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 Free Software Foundation, Inc.
49 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
51 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
52 @c test the generation of HTML documents for the gcc.gnu.org web pages.
54 @c Do not use @footnote{} in this file as it breaks install.texi2html!
56 @c Include everything if we're not making html
60 @set prerequisiteshtml
71 @c Part 2 Summary Description and Copyright
73 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
74 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007,
75 2008 Free Software Foundation, Inc.
77 Permission is granted to copy, distribute and/or modify this document
78 under the terms of the GNU Free Documentation License, Version 1.2 or
79 any later version published by the Free Software Foundation; with no
80 Invariant Sections, the Front-Cover texts being (a) (see below), and
81 with the Back-Cover Texts being (b) (see below). A copy of the
82 license is included in the section entitled ``@uref{./gfdl.html,,GNU
83 Free Documentation License}''.
85 (a) The FSF's Front-Cover Text is:
89 (b) The FSF's Back-Cover Text is:
91 You have freedom to copy and modify this GNU Manual, like GNU
92 software. Copies published by the Free Software Foundation raise
93 funds for GNU development.
98 @dircategory Software development
100 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
103 @c Part 3 Titlepage and Copyright
105 @title Installing GCC
108 @c The following two commands start the copyright page.
110 @vskip 0pt plus 1filll
114 @c Part 4 Top node, Master Menu, and/or Table of Contents
117 @comment node-name, next, Previous, up
120 * Installing GCC:: This document describes the generic installation
121 procedure for GCC as well as detailing some target
122 specific installation instructions.
124 * Specific:: Host/target specific installation notes for GCC.
125 * Binaries:: Where to get pre-compiled binaries.
127 * Old:: Old installation documentation.
129 * GNU Free Documentation License:: How you can copy and share this manual.
130 * Concept Index:: This index has two entries.
138 @c Part 5 The Body of the Document
139 @c ***Installing GCC**********************************************************
141 @comment node-name, next, previous, up
142 @node Installing GCC, Binaries, , Top
146 @chapter Installing GCC
149 The latest version of this document is always available at
150 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
152 This document describes the generic installation procedure for GCC as well
153 as detailing some target specific installation instructions.
155 GCC includes several components that previously were separate distributions
156 with their own installation instructions. This document supersedes all
157 package specific installation instructions.
159 @emph{Before} starting the build/install procedure please check the
161 @ref{Specific, host/target specific installation notes}.
164 @uref{specific.html,,host/target specific installation notes}.
166 We recommend you browse the entire generic installation instructions before
169 Lists of successful builds for released versions of GCC are
170 available at @uref{http://gcc.gnu.org/buildstat.html}.
171 These lists are updated as new information becomes available.
173 The installation procedure itself is broken into five steps.
178 * Downloading the source::
181 * Testing:: (optional)
188 @uref{prerequisites.html,,Prerequisites}
190 @uref{download.html,,Downloading the source}
192 @uref{configure.html,,Configuration}
194 @uref{build.html,,Building}
196 @uref{test.html,,Testing} (optional)
198 @uref{finalinstall.html,,Final install}
202 Please note that GCC does not support @samp{make uninstall} and probably
203 won't do so in the near future as this would open a can of worms. Instead,
204 we suggest that you install GCC into a directory of its own and simply
205 remove that directory when you do not need that specific version of GCC
206 any longer, and, if shared libraries are installed there as well, no
207 more binaries exist that use them.
210 There are also some @uref{old.html,,old installation instructions},
211 which are mostly obsolete but still contain some information which has
212 not yet been merged into the main part of this manual.
220 @uref{./index.html,,Return to the GCC Installation page}
226 @c ***Prerequisites**************************************************
228 @comment node-name, next, previous, up
229 @node Prerequisites, Downloading the source, , Installing GCC
231 @ifset prerequisiteshtml
233 @chapter Prerequisites
235 @cindex Prerequisites
237 GCC requires that various tools and packages be available for use in the
238 build procedure. Modifying GCC sources requires additional tools
241 @heading Tools/packages necessary for building GCC
243 @item ISO C90 compiler
244 Necessary to bootstrap GCC, although versions of GCC prior
245 to 3.4 also allow bootstrapping with a traditional (K&R) C compiler.
247 To build all languages in a cross-compiler or other configuration where
248 3-stage bootstrap is not performed, you need to start with an existing
249 GCC binary (version 2.95 or later) because source code for language
250 frontends other than C might use GCC extensions.
254 In order to build the Ada compiler (GNAT) you must already have GNAT
255 installed because portions of the Ada frontend are written in Ada (with
256 GNAT extensions.) Refer to the Ada installation instructions for more
257 specific information.
259 @item A ``working'' POSIX compatible shell, or GNU bash
261 Necessary when running @command{configure} because some
262 @command{/bin/sh} shells have bugs and may crash when configuring the
263 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
264 have disastrous corner-case performance problems. This
265 can cause target @command{configure} runs to literally take days to
266 complete in some cases.
268 So on some platforms @command{/bin/ksh} is sufficient, on others it
269 isn't. See the host/target specific instructions for your platform, or
270 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
271 environment to your ``good'' shell prior to running
272 @command{configure}/@command{make}.
274 @command{zsh} is not a fully compliant POSIX shell and will not
275 work when configuring GCC@.
277 @item A POSIX or SVR4 awk
279 Necessary for creating some of the generated source files for GCC@.
280 If in doubt, use a recent GNU awk version, as some of the older ones
281 are broken. GNU awk version 3.1.5 is known to work.
285 Necessary in some circumstances, optional in others. See the
286 host/target specific instructions for your platform for the exact
289 @item gzip version 1.2.4 (or later) or
290 @itemx bzip2 version 1.0.2 (or later)
292 Necessary to uncompress GCC @command{tar} files when source code is
293 obtained via FTP mirror sites.
