (either
1885 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
1886 format is default on i386 and x86_64 systems, and the @samp{dpd}
1887 (densely packed decimal) format is default on PowerPC systems.
1889 @item --enable-fixed-point
1890 @itemx --disable-fixed-point
1891 Enable (or disable) support for C fixed-point arithmetic.
1892 This option is enabled by default for some targets (such as MIPS) which
1893 have hardware-support for fixed-point operations. On other targets, you
1894 may enable this option manually.
1896 @item --with-long-double-128
1897 Specify if @code{long double} type should be 128-bit by default on selected
1898 GNU/Linux architectures. If using @code{--without-long-double-128},
1899 @code{long double} will be by default 64-bit, the same as @code{double} type.
1900 When neither of these configure options are used, the default will be
1901 128-bit @code{long double} when built against GNU C Library 2.4 and later,
1902 64-bit @code{long double} otherwise.
1904 @item --with-long-double-format=ibm
1905 @itemx --with-long-double-format=ieee
1906 Specify whether @code{long double} uses the IBM extended double format
1907 or the IEEE 128-bit floating point format on PowerPC Linux systems.
1908 This configuration switch will only work on little endian PowerPC
1909 Linux systems and on big endian 64-bit systems where the default cpu
1910 is at least power7 (i.e. @option{--with-cpu=power7},
1911 @option{--with-cpu=power8}, or @option{--with-cpu=power9} is used).
1913 If you use the @option{--with-long-double-64} configuration option,
1914 the @option{--with-long-double-format=ibm} and
1915 @option{--with-long-double-format=ieee} options are ignored.
1917 The default @code{long double} format is to use IBM extended double.
1918 Until all of the libraries are converted to use IEEE 128-bit floating
1919 point, it is not recommended to use
1920 @option{--with-long-double-format=ieee}.
1922 On little endian PowerPC Linux systems, if you explicitly set the
1923 @code{long double} type, it will build multilibs to allow you to
1924 select either @code{long double} format, unless you disable multilibs
1925 with the @code{--disable-multilib} option. At present,
1926 @code{long double} multilibs are not built on big endian PowerPC Linux
1927 systems. If you are building multilibs, you will need to configure
1928 the compiler using the @option{--with-system-zlib} option.
1930 If you do not set the @code{long double} type explicitly, no multilibs
1933 @item --enable-fdpic
1934 On SH Linux systems, generate ELF FDPIC code.
1936 @item --with-gmp=@var{pathname}
1937 @itemx --with-gmp-include=@var{pathname}
1938 @itemx --with-gmp-lib=@var{pathname}
1939 @itemx --with-mpfr=@var{pathname}
1940 @itemx --with-mpfr-include=@var{pathname}
1941 @itemx --with-mpfr-lib=@var{pathname}
1942 @itemx --with-mpc=@var{pathname}
1943 @itemx --with-mpc-include=@var{pathname}
1944 @itemx --with-mpc-lib=@var{pathname}
1945 If you want to build GCC but do not have the GMP library, the MPFR
1946 library and/or the MPC library installed in a standard location and
1947 do not have their sources present in the GCC source tree then you
1948 can explicitly specify the directory where they are installed
1949 (@samp{--with-gmp=@var{gmpinstalldir}},
1950 @samp{--with-mpfr=@/@var{mpfrinstalldir}},
1951 @samp{--with-mpc=@/@var{mpcinstalldir}}). The
1952 @option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for
1953 @option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and
1954 @option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the
1955 @option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for
1956 @option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and
1957 @option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the
1958 @option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for
1959 @option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and
1960 @option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these
1961 shorthand assumptions are not correct, you can use the explicit
1962 include and lib options directly. You might also need to ensure the
1963 shared libraries can be found by the dynamic linker when building and
1964 using GCC, for example by setting the runtime shared library path
1965 variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems).
1967 These flags are applicable to the host platform only. When building
1968 a cross compiler, they will not be used to configure target libraries.
1970 @item --with-isl=@var{pathname}
1971 @itemx --with-isl-include=@var{pathname}
1972 @itemx --with-isl-lib=@var{pathname}
1973 If you do not have the isl library installed in a standard location and you
1974 want to build GCC, you can explicitly specify the directory where it is
1975 installed (@samp{--with-isl=@/@var{islinstalldir}}). The
1976 @option{--with-isl=@/@var{islinstalldir}} option is shorthand for
1977 @option{--with-isl-lib=@/@var{islinstalldir}/lib} and
1978 @option{--with-isl-include=@/@var{islinstalldir}/include}. If this
1979 shorthand assumption is not correct, you can use the explicit
1980 include and lib options directly.
1982 These flags are applicable to the host platform only. When building
1983 a cross compiler, they will not be used to configure target libraries.
1985 @item --with-stage1-ldflags=@var{flags}
1986 This option may be used to set linker flags to be used when linking
1987 stage 1 of GCC. These are also used when linking GCC if configured with
1988 @option{--disable-bootstrap}. If @option{--with-stage1-libs} is not set to a
1989 value, then the default is @samp{-static-libstdc++ -static-libgcc}, if
1992 @item --with-stage1-libs=@var{libs}
1993 This option may be used to set libraries to be used when linking stage 1
1994 of GCC. These are also used when linking GCC if configured with
1995 @option{--disable-bootstrap}.
1997 @item --with-boot-ldflags=@var{flags}
1998 This option may be used to set linker flags to be used when linking
1999 stage 2 and later when bootstrapping GCC. If --with-boot-libs
2000 is not is set to a value, then the default is
2001 @samp{-static-libstdc++ -static-libgcc}.
2003 @item --with-boot-libs=@var{libs}
2004 This option may be used to set libraries to be used when linking stage 2
2005 and later when bootstrapping GCC.
2007 @item --with-debug-prefix-map=@var{map}
2008 Convert source directory names using @option{-fdebug-prefix-map} when
2009 building runtime libraries. @samp{@var{map}} is a space-separated
2010 list of maps of the form @samp{@var{old}=@var{new}}.
2012 @item --enable-linker-build-id
2013 Tells GCC to pass @option{--build-id} option to the linker for all final
2014 links (links performed without the @option{-r} or @option{--relocatable}
2015 option), if the linker supports it. If you specify
2016 @option{--enable-linker-build-id}, but your linker does not
2017 support @option{--build-id} option, a warning is issued and the
2018 @option{--enable-linker-build-id} option is ignored. The default is off.
2020 @item --with-linker-hash-style=@var{choice}
2021 Tells GCC to pass @option{--hash-style=@var{choice}} option to the
2022 linker for all final links. @var{choice} can be one of
2023 @samp{sysv}, @samp{gnu}, and @samp{both} where @samp{sysv} is the default.
2025 @item --enable-gnu-unique-object
2026 @itemx --disable-gnu-unique-object
2027 Tells GCC to use the gnu_unique_object relocation for C++ template
2028 static data members and inline function local statics. Enabled by
2029 default for a toolchain with an assembler that accepts it and
2030 GLIBC 2.11 or above, otherwise disabled.
2032 @item --with-diagnostics-color=@var{choice}
2033 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-color=}
2034 option (if not used explicitly on the command line). @var{choice}
2035 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2036 where @samp{auto} is the default. @samp{auto-if-env} means that
2037 @option{-fdiagnostics-color=auto} will be the default if @code{GCC_COLORS}
2038 is present and non-empty in the environment, and
2039 @option{-fdiagnostics-color=never} otherwise.
2042 @itemx --disable-lto
2043 Enable support for link-time optimization (LTO). This is enabled by
2044 default, and may be disabled using @option{--disable-lto}.
2046 @item --enable-linker-plugin-configure-flags=FLAGS
2047 @itemx --enable-linker-plugin-flags=FLAGS
2048 By default, linker plugins (such as the LTO plugin) are built for the
2049 host system architecture. For the case that the linker has a
2050 different (but run-time compatible) architecture, these flags can be
2051 specified to build plugins that are compatible to the linker. For
2052 example, if you are building GCC for a 64-bit x86_64
2053 (@samp{x86_64-unknown-linux-gnu}) host system, but have a 32-bit x86
2054 GNU/Linux (@samp{i686-pc-linux-gnu}) linker executable (which is
2055 executable on the former system), you can configure GCC as follows for
2056 getting compatible linker plugins:
2059 % @var{srcdir}/configure \
2060 --host=x86_64-unknown-linux-gnu \
2061 --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \
2062 --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib'
2065 @item --with-plugin-ld=@var{pathname}
2066 Enable an alternate linker to be used at link-time optimization (LTO)
2067 link time when @option{-fuse-linker-plugin} is enabled.
2068 This linker should have plugin support such as gold starting with
2069 version 2.20 or GNU ld starting with version 2.21.
2070 See @option{-fuse-linker-plugin} for details.
2072 @item --enable-canonical-system-headers
2073 @itemx --disable-canonical-system-headers
2074 Enable system header path canonicalization for @file{libcpp}. This can
2075 produce shorter header file paths in diagnostics and dependency output
2076 files, but these changed header paths may conflict with some compilation
2077 environments. Enabled by default, and may be disabled using
2078 @option{--disable-canonical-system-headers}.
2080 @item --with-glibc-version=@var{major}.@var{minor}
2081 Tell GCC that when the GNU C Library (glibc) is used on the target it
2082 will be version @var{major}.@var{minor} or later. Normally this can
2083 be detected from the C library's header files, but this option may be
2084 needed when bootstrapping a cross toolchain without the header files
2085 available for building the initial bootstrap compiler.
2087 If GCC is configured with some multilibs that use glibc and some that
2088 do not, this option applies only to the multilibs that use glibc.
2089 However, such configurations may not work well as not all the relevant
2090 configuration in GCC is on a per-multilib basis.
2092 @item --enable-as-accelerator-for=@var{target}
2093 Build as offload target compiler. Specify offload host triple by @var{target}.
2095 @item --enable-offload-targets=@var{target1}[=@var{path1}],@dots{},@var{targetN}[=@var{pathN}]
2096 Enable offloading to targets @var{target1}, @dots{}, @var{targetN}.
2097 Offload compilers are expected to be already installed. Default search
2098 path for them is @file{@var{exec-prefix}}, but it can be changed by
2099 specifying paths @var{path1}, @dots{}, @var{pathN}.
2102 % @var{srcdir}/configure \
2103 --enable-offload-target=i686-unknown-linux-gnu=/path/to/i686/compiler,x86_64-pc-linux-gnu
2106 If @samp{hsa} is specified as one of the targets, the compiler will be
2107 built with support for HSA GPU accelerators. Because the same
2108 compiler will emit the accelerator code, no path should be specified.
2110 @item --with-hsa-runtime=@var{pathname}
2111 @itemx --with-hsa-runtime-include=@var{pathname}
2112 @itemx --with-hsa-runtime-lib=@var{pathname}
2114 If you configure GCC with HSA offloading but do not have the HSA
2115 run-time library installed in a standard location then you can
2116 explicitly specify the directory where they are installed. The
2117 @option{--with-hsa-runtime=@/@var{hsainstalldir}} option is a
2119 @option{--with-hsa-runtime-lib=@/@var{hsainstalldir}/lib} and
2120 @option{--with-hsa-runtime-include=@/@var{hsainstalldir}/include}.
2123 @itemx --disable-cet
2124 Enable building target run-time libraries with control-flow
2125 instrumentation, see @option{-fcf-protection} option. When
2126 @code{--enable-cet} is specified target libraries are configured
2127 to add @option{-fcf-protection} and, if needed, other target
2128 specific options to a set of building options.
2130 The option is disabled by default. When @code{--enable-cet=auto}
2131 is used, it is enabled on Linux/x86 if target binutils
2132 supports @code{Intel CET} instructions and disabled otherwise.
2133 In this case the target libraries are configured to get additional
2134 @option{-fcf-protection} option.
2137 @subheading Cross-Compiler-Specific Options
2138 The following options only apply to building cross compilers.
2141 @item --with-sysroot
2142 @itemx --with-sysroot=@var{dir}
2143 Tells GCC to consider @var{dir} as the root of a tree that contains
2144 (a subset of) the root filesystem of the target operating system.
2145 Target system headers, libraries and run-time object files will be
2146 searched for in there. More specifically, this acts as if
2147 @option{--sysroot=@var{dir}} was added to the default options of the built
2148 compiler. The specified directory is not copied into the
2149 install tree, unlike the options @option{--with-headers} and
2150 @option{--with-libs} that this option obsoletes. The default value,
2151 in case @option{--with-sysroot} is not given an argument, is
2152 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
2153 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
2154 the GCC binaries if the installation tree is moved.
