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,
49 @c 2010, 2011 Free Software Foundation, Inc.
50 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
52 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
53 @c test the generation of HTML documents for the gcc.gnu.org web pages.
55 @c Do not use @footnote{} in this file as it breaks install.texi2html!
57 @c Include everything if we're not making html
61 @set prerequisiteshtml
72 @c Part 2 Summary Description and Copyright
74 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
75 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
76 2010 Free Software Foundation, Inc.
78 Permission is granted to copy, distribute and/or modify this document
79 under the terms of the GNU Free Documentation License, Version 1.3 or
80 any later version published by the Free Software Foundation; with no
81 Invariant Sections, the Front-Cover texts being (a) (see below), and
82 with the Back-Cover Texts being (b) (see below). A copy of the
83 license is included in the section entitled ``@uref{./gfdl.html,,GNU
84 Free Documentation License}''.
86 (a) The FSF's Front-Cover Text is:
90 (b) The FSF's Back-Cover Text is:
92 You have freedom to copy and modify this GNU Manual, like GNU
93 software. Copies published by the Free Software Foundation raise
94 funds for GNU development.
99 @dircategory Software development
101 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
104 @c Part 3 Titlepage and Copyright
106 @title Installing GCC
109 @c The following two commands start the copyright page.
111 @vskip 0pt plus 1filll
115 @c Part 4 Top node, Master Menu, and/or Table of Contents
118 @comment node-name, next, Previous, up
121 * Installing GCC:: This document describes the generic installation
122 procedure for GCC as well as detailing some target
123 specific installation instructions.
125 * Specific:: Host/target specific installation notes for GCC.
126 * Binaries:: Where to get pre-compiled binaries.
128 * Old:: Old installation documentation.
130 * GNU Free Documentation License:: How you can copy and share this manual.
131 * Concept Index:: This index has two entries.
139 @c Part 5 The Body of the Document
140 @c ***Installing GCC**********************************************************
142 @comment node-name, next, previous, up
143 @node Installing GCC, Binaries, , Top
147 @chapter Installing GCC
150 The latest version of this document is always available at
151 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
153 This document describes the generic installation procedure for GCC as well
154 as detailing some target specific installation instructions.
156 GCC includes several components that previously were separate distributions
157 with their own installation instructions. This document supersedes all
158 package specific installation instructions.
160 @emph{Before} starting the build/install procedure please check the
162 @ref{Specific, host/target specific installation notes}.
165 @uref{specific.html,,host/target specific installation notes}.
167 We recommend you browse the entire generic installation instructions before
170 Lists of successful builds for released versions of GCC are
171 available at @uref{http://gcc.gnu.org/buildstat.html}.
172 These lists are updated as new information becomes available.
174 The installation procedure itself is broken into five steps.
179 * Downloading the source::
182 * Testing:: (optional)
189 @uref{prerequisites.html,,Prerequisites}
191 @uref{download.html,,Downloading the source}
193 @uref{configure.html,,Configuration}
195 @uref{build.html,,Building}
197 @uref{test.html,,Testing} (optional)
199 @uref{finalinstall.html,,Final install}
203 Please note that GCC does not support @samp{make uninstall} and probably
204 won't do so in the near future as this would open a can of worms. Instead,
205 we suggest that you install GCC into a directory of its own and simply
206 remove that directory when you do not need that specific version of GCC
207 any longer, and, if shared libraries are installed there as well, no
208 more binaries exist that use them.
211 There are also some @uref{old.html,,old installation instructions},
212 which are mostly obsolete but still contain some information which has
213 not yet been merged into the main part of this manual.
221 @uref{./index.html,,Return to the GCC Installation page}
227 @c ***Prerequisites**************************************************
229 @comment node-name, next, previous, up
230 @node Prerequisites, Downloading the source, , Installing GCC
232 @ifset prerequisiteshtml
234 @chapter Prerequisites
236 @cindex Prerequisites
238 GCC requires that various tools and packages be available for use in the
239 build procedure. Modifying GCC sources requires additional tools
242 @heading Tools/packages necessary for building GCC
244 @item ISO C90 compiler
245 Necessary to bootstrap GCC, although versions of GCC prior
246 to 3.4 also allow bootstrapping with a traditional (K&R) C compiler.
248 To build all languages in a cross-compiler or other configuration where
249 3-stage bootstrap is not performed, you need to start with an existing
250 GCC binary (version 2.95 or later) because source code for language
251 frontends other than C might use GCC extensions.
255 In order to build the Ada compiler (GNAT) you must already have GNAT
256 installed because portions of the Ada frontend are written in Ada (with
257 GNAT extensions.) Refer to the Ada installation instructions for more
258 specific information.
260 @item A ``working'' POSIX compatible shell, or GNU bash
262 Necessary when running @command{configure} because some
263 @command{/bin/sh} shells have bugs and may crash when configuring the
264 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
265 have disastrous corner-case performance problems. This
266 can cause target @command{configure} runs to literally take days to
267 complete in some cases.
269 So on some platforms @command{/bin/ksh} is sufficient, on others it
270 isn't. See the host/target specific instructions for your platform, or
271 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
272 environment to your ``good'' shell prior to running
273 @command{configure}/@command{make}.
275 @command{zsh} is not a fully compliant POSIX shell and will not
276 work when configuring GCC@.
278 @item A POSIX or SVR4 awk
280 Necessary for creating some of the generated source files for GCC@.
281 If in doubt, use a recent GNU awk version, as some of the older ones
282 are broken. GNU awk version 3.1.5 is known to work.
286 Necessary in some circumstances, optional in others. See the
287 host/target specific instructions for your platform for the exact
290 @item gzip version 1.2.4 (or later) or
291 @itemx bzip2 version 1.0.2 (or later)
293 Necessary to uncompress GCC @command{tar} files when source code is
294 obtained via FTP mirror sites.
296 @item GNU make version 3.80 (or later)
298 You must have GNU make installed to build GCC@.
300 @item GNU tar version 1.14 (or later)
302 Necessary (only on some platforms) to untar the source code. Many
303 systems' @command{tar} programs will also work, only try GNU
304 @command{tar} if you have problems.
306 @item Perl version 5.6.1 (or later)
308 Necessary when targetting Darwin, building @samp{libstdc++},
309 and not using @option{--disable-symvers}.
310 Necessary when targetting Solaris 2 with Sun @command{ld} and not using
311 @option{--disable-symvers}. The bundled @command{perl} in Solaris@tie{}8
314 Necessary when regenerating @file{Makefile} dependencies in libiberty.
315 Necessary when regenerating @file{libiberty/functions.texi}.
316 Necessary when generating manpages from Texinfo manuals.
317 Used by various scripts to generate some files included in SVN (mainly
318 Unicode-related and rarely changing) from source tables.
320 @item @command{jar}, or InfoZIP (@command{zip} and @command{unzip})
322 Necessary to build libgcj, the GCJ runtime.
326 Several support libraries are necessary to build GCC, some are required,
327 others optional. While any sufficiently new version of required tools
328 usually work, library requirements are generally stricter. Newer
329 versions may work in some cases, but it's safer to use the exact
330 versions documented. We appreciate bug reports about problems with
331 newer versions, though.
334 @item GNU Multiple Precision Library (GMP) version 4.3.2 (or later)
336 Necessary to build GCC@. If you do not have it installed in your
337 library search path, you will have to configure with the
338 @option{--with-gmp} configure option. See also @option{--with-gmp-lib}
339 and @option{--with-gmp-include}. Alternatively, if a GMP source
340 distribution is found in a subdirectory of your GCC sources named
341 @file{gmp}, it will be built together with GCC@.
343 @item MPFR Library version 2.4.2 (or later)
345 Necessary to build GCC@. It can be downloaded from
346 @uref{http://www.mpfr.org/}. The @option{--with-mpfr} configure
347 option should be used if your MPFR Library is not installed in your
348 default library search path. See also @option{--with-mpfr-lib} and
349 @option{--with-mpfr-include}. Alternatively, if a MPFR source
350 distribution is found in a subdirectory of your GCC sources named
351 @file{mpfr}, it will be built together with GCC@.
353 @item MPC Library version 0.8.1 (or later)
355 Necessary to build GCC@. It can be downloaded from
356 @uref{http://www.multiprecision.org/}. The @option{--with-mpc}
357 configure option should be used if your MPC Library is not installed
358 in your default library search path. See also @option{--with-mpc-lib}
359 and @option{--with-mpc-include}. Alternatively, if an MPC source
360 distribution is found in a subdirectory of your GCC sources named
361 @file{mpc}, it will be built together with GCC@.
363 @item Parma Polyhedra Library (PPL) version 0.11
365 Necessary to build GCC with the Graphite loop optimizations.
366 It can be downloaded from @uref{http://www.cs.unipr.it/ppl/Download/}.
368 The @option{--with-ppl} configure option should be used if PPL is not
369 installed in your default library search path.
371 @item CLooG-PPL version 0.15 or CLooG 0.16
373 Necessary to build GCC with the Graphite loop optimizations. There
374 are two versions available. CLooG-PPL 0.15 as well as CLooG 0.16.
375 The former is the default right now. It can be downloaded from
376 @uref{ftp://gcc.gnu.org/pub/gcc/infrastructure/} as
377 @file{cloog-ppl-0.15.tar.gz}.
379 CLooG 0.16 support is still in testing stage, but will be the
380 default in future GCC releases. It is also available at
381 @uref{ftp://gcc.gnu.org/pub/gcc/infrastructure/} as
382 @file{cloog-0.16.1.tar.gz}. To use it add the additional configure
383 option @option{--enable-cloog-backend=isl}. Even if CLooG 0.16
384 does not use PPL, PPL is still required for Graphite.
386 In both cases @option{--with-cloog} configure option should be used
387 if CLooG is not installed in your default library search path.
391 @heading Tools/packages necessary for modifying GCC
393 @item autoconf version 2.64
394 @itemx GNU m4 version 1.4.6 (or later)
396 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
397 to regenerate @file{configure} and @file{config.in} files.
399 @item automake version 1.11.1
401 Necessary when modifying a @file{Makefile.am} file to regenerate its
402 associated @file{Makefile.in}.
404 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
405 file. Specifically this applies to the @file{gcc}, @file{intl},
406 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
407 as any of their subdirectories.
409 For directories that use automake, GCC requires the latest release in
410 the 1.11 series, which is currently 1.11.1. When regenerating a directory
411 to a newer version, please update all the directories using an older 1.11
412 to the latest released version.
414 @item gettext version 0.14.5 (or later)
416 Needed to regenerate @file{gcc.pot}.
418 @item gperf version 2.7.2 (or later)
420 Necessary when modifying @command{gperf} input files, e.g.@:
421 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
422 @file{gcc/cp/cfns.h}.
428 Necessary to run the GCC testsuite; see the section on testing for details.
430 @item autogen version 5.5.4 (or later) and
431 @itemx guile version 1.4.1 (or later)
433 Necessary to regenerate @file{fixinc/fixincl.x} from
434 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
436 Necessary to run @samp{make check} for @file{fixinc}.
438 Necessary to regenerate the top level @file{Makefile.in} file from
439 @file{Makefile.tpl} and @file{Makefile.def}.
441 @item Flex version 2.5.4 (or later)
443 Necessary when modifying @file{*.l} files.
445 Necessary to build GCC during development because the generated output
446 files are not included in the SVN repository. They are included in
449 @item Texinfo version 4.7 (or later)
451 Necessary for running @command{makeinfo} when modifying @file{*.texi}
452 files to test your changes.
454 Necessary for running @command{make dvi} or @command{make pdf} to
455 create printable documentation in DVI or PDF format. Texinfo version
456 4.8 or later is required for @command{make pdf}.
458 Necessary to build GCC documentation during development because the
459 generated output files are not included in the SVN repository. They are
460 included in releases.
462 @item @TeX{} (any working version)
464 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
465 are used when running @command{make dvi} or @command{make pdf} to create
466 DVI or PDF files, respectively.
468 @item SVN (any version)
469 @itemx SSH (any version)
471 Necessary to access the SVN repository. Public releases and weekly
472 snapshots of the development sources are also available via FTP@.
474 @item GNU diffutils version 2.7 (or later)
476 Useful when submitting patches for the GCC source code.
478 @item patch version 2.5.4 (or later)
480 Necessary when applying patches, created with @command{diff}, to one's
486 If you wish to modify @file{.java} files in libjava, you will need to
487 configure with @option{--enable-java-maintainer-mode}, and you will need
488 to have executables named @command{ecj1} and @command{gjavah} in your path.
489 The @command{ecj1} executable should run the Eclipse Java compiler via
490 the GCC-specific entry point. You can download a suitable jar from
491 @uref{ftp://sourceware.org/pub/java/}, or by running the script
492 @command{contrib/download_ecj}.
494 @item antlr.jar version 2.7.1 (or later)
497 If you wish to build the @command{gjdoc} binary in libjava, you will
498 need to have an @file{antlr.jar} library available. The library is
499 searched in system locations but can be configured with
500 @option{--with-antlr-jar=} instead. When configuring with
501 @option{--enable-java-maintainer-mode}, you will need to have one of
502 the executables named @command{cantlr}, @command{runantlr} or
503 @command{antlr} in your path.
512 @uref{./index.html,,Return to the GCC Installation page}
516 @c ***Downloading the source**************************************************
518 @comment node-name, next, previous, up
519 @node Downloading the source, Configuration, Prerequisites, Installing GCC
523 @chapter Downloading GCC
525 @cindex Downloading GCC
526 @cindex Downloading the Source
528 GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP
529 tarballs compressed with @command{gzip} or
530 @command{bzip2}. It is possible to download a full distribution or specific
533 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
534 for information on how to obtain GCC@.
536 The full distribution includes the C, C++, Objective-C, Fortran, Java,
537 and Ada (in the case of GCC 3.1 and later) compilers. The full
538 distribution also includes runtime libraries for C++, Objective-C,
539 Fortran, and Java. In GCC 3.0 and later versions, the GNU compiler
540 testsuites are also included in the full distribution.
542 If you choose to download specific components, you must download the core
543 GCC distribution plus any language specific distributions you wish to
544 use. The core distribution includes the C language front end as well as the
545 shared components. Each language has a tarball which includes the language
546 front end as well as the language runtime (when appropriate).
548 Unpack the core distribution as well as any language specific
549 distributions in the same directory.
551 If you also intend to build binutils (either to upgrade an existing
552 installation or for use in place of the corresponding tools of your
553 OS), unpack the binutils distribution either in the same directory or
554 a separate one. In the latter case, add symbolic links to any
555 components of the binutils you intend to build alongside the compiler
556 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
557 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
559 Likewise the GMP, MPFR and MPC libraries can be automatically built
560 together with GCC. Unpack the GMP, MPFR and/or MPC source
561 distributions in the directory containing the GCC sources and rename
562 their directories to @file{gmp}, @file{mpfr} and @file{mpc},
563 respectively (or use symbolic links with the same name).
570 @uref{./index.html,,Return to the GCC Installation page}
574 @c ***Configuration***********************************************************
576 @comment node-name, next, previous, up
577 @node Configuration, Building, Downloading the source, Installing GCC
581 @chapter Installing GCC: Configuration
583 @cindex Configuration
584 @cindex Installing GCC: Configuration
586 Like most GNU software, GCC must be configured before it can be built.
587 This document describes the recommended configuration procedure
588 for both native and cross targets.
590 We use @var{srcdir} to refer to the toplevel source directory for
591 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
593 If you obtained the sources via SVN, @var{srcdir} must refer to the top
594 @file{gcc} directory, the one where the @file{MAINTAINERS} file can be
595 found, and not its @file{gcc} subdirectory, otherwise the build will fail.
597 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
598 file system, the shell's built-in @command{pwd} command will return
599 temporary pathnames. Using these can lead to various sorts of build
600 problems. To avoid this issue, set the @env{PWDCMD} environment
601 variable to an automounter-aware @command{pwd} command, e.g.,
602 @command{pawd} or @samp{amq -w}, during the configuration and build
605 First, we @strong{highly} recommend that GCC be built into a
606 separate directory from the sources which does @strong{not} reside
607 within the source tree. This is how we generally build GCC; building
608 where @var{srcdir} == @var{objdir} should still work, but doesn't
609 get extensive testing; building where @var{objdir} is a subdirectory
610 of @var{srcdir} is unsupported.
612 If you have previously built GCC in the same directory for a
613 different target machine, do @samp{make distclean} to delete all files
614 that might be invalid. One of the files this deletes is @file{Makefile};
615 if @samp{make distclean} complains that @file{Makefile} does not exist
616 or issues a message like ``don't know how to make distclean'' it probably
617 means that the directory is already suitably clean. However, with the
618 recommended method of building in a separate @var{objdir}, you should
619 simply use a different @var{objdir} for each target.
621 Second, when configuring a native system, either @command{cc} or
622 @command{gcc} must be in your path or you must set @env{CC} in
623 your environment before running configure. Otherwise the configuration
627 Note that the bootstrap compiler and the resulting GCC must be link
628 compatible, else the bootstrap will fail with linker errors about
629 incompatible object file formats. Several multilibed targets are
630 affected by this requirement, see
632 @ref{Specific, host/target specific installation notes}.
635 @uref{specific.html,,host/target specific installation notes}.
644 % @var{srcdir}/configure [@var{options}] [@var{target}]
647 @heading Distributor options
649 If you will be distributing binary versions of GCC, with modifications
650 to the source code, you should use the options described in this
651 section to make clear that your version contains modifications.
654 @item --with-pkgversion=@var{version}
655 Specify a string that identifies your package. You may wish
656 to include a build number or build date. This version string will be
657 included in the output of @command{gcc --version}. This suffix does
658 not replace the default version string, only the @samp{GCC} part.
660 The default value is @samp{GCC}.
662 @item --with-bugurl=@var{url}
663 Specify the URL that users should visit if they wish to report a bug.
664 You are of course welcome to forward bugs reported to you to the FSF,
665 if you determine that they are not bugs in your modifications.
667 The default value refers to the FSF's GCC bug tracker.
671 @heading Target specification
674 GCC has code to correctly determine the correct value for @var{target}
675 for nearly all native systems. Therefore, we highly recommend you do
676 not provide a configure target when configuring a native compiler.
679 @var{target} must be specified as @option{--target=@var{target}}
680 when configuring a cross compiler; examples of valid targets would be
681 m68k-elf, sh-elf, etc.
684 Specifying just @var{target} instead of @option{--target=@var{target}}
685 implies that the host defaults to @var{target}.
689 @heading Options specification
691 Use @var{options} to override several configure time options for
692 GCC@. A list of supported @var{options} follows; @samp{configure
693 --help} may list other options, but those not listed below may not
694 work and should not normally be used.
696 Note that each @option{--enable} option has a corresponding
697 @option{--disable} option and that each @option{--with} option has a
698 corresponding @option{--without} option.