295 @item GNU make version 3.80 (or later)
297 You must have GNU make installed to build GCC@.
299 @item GNU tar version 1.14 (or later)
301 Necessary (only on some platforms) to untar the source code. Many
302 systems' @command{tar} programs will also work, only try GNU
303 @command{tar} if you have problems.
305 @item GNU Multiple Precision Library (GMP) version 4.1 (or later)
307 Necessary to build GCC@. If you do not have it installed in your
308 library search path, you will have to configure with the
309 @option{--with-gmp} configure option. See also
310 @option{--with-gmp-lib} and @option{--with-gmp-include}.
312 @item MPFR Library version 2.3.0 (or later)
314 Necessary to build GCC@. It can be downloaded from
315 @uref{http://www.mpfr.org/}. The version of MPFR that is bundled with
316 GMP 4.1.x contains numerous bugs. Although GCC may appear to function
317 with the buggy versions of MPFR, there are a few bugs that will not be
318 fixed when using this version. It is strongly recommended to upgrade
319 to the recommended version of MPFR.
321 The @option{--with-mpfr} configure option should be used if your MPFR
322 Library is not installed in your default library search path. See
323 also @option{--with-mpfr-lib} and @option{--with-mpfr-include}.
325 @item @command{jar}, or InfoZIP (@command{zip} and @command{unzip})
327 Necessary to build libgcj, the GCJ runtime.
332 @heading Tools/packages necessary for modifying GCC
334 @item autoconf version 2.59
335 @itemx GNU m4 version 1.4 (or later)
337 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
338 to regenerate @file{configure} and @file{config.in} files.
340 @item automake version 1.9.6
342 Necessary when modifying a @file{Makefile.am} file to regenerate its
343 associated @file{Makefile.in}.
345 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
346 file. Specifically this applies to the @file{gcc}, @file{intl},
347 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
348 as any of their subdirectories.
350 For directories that use automake, GCC requires the latest release in
351 the 1.9.x series, which is currently 1.9.6. When regenerating a directory
352 to a newer version, please update all the directories using an older 1.9.x
353 to the latest released version.
355 @item gettext version 0.14.5 (or later)
357 Needed to regenerate @file{gcc.pot}.
359 @item gperf version 2.7.2 (or later)
361 Necessary when modifying @command{gperf} input files, e.g.@:
362 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
363 @file{gcc/cp/cfns.h}.
369 Necessary to run the GCC testsuite; see the section on testing for details.
371 @item autogen version 5.5.4 (or later) and
372 @itemx guile version 1.4.1 (or later)
374 Necessary to regenerate @file{fixinc/fixincl.x} from
375 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
377 Necessary to run @samp{make check} for @file{fixinc}.
379 Necessary to regenerate the top level @file{Makefile.in} file from
380 @file{Makefile.tpl} and @file{Makefile.def}.
382 @item Flex version 2.5.4 (or later)
384 Necessary when modifying @file{*.l} files.
386 Necessary to build GCC during development because the generated output
387 files are not included in the SVN repository. They are included in
390 @item Texinfo version 4.7 (or later)
392 Necessary for running @command{makeinfo} when modifying @file{*.texi}
393 files to test your changes.
395 Necessary for running @command{make dvi} or @command{make pdf} to
396 create printable documentation in DVI or PDF format. Texinfo version
397 4.8 or later is required for @command{make pdf}.
399 Necessary to build GCC documentation during development because the
400 generated output files are not included in the SVN repository. They are
401 included in releases.
403 @item @TeX{} (any working version)
405 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
406 are used when running @command{make dvi} or @command{make pdf} to create
407 DVI or PDF files, respectively.
409 @item SVN (any version)
410 @itemx SSH (any version)
412 Necessary to access the SVN repository. Public releases and weekly
413 snapshots of the development sources are also available via FTP@.
415 @item Perl version 5.6.1 (or later)
417 Necessary when regenerating @file{Makefile} dependencies in libiberty.
418 Necessary when regenerating @file{libiberty/functions.texi}.
419 Necessary when generating manpages from Texinfo manuals.
420 Necessary when targetting Darwin, building libstdc++,
421 and not using @option{--disable-symvers}.
422 Used by various scripts to generate some files included in SVN (mainly
423 Unicode-related and rarely changing) from source tables.
425 @item GNU diffutils version 2.7 (or later)
427 Useful when submitting patches for the GCC source code.
429 @item patch version 2.5.4 (or later)
431 Necessary when applying patches, created with @command{diff}, to one's
437 If you wish to modify @file{.java} files in libjava, you will need to
438 configure with @option{--enable-java-maintainer-mode}, and you will need
439 to have executables named @command{ecj1} and @command{gjavah} in your path.
440 The @command{ecj1} executable should run the Eclipse Java compiler via
441 the GCC-specific entry point. You can download a suitable jar from
442 @uref{ftp://sourceware.org/pub/java/}, or by running the script
443 @command{contrib/download_ecj}.
452 @uref{./index.html,,Return to the GCC Installation page}
456 @c ***Downloading the source**************************************************
458 @comment node-name, next, previous, up
459 @node Downloading the source, Configuration, Prerequisites, Installing GCC
463 @chapter Downloading GCC
465 @cindex Downloading GCC
466 @cindex Downloading the Source
468 GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP
469 tarballs compressed with @command{gzip} or
470 @command{bzip2}. It is possible to download a full distribution or specific
473 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
474 for information on how to obtain GCC@.