2156 This option affects the system root for the compiler used to build
2157 target libraries (which runs on the build system) and the compiler newly
2158 installed with @code{make install}; it does not affect the compiler which is
2159 used to build GCC itself.
2161 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2162 option then the compiler will search that directory within @var{dirname} for
2163 native system headers rather than the default @file{/usr/include}.
2165 @item --with-build-sysroot
2166 @itemx --with-build-sysroot=@var{dir}
2167 Tells GCC to consider @var{dir} as the system root (see
2168 @option{--with-sysroot}) while building target libraries, instead of
2169 the directory specified with @option{--with-sysroot}. This option is
2170 only useful when you are already using @option{--with-sysroot}. You
2171 can use @option{--with-build-sysroot} when you are configuring with
2172 @option{--prefix} set to a directory that is different from the one in
2173 which you are installing GCC and your target libraries.
2175 This option affects the system root for the compiler used to build
2176 target libraries (which runs on the build system); it does not affect
2177 the compiler which is used to build GCC itself.
2179 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2180 option then the compiler will search that directory within @var{dirname} for
2181 native system headers rather than the default @file{/usr/include}.
2183 @item --with-headers
2184 @itemx --with-headers=@var{dir}
2185 Deprecated in favor of @option{--with-sysroot}.
2186 Specifies that target headers are available when building a cross compiler.
2187 The @var{dir} argument specifies a directory which has the target include
2188 files. These include files will be copied into the @file{gcc} install
2189 directory. @emph{This option with the @var{dir} argument is required} when
2190 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
2191 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
2192 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
2193 will be run on these files to make them compatible with GCC@.
2195 @item --without-headers
2196 Tells GCC not use any target headers from a libc when building a cross
2197 compiler. When crossing to GNU/Linux, you need the headers so GCC
2198 can build the exception handling for libgcc.
2201 @itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}"
2202 Deprecated in favor of @option{--with-sysroot}.
2203 Specifies a list of directories which contain the target runtime
2204 libraries. These libraries will be copied into the @file{gcc} install
2205 directory. If the directory list is omitted, this option has no
2209 Specifies that @samp{newlib} is
2210 being used as the target C library. This causes @code{__eprintf} to be
2211 omitted from @file{libgcc.a} on the assumption that it will be provided by
2214 @item --with-avrlibc
2215 Specifies that @samp{AVR-Libc} is
2216 being used as the target C library. This causes float support
2217 functions like @code{__addsf3} to be omitted from @file{libgcc.a} on
2218 the assumption that it will be provided by @file{libm.a}. For more
2219 technical details, cf. @uref{http://gcc.gnu.org/PR54461,,PR54461}.
2220 This option is only supported for the AVR target. It is not supported for
2221 RTEMS configurations, which currently use newlib. The option is
2222 supported since version 4.7.2 and is the default in 4.8.0 and newer.
2224 @item --with-nds32-lib=@var{library}
2225 Specifies that @var{library} setting is used for building @file{libgcc.a}.
2226 Currently, the valid @var{library} is @samp{newlib} or @samp{mculib}.
2227 This option is only supported for the NDS32 target.
2229 @item --with-build-time-tools=@var{dir}
2230 Specifies where to find the set of target tools (assembler, linker, etc.)
2231 that will be used while building GCC itself. This option can be useful
2232 if the directory layouts are different between the system you are building
2233 GCC on, and the system where you will deploy it.
2235 For example, on an @samp{ia64-hp-hpux} system, you may have the GNU
2236 assembler and linker in @file{/usr/bin}, and the native tools in a
2237 different path, and build a toolchain that expects to find the
2238 native tools in @file{/usr/bin}.
2240 When you use this option, you should ensure that @var{dir} includes
2241 @command{ar}, @command{as}, @command{ld}, @command{nm},
2242 @command{ranlib} and @command{strip} if necessary, and possibly
2243 @command{objdump}. Otherwise, GCC may use an inconsistent set of
2247 @subsubheading Overriding @command{configure} test results
2249 Sometimes, it might be necessary to override the result of some
2250 @command{configure} test, for example in order to ease porting to a new
2251 system or work around a bug in a test. The toplevel @command{configure}
2252 script provides three variables for this:
2256 @item build_configargs
2257 @cindex @code{build_configargs}
2258 The contents of this variable is passed to all build @command{configure}
2261 @item host_configargs
2262 @cindex @code{host_configargs}
2263 The contents of this variable is passed to all host @command{configure}
2266 @item target_configargs
2267 @cindex @code{target_configargs}
2268 The contents of this variable is passed to all target @command{configure}
2273 In order to avoid shell and @command{make} quoting issues for complex
2274 overrides, you can pass a setting for @env{CONFIG_SITE} and set
2275 variables in the site file.
2277 @subheading Objective-C-Specific Options
2279 The following options apply to the build of the Objective-C runtime library.
2282 @item --enable-objc-gc
2283 Specify that an additional variant of the GNU Objective-C runtime library
2284 is built, using an external build of the Boehm-Demers-Weiser garbage
2285 collector (@uref{http://www.hboehm.info/gc/}). This library needs to be
2286 available for each multilib variant, unless configured with
2287 @option{--enable-objc-gc=@samp{auto}} in which case the build of the
2288 additional runtime library is skipped when not available and the build
2291 @item --with-target-bdw-gc=@var{list}
2292 @itemx --with-target-bdw-gc-include=@var{list}
2293 @itemx --with-target-bdw-gc-lib=@var{list}
2294 Specify search directories for the garbage collector header files and
2295 libraries. @var{list} is a comma separated list of key value pairs of the
2296 form @samp{@var{multilibdir}=@var{path}}, where the default multilib key
2297 is named as @samp{.} (dot), or is omitted (e.g.
2298 @samp{--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32}).
2300 The options @option{--with-target-bdw-gc-include} and
2301 @option{--with-target-bdw-gc-lib} must always be specified together
2302 for each multilib variant and they take precedence over
2303 @option{--with-target-bdw-gc}. If @option{--with-target-bdw-gc-include}
2304 is missing values for a multilib, then the value for the default
2305 multilib is used (e.g. @samp{--with-target-bdw-gc-include=/opt/bdw-gc/include}
2306 @samp{--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32}).
2307 If none of these options are specified, the library is assumed in
2316 @uref{./index.html,,Return to the GCC Installation page}
2320 @c ***Building****************************************************************
2322 @comment node-name, next, previous, up
2323 @node Building, Testing, Configuration, Installing GCC
2329 @cindex Installing GCC: Building
2331 Now that GCC is configured, you are ready to build the compiler and
2334 Some commands executed when making the compiler may fail (return a
2335 nonzero status) and be ignored by @command{make}. These failures, which
2336 are often due to files that were not found, are expected, and can safely
2339 It is normal to have compiler warnings when compiling certain files.
2340 Unless you are a GCC developer, you can generally ignore these warnings
2341 unless they cause compilation to fail. Developers should attempt to fix
2342 any warnings encountered, however they can temporarily continue past
2343 warnings-as-errors by specifying the configure flag
2344 @option{--disable-werror}.
2346 On certain old systems, defining certain environment variables such as
2347 @env{CC} can interfere with the functioning of @command{make}.
2349 If you encounter seemingly strange errors when trying to build the
2350 compiler in a directory other than the source directory, it could be
2351 because you have previously configured the compiler in the source
2352 directory. Make sure you have done all the necessary preparations.
2354 If you build GCC on a BSD system using a directory stored in an old System
2355 V file system, problems may occur in running @command{fixincludes} if the
2356 System V file system doesn't support symbolic links. These problems
2357 result in a failure to fix the declaration of @code{size_t} in
2358 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
2359 that type mismatches occur, this could be the cause.
2361 The solution is not to use such a directory for building GCC@.
2363 Similarly, when building from SVN or snapshots, or if you modify
2364 @file{*.l} files, you need the Flex lexical analyzer generator
2365 installed. If you do not modify @file{*.l} files, releases contain
2366 the Flex-generated files and you do not need Flex installed to build
2367 them. There is still one Flex-based lexical analyzer (part of the
2368 build machinery, not of GCC itself) that is used even if you only
2369 build the C front end.
2371 When building from SVN or snapshots, or if you modify Texinfo
2372 documentation, you need version 4.7 or later of Texinfo installed if you
2373 want Info documentation to be regenerated. Releases contain Info
2374 documentation pre-built for the unmodified documentation in the release.
2376 @section Building a native compiler
2378 For a native build, the default configuration is to perform
2379 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
2380 This will build the entire GCC system and ensure that it compiles
2381 itself correctly. It can be disabled with the @option{--disable-bootstrap}
2382 parameter to @samp{configure}, but bootstrapping is suggested because
2383 the compiler will be tested more completely and could also have
2386 The bootstrapping process will complete the following steps:
2390 Build tools necessary to build the compiler.
2393 Perform a 3-stage bootstrap of the compiler. This includes building
2394 three times the target tools for use by the compiler such as binutils
2395 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
2396 individually linked or moved into the top level GCC source tree before
2400 Perform a comparison test of the stage2 and stage3 compilers.
2403 Build runtime libraries using the stage3 compiler from the previous step.
2407 If you are short on disk space you might consider @samp{make
2408 bootstrap-lean} instead. The sequence of compilation is the
2409 same described above, but object files from the stage1 and
2410 stage2 of the 3-stage bootstrap of the compiler are deleted as
2411 soon as they are no longer needed.
2413 If you wish to use non-default GCC flags when compiling the stage2
2414 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
2415 doing @samp{make}. For example, if you want to save additional space
2416 during the bootstrap and in the final installation as well, you can
2417 build the compiler binaries without debugging information as in the
2418 following example. This will save roughly 40% of disk space both for
2419 the bootstrap and the final installation. (Libraries will still contain
2420 debugging information.)
2423 make BOOT_CFLAGS='-O' bootstrap
2426 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
2427 are less well tested here than the default of @samp{-g -O2}, but should
2428 still work. In a few cases, you may find that you need to specify special
2429 flags such as @option{-msoft-float} here to complete the bootstrap; or,
2430 if the native compiler miscompiles the stage1 compiler, you may need
2431 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
2432 of the stage1 compiler that were miscompiled, or by using @samp{make
2433 bootstrap4} to increase the number of stages of bootstrap.
2435 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
2436 Since these are always compiled with the compiler currently being
2437 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
2438 compilation flags, as for non-bootstrapped target libraries.
2439 Again, if the native compiler miscompiles the stage1 compiler, you may
2440 need to work around this by avoiding non-working parts of the stage1
2441 compiler. Use @code{STAGE1_TFLAGS} to this end.
2443 If you used the flag @option{--enable-languages=@dots{}} to restrict
2444 the compilers to be built, only those you've actually enabled will be
2445 built. This will of course only build those runtime libraries, for
2446 which the particular compiler has been built. Please note,
2447 that re-defining @env{LANGUAGES} when calling @samp{make}
2448 @strong{does not} work anymore!
2450 If the comparison of stage2 and stage3 fails, this normally indicates
2451 that the stage2 compiler has compiled GCC incorrectly, and is therefore
2452 a potentially serious bug which you should investigate and report. (On
2453 a few systems, meaningful comparison of object files is impossible; they
2454 always appear ``different''. If you encounter this problem, you will
2455 need to disable comparison in the @file{Makefile}.)
2457 If you do not want to bootstrap your compiler, you can configure with
2458 @option{--disable-bootstrap}. In particular cases, you may want to
2459 bootstrap your compiler even if the target system is not the same as
2460 the one you are building on: for example, you could build a
2461 @code{powerpc-unknown-linux-gnu} toolchain on a
2462 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
2463 @option{--enable-bootstrap} to the configure script.
2465 @code{BUILD_CONFIG} can be used to bring in additional customization
2466 to the build. It can be set to a whitespace-separated list of names.
2467 For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will
2468 be included by the top-level @file{Makefile}, bringing in any settings
2469 it contains. The default @code{BUILD_CONFIG} can be set using the
2470 configure option @option{--with-build-config=@code{NAME}...}. Some
2471 examples of supported build configurations are:
2474 @item @samp{bootstrap-O1}
2475 Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds
2476 @option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to
2477 @samp{BOOT_CFLAGS='-g -O1'}.