701 @item --prefix=@var{dirname}
702 Specify the toplevel installation
703 directory. This is the recommended way to install the tools into a directory
704 other than the default. The toplevel installation directory defaults to
707 We @strong{highly} recommend against @var{dirname} being the same or a
708 subdirectory of @var{objdir} or vice versa. If specifying a directory
709 beneath a user's home directory tree, some shells will not expand
710 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
713 The following standard @command{autoconf} options are supported. Normally you
714 should not need to use these options.
716 @item --exec-prefix=@var{dirname}
717 Specify the toplevel installation directory for architecture-dependent
718 files. The default is @file{@var{prefix}}.
720 @item --bindir=@var{dirname}
721 Specify the installation directory for the executables called by users
722 (such as @command{gcc} and @command{g++}). The default is
723 @file{@var{exec-prefix}/bin}.
725 @item --libdir=@var{dirname}
726 Specify the installation directory for object code libraries and
727 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
729 @item --libexecdir=@var{dirname}
730 Specify the installation directory for internal executables of GCC@.
731 The default is @file{@var{exec-prefix}/libexec}.
733 @item --with-slibdir=@var{dirname}
734 Specify the installation directory for the shared libgcc library. The
735 default is @file{@var{libdir}}.
737 @item --datarootdir=@var{dirname}
738 Specify the root of the directory tree for read-only architecture-independent
739 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
741 @item --infodir=@var{dirname}
742 Specify the installation directory for documentation in info format.
743 The default is @file{@var{datarootdir}/info}.
745 @item --datadir=@var{dirname}
746 Specify the installation directory for some architecture-independent
747 data files referenced by GCC@. The default is @file{@var{datarootdir}}.
749 @item --docdir=@var{dirname}
750 Specify the installation directory for documentation files (other
751 than Info) for GCC@. The default is @file{@var{datarootdir}/doc}.
753 @item --htmldir=@var{dirname}
754 Specify the installation directory for HTML documentation files.
755 The default is @file{@var{docdir}}.
757 @item --pdfdir=@var{dirname}
758 Specify the installation directory for PDF documentation files.
759 The default is @file{@var{docdir}}.
761 @item --mandir=@var{dirname}
762 Specify the installation directory for manual pages. The default is
763 @file{@var{datarootdir}/man}. (Note that the manual pages are only extracts
764 from the full GCC manuals, which are provided in Texinfo format. The manpages
765 are derived by an automatic conversion process from parts of the full
768 @item --with-gxx-include-dir=@var{dirname}
770 the installation directory for G++ header files. The default depends
771 on other configuration options, and differs between cross and native
776 @item --program-prefix=@var{prefix}
777 GCC supports some transformations of the names of its programs when
778 installing them. This option prepends @var{prefix} to the names of
779 programs to install in @var{bindir} (see above). For example, specifying
780 @option{--program-prefix=foo-} would result in @samp{gcc}
781 being installed as @file{/usr/local/bin/foo-gcc}.
783 @item --program-suffix=@var{suffix}
784 Appends @var{suffix} to the names of programs to install in @var{bindir}
785 (see above). For example, specifying @option{--program-suffix=-3.1}
786 would result in @samp{gcc} being installed as
787 @file{/usr/local/bin/gcc-3.1}.
789 @item --program-transform-name=@var{pattern}
790 Applies the @samp{sed} script @var{pattern} to be applied to the names
791 of programs to install in @var{bindir} (see above). @var{pattern} has to
792 consist of one or more basic @samp{sed} editing commands, separated by
793 semicolons. For example, if you want the @samp{gcc} program name to be
794 transformed to the installed program @file{/usr/local/bin/myowngcc} and
795 the @samp{g++} program name to be transformed to
796 @file{/usr/local/bin/gspecial++} without changing other program names,
797 you could use the pattern
798 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
799 to achieve this effect.
801 All three options can be combined and used together, resulting in more
802 complex conversion patterns. As a basic rule, @var{prefix} (and
803 @var{suffix}) are prepended (appended) before further transformations
804 can happen with a special transformation script @var{pattern}.
806 As currently implemented, this option only takes effect for native
807 builds; cross compiler binaries' names are not transformed even when a
808 transformation is explicitly asked for by one of these options.
810 For native builds, some of the installed programs are also installed
811 with the target alias in front of their name, as in
812 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
813 before the target alias is prepended to the name---so, specifying
814 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
815 resulting binary would be installed as
816 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
818 As a last shortcoming, none of the installed Ada programs are
819 transformed yet, which will be fixed in some time.
821 @item --with-local-prefix=@var{dirname}
823 installation directory for local include files. The default is
824 @file{/usr/local}. Specify this option if you want the compiler to
825 search directory @file{@var{dirname}/include} for locally installed
826 header files @emph{instead} of @file{/usr/local/include}.
828 You should specify @option{--with-local-prefix} @strong{only} if your
829 site has a different convention (not @file{/usr/local}) for where to put
832 The default value for @option{--with-local-prefix} is @file{/usr/local}
833 regardless of the value of @option{--prefix}. Specifying
834 @option{--prefix} has no effect on which directory GCC searches for
835 local header files. This may seem counterintuitive, but actually it is
838 The purpose of @option{--prefix} is to specify where to @emph{install
839 GCC}. The local header files in @file{/usr/local/include}---if you put
840 any in that directory---are not part of GCC@. They are part of other
841 programs---perhaps many others. (GCC installs its own header files in
842 another directory which is based on the @option{--prefix} value.)
844 Both the local-prefix include directory and the GCC-prefix include
845 directory are part of GCC's ``system include'' directories. Although these
846 two directories are not fixed, they need to be searched in the proper
847 order for the correct processing of the include_next directive. The
848 local-prefix include directory is searched before the GCC-prefix
849 include directory. Another characteristic of system include directories
850 is that pedantic warnings are turned off for headers in these directories.
852 Some autoconf macros add @option{-I @var{directory}} options to the
853 compiler command line, to ensure that directories containing installed
854 packages' headers are searched. When @var{directory} is one of GCC's
855 system include directories, GCC will ignore the option so that system
856 directories continue to be processed in the correct order. This
857 may result in a search order different from what was specified but the
858 directory will still be searched.
860 GCC automatically searches for ordinary libraries using
861 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
862 used for both GCC and packages, GCC will automatically search for
863 both headers and libraries. This provides a configuration that is
864 easy to use. GCC behaves in a manner similar to that when it is
865 installed as a system compiler in @file{/usr}.
867 Sites that need to install multiple versions of GCC may not want to
868 use the above simple configuration. It is possible to use the
869 @option{--program-prefix}, @option{--program-suffix} and
870 @option{--program-transform-name} options to install multiple versions
871 into a single directory, but it may be simpler to use different prefixes
872 and the @option{--with-local-prefix} option to specify the location of the
873 site-specific files for each version. It will then be necessary for
874 users to specify explicitly the location of local site libraries
875 (e.g., with @env{LIBRARY_PATH}).
877 The same value can be used for both @option{--with-local-prefix} and
878 @option{--prefix} provided it is not @file{/usr}. This can be used
879 to avoid the default search of @file{/usr/local/include}.
881 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
882 The directory you use for @option{--with-local-prefix} @strong{must not}
883 contain any of the system's standard header files. If it did contain
884 them, certain programs would be miscompiled (including GNU Emacs, on
885 certain targets), because this would override and nullify the header
886 file corrections made by the @command{fixincludes} script.
888 Indications are that people who use this option use it based on mistaken
889 ideas of what it is for. People use it as if it specified where to
890 install part of GCC@. Perhaps they make this assumption because
891 installing GCC creates the directory.
893 @item --enable-shared[=@var{package}[,@dots{}]]
894 Build shared versions of libraries, if shared libraries are supported on
895 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
896 are enabled by default on all platforms that support shared libraries.
898 If a list of packages is given as an argument, build shared libraries
899 only for the listed packages. For other packages, only static libraries
900 will be built. Package names currently recognized in the GCC tree are
901 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
902 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
903 @samp{ada}, @samp{libada}, @samp{libjava}, @samp{libgo}, and @samp{libobjc}.
904 Note @samp{libiberty} does not support shared libraries at all.
906 Use @option{--disable-shared} to build only static libraries. Note that
907 @option{--disable-shared} does not accept a list of package names as
908 argument, only @option{--enable-shared} does.
910 @item @anchor{with-gnu-as}--with-gnu-as
911 Specify that the compiler should assume that the
912 assembler it finds is the GNU assembler. However, this does not modify
913 the rules to find an assembler and will result in confusion if the
914 assembler found is not actually the GNU assembler. (Confusion may also
915 result if the compiler finds the GNU assembler but has not been
916 configured with @option{--with-gnu-as}.) If you have more than one
917 assembler installed on your system, you may want to use this option in
918 connection with @option{--with-as=@var{pathname}} or
919 @option{--with-build-time-tools=@var{pathname}}.
921 The following systems are the only ones where it makes a difference
922 whether you use the GNU assembler. On any other system,
923 @option{--with-gnu-as} has no effect.
926 @item @samp{hppa1.0-@var{any}-@var{any}}
927 @item @samp{hppa1.1-@var{any}-@var{any}}
928 @item @samp{sparc-sun-solaris2.@var{any}}
929 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
932 @item @anchor{with-as}--with-as=@var{pathname}
933 Specify that the compiler should use the assembler pointed to by
934 @var{pathname}, rather than the one found by the standard rules to find
935 an assembler, which are:
938 Unless GCC is being built with a cross compiler, check the
939 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
940 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
941 @var{exec-prefix} defaults to @var{prefix}, which
942 defaults to @file{/usr/local} unless overridden by the
943 @option{--prefix=@var{pathname}} switch described above. @var{target}
944 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
945 @var{version} denotes the GCC version, such as 3.0.
948 If the target system is the same that you are building on, check
949 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
953 Check in the @env{PATH} for a tool whose name is prefixed by the
954 target system triple.
957 Check in the @env{PATH} for a tool whose name is not prefixed by the
958 target system triple, if the host and target system triple are
959 the same (in other words, we use a host tool if it can be used for
963 You may want to use @option{--with-as} if no assembler
964 is installed in the directories listed above, or if you have multiple
965 assemblers installed and want to choose one that is not found by the
968 @item @anchor{with-gnu-ld}--with-gnu-ld
969 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
972 @item --with-ld=@var{pathname}
973 Same as @uref{#with-as,,@option{--with-as}}
977 Specify that stabs debugging
978 information should be used instead of whatever format the host normally
979 uses. Normally GCC uses the same debug format as the host system.
981 On MIPS based systems and on Alphas, you must specify whether you want
982 GCC to create the normal ECOFF debugging format, or to use BSD-style
983 stabs passed through the ECOFF symbol table. The normal ECOFF debug
984 format cannot fully handle languages other than C@. BSD stabs format can
985 handle other languages, but it only works with the GNU debugger GDB@.
987 Normally, GCC uses the ECOFF debugging format by default; if you
988 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
990 No matter which default you choose when you configure GCC, the user
991 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
992 the debug format for a particular compilation.
994 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
995 @option{--with-gas} is used. It selects use of stabs debugging
996 information embedded in COFF output. This kind of debugging information
997 supports C++ well; ordinary COFF debugging information does not.
999 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
1000 selects use of stabs debugging information embedded in ELF output. The
1001 C++ compiler currently (2.6.0) does not support the DWARF debugging
1002 information normally used on 386 SVR4 platforms; stabs provide a
1003 workable alternative. This requires gas and gdb, as the normal SVR4
1004 tools can not generate or interpret stabs.
1006 @item --disable-multilib
1007 Specify that multiple target
1008 libraries to support different target variants, calling
1009 conventions, etc.@: should not be built. The default is to build a
1010 predefined set of them.
1012 Some targets provide finer-grained control over which multilibs are built
1013 (e.g., @option{--disable-softfloat}):
1016 fpu, 26bit, underscore, interwork, biendian, nofmult.
1019 softfloat, m68881, m68000, m68020.
1022 single-float, biendian, softfloat.
1024 @item powerpc*-*-*, rs6000*-*-*
1025 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
1030 @item --with-multilib-list=@var{list}
1031 @itemx --without-multilib-list
1032 Specify what multilibs to build.
1033 Currently only implemented for sh*-*-*.
1035 @var{list} is a comma separated list of CPU names. These must be of the
1036 form @code{sh*} or @code{m*} (in which case they match the compiler option
1037 for that processor). The list should not contain any endian options -
1038 these are handled by @option{--with-endian}.
1040 If @var{list} is empty, then there will be no multilibs for extra
1041 processors. The multilib for the secondary endian remains enabled.
1043 As a special case, if an entry in the list starts with a @code{!}
1044 (exclamation point), then it is added to the list of excluded multilibs.
1045 Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES}
1046 (once the leading @code{!} has been stripped).
1048 If @option{--with-multilib-list} is not given, then a default set of
1049 multilibs is selected based on the value of @option{--target}. This is
1050 usually the complete set of libraries, but some targets imply a more
1053 Example 1: to configure a compiler for SH4A only, but supporting both
1054 endians, with little endian being the default:
1056 --with-cpu=sh4a --with-endian=little,big --with-multilib-list=
1059 Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with
1060 only little endian SH4AL:
1062 --with-cpu=sh4a --with-endian=little,big \
1063 --with-multilib-list=sh4al,!mb/m4al
1066 @item --with-endian=@var{endians}
1067 Specify what endians to use.
1068 Currently only implemented for sh*-*-*.
1070 @var{endians} may be one of the following:
1073 Use big endian exclusively.
1075 Use little endian exclusively.
1077 Use big endian by default. Provide a multilib for little endian.
1079 Use little endian by default. Provide a multilib for big endian.
1082 @item --enable-threads
1083 Specify that the target
1084 supports threads. This affects the Objective-C compiler and runtime
1085 library, and exception handling for other languages like C++ and Java.
1086 On some systems, this is the default.
1088 In general, the best (and, in many cases, the only known) threading
1089 model available will be configured for use. Beware that on some
1090 systems, GCC has not been taught what threading models are generally
1091 available for the system. In this case, @option{--enable-threads} is an
1092 alias for @option{--enable-threads=single}.
1094 @item --disable-threads
1095 Specify that threading support should be disabled for the system.
1096 This is an alias for @option{--enable-threads=single}.
1098 @item --enable-threads=@var{lib}
1100 @var{lib} is the thread support library. This affects the Objective-C
1101 compiler and runtime library, and exception handling for other languages
1102 like C++ and Java. The possibilities for @var{lib} are:
1110 Ada tasking support. For non-Ada programs, this setting is equivalent
1111 to @samp{single}. When used in conjunction with the Ada run time, it
1112 causes GCC to use the same thread primitives as Ada uses. This option
1113 is necessary when using both Ada and the back end exception handling,
1114 which is the default for most Ada targets.
1116 LynxOS thread support.
1118 MIPS SDE thread support.
1120 Novell Kernel Services thread support.
1122 This is an alias for @samp{single}.
1124 Generic POSIX/Unix98 thread support.
1126 Generic POSIX/Unix95 thread support.
1128 RTEMS thread support.
1130 Disable thread support, should work for all platforms.
1134 VxWorks thread support.
1136 Microsoft Win32 API thread support.
1140 Specify that the target supports TLS (Thread Local Storage). Usually
1141 configure can correctly determine if TLS is supported. In cases where
1142 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1143 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1144 the assembler supports TLS but the C library does not, or if the
1145 assumptions made by the configure test are incorrect.
1148 Specify that the target does not support TLS.
1149 This is an alias for @option{--enable-tls=no}.
1151 @item --with-cpu=@var{cpu}
1152 @itemx --with-cpu-32=@var{cpu}
1153 @itemx --with-cpu-64=@var{cpu}
1154 Specify which cpu variant the compiler should generate code for by default.
1155 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1156 This option is only supported on some targets, including ARM, i386, M68k,
1157 PowerPC, and SPARC@. The @option{--with-cpu-32} and
1158 @option{--with-cpu-64} options specify separate default CPUs for
1159 32-bit and 64-bit modes; these options are only supported for i386,
1162 @item --with-schedule=@var{cpu}
1163 @itemx --with-arch=@var{cpu}
1164 @itemx --with-arch-32=@var{cpu}
1165 @itemx --with-arch-64=@var{cpu}
1166 @itemx --with-tune=@var{cpu}
1167 @itemx --with-tune-32=@var{cpu}
1168 @itemx --with-tune-64=@var{cpu}
1169 @itemx --with-abi=@var{abi}
1170 @itemx --with-fpu=@var{type}
1171 @itemx --with-float=@var{type}
1172 These configure options provide default values for the @option{-mschedule=},
1173 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1174 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1175 @option{--with-cpu}, which switches will be accepted and acceptable values
1176 of the arguments depend on the target.
1178 @item --with-mode=@var{mode}
1179 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1180 This option is only supported on ARM targets.
1182 @item --with-fpmath=@var{isa}
1183 This options sets @option{-mfpmath=sse} by default and specifies the default
1184 ISA for floating-point arithmetics. You can select either @samp{sse} which
1185 enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default.
1186 This option is only supported on i386 and x86-64 targets.
1188 @item --with-divide=@var{type}
1189 Specify how the compiler should generate code for checking for
1190 division by zero. This option is only supported on the MIPS target.
1191 The possibilities for @var{type} are:
1194 Division by zero checks use conditional traps (this is the default on
1195 systems that support conditional traps).
1197 Division by zero checks use the break instruction.
1200 @c If you make --with-llsc the default for additional targets,
1201 @c update the --with-llsc description in the MIPS section below.
1204 On MIPS targets, make @option{-mllsc} the default when no
1205 @option{-mno-lsc} option is passed. This is the default for
1206 Linux-based targets, as the kernel will emulate them if the ISA does
1209 @item --without-llsc
1210 On MIPS targets, make @option{-mno-llsc} the default when no
1211 @option{-mllsc} option is passed.
1214 On MIPS targets, make @option{-msynci} the default when no
1215 @option{-mno-synci} option is passed.
1217 @item --without-synci
1218 On MIPS targets, make @option{-mno-synci} the default when no
1219 @option{-msynci} option is passed. This is the default.
1221 @item --with-mips-plt
1222 On MIPS targets, make use of copy relocations and PLTs.
1223 These features are extensions to the traditional
1224 SVR4-based MIPS ABIs and require support from GNU binutils
1225 and the runtime C library.
1227 @item --enable-__cxa_atexit
1228 Define if you want to use __cxa_atexit, rather than atexit, to
1229 register C++ destructors for local statics and global objects.
1230 This is essential for fully standards-compliant handling of
1231 destructors, but requires __cxa_atexit in libc. This option is currently
1232 only available on systems with GNU libc. When enabled, this will cause
1233 @option{-fuse-cxa-atexit} to be passed by default.
1235 @item --enable-indirect-function
1236 Define if you want to enable the @code{ifunc} attribute. This option is
1237 currently only available on systems with GNU libc on certain targets.
1239 @item --enable-target-optspace
1241 libraries should be optimized for code space instead of code speed.