476 The full distribution includes the C, C++, Objective-C, Fortran, Java,
477 and Ada (in the case of GCC 3.1 and later) compilers. The full
478 distribution also includes runtime libraries for C++, Objective-C,
479 Fortran, and Java. In GCC 3.0 and later versions, the GNU compiler
480 testsuites are also included in the full distribution.
482 If you choose to download specific components, you must download the core
483 GCC distribution plus any language specific distributions you wish to
484 use. The core distribution includes the C language front end as well as the
485 shared components. Each language has a tarball which includes the language
486 front end as well as the language runtime (when appropriate).
488 Unpack the core distribution as well as any language specific
489 distributions in the same directory.
491 If you also intend to build binutils (either to upgrade an existing
492 installation or for use in place of the corresponding tools of your
493 OS), unpack the binutils distribution either in the same directory or
494 a separate one. In the latter case, add symbolic links to any
495 components of the binutils you intend to build alongside the compiler
496 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
497 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
504 @uref{./index.html,,Return to the GCC Installation page}
508 @c ***Configuration***********************************************************
510 @comment node-name, next, previous, up
511 @node Configuration, Building, Downloading the source, Installing GCC
515 @chapter Installing GCC: Configuration
517 @cindex Configuration
518 @cindex Installing GCC: Configuration
520 Like most GNU software, GCC must be configured before it can be built.
521 This document describes the recommended configuration procedure
522 for both native and cross targets.
524 We use @var{srcdir} to refer to the toplevel source directory for
525 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
527 If you obtained the sources via SVN, @var{srcdir} must refer to the top
528 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
529 and not its @file{gcc} subdirectory, otherwise the build will fail.
531 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
532 file system, the shell's built-in @command{pwd} command will return
533 temporary pathnames. Using these can lead to various sorts of build
534 problems. To avoid this issue, set the @env{PWDCMD} environment
535 variable to an automounter-aware @command{pwd} command, e.g.,
536 @command{pawd} or @samp{amq -w}, during the configuration and build
539 First, we @strong{highly} recommend that GCC be built into a
540 separate directory than the sources which does @strong{not} reside
541 within the source tree. This is how we generally build GCC; building
542 where @var{srcdir} == @var{objdir} should still work, but doesn't
543 get extensive testing; building where @var{objdir} is a subdirectory
544 of @var{srcdir} is unsupported.
546 If you have previously built GCC in the same directory for a
547 different target machine, do @samp{make distclean} to delete all files
548 that might be invalid. One of the files this deletes is @file{Makefile};
549 if @samp{make distclean} complains that @file{Makefile} does not exist
550 or issues a message like ``don't know how to make distclean'' it probably
551 means that the directory is already suitably clean. However, with the
552 recommended method of building in a separate @var{objdir}, you should
553 simply use a different @var{objdir} for each target.
555 Second, when configuring a native system, either @command{cc} or
556 @command{gcc} must be in your path or you must set @env{CC} in
557 your environment before running configure. Otherwise the configuration
561 Note that the bootstrap compiler and the resulting GCC must be link
562 compatible, else the bootstrap will fail with linker errors about
563 incompatible object file formats. Several multilibed targets are
564 affected by this requirement, see
566 @ref{Specific, host/target specific installation notes}.
569 @uref{specific.html,,host/target specific installation notes}.
578 % @var{srcdir}/configure [@var{options}] [@var{target}]
581 @heading Distributor options
583 If you will be distributing binary versions of GCC, with modifications
584 to the source code, you should use the options described in this
585 section to make clear that your version contains modifications.
588 @item --with-pkgversion=@var{version}
589 Specify a string that identifies your package. You may wish
590 to include a build number or build date. This version string will be
591 included in the output of @command{gcc --version}. This suffix does
592 not replace the default version string, only the @samp{GCC} part.
594 The default value is @samp{GCC}.
596 @item --with-bugurl=@var{url}
597 Specify the URL that users should visit if they wish to report a bug.
598 You are of course welcome to forward bugs reported to you to the FSF,
599 if you determine that they are not bugs in your modifications.
601 The default value refers to the FSF's GCC bug tracker.
605 @heading Target specification
608 GCC has code to correctly determine the correct value for @var{target}
609 for nearly all native systems. Therefore, we highly recommend you not
610 provide a configure target when configuring a native compiler.
613 @var{target} must be specified as @option{--target=@var{target}}
614 when configuring a cross compiler; examples of valid targets would be
615 m68k-coff, sh-elf, etc.
618 Specifying just @var{target} instead of @option{--target=@var{target}}
619 implies that the host defaults to @var{target}.
623 @heading Options specification
625 Use @var{options} to override several configure time options for
626 GCC@. A list of supported @var{options} follows; @samp{configure
627 --help} may list other options, but those not listed below may not
628 work and should not normally be used.
630 Note that each @option{--enable} option has a corresponding
631 @option{--disable} option and that each @option{--with} option has a
632 corresponding @option{--without} option.
635 @item --prefix=@var{dirname}
636 Specify the toplevel installation
637 directory. This is the recommended way to install the tools into a directory
638 other than the default. The toplevel installation directory defaults to
641 We @strong{highly} recommend against @var{dirname} being the same or a
642 subdirectory of @var{objdir} or vice versa. If specifying a directory
643 beneath a user's home directory tree, some shells will not expand
644 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
647 The following standard @command{autoconf} options are supported. Normally you
648 should not need to use these options.
650 @item --exec-prefix=@var{dirname}
651 Specify the toplevel installation directory for architecture-dependent
652 files. The default is @file{@var{prefix}}.
654 @item --bindir=@var{dirname}
655 Specify the installation directory for the executables called by users
656 (such as @command{gcc} and @command{g++}). The default is
657 @file{@var{exec-prefix}/bin}.