2479 @item @samp{bootstrap-O3}
2480 Analogous to @code{bootstrap-O1}.
2482 @item @samp{bootstrap-lto}
2483 Enables Link-Time Optimization for host tools during bootstrapping.
2484 @samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding
2485 @option{-flto} to @samp{BOOT_CFLAGS}. This option assumes that the host
2486 supports the linker plugin (e.g. GNU ld version 2.21 or later or GNU gold
2487 version 2.21 or later).
2489 @item @samp{bootstrap-lto-noplugin}
2490 This option is similar to @code{bootstrap-lto}, but is intended for
2491 hosts that do not support the linker plugin. Without the linker plugin
2492 static libraries are not compiled with link-time optimizations. Since
2493 the GCC middle end and back end are in @file{libbackend.a} this means
2494 that only the front end is actually LTO optimized.
2496 @item @samp{bootstrap-debug}
2497 Verifies that the compiler generates the same executable code, whether
2498 or not it is asked to emit debug information. To this end, this
2499 option builds stage2 host programs without debug information, and uses
2500 @file{contrib/compare-debug} to compare them with the stripped stage3
2501 object files. If @code{BOOT_CFLAGS} is overridden so as to not enable
2502 debug information, stage2 will have it, and stage3 won't. This option
2503 is enabled by default when GCC bootstrapping is enabled, if
2504 @code{strip} can turn object files compiled with and without debug
2505 info into identical object files. In addition to better test
2506 coverage, this option makes default bootstraps faster and leaner.
2508 @item @samp{bootstrap-debug-big}
2509 Rather than comparing stripped object files, as in
2510 @code{bootstrap-debug}, this option saves internal compiler dumps
2511 during stage2 and stage3 and compares them as well, which helps catch
2512 additional potential problems, but at a great cost in terms of disk
2513 space. It can be specified in addition to @samp{bootstrap-debug}.
2515 @item @samp{bootstrap-debug-lean}
2516 This option saves disk space compared with @code{bootstrap-debug-big},
2517 but at the expense of some recompilation. Instead of saving the dumps
2518 of stage2 and stage3 until the final compare, it uses
2519 @option{-fcompare-debug} to generate, compare and remove the dumps
2520 during stage3, repeating the compilation that already took place in
2521 stage2, whose dumps were not saved.
2523 @item @samp{bootstrap-debug-lib}
2524 This option tests executable code invariance over debug information
2525 generation on target libraries, just like @code{bootstrap-debug-lean}
2526 tests it on host programs. It builds stage3 libraries with
2527 @option{-fcompare-debug}, and it can be used along with any of the
2528 @code{bootstrap-debug} options above.
2530 There aren't @code{-lean} or @code{-big} counterparts to this option
2531 because most libraries are only build in stage3, so bootstrap compares
2532 would not get significant coverage. Moreover, the few libraries built
2533 in stage2 are used in stage3 host programs, so we wouldn't want to
2534 compile stage2 libraries with different options for comparison purposes.
2536 @item @samp{bootstrap-debug-ckovw}
2537 Arranges for error messages to be issued if the compiler built on any
2538 stage is run without the option @option{-fcompare-debug}. This is
2539 useful to verify the full @option{-fcompare-debug} testing coverage. It
2540 must be used along with @code{bootstrap-debug-lean} and
2541 @code{bootstrap-debug-lib}.
2543 @item @samp{bootstrap-cet}
2544 This option enables Intel CET for host tools during bootstrapping.
2545 @samp{BUILD_CONFIG=bootstrap-cet} is equivalent to adding
2546 @option{-fcf-protection} to @samp{BOOT_CFLAGS}. This option
2547 assumes that the host supports Intel CET (e.g. GNU assembler version
2550 @item @samp{bootstrap-time}
2551 Arranges for the run time of each program started by the GCC driver,
2552 built in any stage, to be logged to @file{time.log}, in the top level of
2557 @section Building a cross compiler
2559 When building a cross compiler, it is not generally possible to do a
2560 3-stage bootstrap of the compiler. This makes for an interesting problem
2561 as parts of GCC can only be built with GCC@.
2563 To build a cross compiler, we recommend first building and installing a
2564 native compiler. You can then use the native GCC compiler to build the
2565 cross compiler. The installed native compiler needs to be GCC version
2568 Assuming you have already installed a native copy of GCC and configured
2569 your cross compiler, issue the command @command{make}, which performs the
2574 Build host tools necessary to build the compiler.
2577 Build target tools for use by the compiler such as binutils (bfd,
2578 binutils, gas, gprof, ld, and opcodes)
2579 if they have been individually linked or moved into the top level GCC source
2580 tree before configuring.
2583 Build the compiler (single stage only).
2586 Build runtime libraries using the compiler from the previous step.
2589 Note that if an error occurs in any step the make process will exit.
2591 If you are not building GNU binutils in the same source tree as GCC,
2592 you will need a cross-assembler and cross-linker installed before
2593 configuring GCC@. Put them in the directory
2594 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
2595 you should put in this directory:
2599 This should be the cross-assembler.
2602 This should be the cross-linker.
2605 This should be the cross-archiver: a program which can manipulate
2606 archive files (linker libraries) in the target machine's format.
2609 This should be a program to construct a symbol table in an archive file.
2612 The installation of GCC will find these programs in that directory,
2613 and copy or link them to the proper place to for the cross-compiler to
2614 find them when run later.
2616 The easiest way to provide these files is to build the Binutils package.
2617 Configure it with the same @option{--host} and @option{--target}
2618 options that you use for configuring GCC, then build and install
2619 them. They install their executables automatically into the proper
2620 directory. Alas, they do not support all the targets that GCC
2623 If you are not building a C library in the same source tree as GCC,
2624 you should also provide the target libraries and headers before
2625 configuring GCC, specifying the directories with
2626 @option{--with-sysroot} or @option{--with-headers} and
2627 @option{--with-libs}. Many targets also require ``start files'' such
2628 as @file{crt0.o} and
2629 @file{crtn.o} which are linked into each executable. There may be several
2630 alternatives for @file{crt0.o}, for use with profiling or other
2631 compilation options. Check your target's definition of
2632 @code{STARTFILE_SPEC} to find out what start files it uses.
2634 @section Building in parallel
2636 GNU Make 3.80 and above, which is necessary to build GCC, support
2637 building in parallel. To activate this, you can use @samp{make -j 2}
2638 instead of @samp{make}. You can also specify a bigger number, and
2639 in most cases using a value greater than the number of processors in
2640 your machine will result in fewer and shorter I/O latency hits, thus
2641 improving overall throughput; this is especially true for slow drives
2642 and network filesystems.
2644 @section Building the Ada compiler
2646 In order to build GNAT, the Ada compiler, you need a working GNAT
2647 compiler (GCC version 4.0 or later).
2648 This includes GNAT tools such as @command{gnatmake} and
2649 @command{gnatlink}, since the Ada front end is written in Ada and
2650 uses some GNAT-specific extensions.
2652 In order to build a cross compiler, it is suggested to install
2653 the new compiler as native first, and then use it to build the cross
2656 @command{configure} does not test whether the GNAT installation works
2657 and has a sufficiently recent version; if too old a GNAT version is
2658 installed, the build will fail unless @option{--enable-languages} is
2659 used to disable building the Ada front end.
2661 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
2662 must not be set when building the Ada compiler, the Ada tools, or the
2663 Ada runtime libraries. You can check that your build environment is clean
2664 by verifying that @samp{gnatls -v} lists only one explicit path in each
2667 @section Building with profile feedback
2669 It is possible to use profile feedback to optimize the compiler itself. This
2670 should result in a faster compiler binary. Experiments done on x86 using gcc
2671 3.3 showed approximately 7 percent speedup on compiling C programs. To
2672 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
2674 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
2675 compiler. This compiler is used to build a @code{stageprofile} compiler
2676 instrumented to collect execution counts of instruction and branch
2677 probabilities. Training run is done by building @code{stagetrain}
2678 compiler. Finally a @code{stagefeedback} compiler is built
2679 using the information collected.
2681 Unlike standard bootstrap, several additional restrictions apply. The
2682 compiler used to build @code{stage1} needs to support a 64-bit integral type.
2683 It is recommended to only use GCC for this.
2685 On Linux/x86_64 hosts with some restrictions (no virtualization) it is
2686 also possible to do autofdo build with @samp{make
2687 autoprofiledback}. This uses Linux perf to sample branches in the
2688 binary and then rebuild it with feedback derived from the profile.
2689 Linux perf and the @code{autofdo} toolkit needs to be installed for
2692 Only the profile from the current build is used, so when an error
2693 occurs it is recommended to clean before restarting. Otherwise
2694 the code quality may be much worse.
2701 @uref{./index.html,,Return to the GCC Installation page}
2705 @c ***Testing*****************************************************************
2707 @comment node-name, next, previous, up
2708 @node Testing, Final install, Building, Installing GCC
2712 @chapter Installing GCC: Testing
2715 @cindex Installing GCC: Testing
2718 Before you install GCC, we encourage you to run the testsuites and to
2719 compare your results with results from a similar configuration that have
2720 been submitted to the
2721 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
2722 Some of these archived results are linked from the build status lists
2723 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
2724 reports a successful build runs the testsuites and submits the results.
2725 This step is optional and may require you to download additional software,
2726 but it can give you confidence in your new GCC installation or point out
2727 problems before you install and start using your new GCC@.
2729 First, you must have @uref{download.html,,downloaded the testsuites}.
2730 These are part of the full distribution, but if you downloaded the
2731 ``core'' compiler plus any front ends, you must download the testsuites
2734 Second, you must have the testing tools installed. This includes
2735 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
2736 the DejaGnu site has links to these. For running the BRIG frontend
2737 tests, a tool to assemble the binary BRIGs from HSAIL text,
2738 @uref{https://github.com/HSAFoundation/HSAIL-Tools/,,HSAILasm} must
2741 If the directories where @command{runtest} and @command{expect} were
2742 installed are not in the @env{PATH}, you may need to set the following
2743 environment variables appropriately, as in the following example (which
2744 assumes that DejaGnu has been installed under @file{/usr/local}):
2747 TCL_LIBRARY = /usr/local/share/tcl8.0
2748 DEJAGNULIBS = /usr/local/share/dejagnu
2751 (On systems such as Cygwin, these paths are required to be actual
2752 paths, not mounts or links; presumably this is due to some lack of
2753 portability in the DejaGnu code.)
2756 Finally, you can run the testsuite (which may take a long time):
2758 cd @var{objdir}; make -k check
2761 This will test various components of GCC, such as compiler
2762 front ends and runtime libraries. While running the testsuite, DejaGnu
2763 might emit some harmless messages resembling
2764 @samp{WARNING: Couldn't find the global config file.} or
2765 @samp{WARNING: Couldn't find tool init file} that can be ignored.
2767 If you are testing a cross-compiler, you may want to run the testsuite
2768 on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}.
2770 @section How can you run the testsuite on selected tests?
2772 In order to run sets of tests selectively, there are targets
2773 @samp{make check-gcc} and language specific @samp{make check-c},
2774 @samp{make check-c++}, @samp{make check-d} @samp{make check-fortran},
2775 @samp{make check-ada}, @samp{make check-objc}, @samp{make check-obj-c++},
2776 @samp{make check-lto}
2777 in the @file{gcc} subdirectory of the object directory. You can also
2778 just run @samp{make check} in a subdirectory of the object directory.
2781 A more selective way to just run all @command{gcc} execute tests in the
2785 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
2788 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
2789 the testsuite with filenames matching @samp{9805*}, you would use
2792 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
2795 The file-matching expression following @var{filename}@command{.exp=} is treated
2796 as a series of whitespace-delimited glob expressions so that multiple patterns
2797 may be passed, although any whitespace must either be escaped or surrounded by
2798 single quotes if multiple expressions are desired. For example,
2801 make check-g++ RUNTESTFLAGS="old-deja.exp=9805*\ virtual2.c @var{other-options}"
2802 make check-g++ RUNTESTFLAGS="'old-deja.exp=9805* virtual2.c' @var{other-options}"
2805 The @file{*.exp} files are located in the testsuite directories of the GCC
2806 source, the most important ones being @file{compile.exp},
2807 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
2808 To get a list of the possible @file{*.exp} files, pipe the
2809 output of @samp{make check} into a file and look at the
2810 @samp{Running @dots{} .exp} lines.