1242 This is the default for the m32r platform.
1244 @item --with-cpp-install-dir=@var{dirname}
1245 Specify that the user visible @command{cpp} program should be installed
1246 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1248 @item --enable-comdat
1249 Enable COMDAT group support. This is primarily used to override the
1250 automatically detected value.
1252 @item --enable-initfini-array
1253 Force the use of sections @code{.init_array} and @code{.fini_array}
1254 (instead of @code{.init} and @code{.fini}) for constructors and
1255 destructors. Option @option{--disable-initfini-array} has the
1256 opposite effect. If neither option is specified, the configure script
1257 will try to guess whether the @code{.init_array} and
1258 @code{.fini_array} sections are supported and, if they are, use them.
1260 @item --enable-build-with-cxx
1261 Build GCC using a C++ compiler rather than a C compiler. This is an
1262 experimental option which may become the default in a later release.
1264 @item --enable-maintainer-mode
1265 The build rules that regenerate the Autoconf and Automake output files as
1266 well as the GCC master message catalog @file{gcc.pot} are normally
1267 disabled. This is because it can only be rebuilt if the complete source
1268 tree is present. If you have changed the sources and want to rebuild the
1269 catalog, configuring with @option{--enable-maintainer-mode} will enable
1270 this. Note that you need a recent version of the @code{gettext} tools
1273 @item --disable-bootstrap
1274 For a native build, the default configuration is to perform
1275 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1276 testing that GCC can compile itself correctly. If you want to disable
1277 this process, you can configure with @option{--disable-bootstrap}.
1279 @item --enable-bootstrap
1280 In special cases, you may want to perform a 3-stage build
1281 even if the target and host triplets are different.
1282 This is possible when the host can run code compiled for
1283 the target (e.g.@: host is i686-linux, target is i486-linux).
1284 Starting from GCC 4.2, to do this you have to configure explicitly
1285 with @option{--enable-bootstrap}.
1287 @item --enable-generated-files-in-srcdir
1288 Neither the .c and .h files that are generated from Bison and flex nor the
1289 info manuals and man pages that are built from the .texi files are present
1290 in the SVN development tree. When building GCC from that development tree,
1291 or from one of our snapshots, those generated files are placed in your
1292 build directory, which allows for the source to be in a readonly
1295 If you configure with @option{--enable-generated-files-in-srcdir} then those
1296 generated files will go into the source directory. This is mainly intended
1297 for generating release or prerelease tarballs of the GCC sources, since it
1298 is not a requirement that the users of source releases to have flex, Bison,
1301 @item --enable-version-specific-runtime-libs
1303 that runtime libraries should be installed in the compiler specific
1304 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1305 addition, @samp{libstdc++}'s include files will be installed into
1306 @file{@var{libdir}} unless you overruled it by using
1307 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1308 particularly useful if you intend to use several versions of GCC in
1309 parallel. This is currently supported by @samp{libgfortran},
1310 @samp{libjava}, @samp{libmudflap}, @samp{libstdc++}, and @samp{libobjc}.
1312 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1313 Specify that only a particular subset of compilers and
1314 their runtime libraries should be built. For a list of valid values for
1315 @var{langN} you can issue the following command in the
1316 @file{gcc} directory of your GCC source tree:@*
1318 grep language= */config-lang.in
1320 Currently, you can use any of the following:
1321 @code{all}, @code{ada}, @code{c}, @code{c++}, @code{fortran},
1322 @code{go}, @code{java}, @code{objc}, @code{obj-c++}.
1323 Building the Ada compiler has special requirements, see below.
1324 If you do not pass this flag, or specify the option @code{all}, then all
1325 default languages available in the @file{gcc} sub-tree will be configured.
1326 Ada, Go and Objective-C++ are not default languages; the rest are.
1328 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1329 Specify that a particular subset of compilers and their runtime
1330 libraries should be built with the system C compiler during stage 1 of
1331 the bootstrap process, rather than only in later stages with the
1332 bootstrapped C compiler. The list of valid values is the same as for
1333 @option{--enable-languages}, and the option @code{all} will select all
1334 of the languages enabled by @option{--enable-languages}. This option is
1335 primarily useful for GCC development; for instance, when a development
1336 version of the compiler cannot bootstrap due to compiler bugs, or when
1337 one is debugging front ends other than the C front end. When this
1338 option is used, one can then build the target libraries for the
1339 specified languages with the stage-1 compiler by using @command{make
1340 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1341 for the specified languages using @command{make stage1-start check-gcc}.
1343 @item --disable-libada
1344 Specify that the run-time libraries and tools used by GNAT should not
1345 be built. This can be useful for debugging, or for compatibility with
1346 previous Ada build procedures, when it was required to explicitly
1347 do a @samp{make -C gcc gnatlib_and_tools}.
1349 @item --disable-libssp
1350 Specify that the run-time libraries for stack smashing protection
1351 should not be built.
1353 @item --disable-libquadmath
1354 Specify that the GCC quad-precision math library should not be built.
1355 On some systems, the library is required to be linkable when building
1356 the Fortran front end, unless @option{--disable-libquadmath-support}
1359 @item --disable-libquadmath-support
1360 Specify that the Fortran front end and @code{libgfortran} do not add
1361 support for @code{libquadmath} on systems supporting it.
1363 @item --disable-libgomp
1364 Specify that the run-time libraries used by GOMP should not be built.
1367 Specify that the compiler should
1368 use DWARF 2 debugging information as the default.
1370 @item --enable-targets=all
1371 @itemx --enable-targets=@var{target_list}
1372 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1373 These are compilers that are able to generate either 64-bit or 32-bit
1374 code. Typically, the corresponding 32-bit target, e.g.@:
1375 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1376 option enables the 32-bit target to be a bi-arch compiler, which is
1377 useful when you want a bi-arch compiler that defaults to 32-bit, and
1378 you are building a bi-arch or multi-arch binutils in a combined tree.
1379 On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64),
1381 Currently, this option only affects sparc-linux, powerpc-linux, x86-linux
1384 @item --enable-secureplt
1385 This option enables @option{-msecure-plt} by default for powerpc-linux.
1387 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1388 Using the GNU Compiler Collection (GCC)},
1391 See ``RS/6000 and PowerPC Options'' in the main manual
1395 This option enables @option{-mcld} by default for 32-bit x86 targets.
1397 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1398 Using the GNU Compiler Collection (GCC)},
1401 See ``i386 and x86-64 Options'' in the main manual
1404 @item --enable-win32-registry
1405 @itemx --enable-win32-registry=@var{key}
1406 @itemx --disable-win32-registry
1407 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1408 to look up installations paths in the registry using the following key:
1411 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1414 @var{key} defaults to GCC version number, and can be overridden by the
1415 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1416 who use custom installers are encouraged to provide a different key,
1417 perhaps one comprised of vendor name and GCC version number, to
1418 avoid conflict with existing installations. This feature is enabled
1419 by default, and can be disabled by @option{--disable-win32-registry}
1420 option. This option has no effect on the other hosts.
1423 Specify that the machine does not have a floating point unit. This
1424 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1425 system, @option{--nfp} has no effect.
1427 @item --enable-werror
1428 @itemx --disable-werror
1429 @itemx --enable-werror=yes
1430 @itemx --enable-werror=no
1431 When you specify this option, it controls whether certain files in the
1432 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1433 If you don't specify it, @option{-Werror} is turned on for the main
1434 development trunk. However it defaults to off for release branches and
1435 final releases. The specific files which get @option{-Werror} are
1436 controlled by the Makefiles.
1438 @item --enable-checking
1439 @itemx --enable-checking=@var{list}
1440 When you specify this option, the compiler is built to perform internal
1441 consistency checks of the requested complexity. This does not change the
1442 generated code, but adds error checking within the compiler. This will
1443 slow down the compiler and may only work properly if you are building
1444 the compiler with GCC@. This is @samp{yes} by default when building
1445 from SVN or snapshots, but @samp{release} for releases. The default
1446 for building the stage1 compiler is @samp{yes}. More control
1447 over the checks may be had by specifying @var{list}. The categories of
1448 checks available are @samp{yes} (most common checks
1449 @samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at
1450 all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest
1451 checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1452 Individual checks can be enabled with these flags @samp{assert},
1453 @samp{df}, @samp{fold}, @samp{gc}, @samp{gcac} @samp{misc}, @samp{rtl},
1454 @samp{rtlflag}, @samp{runtime}, @samp{tree}, and @samp{valgrind}.
1456 The @samp{valgrind} check requires the external @command{valgrind}
1457 simulator, available from @uref{http://valgrind.org/}. The
1458 @samp{df}, @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive.
1459 To disable all checking, @samp{--disable-checking} or
1460 @samp{--enable-checking=none} must be explicitly requested. Disabling
1461 assertions will make the compiler and runtime slightly faster but
1462 increase the risk of undetected internal errors causing wrong code to be
1465 @item --disable-stage1-checking
1466 @itemx --enable-stage1-checking
1467 @itemx --enable-stage1-checking=@var{list}
1468 If no @option{--enable-checking} option is specified the stage1
1469 compiler will be built with @samp{yes} checking enabled, otherwise
1470 the stage1 checking flags are the same as specified by
1471 @option{--enable-checking}. To build the stage1 compiler with
1472 different checking options use @option{--enable-stage1-checking}.
1473 The list of checking options is the same as for @option{--enable-checking}.
1474 If your system is too slow or too small to bootstrap a released compiler
1475 with checking for stage1 enabled, you can use @samp{--disable-stage1-checking}
1476 to disable checking for the stage1 compiler.
1478 @item --enable-coverage
1479 @itemx --enable-coverage=@var{level}
1480 With this option, the compiler is built to collect self coverage
1481 information, every time it is run. This is for internal development
1482 purposes, and only works when the compiler is being built with gcc. The
1483 @var{level} argument controls whether the compiler is built optimized or
1484 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1485 want to disable optimization, for performance analysis you want to
1486 enable optimization. When coverage is enabled, the default level is
1487 without optimization.
1489 @item --enable-gather-detailed-mem-stats
1490 When this option is specified more detailed information on memory
1491 allocation is gathered. This information is printed when using
1492 @option{-fmem-report}.
1495 @itemx --with-gc=@var{choice}
1496 With this option you can specify the garbage collector implementation
1497 used during the compilation process. @var{choice} can be one of
1498 @samp{page} and @samp{zone}, where @samp{page} is the default.
1501 @itemx --disable-nls
1502 The @option{--enable-nls} option enables Native Language Support (NLS),
1503 which lets GCC output diagnostics in languages other than American
1504 English. Native Language Support is enabled by default if not doing a
1505 canadian cross build. The @option{--disable-nls} option disables NLS@.
1507 @item --with-included-gettext
1508 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1509 procedure to prefer its copy of GNU @command{gettext}.
1511 @item --with-catgets
1512 If NLS is enabled, and if the host lacks @code{gettext} but has the
1513 inferior @code{catgets} interface, the GCC build procedure normally
1514 ignores @code{catgets} and instead uses GCC's copy of the GNU
1515 @code{gettext} library. The @option{--with-catgets} option causes the
1516 build procedure to use the host's @code{catgets} in this situation.
1518 @item --with-libiconv-prefix=@var{dir}
1519 Search for libiconv header files in @file{@var{dir}/include} and
1520 libiconv library files in @file{@var{dir}/lib}.
1522 @item --enable-obsolete
1523 Enable configuration for an obsoleted system. If you attempt to
1524 configure GCC for a system (build, host, or target) which has been
1525 obsoleted, and you do not specify this flag, configure will halt with an
1528 All support for systems which have been obsoleted in one release of GCC
1529 is removed entirely in the next major release, unless someone steps
1530 forward to maintain the port.
1532 @item --enable-decimal-float
1533 @itemx --enable-decimal-float=yes
1534 @itemx --enable-decimal-float=no
1535 @itemx --enable-decimal-float=bid
1536 @itemx --enable-decimal-float=dpd
1537 @itemx --disable-decimal-float
1538 Enable (or disable) support for the C decimal floating point extension
1539 that is in the IEEE 754-2008 standard. This is enabled by default only
1540 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
1541 support it, but require the user to specifically enable it. You can
1542 optionally control which decimal floating point format is used (either
1543 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
1544 format is default on i386 and x86_64 systems, and the @samp{dpd}
1545 (densely packed decimal) format is default on PowerPC systems.
1547 @item --enable-fixed-point
1548 @itemx --disable-fixed-point
1549 Enable (or disable) support for C fixed-point arithmetic.
1550 This option is enabled by default for some targets (such as MIPS) which
1551 have hardware-support for fixed-point operations. On other targets, you
1552 may enable this option manually.
1554 @item --with-long-double-128
1555 Specify if @code{long double} type should be 128-bit by default on selected
1556 GNU/Linux architectures. If using @code{--without-long-double-128},
1557 @code{long double} will be by default 64-bit, the same as @code{double} type.
1558 When neither of these configure options are used, the default will be
1559 128-bit @code{long double} when built against GNU C Library 2.4 and later,
1560 64-bit @code{long double} otherwise.
1562 @item --with-gmp=@var{pathname}
1563 @itemx --with-gmp-include=@var{pathname}
1564 @itemx --with-gmp-lib=@var{pathname}
1565 @itemx --with-mpfr=@var{pathname}
1566 @itemx --with-mpfr-include=@var{pathname}
1567 @itemx --with-mpfr-lib=@var{pathname}
1568 @itemx --with-mpc=@var{pathname}
1569 @itemx --with-mpc-include=@var{pathname}
1570 @itemx --with-mpc-lib=@var{pathname}
1571 If you do not have GMP (the GNU Multiple Precision library), the MPFR
1572 library and/or the MPC library installed in a standard location and
1573 you want to build GCC, you can explicitly specify the directory where
1574 they are installed (@samp{--with-gmp=@var{gmpinstalldir}},
1575 @samp{--with-mpfr=@/@var{mpfrinstalldir}},
1576 @samp{--with-mpc=@/@var{mpcinstalldir}}). The
1577 @option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for
1578 @option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and
1579 @option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the
1580 @option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for
1581 @option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and
1582 @option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the
1583 @option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for
1584 @option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and
1585 @option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these
1586 shorthand assumptions are not correct, you can use the explicit
1587 include and lib options directly. You might also need to ensure the
1588 shared libraries can be found by the dynamic linker when building and
1589 using GCC, for example by setting the runtime shared library path
1590 variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems).
1592 These flags are applicable to the host platform only. When building
1593 a cross compiler, they will not be used to configure target libraries.
1595 @item --with-ppl=@var{pathname}
1596 @itemx --with-ppl-include=@var{pathname}
1597 @itemx --with-ppl-lib=@var{pathname}
1598 @itemx --with-cloog=@var{pathname}
1599 @itemx --with-cloog-include=@var{pathname}
1600 @itemx --with-cloog-lib=@var{pathname}
1601 If you do not have PPL (the Parma Polyhedra Library) and the CLooG
1602 libraries installed in a standard location and you want to build GCC,
1603 you can explicitly specify the directory where they are installed
1604 (@samp{--with-ppl=@/@var{pplinstalldir}},
1605 @samp{--with-cloog=@/@var{clooginstalldir}}). The
1606 @option{--with-ppl=@/@var{pplinstalldir}} option is shorthand for
1607 @option{--with-ppl-lib=@/@var{pplinstalldir}/lib} and
1608 @option{--with-ppl-include=@/@var{pplinstalldir}/include}. Likewise the
1609 @option{--with-cloog=@/@var{clooginstalldir}} option is shorthand for
1610 @option{--with-cloog-lib=@/@var{clooginstalldir}/lib} and
1611 @option{--with-cloog-include=@/@var{clooginstalldir}/include}. If these
1612 shorthand assumptions are not correct, you can use the explicit
1613 include and lib options directly.
1615 These flags are applicable to the host platform only. When building
1616 a cross compiler, they will not be used to configure target libraries.
1618 @item --with-host-libstdcxx=@var{linker-args}
1619 If you are linking with a static copy of PPL, you can use this option
1620 to specify how the linker should find the standard C++ library used
1621 internally by PPL. Typical values of @var{linker-args} might be
1622 @samp{-lstdc++} or @samp{-Wl,-Bstatic,-lstdc++,-Bdynamic -lm}. If you are
1623 linking with a shared copy of PPL, you probably do not need this
1624 option; shared library dependencies will cause the linker to search
1625 for the standard C++ library automatically.
1627 @item --with-stage1-ldflags=@var{flags}
1628 This option may be used to set linker flags to be used when linking
1629 stage 1 of GCC. These are also used when linking GCC if configured with
1630 @option{--disable-bootstrap}. By default no special flags are used.
1632 @item --with-stage1-libs=@var{libs}
1633 This option may be used to set libraries to be used when linking stage 1
1634 of GCC. These are also used when linking GCC if configured with
1635 @option{--disable-bootstrap}. The default is the argument to
1636 @option{--with-host-libstdcxx}, if specified.
1638 @item --with-boot-ldflags=@var{flags}
1639 This option may be used to set linker flags to be used when linking
1640 stage 2 and later when bootstrapping GCC. If neither --with-boot-libs
1641 nor --with-host-libstdcxx is set to a value, then the default is
1642 @samp{-static-libstdc++ -static-libgcc}.
1644 @item --with-boot-libs=@var{libs}
1645 This option may be used to set libraries to be used when linking stage 2
1646 and later when bootstrapping GCC. The default is the argument to
1647 @option{--with-host-libstdcxx}, if specified.
1649 @item --with-debug-prefix-map=@var{map}
1650 Convert source directory names using @option{-fdebug-prefix-map} when
1651 building runtime libraries. @samp{@var{map}} is a space-separated
1652 list of maps of the form @samp{@var{old}=@var{new}}.
1654 @item --enable-linker-build-id
1655 Tells GCC to pass @option{--build-id} option to the linker for all final
1656 links (links performed without the @option{-r} or @option{--relocatable}
1657 option), if the linker supports it. If you specify
1658 @option{--enable-linker-build-id}, but your linker does not
1659 support @option{--build-id} option, a warning is issued and the
1660 @option{--enable-linker-build-id} option is ignored. The default is off.
1662 @item --enable-gnu-unique-object
1663 @itemx --disable-gnu-unique-object
1664 Tells GCC to use the gnu_unique_object relocation for C++ template
1665 static data members and inline function local statics. Enabled by
1666 default for a native toolchain with an assembler that accepts it and
1667 GLIBC 2.11 or above, otherwise disabled.
1670 @itemx --disable-lto
1671 Enable support for link-time optimization (LTO). This is enabled by
1672 default, and may be disabled using @option{--disable-lto}.
1674 @item --with-plugin-ld=@var{pathname}
1675 Enable an alternate linker to be used at link-time optimization (LTO)
1676 link time when @option{-fuse-linker-plugin} is enabled.
1677 This linker should have plugin support such as gold starting with
1678 version 2.20 or GNU ld starting with version 2.21.
