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, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
48 @c 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
49 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
51 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
52 @c test the generation of HTML documents for the gcc.gnu.org web pages.
54 @c Do not use @footnote{} in this file as it breaks install.texi2html!
56 @c Include everything if we're not making html
60 @set prerequisiteshtml
71 @c Part 2 Summary Description and Copyright
73 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
74 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007,
75 2008 Free Software Foundation, Inc.
77 Permission is granted to copy, distribute and/or modify this document
78 under the terms of the GNU Free Documentation License, Version 1.2 or
79 any later version published by the Free Software Foundation; with no
80 Invariant Sections, the Front-Cover texts being (a) (see below), and
81 with the Back-Cover Texts being (b) (see below). A copy of the
82 license is included in the section entitled ``@uref{./gfdl.html,,GNU
83 Free Documentation License}''.
85 (a) The FSF's Front-Cover Text is:
89 (b) The FSF's Back-Cover Text is:
91 You have freedom to copy and modify this GNU Manual, like GNU
92 software. Copies published by the Free Software Foundation raise
93 funds for GNU development.
98 @dircategory Software development
100 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
103 @c Part 3 Titlepage and Copyright
105 @title Installing GCC
108 @c The following two commands start the copyright page.
110 @vskip 0pt plus 1filll
114 @c Part 4 Top node, Master Menu, and/or Table of Contents
117 @comment node-name, next, Previous, up
120 * Installing GCC:: This document describes the generic installation
121 procedure for GCC as well as detailing some target
122 specific installation instructions.
124 * Specific:: Host/target specific installation notes for GCC.
125 * Binaries:: Where to get pre-compiled binaries.
127 * Old:: Old installation documentation.
129 * GNU Free Documentation License:: How you can copy and share this manual.
130 * Concept Index:: This index has two entries.
138 @c Part 5 The Body of the Document
139 @c ***Installing GCC**********************************************************
141 @comment node-name, next, previous, up
142 @node Installing GCC, Binaries, , Top
146 @chapter Installing GCC
149 The latest version of this document is always available at
150 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
152 This document describes the generic installation procedure for GCC as well
153 as detailing some target specific installation instructions.
155 GCC includes several components that previously were separate distributions
156 with their own installation instructions. This document supersedes all
157 package specific installation instructions.
159 @emph{Before} starting the build/install procedure please check the
161 @ref{Specific, host/target specific installation notes}.
164 @uref{specific.html,,host/target specific installation notes}.
166 We recommend you browse the entire generic installation instructions before
169 Lists of successful builds for released versions of GCC are
170 available at @uref{http://gcc.gnu.org/buildstat.html}.
171 These lists are updated as new information becomes available.
173 The installation procedure itself is broken into five steps.
178 * Downloading the source::
181 * Testing:: (optional)
188 @uref{prerequisites.html,,Prerequisites}
190 @uref{download.html,,Downloading the source}
192 @uref{configure.html,,Configuration}
194 @uref{build.html,,Building}
196 @uref{test.html,,Testing} (optional)
198 @uref{finalinstall.html,,Final install}
202 Please note that GCC does not support @samp{make uninstall} and probably
203 won't do so in the near future as this would open a can of worms. Instead,
204 we suggest that you install GCC into a directory of its own and simply
205 remove that directory when you do not need that specific version of GCC
206 any longer, and, if shared libraries are installed there as well, no
207 more binaries exist that use them.
210 There are also some @uref{old.html,,old installation instructions},
211 which are mostly obsolete but still contain some information which has
212 not yet been merged into the main part of this manual.
220 @uref{./index.html,,Return to the GCC Installation page}
226 @c ***Prerequisites**************************************************
228 @comment node-name, next, previous, up
229 @node Prerequisites, Downloading the source, , Installing GCC
231 @ifset prerequisiteshtml
233 @chapter Prerequisites
235 @cindex Prerequisites
237 GCC requires that various tools and packages be available for use in the
238 build procedure. Modifying GCC sources requires additional tools
241 @heading Tools/packages necessary for building GCC
243 @item ISO C90 compiler
244 Necessary to bootstrap GCC, although versions of GCC prior
245 to 3.4 also allow bootstrapping with a traditional (K&R) C compiler.
247 To build all languages in a cross-compiler or other configuration where
248 3-stage bootstrap is not performed, you need to start with an existing
249 GCC binary (version 2.95 or later) because source code for language
250 frontends other than C might use GCC extensions.
254 In order to build the Ada compiler (GNAT) you must already have GNAT
255 installed because portions of the Ada frontend are written in Ada (with
256 GNAT extensions.) Refer to the Ada installation instructions for more
257 specific information.
259 @item A ``working'' POSIX compatible shell, or GNU bash
261 Necessary when running @command{configure} because some
262 @command{/bin/sh} shells have bugs and may crash when configuring the
263 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
264 have disastrous corner-case performance problems. This
265 can cause target @command{configure} runs to literally take days to
266 complete in some cases.
268 So on some platforms @command{/bin/ksh} is sufficient, on others it
269 isn't. See the host/target specific instructions for your platform, or
270 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
271 environment to your ``good'' shell prior to running
272 @command{configure}/@command{make}.
274 @command{zsh} is not a fully compliant POSIX shell and will not
275 work when configuring GCC@.
277 @item A POSIX or SVR4 awk
279 Necessary for creating some of the generated source files for GCC@.
280 If in doubt, use a recent GNU awk version, as some of the older ones
281 are broken. GNU awk version 3.1.5 is known to work.
285 Necessary in some circumstances, optional in others. See the
286 host/target specific instructions for your platform for the exact
289 @item gzip version 1.2.4 (or later) or
290 @itemx bzip2 version 1.0.2 (or later)
292 Necessary to uncompress GCC @command{tar} files when source code is
293 obtained via FTP mirror sites.
295 @item GNU make version 3.80 (or later)
297 You must have GNU make installed to build GCC@.
299 @item GNU tar version 1.14 (or later)
301 Necessary (only on some platforms) to untar the source code. Many
302 systems' @command{tar} programs will also work, only try GNU
303 @command{tar} if you have problems.
305 @item GNU Multiple Precision Library (GMP) version 4.2 (or later)
307 Necessary to build GCC@. If you do not have it installed in your
308 library search path, you will have to configure with the
309 @option{--with-gmp} configure option. See also @option{--with-gmp-lib}
310 and @option{--with-gmp-include}. Alternatively, if a GMP source
311 distribution is found in a subdirectory of your GCC sources named
312 @file{gmp}, it will be built together with GCC@.
314 @item MPFR Library version 2.3.2 (or later)
316 Necessary to build GCC@. It can be downloaded from
317 @uref{http://www.mpfr.org/}. The @option{--with-mpfr} configure
318 option should be used if your MPFR Library is not installed in your
319 default library search path. See also @option{--with-mpfr-lib} and
320 @option{--with-mpfr-include}. Alternatively, if a MPFR source
321 distribution is found in a subdirectory of your GCC sources named
322 @file{mpfr}, it will be built together with GCC@.
324 @item Parma Polyhedra Library (PPL) version 0.10
326 Necessary to build GCC with the Graphite loop optimizations.
327 It can be downloaded from @uref{http://www.cs.unipr.it/ppl/Download/}.
329 The @option{--with-ppl} configure option should be used if PPL is not
330 installed in your default library search path.
332 @item CLooG-PPL version 0.15
334 Necessary to build GCC with the Graphite loop optimizations. It can
335 be downloaded from @uref{ftp://gcc.gnu.org/pub/gcc/infrastructure/}.
336 The code in @file{cloog-ppl-0.15.tar.gz} comes from a branch of CLooG
337 available from @uref{http://repo.or.cz/w/cloog-ppl.git}. CLooG-PPL
338 should be configured with @option{--with-ppl}.
340 The @option{--with-cloog} configure option should be used if CLooG is
341 not installed in your default library search path.
343 @item @command{jar}, or InfoZIP (@command{zip} and @command{unzip})
345 Necessary to build libgcj, the GCJ runtime.
350 @heading Tools/packages necessary for modifying GCC
352 @item autoconf version 2.59
353 @itemx GNU m4 version 1.4 (or later)
355 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
356 to regenerate @file{configure} and @file{config.in} files.
358 @item automake version 1.9.6
360 Necessary when modifying a @file{Makefile.am} file to regenerate its
361 associated @file{Makefile.in}.
363 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
364 file. Specifically this applies to the @file{gcc}, @file{intl},
365 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
366 as any of their subdirectories.
368 For directories that use automake, GCC requires the latest release in
369 the 1.9.x series, which is currently 1.9.6. When regenerating a directory
370 to a newer version, please update all the directories using an older 1.9.x
371 to the latest released version.
373 @item gettext version 0.14.5 (or later)
375 Needed to regenerate @file{gcc.pot}.
377 @item gperf version 2.7.2 (or later)
379 Necessary when modifying @command{gperf} input files, e.g.@:
380 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
381 @file{gcc/cp/cfns.h}.
387 Necessary to run the GCC testsuite; see the section on testing for details.
389 @item autogen version 5.5.4 (or later) and
390 @itemx guile version 1.4.1 (or later)
392 Necessary to regenerate @file{fixinc/fixincl.x} from
393 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
395 Necessary to run @samp{make check} for @file{fixinc}.
397 Necessary to regenerate the top level @file{Makefile.in} file from
398 @file{Makefile.tpl} and @file{Makefile.def}.
400 @item Flex version 2.5.4 (or later)
402 Necessary when modifying @file{*.l} files.
404 Necessary to build GCC during development because the generated output
405 files are not included in the SVN repository. They are included in
408 @item Texinfo version 4.7 (or later)
410 Necessary for running @command{makeinfo} when modifying @file{*.texi}
411 files to test your changes.
413 Necessary for running @command{make dvi} or @command{make pdf} to
414 create printable documentation in DVI or PDF format. Texinfo version
415 4.8 or later is required for @command{make pdf}.
417 Necessary to build GCC documentation during development because the
418 generated output files are not included in the SVN repository. They are
419 included in releases.
421 @item @TeX{} (any working version)
423 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
424 are used when running @command{make dvi} or @command{make pdf} to create
425 DVI or PDF files, respectively.
427 @item SVN (any version)
428 @itemx SSH (any version)
430 Necessary to access the SVN repository. Public releases and weekly
431 snapshots of the development sources are also available via FTP@.
433 @item Perl version 5.6.1 (or later)
435 Necessary when regenerating @file{Makefile} dependencies in libiberty.
436 Necessary when regenerating @file{libiberty/functions.texi}.
437 Necessary when generating manpages from Texinfo manuals.
438 Necessary when targetting Darwin, building libstdc++,
439 and not using @option{--disable-symvers}.
440 Used by various scripts to generate some files included in SVN (mainly
441 Unicode-related and rarely changing) from source tables.
443 @item GNU diffutils version 2.7 (or later)
445 Useful when submitting patches for the GCC source code.
447 @item patch version 2.5.4 (or later)
449 Necessary when applying patches, created with @command{diff}, to one's
455 If you wish to modify @file{.java} files in libjava, you will need to
456 configure with @option{--enable-java-maintainer-mode}, and you will need
457 to have executables named @command{ecj1} and @command{gjavah} in your path.
458 The @command{ecj1} executable should run the Eclipse Java compiler via
459 the GCC-specific entry point. You can download a suitable jar from
460 @uref{ftp://sourceware.org/pub/java/}, or by running the script
461 @command{contrib/download_ecj}.
463 @item antlr.jar version 2.7.1 (or later)
466 If you wish to build the @command{gjdoc} binary in libjava, you will
467 need to have a @file{antlr.jar} library available. The library is
468 searched in system locations but can be configured with
469 @option{--with-antlr-jar=} instead. When configuring with
470 @option{--enable-java-maintainer-mode}, you will need to have one of
471 the executables named @command{cantlr}, @command{runantlr} or
472 @command{antlr} in your path.
481 @uref{./index.html,,Return to the GCC Installation page}
485 @c ***Downloading the source**************************************************
487 @comment node-name, next, previous, up
488 @node Downloading the source, Configuration, Prerequisites, Installing GCC
492 @chapter Downloading GCC
494 @cindex Downloading GCC
495 @cindex Downloading the Source
497 GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP
498 tarballs compressed with @command{gzip} or
499 @command{bzip2}. It is possible to download a full distribution or specific
502 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
503 for information on how to obtain GCC@.
505 The full distribution includes the C, C++, Objective-C, Fortran, Java,
506 and Ada (in the case of GCC 3.1 and later) compilers. The full
507 distribution also includes runtime libraries for C++, Objective-C,
508 Fortran, and Java. In GCC 3.0 and later versions, the GNU compiler
509 testsuites are also included in the full distribution.
511 If you choose to download specific components, you must download the core
512 GCC distribution plus any language specific distributions you wish to
513 use. The core distribution includes the C language front end as well as the
514 shared components. Each language has a tarball which includes the language
515 front end as well as the language runtime (when appropriate).
517 Unpack the core distribution as well as any language specific
518 distributions in the same directory.
520 If you also intend to build binutils (either to upgrade an existing
521 installation or for use in place of the corresponding tools of your
522 OS), unpack the binutils distribution either in the same directory or
523 a separate one. In the latter case, add symbolic links to any
524 components of the binutils you intend to build alongside the compiler
525 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
526 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
528 Likewise, the GMP and MPFR libraries can be automatically built together
529 with GCC. Unpack the GMP and/or MPFR source distributions in the
530 directory containing the GCC sources and rename their directories to
531 @file{gmp} and @file{mpfr}, respectively (or use symbolic links with the
539 @uref{./index.html,,Return to the GCC Installation page}
543 @c ***Configuration***********************************************************
545 @comment node-name, next, previous, up
546 @node Configuration, Building, Downloading the source, Installing GCC
550 @chapter Installing GCC: Configuration
552 @cindex Configuration
553 @cindex Installing GCC: Configuration
555 Like most GNU software, GCC must be configured before it can be built.
556 This document describes the recommended configuration procedure
557 for both native and cross targets.
559 We use @var{srcdir} to refer to the toplevel source directory for
560 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
562 If you obtained the sources via SVN, @var{srcdir} must refer to the top
563 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
564 and not its @file{gcc} subdirectory, otherwise the build will fail.
566 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
567 file system, the shell's built-in @command{pwd} command will return
568 temporary pathnames. Using these can lead to various sorts of build
569 problems. To avoid this issue, set the @env{PWDCMD} environment
570 variable to an automounter-aware @command{pwd} command, e.g.,
571 @command{pawd} or @samp{amq -w}, during the configuration and build
574 First, we @strong{highly} recommend that GCC be built into a
575 separate directory than the sources which does @strong{not} reside
576 within the source tree. This is how we generally build GCC; building
577 where @var{srcdir} == @var{objdir} should still work, but doesn't
578 get extensive testing; building where @var{objdir} is a subdirectory
579 of @var{srcdir} is unsupported.
581 If you have previously built GCC in the same directory for a
582 different target machine, do @samp{make distclean} to delete all files
583 that might be invalid. One of the files this deletes is @file{Makefile};
584 if @samp{make distclean} complains that @file{Makefile} does not exist
585 or issues a message like ``don't know how to make distclean'' it probably
586 means that the directory is already suitably clean. However, with the
587 recommended method of building in a separate @var{objdir}, you should
588 simply use a different @var{objdir} for each target.
590 Second, when configuring a native system, either @command{cc} or
591 @command{gcc} must be in your path or you must set @env{CC} in
592 your environment before running configure. Otherwise the configuration
596 Note that the bootstrap compiler and the resulting GCC must be link
597 compatible, else the bootstrap will fail with linker errors about
598 incompatible object file formats. Several multilibed targets are
599 affected by this requirement, see
601 @ref{Specific, host/target specific installation notes}.
604 @uref{specific.html,,host/target specific installation notes}.
613 % @var{srcdir}/configure [@var{options}] [@var{target}]
616 @heading Distributor options
618 If you will be distributing binary versions of GCC, with modifications
619 to the source code, you should use the options described in this
620 section to make clear that your version contains modifications.
623 @item --with-pkgversion=@var{version}
624 Specify a string that identifies your package. You may wish
625 to include a build number or build date. This version string will be
626 included in the output of @command{gcc --version}. This suffix does
627 not replace the default version string, only the @samp{GCC} part.
629 The default value is @samp{GCC}.
631 @item --with-bugurl=@var{url}
632 Specify the URL that users should visit if they wish to report a bug.
633 You are of course welcome to forward bugs reported to you to the FSF,
634 if you determine that they are not bugs in your modifications.
636 The default value refers to the FSF's GCC bug tracker.
640 @heading Target specification
643 GCC has code to correctly determine the correct value for @var{target}
644 for nearly all native systems. Therefore, we highly recommend you not
645 provide a configure target when configuring a native compiler.
648 @var{target} must be specified as @option{--target=@var{target}}
649 when configuring a cross compiler; examples of valid targets would be
650 m68k-elf, sh-elf, etc.
653 Specifying just @var{target} instead of @option{--target=@var{target}}
654 implies that the host defaults to @var{target}.