659 @item --libdir=@var{dirname}
660 Specify the installation directory for object code libraries and
661 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
663 @item --libexecdir=@var{dirname}
664 Specify the installation directory for internal executables of GCC@.
665 The default is @file{@var{exec-prefix}/libexec}.
667 @item --with-slibdir=@var{dirname}
668 Specify the installation directory for the shared libgcc library. The
669 default is @file{@var{libdir}}.
671 @item --infodir=@var{dirname}
672 Specify the installation directory for documentation in info format.
673 The default is @file{@var{prefix}/info}.
675 @item --datadir=@var{dirname}
676 Specify the installation directory for some architecture-independent
677 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
679 @item --mandir=@var{dirname}
680 Specify the installation directory for manual pages. The default is
681 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
682 the full GCC manuals, which are provided in Texinfo format. The manpages
683 are derived by an automatic conversion process from parts of the full
686 @item --with-gxx-include-dir=@var{dirname}
688 the installation directory for G++ header files. The default is
689 @file{@var{prefix}/include/c++/@var{version}}.
693 @item --program-prefix=@var{prefix}
694 GCC supports some transformations of the names of its programs when
695 installing them. This option prepends @var{prefix} to the names of
696 programs to install in @var{bindir} (see above). For example, specifying
697 @option{--program-prefix=foo-} would result in @samp{gcc}
698 being installed as @file{/usr/local/bin/foo-gcc}.
700 @item --program-suffix=@var{suffix}
701 Appends @var{suffix} to the names of programs to install in @var{bindir}
702 (see above). For example, specifying @option{--program-suffix=-3.1}
703 would result in @samp{gcc} being installed as
704 @file{/usr/local/bin/gcc-3.1}.
706 @item --program-transform-name=@var{pattern}
707 Applies the @samp{sed} script @var{pattern} to be applied to the names
708 of programs to install in @var{bindir} (see above). @var{pattern} has to
709 consist of one or more basic @samp{sed} editing commands, separated by
710 semicolons. For example, if you want the @samp{gcc} program name to be
711 transformed to the installed program @file{/usr/local/bin/myowngcc} and
712 the @samp{g++} program name to be transformed to
713 @file{/usr/local/bin/gspecial++} without changing other program names,
714 you could use the pattern
715 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
716 to achieve this effect.
718 All three options can be combined and used together, resulting in more
719 complex conversion patterns. As a basic rule, @var{prefix} (and
720 @var{suffix}) are prepended (appended) before further transformations
721 can happen with a special transformation script @var{pattern}.
723 As currently implemented, this option only takes effect for native
724 builds; cross compiler binaries' names are not transformed even when a
725 transformation is explicitly asked for by one of these options.
727 For native builds, some of the installed programs are also installed
728 with the target alias in front of their name, as in
729 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
730 before the target alias is prepended to the name---so, specifying
731 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
732 resulting binary would be installed as
733 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
735 As a last shortcoming, none of the installed Ada programs are
736 transformed yet, which will be fixed in some time.
738 @item --with-local-prefix=@var{dirname}
740 installation directory for local include files. The default is
741 @file{/usr/local}. Specify this option if you want the compiler to
742 search directory @file{@var{dirname}/include} for locally installed
743 header files @emph{instead} of @file{/usr/local/include}.
745 You should specify @option{--with-local-prefix} @strong{only} if your
746 site has a different convention (not @file{/usr/local}) for where to put
749 The default value for @option{--with-local-prefix} is @file{/usr/local}
750 regardless of the value of @option{--prefix}. Specifying
751 @option{--prefix} has no effect on which directory GCC searches for
752 local header files. This may seem counterintuitive, but actually it is
755 The purpose of @option{--prefix} is to specify where to @emph{install
756 GCC}. The local header files in @file{/usr/local/include}---if you put
757 any in that directory---are not part of GCC@. They are part of other
758 programs---perhaps many others. (GCC installs its own header files in
759 another directory which is based on the @option{--prefix} value.)
761 Both the local-prefix include directory and the GCC-prefix include
762 directory are part of GCC's ``system include'' directories. Although these
763 two directories are not fixed, they need to be searched in the proper
764 order for the correct processing of the include_next directive. The
765 local-prefix include directory is searched before the GCC-prefix
766 include directory. Another characteristic of system include directories
767 is that pedantic warnings are turned off for headers in these directories.
769 Some autoconf macros add @option{-I @var{directory}} options to the
770 compiler command line, to ensure that directories containing installed
771 packages' headers are searched. When @var{directory} is one of GCC's
772 system include directories, GCC will ignore the option so that system
773 directories continue to be processed in the correct order. This
774 may result in a search order different from what was specified but the
775 directory will still be searched.
777 GCC automatically searches for ordinary libraries using
778 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
779 used for both GCC and packages, GCC will automatically search for
780 both headers and libraries. This provides a configuration that is
781 easy to use. GCC behaves in a manner similar to that when it is
782 installed as a system compiler in @file{/usr}.
784 Sites that need to install multiple versions of GCC may not want to
785 use the above simple configuration. It is possible to use the
786 @option{--program-prefix}, @option{--program-suffix} and
787 @option{--program-transform-name} options to install multiple versions
788 into a single directory, but it may be simpler to use different prefixes
789 and the @option{--with-local-prefix} option to specify the location of the
790 site-specific files for each version. It will then be necessary for
791 users to specify explicitly the location of local site libraries
792 (e.g., with @env{LIBRARY_PATH}).
794 The same value can be used for both @option{--with-local-prefix} and
795 @option{--prefix} provided it is not @file{/usr}. This can be used
796 to avoid the default search of @file{/usr/local/include}.