1679 See @option{-fuse-linker-plugin} for details.
1682 @subheading Cross-Compiler-Specific Options
1683 The following options only apply to building cross compilers.
1686 @item --with-sysroot
1687 @itemx --with-sysroot=@var{dir}
1688 Tells GCC to consider @var{dir} as the root of a tree that contains
1689 (a subset of) the root filesystem of the target operating system.
1690 Target system headers, libraries and run-time object files will be
1691 searched in there. More specifically, this acts as if
1692 @option{--sysroot=@var{dir}} was added to the default options of the built
1693 compiler. The specified directory is not copied into the
1694 install tree, unlike the options @option{--with-headers} and
1695 @option{--with-libs} that this option obsoletes. The default value,
1696 in case @option{--with-sysroot} is not given an argument, is
1697 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1698 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1699 the GCC binaries if the installation tree is moved.
1701 This option affects the system root for the compiler used to build
1702 target libraries (which runs on the build system) and the compiler newly
1703 installed with @code{make install}; it does not affect the compiler which is
1704 used to build GCC itself.
1706 @item --with-build-sysroot
1707 @itemx --with-build-sysroot=@var{dir}
1708 Tells GCC to consider @var{dir} as the system root (see
1709 @option{--with-sysroot}) while building target libraries, instead of
1710 the directory specified with @option{--with-sysroot}. This option is
1711 only useful when you are already using @option{--with-sysroot}. You
1712 can use @option{--with-build-sysroot} when you are configuring with
1713 @option{--prefix} set to a directory that is different from the one in
1714 which you are installing GCC and your target libraries.
1716 This option affects the system root for the compiler used to build
1717 target libraries (which runs on the build system); it does not affect
1718 the compiler which is used to build GCC itself.
1720 @item --with-headers
1721 @itemx --with-headers=@var{dir}
1722 Deprecated in favor of @option{--with-sysroot}.
1723 Specifies that target headers are available when building a cross compiler.
1724 The @var{dir} argument specifies a directory which has the target include
1725 files. These include files will be copied into the @file{gcc} install
1726 directory. @emph{This option with the @var{dir} argument is required} when
1727 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1728 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1729 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1730 will be run on these files to make them compatible with GCC@.
1732 @item --without-headers
1733 Tells GCC not use any target headers from a libc when building a cross
1734 compiler. When crossing to GNU/Linux, you need the headers so GCC
1735 can build the exception handling for libgcc.
1738 @itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}"
1739 Deprecated in favor of @option{--with-sysroot}.
1740 Specifies a list of directories which contain the target runtime
1741 libraries. These libraries will be copied into the @file{gcc} install
1742 directory. If the directory list is omitted, this option has no
1746 Specifies that @samp{newlib} is
1747 being used as the target C library. This causes @code{__eprintf} to be
1748 omitted from @file{libgcc.a} on the assumption that it will be provided by
1751 @item --with-build-time-tools=@var{dir}
1752 Specifies where to find the set of target tools (assembler, linker, etc.)
1753 that will be used while building GCC itself. This option can be useful
1754 if the directory layouts are different between the system you are building
1755 GCC on, and the system where you will deploy it.
1757 For example, on an @samp{ia64-hp-hpux} system, you may have the GNU
1758 assembler and linker in @file{/usr/bin}, and the native tools in a
1759 different path, and build a toolchain that expects to find the
1760 native tools in @file{/usr/bin}.
1762 When you use this option, you should ensure that @var{dir} includes
1763 @command{ar}, @command{as}, @command{ld}, @command{nm},
1764 @command{ranlib} and @command{strip} if necessary, and possibly
1765 @command{objdump}. Otherwise, GCC may use an inconsistent set of
1769 @subheading Java-Specific Options
1771 The following option applies to the build of the Java front end.
1774 @item --disable-libgcj
1775 Specify that the run-time libraries
1776 used by GCJ should not be built. This is useful in case you intend
1777 to use GCJ with some other run-time, or you're going to install it
1778 separately, or it just happens not to build on your particular
1779 machine. In general, if the Java front end is enabled, the GCJ
1780 libraries will be enabled too, unless they're known to not work on
1781 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1782 may need to port it; in this case, before modifying the top-level
1783 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1784 you may use @option{--enable-libgcj} to override the default.
1788 The following options apply to building @samp{libgcj}.
1790 @subsubheading General Options
1793 @item --enable-java-maintainer-mode
1794 By default the @samp{libjava} build will not attempt to compile the
1795 @file{.java} source files to @file{.class}. Instead, it will use the
1796 @file{.class} files from the source tree. If you use this option you
1797 must have executables named @command{ecj1} and @command{gjavah} in your path
1798 for use by the build. You must use this option if you intend to
1799 modify any @file{.java} files in @file{libjava}.
1801 @item --with-java-home=@var{dirname}
1802 This @samp{libjava} option overrides the default value of the
1803 @samp{java.home} system property. It is also used to set
1804 @samp{sun.boot.class.path} to @file{@var{dirname}/lib/rt.jar}. By
1805 default @samp{java.home} is set to @file{@var{prefix}} and
1806 @samp{sun.boot.class.path} to
1807 @file{@var{datadir}/java/libgcj-@var{version}.jar}.
1809 @item --with-ecj-jar=@var{filename}
1810 This option can be used to specify the location of an external jar
1811 file containing the Eclipse Java compiler. A specially modified
1812 version of this compiler is used by @command{gcj} to parse
1813 @file{.java} source files. If this option is given, the
1814 @samp{libjava} build will create and install an @file{ecj1} executable
1815 which uses this jar file at runtime.
1817 If this option is not given, but an @file{ecj.jar} file is found in
1818 the topmost source tree at configure time, then the @samp{libgcj}
1819 build will create and install @file{ecj1}, and will also install the
1820 discovered @file{ecj.jar} into a suitable place in the install tree.
1822 If @file{ecj1} is not installed, then the user will have to supply one
1823 on his path in order for @command{gcj} to properly parse @file{.java}
1824 source files. A suitable jar is available from
1825 @uref{ftp://sourceware.org/pub/java/}.
1827 @item --disable-getenv-properties
1828 Don't set system properties from @env{GCJ_PROPERTIES}.
1830 @item --enable-hash-synchronization
1831 Use a global hash table for monitor locks. Ordinarily,
1832 @samp{libgcj}'s @samp{configure} script automatically makes
1833 the correct choice for this option for your platform. Only use
1834 this if you know you need the library to be configured differently.
1836 @item --enable-interpreter
1837 Enable the Java interpreter. The interpreter is automatically
1838 enabled by default on all platforms that support it. This option
1839 is really only useful if you want to disable the interpreter
1840 (using @option{--disable-interpreter}).
1842 @item --disable-java-net
1843 Disable java.net. This disables the native part of java.net only,
1844 using non-functional stubs for native method implementations.
1846 @item --disable-jvmpi
1847 Disable JVMPI support.
1849 @item --disable-libgcj-bc
1850 Disable BC ABI compilation of certain parts of libgcj. By default,
1851 some portions of libgcj are compiled with @option{-findirect-dispatch}
1852 and @option{-fno-indirect-classes}, allowing them to be overridden at
1855 If @option{--disable-libgcj-bc} is specified, libgcj is built without
1856 these options. This allows the compile-time linker to resolve
1857 dependencies when statically linking to libgcj. However it makes it
1858 impossible to override the affected portions of libgcj at run-time.
1860 @item --enable-reduced-reflection
1861 Build most of libgcj with @option{-freduced-reflection}. This reduces
1862 the size of libgcj at the expense of not being able to do accurate
1863 reflection on the classes it contains. This option is safe if you
1864 know that code using libgcj will never use reflection on the standard
1865 runtime classes in libgcj (including using serialization, RMI or CORBA).
1868 Enable runtime eCos target support.
1870 @item --without-libffi
1871 Don't use @samp{libffi}. This will disable the interpreter and JNI
1872 support as well, as these require @samp{libffi} to work.
1874 @item --enable-libgcj-debug
1875 Enable runtime debugging code.
1877 @item --enable-libgcj-multifile
1878 If specified, causes all @file{.java} source files to be
1879 compiled into @file{.class} files in one invocation of
1880 @samp{gcj}. This can speed up build time, but is more
1881 resource-intensive. If this option is unspecified or
1882 disabled, @samp{gcj} is invoked once for each @file{.java}
1883 file to compile into a @file{.class} file.
1885 @item --with-libiconv-prefix=DIR
1886 Search for libiconv in @file{DIR/include} and @file{DIR/lib}.
1888 @item --enable-sjlj-exceptions
1889 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1890 @samp{configure} ordinarily picks the correct value based on the platform.
1891 Only use this option if you are sure you need a different setting.
1893 @item --with-system-zlib
1894 Use installed @samp{zlib} rather than that included with GCC@.
1896 @item --with-win32-nlsapi=ansi, unicows or unicode
1897 Indicates how MinGW @samp{libgcj} translates between UNICODE
1898 characters and the Win32 API@.
1900 @item --enable-java-home
1901 If enabled, this creates a JPackage compatible SDK environment during install.
1902 Note that if --enable-java-home is used, --with-arch-directory=ARCH must also
1905 @item --with-arch-directory=ARCH
1906 Specifies the name to use for the @file{jre/lib/ARCH} directory in the SDK
1907 environment created when --enable-java-home is passed. Typical names for this
1908 directory include i386, amd64, ia64, etc.
1910 @item --with-os-directory=DIR
1911 Specifies the OS directory for the SDK include directory. This is set to auto
1912 detect, and is typically 'linux'.
1914 @item --with-origin-name=NAME
1915 Specifies the JPackage origin name. This defaults to the 'gcj' in
1918 @item --with-arch-suffix=SUFFIX
1919 Specifies the suffix for the sdk directory. Defaults to the empty string.
1920 Examples include '.x86_64' in 'java-1.5.0-gcj-1.5.0.0.x86_64'.
1922 @item --with-jvm-root-dir=DIR
1923 Specifies where to install the SDK. Default is $(prefix)/lib/jvm.
1925 @item --with-jvm-jar-dir=DIR
1926 Specifies where to install jars. Default is $(prefix)/lib/jvm-exports.
1928 @item --with-python-dir=DIR
1929 Specifies where to install the Python modules used for aot-compile. DIR should
1930 not include the prefix used in installation. For example, if the Python modules
1931 are to be installed in /usr/lib/python2.5/site-packages, then
1932 --with-python-dir=/lib/python2.5/site-packages should be passed. If this is
1933 not specified, then the Python modules are installed in $(prefix)/share/python.
1935 @item --enable-aot-compile-rpm
1936 Adds aot-compile-rpm to the list of installed scripts.
1938 @item --enable-browser-plugin
1939 Build the gcjwebplugin web browser plugin.
1943 Use the single-byte @code{char} and the Win32 A functions natively,
1944 translating to and from UNICODE when using these functions. If
1945 unspecified, this is the default.
1948 Use the @code{WCHAR} and Win32 W functions natively. Adds
1949 @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}.
1950 @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines
1951 running built executables. @file{libunicows.a}, an open-source
1952 import library around Microsoft's @code{unicows.dll}, is obtained from
1953 @uref{http://libunicows.sourceforge.net/}, which also gives details
1954 on getting @file{unicows.dll} from Microsoft.
1957 Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not}
1958 add @code{-lunicows} to @file{libgcj.spec}. The built executables will
1959 only run on Microsoft Windows NT and above.
1963 @subsubheading AWT-Specific Options
1967 Use the X Window System.
1969 @item --enable-java-awt=PEER(S)
1970 Specifies the AWT peer library or libraries to build alongside
1971 @samp{libgcj}. If this option is unspecified or disabled, AWT
1972 will be non-functional. Current valid values are @option{gtk} and
1973 @option{xlib}. Multiple libraries should be separated by a
1974 comma (i.e.@: @option{--enable-java-awt=gtk,xlib}).
1976 @item --enable-gtk-cairo
1977 Build the cairo Graphics2D implementation on GTK@.
1979 @item --enable-java-gc=TYPE
1980 Choose garbage collector. Defaults to @option{boehm} if unspecified.
1982 @item --disable-gtktest
1983 Do not try to compile and run a test GTK+ program.
1985 @item --disable-glibtest
1986 Do not try to compile and run a test GLIB program.
1988 @item --with-libart-prefix=PFX
1989 Prefix where libart is installed (optional).
1991 @item --with-libart-exec-prefix=PFX
1992 Exec prefix where libart is installed (optional).
1994 @item --disable-libarttest
1995 Do not try to compile and run a test libart program.
1999 @subsubheading Overriding @command{configure} test results
2001 Sometimes, it might be necessary to override the result of some
2002 @command{configure} test, for example in order to ease porting to a new
2003 system or work around a bug in a test. The toplevel @command{configure}
2004 script provides three variables for this:
2008 @item build_configargs
2009 @cindex @code{build_configargs}
2010 The contents of this variable is passed to all build @command{configure}
2013 @item host_configargs
2014 @cindex @code{host_configargs}
2015 The contents of this variable is passed to all host @command{configure}
2018 @item target_configargs
2019 @cindex @code{target_configargs}
2020 The contents of this variable is passed to all target @command{configure}
2025 In order to avoid shell and @command{make} quoting issues for complex
2026 overrides, you can pass a setting for @env{CONFIG_SITE} and set
2027 variables in the site file.
2034 @uref{./index.html,,Return to the GCC Installation page}
2038 @c ***Building****************************************************************
2040 @comment node-name, next, previous, up
2041 @node Building, Testing, Configuration, Installing GCC
2047 @cindex Installing GCC: Building
2049 Now that GCC is configured, you are ready to build the compiler and
2052 Some commands executed when making the compiler may fail (return a
2053 nonzero status) and be ignored by @command{make}. These failures, which
2054 are often due to files that were not found, are expected, and can safely
2057 It is normal to have compiler warnings when compiling certain files.
2058 Unless you are a GCC developer, you can generally ignore these warnings
2059 unless they cause compilation to fail. Developers should attempt to fix
2060 any warnings encountered, however they can temporarily continue past
2061 warnings-as-errors by specifying the configure flag
2062 @option{--disable-werror}.
2064 On certain old systems, defining certain environment variables such as
2065 @env{CC} can interfere with the functioning of @command{make}.
2067 If you encounter seemingly strange errors when trying to build the
2068 compiler in a directory other than the source directory, it could be
2069 because you have previously configured the compiler in the source
2070 directory. Make sure you have done all the necessary preparations.
2072 If you build GCC on a BSD system using a directory stored in an old System
2073 V file system, problems may occur in running @command{fixincludes} if the
2074 System V file system doesn't support symbolic links. These problems
2075 result in a failure to fix the declaration of @code{size_t} in
2076 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
2077 that type mismatches occur, this could be the cause.
2079 The solution is not to use such a directory for building GCC@.
2081 Similarly, when building from SVN or snapshots, or if you modify
2082 @file{*.l} files, you need the Flex lexical analyzer generator
2083 installed. If you do not modify @file{*.l} files, releases contain
2084 the Flex-generated files and you do not need Flex installed to build
2085 them. There is still one Flex-based lexical analyzer (part of the
2086 build machinery, not of GCC itself) that is used even if you only
2087 build the C front end.
2089 When building from SVN or snapshots, or if you modify Texinfo
2090 documentation, you need version 4.7 or later of Texinfo installed if you
2091 want Info documentation to be regenerated. Releases contain Info
2092 documentation pre-built for the unmodified documentation in the release.
2094 @section Building a native compiler
2096 For a native build, the default configuration is to perform
2097 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
2098 This will build the entire GCC system and ensure that it compiles
2099 itself correctly. It can be disabled with the @option{--disable-bootstrap}
2100 parameter to @samp{configure}, but bootstrapping is suggested because
2101 the compiler will be tested more completely and could also have
2104 The bootstrapping process will complete the following steps:
2108 Build tools necessary to build the compiler.
2111 Perform a 3-stage bootstrap of the compiler. This includes building
2112 three times the target tools for use by the compiler such as binutils
2113 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
2114 individually linked or moved into the top level GCC source tree before
2118 Perform a comparison test of the stage2 and stage3 compilers.
2121 Build runtime libraries using the stage3 compiler from the previous step.
2125 If you are short on disk space you might consider @samp{make
2126 bootstrap-lean} instead. The sequence of compilation is the
2127 same described above, but object files from the stage1 and
2128 stage2 of the 3-stage bootstrap of the compiler are deleted as
2129 soon as they are no longer needed.
2131 If you wish to use non-default GCC flags when compiling the stage2
2132 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
2133 doing @samp{make}. For example, if you want to save additional space
2134 during the bootstrap and in the final installation as well, you can
2135 build the compiler binaries without debugging information as in the
2136 following example. This will save roughly 40% of disk space both for
2137 the bootstrap and the final installation. (Libraries will still contain
2138 debugging information.)
2141 make BOOT_CFLAGS='-O' bootstrap
2144 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
2145 are less well tested here than the default of @samp{-g -O2}, but should
2146 still work. In a few cases, you may find that you need to specify special
2147 flags such as @option{-msoft-float} here to complete the bootstrap; or,
2148 if the native compiler miscompiles the stage1 compiler, you may need
2149 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
2150 of the stage1 compiler that were miscompiled, or by using @samp{make
2151 bootstrap4} to increase the number of stages of bootstrap.
2153 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
2154 Since these are always compiled with the compiler currently being
2155 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
2156 compilation flags, as for non-bootstrapped target libraries.
2157 Again, if the native compiler miscompiles the stage1 compiler, you may
2158 need to work around this by avoiding non-working parts of the stage1
2159 compiler. Use @code{STAGE1_TFLAGS} to this end.
2161 If you used the flag @option{--enable-languages=@dots{}} to restrict
2162 the compilers to be built, only those you've actually enabled will be
2163 built. This will of course only build those runtime libraries, for
2164 which the particular compiler has been built. Please note,
2165 that re-defining @env{LANGUAGES} when calling @samp{make}
2166 @strong{does not} work anymore!
2168 If the comparison of stage2 and stage3 fails, this normally indicates
2169 that the stage2 compiler has compiled GCC incorrectly, and is therefore
2170 a potentially serious bug which you should investigate and report. (On
2171 a few systems, meaningful comparison of object files is impossible; they
2172 always appear ``different''. If you encounter this problem, you will
2173 need to disable comparison in the @file{Makefile}.)
2175 If you do not want to bootstrap your compiler, you can configure with
2176 @option{--disable-bootstrap}. In particular cases, you may want to
2177 bootstrap your compiler even if the target system is not the same as
2178 the one you are building on: for example, you could build a
2179 @code{powerpc-unknown-linux-gnu} toolchain on a
2180 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
2181 @option{--enable-bootstrap} to the configure script.
2183 @code{BUILD_CONFIG} can be used to bring in additional customization
2184 to the build. It can be set to a whitespace-separated list of names.