658 @heading Options specification
660 Use @var{options} to override several configure time options for
661 GCC@. A list of supported @var{options} follows; @samp{configure
662 --help} may list other options, but those not listed below may not
663 work and should not normally be used.
665 Note that each @option{--enable} option has a corresponding
666 @option{--disable} option and that each @option{--with} option has a
667 corresponding @option{--without} option.
670 @item --prefix=@var{dirname}
671 Specify the toplevel installation
672 directory. This is the recommended way to install the tools into a directory
673 other than the default. The toplevel installation directory defaults to
676 We @strong{highly} recommend against @var{dirname} being the same or a
677 subdirectory of @var{objdir} or vice versa. If specifying a directory
678 beneath a user's home directory tree, some shells will not expand
679 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
682 The following standard @command{autoconf} options are supported. Normally you
683 should not need to use these options.
685 @item --exec-prefix=@var{dirname}
686 Specify the toplevel installation directory for architecture-dependent
687 files. The default is @file{@var{prefix}}.
689 @item --bindir=@var{dirname}
690 Specify the installation directory for the executables called by users
691 (such as @command{gcc} and @command{g++}). The default is
692 @file{@var{exec-prefix}/bin}.
694 @item --libdir=@var{dirname}
695 Specify the installation directory for object code libraries and
696 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
698 @item --libexecdir=@var{dirname}
699 Specify the installation directory for internal executables of GCC@.
700 The default is @file{@var{exec-prefix}/libexec}.
702 @item --with-slibdir=@var{dirname}
703 Specify the installation directory for the shared libgcc library. The
704 default is @file{@var{libdir}}.
706 @item --infodir=@var{dirname}
707 Specify the installation directory for documentation in info format.
708 The default is @file{@var{prefix}/info}.
710 @item --datadir=@var{dirname}
711 Specify the installation directory for some architecture-independent
712 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
714 @item --mandir=@var{dirname}
715 Specify the installation directory for manual pages. The default is
716 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
717 the full GCC manuals, which are provided in Texinfo format. The manpages
718 are derived by an automatic conversion process from parts of the full
721 @item --with-gxx-include-dir=@var{dirname}
723 the installation directory for G++ header files. The default is
724 @file{@var{prefix}/include/c++/@var{version}}.
728 @item --program-prefix=@var{prefix}
729 GCC supports some transformations of the names of its programs when
730 installing them. This option prepends @var{prefix} to the names of
731 programs to install in @var{bindir} (see above). For example, specifying
732 @option{--program-prefix=foo-} would result in @samp{gcc}
733 being installed as @file{/usr/local/bin/foo-gcc}.
735 @item --program-suffix=@var{suffix}
736 Appends @var{suffix} to the names of programs to install in @var{bindir}
737 (see above). For example, specifying @option{--program-suffix=-3.1}
738 would result in @samp{gcc} being installed as
739 @file{/usr/local/bin/gcc-3.1}.
741 @item --program-transform-name=@var{pattern}
742 Applies the @samp{sed} script @var{pattern} to be applied to the names
743 of programs to install in @var{bindir} (see above). @var{pattern} has to
744 consist of one or more basic @samp{sed} editing commands, separated by
745 semicolons. For example, if you want the @samp{gcc} program name to be
746 transformed to the installed program @file{/usr/local/bin/myowngcc} and
747 the @samp{g++} program name to be transformed to
748 @file{/usr/local/bin/gspecial++} without changing other program names,
749 you could use the pattern
750 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
751 to achieve this effect.
753 All three options can be combined and used together, resulting in more
754 complex conversion patterns. As a basic rule, @var{prefix} (and
755 @var{suffix}) are prepended (appended) before further transformations
756 can happen with a special transformation script @var{pattern}.
758 As currently implemented, this option only takes effect for native
759 builds; cross compiler binaries' names are not transformed even when a
760 transformation is explicitly asked for by one of these options.
762 For native builds, some of the installed programs are also installed
763 with the target alias in front of their name, as in
764 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
765 before the target alias is prepended to the name---so, specifying
766 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
767 resulting binary would be installed as
768 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
770 As a last shortcoming, none of the installed Ada programs are
771 transformed yet, which will be fixed in some time.
773 @item --with-local-prefix=@var{dirname}
775 installation directory for local include files. The default is
776 @file{/usr/local}. Specify this option if you want the compiler to
777 search directory @file{@var{dirname}/include} for locally installed
778 header files @emph{instead} of @file{/usr/local/include}.
780 You should specify @option{--with-local-prefix} @strong{only} if your
781 site has a different convention (not @file{/usr/local}) for where to put
784 The default value for @option{--with-local-prefix} is @file{/usr/local}
785 regardless of the value of @option{--prefix}. Specifying
786 @option{--prefix} has no effect on which directory GCC searches for
787 local header files. This may seem counterintuitive, but actually it is
790 The purpose of @option{--prefix} is to specify where to @emph{install
791 GCC}. The local header files in @file{/usr/local/include}---if you put
792 any in that directory---are not part of GCC@. They are part of other
793 programs---perhaps many others. (GCC installs its own header files in
794 another directory which is based on the @option{--prefix} value.)
796 Both the local-prefix include directory and the GCC-prefix include
797 directory are part of GCC's ``system include'' directories. Although these
798 two directories are not fixed, they need to be searched in the proper
799 order for the correct processing of the include_next directive. The
800 local-prefix include directory is searched before the GCC-prefix
801 include directory. Another characteristic of system include directories
802 is that pedantic warnings are turned off for headers in these directories.
804 Some autoconf macros add @option{-I @var{directory}} options to the
805 compiler command line, to ensure that directories containing installed
806 packages' headers are searched. When @var{directory} is one of GCC's
807 system include directories, GCC will ignore the option so that system
808 directories continue to be processed in the correct order. This
809 may result in a search order different from what was specified but the
810 directory will still be searched.
812 GCC automatically searches for ordinary libraries using
813 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
814 used for both GCC and packages, GCC will automatically search for
815 both headers and libraries. This provides a configuration that is
816 easy to use. GCC behaves in a manner similar to that when it is
817 installed as a system compiler in @file{/usr}.
819 Sites that need to install multiple versions of GCC may not want to
820 use the above simple configuration. It is possible to use the
821 @option{--program-prefix}, @option{--program-suffix} and
822 @option{--program-transform-name} options to install multiple versions
823 into a single directory, but it may be simpler to use different prefixes
824 and the @option{--with-local-prefix} option to specify the location of the
825 site-specific files for each version. It will then be necessary for
826 users to specify explicitly the location of local site libraries
827 (e.g., with @env{LIBRARY_PATH}).
829 The same value can be used for both @option{--with-local-prefix} and
830 @option{--prefix} provided it is not @file{/usr}. This can be used
831 to avoid the default search of @file{/usr/local/include}.
833 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
834 The directory you use for @option{--with-local-prefix} @strong{must not}
835 contain any of the system's standard header files. If it did contain
836 them, certain programs would be miscompiled (including GNU Emacs, on
837 certain targets), because this would override and nullify the header
838 file corrections made by the @command{fixincludes} script.
840 Indications are that people who use this option use it based on mistaken
841 ideas of what it is for. People use it as if it specified where to
842 install part of GCC@. Perhaps they make this assumption because
843 installing GCC creates the directory.
845 @item --enable-shared[=@var{package}[,@dots{}]]
846 Build shared versions of libraries, if shared libraries are supported on
847 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
848 are enabled by default on all platforms that support shared libraries.
850 If a list of packages is given as an argument, build shared libraries
851 only for the listed packages. For other packages, only static libraries
852 will be built. Package names currently recognized in the GCC tree are
853 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
854 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
855 @samp{ada}, @samp{libada}, @samp{libjava} and @samp{libobjc}.
856 Note @samp{libiberty} does not support shared libraries at all.
858 Use @option{--disable-shared} to build only static libraries. Note that
859 @option{--disable-shared} does not accept a list of package names as
860 argument, only @option{--enable-shared} does.
862 @item @anchor{with-gnu-as}--with-gnu-as
863 Specify that the compiler should assume that the
864 assembler it finds is the GNU assembler. However, this does not modify
865 the rules to find an assembler and will result in confusion if the
866 assembler found is not actually the GNU assembler. (Confusion may also
867 result if the compiler finds the GNU assembler but has not been
868 configured with @option{--with-gnu-as}.) If you have more than one
869 assembler installed on your system, you may want to use this option in
870 connection with @option{--with-as=@var{pathname}} or
871 @option{--with-build-time-tools=@var{pathname}}.
873 The following systems are the only ones where it makes a difference
874 whether you use the GNU assembler. On any other system,
875 @option{--with-gnu-as} has no effect.
878 @item @samp{hppa1.0-@var{any}-@var{any}}
879 @item @samp{hppa1.1-@var{any}-@var{any}}
880 @item @samp{sparc-sun-solaris2.@var{any}}
881 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
884 @item @anchor{with-as}--with-as=@var{pathname}
885 Specify that the compiler should use the assembler pointed to by
886 @var{pathname}, rather than the one found by the standard rules to find
887 an assembler, which are:
890 Unless GCC is being built with a cross compiler, check the
891 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
892 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
893 @var{exec-prefix} defaults to @var{prefix}, which
894 defaults to @file{/usr/local} unless overridden by the
895 @option{--prefix=@var{pathname}} switch described above. @var{target}
896 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
897 @var{version} denotes the GCC version, such as 3.0.
900 If the target system is the same that you are building on, check
901 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
905 Check in the @env{PATH} for a tool whose name is prefixed by the
906 target system triple.
909 Check in the @env{PATH} for a tool whose name is not prefixed by the
910 target system triple, if the host and target system triple are
911 the same (in other words, we use a host tool if it can be used for
915 You may want to use @option{--with-as} if no assembler
916 is installed in the directories listed above, or if you have multiple
917 assemblers installed and want to choose one that is not found by the
920 @item @anchor{with-gnu-ld}--with-gnu-ld
921 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
924 @item --with-ld=@var{pathname}
925 Same as @uref{#with-as,,@option{--with-as}}
929 Specify that stabs debugging
930 information should be used instead of whatever format the host normally
931 uses. Normally GCC uses the same debug format as the host system.
933 On MIPS based systems and on Alphas, you must specify whether you want
934 GCC to create the normal ECOFF debugging format, or to use BSD-style
935 stabs passed through the ECOFF symbol table. The normal ECOFF debug
936 format cannot fully handle languages other than C@. BSD stabs format can
937 handle other languages, but it only works with the GNU debugger GDB@.
939 Normally, GCC uses the ECOFF debugging format by default; if you
940 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
942 No matter which default you choose when you configure GCC, the user
943 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
944 the debug format for a particular compilation.
946 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
947 @option{--with-gas} is used. It selects use of stabs debugging
948 information embedded in COFF output. This kind of debugging information
949 supports C++ well; ordinary COFF debugging information does not.
951 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
952 selects use of stabs debugging information embedded in ELF output. The
953 C++ compiler currently (2.6.0) does not support the DWARF debugging
954 information normally used on 386 SVR4 platforms; stabs provide a
955 workable alternative. This requires gas and gdb, as the normal SVR4
956 tools can not generate or interpret stabs.
958 @item --disable-multilib
959 Specify that multiple target
960 libraries to support different target variants, calling
961 conventions, etc.@: should not be built. The default is to build a
962 predefined set of them.
964 Some targets provide finer-grained control over which multilibs are built
965 (e.g., @option{--disable-softfloat}):
971 fpu, 26bit, underscore, interwork, biendian, nofmult.
974 softfloat, m68881, m68000, m68020.
977 single-float, biendian, softfloat.
979 @item powerpc*-*-*, rs6000*-*-*
980 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
985 @item --enable-threads
986 Specify that the target
987 supports threads. This affects the Objective-C compiler and runtime
988 library, and exception handling for other languages like C++ and Java.
989 On some systems, this is the default.
991 In general, the best (and, in many cases, the only known) threading
992 model available will be configured for use. Beware that on some
993 systems, GCC has not been taught what threading models are generally
994 available for the system. In this case, @option{--enable-threads} is an
995 alias for @option{--enable-threads=single}.
997 @item --disable-threads
998 Specify that threading support should be disabled for the system.
999 This is an alias for @option{--enable-threads=single}.
1001 @item --enable-threads=@var{lib}
1003 @var{lib} is the thread support library. This affects the Objective-C
1004 compiler and runtime library, and exception handling for other languages
1005 like C++ and Java. The possibilities for @var{lib} are:
1013 Ada tasking support. For non-Ada programs, this setting is equivalent
1014 to @samp{single}. When used in conjunction with the Ada run time, it
1015 causes GCC to use the same thread primitives as Ada uses. This option
1016 is necessary when using both Ada and the back end exception handling,
1017 which is the default for most Ada targets.
1019 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
1020 that the file needed to support this configuration, @file{gthr-mach.h}, is
1021 missing and thus this setting will cause a known bootstrap failure.)
1023 This is an alias for @samp{single}.
1025 Generic POSIX/Unix98 thread support.
1027 Generic POSIX/Unix95 thread support.
1029 RTEMS thread support.
1031 Disable thread support, should work for all platforms.
1033 Sun Solaris 2 thread support.
1035 VxWorks thread support.
1037 Microsoft Win32 API thread support.
1039 Novell Kernel Services thread support.
1043 Specify that the target supports TLS (Thread Local Storage). Usually
1044 configure can correctly determine if TLS is supported. In cases where
1045 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1046 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1047 the assembler supports TLS but the C library does not, or if the
1048 assumptions made by the configure test are incorrect.
1051 Specify that the target does not support TLS.
1052 This is an alias for @option{--enable-tls=no}.
1054 @item --with-cpu=@var{cpu}
1055 @itemx --with-cpu-32=@var{cpu}
1056 @itemx --with-cpu-64=@var{cpu}
1057 Specify which cpu variant the compiler should generate code for by default.
1058 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1059 This option is only supported on some targets, including ARM, i386, M68k,
1060 PowerPC, and SPARC@. The @option{--with-cpu-32} and
1061 @option{--with-cpu-64} options specify separate default CPUs for
1062 32-bit and 64-bit modes; these options are only supported for i386 and
1065 @item --with-schedule=@var{cpu}
1066 @itemx --with-arch=@var{cpu}
1067 @itemx --with-arch-32=@var{cpu}
1068 @itemx --with-arch-64=@var{cpu}
1069 @itemx --with-tune=@var{cpu}
1070 @itemx --with-tune-32=@var{cpu}
1071 @itemx --with-tune-64=@var{cpu}
1072 @itemx --with-abi=@var{abi}
1073 @itemx --with-fpu=@var{type}
1074 @itemx --with-float=@var{type}
1075 These configure options provide default values for the @option{-mschedule=},
1076 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1077 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1078 @option{--with-cpu}, which switches will be accepted and acceptable values
1079 of the arguments depend on the target.
1081 @item --with-mode=@var{mode}
1082 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1083 This option is only supported on ARM targets.
1085 @item --with-divide=@var{type}
1086 Specify how the compiler should generate code for checking for
1087 division by zero. This option is only supported on the MIPS target.
1088 The possibilities for @var{type} are:
1091 Division by zero checks use conditional traps (this is the default on
1092 systems that support conditional traps).
1094 Division by zero checks use the break instruction.
1097 @c If you make --with-llsc the default for additional targets,
1098 @c update the --with-llsc description in the MIPS section below.
1101 On MIPS targets, make @option{-mllsc} the default when no
1102 @option{-mno-lsc} option is passed. This is the default for
1103 Linux-based targets, as the kernel will emulate them if the ISA does
1106 @item --without-llsc
1107 On MIPS targets, make @option{-mno-llsc} the default when no
1108 @option{-mllsc} option is passed.
1110 @item --with-mips-plt
1111 On MIPS targets, make use of copy relocations and PLTs.
1112 These features are extensions to the traditional
1113 SVR4-based MIPS ABIs and require support from GNU binutils
1114 and the runtime C library.
1116 @item --enable-__cxa_atexit
1117 Define if you want to use __cxa_atexit, rather than atexit, to
1118 register C++ destructors for local statics and global objects.
1119 This is essential for fully standards-compliant handling of
1120 destructors, but requires __cxa_atexit in libc. This option is currently
1121 only available on systems with GNU libc. When enabled, this will cause
1122 @option{-fuse-cxa-atexit} to be passed by default.
1124 @item --enable-target-optspace
1126 libraries should be optimized for code space instead of code speed.
1127 This is the default for the m32r platform.
1130 Specify that a user visible @command{cpp} program should not be installed.
1132 @item --with-cpp-install-dir=@var{dirname}
1133 Specify that the user visible @command{cpp} program should be installed
1134 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1136 @item --enable-initfini-array
1137 Force the use of sections @code{.init_array} and @code{.fini_array}
1138 (instead of @code{.init} and @code{.fini}) for constructors and
1139 destructors. Option @option{--disable-initfini-array} has the
1140 opposite effect. If neither option is specified, the configure script
1141 will try to guess whether the @code{.init_array} and
1142 @code{.fini_array} sections are supported and, if they are, use them.