798 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
799 The directory you use for @option{--with-local-prefix} @strong{must not}
800 contain any of the system's standard header files. If it did contain
801 them, certain programs would be miscompiled (including GNU Emacs, on
802 certain targets), because this would override and nullify the header
803 file corrections made by the @command{fixincludes} script.
805 Indications are that people who use this option use it based on mistaken
806 ideas of what it is for. People use it as if it specified where to
807 install part of GCC@. Perhaps they make this assumption because
808 installing GCC creates the directory.
810 @item --enable-shared[=@var{package}[,@dots{}]]
811 Build shared versions of libraries, if shared libraries are supported on
812 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
813 are enabled by default on all platforms that support shared libraries.
815 If a list of packages is given as an argument, build shared libraries
816 only for the listed packages. For other packages, only static libraries
817 will be built. Package names currently recognized in the GCC tree are
818 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
819 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
820 @samp{ada}, @samp{libada}, @samp{libjava} and @samp{libobjc}.
821 Note @samp{libiberty} does not support shared libraries at all.
823 Use @option{--disable-shared} to build only static libraries. Note that
824 @option{--disable-shared} does not accept a list of package names as
825 argument, only @option{--enable-shared} does.
827 @item @anchor{with-gnu-as}--with-gnu-as
828 Specify that the compiler should assume that the
829 assembler it finds is the GNU assembler. However, this does not modify
830 the rules to find an assembler and will result in confusion if the
831 assembler found is not actually the GNU assembler. (Confusion may also
832 result if the compiler finds the GNU assembler but has not been
833 configured with @option{--with-gnu-as}.) If you have more than one
834 assembler installed on your system, you may want to use this option in
835 connection with @option{--with-as=@var{pathname}} or
836 @option{--with-build-time-tools=@var{pathname}}.
838 The following systems are the only ones where it makes a difference
839 whether you use the GNU assembler. On any other system,
840 @option{--with-gnu-as} has no effect.
843 @item @samp{hppa1.0-@var{any}-@var{any}}
844 @item @samp{hppa1.1-@var{any}-@var{any}}
845 @item @samp{sparc-sun-solaris2.@var{any}}
846 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
849 On the systems listed above (except for the HP-PA, the SPARC, for ISC on
850 the 386, if you use the GNU assembler, you should also use the GNU linker
851 (and specify @option{--with-gnu-ld}).
853 @item @anchor{with-as}--with-as=@var{pathname}
854 Specify that the compiler should use the assembler pointed to by
855 @var{pathname}, rather than the one found by the standard rules to find
856 an assembler, which are:
859 Unless GCC is being built with a cross compiler, check the
860 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
861 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
862 @var{exec-prefix} defaults to @var{prefix}, which
863 defaults to @file{/usr/local} unless overridden by the
864 @option{--prefix=@var{pathname}} switch described above. @var{target}
865 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
866 @var{version} denotes the GCC version, such as 3.0.
869 If the target system is the same that you are building on, check
870 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
874 Check in the @env{PATH} for a tool whose name is prefixed by the
875 target system triple.
878 Check in the @env{PATH} for a tool whose name is not prefixed by the
879 target system triple, if the host and target system triple are
880 the same (in other words, we use a host tool if it can be used for
884 You may want to use @option{--with-as} if no assembler
885 is installed in the directories listed above, or if you have multiple
886 assemblers installed and want to choose one that is not found by the
889 @item @anchor{with-gnu-ld}--with-gnu-ld
890 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
893 @item --with-ld=@var{pathname}
894 Same as @uref{#with-as,,@option{--with-as}}
898 Specify that stabs debugging
899 information should be used instead of whatever format the host normally
900 uses. Normally GCC uses the same debug format as the host system.
902 On MIPS based systems and on Alphas, you must specify whether you want
903 GCC to create the normal ECOFF debugging format, or to use BSD-style
904 stabs passed through the ECOFF symbol table. The normal ECOFF debug
905 format cannot fully handle languages other than C@. BSD stabs format can
906 handle other languages, but it only works with the GNU debugger GDB@.
908 Normally, GCC uses the ECOFF debugging format by default; if you
909 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
911 No matter which default you choose when you configure GCC, the user
912 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
913 the debug format for a particular compilation.
915 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
916 @option{--with-gas} is used. It selects use of stabs debugging
917 information embedded in COFF output. This kind of debugging information
918 supports C++ well; ordinary COFF debugging information does not.
920 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
921 selects use of stabs debugging information embedded in ELF output. The
922 C++ compiler currently (2.6.0) does not support the DWARF debugging
923 information normally used on 386 SVR4 platforms; stabs provide a
924 workable alternative. This requires gas and gdb, as the normal SVR4
925 tools can not generate or interpret stabs.
927 @item --disable-multilib
928 Specify that multiple target
929 libraries to support different target variants, calling
930 conventions, etc.@: should not be built. The default is to build a
931 predefined set of them.
933 Some targets provide finer-grained control over which multilibs are built
934 (e.g., @option{--disable-softfloat}):
940 fpu, 26bit, underscore, interwork, biendian, nofmult.
943 softfloat, m68881, m68000, m68020.
946 single-float, biendian, softfloat.
948 @item powerpc*-*-*, rs6000*-*-*
949 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
954 @item --enable-threads
955 Specify that the target
956 supports threads. This affects the Objective-C compiler and runtime
957 library, and exception handling for other languages like C++ and Java.
958 On some systems, this is the default.
960 In general, the best (and, in many cases, the only known) threading
961 model available will be configured for use. Beware that on some
962 systems, GCC has not been taught what threading models are generally
963 available for the system. In this case, @option{--enable-threads} is an
964 alias for @option{--enable-threads=single}.