2812 @section Passing options and running multiple testsuites
2814 You can pass multiple options to the testsuite using the
2815 @samp{--target_board} option of DejaGNU, either passed as part of
2816 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
2817 work outside the makefiles. For example,
2820 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
2823 will run the standard @command{g++} testsuites (``unix'' is the target name
2824 for a standard native testsuite situation), passing
2825 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
2826 slashes separate options.
2828 You can run the testsuites multiple times using combinations of options
2829 with a syntax similar to the brace expansion of popular shells:
2832 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
2835 (Note the empty option caused by the trailing comma in the final group.)
2836 The following will run each testsuite eight times using the @samp{arm-sim}
2837 target, as if you had specified all possible combinations yourself:
2840 --target_board='arm-sim/-mhard-float/-O1 \
2841 arm-sim/-mhard-float/-O2 \
2842 arm-sim/-mhard-float/-O3 \
2843 arm-sim/-mhard-float \
2844 arm-sim/-msoft-float/-O1 \
2845 arm-sim/-msoft-float/-O2 \
2846 arm-sim/-msoft-float/-O3 \
2847 arm-sim/-msoft-float'
2850 They can be combined as many times as you wish, in arbitrary ways. This
2854 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
2857 will generate four combinations, all involving @samp{-Wextra}.
2859 The disadvantage to this method is that the testsuites are run in serial,
2860 which is a waste on multiprocessor systems. For users with GNU Make and
2861 a shell which performs brace expansion, you can run the testsuites in
2862 parallel by having the shell perform the combinations and @command{make}
2863 do the parallel runs. Instead of using @samp{--target_board}, use a
2864 special makefile target:
2867 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
2873 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
2876 will run three concurrent ``make-gcc'' testsuites, eventually testing all
2877 ten combinations as described above. Note that this is currently only
2878 supported in the @file{gcc} subdirectory. (To see how this works, try
2879 typing @command{echo} before the example given here.)
2882 @section How to interpret test results
2884 The result of running the testsuite are various @file{*.sum} and @file{*.log}
2885 files in the testsuite subdirectories. The @file{*.log} files contain a
2886 detailed log of the compiler invocations and the corresponding
2887 results, the @file{*.sum} files summarize the results. These summaries
2888 contain status codes for all tests:
2892 PASS: the test passed as expected
2894 XPASS: the test unexpectedly passed
2896 FAIL: the test unexpectedly failed
2898 XFAIL: the test failed as expected
2900 UNSUPPORTED: the test is not supported on this platform
2902 ERROR: the testsuite detected an error
2904 WARNING: the testsuite detected a possible problem
2907 It is normal for some tests to report unexpected failures. At the
2908 current time the testing harness does not allow fine grained control
2909 over whether or not a test is expected to fail. This problem should
2910 be fixed in future releases.
2913 @section Submitting test results
2915 If you want to report the results to the GCC project, use the
2916 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
2919 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
2920 -m gcc-testresults@@gcc.gnu.org |sh
2923 This script uses the @command{Mail} program to send the results, so
2924 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
2925 prepended to the testsuite summary and should contain any special
2926 remarks you have on your results or your build environment. Please
2927 do not edit the testsuite result block or the subject line, as these
2928 messages may be automatically processed.
2935 @uref{./index.html,,Return to the GCC Installation page}
2939 @c ***Final install***********************************************************
2941 @comment node-name, next, previous, up
2942 @node Final install, , Testing, Installing GCC
2944 @ifset finalinstallhtml
2946 @chapter Installing GCC: Final installation
2949 Now that GCC has been built (and optionally tested), you can install it with
2951 cd @var{objdir} && make install
2954 We strongly recommend to install into a target directory where there is
2955 no previous version of GCC present. Also, the GNAT runtime should not
2956 be stripped, as this would break certain features of the debugger that
2957 depend on this debugging information (catching Ada exceptions for
2960 That step completes the installation of GCC; user level binaries can
2961 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
2962 you specified with the @option{--prefix} to configure (or
2963 @file{/usr/local} by default). (If you specified @option{--bindir},
2964 that directory will be used instead; otherwise, if you specified
2965 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
2966 Headers for the C++ library are installed in
2967 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
2968 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
2969 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
2970 in info format in @file{@var{infodir}} (normally
2971 @file{@var{prefix}/info}).
2973 When installing cross-compilers, GCC's executables
2974 are not only installed into @file{@var{bindir}}, that
2975 is, @file{@var{exec-prefix}/bin}, but additionally into
2976 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
2977 exists. Typically, such @dfn{tooldirs} hold target-specific
2978 binutils, including assembler and linker.
2980 Installation into a temporary staging area or into a @command{chroot}
2981 jail can be achieved with the command
2984 make DESTDIR=@var{path-to-rootdir} install
2988 where @var{path-to-rootdir} is the absolute path of
2989 a directory relative to which all installation paths will be
2990 interpreted. Note that the directory specified by @code{DESTDIR}
2991 need not exist yet; it will be created if necessary.
2993 There is a subtle point with tooldirs and @code{DESTDIR}:
2994 If you relocate a cross-compiler installation with
2995 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
2996 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
2997 be filled with duplicated GCC executables only if it already exists,
2998 it will not be created otherwise. This is regarded as a feature,
2999 not as a bug, because it gives slightly more control to the packagers
3000 using the @code{DESTDIR} feature.
3002 You can install stripped programs and libraries with
3008 If you are bootstrapping a released version of GCC then please
3009 quickly review the build status page for your release, available from
3010 @uref{http://gcc.gnu.org/buildstat.html}.
3011 If your system is not listed for the version of GCC that you built,
3013 @email{gcc@@gcc.gnu.org} indicating
3014 that you successfully built and installed GCC@.
3015 Include the following information:
3019 Output from running @file{@var{srcdir}/config.guess}. Do not send
3020 that file itself, just the one-line output from running it.
3023 The output of @samp{gcc -v} for your newly installed @command{gcc}.
3024 This tells us which version of GCC you built and the options you passed to
3028 Whether you enabled all languages or a subset of them. If you used a
3029 full distribution then this information is part of the configure
3030 options in the output of @samp{gcc -v}, but if you downloaded the
3031 ``core'' compiler plus additional front ends then it isn't apparent
3032 which ones you built unless you tell us about it.
3035 If the build was for GNU/Linux, also include:
3038 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
3039 this information should be available from @file{/etc/issue}.
3042 The version of the Linux kernel, available from @samp{uname --version}
3046 The version of glibc you used; for RPM-based systems like Red Hat,
3047 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
3048 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
3050 For other systems, you can include similar information if you think it is
3054 Any other information that you think would be useful to people building
3055 GCC on the same configuration. The new entry in the build status list
3056 will include a link to the archived copy of your message.
3059 We'd also like to know if the
3061 @ref{Specific, host/target specific installation notes}
3064 @uref{specific.html,,host/target specific installation notes}
3066 didn't include your host/target information or if that information is
3067 incomplete or out of date. Send a note to
3068 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
3070 If you find a bug, please report it following the
3071 @uref{../bugs/,,bug reporting guidelines}.
3073 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
3074 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
3075 and @TeX{} installed. This creates a number of @file{.dvi} files in
3076 subdirectories of @file{@var{objdir}}; these may be converted for
3077 printing with programs such as @command{dvips}. Alternately, by using
3078 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
3079 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
3080 is included with Texinfo version 4.8 and later. You can also
3081 @uref{https://shop.fsf.org/,,buy printed manuals from the
3082 Free Software Foundation}, though such manuals may not be for the most
3083 recent version of GCC@.
3085 If you would like to generate online HTML documentation, do @samp{cd
3086 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
3087 @file{@var{objdir}/gcc/HTML}.
3094 @uref{./index.html,,Return to the GCC Installation page}
3098 @c ***Binaries****************************************************************
3100 @comment node-name, next, previous, up
3101 @node Binaries, Specific, Installing GCC, Top
3105 @chapter Installing GCC: Binaries
3108 @cindex Installing GCC: Binaries
3110 We are often asked about pre-compiled versions of GCC@. While we cannot
3111 provide these for all platforms, below you'll find links to binaries for
3112 various platforms where creating them by yourself is not easy due to various
3115 Please note that we did not create these binaries, nor do we
3116 support them. If you have any problems installing them, please
3117 contact their makers.
3124 @uref{http://www.bullfreeware.com,,Bull's Open Source Software Archive for
3125 for AIX 5L and AIX 6};
3128 @uref{http://www.perzl.org/aix/,,AIX Open Source Packages (AIX5L AIX 6.1
3133 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
3139 @uref{http://hpux.connect.org.uk/,,HP-UX Porting Center};
3143 Solaris 2 (SPARC, Intel):
3146 @uref{https://www.opencsw.org/,,OpenCSW}
3149 @uref{http://jupiterrise.com/tgcware/,,TGCware}
3156 The @uref{https://brew.sh,,Homebrew} package manager;
3158 @uref{https://www.macports.org,,MacPorts}.
3165 The @uref{https://sourceware.org/cygwin/,,Cygwin} project;
3167 The @uref{http://www.mingw.org/,,MinGW} and
3168 @uref{http://mingw-w64.org/doku.php,,mingw-w64} projects.
3172 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
3173 number of platforms.
3176 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
3177 links to GNU Fortran binaries for several platforms.
3185 @uref{./index.html,,Return to the GCC Installation page}
3189 @c ***Specific****************************************************************
3191 @comment node-name, next, previous, up
3192 @node Specific, Old, Binaries, Top
3196 @chapter Host/target specific installation notes for GCC
3199 @cindex Specific installation notes
3200 @cindex Target specific installation
3201 @cindex Host specific installation
3202 @cindex Target specific installation notes
3204 Please read this document carefully @emph{before} installing the
3205 GNU Compiler Collection on your machine.
3207 Note that this list of install notes is @emph{not} a list of supported
3208 hosts or targets. Not all supported hosts and targets are listed
3209 here, only the ones that require host-specific or target-specific
3210 information have to.
3215 @uref{#aarch64-x-x,,aarch64*-*-*}
3217 @uref{#alpha-x-x,,alpha*-*-*}
3219 @uref{#amd64-x-solaris210,,amd64-*-solaris2.10}
3221 @uref{#arm-x-eabi,,arm-*-eabi}
3225 @uref{#bfin,,Blackfin}
3229 @uref{#x-x-freebsd,,*-*-freebsd*}
3231 @uref{#h8300-hms,,h8300-hms}
3233 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
3235 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
3237 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
3239 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
3241 @uref{#ix86-x-linux,,i?86-*-linux*}
3243 @uref{#ix86-x-solaris210,,i?86-*-solaris2.10}
3245 @uref{#ia64-x-linux,,ia64-*-linux}
3247 @uref{#ia64-x-hpux,,ia64-*-hpux*}
3249 @uref{#x-ibm-aix,,*-ibm-aix*}
3251 @uref{#iq2000-x-elf,,iq2000-*-elf}
3253 @uref{#lm32-x-elf,,lm32-*-elf}
3255 @uref{#lm32-x-uclinux,,lm32-*-uclinux}
3257 @uref{#m32c-x-elf,,m32c-*-elf}
3259 @uref{#m32r-x-elf,,m32r-*-elf}
3261 @uref{#m68k-x-x,,m68k-*-*}
3263 @uref{#m68k-uclinux,,m68k-uclinux}
3265 @uref{#microblaze-x-elf,,microblaze-*-elf}
3267 @uref{#mips-x-x,,mips-*-*}
3269 @uref{#nds32le-x-elf,,nds32le-*-elf}
3271 @uref{#nds32be-x-elf,,nds32be-*-elf}
3273 @uref{#nvptx-x-none,,nvptx-*-none}
3275 @uref{#powerpc-x-x,,powerpc*-*-*}
3277 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
3279 @uref{#powerpc-x-elf,,powerpc-*-elf}
3281 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
3283 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
3285 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
3287 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
3289 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
3291 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
3293 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
3295 @uref{#riscv32-x-elf,,riscv32-*-elf}
3297 @uref{#riscv32-x-linux,,riscv32-*-linux}
3299 @uref{#riscv64-x-elf,,riscv64-*-elf}
3301 @uref{#riscv64-x-linux,,riscv64-*-linux}
3303 @uref{#s390-x-linux,,s390-*-linux*}
3305 @uref{#s390x-x-linux,,s390x-*-linux*}
3307 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
3309 @uref{#x-x-solaris2,,*-*-solaris2*}
3311 @uref{#sparc-x-x,,sparc*-*-*}
3313 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
3315 @uref{#sparc-sun-solaris210,,sparc-sun-solaris2.10}
3317 @uref{#sparc-x-linux,,sparc-*-linux*}
3319 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
3321 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
3323 @uref{#c6x-x-x,,c6x-*-*}
3325 @uref{#tilegx-x-linux,,tilegx-*-linux*}
3327 @uref{#tilegxbe-x-linux,,tilegxbe-*-linux*}
3329 @uref{#tilepro-x-linux,,tilepro-*-linux*}
3331 @uref{#visium-x-elf, visium-*-elf}
3333 @uref{#x-x-vxworks,,*-*-vxworks*}
3335 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
3337 @uref{#x86-64-x-solaris210,,x86_64-*-solaris2.1[0-9]*}
3339 @uref{#xtensa-x-elf,,xtensa*-*-elf}
3341 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
3343 @uref{#windows,,Microsoft Windows}
3345 @uref{#x-x-cygwin,,*-*-cygwin}
3347 @uref{#x-x-mingw32,,*-*-mingw32}
3351 @uref{#older,,Older systems}
3356 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
3362 <!-- -------- host/target specific issues start here ---------------- -->
3365 @anchor{aarch64-x-x}
3366 @heading aarch64*-*-*
3367 Binutils pre 2.24 does not have support for selecting @option{-mabi} and
3368 does not support ILP32. If it is used to build GCC 4.9 or later, GCC will
3369 not support option @option{-mabi=ilp32}.