1144 @item --enable-maintainer-mode
1145 The build rules that
1146 regenerate the GCC master message catalog @file{gcc.pot} are normally
1147 disabled. This is because it can only be rebuilt if the complete source
1148 tree is present. If you have changed the sources and want to rebuild the
1149 catalog, configuring with @option{--enable-maintainer-mode} will enable
1150 this. Note that you need a recent version of the @code{gettext} tools
1153 @item --disable-bootstrap
1154 For a native build, the default configuration is to perform
1155 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1156 testing that GCC can compile itself correctly. If you want to disable
1157 this process, you can configure with @option{--disable-bootstrap}.
1159 @item --enable-bootstrap
1160 In special cases, you may want to perform a 3-stage build
1161 even if the target and host triplets are different.
1162 This could happen when the host can run code compiled for
1163 the target (e.g.@: host is i686-linux, target is i486-linux).
1164 Starting from GCC 4.2, to do this you have to configure explicitly
1165 with @option{--enable-bootstrap}.
1167 @item --enable-generated-files-in-srcdir
1168 Neither the .c and .h files that are generated from Bison and flex nor the
1169 info manuals and man pages that are built from the .texi files are present
1170 in the SVN development tree. When building GCC from that development tree,
1171 or from one of our snapshots, those generated files are placed in your
1172 build directory, which allows for the source to be in a readonly
1175 If you configure with @option{--enable-generated-files-in-srcdir} then those
1176 generated files will go into the source directory. This is mainly intended
1177 for generating release or prerelease tarballs of the GCC sources, since it
1178 is not a requirement that the users of source releases to have flex, Bison,
1181 @item --enable-version-specific-runtime-libs
1183 that runtime libraries should be installed in the compiler specific
1184 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1185 addition, @samp{libstdc++}'s include files will be installed into
1186 @file{@var{libdir}} unless you overruled it by using
1187 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1188 particularly useful if you intend to use several versions of GCC in
1189 parallel. This is currently supported by @samp{libgfortran},
1190 @samp{libjava}, @samp{libmudflap}, @samp{libstdc++}, and @samp{libobjc}.
1192 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1193 Specify that only a particular subset of compilers and
1194 their runtime libraries should be built. For a list of valid values for
1195 @var{langN} you can issue the following command in the
1196 @file{gcc} directory of your GCC source tree:@*
1198 grep language= */config-lang.in
1200 Currently, you can use any of the following:
1201 @code{all}, @code{ada}, @code{c}, @code{c++}, @code{fortran}, @code{java},
1202 @code{objc}, @code{obj-c++}.
1203 Building the Ada compiler has special requirements, see below.
1204 If you do not pass this flag, or specify the option @code{all}, then all
1205 default languages available in the @file{gcc} sub-tree will be configured.
1206 Ada and Objective-C++ are not default languages; the rest are.
1207 Re-defining @code{LANGUAGES} when calling @samp{make} @strong{does not}
1208 work anymore, as those language sub-directories might not have been
1211 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1212 Specify that a particular subset of compilers and their runtime
1213 libraries should be built with the system C compiler during stage 1 of
1214 the bootstrap process, rather than only in later stages with the
1215 bootstrapped C compiler. The list of valid values is the same as for
1216 @option{--enable-languages}, and the option @code{all} will select all
1217 of the languages enabled by @option{--enable-languages}. This option is
1218 primarily useful for GCC development; for instance, when a development
1219 version of the compiler cannot bootstrap due to compiler bugs, or when
1220 one is debugging front ends other than the C front end. When this
1221 option is used, one can then build the target libraries for the
1222 specified languages with the stage-1 compiler by using @command{make
1223 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1224 for the specified languages using @command{make stage1-start check-gcc}.
1226 @item --disable-libada
1227 Specify that the run-time libraries and tools used by GNAT should not
1228 be built. This can be useful for debugging, or for compatibility with
1229 previous Ada build procedures, when it was required to explicitly
1230 do a @samp{make -C gcc gnatlib_and_tools}.
1232 @item --disable-libssp
1233 Specify that the run-time libraries for stack smashing protection
1234 should not be built.
1236 @item --disable-libgomp
1237 Specify that the run-time libraries used by GOMP should not be built.
1240 Specify that the compiler should
1241 use DWARF 2 debugging information as the default.
1243 @item --enable-targets=all
1244 @itemx --enable-targets=@var{target_list}
1245 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1246 These are compilers that are able to generate either 64-bit or 32-bit
1247 code. Typically, the corresponding 32-bit target, e.g.@:
1248 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1249 option enables the 32-bit target to be a bi-arch compiler, which is
1250 useful when you want a bi-arch compiler that defaults to 32-bit, and
1251 you are building a bi-arch or multi-arch binutils in a combined tree.
1252 Currently, this option only affects sparc-linux, powerpc-linux and
1255 @item --enable-secureplt
1256 This option enables @option{-msecure-plt} by default for powerpc-linux.
1258 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1259 Using the GNU Compiler Collection (GCC)},
1262 See ``RS/6000 and PowerPC Options'' in the main manual
1266 This option enables @option{-mcld} by default for 32-bit x86 targets.
1268 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1269 Using the GNU Compiler Collection (GCC)},
1272 See ``i386 and x86-64 Options'' in the main manual
1275 @item --enable-win32-registry
1276 @itemx --enable-win32-registry=@var{key}
1277 @itemx --disable-win32-registry
1278 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1279 to look up installations paths in the registry using the following key:
1282 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1285 @var{key} defaults to GCC version number, and can be overridden by the
1286 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1287 who use custom installers are encouraged to provide a different key,
1288 perhaps one comprised of vendor name and GCC version number, to
1289 avoid conflict with existing installations. This feature is enabled
1290 by default, and can be disabled by @option{--disable-win32-registry}
1291 option. This option has no effect on the other hosts.
1294 Specify that the machine does not have a floating point unit. This
1295 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1296 system, @option{--nfp} has no effect.
1298 @item --enable-werror
1299 @itemx --disable-werror
1300 @itemx --enable-werror=yes
1301 @itemx --enable-werror=no
1302 When you specify this option, it controls whether certain files in the
1303 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1304 If you don't specify it, @option{-Werror} is turned on for the main
1305 development trunk. However it defaults to off for release branches and
1306 final releases. The specific files which get @option{-Werror} are
1307 controlled by the Makefiles.
1309 @item --enable-checking
1310 @itemx --enable-checking=@var{list}
1311 When you specify this option, the compiler is built to perform internal
1312 consistency checks of the requested complexity. This does not change the
1313 generated code, but adds error checking within the compiler. This will
1314 slow down the compiler and may only work properly if you are building
1315 the compiler with GCC@. This is @samp{yes} by default when building
1316 from SVN or snapshots, but @samp{release} for releases. The default
1317 for building the stage1 compiler is @samp{yes}. More control
1318 over the checks may be had by specifying @var{list}. The categories of
1319 checks available are @samp{yes} (most common checks
1320 @samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at
1321 all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest
1322 checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1323 Individual checks can be enabled with these flags @samp{assert},
1324 @samp{df}, @samp{fold}, @samp{gc}, @samp{gcac} @samp{misc}, @samp{rtl},
1325 @samp{rtlflag}, @samp{runtime}, @samp{tree}, and @samp{valgrind}.
1327 The @samp{valgrind} check requires the external @command{valgrind}
1328 simulator, available from @uref{http://valgrind.org/}. The
1329 @samp{df}, @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive.
1330 To disable all checking, @samp{--disable-checking} or
1331 @samp{--enable-checking=none} must be explicitly requested. Disabling
1332 assertions will make the compiler and runtime slightly faster but
1333 increase the risk of undetected internal errors causing wrong code to be
1336 @item --disable-stage1-checking
1337 @item --enable-stage1-checking
1338 @itemx --enable-stage1-checking=@var{list}
1339 If no @option{--enable-checking} option is specified the stage1
1340 compiler will be built with @samp{yes} checking enabled, otherwise
1341 the stage1 checking flags are the same as specified by
1342 @option{--enable-checking}. To build the stage1 compiler with
1343 different checking options use @option{--enable-stage1-checking}.
1344 The list of checking options is the same as for @option{--enable-checking}.
1345 If your system is too slow or too small to bootstrap a released compiler
1346 with checking for stage1 enabled, you can use @samp{--disable-stage1-checking}
1347 to disable checking for the stage1 compiler.
1349 @item --enable-coverage
1350 @itemx --enable-coverage=@var{level}
1351 With this option, the compiler is built to collect self coverage
1352 information, every time it is run. This is for internal development
1353 purposes, and only works when the compiler is being built with gcc. The
1354 @var{level} argument controls whether the compiler is built optimized or
1355 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1356 want to disable optimization, for performance analysis you want to
1357 enable optimization. When coverage is enabled, the default level is
1358 without optimization.
1360 @item --enable-gather-detailed-mem-stats
1361 When this option is specified more detailed information on memory
1362 allocation is gathered. This information is printed when using
1363 @option{-fmem-report}.
1366 @itemx --with-gc=@var{choice}
1367 With this option you can specify the garbage collector implementation
1368 used during the compilation process. @var{choice} can be one of
1369 @samp{page} and @samp{zone}, where @samp{page} is the default.
1372 @itemx --disable-nls
1373 The @option{--enable-nls} option enables Native Language Support (NLS),
1374 which lets GCC output diagnostics in languages other than American
1375 English. Native Language Support is enabled by default if not doing a
1376 canadian cross build. The @option{--disable-nls} option disables NLS@.
1378 @item --with-included-gettext
1379 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1380 procedure to prefer its copy of GNU @command{gettext}.
1382 @item --with-catgets
1383 If NLS is enabled, and if the host lacks @code{gettext} but has the
1384 inferior @code{catgets} interface, the GCC build procedure normally
1385 ignores @code{catgets} and instead uses GCC's copy of the GNU
1386 @code{gettext} library. The @option{--with-catgets} option causes the
1387 build procedure to use the host's @code{catgets} in this situation.
1389 @item --with-libiconv-prefix=@var{dir}
1390 Search for libiconv header files in @file{@var{dir}/include} and
1391 libiconv library files in @file{@var{dir}/lib}.
1393 @item --enable-obsolete
1394 Enable configuration for an obsoleted system. If you attempt to
1395 configure GCC for a system (build, host, or target) which has been
1396 obsoleted, and you do not specify this flag, configure will halt with an
1399 All support for systems which have been obsoleted in one release of GCC
1400 is removed entirely in the next major release, unless someone steps
1401 forward to maintain the port.
1403 @item --enable-decimal-float
1404 @itemx --enable-decimal-float=yes
1405 @itemx --enable-decimal-float=no
1406 @itemx --enable-decimal-float=bid
1407 @itemx --enable-decimal-float=dpd
1408 @itemx --disable-decimal-float
1409 Enable (or disable) support for the C decimal floating point extension
1410 that is in the IEEE 754-2008 standard. This is enabled by default only
1411 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
1412 support it, but require the user to specifically enable it. You can
1413 optionally control which decimal floating point format is used (either
1414 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
1415 format is default on i386 and x86_64 systems, and the @samp{dpd}
1416 (densely packed decimal) format is default on PowerPC systems.
1418 @item --enable-fixed-point
1419 @itemx --disable-fixed-point
1420 Enable (or disable) support for C fixed-point arithmetic.
1421 This option is enabled by default for some targets (such as MIPS) which
1422 have hardware-support for fixed-point operations. On other targets, you
1423 may enable this option manually.
1425 @item --with-long-double-128
1426 Specify if @code{long double} type should be 128-bit by default on selected
1427 GNU/Linux architectures. If using @code{--without-long-double-128},
1428 @code{long double} will be by default 64-bit, the same as @code{double} type.
1429 When neither of these configure options are used, the default will be
1430 128-bit @code{long double} when built against GNU C Library 2.4 and later,
1431 64-bit @code{long double} otherwise.
1433 @item --with-gmp=@var{pathname}
1434 @itemx --with-gmp-include=@var{pathname}
1435 @itemx --with-gmp-lib=@var{pathname}
1436 @itemx --with-mpfr=@var{pathname}
1437 @itemx --with-mpfr-include=@var{pathname}
1438 @itemx --with-mpfr-lib=@var{pathname}
1439 If you do not have GMP (the GNU Multiple Precision library) and the
1440 MPFR Libraries installed in a standard location and you want to build
1441 GCC, you can explicitly specify the directory where they are installed
1442 (@samp{--with-gmp=@var{gmpinstalldir}},
1443 @samp{--with-mpfr=@var{mpfrinstalldir}}). The
1444 @option{--with-gmp=@var{gmpinstalldir}} option is shorthand for
1445 @option{--with-gmp-lib=@var{gmpinstalldir}/lib} and
1446 @option{--with-gmp-include=@var{gmpinstalldir}/include}. Likewise the
1447 @option{--with-mpfr=@var{mpfrinstalldir}} option is shorthand for
1448 @option{--with-mpfr-lib=@var{mpfrinstalldir}/lib} and
1449 @option{--with-mpfr-include=@var{mpfrinstalldir}/include}. If these
1450 shorthand assumptions are not correct, you can use the explicit
1451 include and lib options directly.
1453 @item --with-ppl=@var{pathname}
1454 @itemx --with-ppl-include=@var{pathname}
1455 @itemx --with-ppl-lib=@var{pathname}
1456 @itemx --with-cloog=@var{pathname}
1457 @itemx --with-cloog-include=@var{pathname}
1458 @itemx --with-cloog-lib=@var{pathname}
1459 If you do not have PPL (the Parma Polyhedra Library) and the CLooG
1460 libraries installed in a standard location and you want to build GCC,
1461 you can explicitly specify the directory where they are installed
1462 (@samp{--with-ppl=@var{pplinstalldir}},
1463 @samp{--with-cloog=@var{clooginstalldir}}). The
1464 @option{--with-ppl=@var{pplinstalldir}} option is shorthand for
1465 @option{--with-ppl-lib=@var{pplinstalldir}/lib} and
1466 @option{--with-ppl-include=@var{pplinstalldir}/include}. Likewise the
1467 @option{--with-cloog=@var{clooginstalldir}} option is shorthand for
1468 @option{--with-cloog-lib=@var{clooginstalldir}/lib} and
1469 @option{--with-cloog-include=@var{clooginstalldir}/include}. If these
1470 shorthand assumptions are not correct, you can use the explicit
1471 include and lib options directly.
1473 @item --with-host-libstdcxx=@var{linker-args}
1474 If you are linking with a static copy of PPL, you can use this option
1475 to specify how the linker should find the standard C++ library used
1476 internally by PPL. Typical values of @var{linker-args} might be
1477 @samp{-lstdc++} or @samp{-Wl,-Bstatic,-lstdc++,-Bdynamic -lm}. If you are
1478 linking with a shared copy of PPL, you probably do not need this
1479 option; shared library dependencies will cause the linker to search
1480 for the standard C++ library automatically.
1482 @item --with-debug-prefix-map=@var{map}
1483 Convert source directory names using @option{-fdebug-prefix-map} when
1484 building runtime libraries. @samp{@var{map}} is a space-separated
1485 list of maps of the form @samp{@var{old}=@var{new}}.
1489 @subheading Cross-Compiler-Specific Options
1490 The following options only apply to building cross compilers.
1492 @item --with-sysroot
1493 @itemx --with-sysroot=@var{dir}
1494 Tells GCC to consider @var{dir} as the root of a tree that contains a
1495 (subset of) the root filesystem of the target operating system.
1496 Target system headers, libraries and run-time object files will be
1497 searched in there. The specified directory is not copied into the
1498 install tree, unlike the options @option{--with-headers} and
1499 @option{--with-libs} that this option obsoletes. The default value,
1500 in case @option{--with-sysroot} is not given an argument, is
1501 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1502 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1503 the GCC binaries if the installation tree is moved.
1505 @item --with-build-sysroot
1506 @itemx --with-build-sysroot=@var{dir}
1507 Tells GCC to consider @var{dir} as the system root (see
1508 @option{--with-sysroot}) while building target libraries, instead of
1509 the directory specified with @option{--with-sysroot}. This option is
1510 only useful when you are already using @option{--with-sysroot}. You
1511 can use @option{--with-build-sysroot} when you are configuring with
1512 @option{--prefix} set to a directory that is different from the one in
1513 which you are installing GCC and your target libraries.
1515 This option affects the system root for the compiler used to build
1516 target libraries (which runs on the build system); it does not affect
1517 the compiler which is used to build GCC itself.
1519 @item --with-headers
1520 @itemx --with-headers=@var{dir}
1521 Deprecated in favor of @option{--with-sysroot}.
1522 Specifies that target headers are available when building a cross compiler.
1523 The @var{dir} argument specifies a directory which has the target include
1524 files. These include files will be copied into the @file{gcc} install
1525 directory. @emph{This option with the @var{dir} argument is required} when
1526 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1527 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1528 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1529 will be run on these files to make them compatible with GCC@.
1531 @item --without-headers
1532 Tells GCC not use any target headers from a libc when building a cross
1533 compiler. When crossing to GNU/Linux, you need the headers so GCC
1534 can build the exception handling for libgcc.