966 @item --disable-threads
967 Specify that threading support should be disabled for the system.
968 This is an alias for @option{--enable-threads=single}.
970 @item --enable-threads=@var{lib}
972 @var{lib} is the thread support library. This affects the Objective-C
973 compiler and runtime library, and exception handling for other languages
974 like C++ and Java. The possibilities for @var{lib} are:
982 Ada tasking support. For non-Ada programs, this setting is equivalent
983 to @samp{single}. When used in conjunction with the Ada run time, it
984 causes GCC to use the same thread primitives as Ada uses. This option
985 is necessary when using both Ada and the back end exception handling,
986 which is the default for most Ada targets.
988 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
989 that the file needed to support this configuration, @file{gthr-mach.h}, is
990 missing and thus this setting will cause a known bootstrap failure.)
992 This is an alias for @samp{single}.
994 Generic POSIX/Unix98 thread support.
996 Generic POSIX/Unix95 thread support.
998 RTEMS thread support.
1000 Disable thread support, should work for all platforms.
1002 Sun Solaris 2 thread support.
1004 VxWorks thread support.
1006 Microsoft Win32 API thread support.
1008 Novell Kernel Services thread support.
1012 Specify that the target supports TLS (Thread Local Storage). Usually
1013 configure can correctly determine if TLS is supported. In cases where
1014 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1015 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1016 the assembler supports TLS but the C library does not, or if the
1017 assumptions made by the configure test are incorrect.
1020 Specify that the target does not support TLS.
1021 This is an alias for @option{--enable-tls=no}.
1023 @item --with-cpu=@var{cpu}
1024 @itemx --with-cpu-32=@var{cpu}
1025 @itemx --with-cpu-64=@var{cpu}
1026 Specify which cpu variant the compiler should generate code for by default.
1027 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1028 This option is only supported on some targets, including ARM, i386, M68k,
1029 PowerPC, and SPARC@. The @option{--with-cpu-32} and
1030 @option{--with-cpu-64} options specify separate default CPUs for
1031 32-bit and 64-bit modes; these options are only supported for i386 and
1034 @item --with-schedule=@var{cpu}
1035 @itemx --with-arch=@var{cpu}
1036 @itemx --with-arch-32=@var{cpu}
1037 @itemx --with-arch-64=@var{cpu}
1038 @itemx --with-tune=@var{cpu}
1039 @itemx --with-tune-32=@var{cpu}
1040 @itemx --with-tune-64=@var{cpu}
1041 @itemx --with-abi=@var{abi}
1042 @itemx --with-fpu=@var{type}
1043 @itemx --with-float=@var{type}
1044 These configure options provide default values for the @option{-mschedule=},
1045 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1046 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1047 @option{--with-cpu}, which switches will be accepted and acceptable values
1048 of the arguments depend on the target.
1050 @item --with-mode=@var{mode}
1051 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1052 This option is only supported on ARM targets.
1054 @item --with-divide=@var{type}
1055 Specify how the compiler should generate code for checking for
1056 division by zero. This option is only supported on the MIPS target.
1057 The possibilities for @var{type} are:
1060 Division by zero checks use conditional traps (this is the default on
1061 systems that support conditional traps).
1063 Division by zero checks use the break instruction.
1066 @c If you make --with-llsc the default for additional targets,
1067 @c update the --with-llsc description in the MIPS section below.
1070 On MIPS targets, make @option{-mllsc} the default when no
1071 @option{-mno-lsc} option is passed. This is the default for
1072 Linux-based targets, as the kernel will emulate them if the ISA does
1075 @item --without-llsc
1076 On MIPS targets, make @option{-mno-llsc} the default when no
1077 @option{-mllsc} option is passed.
1079 @item --with-mips-plt
1080 On MIPS targets, make use of copy relocations and PLTs.
1081 These features are extensions to the traditional
1082 SVR4-based MIPS ABIs and require support from GNU binutils
1083 and the runtime C library.
1085 @item --enable-__cxa_atexit
1086 Define if you want to use __cxa_atexit, rather than atexit, to
1087 register C++ destructors for local statics and global objects.
1088 This is essential for fully standards-compliant handling of
1089 destructors, but requires __cxa_atexit in libc. This option is currently
1090 only available on systems with GNU libc. When enabled, this will cause
1091 @option{-fuse-cxa-atexit} to be passed by default.
1093 @item --enable-target-optspace
1095 libraries should be optimized for code space instead of code speed.
1096 This is the default for the m32r platform.
1099 Specify that a user visible @command{cpp} program should not be installed.
1101 @item --with-cpp-install-dir=@var{dirname}
1102 Specify that the user visible @command{cpp} program should be installed
1103 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1105 @item --enable-initfini-array
1106 Force the use of sections @code{.init_array} and @code{.fini_array}
1107 (instead of @code{.init} and @code{.fini}) for constructors and
1108 destructors. Option @option{--disable-initfini-array} has the
1109 opposite effect. If neither option is specified, the configure script
1110 will try to guess whether the @code{.init_array} and
1111 @code{.fini_array} sections are supported and, if they are, use them.
1113 @item --enable-maintainer-mode
1114 The build rules that
1115 regenerate the GCC master message catalog @file{gcc.pot} are normally
1116 disabled. This is because it can only be rebuilt if the complete source
1117 tree is present. If you have changed the sources and want to rebuild the
1118 catalog, configuring with @option{--enable-maintainer-mode} will enable
1119 this. Note that you need a recent version of the @code{gettext} tools
1122 @item --disable-bootstrap
1123 For a native build, the default configuration is to perform
1124 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1125 testing that GCC can compile itself correctly. If you want to disable
1126 this process, you can configure with @option{--disable-bootstrap}.