2185 For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will
2186 be included by the top-level @file{Makefile}, bringing in any settings
2187 it contains. The default @code{BUILD_CONFIG} can be set using the
2188 configure option @option{--with-build-config=@code{NAME}...}. Some
2189 examples of supported build configurations are:
2192 @item @samp{bootstrap-O1}
2193 Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds
2194 @option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to
2195 @samp{BOOT_CFLAGS='-g -O1'}.
2197 @item @samp{bootstrap-O3}
2198 Analogous to @code{bootstrap-O1}.
2200 @item @samp{bootstrap-lto}
2201 Enables Link-Time Optimization for host tools during bootstrapping.
2202 @samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding
2203 @option{-flto} to @samp{BOOT_CFLAGS}.
2205 @item @samp{bootstrap-debug}
2206 Verifies that the compiler generates the same executable code, whether
2207 or not it is asked to emit debug information. To this end, this
2208 option builds stage2 host programs without debug information, and uses
2209 @file{contrib/compare-debug} to compare them with the stripped stage3
2210 object files. If @code{BOOT_CFLAGS} is overridden so as to not enable
2211 debug information, stage2 will have it, and stage3 won't. This option
2212 is enabled by default when GCC bootstrapping is enabled, if
2213 @code{strip} can turn object files compiled with and without debug
2214 info into identical object files. In addition to better test
2215 coverage, this option makes default bootstraps faster and leaner.
2217 @item @samp{bootstrap-debug-big}
2218 Rather than comparing stripped object files, as in
2219 @code{bootstrap-debug}, this option saves internal compiler dumps
2220 during stage2 and stage3 and compares them as well, which helps catch
2221 additional potential problems, but at a great cost in terms of disk
2222 space. It can be specified in addition to @samp{bootstrap-debug}.
2224 @item @samp{bootstrap-debug-lean}
2225 This option saves disk space compared with @code{bootstrap-debug-big},
2226 but at the expense of some recompilation. Instead of saving the dumps
2227 of stage2 and stage3 until the final compare, it uses
2228 @option{-fcompare-debug} to generate, compare and remove the dumps
2229 during stage3, repeating the compilation that already took place in
2230 stage2, whose dumps were not saved.
2232 @item @samp{bootstrap-debug-lib}
2233 This option tests executable code invariance over debug information
2234 generation on target libraries, just like @code{bootstrap-debug-lean}
2235 tests it on host programs. It builds stage3 libraries with
2236 @option{-fcompare-debug}, and it can be used along with any of the
2237 @code{bootstrap-debug} options above.
2239 There aren't @code{-lean} or @code{-big} counterparts to this option
2240 because most libraries are only build in stage3, so bootstrap compares
2241 would not get significant coverage. Moreover, the few libraries built
2242 in stage2 are used in stage3 host programs, so we wouldn't want to
2243 compile stage2 libraries with different options for comparison purposes.
2245 @item @samp{bootstrap-debug-ckovw}
2246 Arranges for error messages to be issued if the compiler built on any
2247 stage is run without the option @option{-fcompare-debug}. This is
2248 useful to verify the full @option{-fcompare-debug} testing coverage. It
2249 must be used along with @code{bootstrap-debug-lean} and
2250 @code{bootstrap-debug-lib}.
2252 @item @samp{bootstrap-time}
2253 Arranges for the run time of each program started by the GCC driver,
2254 built in any stage, to be logged to @file{time.log}, in the top level of
2259 @section Building a cross compiler
2261 When building a cross compiler, it is not generally possible to do a
2262 3-stage bootstrap of the compiler. This makes for an interesting problem
2263 as parts of GCC can only be built with GCC@.
2265 To build a cross compiler, we recommend first building and installing a
2266 native compiler. You can then use the native GCC compiler to build the
2267 cross compiler. The installed native compiler needs to be GCC version
2270 If the cross compiler is to be built with support for the Java
2271 programming language and the ability to compile .java source files is
2272 desired, the installed native compiler used to build the cross
2273 compiler needs to be the same GCC version as the cross compiler. In
2274 addition the cross compiler needs to be configured with
2275 @option{--with-ecj-jar=@dots{}}.
2277 Assuming you have already installed a native copy of GCC and configured
2278 your cross compiler, issue the command @command{make}, which performs the
2283 Build host tools necessary to build the compiler.
2286 Build target tools for use by the compiler such as binutils (bfd,
2287 binutils, gas, gprof, ld, and opcodes)
2288 if they have been individually linked or moved into the top level GCC source
2289 tree before configuring.
2292 Build the compiler (single stage only).
2295 Build runtime libraries using the compiler from the previous step.
2298 Note that if an error occurs in any step the make process will exit.
2300 If you are not building GNU binutils in the same source tree as GCC,
2301 you will need a cross-assembler and cross-linker installed before
2302 configuring GCC@. Put them in the directory
2303 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
2304 you should put in this directory:
2308 This should be the cross-assembler.
2311 This should be the cross-linker.
2314 This should be the cross-archiver: a program which can manipulate
2315 archive files (linker libraries) in the target machine's format.
2318 This should be a program to construct a symbol table in an archive file.
2321 The installation of GCC will find these programs in that directory,
2322 and copy or link them to the proper place to for the cross-compiler to
2323 find them when run later.
2325 The easiest way to provide these files is to build the Binutils package.
2326 Configure it with the same @option{--host} and @option{--target}
2327 options that you use for configuring GCC, then build and install
2328 them. They install their executables automatically into the proper
2329 directory. Alas, they do not support all the targets that GCC
2332 If you are not building a C library in the same source tree as GCC,
2333 you should also provide the target libraries and headers before
2334 configuring GCC, specifying the directories with
2335 @option{--with-sysroot} or @option{--with-headers} and
2336 @option{--with-libs}. Many targets also require ``start files'' such
2337 as @file{crt0.o} and
2338 @file{crtn.o} which are linked into each executable. There may be several
2339 alternatives for @file{crt0.o}, for use with profiling or other
2340 compilation options. Check your target's definition of
2341 @code{STARTFILE_SPEC} to find out what start files it uses.
2343 @section Building in parallel
2345 GNU Make 3.80 and above, which is necessary to build GCC, support
2346 building in parallel. To activate this, you can use @samp{make -j 2}
2347 instead of @samp{make}. You can also specify a bigger number, and
2348 in most cases using a value greater than the number of processors in
2349 your machine will result in fewer and shorter I/O latency hits, thus
2350 improving overall throughput; this is especially true for slow drives
2351 and network filesystems.
2353 @section Building the Ada compiler
2355 In order to build GNAT, the Ada compiler, you need a working GNAT
2356 compiler (GCC version 4.0 or later).
2357 This includes GNAT tools such as @command{gnatmake} and
2358 @command{gnatlink}, since the Ada front end is written in Ada and
2359 uses some GNAT-specific extensions.
2361 In order to build a cross compiler, it is suggested to install
2362 the new compiler as native first, and then use it to build the cross
2365 @command{configure} does not test whether the GNAT installation works
2366 and has a sufficiently recent version; if too old a GNAT version is
2367 installed, the build will fail unless @option{--enable-languages} is
2368 used to disable building the Ada front end.
2370 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
2371 must not be set when building the Ada compiler, the Ada tools, or the
2372 Ada runtime libraries. You can check that your build environment is clean
2373 by verifying that @samp{gnatls -v} lists only one explicit path in each
2376 @section Building with profile feedback
2378 It is possible to use profile feedback to optimize the compiler itself. This
2379 should result in a faster compiler binary. Experiments done on x86 using gcc
2380 3.3 showed approximately 7 percent speedup on compiling C programs. To
2381 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
2383 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
2384 compiler. This compiler is used to build a @code{stageprofile} compiler
2385 instrumented to collect execution counts of instruction and branch
2386 probabilities. Then runtime libraries are compiled with profile collected.
2387 Finally a @code{stagefeedback} compiler is built using the information collected.
2389 Unlike standard bootstrap, several additional restrictions apply. The
2390 compiler used to build @code{stage1} needs to support a 64-bit integral type.
2391 It is recommended to only use GCC for this. Also parallel make is currently
2392 not supported since collisions in profile collecting may occur.
2399 @uref{./index.html,,Return to the GCC Installation page}
2403 @c ***Testing*****************************************************************
2405 @comment node-name, next, previous, up
2406 @node Testing, Final install, Building, Installing GCC
2410 @chapter Installing GCC: Testing
2413 @cindex Installing GCC: Testing
2416 Before you install GCC, we encourage you to run the testsuites and to
2417 compare your results with results from a similar configuration that have
2418 been submitted to the
2419 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
2420 Some of these archived results are linked from the build status lists
2421 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
2422 reports a successful build runs the testsuites and submits the results.
2423 This step is optional and may require you to download additional software,
2424 but it can give you confidence in your new GCC installation or point out
2425 problems before you install and start using your new GCC@.
2427 First, you must have @uref{download.html,,downloaded the testsuites}.
2428 These are part of the full distribution, but if you downloaded the
2429 ``core'' compiler plus any front ends, you must download the testsuites
2432 Second, you must have the testing tools installed. This includes
2433 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
2434 the DejaGnu site has links to these.
2436 If the directories where @command{runtest} and @command{expect} were
2437 installed are not in the @env{PATH}, you may need to set the following
2438 environment variables appropriately, as in the following example (which
2439 assumes that DejaGnu has been installed under @file{/usr/local}):
2442 TCL_LIBRARY = /usr/local/share/tcl8.0
2443 DEJAGNULIBS = /usr/local/share/dejagnu
2446 (On systems such as Cygwin, these paths are required to be actual
2447 paths, not mounts or links; presumably this is due to some lack of
2448 portability in the DejaGnu code.)
2451 Finally, you can run the testsuite (which may take a long time):
2453 cd @var{objdir}; make -k check
2456 This will test various components of GCC, such as compiler
2457 front ends and runtime libraries. While running the testsuite, DejaGnu
2458 might emit some harmless messages resembling
2459 @samp{WARNING: Couldn't find the global config file.} or
2460 @samp{WARNING: Couldn't find tool init file} that can be ignored.
2462 If you are testing a cross-compiler, you may want to run the testsuite
2463 on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}.
2465 @section How can you run the testsuite on selected tests?
2467 In order to run sets of tests selectively, there are targets
2468 @samp{make check-gcc} and @samp{make check-g++}
2469 in the @file{gcc} subdirectory of the object directory. You can also
2470 just run @samp{make check} in a subdirectory of the object directory.
2473 A more selective way to just run all @command{gcc} execute tests in the
2477 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
2480 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
2481 the testsuite with filenames matching @samp{9805*}, you would use
2484 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
2487 The @file{*.exp} files are located in the testsuite directories of the GCC
2488 source, the most important ones being @file{compile.exp},
2489 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
2490 To get a list of the possible @file{*.exp} files, pipe the
2491 output of @samp{make check} into a file and look at the
2492 @samp{Running @dots{} .exp} lines.
2494 @section Passing options and running multiple testsuites
2496 You can pass multiple options to the testsuite using the
2497 @samp{--target_board} option of DejaGNU, either passed as part of
2498 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
2499 work outside the makefiles. For example,
2502 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
2505 will run the standard @command{g++} testsuites (``unix'' is the target name
2506 for a standard native testsuite situation), passing
2507 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
2508 slashes separate options.
2510 You can run the testsuites multiple times using combinations of options
2511 with a syntax similar to the brace expansion of popular shells:
2514 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
2517 (Note the empty option caused by the trailing comma in the final group.)
2518 The following will run each testsuite eight times using the @samp{arm-sim}
2519 target, as if you had specified all possible combinations yourself:
2522 --target_board=arm-sim/-mhard-float/-O1
2523 --target_board=arm-sim/-mhard-float/-O2
2524 --target_board=arm-sim/-mhard-float/-O3
2525 --target_board=arm-sim/-mhard-float
2526 --target_board=arm-sim/-msoft-float/-O1
2527 --target_board=arm-sim/-msoft-float/-O2
2528 --target_board=arm-sim/-msoft-float/-O3
2529 --target_board=arm-sim/-msoft-float
2532 They can be combined as many times as you wish, in arbitrary ways. This
2536 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
2539 will generate four combinations, all involving @samp{-Wextra}.
2541 The disadvantage to this method is that the testsuites are run in serial,
2542 which is a waste on multiprocessor systems. For users with GNU Make and
2543 a shell which performs brace expansion, you can run the testsuites in
2544 parallel by having the shell perform the combinations and @command{make}
2545 do the parallel runs. Instead of using @samp{--target_board}, use a
2546 special makefile target:
2549 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
2555 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
2558 will run three concurrent ``make-gcc'' testsuites, eventually testing all
2559 ten combinations as described above. Note that this is currently only
2560 supported in the @file{gcc} subdirectory. (To see how this works, try
2561 typing @command{echo} before the example given here.)
2564 @section Additional testing for Java Class Libraries
2566 The Java runtime tests can be executed via @samp{make check}
2567 in the @file{@var{target}/libjava/testsuite} directory in
2570 The @uref{http://sourceware.org/mauve/,,Mauve Project} provides
2571 a suite of tests for the Java Class Libraries. This suite can be run
2572 as part of libgcj testing by placing the Mauve tree within the libjava
2573 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
2574 specifying the location of that tree when invoking @samp{make}, as in
2575 @samp{make MAUVEDIR=~/mauve check}.
2577 @section How to interpret test results
2579 The result of running the testsuite are various @file{*.sum} and @file{*.log}
2580 files in the testsuite subdirectories. The @file{*.log} files contain a
2581 detailed log of the compiler invocations and the corresponding
2582 results, the @file{*.sum} files summarize the results. These summaries
2583 contain status codes for all tests:
2587 PASS: the test passed as expected
2589 XPASS: the test unexpectedly passed
2591 FAIL: the test unexpectedly failed
2593 XFAIL: the test failed as expected
2595 UNSUPPORTED: the test is not supported on this platform
2597 ERROR: the testsuite detected an error
2599 WARNING: the testsuite detected a possible problem
2602 It is normal for some tests to report unexpected failures. At the
2603 current time the testing harness does not allow fine grained control
2604 over whether or not a test is expected to fail. This problem should
2605 be fixed in future releases.
2608 @section Submitting test results
2610 If you want to report the results to the GCC project, use the
2611 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
2614 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
2615 -m gcc-testresults@@gcc.gnu.org |sh
2618 This script uses the @command{Mail} program to send the results, so
2619 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
2620 prepended to the testsuite summary and should contain any special
2621 remarks you have on your results or your build environment. Please
2622 do not edit the testsuite result block or the subject line, as these
2623 messages may be automatically processed.
2630 @uref{./index.html,,Return to the GCC Installation page}
2634 @c ***Final install***********************************************************
2636 @comment node-name, next, previous, up
2637 @node Final install, , Testing, Installing GCC
2639 @ifset finalinstallhtml
2641 @chapter Installing GCC: Final installation
2644 Now that GCC has been built (and optionally tested), you can install it with
2646 cd @var{objdir} && make install
2649 We strongly recommend to install into a target directory where there is
2650 no previous version of GCC present. Also, the GNAT runtime should not
2651 be stripped, as this would break certain features of the debugger that
2652 depend on this debugging information (catching Ada exceptions for
2655 That step completes the installation of GCC; user level binaries can
2656 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
2657 you specified with the @option{--prefix} to configure (or
2658 @file{/usr/local} by default). (If you specified @option{--bindir},
2659 that directory will be used instead; otherwise, if you specified
2660 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
2661 Headers for the C++ and Java libraries are installed in
2662 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
2663 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
2664 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
2665 in info format in @file{@var{infodir}} (normally
2666 @file{@var{prefix}/info}).
2668 When installing cross-compilers, GCC's executables
2669 are not only installed into @file{@var{bindir}}, that
2670 is, @file{@var{exec-prefix}/bin}, but additionally into
2671 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
2672 exists. Typically, such @dfn{tooldirs} hold target-specific
2673 binutils, including assembler and linker.
2675 Installation into a temporary staging area or into a @command{chroot}
2676 jail can be achieved with the command
2679 make DESTDIR=@var{path-to-rootdir} install
2683 where @var{path-to-rootdir} is the absolute path of
2684 a directory relative to which all installation paths will be
2685 interpreted. Note that the directory specified by @code{DESTDIR}
2686 need not exist yet; it will be created if necessary.
2688 There is a subtle point with tooldirs and @code{DESTDIR}:
2689 If you relocate a cross-compiler installation with
2690 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
2691 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
2692 be filled with duplicated GCC executables only if it already exists,
2693 it will not be created otherwise. This is regarded as a feature,
2694 not as a bug, because it gives slightly more control to the packagers
2695 using the @code{DESTDIR} feature.
2697 You can install stripped programs and libraries with
2703 If you are bootstrapping a released version of GCC then please
2704 quickly review the build status page for your release, available from
2705 @uref{http://gcc.gnu.org/buildstat.html}.
2706 If your system is not listed for the version of GCC that you built,
2708 @email{gcc@@gcc.gnu.org} indicating
2709 that you successfully built and installed GCC@.
2710 Include the following information:
2714 Output from running @file{@var{srcdir}/config.guess}. Do not send
2715 that file itself, just the one-line output from running it.
2718 The output of @samp{gcc -v} for your newly installed @command{gcc}.
2719 This tells us which version of GCC you built and the options you passed to
2723 Whether you enabled all languages or a subset of them. If you used a
2724 full distribution then this information is part of the configure
2725 options in the output of @samp{gcc -v}, but if you downloaded the
2726 ``core'' compiler plus additional front ends then it isn't apparent
2727 which ones you built unless you tell us about it.
2730 If the build was for GNU/Linux, also include:
2733 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
2734 this information should be available from @file{/etc/issue}.
2737 The version of the Linux kernel, available from @samp{uname --version}
2741 The version of glibc you used; for RPM-based systems like Red Hat,
2742 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
2743 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
2745 For other systems, you can include similar information if you think it is
2749 Any other information that you think would be useful to people building
2750 GCC on the same configuration. The new entry in the build status list
2751 will include a link to the archived copy of your message.
2754 We'd also like to know if the
2756 @ref{Specific, host/target specific installation notes}
2759 @uref{specific.html,,host/target specific installation notes}
2761 didn't include your host/target information or if that information is
2762 incomplete or out of date. Send a note to
2763 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
2765 If you find a bug, please report it following the
2766 @uref{../bugs/,,bug reporting guidelines}.
2768 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
2769 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
2770 and @TeX{} installed. This creates a number of @file{.dvi} files in
2771 subdirectories of @file{@var{objdir}}; these may be converted for
2772 printing with programs such as @command{dvips}. Alternately, by using
2773 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
2774 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
2775 is included with Texinfo version 4.8 and later. You can also
2776 @uref{http://shop.fsf.org/,,buy printed manuals from the
2777 Free Software Foundation}, though such manuals may not be for the most
2778 recent version of GCC@.
2780 If you would like to generate online HTML documentation, do @samp{cd
2781 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
2782 @file{@var{objdir}/gcc/HTML}.