1537 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1538 Deprecated in favor of @option{--with-sysroot}.
1539 Specifies a list of directories which contain the target runtime
1540 libraries. These libraries will be copied into the @file{gcc} install
1541 directory. If the directory list is omitted, this option has no
1545 Specifies that @samp{newlib} is
1546 being used as the target C library. This causes @code{__eprintf} to be
1547 omitted from @file{libgcc.a} on the assumption that it will be provided by
1550 @item --with-build-time-tools=@var{dir}
1551 Specifies where to find the set of target tools (assembler, linker, etc.)
1552 that will be used while building GCC itself. This option can be useful
1553 if the directory layouts are different between the system you are building
1554 GCC on, and the system where you will deploy it.
1556 For example, on a @option{ia64-hp-hpux} system, you may have the GNU
1557 assembler and linker in @file{/usr/bin}, and the native tools in a
1558 different path, and build a toolchain that expects to find the
1559 native tools in @file{/usr/bin}.
1561 When you use this option, you should ensure that @var{dir} includes
1562 @command{ar}, @command{as}, @command{ld}, @command{nm},
1563 @command{ranlib} and @command{strip} if necessary, and possibly
1564 @command{objdump}. Otherwise, GCC may use an inconsistent set of
1568 @subheading Java-Specific Options
1570 The following option applies to the build of the Java front end.
1573 @item --disable-libgcj
1574 Specify that the run-time libraries
1575 used by GCJ should not be built. This is useful in case you intend
1576 to use GCJ with some other run-time, or you're going to install it
1577 separately, or it just happens not to build on your particular
1578 machine. In general, if the Java front end is enabled, the GCJ
1579 libraries will be enabled too, unless they're known to not work on
1580 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1581 may need to port it; in this case, before modifying the top-level
1582 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1583 you may use @option{--enable-libgcj} to override the default.
1587 The following options apply to building @samp{libgcj}.
1589 @subsubheading General Options
1592 @item --enable-java-maintainer-mode
1593 By default the @samp{libjava} build will not attempt to compile the
1594 @file{.java} source files to @file{.class}. Instead, it will use the
1595 @file{.class} files from the source tree. If you use this option you
1596 must have executables named @command{ecj1} and @command{gjavah} in your path
1597 for use by the build. You must use this option if you intend to
1598 modify any @file{.java} files in @file{libjava}.
1600 @item --with-java-home=@var{dirname}
1601 This @samp{libjava} option overrides the default value of the
1602 @samp{java.home} system property. It is also used to set
1603 @samp{sun.boot.class.path} to @file{@var{dirname}/lib/rt.jar}. By
1604 default @samp{java.home} is set to @file{@var{prefix}} and
1605 @samp{sun.boot.class.path} to
1606 @file{@var{datadir}/java/libgcj-@var{version}.jar}.
1608 @item --with-ecj-jar=@var{filename}
1609 This option can be used to specify the location of an external jar
1610 file containing the Eclipse Java compiler. A specially modified
1611 version of this compiler is used by @command{gcj} to parse
1612 @file{.java} source files. If this option is given, the
1613 @samp{libjava} build will create and install an @file{ecj1} executable
1614 which uses this jar file at runtime.
1616 If this option is not given, but an @file{ecj.jar} file is found in
1617 the topmost source tree at configure time, then the @samp{libgcj}
1618 build will create and install @file{ecj1}, and will also install the
1619 discovered @file{ecj.jar} into a suitable place in the install tree.
1621 If @file{ecj1} is not installed, then the user will have to supply one
1622 on his path in order for @command{gcj} to properly parse @file{.java}
1623 source files. A suitable jar is available from
1624 @uref{ftp://sourceware.org/pub/java/}.
1626 @item --disable-getenv-properties
1627 Don't set system properties from @env{GCJ_PROPERTIES}.
1629 @item --enable-hash-synchronization
1630 Use a global hash table for monitor locks. Ordinarily,
1631 @samp{libgcj}'s @samp{configure} script automatically makes
1632 the correct choice for this option for your platform. Only use
1633 this if you know you need the library to be configured differently.
1635 @item --enable-interpreter
1636 Enable the Java interpreter. The interpreter is automatically
1637 enabled by default on all platforms that support it. This option
1638 is really only useful if you want to disable the interpreter
1639 (using @option{--disable-interpreter}).
1641 @item --disable-java-net
1642 Disable java.net. This disables the native part of java.net only,
1643 using non-functional stubs for native method implementations.
1645 @item --disable-jvmpi
1646 Disable JVMPI support.
1648 @item --disable-libgcj-bc
1649 Disable BC ABI compilation of certain parts of libgcj. By default,
1650 some portions of libgcj are compiled with @option{-findirect-dispatch}
1651 and @option{-fno-indirect-classes}, allowing them to be overridden at
1654 If @option{--disable-libgcj-bc} is specified, libgcj is built without
1655 these options. This allows the compile-time linker to resolve
1656 dependencies when statically linking to libgcj. However it makes it
1657 impossible to override the affected portions of libgcj at run-time.
1659 @item --enable-reduced-reflection
1660 Build most of libgcj with @option{-freduced-reflection}. This reduces
1661 the size of libgcj at the expense of not being able to do accurate
1662 reflection on the classes it contains. This option is safe if you
1663 know that code using libgcj will never use reflection on the standard
1664 runtime classes in libgcj (including using serialization, RMI or CORBA).
1667 Enable runtime eCos target support.
1669 @item --without-libffi
1670 Don't use @samp{libffi}. This will disable the interpreter and JNI
1671 support as well, as these require @samp{libffi} to work.
1673 @item --enable-libgcj-debug
1674 Enable runtime debugging code.
1676 @item --enable-libgcj-multifile
1677 If specified, causes all @file{.java} source files to be
1678 compiled into @file{.class} files in one invocation of
1679 @samp{gcj}. This can speed up build time, but is more
1680 resource-intensive. If this option is unspecified or
1681 disabled, @samp{gcj} is invoked once for each @file{.java}
1682 file to compile into a @file{.class} file.
1684 @item --with-libiconv-prefix=DIR
1685 Search for libiconv in @file{DIR/include} and @file{DIR/lib}.
1687 @item --enable-sjlj-exceptions
1688 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1689 @samp{configure} ordinarily picks the correct value based on the platform.
1690 Only use this option if you are sure you need a different setting.
1692 @item --with-system-zlib
1693 Use installed @samp{zlib} rather than that included with GCC@.
1695 @item --with-win32-nlsapi=ansi, unicows or unicode
1696 Indicates how MinGW @samp{libgcj} translates between UNICODE
1697 characters and the Win32 API@.
1699 @item --enable-java-home
1700 If enabled, this creates a JPackage compatible SDK environment during install.
1701 Note that if --enable-java-home is used, --with-arch-directory=ARCH must also
1704 @item --with-arch-directory=ARCH
1705 Specifies the name to use for the @file{jre/lib/ARCH} directory in the SDK
1706 environment created when --enable-java-home is passed. Typical names for this
1707 directory include i386, amd64, ia64, etc.
1709 @item --with-os-directory=DIR
1710 Specifies the OS directory for the SDK include directory. This is set to auto
1711 detect, and is typically 'linux'.
1713 @item --with-origin-name=NAME
1714 Specifies the JPackage origin name. This defaults to the 'gcj' in
1717 @item --with-arch-suffix=SUFFIX
1718 Specifies the suffix for the sdk directory. Defaults to the empty string.
1719 Examples include '.x86_64' in 'java-1.5.0-gcj-1.5.0.0.x86_64'.
1721 @item --with-jvm-root-dir=DIR
1722 Specifies where to install the SDK. Default is $(prefix)/lib/jvm.
1724 @item --with-jvm-jar-dir=DIR
1725 Specifies where to install jars. Default is $(prefix)/lib/jvm-exports.
1727 @item --with-python-dir=DIR
1728 Specifies where to install the Python modules used for aot-compile. DIR should
1729 not include the prefix used in installation. For example, if the Python modules
1730 are to be installed in /usr/lib/python2.5/site-packages, then
1731 --with-python-dir=/lib/python2.5/site-packages should be passed. If this is
1732 not specified, then the Python modules are installed in $(prefix)/share/python.
1734 @item --enable-aot-compile-rpm
1735 Adds aot-compile-rpm to the list of installed scripts.
1739 Use the single-byte @code{char} and the Win32 A functions natively,
1740 translating to and from UNICODE when using these functions. If
1741 unspecified, this is the default.
1744 Use the @code{WCHAR} and Win32 W functions natively. Adds
1745 @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}.
1746 @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines
1747 running built executables. @file{libunicows.a}, an open-source
1748 import library around Microsoft's @code{unicows.dll}, is obtained from
1749 @uref{http://libunicows.sourceforge.net/}, which also gives details
1750 on getting @file{unicows.dll} from Microsoft.
1753 Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not}
1754 add @code{-lunicows} to @file{libgcj.spec}. The built executables will
1755 only run on Microsoft Windows NT and above.
1759 @subsubheading AWT-Specific Options
1763 Use the X Window System.
1765 @item --enable-java-awt=PEER(S)
1766 Specifies the AWT peer library or libraries to build alongside
1767 @samp{libgcj}. If this option is unspecified or disabled, AWT
1768 will be non-functional. Current valid values are @option{gtk} and
1769 @option{xlib}. Multiple libraries should be separated by a
1770 comma (i.e.@: @option{--enable-java-awt=gtk,xlib}).
1772 @item --enable-gtk-cairo
1773 Build the cairo Graphics2D implementation on GTK@.
1775 @item --enable-java-gc=TYPE
1776 Choose garbage collector. Defaults to @option{boehm} if unspecified.
1778 @item --disable-gtktest
1779 Do not try to compile and run a test GTK+ program.
1781 @item --disable-glibtest
1782 Do not try to compile and run a test GLIB program.
1784 @item --with-libart-prefix=PFX
1785 Prefix where libart is installed (optional).
1787 @item --with-libart-exec-prefix=PFX
1788 Exec prefix where libart is installed (optional).
1790 @item --disable-libarttest
1791 Do not try to compile and run a test libart program.
1800 @uref{./index.html,,Return to the GCC Installation page}
1804 @c ***Building****************************************************************
1806 @comment node-name, next, previous, up
1807 @node Building, Testing, Configuration, Installing GCC
1813 @cindex Installing GCC: Building
1815 Now that GCC is configured, you are ready to build the compiler and
1818 Some commands executed when making the compiler may fail (return a
1819 nonzero status) and be ignored by @command{make}. These failures, which
1820 are often due to files that were not found, are expected, and can safely
1823 It is normal to have compiler warnings when compiling certain files.
1824 Unless you are a GCC developer, you can generally ignore these warnings
1825 unless they cause compilation to fail. Developers should attempt to fix
1826 any warnings encountered, however they can temporarily continue past
1827 warnings-as-errors by specifying the configure flag
1828 @option{--disable-werror}.
1830 On certain old systems, defining certain environment variables such as
1831 @env{CC} can interfere with the functioning of @command{make}.
1833 If you encounter seemingly strange errors when trying to build the
1834 compiler in a directory other than the source directory, it could be
1835 because you have previously configured the compiler in the source
1836 directory. Make sure you have done all the necessary preparations.
1838 If you build GCC on a BSD system using a directory stored in an old System
1839 V file system, problems may occur in running @command{fixincludes} if the
1840 System V file system doesn't support symbolic links. These problems
1841 result in a failure to fix the declaration of @code{size_t} in
1842 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1843 that type mismatches occur, this could be the cause.
1845 The solution is not to use such a directory for building GCC@.
1847 Similarly, when building from SVN or snapshots, or if you modify
1848 @file{*.l} files, you need the Flex lexical analyzer generator
1849 installed. If you do not modify @file{*.l} files, releases contain
1850 the Flex-generated files and you do not need Flex installed to build
1851 them. There is still one Flex-based lexical analyzer (part of the
1852 build machinery, not of GCC itself) that is used even if you only
1853 build the C front end.
1855 When building from SVN or snapshots, or if you modify Texinfo
1856 documentation, you need version 4.7 or later of Texinfo installed if you
1857 want Info documentation to be regenerated. Releases contain Info
1858 documentation pre-built for the unmodified documentation in the release.
1860 @section Building a native compiler
1862 For a native build, the default configuration is to perform
1863 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
1864 This will build the entire GCC system and ensure that it compiles
1865 itself correctly. It can be disabled with the @option{--disable-bootstrap}
1866 parameter to @samp{configure}, but bootstrapping is suggested because
1867 the compiler will be tested more completely and could also have
1870 The bootstrapping process will complete the following steps:
1874 Build tools necessary to build the compiler.
1877 Perform a 3-stage bootstrap of the compiler. This includes building
1878 three times the target tools for use by the compiler such as binutils
1879 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
1880 individually linked or moved into the top level GCC source tree before
1884 Perform a comparison test of the stage2 and stage3 compilers.
1887 Build runtime libraries using the stage3 compiler from the previous step.
1891 If you are short on disk space you might consider @samp{make
1892 bootstrap-lean} instead. The sequence of compilation is the
1893 same described above, but object files from the stage1 and
1894 stage2 of the 3-stage bootstrap of the compiler are deleted as
1895 soon as they are no longer needed.
1897 If you wish to use non-default GCC flags when compiling the stage2
1898 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
1899 doing @samp{make}. For example, if you want to save additional space
1900 during the bootstrap and in the final installation as well, you can
1901 build the compiler binaries without debugging information as in the
1902 following example. This will save roughly 40% of disk space both for
1903 the bootstrap and the final installation. (Libraries will still contain
1904 debugging information.)
1907 make BOOT_CFLAGS='-O' bootstrap
1910 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
1911 are less well tested here than the default of @samp{-g -O2}, but should
1912 still work. In a few cases, you may find that you need to specify special
1913 flags such as @option{-msoft-float} here to complete the bootstrap; or,
1914 if the native compiler miscompiles the stage1 compiler, you may need
1915 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
1916 of the stage1 compiler that were miscompiled, or by using @samp{make
1917 bootstrap4} to increase the number of stages of bootstrap.
1919 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
1920 Since these are always compiled with the compiler currently being
1921 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
1922 compilation flags, as for non-bootstrapped target libraries.
1923 Again, if the native compiler miscompiles the stage1 compiler, you may
1924 need to work around this by avoiding non-working parts of the stage1
1925 compiler. Use @code{STAGE1_LIBCFLAGS} to this end.
1927 If you used the flag @option{--enable-languages=@dots{}} to restrict
1928 the compilers to be built, only those you've actually enabled will be
1929 built. This will of course only build those runtime libraries, for
1930 which the particular compiler has been built. Please note,
1931 that re-defining @env{LANGUAGES} when calling @samp{make}
1932 @strong{does not} work anymore!
1934 If the comparison of stage2 and stage3 fails, this normally indicates
1935 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1936 a potentially serious bug which you should investigate and report. (On
1937 a few systems, meaningful comparison of object files is impossible; they
1938 always appear ``different''. If you encounter this problem, you will
1939 need to disable comparison in the @file{Makefile}.)
1941 If you do not want to bootstrap your compiler, you can configure with
1942 @option{--disable-bootstrap}. In particular cases, you may want to
1943 bootstrap your compiler even if the target system is not the same as
1944 the one you are building on: for example, you could build a
1945 @code{powerpc-unknown-linux-gnu} toolchain on a
1946 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
1947 @option{--enable-bootstrap} to the configure script.
1950 @section Building a cross compiler
1952 When building a cross compiler, it is not generally possible to do a
1953 3-stage bootstrap of the compiler. This makes for an interesting problem
1954 as parts of GCC can only be built with GCC@.
1956 To build a cross compiler, we first recommend building and installing a
1957 native compiler. You can then use the native GCC compiler to build the
1958 cross compiler. The installed native compiler needs to be GCC version
1961 If the cross compiler is to be built with support for the Java
1962 programming language and the ability to compile .java source files is
1963 desired, the installed native compiler used to build the cross
1964 compiler needs to be the same GCC version as the cross compiler. In
1965 addition the cross compiler needs to be configured with
1966 @option{--with-ecj-jar=@dots{}}.
1968 Assuming you have already installed a native copy of GCC and configured
1969 your cross compiler, issue the command @command{make}, which performs the
1974 Build host tools necessary to build the compiler.
1977 Build target tools for use by the compiler such as binutils (bfd,
1978 binutils, gas, gprof, ld, and opcodes)
1979 if they have been individually linked or moved into the top level GCC source
1980 tree before configuring.
1983 Build the compiler (single stage only).
1986 Build runtime libraries using the compiler from the previous step.
1989 Note that if an error occurs in any step the make process will exit.
1991 If you are not building GNU binutils in the same source tree as GCC,
1992 you will need a cross-assembler and cross-linker installed before
1993 configuring GCC@. Put them in the directory
1994 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
1995 you should put in this directory:
1999 This should be the cross-assembler.
2002 This should be the cross-linker.
2005 This should be the cross-archiver: a program which can manipulate
2006 archive files (linker libraries) in the target machine's format.
2009 This should be a program to construct a symbol table in an archive file.
2012 The installation of GCC will find these programs in that directory,
2013 and copy or link them to the proper place to for the cross-compiler to
2014 find them when run later.