3371 To enable a workaround for the Cortex-A53 erratum number 835769 by default
3372 (for all CPUs regardless of -mcpu option given) at configure time use the
3373 @option{--enable-fix-cortex-a53-835769} option. This will enable the fix by
3374 default and can be explicitly disabled during compilation by passing the
3375 @option{-mno-fix-cortex-a53-835769} option. Conversely,
3376 @option{--disable-fix-cortex-a53-835769} will disable the workaround by
3377 default. The workaround is disabled by default if neither of
3378 @option{--enable-fix-cortex-a53-835769} or
3379 @option{--disable-fix-cortex-a53-835769} is given at configure time.
3381 To enable a workaround for the Cortex-A53 erratum number 843419 by default
3382 (for all CPUs regardless of -mcpu option given) at configure time use the
3383 @option{--enable-fix-cortex-a53-843419} option. This workaround is applied at
3384 link time. Enabling the workaround will cause GCC to pass the relevant option
3385 to the linker. It can be explicitly disabled during compilation by passing the
3386 @option{-mno-fix-cortex-a53-843419} option. Conversely,
3387 @option{--disable-fix-cortex-a53-843419} will disable the workaround by default.
3388 The workaround is disabled by default if neither of
3389 @option{--enable-fix-cortex-a53-843419} or
3390 @option{--disable-fix-cortex-a53-843419} is given at configure time.
3397 This section contains general configuration information for all
3398 Alpha-based platforms using ELF@. In addition to reading this
3399 section, please read all other sections that match your target.
3401 We require binutils 2.11.2 or newer.
3402 Previous binutils releases had a number of problems with DWARF 2
3403 debugging information, not the least of which is incorrect linking of
3409 @anchor{amd64-x-solaris210}
3410 @heading amd64-*-solaris2.1[0-9]*
3411 This is a synonym for @samp{x86_64-*-solaris2.1[0-9]*}.
3416 @anchor{arc-x-elf32}
3417 @heading arc-*-elf32
3419 Use @samp{configure --target=arc-elf32 --with-cpu=@var{cpu} --enable-languages="c,c++"}
3420 to configure GCC, with @var{cpu} being one of @samp{arc600}, @samp{arc601},
3426 @anchor{arc-linux-uclibc}
3427 @heading arc-linux-uclibc
3429 Use @samp{configure --target=arc-linux-uclibc --with-cpu=arc700 --enable-languages="c,c++"} to configure GCC@.
3436 ARM-family processors.
3438 Building the Ada frontend commonly fails (an infinite loop executing
3439 @code{xsinfo}) if the host compiler is GNAT 4.8. Host compilers built from the
3440 GNAT 4.6, 4.9 or 5 release branches are known to succeed.
3447 ATMEL AVR-family micro controllers. These are used in embedded
3448 applications. There are no standard Unix configurations.
3450 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
3454 See ``AVR Options'' in the main manual
3456 for the list of supported MCU types.
3458 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
3460 Further installation notes and other useful information about AVR tools
3461 can also be obtained from:
3465 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
3467 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
3470 The following error:
3472 Error: register required
3475 indicates that you should upgrade to a newer version of the binutils.
3482 The Blackfin processor, an Analog Devices DSP.
3484 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
3488 See ``Blackfin Options'' in the main manual
3491 More information, and a version of binutils with support for this processor,
3492 is available at @uref{https://blackfin.uclinux.org}
3499 The CR16 CompactRISC architecture is a 16-bit architecture. This
3500 architecture is used in embedded applications.
3503 @xref{CR16 Options,, CR16 Options, gcc, Using and Porting the GNU Compiler
3508 See ``CR16 Options'' in the main manual for a list of CR16-specific options.
3511 Use @samp{configure --target=cr16-elf --enable-languages=c,c++} to configure
3512 GCC@ for building a CR16 elf cross-compiler.
3514 Use @samp{configure --target=cr16-uclinux --enable-languages=c,c++} to
3515 configure GCC@ for building a CR16 uclinux cross-compiler.
3522 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
3523 series. These are used in embedded applications.
3526 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
3530 See ``CRIS Options'' in the main manual
3532 for a list of CRIS-specific options.
3534 There are a few different CRIS targets:
3537 Mainly for monolithic embedded systems. Includes a multilib for the
3538 @samp{v10} core used in @samp{ETRAX 100 LX}.
3539 @item cris-axis-linux-gnu
3540 A GNU/Linux port for the CRIS architecture, currently targeting
3541 @samp{ETRAX 100 LX} by default.
3544 Pre-packaged tools can be obtained from
3545 @uref{ftp://ftp.axis.com/@/pub/@/axis/@/tools/@/cris/@/compiler-kit/}. More
3546 information about this platform is available at
3547 @uref{http://developer.axis.com/}.
3554 Please have a look at the @uref{binaries.html,,binaries page}.
3556 You cannot install GCC by itself on MSDOS; it will not compile under
3557 any MSDOS compiler except itself. You need to get the complete
3558 compilation package DJGPP, which includes binaries as well as sources,
3559 and includes all the necessary compilation tools and libraries.
3564 @anchor{epiphany-x-elf}
3565 @heading epiphany-*-elf
3567 This configuration is intended for embedded systems.
3572 @anchor{x-x-freebsd}
3573 @heading *-*-freebsd*
3574 Support for FreeBSD 1 was discontinued in GCC 3.2. Support for
3575 FreeBSD 2 (and any mutant a.out variants of FreeBSD 3) was
3576 discontinued in GCC 4.0.
3578 In order to better utilize FreeBSD base system functionality and match
3579 the configuration of the system compiler, GCC 4.5 and above as well as
3580 GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present
3581 on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default
3582 (on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside
3583 @file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled
3584 by GCC 4.5 and above.
3586 We support FreeBSD using the ELF file format with DWARF 2 debugging
3587 for all CPU architectures. You may use @option{-gstabs} instead of
3588 @option{-g}, if you really want the old debugging format. There are
3589 no known issues with mixing object files and libraries with different
3590 debugging formats. Otherwise, this release of GCC should now match
3591 more of the configuration used in the stock FreeBSD configuration of
3592 GCC@. In particular, @option{--enable-threads} is now configured by
3593 default. However, as a general user, do not attempt to replace the
3594 system compiler with this release. Known to bootstrap and check with
3595 good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap
3596 and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4,
3597 4.5, 4.8, 4.9 and 5-CURRENT@.
3599 The version of binutils installed in @file{/usr/bin} probably works
3600 with this release of GCC@. Bootstrapping against the latest GNU
3601 binutils and/or the version found in @file{/usr/ports/devel/binutils} has
3602 been known to enable additional features and improve overall testsuite
3603 results. However, it is currently known that boehm-gc may not configure
3604 properly on FreeBSD prior to the FreeBSD 7.0 release with GNU binutils
3613 This configuration is intended for embedded systems.
3620 Renesas H8/300 series of processors.
3622 Please have a look at the @uref{binaries.html,,binaries page}.
3624 The calling convention and structure layout has changed in release 2.6.
3625 All code must be recompiled. The calling convention now passes the
3626 first three arguments in function calls in registers. Structures are no
3627 longer a multiple of 2 bytes.
3632 @anchor{hppa-hp-hpux}
3633 @heading hppa*-hp-hpux*
3634 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
3636 We require using gas/binutils on all hppa platforms. Version 2.19 or
3637 later is recommended.
3639 It may be helpful to configure GCC with the
3640 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
3641 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
3643 The HP assembler should not be used with GCC. It is rarely tested and may
3644 not work. It shouldn't be used with any languages other than C due to its
3647 Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging
3648 format which GCC does not know about). It also inserts timestamps
3649 into each object file it creates, causing the 3-stage comparison test to
3650 fail during a bootstrap. You should be able to continue by saying
3651 @samp{make all-host all-target} after getting the failure from @samp{make}.
3653 Various GCC features are not supported. For example, it does not support weak
3654 symbols or alias definitions. As a result, explicit template instantiations
3655 are required when using C++. This makes it difficult if not impossible to
3656 build many C++ applications.
3658 There are two default scheduling models for instructions. These are
3659 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
3660 architecture specified for the target machine when configuring.
3661 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
3662 the target is a @samp{hppa1*} machine.
3664 The PROCESSOR_8000 model is not well suited to older processors. Thus,
3665 it is important to completely specify the machine architecture when
3666 configuring if you want a model other than PROCESSOR_8000. The macro
3667 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
3668 default scheduling model is desired.
3670 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
3671 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
3672 This namespace change might cause problems when bootstrapping with
3673 an earlier version of GCC or the HP compiler as essentially the same
3674 namespace is required for an entire build. This problem can be avoided
3675 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
3676 or @samp{98}. Another way is to add an appropriate set of predefines
3677 to @env{CC}. The description for the @option{munix=} option contains
3678 a list of the predefines used with each standard.
3680 More specific information to @samp{hppa*-hp-hpux*} targets follows.
3685 @anchor{hppa-hp-hpux10}
3686 @heading hppa*-hp-hpux10
3687 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
3688 @code{PHCO_19798} from HP@.
3690 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
3691 used for one-only code and data. This resolves many of the previous
3692 problems in using C++ on this target. However, the ABI is not compatible
3693 with the one implemented under HP-UX 11 using secondary definitions.
3698 @anchor{hppa-hp-hpux11}
3699 @heading hppa*-hp-hpux11
3700 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
3701 be used to compile GCC 3.0 and up.
3703 The libffi library haven't been ported to 64-bit HP-UX@ and doesn't build.
3705 Refer to @uref{binaries.html,,binaries} for information about obtaining
3706 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
3707 to build the Ada language as it cannot be bootstrapped using C@. Ada is
3708 only available for the 32-bit PA-RISC runtime.
3710 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
3711 bundled compiler supports only traditional C; you will need either HP's
3712 unbundled compiler, or a binary distribution of GCC@.
3714 It is possible to build GCC 3.3 starting with the bundled HP compiler,
3715 but the process requires several steps. GCC 3.3 can then be used to
3716 build later versions.
3718 There are several possible approaches to building the distribution.
3719 Binutils can be built first using the HP tools. Then, the GCC
3720 distribution can be built. The second approach is to build GCC
3721 first using the HP tools, then build binutils, then rebuild GCC@.
3722 There have been problems with various binary distributions, so it
3723 is best not to start from a binary distribution.
3725 On 64-bit capable systems, there are two distinct targets. Different
3726 installation prefixes must be used if both are to be installed on
3727 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
3728 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
3729 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
3730 PA-RISC 2.0 architecture.
3732 The script config.guess now selects the target type based on the compiler
3733 detected during configuration. You must define @env{PATH} or @env{CC} so
3734 that configure finds an appropriate compiler for the initial bootstrap.