2789 @uref{./index.html,,Return to the GCC Installation page}
2793 @c ***Binaries****************************************************************
2795 @comment node-name, next, previous, up
2796 @node Binaries, Specific, Installing GCC, Top
2800 @chapter Installing GCC: Binaries
2803 @cindex Installing GCC: Binaries
2805 We are often asked about pre-compiled versions of GCC@. While we cannot
2806 provide these for all platforms, below you'll find links to binaries for
2807 various platforms where creating them by yourself is not easy due to various
2810 Please note that we did not create these binaries, nor do we
2811 support them. If you have any problems installing them, please
2812 contact their makers.
2819 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
2822 @uref{http://pware.hvcc.edu,,Hudson Valley Community College Open Source Software for IBM System p};
2825 @uref{http://www.perzl.org/aix/,,AIX 5L and 6 Open Source Packages}.
2829 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
2832 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
2833 Development Tools for the Renesas H8/300[HS] Series}.
2839 @uref{http://hpux.connect.org.uk/,,HP-UX Porting Center};
2842 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
2846 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
2847 OpenServer/Unixware}.
2850 Solaris 2 (SPARC, Intel):
2853 @uref{http://www.sunfreeware.com/,,Sunfreeware}
2856 @uref{http://www.blastwave.org/,,Blastwave}
2859 @uref{http://www.opencsw.org/,,OpenCSW}
2862 @uref{http://jupiterrise.com/tgcware/,,TGCware}
2869 @uref{http://nekochan.net/,,Nekoware}
2872 @uref{http://jupiterrise.com/tgcware/,,TGCware}
2879 The @uref{http://sourceware.org/cygwin/,,Cygwin} project;
2881 The @uref{http://www.mingw.org/,,MinGW} project.
2885 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
2886 Written Word} offers binaries for
2887 AIX 4.3.3, 5.1 and 5.2,
2889 Tru64 UNIX 4.0D and 5.1,
2891 HP-UX 10.20, 11.00, and 11.11, and
2892 Solaris/SPARC 2.5.1, 2.6, 7, 8, 9 and 10.
2895 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
2896 number of platforms.
2899 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
2900 links to GNU Fortran binaries for several platforms.
2908 @uref{./index.html,,Return to the GCC Installation page}
2912 @c ***Specific****************************************************************
2914 @comment node-name, next, previous, up
2915 @node Specific, Old, Binaries, Top
2919 @chapter Host/target specific installation notes for GCC
2922 @cindex Specific installation notes
2923 @cindex Target specific installation
2924 @cindex Host specific installation
2925 @cindex Target specific installation notes
2927 Please read this document carefully @emph{before} installing the
2928 GNU Compiler Collection on your machine.
2930 Note that this list of install notes is @emph{not} a list of supported
2931 hosts or targets. Not all supported hosts and targets are listed
2932 here, only the ones that require host-specific or target-specific
2938 @uref{#alpha-x-x,,alpha*-*-*}
2940 @uref{#alpha-dec-osf51,,alpha*-dec-osf5.1}
2942 @uref{#arm-x-elf,,arm-*-elf}
2946 @uref{#bfin,,Blackfin}
2950 @uref{#x-x-freebsd,,*-*-freebsd*}
2952 @uref{#h8300-hms,,h8300-hms}
2954 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
2956 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
2958 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
2960 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
2962 @uref{#ix86-x-linux,,i?86-*-linux*}
2964 @uref{#ix86-x-solaris289,,i?86-*-solaris2.[89]}
2966 @uref{#ix86-x-solaris210,,i?86-*-solaris2.10}
2968 @uref{#ia64-x-linux,,ia64-*-linux}
2970 @uref{#ia64-x-hpux,,ia64-*-hpux*}
2972 @uref{#x-ibm-aix,,*-ibm-aix*}
2974 @uref{#iq2000-x-elf,,iq2000-*-elf}
2976 @uref{#lm32-x-elf,,lm32-*-elf}
2978 @uref{#lm32-x-uclinux,,lm32-*-uclinux}
2980 @uref{#m32c-x-elf,,m32c-*-elf}
2982 @uref{#m32r-x-elf,,m32r-*-elf}
2984 @uref{#m68k-x-x,,m68k-*-*}
2986 @uref{#m68k-uclinux,,m68k-uclinux}
2988 @uref{#mep-x-elf,,mep-*-elf}
2990 @uref{#microblaze-x-elf,,microblaze-*-elf}
2992 @uref{#mips-x-x,,mips-*-*}
2994 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2996 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2998 @uref{#powerpc-x-x,,powerpc*-*-*}
3000 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
3002 @uref{#powerpc-x-elf,,powerpc-*-elf}
3004 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
3006 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
3008 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
3010 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
3012 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
3014 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
3016 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
3018 @uref{#s390-x-linux,,s390-*-linux*}
3020 @uref{#s390x-x-linux,,s390x-*-linux*}
3022 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
3024 @uref{#x-x-solaris2,,*-*-solaris2*}
3026 @uref{#sparc-x-x,,sparc*-*-*}
3028 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
3030 @uref{#sparc-sun-solaris210,,sparc-sun-solaris2.10}
3032 @uref{#sparc-x-linux,,sparc-*-linux*}
3034 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
3036 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
3038 @uref{#x-x-vxworks,,*-*-vxworks*}
3040 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
3042 @uref{#xtensa-x-elf,,xtensa*-*-elf}
3044 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
3046 @uref{#windows,,Microsoft Windows}
3048 @uref{#x-x-cygwin,,*-*-cygwin}
3050 @uref{#x-x-interix,,*-*-interix}
3052 @uref{#x-x-mingw32,,*-*-mingw32}
3056 @uref{#older,,Older systems}
3061 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
3067 <!-- -------- host/target specific issues start here ---------------- -->
3070 @heading @anchor{alpha-x-x}alpha*-*-*
3072 This section contains general configuration information for all
3073 alpha-based platforms using ELF (in particular, ignore this section for
3074 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
3075 section, please read all other sections that match your target.
3077 We require binutils 2.11.2 or newer.
3078 Previous binutils releases had a number of problems with DWARF 2
3079 debugging information, not the least of which is incorrect linking of
3085 @heading @anchor{alpha-dec-osf51}alpha*-dec-osf5.1
3086 Systems using processors that implement the DEC Alpha architecture and
3087 are running the DEC/Compaq/HP Unix (DEC OSF/1, Digital UNIX, or Compaq/HP
3088 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
3090 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
3091 supported. (These are the versions which identify themselves as DEC
3092 OSF/1.) As of GCC 4.6, support for Tru64 UNIX V4.0 and V5.0 has been
3095 On Tru64 UNIX, virtual memory exhausted bootstrap failures
3096 may be fixed by reconfiguring Kernel Virtual Memory and Swap parameters
3097 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
3098 or applying the patch in
3099 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}. Depending on
3100 the OS version used, you need a data segment size between 512 MB and
3101 1 GB, so simply use @command{ulimit -Sd unlimited}.
3103 As of GNU binutils 2.21, neither GNU @command{as} nor GNU @command{ld}
3104 are supported on Tru64 UNIX, so you must not configure GCC with
3105 @option{--with-gnu-as} or @option{--with-gnu-ld}.
3107 GCC writes a @samp{.verstamp} directive to the assembler output file
3108 unless it is built as a cross-compiler. It gets the version to use from
3109 the system header file @file{/usr/include/stamp.h}. If you install a
3110 new version of Tru64 UNIX, you should rebuild GCC to pick up the new version
3113 GCC now supports both the native (ECOFF) debugging format used by DBX
3114 and GDB and an encapsulated STABS format for use only with GDB@. See the
3115 discussion of the @option{--with-stabs} option of @file{configure} above
3116 for more information on these formats and how to select them.
3117 @c FIXME: does this work at all? If so, perhaps make default.
3119 There is a bug in DEC's assembler that produces incorrect line numbers
3120 for ECOFF format when the @samp{.align} directive is used. To work
3121 around this problem, GCC will not emit such alignment directives
3122 while writing ECOFF format debugging information even if optimization is
3123 being performed. Unfortunately, this has the very undesirable
3124 side-effect that code addresses when @option{-O} is specified are
3125 different depending on whether or not @option{-g} is also specified.
3127 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
3128 DBX@. DEC is now aware of this problem with the assembler and hopes to
3129 provide a fix shortly.
3131 @c FIXME: still applicable?
3136 @heading @anchor{arm-x-elf}arm-*-elf
3137 ARM-family processors. Subtargets that use the ELF object format
3138 require GNU binutils 2.13 or newer. Such subtargets include:
3139 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux}
3140 and @code{arm-*-rtems}.
3145 @heading @anchor{avr}avr
3147 ATMEL AVR-family micro controllers. These are used in embedded
3148 applications. There are no standard Unix configurations.
3150 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
3154 See ``AVR Options'' in the main manual
3156 for the list of supported MCU types.
3158 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
3160 Further installation notes and other useful information about AVR tools
3161 can also be obtained from:
3165 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
3167 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
3170 We @emph{strongly} recommend using binutils 2.13 or newer.
3172 The following error:
3174 Error: register required
3177 indicates that you should upgrade to a newer version of the binutils.
3182 @heading @anchor{bfin}Blackfin
3184 The Blackfin processor, an Analog Devices DSP.
3186 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
3190 See ``Blackfin Options'' in the main manual
3193 More information, and a version of binutils with support for this processor,
3194 is available at @uref{http://blackfin.uclinux.org}
3199 @heading @anchor{cris}CRIS
3201 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
3202 series. These are used in embedded applications.
3205 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
3209 See ``CRIS Options'' in the main manual
3211 for a list of CRIS-specific options.
3213 There are a few different CRIS targets:
3216 Mainly for monolithic embedded systems. Includes a multilib for the
3217 @samp{v10} core used in @samp{ETRAX 100 LX}.
3218 @item cris-axis-linux-gnu
3219 A GNU/Linux port for the CRIS architecture, currently targeting
3220 @samp{ETRAX 100 LX} by default.
3223 For @code{cris-axis-elf} you need binutils 2.11
3224 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
3226 Pre-packaged tools can be obtained from
3227 @uref{ftp://ftp.axis.com/@/pub/@/axis/@/tools/@/cris/@/compiler-kit/}. More
3228 information about this platform is available at
3229 @uref{http://developer.axis.com/}.
3234 @heading @anchor{dos}DOS
3236 Please have a look at the @uref{binaries.html,,binaries page}.
3238 You cannot install GCC by itself on MSDOS; it will not compile under
3239 any MSDOS compiler except itself. You need to get the complete
3240 compilation package DJGPP, which includes binaries as well as sources,
3241 and includes all the necessary compilation tools and libraries.
3246 @heading @anchor{x-x-freebsd}*-*-freebsd*
3248 Support for FreeBSD 1 was discontinued in GCC 3.2. Support for
3249 FreeBSD 2 (and any mutant a.out variants of FreeBSD 3) was
3250 discontinued in GCC 4.0.
3252 In order to better utilize FreeBSD base system functionality and match
3253 the configuration of the system compiler, GCC 4.5 and above as well as
3254 GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present
3255 on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default
3256 (on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside
3257 @file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled
3258 by GCC 4.5 and above.
3260 We support FreeBSD using the ELF file format with DWARF 2 debugging
3261 for all CPU architectures. You may use @option{-gstabs} instead of
3262 @option{-g}, if you really want the old debugging format. There are
3263 no known issues with mixing object files and libraries with different
3264 debugging formats. Otherwise, this release of GCC should now match
3265 more of the configuration used in the stock FreeBSD configuration of
3266 GCC@. In particular, @option{--enable-threads} is now configured by
3267 default. However, as a general user, do not attempt to replace the
3268 system compiler with this release. Known to bootstrap and check with
3269 good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap
3270 and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4,
3271 4.5, 4.8, 4.9 and 5-CURRENT@.
3273 The version of binutils installed in @file{/usr/bin} probably works
3274 with this release of GCC@. Bootstrapping against the latest GNU
3275 binutils and/or the version found in @file{/usr/ports/devel/binutils} has
3276 been known to enable additional features and improve overall testsuite
3277 results. However, it is currently known that boehm-gc (which itself
3278 is required for java) may not configure properly on FreeBSD prior to
3279 the FreeBSD 7.0 release with GNU binutils after 2.16.1.
3284 @heading @anchor{h8300-hms}h8300-hms
3285 Renesas H8/300 series of processors.
3287 Please have a look at the @uref{binaries.html,,binaries page}.
3289 The calling convention and structure layout has changed in release 2.6.
3290 All code must be recompiled. The calling convention now passes the
3291 first three arguments in function calls in registers. Structures are no
3292 longer a multiple of 2 bytes.
3297 @heading @anchor{hppa-hp-hpux}hppa*-hp-hpux*
3298 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
3300 We require using gas/binutils on all hppa platforms. Version 2.19 or
3301 later is recommended.
3303 It may be helpful to configure GCC with the
3304 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
3305 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
3307 The HP assembler should not be used with GCC. It is rarely tested and may
3308 not work. It shouldn't be used with any languages other than C due to its
3311 Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging
3312 format which GCC does not know about). It also inserts timestamps
3313 into each object file it creates, causing the 3-stage comparison test to
3314 fail during a bootstrap. You should be able to continue by saying
3315 @samp{make all-host all-target} after getting the failure from @samp{make}.
3317 Various GCC features are not supported. For example, it does not support weak
3318 symbols or alias definitions. As a result, explicit template instantiations
3319 are required when using C++. This makes it difficult if not impossible to
3320 build many C++ applications.
3322 There are two default scheduling models for instructions. These are
3323 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
3324 architecture specified for the target machine when configuring.
3325 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
3326 the target is a @samp{hppa1*} machine.
3328 The PROCESSOR_8000 model is not well suited to older processors. Thus,
3329 it is important to completely specify the machine architecture when
3330 configuring if you want a model other than PROCESSOR_8000. The macro
3331 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
3332 default scheduling model is desired.
3334 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
3335 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
3336 This namespace change might cause problems when bootstrapping with
3337 an earlier version of GCC or the HP compiler as essentially the same
3338 namespace is required for an entire build. This problem can be avoided
3339 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
3340 or @samp{98}. Another way is to add an appropriate set of predefines
3341 to @env{CC}. The description for the @option{munix=} option contains
3342 a list of the predefines used with each standard.
3344 More specific information to @samp{hppa*-hp-hpux*} targets follows.
3349 @heading @anchor{hppa-hp-hpux10}hppa*-hp-hpux10
3351 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
3352 @code{PHCO_19798} from HP@.
3354 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
3355 used for one-only code and data. This resolves many of the previous
3356 problems in using C++ on this target. However, the ABI is not compatible
3357 with the one implemented under HP-UX 11 using secondary definitions.
3362 @heading @anchor{hppa-hp-hpux11}hppa*-hp-hpux11
3364 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
3365 be used to compile GCC 3.0 and up.
3367 The libffi and libjava libraries haven't been ported to 64-bit HP-UX@
3370 Refer to @uref{binaries.html,,binaries} for information about obtaining
3371 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
3372 to build the Ada language as it can't be bootstrapped using C@. Ada is
3373 only available for the 32-bit PA-RISC runtime.
3375 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
3376 bundled compiler supports only traditional C; you will need either HP's
3377 unbundled compiler, or a binary distribution of GCC@.
3379 It is possible to build GCC 3.3 starting with the bundled HP compiler,
3380 but the process requires several steps. GCC 3.3 can then be used to
3381 build later versions. The fastjar program contains ISO C code and
3382 can't be built with the HP bundled compiler. This problem can be
3383 avoided by not building the Java language. For example, use the
3384 @option{--enable-languages="c,c++,f77,objc"} option in your configure
3387 There are several possible approaches to building the distribution.
3388 Binutils can be built first using the HP tools. Then, the GCC
3389 distribution can be built. The second approach is to build GCC
3390 first using the HP tools, then build binutils, then rebuild GCC@.
3391 There have been problems with various binary distributions, so it
3392 is best not to start from a binary distribution.
3394 On 64-bit capable systems, there are two distinct targets. Different
3395 installation prefixes must be used if both are to be installed on
3396 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
3397 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
3398 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
3399 PA-RISC 2.0 architecture.
3401 The script config.guess now selects the target type based on the compiler
3402 detected during configuration. You must define @env{PATH} or @env{CC} so
3403 that configure finds an appropriate compiler for the initial bootstrap.
3404 When @env{CC} is used, the definition should contain the options that are
3405 needed whenever @env{CC} is used.
3407 Specifically, options that determine the runtime architecture must be
3408 in @env{CC} to correctly select the target for the build. It is also
3409 convenient to place many other compiler options in @env{CC}. For example,
3410 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
3411 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
3412 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
3413 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
3414 macro definition table of cpp needs to be increased for a successful
3415 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
3416 be defined when building with the bundled compiler, or when using the
3417 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
3419 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
3420 with the @option{--with-ld=@dots{}} option. This overrides the standard
3421 search for ld. The two linkers supported on this target require different
3422 commands. The default linker is determined during configuration. As a
3423 result, it's not possible to switch linkers in the middle of a GCC build.
3424 This has been reported to sometimes occur in unified builds of binutils
3427 A recent linker patch must be installed for the correct operation of
3428 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
3429 oldest linker patches that are known to work. They are for HP-UX
3430 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
3431 @code{PHSS_24304}, might be usable but it hasn't been tested. These
3432 patches have been superseded. Consult the HP patch database to obtain
3433 the currently recommended linker patch for your system.
3435 The patches are necessary for the support of weak symbols on the
3436 32-bit port, and for the running of initializers and finalizers. Weak
3437 symbols are implemented using SOM secondary definition symbols. Prior
3438 to HP-UX 11, there are bugs in the linker support for secondary symbols.
3439 The patches correct a problem of linker core dumps creating shared
3440 libraries containing secondary symbols, as well as various other
3441 linking issues involving secondary symbols.
3443 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
3444 run initializers and finalizers on the 64-bit port. The 32-bit port
3445 uses the linker @option{+init} and @option{+fini} options for the same
3446 purpose. The patches correct various problems with the +init/+fini
3447 options, including program core dumps. Binutils 2.14 corrects a
3448 problem on the 64-bit port resulting from HP's non-standard use of
3449 the .init and .fini sections for array initializers and finalizers.
3451 Although the HP and GNU linkers are both supported for the
3452 @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the
3453 HP linker be used for link editing on this target.
3455 At this time, the GNU linker does not support the creation of long
3456 branch stubs. As a result, it can't successfully link binaries
3457 containing branch offsets larger than 8 megabytes. In addition,
3458 there are problems linking shared libraries, linking executables
3459 with @option{-static}, and with dwarf2 unwind and exception support.
3460 It also doesn't provide stubs for internal calls to global functions
3461 in shared libraries, so these calls can't be overloaded.
3463 The HP dynamic loader does not support GNU symbol versioning, so symbol
3464 versioning is not supported. It may be necessary to disable symbol
3465 versioning with @option{--disable-symvers} when using GNU ld.