2016 The easiest way to provide these files is to build the Binutils package.
2017 Configure it with the same @option{--host} and @option{--target}
2018 options that you use for configuring GCC, then build and install
2019 them. They install their executables automatically into the proper
2020 directory. Alas, they do not support all the targets that GCC
2023 If you are not building a C library in the same source tree as GCC,
2024 you should also provide the target libraries and headers before
2025 configuring GCC, specifying the directories with
2026 @option{--with-sysroot} or @option{--with-headers} and
2027 @option{--with-libs}. Many targets also require ``start files'' such
2028 as @file{crt0.o} and
2029 @file{crtn.o} which are linked into each executable. There may be several
2030 alternatives for @file{crt0.o}, for use with profiling or other
2031 compilation options. Check your target's definition of
2032 @code{STARTFILE_SPEC} to find out what start files it uses.
2034 @section Building in parallel
2036 GNU Make 3.79 and above, which is necessary to build GCC, support
2037 building in parallel. To activate this, you can use @samp{make -j 2}
2038 instead of @samp{make}. You can also specify a bigger number, and
2039 in most cases using a value greater than the number of processors in
2040 your machine will result in fewer and shorter I/O latency hits, thus
2041 improving overall throughput; this is especially true for slow drives
2042 and network filesystems.
2044 @section Building the Ada compiler
2046 In order to build GNAT, the Ada compiler, you need a working GNAT
2047 compiler (GCC version 3.4 or later).
2048 This includes GNAT tools such as @command{gnatmake} and
2049 @command{gnatlink}, since the Ada front end is written in Ada and
2050 uses some GNAT-specific extensions.
2052 In order to build a cross compiler, it is suggested to install
2053 the new compiler as native first, and then use it to build the cross
2056 @command{configure} does not test whether the GNAT installation works
2057 and has a sufficiently recent version; if too old a GNAT version is
2058 installed, the build will fail unless @option{--enable-languages} is
2059 used to disable building the Ada front end.
2061 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
2062 must not be set when building the Ada compiler, the Ada tools, or the
2063 Ada runtime libraries. You can check that your build environment is clean
2064 by verifying that @samp{gnatls -v} lists only one explicit path in each
2067 @section Building with profile feedback
2069 It is possible to use profile feedback to optimize the compiler itself. This
2070 should result in a faster compiler binary. Experiments done on x86 using gcc
2071 3.3 showed approximately 7 percent speedup on compiling C programs. To
2072 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
2074 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
2075 compiler. This compiler is used to build a @code{stageprofile} compiler
2076 instrumented to collect execution counts of instruction and branch
2077 probabilities. Then runtime libraries are compiled with profile collected.
2078 Finally a @code{stagefeedback} compiler is built using the information collected.
2080 Unlike standard bootstrap, several additional restrictions apply. The
2081 compiler used to build @code{stage1} needs to support a 64-bit integral type.
2082 It is recommended to only use GCC for this. Also parallel make is currently
2083 not supported since collisions in profile collecting may occur.
2090 @uref{./index.html,,Return to the GCC Installation page}
2094 @c ***Testing*****************************************************************
2096 @comment node-name, next, previous, up
2097 @node Testing, Final install, Building, Installing GCC
2101 @chapter Installing GCC: Testing
2104 @cindex Installing GCC: Testing
2107 Before you install GCC, we encourage you to run the testsuites and to
2108 compare your results with results from a similar configuration that have
2109 been submitted to the
2110 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
2111 Some of these archived results are linked from the build status lists
2112 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
2113 reports a successful build runs the testsuites and submits the results.
2114 This step is optional and may require you to download additional software,
2115 but it can give you confidence in your new GCC installation or point out
2116 problems before you install and start using your new GCC@.
2118 First, you must have @uref{download.html,,downloaded the testsuites}.
2119 These are part of the full distribution, but if you downloaded the
2120 ``core'' compiler plus any front ends, you must download the testsuites
2123 Second, you must have the testing tools installed. This includes
2124 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
2125 the DejaGnu site has links to these.
2127 If the directories where @command{runtest} and @command{expect} were
2128 installed are not in the @env{PATH}, you may need to set the following
2129 environment variables appropriately, as in the following example (which
2130 assumes that DejaGnu has been installed under @file{/usr/local}):
2133 TCL_LIBRARY = /usr/local/share/tcl8.0
2134 DEJAGNULIBS = /usr/local/share/dejagnu
2137 (On systems such as Cygwin, these paths are required to be actual
2138 paths, not mounts or links; presumably this is due to some lack of
2139 portability in the DejaGnu code.)
2142 Finally, you can run the testsuite (which may take a long time):
2144 cd @var{objdir}; make -k check
2147 This will test various components of GCC, such as compiler
2148 front ends and runtime libraries. While running the testsuite, DejaGnu
2149 might emit some harmless messages resembling
2150 @samp{WARNING: Couldn't find the global config file.} or
2151 @samp{WARNING: Couldn't find tool init file} that can be ignored.
2153 If you are testing a cross-compiler, you may want to run the testsuite
2154 on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}.
2156 @section How can you run the testsuite on selected tests?
2158 In order to run sets of tests selectively, there are targets
2159 @samp{make check-gcc} and @samp{make check-g++}
2160 in the @file{gcc} subdirectory of the object directory. You can also
2161 just run @samp{make check} in a subdirectory of the object directory.
2164 A more selective way to just run all @command{gcc} execute tests in the
2168 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
2171 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
2172 the testsuite with filenames matching @samp{9805*}, you would use
2175 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
2178 The @file{*.exp} files are located in the testsuite directories of the GCC
2179 source, the most important ones being @file{compile.exp},
2180 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
2181 To get a list of the possible @file{*.exp} files, pipe the
2182 output of @samp{make check} into a file and look at the
2183 @samp{Running @dots{} .exp} lines.
2185 @section Passing options and running multiple testsuites
2187 You can pass multiple options to the testsuite using the
2188 @samp{--target_board} option of DejaGNU, either passed as part of
2189 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
2190 work outside the makefiles. For example,
2193 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
2196 will run the standard @command{g++} testsuites (``unix'' is the target name
2197 for a standard native testsuite situation), passing
2198 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
2199 slashes separate options.
2201 You can run the testsuites multiple times using combinations of options
2202 with a syntax similar to the brace expansion of popular shells:
2205 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
2208 (Note the empty option caused by the trailing comma in the final group.)
2209 The following will run each testsuite eight times using the @samp{arm-sim}
2210 target, as if you had specified all possible combinations yourself:
2213 --target_board=arm-sim/-mhard-float/-O1
2214 --target_board=arm-sim/-mhard-float/-O2
2215 --target_board=arm-sim/-mhard-float/-O3
2216 --target_board=arm-sim/-mhard-float
2217 --target_board=arm-sim/-msoft-float/-O1
2218 --target_board=arm-sim/-msoft-float/-O2
2219 --target_board=arm-sim/-msoft-float/-O3
2220 --target_board=arm-sim/-msoft-float
2223 They can be combined as many times as you wish, in arbitrary ways. This
2227 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
2230 will generate four combinations, all involving @samp{-Wextra}.
2232 The disadvantage to this method is that the testsuites are run in serial,
2233 which is a waste on multiprocessor systems. For users with GNU Make and
2234 a shell which performs brace expansion, you can run the testsuites in
2235 parallel by having the shell perform the combinations and @command{make}
2236 do the parallel runs. Instead of using @samp{--target_board}, use a
2237 special makefile target:
2240 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
2246 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
2249 will run three concurrent ``make-gcc'' testsuites, eventually testing all
2250 ten combinations as described above. Note that this is currently only
2251 supported in the @file{gcc} subdirectory. (To see how this works, try
2252 typing @command{echo} before the example given here.)
2255 @section Additional testing for Java Class Libraries
2257 The Java runtime tests can be executed via @samp{make check}
2258 in the @file{@var{target}/libjava/testsuite} directory in
2261 The @uref{http://sourceware.org/mauve/,,Mauve Project} provides
2262 a suite of tests for the Java Class Libraries. This suite can be run
2263 as part of libgcj testing by placing the Mauve tree within the libjava
2264 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
2265 specifying the location of that tree when invoking @samp{make}, as in
2266 @samp{make MAUVEDIR=~/mauve check}.
2268 @section How to interpret test results
2270 The result of running the testsuite are various @file{*.sum} and @file{*.log}
2271 files in the testsuite subdirectories. The @file{*.log} files contain a
2272 detailed log of the compiler invocations and the corresponding
2273 results, the @file{*.sum} files summarize the results. These summaries
2274 contain status codes for all tests:
2278 PASS: the test passed as expected
2280 XPASS: the test unexpectedly passed
2282 FAIL: the test unexpectedly failed
2284 XFAIL: the test failed as expected
2286 UNSUPPORTED: the test is not supported on this platform
2288 ERROR: the testsuite detected an error
2290 WARNING: the testsuite detected a possible problem
2293 It is normal for some tests to report unexpected failures. At the
2294 current time the testing harness does not allow fine grained control
2295 over whether or not a test is expected to fail. This problem should
2296 be fixed in future releases.
2299 @section Submitting test results
2301 If you want to report the results to the GCC project, use the
2302 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
2305 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
2306 -m gcc-testresults@@gcc.gnu.org |sh
2309 This script uses the @command{Mail} program to send the results, so
2310 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
2311 prepended to the testsuite summary and should contain any special
2312 remarks you have on your results or your build environment. Please
2313 do not edit the testsuite result block or the subject line, as these
2314 messages may be automatically processed.
2321 @uref{./index.html,,Return to the GCC Installation page}
2325 @c ***Final install***********************************************************
2327 @comment node-name, next, previous, up
2328 @node Final install, , Testing, Installing GCC
2330 @ifset finalinstallhtml
2332 @chapter Installing GCC: Final installation
2335 Now that GCC has been built (and optionally tested), you can install it with
2337 cd @var{objdir}; make install
2340 We strongly recommend to install into a target directory where there is
2341 no previous version of GCC present. Also, the GNAT runtime should not
2342 be stripped, as this would break certain features of the debugger that
2343 depend on this debugging information (catching Ada exceptions for
2346 That step completes the installation of GCC; user level binaries can
2347 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
2348 you specified with the @option{--prefix} to configure (or
2349 @file{/usr/local} by default). (If you specified @option{--bindir},
2350 that directory will be used instead; otherwise, if you specified
2351 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
2352 Headers for the C++ and Java libraries are installed in
2353 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
2354 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
2355 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
2356 in info format in @file{@var{infodir}} (normally
2357 @file{@var{prefix}/info}).
2359 When installing cross-compilers, GCC's executables
2360 are not only installed into @file{@var{bindir}}, that
2361 is, @file{@var{exec-prefix}/bin}, but additionally into
2362 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
2363 exists. Typically, such @dfn{tooldirs} hold target-specific
2364 binutils, including assembler and linker.
2366 Installation into a temporary staging area or into a @command{chroot}
2367 jail can be achieved with the command
2370 make DESTDIR=@var{path-to-rootdir} install
2373 @noindent where @var{path-to-rootdir} is the absolute path of
2374 a directory relative to which all installation paths will be
2375 interpreted. Note that the directory specified by @code{DESTDIR}
2376 need not exist yet; it will be created if necessary.
2378 There is a subtle point with tooldirs and @code{DESTDIR}:
2379 If you relocate a cross-compiler installation with
2380 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
2381 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
2382 be filled with duplicated GCC executables only if it already exists,
2383 it will not be created otherwise. This is regarded as a feature,
2384 not as a bug, because it gives slightly more control to the packagers
2385 using the @code{DESTDIR} feature.
2387 If you are bootstrapping a released version of GCC then please
2388 quickly review the build status page for your release, available from
2389 @uref{http://gcc.gnu.org/buildstat.html}.
2390 If your system is not listed for the version of GCC that you built,
2392 @email{gcc@@gcc.gnu.org} indicating
2393 that you successfully built and installed GCC@.
2394 Include the following information:
2398 Output from running @file{@var{srcdir}/config.guess}. Do not send
2399 that file itself, just the one-line output from running it.
2402 The output of @samp{gcc -v} for your newly installed @command{gcc}.
2403 This tells us which version of GCC you built and the options you passed to
2407 Whether you enabled all languages or a subset of them. If you used a
2408 full distribution then this information is part of the configure
2409 options in the output of @samp{gcc -v}, but if you downloaded the
2410 ``core'' compiler plus additional front ends then it isn't apparent
2411 which ones you built unless you tell us about it.
2414 If the build was for GNU/Linux, also include:
2417 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
2418 this information should be available from @file{/etc/issue}.
2421 The version of the Linux kernel, available from @samp{uname --version}
2425 The version of glibc you used; for RPM-based systems like Red Hat,
2426 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
2427 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
2429 For other systems, you can include similar information if you think it is
2433 Any other information that you think would be useful to people building
2434 GCC on the same configuration. The new entry in the build status list
2435 will include a link to the archived copy of your message.
2438 We'd also like to know if the
2440 @ref{Specific, host/target specific installation notes}
2443 @uref{specific.html,,host/target specific installation notes}
2445 didn't include your host/target information or if that information is
2446 incomplete or out of date. Send a note to
2447 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
2449 If you find a bug, please report it following the
2450 @uref{../bugs.html,,bug reporting guidelines}.
2452 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
2453 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
2454 and @TeX{} installed. This creates a number of @file{.dvi} files in
2455 subdirectories of @file{@var{objdir}}; these may be converted for
2456 printing with programs such as @command{dvips}. Alternately, by using
2457 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
2458 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
2459 is included with Texinfo version 4.8 and later. You can also
2460 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
2461 Free Software Foundation}, though such manuals may not be for the most
2462 recent version of GCC@.
2464 If you would like to generate online HTML documentation, do @samp{cd
2465 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
2466 @file{@var{objdir}/gcc/HTML}.
2473 @uref{./index.html,,Return to the GCC Installation page}
2477 @c ***Binaries****************************************************************
2479 @comment node-name, next, previous, up
2480 @node Binaries, Specific, Installing GCC, Top
2484 @chapter Installing GCC: Binaries
2487 @cindex Installing GCC: Binaries
2489 We are often asked about pre-compiled versions of GCC@. While we cannot
2490 provide these for all platforms, below you'll find links to binaries for
2491 various platforms where creating them by yourself is not easy due to various
2494 Please note that we did not create these binaries, nor do we
2495 support them. If you have any problems installing them, please
2496 contact their makers.
2503 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
2506 @uref{http://pware.hvcc.edu,,Hudson Valley Community College Open Source Software for IBM System p};
2509 @uref{http://www.perzl.org/aix,,AIX 5L and 6 Open Source Packages}.
2513 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
2516 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
2517 Development Tools for the Renesas H8/300[HS] Series}.
2523 @uref{http://hpux.cs.utah.edu/,,HP-UX Porting Center};
2526 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
2530 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
2531 Development Tools for the Motorola 68HC11/68HC12}.
2534 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
2535 OpenServer/Unixware}.
2538 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
2541 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
2547 The @uref{http://sourceware.org/cygwin/,,Cygwin} project;
2549 The @uref{http://www.mingw.org/,,MinGW} project.
2553 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
2554 Written Word} offers binaries for
2555 AIX 4.3.3, 5.1 and 5.2,
2557 Tru64 UNIX 4.0D and 5.1,
2559 HP-UX 10.20, 11.00, and 11.11, and
2560 Solaris/SPARC 2.5.1, 2.6, 7, 8, 9 and 10.
2563 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
2564 number of platforms.
2567 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
2568 links to GNU Fortran binaries for several platforms.
2571 In addition to those specific offerings, you can get a binary
2572 distribution CD-ROM from the
2573 @uref{http://www.gnu.org/order/order.html,,Free Software Foundation}.