3735 When @env{CC} is used, the definition should contain the options that are
3736 needed whenever @env{CC} is used.
3738 Specifically, options that determine the runtime architecture must be
3739 in @env{CC} to correctly select the target for the build. It is also
3740 convenient to place many other compiler options in @env{CC}. For example,
3741 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
3742 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
3743 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
3744 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
3745 macro definition table of cpp needs to be increased for a successful
3746 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
3747 be defined when building with the bundled compiler, or when using the
3748 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
3750 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
3751 with the @option{--with-ld=@dots{}} option. This overrides the standard
3752 search for ld. The two linkers supported on this target require different
3753 commands. The default linker is determined during configuration. As a
3754 result, it's not possible to switch linkers in the middle of a GCC build.
3755 This has been reported to sometimes occur in unified builds of binutils
3758 A recent linker patch must be installed for the correct operation of
3759 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
3760 oldest linker patches that are known to work. They are for HP-UX
3761 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
3762 @code{PHSS_24304}, might be usable but it hasn't been tested. These
3763 patches have been superseded. Consult the HP patch database to obtain
3764 the currently recommended linker patch for your system.
3766 The patches are necessary for the support of weak symbols on the
3767 32-bit port, and for the running of initializers and finalizers. Weak
3768 symbols are implemented using SOM secondary definition symbols. Prior
3769 to HP-UX 11, there are bugs in the linker support for secondary symbols.
3770 The patches correct a problem of linker core dumps creating shared
3771 libraries containing secondary symbols, as well as various other
3772 linking issues involving secondary symbols.
3774 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
3775 run initializers and finalizers on the 64-bit port. The 32-bit port
3776 uses the linker @option{+init} and @option{+fini} options for the same
3777 purpose. The patches correct various problems with the +init/+fini
3778 options, including program core dumps. Binutils 2.14 corrects a
3779 problem on the 64-bit port resulting from HP's non-standard use of
3780 the .init and .fini sections for array initializers and finalizers.
3782 Although the HP and GNU linkers are both supported for the
3783 @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the
3784 HP linker be used for link editing on this target.
3786 At this time, the GNU linker does not support the creation of long
3787 branch stubs. As a result, it cannot successfully link binaries
3788 containing branch offsets larger than 8 megabytes. In addition,
3789 there are problems linking shared libraries, linking executables
3790 with @option{-static}, and with dwarf2 unwind and exception support.
3791 It also doesn't provide stubs for internal calls to global functions
3792 in shared libraries, so these calls cannot be overloaded.
3794 The HP dynamic loader does not support GNU symbol versioning, so symbol
3795 versioning is not supported. It may be necessary to disable symbol
3796 versioning with @option{--disable-symvers} when using GNU ld.
3798 POSIX threads are the default. The optional DCE thread library is not
3799 supported, so @option{--enable-threads=dce} does not work.
3804 @anchor{x-x-linux-gnu}
3805 @heading *-*-linux-gnu
3806 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
3807 in glibc 2.2.5 and later. More information is available in the
3808 libstdc++-v3 documentation.
3813 @anchor{ix86-x-linux}
3814 @heading i?86-*-linux*
3815 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
3816 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
3818 If you receive Signal 11 errors when building on GNU/Linux, then it is
3819 possible you have a hardware problem. Further information on this can be
3820 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
3825 @anchor{ix86-x-solaris210}
3826 @heading i?86-*-solaris2.10
3827 Use this for Solaris 10 or later on x86 and x86-64 systems. Starting
3828 with GCC 4.7, there is also a 64-bit @samp{amd64-*-solaris2.1[0-9]*} or
3829 @samp{x86_64-*-solaris2.1[0-9]*} configuration that corresponds to
3830 @samp{sparcv9-sun-solaris2*}.
3832 It is recommended that you configure GCC to use the GNU assembler. The
3833 versions included in Solaris 10, from GNU binutils 2.15 (in
3834 @file{/usr/sfw/bin/gas}), and Solaris 11, from GNU binutils 2.19 or
3835 newer (also available as @file{/usr/bin/gas} and
3836 @file{/usr/gnu/bin/as}), work fine. The current version, from GNU
3837 binutils 2.29, is known to work, but the version from GNU binutils 2.26
3838 must be avoided. Recent versions of the Solaris assembler in
3839 @file{/usr/ccs/bin/as} work almost as well, though.
3840 @c FIXME: as patch requirements?
3842 For linking, the Solaris linker, is preferred. If you want to use the GNU
3843 linker instead, note that due to a packaging bug the version in Solaris
3844 10, from GNU binutils 2.15 (in @file{/usr/sfw/bin/gld}), cannot be used,
3845 while the version in Solaris 11, from GNU binutils 2.19 or newer (also
3846 in @file{/usr/gnu/bin/ld} and @file{/usr/bin/gld}), works, as does the
3847 latest version, from GNU binutils 2.29.
3849 To use GNU @command{as}, configure with the options
3850 @option{--with-gnu-as --with-as=@//usr/@/sfw/@/bin/@/gas}. It may be necessary
3851 to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to
3852 guarantee use of Sun @command{ld}.
3853 @c FIXME: why --without-gnu-ld --with-ld?
3858 @anchor{ia64-x-linux}
3859 @heading ia64-*-linux
3860 IA-64 processor (also known as IPF, or Itanium Processor Family)
3863 If you are using the installed system libunwind library with
3864 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
3867 None of the following versions of GCC has an ABI that is compatible
3868 with any of the other versions in this list, with the exception that
3869 Red Hat 2.96 and Trillian 000171 are compatible with each other:
3870 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
3871 This primarily affects C++ programs and programs that create shared libraries.
3872 GCC 3.1 or later is recommended for compiling linux, the kernel.
3873 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
3874 more major ABI changes are expected.
3879 @anchor{ia64-x-hpux}
3880 @heading ia64-*-hpux*
3881 Building GCC on this target requires the GNU Assembler. The bundled HP
3882 assembler will not work. To prevent GCC from using the wrong assembler,
3883 the option @option{--with-gnu-as} may be necessary.
3885 The GCC libunwind library has not been ported to HPUX@. This means that for
3886 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
3887 is required to build GCC@. For GCC 3.3 and later, this is the default.
3888 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
3889 removed and the system libunwind library will always be used.
3893 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
3897 Support for AIX version 3 and older was discontinued in GCC 3.4.
3898 Support for AIX version 4.2 and older was discontinued in GCC 4.5.
3900 ``out of memory'' bootstrap failures may indicate a problem with
3901 process resource limits (ulimit). Hard limits are configured in the
3902 @file{/etc/security/limits} system configuration file.
3904 GCC 4.9 and above require a C++ compiler for bootstrap. IBM VAC++ / xlC
3905 cannot bootstrap GCC. xlc can bootstrap an older version of GCC and
3906 G++ can bootstrap recent releases of GCC.
3908 GCC can bootstrap with recent versions of IBM XLC, but bootstrapping
3909 with an earlier release of GCC is recommended. Bootstrapping with XLC
3910 requires a larger data segment, which can be enabled through the
3911 @var{LDR_CNTRL} environment variable, e.g.,
3914 % LDR_CNTRL=MAXDATA=0x50000000
3918 One can start with a pre-compiled version of GCC to build from
3919 sources. One may delete GCC's ``fixed'' header files when starting
3920 with a version of GCC built for an earlier release of AIX.
3922 To speed up the configuration phases of bootstrapping and installing GCC,
3923 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
3926 % CONFIG_SHELL=/opt/freeware/bin/bash
3927 % export CONFIG_SHELL
3930 and then proceed as described in @uref{build.html,,the build
3931 instructions}, where we strongly recommend specifying an absolute path
3932 to invoke @var{srcdir}/configure.
3934 Because GCC on AIX is built as a 32-bit executable by default,
3935 (although it can generate 64-bit programs) the GMP and MPFR libraries
3936 required by gfortran must be 32-bit libraries. Building GMP and MPFR
3937 as static archive libraries works better than shared libraries.
3939 Errors involving @code{alloca} when building GCC generally are due
3940 to an incorrect definition of @code{CC} in the Makefile or mixing files
3941 compiled with the native C compiler and GCC@. During the stage1 phase of
3942 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
3943 (not @command{xlc}). Once @command{configure} has been informed of
3944 @command{xlc}, one needs to use @samp{make distclean} to remove the
3945 configure cache files and ensure that @env{CC} environment variable
3946 does not provide a definition that will confuse @command{configure}.
3947 If this error occurs during stage2 or later, then the problem most likely
3948 is the version of Make (see above).
3950 The native @command{as} and @command{ld} are recommended for
3951 bootstrapping on AIX@. The GNU Assembler, GNU Linker, and GNU
3952 Binutils version 2.20 is the minimum level that supports bootstrap on
3953 AIX 5@. The GNU Assembler has not been updated to support AIX 6@ or
3954 AIX 7. The native AIX tools do interoperate with GCC@.
3956 AIX 7.1 added partial support for DWARF debugging, but full support
3957 requires AIX 7.1 TL03 SP7 that supports additional DWARF sections and
3958 fixes a bug in the assembler. AIX 7.1 TL03 SP5 distributed a version
3959 of libm.a missing important symbols; a fix for IV77796 will be
3962 AIX 5.3 TL10, AIX 6.1 TL05 and AIX 7.1 TL00 introduced an AIX
3963 assembler change that sometimes produces corrupt assembly files
3964 causing AIX linker errors. The bug breaks GCC bootstrap on AIX and
3965 can cause compilation failures with existing GCC installations. An
3966 AIX iFix for AIX 5.3 is available (APAR IZ98385 for AIX 5.3 TL10, APAR
3967 IZ98477 for AIX 5.3 TL11 and IZ98134 for AIX 5.3 TL12). AIX 5.3 TL11 SP8,
3968 AIX 5.3 TL12 SP5, AIX 6.1 TL04 SP11, AIX 6.1 TL05 SP7, AIX 6.1 TL06 SP6,
3969 AIX 6.1 TL07 and AIX 7.1 TL01 should include the fix.
3971 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
3972 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
3973 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
3974 referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1)
3976 @anchor{TransferAixShobj}
3977 @samp{libstdc++} in GCC 3.4 increments the major version number of the
3978 shared object and GCC installation places the @file{libstdc++.a}
3979 shared library in a common location which will overwrite the and GCC
3980 3.3 version of the shared library. Applications either need to be
3981 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
3982 versions of the @samp{libstdc++} shared object needs to be available
3983 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
3984 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
3985 installed for runtime dynamic loading using the following steps to set
3986 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3987 multilib @file{libstdc++.a} installed:
3989 Extract the shared objects from the currently installed
3990 @file{libstdc++.a} archive:
3992 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3995 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3996 available for runtime dynamic loading, but not linking:
3998 % strip -e libstdc++.so.4 libstdc++.so.5
4001 Archive the runtime-only shared object in the GCC 3.4
4002 @file{libstdc++.a} archive:
4004 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
4008 @uref{./configure.html#WithAixSoname,,@option{--with-aix-soname=svr4}}
4009 configure option may drop the need for this procedure for libraries that
4012 Linking executables and shared libraries may produce warnings of
4013 duplicate symbols. The assembly files generated by GCC for AIX always
4014 have included multiple symbol definitions for certain global variable
4015 and function declarations in the original program. The warnings should
4016 not prevent the linker from producing a correct library or runnable
4019 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
4020 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
4021 to parse archive libraries did not handle the new format correctly.
4022 These routines are used by GCC and result in error messages during
4023 linking such as ``not a COFF file''. The version of the routines shipped
4024 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
4025 option of the archive command may be used to create archives of 32-bit
4026 objects using the original ``small format''. A correct version of the
4027 routines is shipped with AIX 4.3.2 and above.
4029 Some versions of the AIX binder (linker) can fail with a relocation
4030 overflow severe error when the @option{-bbigtoc} option is used to link
4031 GCC-produced object files into an executable that overflows the TOC@. A fix
4032 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
4033 available from IBM Customer Support and from its
4034 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4035 website as PTF U455193.
4037 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
4038 with a segmentation fault when invoked by any version of GCC@. A fix for
4039 APAR IX87327 is available from IBM Customer Support and from its
4040 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4041 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
4043 The initial assembler shipped with AIX 4.3.0 generates incorrect object
4044 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
4045 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
4046 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4047 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
4049 AIX provides National Language Support (NLS)@. Compilers and assemblers
4050 use NLS to support locale-specific representations of various data
4051 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
4052 separating decimal fractions). There have been problems reported where
4053 GCC does not produce the same floating-point formats that the assembler
4054 expects. If one encounters this problem, set the @env{LANG}
4055 environment variable to @samp{C} or @samp{En_US}.