3467 POSIX threads are the default. The optional DCE thread library is not
3468 supported, so @option{--enable-threads=dce} does not work.
3473 @heading @anchor{x-x-linux-gnu}*-*-linux-gnu
3475 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
3476 in glibc 2.2.5 and later. More information is available in the
3477 libstdc++-v3 documentation.
3482 @heading @anchor{ix86-x-linux}i?86-*-linux*
3484 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
3485 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
3487 If you receive Signal 11 errors when building on GNU/Linux, then it is
3488 possible you have a hardware problem. Further information on this can be
3489 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
3494 @heading @anchor{ix86-x-solaris289}i?86-*-solaris2.[89]
3495 The Sun assembler in Solaris 8 and 9 has several bugs and limitations.
3496 While GCC works around them, several features are missing, so it is
3497 @c FIXME: which ones?
3498 recommended to use the GNU assembler instead. There is no bundled
3499 version, but the current version, from GNU binutils 2.21, is known to
3502 Solaris@tie{}2/x86 doesn't support the execution of SSE/SSE2 instructions
3503 before Solaris@tie{}9 4/04, even if the CPU supports them. Programs will
3504 receive @code{SIGILL} if they try. The fix is available both in
3505 Solaris@tie{}9 Update@tie{}6 and kernel patch 112234-12 or newer. There is no
3506 corresponding patch for Solaris 8. To avoid this problem,
3507 @option{-march} defaults to @samp{pentiumpro} on Solaris 8 and 9. If
3508 you have the patch installed, you can configure GCC with an appropriate
3509 @option{--with-arch} option, but need GNU @command{as} for SSE2 support.
3514 @heading @anchor{ix86-x-solaris210}i?86-*-solaris2.10
3515 Use this for Solaris 10 or later on x86 and x86-64 systems. This
3516 configuration is supported by GCC 4.0 and later versions only. Unlike
3517 @samp{sparcv9-sun-solaris2*}, there is no corresponding 64-bit
3518 configuration like @samp{amd64-*-solaris2*} or @samp{x86_64-*-solaris2*}.
3519 @c FIXME: will there ever be?
3521 It is recommended that you configure GCC to use the GNU assembler, in
3522 @file{/usr/sfw/bin/gas}. The versions included in Solaris 10, from GNU
3523 binutils 2.15, and Solaris 11, from GNU binutils 2.19, work fine,
3524 although the current version, from GNU binutils
3525 2.21, is known to work, too. Recent versions of the Sun assembler in
3526 @file{/usr/ccs/bin/as} work almost as well, though.
3527 @c FIXME: as patch requirements?
3529 For linking, the Sun linker, is preferred. If you want to use the GNU
3530 linker instead, which is available in @file{/usr/sfw/bin/gld}, note that
3531 due to a packaging bug the version in Solaris 10, from GNU binutils
3532 2.15, cannot be used, while the version in Solaris 11, from GNU binutils
3533 2.19, works, as does the latest version, from GNU binutils 2.21.
3535 To use GNU @command{as}, configure with the options
3536 @option{--with-gnu-as --with-as=@//usr/@/sfw/@/bin/@/gas}. It may be necessary
3537 to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to
3538 guarantee use of Sun @command{ld}.
3539 @c FIXME: why --without-gnu-ld --with-ld?
3544 @heading @anchor{ia64-x-linux}ia64-*-linux
3545 IA-64 processor (also known as IPF, or Itanium Processor Family)
3548 If you are using the installed system libunwind library with
3549 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
3552 None of the following versions of GCC has an ABI that is compatible
3553 with any of the other versions in this list, with the exception that
3554 Red Hat 2.96 and Trillian 000171 are compatible with each other:
3555 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
3556 This primarily affects C++ programs and programs that create shared libraries.
3557 GCC 3.1 or later is recommended for compiling linux, the kernel.
3558 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
3559 more major ABI changes are expected.
3564 @heading @anchor{ia64-x-hpux}ia64-*-hpux*
3565 Building GCC on this target requires the GNU Assembler. The bundled HP
3566 assembler will not work. To prevent GCC from using the wrong assembler,
3567 the option @option{--with-gnu-as} may be necessary.
3569 The GCC libunwind library has not been ported to HPUX@. This means that for
3570 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
3571 is required to build GCC@. For GCC 3.3 and later, this is the default.
3572 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
3573 removed and the system libunwind library will always be used.
3577 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
3579 @heading @anchor{x-ibm-aix}*-ibm-aix*
3580 Support for AIX version 3 and older was discontinued in GCC 3.4.
3581 Support for AIX version 4.2 and older was discontinued in GCC 4.5.
3583 ``out of memory'' bootstrap failures may indicate a problem with
3584 process resource limits (ulimit). Hard limits are configured in the
3585 @file{/etc/security/limits} system configuration file.
3587 GCC can bootstrap with recent versions of IBM XLC, but bootstrapping
3588 with an earlier release of GCC is recommended. Bootstrapping with XLC
3589 requires a larger data segment, which can be enabled through the
3590 @var{LDR_CNTRL} environment variable, e.g.,
3593 % LDR_CNTRL=MAXDATA=0x50000000
3597 One can start with a pre-compiled version of GCC to build from
3598 sources. One may delete GCC's ``fixed'' header files when starting
3599 with a version of GCC built for an earlier release of AIX.
3601 To speed up the configuration phases of bootstrapping and installing GCC,
3602 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
3605 % CONFIG_SHELL=/opt/freeware/bin/bash
3606 % export CONFIG_SHELL
3609 and then proceed as described in @uref{build.html,,the build
3610 instructions}, where we strongly recommend specifying an absolute path
3611 to invoke @var{srcdir}/configure.
3613 Because GCC on AIX is built as a 32-bit executable by default,
3614 (although it can generate 64-bit programs) the GMP and MPFR libraries
3615 required by gfortran must be 32-bit libraries. Building GMP and MPFR
3616 as static archive libraries works better than shared libraries.
3618 Errors involving @code{alloca} when building GCC generally are due
3619 to an incorrect definition of @code{CC} in the Makefile or mixing files
3620 compiled with the native C compiler and GCC@. During the stage1 phase of
3621 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
3622 (not @command{xlc}). Once @command{configure} has been informed of
3623 @command{xlc}, one needs to use @samp{make distclean} to remove the
3624 configure cache files and ensure that @env{CC} environment variable
3625 does not provide a definition that will confuse @command{configure}.
3626 If this error occurs during stage2 or later, then the problem most likely
3627 is the version of Make (see above).
3629 The native @command{as} and @command{ld} are recommended for bootstrapping
3630 on AIX@. The GNU Assembler, GNU Linker, and GNU Binutils version 2.20
3631 is required to bootstrap on AIX 5@. The native AIX tools do
3632 interoperate with GCC@.
3634 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
3635 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
3636 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
3637 referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1)
3639 @samp{libstdc++} in GCC 3.4 increments the major version number of the
3640 shared object and GCC installation places the @file{libstdc++.a}
3641 shared library in a common location which will overwrite the and GCC
3642 3.3 version of the shared library. Applications either need to be
3643 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
3644 versions of the @samp{libstdc++} shared object needs to be available
3645 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
3646 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
3647 installed for runtime dynamic loading using the following steps to set
3648 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3649 multilib @file{libstdc++.a} installed:
3651 Extract the shared objects from the currently installed
3652 @file{libstdc++.a} archive:
3654 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3657 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3658 available for runtime dynamic loading, but not linking:
3660 % strip -e libstdc++.so.4 libstdc++.so.5
3663 Archive the runtime-only shared object in the GCC 3.4
3664 @file{libstdc++.a} archive:
3666 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
3669 Linking executables and shared libraries may produce warnings of
3670 duplicate symbols. The assembly files generated by GCC for AIX always
3671 have included multiple symbol definitions for certain global variable
3672 and function declarations in the original program. The warnings should
3673 not prevent the linker from producing a correct library or runnable
3676 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
3677 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
3678 to parse archive libraries did not handle the new format correctly.
3679 These routines are used by GCC and result in error messages during
3680 linking such as ``not a COFF file''. The version of the routines shipped
3681 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
3682 option of the archive command may be used to create archives of 32-bit
3683 objects using the original ``small format''. A correct version of the
3684 routines is shipped with AIX 4.3.2 and above.
3686 Some versions of the AIX binder (linker) can fail with a relocation
3687 overflow severe error when the @option{-bbigtoc} option is used to link
3688 GCC-produced object files into an executable that overflows the TOC@. A fix
3689 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
3690 available from IBM Customer Support and from its
3691 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3692 website as PTF U455193.
3694 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
3695 with a segmentation fault when invoked by any version of GCC@. A fix for
3696 APAR IX87327 is available from IBM Customer Support and from its
3697 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3698 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
3700 The initial assembler shipped with AIX 4.3.0 generates incorrect object
3701 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
3702 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
3703 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3704 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
3706 AIX provides National Language Support (NLS)@. Compilers and assemblers
3707 use NLS to support locale-specific representations of various data
3708 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
3709 separating decimal fractions). There have been problems reported where
3710 GCC does not produce the same floating-point formats that the assembler
3711 expects. If one encounters this problem, set the @env{LANG}
3712 environment variable to @samp{C} or @samp{En_US}.
3714 A default can be specified with the @option{-mcpu=@var{cpu_type}}
3715 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
3720 @heading @anchor{iq2000-x-elf}iq2000-*-elf
3721 Vitesse IQ2000 processors. These are used in embedded
3722 applications. There are no standard Unix configurations.
3727 @heading @anchor{lm32-x-elf}lm32-*-elf
3728 Lattice Mico32 processor.
3729 This configuration is intended for embedded systems.
3734 @heading @anchor{lm32-x-uclinux}lm32-*-uclinux
3735 Lattice Mico32 processor.
3736 This configuration is intended for embedded systems running uClinux.
3741 @heading @anchor{m32c-x-elf}m32c-*-elf
3742 Renesas M32C processor.
3743 This configuration is intended for embedded systems.
3748 @heading @anchor{m32r-x-elf}m32r-*-elf
3749 Renesas M32R processor.
3750 This configuration is intended for embedded systems.
3755 @heading @anchor{m68k-x-x}m68k-*-*
3757 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
3759 build libraries for both M680x0 and ColdFire processors. If you only
3760 need the M680x0 libraries, you can omit the ColdFire ones by passing
3761 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
3762 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
3763 @command{configure}. These targets default to 5206 or 5475 code as
3764 appropriate for the target system when
3765 configured with @option{--with-arch=cf} and 68020 code otherwise.
3767 The @samp{m68k-*-netbsd} and
3768 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
3769 option. They will generate ColdFire CFV4e code when configured with
3770 @option{--with-arch=cf} and 68020 code otherwise.
3772 You can override the default processors listed above by configuring
3773 with @option{--with-cpu=@var{target}}. This @var{target} can either
3774 be a @option{-mcpu} argument or one of the following values:
3775 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
3776 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
3778 GCC requires at least binutils version 2.17 on these targets.
3783 @heading @anchor{m68k-x-uclinux}m68k-*-uclinux
3784 GCC 4.3 changed the uClinux configuration so that it uses the
3785 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
3786 It also added improved support for C++ and flat shared libraries,
3787 both of which were ABI changes.
3793 @heading @anchor{mep-x-elf}mep-*-elf
3794 Toshiba Media embedded Processor.
3795 This configuration is intended for embedded systems.
3800 @heading @anchor{microblaze-x-elf}microblaze-*-elf
3801 Xilinx MicroBlaze processor.
3802 This configuration is intended for embedded systems.
3807 @heading @anchor{mips-x-x}mips-*-*
3808 If on a MIPS system you get an error message saying ``does not have gp
3809 sections for all it's [sic] sectons [sic]'', don't worry about it. This
3810 happens whenever you use GAS with the MIPS linker, but there is not
3811 really anything wrong, and it is okay to use the output file. You can
3812 stop such warnings by installing the GNU linker.
3814 It would be nice to extend GAS to produce the gp tables, but they are
3815 optional, and there should not be a warning about their absence.
3817 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
3818 and later. A patch went in just after the GCC 3.3 release to
3819 make @samp{mips*-*-*} use the generic implementation instead. You can also
3820 configure for @samp{mipsel-elf} as a workaround. The
3821 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
3822 work on this is expected in future releases.
3824 @c If you make --with-llsc the default for another target, please also
3825 @c update the description of the --with-llsc option.
3827 The built-in @code{__sync_*} functions are available on MIPS II and
3828 later systems and others that support the @samp{ll}, @samp{sc} and
3829 @samp{sync} instructions. This can be overridden by passing
3830 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
3831 Since the Linux kernel emulates these instructions if they are
3832 missing, the default for @samp{mips*-*-linux*} targets is
3833 @option{--with-llsc}. The @option{--with-llsc} and
3834 @option{--without-llsc} configure options may be overridden at compile
3835 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
3838 MIPS systems check for division by zero (unless
3839 @option{-mno-check-zero-division} is passed to the compiler) by
3840 generating either a conditional trap or a break instruction. Using
3841 trap results in smaller code, but is only supported on MIPS II and
3842 later. Also, some versions of the Linux kernel have a bug that
3843 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
3844 the use of break, use the @option{--with-divide=breaks}
3845 @command{configure} option when configuring GCC@. The default is to
3846 use traps on systems that support them.
3848 Cross-compilers for the MIPS as target using the MIPS assembler
3849 currently do not work, because the auxiliary programs
3850 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
3851 anything but a MIPS@. It does work to cross compile for a MIPS
3852 if you use the GNU assembler and linker.
3854 The assembler from GNU binutils 2.17 and earlier has a bug in the way
3855 it sorts relocations for REL targets (o32, o64, EABI). This can cause
3856 bad code to be generated for simple C++ programs. Also the linker
3857 from GNU binutils versions prior to 2.17 has a bug which causes the
3858 runtime linker stubs in very large programs, like @file{libgcj.so}, to
3859 be incorrectly generated. GNU Binutils 2.18 and later (and snapshots
3860 made after Nov. 9, 2006) should be free from both of these problems.
3865 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
3867 Support for IRIX 5 has been removed in GCC 4.6.
3872 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3874 Support for IRIX 6 releases before 6.5 has been removed in GCC 4.6, as
3876 the O32 ABI. It is @emph{strongly} recommended to upgrade to at least
3877 IRIX 6.5.18. This release introduced full ISO C99 support, though for
3878 the N32 and N64 ABIs only.
3880 To build and use GCC on IRIX 6.5, you need the IRIX Development Foundation
3881 (IDF) and IRIX Development Libraries (IDL). They are included with the
3884 If you are using SGI's MIPSpro @command{cc} as your bootstrap compiler, you must
3885 ensure that the N32 ABI is in use. To test this, compile a simple C
3886 file with @command{cc} and then run @command{file} on the
3887 resulting object file. The output should look like:
3890 test.o: ELF N32 MSB @dots{}
3897 test.o: ELF 32-bit MSB @dots{}
3904 test.o: ELF 64-bit MSB @dots{}
3908 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3909 should set the environment variable @env{CC} to @samp{cc -n32}
3910 before configuring GCC@.
3912 If you want the resulting @command{gcc} to run on old 32-bit systems
3913 with the MIPS R4400 CPU, you need to ensure that only code for the @samp{mips3}
3914 instruction set architecture (ISA) is generated. While GCC 3.x does
3915 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3916 the ISA depending on the machine where GCC is built. Using one of them
3917 as the bootstrap compiler may result in @samp{mips4} code, which won't run at
3918 all on @samp{mips3}-only systems. For the test program above, you should see:
3921 test.o: ELF N32 MSB mips-3 @dots{}
3928 test.o: ELF N32 MSB mips-4 @dots{}
3932 instead, you should set the environment variable @env{CC} to @samp{cc
3933 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3935 MIPSpro C 7.4 may cause bootstrap failures, due to a bug when inlining
3936 @code{memcmp}. Either add @code{-U__INLINE_INTRINSICS} to the @env{CC}
3937 environment variable as a workaround or upgrade to MIPSpro C 7.4.1m.
3939 GCC on IRIX 6.5 is usually built to support the N32 and N64 ABIs. If
3940 you build GCC on a system that doesn't have the N64 libraries installed
3941 or cannot run 64-bit binaries,
3942 you need to configure with @option{--disable-multilib} so GCC doesn't
3944 Look for @file{/usr/lib64/libc.so.1} to see if you
3945 have the 64-bit libraries installed.
3947 GCC must be configured with GNU @command{as}. The latest version, from GNU
3948 binutils 2.21, is known to work. On the other hand, bootstrap fails
3949 with GNU @command{ld} at least since GNU binutils 2.17.
3951 The @option{--enable-libgcj}
3952 option is disabled by default: IRIX 6 uses a very low default limit
3953 (20480) for the command line length. Although @command{libtool} contains a
3954 workaround for this problem, at least the N64 @samp{libgcj} is known not
3955 to build despite this, running into an internal error of the native
3956 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3957 its maximum of 262144 bytes. If you have root access, you can use the
3958 @command{systune} command to do this.
3959 @c FIXME: does this work with current libtool?
3961 @code{wchar_t} support in @samp{libstdc++} is not available for old
3962 IRIX 6.5.x releases, @math{x < 19}. The problem cannot be autodetected
3963 and in order to build GCC for such targets you need to configure with
3964 @option{--disable-wchar_t}.
3969 @heading @anchor{moxie-x-elf}moxie-*-elf
3970 The moxie processor. See @uref{http://moxielogic.org/} for more
3971 information about this processor.
3976 @heading @anchor{powerpc-x-x}powerpc-*-*
3978 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3979 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3982 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.15}
3983 or newer for a working GCC@.
3988 @heading @anchor{powerpc-x-darwin}powerpc-*-darwin*
3989 PowerPC running Darwin (Mac OS X kernel).
3991 Pre-installed versions of Mac OS X may not include any developer tools,
3992 meaning that you will not be able to build GCC from source. Tool
3993 binaries are available at
3994 @uref{http://opensource.apple.com/}.
3996 This version of GCC requires at least cctools-590.36. The
3997 cctools-590.36 package referenced from
3998 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
3999 on systems older than 10.3.9 (aka darwin7.9.0).
4004 @heading @anchor{powerpc-x-elf}powerpc-*-elf
4005 PowerPC system in big endian mode, running System V.4.
4010 @heading @anchor{powerpc-x-linux-gnu}powerpc*-*-linux-gnu*
4012 PowerPC system in big endian mode running Linux.
4017 @heading @anchor{powerpc-x-netbsd}powerpc-*-netbsd*
4018 PowerPC system in big endian mode running NetBSD@.
4023 @heading @anchor{powerpc-x-eabisim}powerpc-*-eabisim
4024 Embedded PowerPC system in big endian mode for use in running under the
4030 @heading @anchor{powerpc-x-eabi}powerpc-*-eabi
4031 Embedded PowerPC system in big endian mode.
4036 @heading @anchor{powerpcle-x-elf}powerpcle-*-elf
4037 PowerPC system in little endian mode, running System V.4.
4042 @heading @anchor{powerpcle-x-eabisim}powerpcle-*-eabisim
4043 Embedded PowerPC system in little endian mode for use in running under
4049 @heading @anchor{powerpcle-x-eabi}powerpcle-*-eabi
4050 Embedded PowerPC system in little endian mode.
4055 @heading @anchor{rx-x-elf}rx-*-elf
4056 The Renesas RX processor. See
4057 @uref{http://eu.renesas.com/fmwk.jsp?cnt=rx600_series_landing.jsp&fp=/products/mpumcu/rx_family/rx600_series}
4058 for more information about this processor.