2574 It contains binaries for a number of platforms, and
2575 includes not only GCC, but other stuff as well. The current CD does
2576 not contain the latest version of GCC, but it should allow
2577 bootstrapping the compiler. An updated version of that disk is in the
2585 @uref{./index.html,,Return to the GCC Installation page}
2589 @c ***Specific****************************************************************
2591 @comment node-name, next, previous, up
2592 @node Specific, Old, Binaries, Top
2596 @chapter Host/target specific installation notes for GCC
2599 @cindex Specific installation notes
2600 @cindex Target specific installation
2601 @cindex Host specific installation
2602 @cindex Target specific installation notes
2604 Please read this document carefully @emph{before} installing the
2605 GNU Compiler Collection on your machine.
2607 Note that this list of install notes is @emph{not} a list of supported
2608 hosts or targets. Not all supported hosts and targets are listed
2609 here, only the ones that require host-specific or target-specific
2615 @uref{#alpha-x-x,,alpha*-*-*}
2617 @uref{#alpha-dec-osf,,alpha*-dec-osf*}
2619 @uref{#arc-x-elf,,arc-*-elf}
2621 @uref{#arm-x-elf,,arm-*-elf}
2625 @uref{#bfin,,Blackfin}
2629 @uref{#x-x-freebsd,,*-*-freebsd*}
2631 @uref{#h8300-hms,,h8300-hms}
2633 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
2635 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
2637 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
2639 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
2641 @uref{#ix86-x-linux,,i?86-*-linux*}
2643 @uref{#ix86-x-solaris210,,i?86-*-solaris2.10}
2645 @uref{#ia64-x-linux,,ia64-*-linux}
2647 @uref{#ia64-x-hpux,,ia64-*-hpux*}
2649 @uref{#x-ibm-aix,,*-ibm-aix*}
2651 @uref{#iq2000-x-elf,,iq2000-*-elf}
2653 @uref{#m32c-x-elf,,m32c-*-elf}
2655 @uref{#m32r-x-elf,,m32r-*-elf}
2657 @uref{#m6811-elf,,m6811-elf}
2659 @uref{#m6812-elf,,m6812-elf}
2661 @uref{#m68k-x-x,,m68k-*-*}
2663 @uref{#m68k-uclinux,,m68k-uclinux}
2665 @uref{#mips-x-x,,mips-*-*}
2667 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2669 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2671 @uref{#powerpc-x-x,,powerpc*-*-*}
2673 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
2675 @uref{#powerpc-x-elf,,powerpc-*-elf}
2677 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
2679 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
2681 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
2683 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
2685 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
2687 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
2689 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
2691 @uref{#s390-x-linux,,s390-*-linux*}
2693 @uref{#s390x-x-linux,,s390x-*-linux*}
2695 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
2697 @uref{#x-x-solaris2,,*-*-solaris2*}
2699 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
2701 @uref{#sparc-sun-solaris27,,sparc-sun-solaris2.7}
2703 @uref{#sparc-x-linux,,sparc-*-linux*}
2705 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
2707 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
2709 @uref{#x-x-vxworks,,*-*-vxworks*}
2711 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
2713 @uref{#xtensa-x-elf,,xtensa*-*-elf}
2715 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
2717 @uref{#windows,,Microsoft Windows}
2719 @uref{#x-x-cygwin,,*-*-cygwin}
2721 @uref{#x-x-interix,,*-*-interix}
2723 @uref{#x-x-mingw32,,*-*-mingw32}
2727 @uref{#older,,Older systems}
2732 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
2738 <!-- -------- host/target specific issues start here ---------------- -->
2741 @heading @anchor{alpha-x-x}alpha*-*-*
2743 This section contains general configuration information for all
2744 alpha-based platforms using ELF (in particular, ignore this section for
2745 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2746 section, please read all other sections that match your target.
2748 We require binutils 2.11.2 or newer.
2749 Previous binutils releases had a number of problems with DWARF 2
2750 debugging information, not the least of which is incorrect linking of
2756 @heading @anchor{alpha-dec-osf}alpha*-dec-osf*
2757 Systems using processors that implement the DEC Alpha architecture and
2758 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2759 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2761 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2762 supported. (These are the versions which identify themselves as DEC
2765 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2766 may be fixed by configuring with @option{--with-gc=simple},
2767 reconfiguring Kernel Virtual Memory and Swap parameters
2768 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2769 or applying the patch in
2770 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2772 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2773 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2774 we need to use the old assembler, invoked via the barely documented
2775 @option{-oldas} option. To bootstrap GCC, you either need to use the
2779 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2782 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2785 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2788 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2789 are supported on Tru64 UNIX, so you must not configure GCC with
2790 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2792 GCC writes a @samp{.verstamp} directive to the assembler output file
2793 unless it is built as a cross-compiler. It gets the version to use from
2794 the system header file @file{/usr/include/stamp.h}. If you install a
2795 new version of DEC Unix, you should rebuild GCC to pick up the new version
2798 @samp{make compare} may fail on old versions of DEC Unix unless you add
2799 @option{-save-temps} to @code{BOOT_CFLAGS}. On these systems, the name
2800 of the assembler input file is stored in the object file, and that makes
2801 comparison fail if it differs between the @code{stage1} and
2802 @code{stage2} compilations. The option @option{-save-temps} forces a
2803 fixed name to be used for the assembler input file, instead of a
2804 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2805 unless the comparisons fail without that option. If you add
2806 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2807 @samp{.s} files after each series of compilations.
2809 GCC now supports both the native (ECOFF) debugging format used by DBX
2810 and GDB and an encapsulated STABS format for use only with GDB@. See the
2811 discussion of the @option{--with-stabs} option of @file{configure} above
2812 for more information on these formats and how to select them.
2814 There is a bug in DEC's assembler that produces incorrect line numbers
2815 for ECOFF format when the @samp{.align} directive is used. To work
2816 around this problem, GCC will not emit such alignment directives
2817 while writing ECOFF format debugging information even if optimization is
2818 being performed. Unfortunately, this has the very undesirable
2819 side-effect that code addresses when @option{-O} is specified are
2820 different depending on whether or not @option{-g} is also specified.
2822 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2823 DBX@. DEC is now aware of this problem with the assembler and hopes to
2824 provide a fix shortly.
2829 @heading @anchor{arc-x-elf}arc-*-elf
2830 Argonaut ARC processor.
2831 This configuration is intended for embedded systems.
2836 @heading @anchor{arm-x-elf}arm-*-elf
2837 ARM-family processors. Subtargets that use the ELF object format
2838 require GNU binutils 2.13 or newer. Such subtargets include:
2839 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux}
2840 and @code{arm-*-rtems}.
2845 @heading @anchor{avr}avr
2847 ATMEL AVR-family micro controllers. These are used in embedded
2848 applications. There are no standard Unix configurations.
2850 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
2854 See ``AVR Options'' in the main manual
2856 for the list of supported MCU types.
2858 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2860 Further installation notes and other useful information about AVR tools
2861 can also be obtained from:
2865 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
2867 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2869 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2872 We @emph{strongly} recommend using binutils 2.13 or newer.
2874 The following error:
2876 Error: register required
2879 indicates that you should upgrade to a newer version of the binutils.
2884 @heading @anchor{bfin}Blackfin
2886 The Blackfin processor, an Analog Devices DSP.
2888 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
2892 See ``Blackfin Options'' in the main manual
2895 More information, and a version of binutils with support for this processor,
2896 is available at @uref{http://blackfin.uclinux.org}
2901 @heading @anchor{cris}CRIS
2903 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2904 series. These are used in embedded applications.
2907 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
2911 See ``CRIS Options'' in the main manual
2913 for a list of CRIS-specific options.
2915 There are a few different CRIS targets:
2918 Mainly for monolithic embedded systems. Includes a multilib for the
2919 @samp{v10} core used in @samp{ETRAX 100 LX}.
2920 @item cris-axis-linux-gnu
2921 A GNU/Linux port for the CRIS architecture, currently targeting
2922 @samp{ETRAX 100 LX} by default.
2925 For @code{cris-axis-elf} you need binutils 2.11
2926 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2928 Pre-packaged tools can be obtained from
2929 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2930 information about this platform is available at
2931 @uref{http://developer.axis.com/}.
2936 @heading @anchor{crx}CRX
2938 The CRX CompactRISC architecture is a low-power 32-bit architecture with
2939 fast context switching and architectural extensibility features.
2942 @xref{CRX Options,, CRX Options, gcc, Using and Porting the GNU Compiler
2947 See ``CRX Options'' in the main manual for a list of CRX-specific options.
2950 Use @samp{configure --target=crx-elf --enable-languages=c,c++} to configure
2951 GCC@ for building a CRX cross-compiler. The option @samp{--target=crx-elf}
2952 is also used to build the @samp{newlib} C library for CRX.
2954 It is also possible to build libstdc++-v3 for the CRX architecture. This
2955 needs to be done in a separate step with the following configure settings:
2956 @samp{gcc/libstdc++-v3/configure --host=crx-elf --with-newlib
2957 --enable-sjlj-exceptions --enable-cxx-flags='-fexceptions -frtti'}
2962 @heading @anchor{dos}DOS
2964 Please have a look at the @uref{binaries.html,,binaries page}.
2966 You cannot install GCC by itself on MSDOS; it will not compile under
2967 any MSDOS compiler except itself. You need to get the complete
2968 compilation package DJGPP, which includes binaries as well as sources,
2969 and includes all the necessary compilation tools and libraries.
2974 @heading @anchor{x-x-freebsd}*-*-freebsd*
2976 The version of binutils installed in @file{/usr/bin} probably works with
2977 this release of GCC@. However, on FreeBSD 4, bootstrapping against the
2978 latest FSF binutils is known to improve overall testsuite results; and,
2979 on FreeBSD/alpha, using binutils 2.14 or later is required to build libjava.
2981 Support for FreeBSD 1 was discontinued in GCC 3.2.
2983 Support for FreeBSD 2 will be discontinued after GCC 3.4. The
2984 following was true for GCC 3.1 but the current status is unknown.
2985 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2986 configuration support and files as shipped with GCC 2.95 are still in
2987 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2988 it is unknown which version of binutils was used (it is assumed that it
2989 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2991 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2992 default for all CPU architectures. It had been the default on
2993 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2994 of @option{-g}, if you really want the old debugging format. There are
2995 no known issues with mixing object files and libraries with different
2996 debugging formats. Otherwise, this release of GCC should now match more
2997 of the configuration used in the stock FreeBSD configuration of GCC@. In
2998 particular, @option{--enable-threads} is now configured by default.
2999 However, as a general user, do not attempt to replace the system
3000 compiler with this release. Known to bootstrap and check with good
3001 results on FreeBSD 4.9-STABLE and 5-CURRENT@. In the past, known to
3002 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
3003 4.3, 4.4, 4.5, 4.8-STABLE@.
3005 In principle, @option{--enable-threads} is now compatible with
3006 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
3007 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
3009 library may be incorrectly built (symbols are missing at link time).
3010 There is a rare timing-based startup hang (probably involves an
3011 assumption about the thread library). Multi-threaded boehm-gc (required for
3012 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
3013 4.5-RELEASE@. Other CPU architectures
3014 supported by FreeBSD will require additional configuration tuning in, at
3015 the very least, both boehm-gc and libffi.
3017 Shared @file{libgcc_s.so} is now built and installed by default.
3022 @heading @anchor{h8300-hms}h8300-hms
3023 Renesas H8/300 series of processors.
3025 Please have a look at the @uref{binaries.html,,binaries page}.
3027 The calling convention and structure layout has changed in release 2.6.
3028 All code must be recompiled. The calling convention now passes the
3029 first three arguments in function calls in registers. Structures are no
3030 longer a multiple of 2 bytes.
3035 @heading @anchor{hppa-hp-hpux}hppa*-hp-hpux*
3036 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
3038 We require using gas/binutils on all hppa platforms. Version 2.19 or
3039 later is recommended.
3041 It may be helpful to configure GCC with the
3042 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
3043 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
3045 The HP assembler should not be used with GCC. It is rarely tested and may
3046 not work. It shouldn't be used with any languages other than C due to its
3049 Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging
3050 format which GCC does not know about). It also inserts timestamps
3051 into each object file it creates, causing the 3-stage comparison test to
3052 fail during a bootstrap. You should be able to continue by saying
3053 @samp{make all-host all-target} after getting the failure from @samp{make}.
3055 Various GCC features are not supported. For example, it does not support weak
3056 symbols or alias definitions. As a result, explicit template instantiations
3057 are required when using C++. This makes it difficult if not impossible to
3058 build many C++ applications.
3060 There are two default scheduling models for instructions. These are
3061 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
3062 architecture specified for the target machine when configuring.
3063 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
3064 the target is a @samp{hppa1*} machine.
3066 The PROCESSOR_8000 model is not well suited to older processors. Thus,
3067 it is important to completely specify the machine architecture when
3068 configuring if you want a model other than PROCESSOR_8000. The macro
3069 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
3070 default scheduling model is desired.
3072 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
3073 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
3074 This namespace change might cause problems when bootstrapping with
3075 an earlier version of GCC or the HP compiler as essentially the same
3076 namespace is required for an entire build. This problem can be avoided
3077 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
3078 or @samp{98}. Another way is to add an appropriate set of predefines
3079 to @env{CC}. The description for the @option{munix=} option contains
3080 a list of the predefines used with each standard.
3082 More specific information to @samp{hppa*-hp-hpux*} targets follows.
3087 @heading @anchor{hppa-hp-hpux10}hppa*-hp-hpux10
3089 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
3090 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
3096 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
3100 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
3104 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
3107 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
3108 used for one-only code and data. This resolves many of the previous
3109 problems in using C++ on this target. However, the ABI is not compatible
3110 with the one implemented under HP-UX 11 using secondary definitions.
3115 @heading @anchor{hppa-hp-hpux11}hppa*-hp-hpux11
3117 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
3118 be used to compile GCC 3.0 and up.
3120 The libffi and libjava libraries haven't been ported to 64-bit HP-UX@
3123 Refer to @uref{binaries.html,,binaries} for information about obtaining
3124 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
3125 to build the Ada language as it can't be bootstrapped using C@. Ada is
3126 only available for the 32-bit PA-RISC runtime.
3128 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
3129 bundled compiler supports only traditional C; you will need either HP's
3130 unbundled compiler, or a binary distribution of GCC@.
3132 It is possible to build GCC 3.3 starting with the bundled HP compiler,
3133 but the process requires several steps. GCC 3.3 can then be used to
3134 build later versions. The fastjar program contains ISO C code and
3135 can't be built with the HP bundled compiler. This problem can be
3136 avoided by not building the Java language. For example, use the
3137 @option{--enable-languages="c,c++,f77,objc"} option in your configure
3140 There are several possible approaches to building the distribution.
3141 Binutils can be built first using the HP tools. Then, the GCC
3142 distribution can be built. The second approach is to build GCC
3143 first using the HP tools, then build binutils, then rebuild GCC@.
3144 There have been problems with various binary distributions, so it
3145 is best not to start from a binary distribution.
3147 On 64-bit capable systems, there are two distinct targets. Different
3148 installation prefixes must be used if both are to be installed on
3149 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
3150 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
3151 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
3152 PA-RISC 2.0 architecture.
3154 The script config.guess now selects the target type based on the compiler
3155 detected during configuration. You must define @env{PATH} or @env{CC} so
3156 that configure finds an appropriate compiler for the initial bootstrap.
3157 When @env{CC} is used, the definition should contain the options that are
3158 needed whenever @env{CC} is used.
3160 Specifically, options that determine the runtime architecture must be
3161 in @env{CC} to correctly select the target for the build. It is also
3162 convenient to place many other compiler options in @env{CC}. For example,
3163 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
3164 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
3165 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
3166 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
3167 macro definition table of cpp needs to be increased for a successful
3168 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
3169 be defined when building with the bundled compiler, or when using the
3170 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
3172 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
3173 with the @option{--with-ld=@dots{}} option. This overrides the standard
3174 search for ld. The two linkers supported on this target require different
3175 commands. The default linker is determined during configuration. As a
3176 result, it's not possible to switch linkers in the middle of a GCC build.
3177 This has been reported to sometimes occur in unified builds of binutils
3180 A recent linker patch must be installed for the correct operation of
3181 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
3182 oldest linker patches that are known to work. They are for HP-UX
3183 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
3184 @code{PHSS_24304}, might be usable but it hasn't been tested. These
3185 patches have been superseded. Consult the HP patch database to obtain
3186 the currently recommended linker patch for your system.
3188 The patches are necessary for the support of weak symbols on the
3189 32-bit port, and for the running of initializers and finalizers. Weak
3190 symbols are implemented using SOM secondary definition symbols. Prior
3191 to HP-UX 11, there are bugs in the linker support for secondary symbols.
3192 The patches correct a problem of linker core dumps creating shared
3193 libraries containing secondary symbols, as well as various other
3194 linking issues involving secondary symbols.
3196 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
3197 run initializers and finalizers on the 64-bit port. The 32-bit port
3198 uses the linker @option{+init} and @option{+fini} options for the same
3199 purpose. The patches correct various problems with the +init/+fini
3200 options, including program core dumps. Binutils 2.14 corrects a
3201 problem on the 64-bit port resulting from HP's non-standard use of
3202 the .init and .fini sections for array initializers and finalizers.
3204 Although the HP and GNU linkers are both supported for the
3205 @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the
3206 HP linker be used for link editing on this target.
3208 At this time, the GNU linker does not support the creation of long
3209 branch stubs. As a result, it can't successfully link binaries
3210 containing branch offsets larger than 8 megabytes. In addition,
3211 there are problems linking shared libraries, linking executables
3212 with @option{-static}, and with dwarf2 unwind and exception support.
3213 It also doesn't provide stubs for internal calls to global functions
3214 in shared libraries, so these calls can't be overloaded.
3216 The HP dynamic loader does not support GNU symbol versioning, so symbol
3217 versioning is not supported. It may be necessary to disable symbol
3218 versioning with @option{--disable-symvers} when using GNU ld.
3220 POSIX threads are the default. The optional DCE thread library is not
3221 supported, so @option{--enable-threads=dce} does not work.
3226 @heading @anchor{x-x-linux-gnu}*-*-linux-gnu
3228 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
3229 in glibc 2.2.5 and later. More information is available in the
3230 libstdc++-v3 documentation.