1128 @item --enable-bootstrap
1129 In special cases, you may want to perform a 3-stage build
1130 even if the target and host triplets are different.
1131 This could happen when the host can run code compiled for
1132 the target (e.g.@: host is i686-linux, target is i486-linux).
1133 Starting from GCC 4.2, to do this you have to configure explicitly
1134 with @option{--enable-bootstrap}.
1136 @item --enable-generated-files-in-srcdir
1137 Neither the .c and .h files that are generated from Bison and flex nor the
1138 info manuals and man pages that are built from the .texi files are present
1139 in the SVN development tree. When building GCC from that development tree,
1140 or from one of our snapshots, those generated files are placed in your
1141 build directory, which allows for the source to be in a readonly
1144 If you configure with @option{--enable-generated-files-in-srcdir} then those
1145 generated files will go into the source directory. This is mainly intended
1146 for generating release or prerelease tarballs of the GCC sources, since it
1147 is not a requirement that the users of source releases to have flex, Bison,
1150 @item --enable-version-specific-runtime-libs
1152 that runtime libraries should be installed in the compiler specific
1153 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1154 addition, @samp{libstdc++}'s include files will be installed into
1155 @file{@var{libdir}} unless you overruled it by using
1156 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1157 particularly useful if you intend to use several versions of GCC in
1158 parallel. This is currently supported by @samp{libgfortran},
1159 @samp{libjava}, @samp{libmudflap}, @samp{libstdc++}, and @samp{libobjc}.
1161 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1162 Specify that only a particular subset of compilers and
1163 their runtime libraries should be built. For a list of valid values for
1164 @var{langN} you can issue the following command in the
1165 @file{gcc} directory of your GCC source tree:@*
1167 grep language= */config-lang.in
1169 Currently, you can use any of the following:
1170 @code{all}, @code{ada}, @code{c}, @code{c++}, @code{fortran}, @code{java},
1171 @code{objc}, @code{obj-c++}.
1172 Building the Ada compiler has special requirements, see below.
1173 If you do not pass this flag, or specify the option @code{all}, then all
1174 default languages available in the @file{gcc} sub-tree will be configured.
1175 Ada and Objective-C++ are not default languages; the rest are.
1176 Re-defining @code{LANGUAGES} when calling @samp{make} @strong{does not}
1177 work anymore, as those language sub-directories might not have been
1180 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1181 Specify that a particular subset of compilers and their runtime
1182 libraries should be built with the system C compiler during stage 1 of
1183 the bootstrap process, rather than only in later stages with the
1184 bootstrapped C compiler. The list of valid values is the same as for
1185 @option{--enable-languages}, and the option @code{all} will select all
1186 of the languages enabled by @option{--enable-languages}. This option is
1187 primarily useful for GCC development; for instance, when a development
1188 version of the compiler cannot bootstrap due to compiler bugs, or when
1189 one is debugging front ends other than the C front end. When this
1190 option is used, one can then build the target libraries for the
1191 specified languages with the stage-1 compiler by using @command{make
1192 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1193 for the specified languages using @command{make stage1-start check-gcc}.
1195 @item --disable-libada
1196 Specify that the run-time libraries and tools used by GNAT should not
1197 be built. This can be useful for debugging, or for compatibility with
1198 previous Ada build procedures, when it was required to explicitly
1199 do a @samp{make -C gcc gnatlib_and_tools}.
1201 @item --disable-libssp
1202 Specify that the run-time libraries for stack smashing protection
1203 should not be built.
1205 @item --disable-libgomp
1206 Specify that the run-time libraries used by GOMP should not be built.
1209 Specify that the compiler should
1210 use DWARF 2 debugging information as the default.
1212 @item --enable-targets=all
1213 @itemx --enable-targets=@var{target_list}
1214 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1215 These are compilers that are able to generate either 64-bit or 32-bit
1216 code. Typically, the corresponding 32-bit target, e.g.@:
1217 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1218 option enables the 32-bit target to be a bi-arch compiler, which is
1219 useful when you want a bi-arch compiler that defaults to 32-bit, and
1220 you are building a bi-arch or multi-arch binutils in a combined tree.
1221 Currently, this option only affects sparc-linux, powerpc-linux and
1224 @item --enable-secureplt
1225 This option enables @option{-msecure-plt} by default for powerpc-linux.
1227 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1228 Using the GNU Compiler Collection (GCC)},
1231 See ``RS/6000 and PowerPC Options'' in the main manual
1235 This option enables @option{-mcld} by default for 32-bit x86 targets.
1237 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1238 Using the GNU Compiler Collection (GCC)},
1241 See ``i386 and x86-64 Options'' in the main manual
1244 @item --enable-win32-registry
1245 @itemx --enable-win32-registry=@var{key}
1246 @itemx --disable-win32-registry
1247 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1248 to look up installations paths in the registry using the following key:
1251 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1254 @var{key} defaults to GCC version number, and can be overridden by the
1255 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1256 who use custom installers are encouraged to provide a different key,
1257 perhaps one comprised of vendor name and GCC version number, to
1258 avoid conflict with existing installations. This feature is enabled
1259 by default, and can be disabled by @option{--disable-win32-registry}
1260 option. This option has no effect on the other hosts.
1263 Specify that the machine does not have a floating point unit. This
1264 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1265 system, @option{--nfp} has no effect.
1267 @item --enable-werror
1268 @itemx --disable-werror
1269 @itemx --enable-werror=yes
1270 @itemx --enable-werror=no
1271 When you specify this option, it controls whether certain files in the
1272 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1273 If you don't specify it, @option{-Werror} is turned on for the main
1274 development trunk. However it defaults to off for release branches and
1275 final releases. The specific files which get @option{-Werror} are
1276 controlled by the Makefiles.