4057 A default can be specified with the @option{-mcpu=@var{cpu_type}}
4058 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
4063 @anchor{iq2000-x-elf}
4064 @heading iq2000-*-elf
4065 Vitesse IQ2000 processors. These are used in embedded
4066 applications. There are no standard Unix configurations.
4073 Lattice Mico32 processor.
4074 This configuration is intended for embedded systems.
4079 @anchor{lm32-x-uclinux}
4080 @heading lm32-*-uclinux
4081 Lattice Mico32 processor.
4082 This configuration is intended for embedded systems running uClinux.
4089 Renesas M32C processor.
4090 This configuration is intended for embedded systems.
4097 Renesas M32R processor.
4098 This configuration is intended for embedded systems.
4106 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
4108 build libraries for both M680x0 and ColdFire processors. If you only
4109 need the M680x0 libraries, you can omit the ColdFire ones by passing
4110 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
4111 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
4112 @command{configure}. These targets default to 5206 or 5475 code as
4113 appropriate for the target system when
4114 configured with @option{--with-arch=cf} and 68020 code otherwise.
4116 The @samp{m68k-*-netbsd} and
4117 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
4118 option. They will generate ColdFire CFV4e code when configured with
4119 @option{--with-arch=cf} and 68020 code otherwise.
4121 You can override the default processors listed above by configuring
4122 with @option{--with-cpu=@var{target}}. This @var{target} can either
4123 be a @option{-mcpu} argument or one of the following values:
4124 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
4125 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
4127 GCC requires at least binutils version 2.17 on these targets.
4132 @anchor{m68k-x-uclinux}
4133 @heading m68k-*-uclinux
4134 GCC 4.3 changed the uClinux configuration so that it uses the
4135 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
4136 It also added improved support for C++ and flat shared libraries,
4137 both of which were ABI changes.
4142 @anchor{microblaze-x-elf}
4143 @heading microblaze-*-elf
4144 Xilinx MicroBlaze processor.
4145 This configuration is intended for embedded systems.
4152 If on a MIPS system you get an error message saying ``does not have gp
4153 sections for all it's [sic] sectons [sic]'', don't worry about it. This
4154 happens whenever you use GAS with the MIPS linker, but there is not
4155 really anything wrong, and it is okay to use the output file. You can
4156 stop such warnings by installing the GNU linker.
4158 It would be nice to extend GAS to produce the gp tables, but they are
4159 optional, and there should not be a warning about their absence.
4161 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
4162 and later. A patch went in just after the GCC 3.3 release to
4163 make @samp{mips*-*-*} use the generic implementation instead. You can also
4164 configure for @samp{mipsel-elf} as a workaround. The
4165 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
4166 work on this is expected in future releases.
4168 @c If you make --with-llsc the default for another target, please also
4169 @c update the description of the --with-llsc option.
4171 The built-in @code{__sync_*} functions are available on MIPS II and
4172 later systems and others that support the @samp{ll}, @samp{sc} and
4173 @samp{sync} instructions. This can be overridden by passing
4174 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
4175 Since the Linux kernel emulates these instructions if they are
4176 missing, the default for @samp{mips*-*-linux*} targets is
4177 @option{--with-llsc}. The @option{--with-llsc} and
4178 @option{--without-llsc} configure options may be overridden at compile
4179 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
4182 MIPS systems check for division by zero (unless
4183 @option{-mno-check-zero-division} is passed to the compiler) by
4184 generating either a conditional trap or a break instruction. Using
4185 trap results in smaller code, but is only supported on MIPS II and
4186 later. Also, some versions of the Linux kernel have a bug that
4187 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
4188 the use of break, use the @option{--with-divide=breaks}
4189 @command{configure} option when configuring GCC@. The default is to
4190 use traps on systems that support them.
4195 @anchor{moxie-x-elf}
4196 @heading moxie-*-elf
4197 The moxie processor.
4202 @anchor{msp430-x-elf}
4203 @heading msp430-*-elf
4204 TI MSP430 processor.
4205 This configuration is intended for embedded systems.
4210 @anchor{nds32le-x-elf}
4211 @heading nds32le-*-elf
4212 Andes NDS32 target in little endian mode.
4217 @anchor{nds32be-x-elf}
4218 @heading nds32be-*-elf
4219 Andes NDS32 target in big endian mode.
4224 @anchor{nvptx-x-none}
4225 @heading nvptx-*-none
4228 Instead of GNU binutils, you will need to install
4229 @uref{https://github.com/MentorEmbedded/nvptx-tools/,,nvptx-tools}.
4230 Tell GCC where to find it:
4231 @option{--with-build-time-tools=[install-nvptx-tools]/nvptx-none/bin}.
4233 You will need newlib 3.0 git revision
4234 cd31fbb2aea25f94d7ecedc9db16dfc87ab0c316 or later. It can be
4235 automatically built together with GCC@. For this, add a symbolic link
4236 to nvptx-newlib's @file{newlib} directory to the directory containing
4239 Use the @option{--disable-sjlj-exceptions} and
4240 @option{--enable-newlib-io-long-long} options when configuring.
4245 @anchor{powerpc-x-x}
4246 @heading powerpc-*-*
4247 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
4248 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
4250 You will need GNU binutils 2.15 or newer.
4255 @anchor{powerpc-x-darwin}
4256 @heading powerpc-*-darwin*
4257 PowerPC running Darwin (Mac OS X kernel).
4259 Pre-installed versions of Mac OS X may not include any developer tools,
4260 meaning that you will not be able to build GCC from source. Tool
4261 binaries are available at
4262 @uref{https://opensource.apple.com}.
4264 This version of GCC requires at least cctools-590.36. The
4265 cctools-590.36 package referenced from
4266 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
4267 on systems older than 10.3.9 (aka darwin7.9.0).
4272 @anchor{powerpc-x-elf}
4273 @heading powerpc-*-elf
4274 PowerPC system in big endian mode, running System V.4.
4279 @anchor{powerpc-x-linux-gnu}
4280 @heading powerpc*-*-linux-gnu*
4281 PowerPC system in big endian mode running Linux.
4286 @anchor{powerpc-x-netbsd}
4287 @heading powerpc-*-netbsd*
4288 PowerPC system in big endian mode running NetBSD@.
4293 @anchor{powerpc-x-eabisim}
4294 @heading powerpc-*-eabisim
4295 Embedded PowerPC system in big endian mode for use in running under the
4301 @anchor{powerpc-x-eabi}
4302 @heading powerpc-*-eabi
4303 Embedded PowerPC system in big endian mode.
4308 @anchor{powerpcle-x-elf}
4309 @heading powerpcle-*-elf
4310 PowerPC system in little endian mode, running System V.4.
4315 @anchor{powerpcle-x-eabisim}
4316 @heading powerpcle-*-eabisim
4317 Embedded PowerPC system in little endian mode for use in running under
4323 @anchor{powerpcle-x-eabi}
4324 @heading powerpcle-*-eabi
4325 Embedded PowerPC system in little endian mode.
4332 The Renesas RL78 processor.
4333 This configuration is intended for embedded systems.
4338 @anchor{riscv32-x-elf}
4339 @heading riscv32-*-elf
4340 The RISC-V RV32 instruction set.
4341 This configuration is intended for embedded systems.
4342 This (and all other RISC-V) targets are supported upstream as of the
4343 binutils 2.28 release.
4348 @anchor{riscv32-x-linux}
4349 @heading riscv32-*-linux
4350 The RISC-V RV32 instruction set running GNU/Linux.
4351 This (and all other RISC-V) targets are supported upstream as of the
4352 binutils 2.28 release.
4357 @anchor{riscv64-x-elf}
4358 @heading riscv64-*-elf
4359 The RISC-V RV64 instruction set.
4360 This configuration is intended for embedded systems.
4361 This (and all other RISC-V) targets are supported upstream as of the
4362 binutils 2.28 release.
4367 @anchor{riscv64-x-linux}
4368 @heading riscv64-*-linux
4369 The RISC-V RV64 instruction set running GNU/Linux.
4370 This (and all other RISC-V) targets are supported upstream as of the
4371 binutils 2.28 release.
4378 The Renesas RX processor.
4383 @anchor{s390-x-linux}
4384 @heading s390-*-linux*
4385 S/390 system running GNU/Linux for S/390@.
4390 @anchor{s390x-x-linux}
4391 @heading s390x-*-linux*
4392 zSeries system (64-bit) running GNU/Linux for zSeries@.
4397 @anchor{s390x-ibm-tpf}
4398 @heading s390x-ibm-tpf*
4399 zSeries system (64-bit) running TPF@. This platform is
4400 supported as cross-compilation target only.
4405 @c Please use Solaris 2 to refer to all release of Solaris, starting
4406 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
4407 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
4408 @c alone is too unspecific and must be avoided.
4409 @anchor{x-x-solaris2}
4410 @heading *-*-solaris2*
4411 Support for Solaris 10 has been obsoleted in GCC 9, but can still be
4412 enabled by configuring with @option{--enable-obsolete}. Support will be
4413 removed in GCC 10. Support for Solaris 9 has been removed in GCC 5.
4414 Support for Solaris 8 has been removed in GCC 4.8. Support for Solaris
4415 7 has been removed in GCC 4.6.
4417 Sun does not ship a C compiler with Solaris 2 before Solaris 10, though
4418 you can download the Sun Studio compilers for free. In Solaris 10 and
4419 11, GCC 3.4.3 is available as @command{/usr/sfw/bin/gcc}. Solaris 11
4420 also provides GCC 4.5.2, 4.7.3, and 4.8.2 as
4421 @command{/usr/gcc/4.5/bin/gcc} or similar. Alternatively,
4422 you can install a pre-built GCC to bootstrap and install GCC. See the
4423 @uref{binaries.html,,binaries page} for details.
4425 The Solaris 2 @command{/bin/sh} will often fail to configure
4426 @samp{libstdc++-v3}or @samp{boehm-gc}. We therefore recommend using the
4427 following initial sequence of commands
4430 % CONFIG_SHELL=/bin/ksh
4431 % export CONFIG_SHELL
4435 and proceed as described in @uref{configure.html,,the configure instructions}.
4436 In addition we strongly recommend specifying an absolute path to invoke
4437 @command{@var{srcdir}/configure}.
4439 Solaris 10 comes with a number of optional OS packages. Some of these
4440 are needed to use GCC fully, namely @code{SUNWarc},
4441 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
4442 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
4443 optional packages when installing Solaris 10, you will need to verify that
4444 the packages that GCC needs are installed.
4445 To check whether an optional package is installed, use
4446 the @command{pkginfo} command. To add an optional package, use the
4447 @command{pkgadd} command. For further details, see the Solaris 10
4450 Starting with Solaris 11, the package management has changed, so you
4451 need to check for @code{system/header}, @code{system/linker}, and
4452 @code{developer/assembler} packages. Checking for and installing
4453 packages is done with the @command{pkg} command now.
4455 Trying to use the linker and other tools in
4456 @file{/usr/ucb} to install GCC has been observed to cause trouble.
4457 For example, the linker may hang indefinitely. The fix is to remove
4458 @file{/usr/ucb} from your @env{PATH}.
4460 The build process works more smoothly with the legacy Sun tools so, if you
4461 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
4462 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
4464 We recommend the use of the Solaris assembler or the GNU assembler, in
4465 conjunction with the Solaris linker. The GNU @command{as}
4466 versions included in Solaris 10, from GNU binutils 2.15 (in
4467 @file{/usr/sfw/bin/gas}), and Solaris 11,
4468 from GNU binutils 2.19 or newer (also in @file{/usr/bin/gas} and
4469 @file{/usr/gnu/bin/as}), are known to work.
4470 The current version, from GNU binutils 2.29,
4471 is known to work as well. Note that your mileage may vary
4472 if you use a combination of the GNU tools and the Solaris tools: while the
4473 combination GNU @command{as} + Sun @command{ld} should reasonably work,
4474 the reverse combination Sun @command{as} + GNU @command{ld} may fail to
4475 build or cause memory corruption at runtime in some cases for C++ programs.