3235 @heading @anchor{ix86-x-linux}i?86-*-linux*
3237 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
3238 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
3240 If you receive Signal 11 errors when building on GNU/Linux, then it is
3241 possible you have a hardware problem. Further information on this can be
3242 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
3247 @heading @anchor{ix86-x-solaris210}i?86-*-solaris2.10
3248 Use this for Solaris 10 or later on x86 and x86-64 systems. This
3249 configuration is supported by GCC 4.0 and later versions only.
3251 It is recommended that you configure GCC to use the GNU assembler in
3252 @file{/usr/sfw/bin/gas} but the Sun linker, using the options
3253 @option{--with-gnu-as --with-as=/usr/sfw/bin/gas --without-gnu-ld
3254 --with-ld=/usr/ccs/bin/ld}.
3259 @heading @anchor{ia64-x-linux}ia64-*-linux
3260 IA-64 processor (also known as IPF, or Itanium Processor Family)
3263 If you are using the installed system libunwind library with
3264 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
3267 None of the following versions of GCC has an ABI that is compatible
3268 with any of the other versions in this list, with the exception that
3269 Red Hat 2.96 and Trillian 000171 are compatible with each other:
3270 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
3271 This primarily affects C++ programs and programs that create shared libraries.
3272 GCC 3.1 or later is recommended for compiling linux, the kernel.
3273 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
3274 more major ABI changes are expected.
3279 @heading @anchor{ia64-x-hpux}ia64-*-hpux*
3280 Building GCC on this target requires the GNU Assembler. The bundled HP
3281 assembler will not work. To prevent GCC from using the wrong assembler,
3282 the option @option{--with-gnu-as} may be necessary.
3284 The GCC libunwind library has not been ported to HPUX@. This means that for
3285 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
3286 is required to build GCC@. For GCC 3.3 and later, this is the default.
3287 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
3288 removed and the system libunwind library will always be used.
3292 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
3294 @heading @anchor{x-ibm-aix}*-ibm-aix*
3295 Support for AIX version 3 and older was discontinued in GCC 3.4.
3296 Support for AIX version 4.2 and older was discontinued in GCC 4.5.
3298 ``out of memory'' bootstrap failures may indicate a problem with
3299 process resource limits (ulimit). Hard limits are configured in the
3300 @file{/etc/security/limits} system configuration file.
3302 To speed up the configuration phases of bootstrapping and installing GCC,
3303 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
3306 % CONFIG_SHELL=/opt/freeware/bin/bash
3307 % export CONFIG_SHELL
3310 and then proceed as described in @uref{build.html,,the build
3311 instructions}, where we strongly recommend specifying an absolute path
3312 to invoke @var{srcdir}/configure.
3314 Because GCC on AIX is built as a 32-bit executable by default,
3315 (although it can generate 64-bit programs) the GMP and MPFR libraries
3316 required by gfortran must be 32-bit libraries. Building GMP and MPFR
3317 as static archive libraries works better than shared libraries.
3319 Errors involving @code{alloca} when building GCC generally are due
3320 to an incorrect definition of @code{CC} in the Makefile or mixing files
3321 compiled with the native C compiler and GCC@. During the stage1 phase of
3322 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
3323 (not @command{xlc}). Once @command{configure} has been informed of
3324 @command{xlc}, one needs to use @samp{make distclean} to remove the
3325 configure cache files and ensure that @env{CC} environment variable
3326 does not provide a definition that will confuse @command{configure}.
3327 If this error occurs during stage2 or later, then the problem most likely
3328 is the version of Make (see above).
3330 The native @command{as} and @command{ld} are recommended for bootstrapping
3331 on AIX 4 and required for bootstrapping on AIX 5L@. The GNU Assembler
3332 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
3333 utilize weak symbol functionality although it is not supported. The GNU
3334 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC@.
3335 The native AIX tools do interoperate with GCC@.
3337 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
3338 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
3339 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
3340 referenced as APAR IY53606 (AIX 5.2) or a APAR IY54774 (AIX 5.1)
3342 @samp{libstdc++} in GCC 3.4 increments the major version number of the
3343 shared object and GCC installation places the @file{libstdc++.a}
3344 shared library in a common location which will overwrite the and GCC
3345 3.3 version of the shared library. Applications either need to be
3346 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
3347 versions of the @samp{libstdc++} shared object needs to be available
3348 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
3349 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
3350 installed for runtime dynamic loading using the following steps to set
3351 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3352 multilib @file{libstdc++.a} installed:
3354 Extract the shared objects from the currently installed
3355 @file{libstdc++.a} archive:
3357 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3360 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3361 available for runtime dynamic loading, but not linking:
3363 % strip -e libstdc++.so.4 libstdc++.so.5
3366 Archive the runtime-only shared object in the GCC 3.4
3367 @file{libstdc++.a} archive:
3369 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
3372 Linking executables and shared libraries may produce warnings of
3373 duplicate symbols. The assembly files generated by GCC for AIX always
3374 have included multiple symbol definitions for certain global variable
3375 and function declarations in the original program. The warnings should
3376 not prevent the linker from producing a correct library or runnable
3379 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
3380 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
3381 to parse archive libraries did not handle the new format correctly.
3382 These routines are used by GCC and result in error messages during
3383 linking such as ``not a COFF file''. The version of the routines shipped
3384 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
3385 option of the archive command may be used to create archives of 32-bit
3386 objects using the original ``small format''. A correct version of the
3387 routines is shipped with AIX 4.3.2 and above.
3389 Some versions of the AIX binder (linker) can fail with a relocation
3390 overflow severe error when the @option{-bbigtoc} option is used to link
3391 GCC-produced object files into an executable that overflows the TOC@. A fix
3392 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
3393 available from IBM Customer Support and from its
3394 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3395 website as PTF U455193.
3397 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
3398 with a segmentation fault when invoked by any version of GCC@. A fix for
3399 APAR IX87327 is available from IBM Customer Support and from its
3400 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3401 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
3403 The initial assembler shipped with AIX 4.3.0 generates incorrect object
3404 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
3405 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
3406 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3407 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
3409 AIX provides National Language Support (NLS)@. Compilers and assemblers
3410 use NLS to support locale-specific representations of various data
3411 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
3412 separating decimal fractions). There have been problems reported where
3413 GCC does not produce the same floating-point formats that the assembler
3414 expects. If one encounters this problem, set the @env{LANG}
3415 environment variable to @samp{C} or @samp{En_US}.
3417 A default can be specified with the @option{-mcpu=@var{cpu_type}}
3418 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
3423 @heading @anchor{iq2000-x-elf}iq2000-*-elf
3424 Vitesse IQ2000 processors. These are used in embedded
3425 applications. There are no standard Unix configurations.
3430 @heading @anchor{m32c-x-elf}m32c-*-elf
3431 Renesas M32C processor.
3432 This configuration is intended for embedded systems.
3437 @heading @anchor{m32r-x-elf}m32r-*-elf
3438 Renesas M32R processor.
3439 This configuration is intended for embedded systems.
3444 @heading @anchor{m6811-elf}m6811-elf
3445 Motorola 68HC11 family micro controllers. These are used in embedded
3446 applications. There are no standard Unix configurations.
3451 @heading @anchor{m6812-elf}m6812-elf
3452 Motorola 68HC12 family micro controllers. These are used in embedded
3453 applications. There are no standard Unix configurations.
3458 @heading @anchor{m68k-x-x}m68k-*-*
3460 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
3462 build libraries for both M680x0 and ColdFire processors. If you only
3463 need the M680x0 libraries, you can omit the ColdFire ones by passing
3464 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
3465 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
3466 @command{configure}. These targets default to 5206 or 5475 code as
3467 appropriate for the target system when
3468 configured with @option{--with-arch=cf} and 68020 code otherwise.
3470 The @samp{m68k-*-netbsd} and
3471 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
3472 option. They will generate ColdFire CFV4e code when configured with
3473 @option{--with-arch=cf} and 68020 code otherwise.
3475 You can override the default processors listed above by configuring
3476 with @option{--with-cpu=@var{target}}. This @var{target} can either
3477 be a @option{-mcpu} argument or one of the following values:
3478 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
3479 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
3484 @heading @anchor{m68k-x-uclinux}m68k-*-uclinux
3485 GCC 4.3 changed the uClinux configuration so that it uses the
3486 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
3487 It also added improved support for C++ and flat shared libraries,
3488 both of which were ABI changes. However, you can still use the
3489 original ABI by configuring for @samp{m68k-uclinuxoldabi} or
3490 @samp{m68k-@var{vendor}-uclinuxoldabi}.
3495 @heading @anchor{mips-x-x}mips-*-*
3496 If on a MIPS system you get an error message saying ``does not have gp
3497 sections for all it's [sic] sectons [sic]'', don't worry about it. This
3498 happens whenever you use GAS with the MIPS linker, but there is not
3499 really anything wrong, and it is okay to use the output file. You can
3500 stop such warnings by installing the GNU linker.
3502 It would be nice to extend GAS to produce the gp tables, but they are
3503 optional, and there should not be a warning about their absence.
3505 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
3506 and later. A patch went in just after the GCC 3.3 release to
3507 make @samp{mips*-*-*} use the generic implementation instead. You can also
3508 configure for @samp{mipsel-elf} as a workaround. The
3509 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
3510 work on this is expected in future releases.
3512 @c If you make --with-llsc the default for another target, please also
3513 @c update the description of the --with-llsc option.
3515 The built-in @code{__sync_*} functions are available on MIPS II and
3516 later systems and others that support the @samp{ll}, @samp{sc} and
3517 @samp{sync} instructions. This can be overridden by passing
3518 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
3519 Since the Linux kernel emulates these instructions if they are
3520 missing, the default for @samp{mips*-*-linux*} targets is
3521 @option{--with-llsc}. The @option{--with-llsc} and
3522 @option{--without-llsc} configure options may be overridden at compile
3523 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
3526 MIPS systems check for division by zero (unless
3527 @option{-mno-check-zero-division} is passed to the compiler) by
3528 generating either a conditional trap or a break instruction. Using
3529 trap results in smaller code, but is only supported on MIPS II and
3530 later. Also, some versions of the Linux kernel have a bug that
3531 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
3532 the use of break, use the @option{--with-divide=breaks}
3533 @command{configure} option when configuring GCC@. The default is to
3534 use traps on systems that support them.
3536 Cross-compilers for the MIPS as target using the MIPS assembler
3537 currently do not work, because the auxiliary programs
3538 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
3539 anything but a MIPS@. It does work to cross compile for a MIPS
3540 if you use the GNU assembler and linker.
3542 The assembler from GNU binutils 2.17 and earlier has a bug in the way
3543 it sorts relocations for REL targets (o32, o64, EABI). This can cause
3544 bad code to be generated for simple C++ programs. Also the linker
3545 from GNU binutils versions prior to 2.17 has a bug which causes the
3546 runtime linker stubs in very large programs, like @file{libgcj.so}, to
3547 be incorrectly generated. GNU Binutils 2.18 and later (and snapshots
3548 made after Nov. 9, 2006) should be free from both of these problems.
3553 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
3555 In order to compile GCC on an SGI running IRIX 5, the @samp{compiler_dev.hdr}
3556 subsystem must be installed from the IDO CD-ROM supplied by SGI@.
3557 It is also available for download from
3558 @uref{ftp://ftp.sgi.com/sgi/IRIX5.3/iris-development-option-5.3.tardist}.
3560 If you use the MIPS C compiler to bootstrap, it may be necessary
3561 to increase its table size for switch statements with the
3562 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
3563 optimization option, you also need to use @option{-Olimit 3000}.
3565 To enable debugging under IRIX 5, you must use GNU binutils 2.15 or
3566 later, and use the @option{--with-gnu-ld} @command{configure} option
3567 when configuring GCC@. You need to use GNU @command{ar} and @command{nm},
3568 also distributed with GNU binutils.
3570 Some users have reported that @command{/bin/sh} will hang during bootstrap.
3571 This problem can be avoided by running the commands:
3574 % CONFIG_SHELL=/bin/ksh
3575 % export CONFIG_SHELL
3578 before starting the build.
3583 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3585 If you are using SGI's MIPSpro @command{cc} as your bootstrap compiler, you must
3586 ensure that the N32 ABI is in use. To test this, compile a simple C
3587 file with @command{cc} and then run @command{file} on the
3588 resulting object file. The output should look like:
3591 test.o: ELF N32 MSB @dots{}
3597 test.o: ELF 32-bit MSB @dots{}
3603 test.o: ELF 64-bit MSB @dots{}
3606 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3607 should set the environment variable @env{CC} to @samp{cc -n32}
3608 before configuring GCC@.
3610 If you want the resulting @command{gcc} to run on old 32-bit systems
3611 with the MIPS R4400 CPU, you need to ensure that only code for the @samp{mips3}
3612 instruction set architecture (ISA) is generated. While GCC 3.x does
3613 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3614 the ISA depending on the machine where GCC is built. Using one of them
3615 as the bootstrap compiler may result in @samp{mips4} code, which won't run at
3616 all on @samp{mips3}-only systems. For the test program above, you should see:
3619 test.o: ELF N32 MSB mips-3 @dots{}
3625 test.o: ELF N32 MSB mips-4 @dots{}
3628 instead, you should set the environment variable @env{CC} to @samp{cc
3629 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3631 MIPSpro C 7.4 may cause bootstrap failures, due to a bug when inlining
3632 @code{memcmp}. Either add @code{-U__INLINE_INTRINSICS} to the @env{CC}
3633 environment variable as a workaround or upgrade to MIPSpro C 7.4.1m.
3635 GCC on IRIX 6 is usually built to support the N32, O32 and N64 ABIs. If
3636 you build GCC on a system that doesn't have the N64 libraries installed
3637 or cannot run 64-bit binaries,
3638 you need to configure with @option{--disable-multilib} so GCC doesn't
3639 try to use them. This will disable building the O32 libraries, too.
3640 Look for @file{/usr/lib64/libc.so.1} to see if you
3641 have the 64-bit libraries installed.
3643 To enable debugging for the O32 ABI, you must use GNU @command{as} from
3644 GNU binutils 2.15 or later. You may also use GNU @command{ld}, but
3645 this is not required and currently causes some problems with Ada.
3647 The @option{--enable-libgcj}
3648 option is disabled by default: IRIX 6 uses a very low default limit
3649 (20480) for the command line length. Although @command{libtool} contains a
3650 workaround for this problem, at least the N64 @samp{libgcj} is known not
3651 to build despite this, running into an internal error of the native
3652 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3653 its maximum of 262144 bytes. If you have root access, you can use the
3654 @command{systune} command to do this.
3656 @code{wchar_t} support in @samp{libstdc++} is not available for old
3657 IRIX 6.5.x releases, @math{x < 19}. The problem cannot be autodetected
3658 and in order to build GCC for such targets you need to configure with
3659 @option{--disable-wchar_t}.
3661 See @uref{http://freeware.sgi.com/} for more
3662 information about using GCC on IRIX platforms.
3667 @heading @anchor{powerpc-x-x}powerpc-*-*
3669 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3670 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3673 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.15}
3674 or newer for a working GCC@.
3679 @heading @anchor{powerpc-x-darwin}powerpc-*-darwin*
3680 PowerPC running Darwin (Mac OS X kernel).
3682 Pre-installed versions of Mac OS X may not include any developer tools,
3683 meaning that you will not be able to build GCC from source. Tool
3684 binaries are available at
3685 @uref{http://developer.apple.com/darwin/projects/compiler/} (free
3686 registration required).
3688 This version of GCC requires at least cctools-590.36. The
3689 cctools-590.36 package referenced from
3690 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
3691 on systems older than 10.3.9 (aka darwin7.9.0).
3696 @heading @anchor{powerpc-x-elf}powerpc-*-elf
3697 PowerPC system in big endian mode, running System V.4.
3702 @heading @anchor{powerpc-x-linux-gnu}powerpc*-*-linux-gnu*
3704 PowerPC system in big endian mode running Linux.
3709 @heading @anchor{powerpc-x-netbsd}powerpc-*-netbsd*
3710 PowerPC system in big endian mode running NetBSD@.
3715 @heading @anchor{powerpc-x-eabisim}powerpc-*-eabisim
3716 Embedded PowerPC system in big endian mode for use in running under the
3722 @heading @anchor{powerpc-x-eabi}powerpc-*-eabi
3723 Embedded PowerPC system in big endian mode.
3728 @heading @anchor{powerpcle-x-elf}powerpcle-*-elf
3729 PowerPC system in little endian mode, running System V.4.
3734 @heading @anchor{powerpcle-x-eabisim}powerpcle-*-eabisim
3735 Embedded PowerPC system in little endian mode for use in running under
3741 @heading @anchor{powerpcle-x-eabi}powerpcle-*-eabi
3742 Embedded PowerPC system in little endian mode.
3747 @heading @anchor{s390-x-linux}s390-*-linux*
3748 S/390 system running GNU/Linux for S/390@.
3753 @heading @anchor{s390x-x-linux}s390x-*-linux*
3754 zSeries system (64-bit) running GNU/Linux for zSeries@.