4063 @heading @anchor{s390-x-linux}s390-*-linux*
4064 S/390 system running GNU/Linux for S/390@.
4069 @heading @anchor{s390x-x-linux}s390x-*-linux*
4070 zSeries system (64-bit) running GNU/Linux for zSeries@.
4075 @heading @anchor{s390x-ibm-tpf}s390x-ibm-tpf*
4076 zSeries system (64-bit) running TPF@. This platform is
4077 supported as cross-compilation target only.
4082 @c Please use Solaris 2 to refer to all release of Solaris, starting
4083 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
4084 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
4085 @c alone is too unspecific and must be avoided.
4086 @heading @anchor{x-x-solaris2}*-*-solaris2*
4088 Support for Solaris 8 has been obsoleted in GCC 4.7, but can still be
4089 enabled by configuring with @option{--enable-obsolete}. Support will be
4090 removed in GCC 4.8. Support for Solaris 7 has been removed in GCC 4.6.
4092 Sun does not ship a C compiler with Solaris 2, though you can download
4093 the Sun Studio compilers for free. Alternatively,
4094 you can install a pre-built GCC to bootstrap and install GCC. See the
4095 @uref{binaries.html,,binaries page} for details.
4097 The Solaris 2 @command{/bin/sh} will often fail to configure
4098 @samp{libstdc++-v3}, @samp{boehm-gc} or @samp{libjava}. We therefore
4099 recommend using the following initial sequence of commands
4102 % CONFIG_SHELL=/bin/ksh
4103 % export CONFIG_SHELL
4107 and proceed as described in @uref{configure.html,,the configure instructions}.
4108 In addition we strongly recommend specifying an absolute path to invoke
4109 @command{@var{srcdir}/configure}.
4111 Solaris 2 comes with a number of optional OS packages. Some of these
4112 are needed to use GCC fully, namely @code{SUNWarc},
4113 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
4114 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
4115 optional packages when installing Solaris 2, you will need to verify that
4116 the packages that GCC needs are installed.
4118 To check whether an optional package is installed, use
4119 the @command{pkginfo} command. To add an optional package, use the
4120 @command{pkgadd} command. For further details, see the Solaris 2
4123 Trying to use the linker and other tools in
4124 @file{/usr/ucb} to install GCC has been observed to cause trouble.
4125 For example, the linker may hang indefinitely. The fix is to remove
4126 @file{/usr/ucb} from your @env{PATH}.
4128 The build process works more smoothly with the legacy Sun tools so, if you
4129 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
4130 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
4132 We recommend the use of the Sun assembler or the GNU assembler, in
4133 conjunction with the Sun linker. The GNU @command{as}
4134 versions included in Solaris 10, from GNU binutils 2.15, and Solaris 11,
4135 from GNU binutils 2.19, are known to work. They can be found in
4136 @file{/usr/sfw/bin/gas}. Current versions of GNU binutils (2.21)
4137 are known to work as well. Note that your mileage may vary
4138 if you use a combination of the GNU tools and the Sun tools: while the
4139 combination GNU @command{as} + Sun @command{ld} should reasonably work,
4140 the reverse combination Sun @command{as} + GNU @command{ld} is known to
4141 cause memory corruption at runtime in some cases for C++ programs.
4143 GNU @command{ld} usually works as well, although the version included in
4144 Solaris 10 cannot be used due to several bugs. Again, the current
4145 version (2.21) is known to work, but generally lacks platform specific
4146 features, so better stay with Sun @command{ld}.
4148 To enable symbol versioning in @samp{libstdc++} with Sun @command{ld},
4149 you need to have any version of GNU @command{c++filt}, which is part of
4150 GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no
4151 appropriate version is found. Sun @command{c++filt} from the Sun Studio
4152 compilers does @emph{not} work.
4154 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
4155 newer: @command{g++} will complain that types are missing. These headers
4156 assume that omitting the type means @code{int}; this assumption worked for
4157 C90 but is wrong for C++, and is now wrong for C99 also.
4159 @command{g++} accepts such (invalid) constructs with the option
4160 @option{-fpermissive}; it will assume that any missing type is @code{int}
4161 (as defined by C90).
4163 There are patches for Solaris 8 (108652-24 or newer for SPARC,
4164 108653-22 for Intel) that fix this bug.
4166 Sun bug 4927647 sometimes causes random spurious testsuite failures
4167 related to missing diagnostic output. This bug doesn't affect GCC
4168 itself, rather it is a kernel bug triggered by the @command{expect}
4169 program which is used only by the GCC testsuite driver. When the bug
4170 causes the @command{expect} program to miss anticipated output, extra
4171 testsuite failures appear.
4173 There are patches for Solaris 8 (117350-12 or newer for SPARC,
4174 117351-12 or newer for Intel) and Solaris 9 (117171-11 or newer for
4175 SPARC, 117172-11 or newer for Intel) that address this problem.
4177 Solaris@tie{}8 provides an alternate implementation of the thread libraries,
4178 @samp{libpthread} and @samp{libthread}. They are required for TLS
4179 support and have been made the default in Solaris@tie{}9, so they are always
4180 used on Solaris@tie{}8.
4182 Thread-local storage (TLS) is supported in Solaris@tie{}8 and 9, but requires
4183 some patches. The @samp{libthread} patches provide the
4184 @code{__tls_get_addr} (SPARC, 64-bit x86) resp.@ @code{___tls_get_addr}
4185 (32-bit x86) functions. On Solaris@tie{}8, you need 108993-26 or newer on
4186 SPARC, 108994-26 or newer on Intel. On Solaris@tie{}9, the necessary support
4187 on SPARC is present since FCS, while 114432-05 or newer is required on
4188 Intel. Additionally, on Solaris@tie{}8, patch 109147-14 or newer on SPARC or
4189 109148-22 or newer on Intel are required for the Sun @command{ld} and
4190 runtime linker (@command{ld.so.1}) support. Again, Solaris@tie{}9/SPARC
4191 works since FCS, while 113986-02 is required on Intel. The linker
4192 patches must be installed even if GNU @command{ld} is used. Sun
4193 @command{as} in Solaris@tie{}8 and 9 doesn't support the necessary
4194 relocations, so GNU @command{as} must be used. The @command{configure}
4195 script checks for those prerequisites and automatically enables TLS
4196 support if they are met. Although those minimal patch versions should
4197 work, it is recommended to use the latest patch versions which include
4198 additional bug fixes.
4203 @heading @anchor{sparc-x-x}sparc*-*-*
4205 This section contains general configuration information for all
4206 SPARC-based platforms. In addition to reading this section, please
4207 read all other sections that match your target.
4209 Newer versions of the GNU Multiple Precision Library (GMP), the MPFR
4210 library and the MPC library are known to be miscompiled by earlier
4211 versions of GCC on these platforms. We therefore recommend the use
4212 of the exact versions of these libraries listed as minimal versions
4213 in @uref{prerequisites.html,,the prerequisites}.
4218 @heading @anchor{sparc-sun-solaris2}sparc-sun-solaris2*
4220 When GCC is configured to use GNU binutils 2.14 or later, the binaries
4221 produced are smaller than the ones produced using Sun's native tools;
4222 this difference is quite significant for binaries containing debugging
4225 Starting with Solaris 7, the operating system is capable of executing
4226 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
4227 this; the @option{-m64} option enables 64-bit code generation.
4228 However, if all you want is code tuned for the UltraSPARC CPU, you
4229 should try the @option{-mtune=ultrasparc} option instead, which produces
4230 code that, unlike full 64-bit code, can still run on non-UltraSPARC
4233 When configuring on a Solaris 7 or later system that is running a kernel
4234 that supports only 32-bit binaries, one must configure with
4235 @option{--disable-multilib}, since we will not be able to build the
4236 64-bit target libraries.
4238 GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
4239 the GNU compiler (especially GCC 3.0.x versions), which lead to the
4240 miscompilation of the stage1 compiler and the subsequent failure of the
4241 bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
4242 stage, i.e.@: to bootstrap that compiler with the base compiler and then
4243 use it to bootstrap the final compiler.
4245 GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
4246 and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
4247 failure in form of a miscompilation of the stage1 compiler by the Sun
4248 compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
4250 GCC 3.4 changed the default debugging format from Stabs to DWARF-2 for
4251 32-bit code on Solaris 7 and later. If you use the Sun assembler, this
4252 change apparently runs afoul of Sun bug 4910101 (which is referenced as
4253 an x86-only problem by Sun, probably because they do not use DWARF-2).
4254 A symptom of the problem is that you cannot compile C++ programs like
4255 @command{groff} 1.19.1 without getting messages similar to the following:
4258 ld: warning: relocation error: R_SPARC_UA32: @dots{}
4259 external symbolic relocation against non-allocatable section
4260 .debug_info cannot be processed at runtime: relocation ignored.
4264 To work around this problem, compile with @option{-gstabs+} instead of
4267 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4268 library or the MPC library on a Solaris 7 or later system, the canonical
4269 target triplet must be specified as the @command{build} parameter on the
4270 configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and
4271 not that of GMP or MPFR or MPC). For example on a Solaris 9 system:
4274 % ./configure --build=sparc-sun-solaris2.9 --prefix=xxx
4280 @heading @anchor{sparc-sun-solaris210}sparc-sun-solaris2.10
4282 There is a bug in older versions of the Sun assembler which breaks
4283 thread-local storage (TLS). A typical error message is
4286 ld: fatal: relocation error: R_SPARC_TLS_LE_HIX22: file /var/tmp//ccamPA1v.o:
4287 symbol <unknown>: bad symbol type SECT: symbol type must be TLS
4291 This bug is fixed in Sun patch 118683-03 or later.
4296 @heading @anchor{sparc-x-linux}sparc-*-linux*
4298 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
4299 or newer on this platform. All earlier binutils and glibc
4300 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
4306 @heading @anchor{sparc64-x-solaris2}sparc64-*-solaris2*
4308 When configuring the GNU Multiple Precision Library (GMP) or the
4309 MPFR library, the canonical target triplet must be specified as
4310 the @command{build} parameter on the configure line. For example
4311 on a Solaris 9 system:
4314 % ./configure --build=sparc64-sun-solaris2.9 --prefix=xxx
4317 The following compiler flags must be specified in the configure
4318 step in order to bootstrap this target with the Sun compiler:
4321 % CC="cc -xarch=v9 -xildoff" @var{srcdir}/configure [@var{options}] [@var{target}]
4325 @option{-xarch=v9} specifies the SPARC-V9 architecture to the Sun toolchain
4326 and @option{-xildoff} turns off the incremental linker.
4331 @heading @anchor{sparcv9-x-solaris2}sparcv9-*-solaris2*
4333 This is a synonym for @samp{sparc64-*-solaris2*}.
4338 @heading @anchor{x-x-vxworks}*-*-vxworks*
4339 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4340 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4341 We welcome patches for other architectures supported by VxWorks 5.5.
4342 Support for VxWorks AE would also be welcome; we believe this is merely
4343 a matter of writing an appropriate ``configlette'' (see below). We are
4344 not interested in supporting older, a.out or COFF-based, versions of
4347 VxWorks comes with an older version of GCC installed in
4348 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4349 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4350 Before running @command{configure}, create the directories @file{@var{prefix}}
4351 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4352 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4353 include that directory while running both @command{configure} and
4356 You must give @command{configure} the
4357 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4358 find the VxWorks system headers. Since VxWorks is a cross compilation
4359 target only, you must also specify @option{--target=@var{target}}.
4360 @command{configure} will attempt to create the directory
4361 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4362 make sure the user running @command{configure} has sufficient privilege
4365 GCC's exception handling runtime requires a special ``configlette''
4366 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4367 that file to add the module to your kernel build. (Future versions of
4368 VxWorks will incorporate this module.)
4373 @heading @anchor{x86-64-x-x}x86_64-*-*, amd64-*-*
4375 GCC supports the x86-64 architecture implemented by the AMD64 processor
4376 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4377 On GNU/Linux the default is a bi-arch compiler which is able to generate
4378 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4383 @heading @anchor{xtensa-x-elf}xtensa*-*-elf
4385 This target is intended for embedded Xtensa systems using the
4386 @samp{newlib} C library. It uses ELF but does not support shared
4387 objects. Designed-defined instructions specified via the
4388 Tensilica Instruction Extension (TIE) language are only supported
4389 through inline assembly.
4391 The Xtensa configuration information must be specified prior to
4392 building GCC@. The @file{include/xtensa-config.h} header
4393 file contains the configuration information. If you created your
4394 own Xtensa configuration with the Xtensa Processor Generator, the
4395 downloaded files include a customized copy of this header file,
4396 which you can use to replace the default header file.
4401 @heading @anchor{xtensa-x-linux}xtensa*-*-linux*
4403 This target is for Xtensa systems running GNU/Linux. It supports ELF
4404 shared objects and the GNU C library (glibc). It also generates
4405 position-independent code (PIC) regardless of whether the
4406 @option{-fpic} or @option{-fPIC} options are used. In other
4407 respects, this target is the same as the
4408 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
4413 @heading @anchor{windows}Microsoft Windows
4415 @subheading Intel 16-bit versions
4416 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
4419 However, the 32-bit port has limited support for Microsoft
4420 Windows 3.11 in the Win32s environment, as a target only. See below.
4422 @subheading Intel 32-bit versions
4424 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
4425 XP, and Windows Vista, are supported by several different target
4426 platforms. These targets differ in which Windows subsystem they target
4427 and which C libraries are used.
4430 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
4431 Linux API emulation layer in the Win32 subsystem.
4432 @item Interix @uref{#x-x-interix,,*-*-interix}: The Interix subsystem
4433 provides native support for POSIX.
4434 @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for
4435 the Win32 subsystem that provides a subset of POSIX.
4436 @item MKS i386-pc-mks: NuTCracker from MKS. See
4437 @uref{http://www.mkssoftware.com/} for more information.
4440 @subheading Intel 64-bit versions
4442 GCC contains support for x86-64 using the mingw-w64
4443 runtime library, available from @uref{http://mingw-w64.sourceforge.net/}.
4444 This library should be used with the target triple x86_64-pc-mingw32.
4446 Presently Windows for Itanium is not supported.
4448 @subheading Windows CE
4450 Windows CE is supported as a target only on ARM (arm-wince-pe), Hitachi
4451 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
4453 @subheading Other Windows Platforms
4455 GCC no longer supports Windows NT on the Alpha or PowerPC.
4457 GCC no longer supports the Windows POSIX subsystem. However, it does
4458 support the Interix subsystem. See above.
4460 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
4462 PW32 (i386-pc-pw32) support was never completed, and the project seems to
4463 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
4465 UWIN support has been removed due to a lack of maintenance.
4470 @heading @anchor{x-x-cygwin}*-*-cygwin
4472 Ports of GCC are included with the
4473 @uref{http://www.cygwin.com/,,Cygwin environment}.
4475 GCC will build under Cygwin without modification; it does not build
4476 with Microsoft's C++ compiler and there are no plans to make it do so.
4478 The Cygwin native compiler can be configured to target any 32-bit x86
4479 cpu architecture desired; the default is i686-pc-cygwin. It should be
4480 used with as up-to-date a version of binutils as possible; use either
4481 the latest official GNU binutils release in the Cygwin distribution,
4482 or version 2.20 or above if building your own.
4487 @heading @anchor{x-x-interix}*-*-interix
4489 The Interix target is used by OpenNT, Interix, Services For UNIX (SFU),
4490 and Subsystem for UNIX-based Applications (SUA). Applications compiled
4491 with this target run in the Interix subsystem, which is separate from
4492 the Win32 subsystem. This target was last known to work in GCC 3.3.
4497 @heading @anchor{x-x-mingw32}*-*-mingw32
4499 GCC will build with and support only MinGW runtime 3.12 and later.
4500 Earlier versions of headers are incompatible with the new default semantics
4501 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
4506 @heading @anchor{older}Older systems
4508 GCC contains support files for many older (1980s and early
4509 1990s) Unix variants. For the most part, support for these systems
4510 has not been deliberately removed, but it has not been maintained for
4511 several years and may suffer from bitrot.
4513 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
4514 Support for these systems is still present in that release, but
4515 @command{configure} will fail unless the @option{--enable-obsolete}
4516 option is given. Unless a maintainer steps forward, support for these
4517 systems will be removed from the next release of GCC@.
4519 Support for old systems as hosts for GCC can cause problems if the
4520 workarounds for compiler, library and operating system bugs affect the
4521 cleanliness or maintainability of the rest of GCC@. In some cases, to
4522 bring GCC up on such a system, if still possible with current GCC, may
4523 require first installing an old version of GCC which did work on that
4524 system, and using it to compile a more recent GCC, to avoid bugs in the
4525 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
4526 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
4527 sites}. Header bugs may generally be avoided using
4528 @command{fixincludes}, but bugs or deficiencies in libraries and the
4529 operating system may still cause problems.
4531 Support for older systems as targets for cross-compilation is less
4532 problematic than support for them as hosts for GCC; if an enthusiast
4533 wishes to make such a target work again (including resurrecting any of
4534 the targets that never worked with GCC 2, starting from the last
4535 version before they were removed), patches
4536 @uref{../contribute.html,,following the usual requirements} would be
4537 likely to be accepted, since they should not affect the support for more
4540 For some systems, old versions of GNU binutils may also be useful,
4541 and are available from @file{pub/binutils/old-releases} on
4542 @uref{http://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
4544 Some of the information on specific systems above relates to
4545 such older systems, but much of the information
4546 about GCC on such systems (which may no longer be applicable to
4547 current GCC) is to be found in the GCC texinfo manual.
4552 @heading @anchor{elf}all ELF targets (SVR4, Solaris 2, etc.)
4554 C++ support is significantly better on ELF targets if you use the
4555 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
4556 inlines, vtables and template instantiations will be discarded
4565 @uref{./index.html,,Return to the GCC Installation page}
4569 @c ***Old documentation******************************************************
4571 @include install-old.texi
4577 @uref{./index.html,,Return to the GCC Installation page}
4581 @c ***GFDL********************************************************************
4589 @uref{./index.html,,Return to the GCC Installation page}
4593 @c ***************************************************************************
4594 @c Part 6 The End of the Document
4596 @comment node-name, next, previous, up
4597 @node Concept Index, , GNU Free Documentation License, Top
4601 @unnumbered Concept Index