4477 GNU @command{ld} usually works as well, although the version included in
4478 Solaris 10 cannot be used due to several bugs. Again, the current
4479 version (2.29) is known to work, but generally lacks platform specific
4480 features, so better stay with Solaris @command{ld}. To use the LTO linker
4481 plugin (@option{-fuse-linker-plugin}) with GNU @command{ld}, GNU
4482 binutils @emph{must} be configured with @option{--enable-largefile}.
4484 To enable symbol versioning in @samp{libstdc++} with the Solaris linker,
4485 you need to have any version of GNU @command{c++filt}, which is part of
4486 GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no
4487 appropriate version is found. Solaris @command{c++filt} from the Solaris
4488 Studio compilers does @emph{not} work.
4490 Sun bug 4927647 sometimes causes random spurious testsuite failures
4491 related to missing diagnostic output. This bug doesn't affect GCC
4492 itself, rather it is a kernel bug triggered by the @command{expect}
4493 program which is used only by the GCC testsuite driver. When the bug
4494 causes the @command{expect} program to miss anticipated output, extra
4495 testsuite failures appear.
4502 This section contains general configuration information for all
4503 SPARC-based platforms. In addition to reading this section, please
4504 read all other sections that match your target.
4506 Newer versions of the GNU Multiple Precision Library (GMP), the MPFR
4507 library and the MPC library are known to be miscompiled by earlier
4508 versions of GCC on these platforms. We therefore recommend the use
4509 of the exact versions of these libraries listed as minimal versions
4510 in @uref{prerequisites.html,,the prerequisites}.
4515 @anchor{sparc-sun-solaris2}
4516 @heading sparc-sun-solaris2*
4517 When GCC is configured to use GNU binutils 2.14 or later, the binaries
4518 produced are smaller than the ones produced using Sun's native tools;
4519 this difference is quite significant for binaries containing debugging
4522 Starting with Solaris 7, the operating system is capable of executing
4523 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
4524 this; the @option{-m64} option enables 64-bit code generation.
4525 However, if all you want is code tuned for the UltraSPARC CPU, you
4526 should try the @option{-mtune=ultrasparc} option instead, which produces
4527 code that, unlike full 64-bit code, can still run on non-UltraSPARC
4530 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4531 library or the MPC library on a Solaris 7 or later system, the canonical
4532 target triplet must be specified as the @command{build} parameter on the
4533 configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and
4534 not that of GMP or MPFR or MPC). For example on a Solaris 9 system:
4537 % ./configure --build=sparc-sun-solaris2.9 --prefix=xxx
4543 @anchor{sparc-sun-solaris210}
4544 @heading sparc-sun-solaris2.10
4545 There is a bug in older versions of the Sun assembler which breaks
4546 thread-local storage (TLS). A typical error message is
4549 ld: fatal: relocation error: R_SPARC_TLS_LE_HIX22: file /var/tmp//ccamPA1v.o:
4550 symbol <unknown>: bad symbol type SECT: symbol type must be TLS
4554 This bug is fixed in Sun patch 118683-03 or later.
4559 @anchor{sparc-x-linux}
4560 @heading sparc-*-linux*
4565 @anchor{sparc64-x-solaris2}
4566 @heading sparc64-*-solaris2*
4567 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4568 library or the MPC library, the canonical target triplet must be specified
4569 as the @command{build} parameter on the configure line. For example
4570 on a Solaris 9 system:
4573 % ./configure --build=sparc64-sun-solaris2.9 --prefix=xxx
4579 @anchor{sparcv9-x-solaris2}
4580 @heading sparcv9-*-solaris2*
4581 This is a synonym for @samp{sparc64-*-solaris2*}.
4588 The C6X family of processors. This port requires binutils-2.22 or newer.
4593 @anchor{tilegx-*-linux}
4594 @heading tilegx-*-linux*
4595 The TILE-Gx processor in little endian mode, running GNU/Linux. This
4596 port requires binutils-2.22 or newer.
4601 @anchor{tilegxbe-*-linux}
4602 @heading tilegxbe-*-linux*
4603 The TILE-Gx processor in big endian mode, running GNU/Linux. This
4604 port requires binutils-2.23 or newer.
4609 @anchor{tilepro-*-linux}
4610 @heading tilepro-*-linux*
4611 The TILEPro processor running GNU/Linux. This port requires
4612 binutils-2.22 or newer.
4617 @anchor{visium-x-elf}
4618 @heading visium-*-elf
4619 CDS VISIUMcore processor.
4620 This configuration is intended for embedded systems.
4625 @anchor{x-x-vxworks}
4626 @heading *-*-vxworks*
4627 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4628 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4629 We welcome patches for other architectures supported by VxWorks 5.5.
4630 Support for VxWorks AE would also be welcome; we believe this is merely
4631 a matter of writing an appropriate ``configlette'' (see below). We are
4632 not interested in supporting older, a.out or COFF-based, versions of
4635 VxWorks comes with an older version of GCC installed in
4636 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4637 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4638 Before running @command{configure}, create the directories @file{@var{prefix}}
4639 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4640 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4641 include that directory while running both @command{configure} and
4644 You must give @command{configure} the
4645 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4646 find the VxWorks system headers. Since VxWorks is a cross compilation
4647 target only, you must also specify @option{--target=@var{target}}.
4648 @command{configure} will attempt to create the directory
4649 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4650 make sure the user running @command{configure} has sufficient privilege
4653 GCC's exception handling runtime requires a special ``configlette''
4654 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4655 that file to add the module to your kernel build. (Future versions of
4656 VxWorks will incorporate this module.)
4662 @heading x86_64-*-*, amd64-*-*
4663 GCC supports the x86-64 architecture implemented by the AMD64 processor
4664 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4665 On GNU/Linux the default is a bi-arch compiler which is able to generate
4666 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4671 @anchor{x86-64-x-solaris210}
4672 @heading x86_64-*-solaris2.1[0-9]*
4673 GCC also supports the x86-64 architecture implemented by the AMD64
4674 processor (@samp{amd64-*-*} is an alias for @samp{x86_64-*-*}) on
4675 Solaris 10 or later. Unlike other systems, without special options a
4676 bi-arch compiler is built which generates 32-bit code by default, but
4677 can generate 64-bit x86-64 code with the @option{-m64} switch. Since
4678 GCC 4.7, there is also a configuration that defaults to 64-bit code, but
4679 can generate 32-bit code with @option{-m32}. To configure and build
4680 this way, you have to provide all support libraries like @file{libgmp}
4681 as 64-bit code, configure with @option{--target=x86_64-pc-solaris2.1x}
4682 and @samp{CC=gcc -m64}.
4687 @anchor{xtensa-x-elf}
4688 @heading xtensa*-*-elf
4689 This target is intended for embedded Xtensa systems using the
4690 @samp{newlib} C library. It uses ELF but does not support shared
4691 objects. Designed-defined instructions specified via the
4692 Tensilica Instruction Extension (TIE) language are only supported
4693 through inline assembly.
4695 The Xtensa configuration information must be specified prior to
4696 building GCC@. The @file{include/xtensa-config.h} header
4697 file contains the configuration information. If you created your
4698 own Xtensa configuration with the Xtensa Processor Generator, the
4699 downloaded files include a customized copy of this header file,
4700 which you can use to replace the default header file.
4705 @anchor{xtensa-x-linux}
4706 @heading xtensa*-*-linux*
4707 This target is for Xtensa systems running GNU/Linux. It supports ELF
4708 shared objects and the GNU C library (glibc). It also generates
4709 position-independent code (PIC) regardless of whether the
4710 @option{-fpic} or @option{-fPIC} options are used. In other
4711 respects, this target is the same as the
4712 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
4718 @heading Microsoft Windows
4720 @subheading Intel 16-bit versions
4721 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
4724 However, the 32-bit port has limited support for Microsoft
4725 Windows 3.11 in the Win32s environment, as a target only. See below.
4727 @subheading Intel 32-bit versions
4728 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
4729 XP, and Windows Vista, are supported by several different target
4730 platforms. These targets differ in which Windows subsystem they target
4731 and which C libraries are used.
4734 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
4735 Linux API emulation layer in the Win32 subsystem.
4736 @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for
4737 the Win32 subsystem that provides a subset of POSIX.
4738 @item MKS i386-pc-mks: NuTCracker from MKS. See
4739 @uref{https://www.mkssoftware.com} for more information.
4742 @subheading Intel 64-bit versions
4743 GCC contains support for x86-64 using the mingw-w64
4744 runtime library, available from @uref{http://mingw-w64.org/doku.php}.
4745 This library should be used with the target triple x86_64-pc-mingw32.
4747 Presently Windows for Itanium is not supported.
4749 @subheading Windows CE
4750 Windows CE is supported as a target only on Hitachi
4751 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
4753 @subheading Other Windows Platforms
4754 GCC no longer supports Windows NT on the Alpha or PowerPC.
4756 GCC no longer supports the Windows POSIX subsystem. However, it does
4757 support the Interix subsystem. See above.
4759 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
4761 PW32 (i386-pc-pw32) support was never completed, and the project seems to
4762 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
4764 UWIN support has been removed due to a lack of maintenance.
4771 Ports of GCC are included with the
4772 @uref{http://www.cygwin.com/,,Cygwin environment}.
4774 GCC will build under Cygwin without modification; it does not build
4775 with Microsoft's C++ compiler and there are no plans to make it do so.
4777 The Cygwin native compiler can be configured to target any 32-bit x86
4778 cpu architecture desired; the default is i686-pc-cygwin. It should be
4779 used with as up-to-date a version of binutils as possible; use either
4780 the latest official GNU binutils release in the Cygwin distribution,
4781 or version 2.20 or above if building your own.
4786 @anchor{x-x-mingw32}
4787 @heading *-*-mingw32
4788 GCC will build with and support only MinGW runtime 3.12 and later.
4789 Earlier versions of headers are incompatible with the new default semantics
4790 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
4796 @heading Older systems
4797 GCC contains support files for many older (1980s and early
4798 1990s) Unix variants. For the most part, support for these systems
4799 has not been deliberately removed, but it has not been maintained for
4800 several years and may suffer from bitrot.
4802 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
4803 Support for these systems is still present in that release, but
4804 @command{configure} will fail unless the @option{--enable-obsolete}
4805 option is given. Unless a maintainer steps forward, support for these
4806 systems will be removed from the next release of GCC@.
4808 Support for old systems as hosts for GCC can cause problems if the
4809 workarounds for compiler, library and operating system bugs affect the
4810 cleanliness or maintainability of the rest of GCC@. In some cases, to
4811 bring GCC up on such a system, if still possible with current GCC, may
4812 require first installing an old version of GCC which did work on that
4813 system, and using it to compile a more recent GCC, to avoid bugs in the
4814 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
4815 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
4816 sites}. Header bugs may generally be avoided using
4817 @command{fixincludes}, but bugs or deficiencies in libraries and the
4818 operating system may still cause problems.
4820 Support for older systems as targets for cross-compilation is less
4821 problematic than support for them as hosts for GCC; if an enthusiast
4822 wishes to make such a target work again (including resurrecting any of
4823 the targets that never worked with GCC 2, starting from the last
4824 version before they were removed), patches
4825 @uref{../contribute.html,,following the usual requirements} would be
4826 likely to be accepted, since they should not affect the support for more
4829 For some systems, old versions of GNU binutils may also be useful,
4830 and are available from @file{pub/binutils/old-releases} on
4831 @uref{https://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
4833 Some of the information on specific systems above relates to
4834 such older systems, but much of the information
4835 about GCC on such systems (which may no longer be applicable to
4836 current GCC) is to be found in the GCC texinfo manual.
4842 @heading all ELF targets (SVR4, Solaris 2, etc.)
4843 C++ support is significantly better on ELF targets if you use the
4844 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
4845 inlines, vtables and template instantiations will be discarded
4854 @uref{./index.html,,Return to the GCC Installation page}
4858 @c ***Old documentation******************************************************
4860 @include install-old.texi
4866 @uref{./index.html,,Return to the GCC Installation page}
4870 @c ***GFDL********************************************************************
4878 @uref{./index.html,,Return to the GCC Installation page}
4882 @c ***************************************************************************
4883 @c Part 6 The End of the Document
4885 @comment node-name, next, previous, up
4886 @node Concept Index, , GNU Free Documentation License, Top
4890 @unnumbered Concept Index