3759 @heading @anchor{s390x-ibm-tpf}s390x-ibm-tpf*
3760 zSeries system (64-bit) running TPF@. This platform is
3761 supported as cross-compilation target only.
3766 @c Please use Solaris 2 to refer to all release of Solaris, starting
3767 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
3768 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3769 @c alone is too unspecific and must be avoided.
3770 @heading @anchor{x-x-solaris2}*-*-solaris2*
3772 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3773 GCC you first have to install a pre-built compiler, see the
3774 @uref{binaries.html,,binaries page} for details.
3776 The Solaris 2 @command{/bin/sh} will often fail to configure
3777 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3778 recommend using the following initial sequence of commands
3781 % CONFIG_SHELL=/bin/ksh
3782 % export CONFIG_SHELL
3785 and proceed as described in @uref{configure.html,,the configure instructions}.
3786 In addition we strongly recommend specifying an absolute path to invoke
3787 @var{srcdir}/configure.
3789 Solaris 2 comes with a number of optional OS packages. Some of these
3790 are needed to use GCC fully, namely @code{SUNWarc},
3791 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3792 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3793 optional packages when installing Solaris 2, you will need to verify that
3794 the packages that GCC needs are installed.
3796 To check whether an optional package is installed, use
3797 the @command{pkginfo} command. To add an optional package, use the
3798 @command{pkgadd} command. For further details, see the Solaris 2
3801 Trying to use the linker and other tools in
3802 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3803 For example, the linker may hang indefinitely. The fix is to remove
3804 @file{/usr/ucb} from your @env{PATH}.
3806 The build process works more smoothly with the legacy Sun tools so, if you
3807 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3808 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3810 We recommend the use of GNU binutils 2.14 or later, or the vendor tools
3811 (Sun @command{as}, Sun @command{ld}). Note that your mileage may vary
3812 if you use a combination of the GNU tools and the Sun tools: while the
3813 combination GNU @command{as} + Sun @command{ld} should reasonably work,
3814 the reverse combination Sun @command{as} + GNU @command{ld} is known to
3815 cause memory corruption at runtime in some cases for C++ programs.
3817 The stock GNU binutils 2.15 release is broken on this platform because of a
3818 single bug. It has been fixed on the 2.15 branch in the CVS repository.
3819 You can obtain a working version by checking out the binutils-2_15-branch
3820 from the CVS repository or applying the patch
3821 @uref{http://sourceware.org/ml/binutils-cvs/2004-09/msg00036.html} to the
3824 We recommend the use of GNU binutils 2.16 or later in conjunction with GCC
3825 4.x, or the vendor tools (Sun @command{as}, Sun @command{ld}). However,
3826 for Solaris 10 and above, an additional patch is required in order for the
3827 GNU linker to be able to cope with a new flavor of shared libraries. You
3828 can obtain a working version by checking out the binutils-2_16-branch from
3829 the CVS repository or applying the patch
3830 @uref{http://sourceware.org/ml/binutils-cvs/2005-07/msg00122.html} to the
3833 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3834 newer: @command{g++} will complain that types are missing. These headers
3835 assume that omitting the type means @code{int}; this assumption worked for
3836 C89 but is wrong for C++, and is now wrong for C99 also.
3838 @command{g++} accepts such (invalid) constructs with the option
3839 @option{-fpermissive}; it will assume that any missing type is @code{int}
3840 (as defined by C89).
3842 There are patches for Solaris 7 (108376-21 or newer for SPARC,
3843 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3844 108653-22 for Intel) that fix this bug.
3846 Sun bug 4927647 sometimes causes random spurious testsuite failures
3847 related to missing diagnostic output. This bug doesn't affect GCC
3848 itself, rather it is a kernel bug triggered by the @command{expect}
3849 program which is used only by the GCC testsuite driver. When the bug
3850 causes the @command{expect} program to miss anticipated output, extra
3851 testsuite failures appear.
3853 There are patches for Solaris 8 (117350-12 or newer for SPARC,
3854 117351-12 or newer for Intel) and Solaris 9 (117171-11 or newer for
3855 SPARC, 117172-11 or newer for Intel) that address this problem.
3860 @heading @anchor{sparc-sun-solaris2}sparc-sun-solaris2*
3862 When GCC is configured to use binutils 2.14 or later the binaries
3863 produced are smaller than the ones produced using Sun's native tools;
3864 this difference is quite significant for binaries containing debugging
3867 Starting with Solaris 7, the operating system is capable of executing
3868 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3869 this; the @option{-m64} option enables 64-bit code generation.
3870 However, if all you want is code tuned for the UltraSPARC CPU, you
3871 should try the @option{-mtune=ultrasparc} option instead, which produces
3872 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3875 When configuring on a Solaris 7 or later system that is running a kernel
3876 that supports only 32-bit binaries, one must configure with
3877 @option{--disable-multilib}, since we will not be able to build the
3878 64-bit target libraries.
3880 GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
3881 the GNU compiler (especially GCC 3.0.x versions), which lead to the
3882 miscompilation of the stage1 compiler and the subsequent failure of the
3883 bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
3884 stage, i.e.@: to bootstrap that compiler with the base compiler and then
3885 use it to bootstrap the final compiler.
3887 GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
3888 and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
3889 failure in form of a miscompilation of the stage1 compiler by the Sun
3890 compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
3892 GCC 3.4 changed the default debugging format from STABS to DWARF-2 for
3893 32-bit code on Solaris 7 and later. If you use the Sun assembler, this
3894 change apparently runs afoul of Sun bug 4910101 (which is referenced as
3895 a x86-only problem by Sun, probably because they do not use DWARF-2).
3896 A symptom of the problem is that you cannot compile C++ programs like
3897 @command{groff} 1.19.1 without getting messages similar to the following:
3900 ld: warning: relocation error: R_SPARC_UA32: @dots{}
3901 external symbolic relocation against non-allocatable section
3902 .debug_info cannot be processed at runtime: relocation ignored.
3905 To work around this problem, compile with @option{-gstabs+} instead of
3908 When configuring the GNU Multiple Precision Library (GMP) or the MPFR
3909 library on a Solaris 7 or later system, the canonical target triplet
3910 must be specified as the @command{build} parameter on the configure
3911 line. This triplet can be obtained by invoking ./config.guess in
3912 the toplevel source directory of GCC (and not that of GMP or MPFR).
3913 For example on a Solaris 7 system:
3916 % ./configure --build=sparc-sun-solaris2.7 --prefix=xxx
3922 @heading @anchor{sparc-sun-solaris27}sparc-sun-solaris2.7
3924 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3925 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3926 and later, including all EGCS releases. Sun formerly recommended
3927 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3928 recommend it only for people who use Sun's compilers.
3930 Here are some workarounds to this problem:
3933 Do not install Sun patch 107058-01 until after Sun releases a
3934 complete patch for bug 4210064. This is the simplest course to take,
3935 unless you must also use Sun's C compiler. Unfortunately 107058-01
3936 is preinstalled on some new Solaris 7-based hosts, so you may have to
3940 Copy the original, unpatched Solaris 7
3941 @command{/usr/ccs/bin/as} into
3942 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
3943 adjusting the latter name to fit your local conventions and software
3947 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3948 both 107058-01 and 106950-03 installed has reported the bug with GCC
3949 and Sun's dynamic linker. This last course of action is riskiest,
3950 for two reasons. First, you must install 106950 on all hosts that
3951 run code generated by GCC; it doesn't suffice to install it only on
3952 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3953 only a partial fix for bug 4210064, but Sun doesn't know whether the
3954 partial fix is adequate for GCC@. Revision -08 or later should fix
3955 the bug. The current (as of 2004-05-23) revision is -24, and is included in
3956 the Solaris 7 Recommended Patch Cluster.
3959 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
3960 which causes a bootstrap failure when linking the 64-bit shared version of
3961 libgcc. A typical error message is:
3964 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
3965 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
3968 This bug has been fixed in the final 5.0 version of the assembler.
3970 A similar problem was reported for version Sun WorkShop 6 99/08/18 of the
3971 Sun assembler, which causes a bootstrap failure with GCC 4.0.0:
3974 ld: fatal: relocation error: R_SPARC_DISP32:
3975 file .libs/libstdc++.lax/libsupc++convenience.a/vterminate.o:
3976 symbol <unknown>: offset 0xfccd33ad is non-aligned
3979 This bug has been fixed in more recent revisions of the assembler.
3984 @heading @anchor{sparc-x-linux}sparc-*-linux*
3986 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
3987 or newer on this platform. All earlier binutils and glibc
3988 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
3994 @heading @anchor{sparc64-x-solaris2}sparc64-*-solaris2*
3996 When configuring the GNU Multiple Precision Library (GMP) or the
3997 MPFR library, the canonical target triplet must be specified as
3998 the @command{build} parameter on the configure line. For example
3999 on a Solaris 7 system:
4002 % ./configure --build=sparc64-sun-solaris2.7 --prefix=xxx
4005 The following compiler flags must be specified in the configure
4006 step in order to bootstrap this target with the Sun compiler:
4009 % CC="cc -xarch=v9 -xildoff" @var{srcdir}/configure [@var{options}] [@var{target}]
4012 @option{-xarch=v9} specifies the SPARC-V9 architecture to the Sun toolchain
4013 and @option{-xildoff} turns off the incremental linker.
4018 @heading @anchor{sparcv9-x-solaris2}sparcv9-*-solaris2*
4020 This is a synonym for sparc64-*-solaris2*.
4025 @heading @anchor{x-x-vxworks}*-*-vxworks*
4026 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4027 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4028 We welcome patches for other architectures supported by VxWorks 5.5.
4029 Support for VxWorks AE would also be welcome; we believe this is merely
4030 a matter of writing an appropriate ``configlette'' (see below). We are
4031 not interested in supporting older, a.out or COFF-based, versions of
4034 VxWorks comes with an older version of GCC installed in
4035 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4036 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4037 Before running @command{configure}, create the directories @file{@var{prefix}}
4038 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4039 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4040 include that directory while running both @command{configure} and
4043 You must give @command{configure} the
4044 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4045 find the VxWorks system headers. Since VxWorks is a cross compilation
4046 target only, you must also specify @option{--target=@var{target}}.
4047 @command{configure} will attempt to create the directory
4048 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4049 make sure the user running @command{configure} has sufficient privilege
4052 GCC's exception handling runtime requires a special ``configlette''
4053 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4054 that file to add the module to your kernel build. (Future versions of
4055 VxWorks will incorporate this module.)
4060 @heading @anchor{x86-64-x-x}x86_64-*-*, amd64-*-*
4062 GCC supports the x86-64 architecture implemented by the AMD64 processor
4063 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4064 On GNU/Linux the default is a bi-arch compiler which is able to generate
4065 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4070 @heading @anchor{xtensa-x-elf}xtensa*-*-elf
4072 This target is intended for embedded Xtensa systems using the
4073 @samp{newlib} C library. It uses ELF but does not support shared
4074 objects. Designed-defined instructions specified via the
4075 Tensilica Instruction Extension (TIE) language are only supported
4076 through inline assembly.
4078 The Xtensa configuration information must be specified prior to
4079 building GCC@. The @file{include/xtensa-config.h} header
4080 file contains the configuration information. If you created your
4081 own Xtensa configuration with the Xtensa Processor Generator, the
4082 downloaded files include a customized copy of this header file,
4083 which you can use to replace the default header file.
4088 @heading @anchor{xtensa-x-linux}xtensa*-*-linux*
4090 This target is for Xtensa systems running GNU/Linux. It supports ELF
4091 shared objects and the GNU C library (glibc). It also generates
4092 position-independent code (PIC) regardless of whether the
4093 @option{-fpic} or @option{-fPIC} options are used. In other
4094 respects, this target is the same as the
4095 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
4100 @heading @anchor{windows}Microsoft Windows
4102 @subheading Intel 16-bit versions
4103 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
4106 However, the 32-bit port has limited support for Microsoft
4107 Windows 3.11 in the Win32s environment, as a target only. See below.
4109 @subheading Intel 32-bit versions
4111 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
4112 XP, and Windows Vista, are supported by several different target
4113 platforms. These targets differ in which Windows subsystem they target
4114 and which C libraries are used.
4117 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
4118 Linux API emulation layer in the Win32 subsystem.
4119 @item Interix @uref{#x-x-interix,,*-*-interix}: The Interix subsystem
4120 provides native support for POSIX.
4121 @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for
4122 the Win32 subsystem that provides a subset of POSIX.
4123 @item MKS i386-pc-mks: NuTCracker from MKS. See
4124 @uref{http://www.mkssoftware.com/} for more information.
4127 @subheading Intel 64-bit versions
4129 GCC contains support for x86-64 using the mingw-w64
4130 runtime library, available from @uref{http://mingw-w64.sourceforge.net/}.
4131 This library should be used with the target triple x86_64-pc-mingw32.
4133 Presently Windows for Itanium is not supported.
4135 @subheading Windows CE
4137 Windows CE is supported as a target only on ARM (arm-wince-pe), Hitachi
4138 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
4140 @subheading Other Windows Platforms
4142 GCC no longer supports Windows NT on the Alpha or PowerPC.
4144 GCC no longer supports the Windows POSIX subsystem. However, it does
4145 support the Interix subsystem. See above.
4147 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
4149 PW32 (i386-pc-pw32) support was never completed, and the project seems to
4150 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
4152 UWIN support has been removed due to a lack of maintenance.
4157 @heading @anchor{x-x-cygwin}*-*-cygwin
4159 Ports of GCC are included with the
4160 @uref{http://www.cygwin.com/,,Cygwin environment}.
4162 GCC will build under Cygwin without modification; it does not build
4163 with Microsoft's C++ compiler and there are no plans to make it do so.
4165 Cygwin can be compiled with i?86-pc-cygwin.
4170 @heading @anchor{x-x-interix}*-*-interix
4172 The Interix target is used by OpenNT, Interix, Services For UNIX (SFU),
4173 and Subsystem for UNIX-based Applications (SUA). Applications compiled
4174 with this target run in the Interix subsystem, which is separate from
4175 the Win32 subsystem. This target was last known to work in GCC 3.3.
4177 For more information, see @uref{http://www.interix.com/}.
4182 @heading @anchor{x-x-mingw32}*-*-mingw32
4184 GCC will build with and support only MinGW runtime 3.12 and later.
4185 Earlier versions of headers are incompatible with the new default semantics
4186 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
4191 @heading @anchor{os2}OS/2
4193 GCC does not currently support OS/2. However, Andrew Zabolotny has been
4194 working on a generic OS/2 port with pgcc. The current code can be found
4195 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
4200 @heading @anchor{older}Older systems
4202 GCC contains support files for many older (1980s and early
4203 1990s) Unix variants. For the most part, support for these systems
4204 has not been deliberately removed, but it has not been maintained for
4205 several years and may suffer from bitrot.
4207 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
4208 Support for these systems is still present in that release, but
4209 @command{configure} will fail unless the @option{--enable-obsolete}
4210 option is given. Unless a maintainer steps forward, support for these
4211 systems will be removed from the next release of GCC@.
4213 Support for old systems as hosts for GCC can cause problems if the
4214 workarounds for compiler, library and operating system bugs affect the
4215 cleanliness or maintainability of the rest of GCC@. In some cases, to
4216 bring GCC up on such a system, if still possible with current GCC, may
4217 require first installing an old version of GCC which did work on that
4218 system, and using it to compile a more recent GCC, to avoid bugs in the
4219 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
4220 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
4221 sites}. Header bugs may generally be avoided using
4222 @command{fixincludes}, but bugs or deficiencies in libraries and the
4223 operating system may still cause problems.
4225 Support for older systems as targets for cross-compilation is less
4226 problematic than support for them as hosts for GCC; if an enthusiast
4227 wishes to make such a target work again (including resurrecting any of
4228 the targets that never worked with GCC 2, starting from the last
4229 version before they were removed), patches
4230 @uref{../contribute.html,,following the usual requirements} would be
4231 likely to be accepted, since they should not affect the support for more
4234 For some systems, old versions of GNU binutils may also be useful,
4235 and are available from @file{pub/binutils/old-releases} on
4236 @uref{http://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
4238 Some of the information on specific systems above relates to
4239 such older systems, but much of the information
4240 about GCC on such systems (which may no longer be applicable to
4241 current GCC) is to be found in the GCC texinfo manual.
4246 @heading @anchor{elf}all ELF targets (SVR4, Solaris 2, etc.)
4248 C++ support is significantly better on ELF targets if you use the
4249 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
4250 inlines, vtables and template instantiations will be discarded
4259 @uref{./index.html,,Return to the GCC Installation page}
4263 @c ***Old documentation******************************************************
4265 @include install-old.texi
4271 @uref{./index.html,,Return to the GCC Installation page}
4275 @c ***GFDL********************************************************************
4283 @uref{./index.html,,Return to the GCC Installation page}
4287 @c ***************************************************************************
4288 @c Part 6 The End of the Document
4290 @comment node-name, next, previous, up
4291 @node Concept Index, , GNU Free Documentation License, Top
4295 @unnumbered Concept Index