1278 @item --enable-checking
1279 @itemx --enable-checking=@var{list}
1280 When you specify this option, the compiler is built to perform internal
1281 consistency checks of the requested complexity. This does not change the
1282 generated code, but adds error checking within the compiler. This will
1283 slow down the compiler and may only work properly if you are building
1284 the compiler with GCC@. This is @samp{yes} by default when building
1285 from SVN or snapshots, but @samp{release} for releases. More control
1286 over the checks may be had by specifying @var{list}. The categories of
1287 checks available are @samp{yes} (most common checks
1288 @samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at
1289 all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest
1290 checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1291 Individual checks can be enabled with these flags @samp{assert},
1292 @samp{df}, @samp{fold}, @samp{gc}, @samp{gcac} @samp{misc}, @samp{rtl},
1293 @samp{rtlflag}, @samp{runtime}, @samp{tree}, and @samp{valgrind}.
1295 The @samp{valgrind} check requires the external @command{valgrind}
1296 simulator, available from @uref{http://valgrind.org/}. The
1297 @samp{df}, @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive.
1298 To disable all checking, @samp{--disable-checking} or
1299 @samp{--enable-checking=none} must be explicitly requested. Disabling
1300 assertions will make the compiler and runtime slightly faster but
1301 increase the risk of undetected internal errors causing wrong code to be
1304 @item --enable-coverage
1305 @itemx --enable-coverage=@var{level}
1306 With this option, the compiler is built to collect self coverage
1307 information, every time it is run. This is for internal development
1308 purposes, and only works when the compiler is being built with gcc. The
1309 @var{level} argument controls whether the compiler is built optimized or
1310 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1311 want to disable optimization, for performance analysis you want to
1312 enable optimization. When coverage is enabled, the default level is
1313 without optimization.
1315 @item --enable-gather-detailed-mem-stats
1316 When this option is specified more detailed information on memory
1317 allocation is gathered. This information is printed when using
1318 @option{-fmem-report}.
1321 @itemx --with-gc=@var{choice}
1322 With this option you can specify the garbage collector implementation
1323 used during the compilation process. @var{choice} can be one of
1324 @samp{page} and @samp{zone}, where @samp{page} is the default.
1327 @itemx --disable-nls
1328 The @option{--enable-nls} option enables Native Language Support (NLS),
1329 which lets GCC output diagnostics in languages other than American
1330 English. Native Language Support is enabled by default if not doing a
1331 canadian cross build. The @option{--disable-nls} option disables NLS@.
1333 @item --with-included-gettext
1334 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1335 procedure to prefer its copy of GNU @command{gettext}.
1337 @item --with-catgets
1338 If NLS is enabled, and if the host lacks @code{gettext} but has the
1339 inferior @code{catgets} interface, the GCC build procedure normally
1340 ignores @code{catgets} and instead uses GCC's copy of the GNU
1341 @code{gettext} library. The @option{--with-catgets} option causes the
1342 build procedure to use the host's @code{catgets} in this situation.
1344 @item --with-libiconv-prefix=@var{dir}
1345 Search for libiconv header files in @file{@var{dir}/include} and
1346 libiconv library files in @file{@var{dir}/lib}.
1348 @item --enable-obsolete
1349 Enable configuration for an obsoleted system. If you attempt to
1350 configure GCC for a system (build, host, or target) which has been
1351 obsoleted, and you do not specify this flag, configure will halt with an
1354 All support for systems which have been obsoleted in one release of GCC
1355 is removed entirely in the next major release, unless someone steps
1356 forward to maintain the port.
1358 @item --enable-decimal-float
1359 @itemx --enable-decimal-float=yes
1360 @itemx --enable-decimal-float=no
1361 @itemx --enable-decimal-float=bid
1362 @itemx --enable-decimal-float=dpd
1363 @itemx --disable-decimal-float
1364 Enable (or disable) support for the C decimal floating point extension
1365 that is in the IEEE 754-2008 standard. This is enabled by default only
1366 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
1367 support it, but require the user to specifically enable it. You can
1368 optionally control which decimal floating point format is used (either
1369 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
1370 format is default on i386 and x86_64 systems, and the @samp{dpd}
1371 (densely packed decimal) format is default on PowerPC systems.
1373 @item --enable-fixed-point
1374 @itemx --disable-fixed-point
1375 Enable (or disable) support for C fixed-point arithmetic.
1376 This option is enabled by default for some targets (such as MIPS) which
1377 have hardware-support for fixed-point operations. On other targets, you
1378 may enable this option manually.
1380 @item --with-long-double-128
1381 Specify if @code{long double} type should be 128-bit by default on selected
1382 GNU/Linux architectures. If using @code{--without-long-double-128},
1383 @code{long double} will be by default 64-bit, the same as @code{double} type.
1384 When neither of these configure options are used, the default will be
1385 128-bit @code{long double} when built against GNU C Library 2.4 and later,
1386 64-bit @code{long double} otherwise.
1388 @item --with-gmp=@var{pathname}
1389 @itemx --with-gmp-include=@var{pathname}
1390 @itemx --with-gmp-lib=@var{pathname}
1391 @itemx --with-mpfr=@var{pathname}
1392 @itemx --with-mpfr-include=@var{pathname}
1393 @itemx --with-mpfr-lib=@var{pathname}
1394 If you do not have GMP (the GNU Multiple Precision library) and the
1395 MPFR Libraries installed in a standard location and you want to build
1396 GCC, you can explicitly specify the directory where they are installed
1397 (@samp{--with-gmp=@var{gmpinstalldir}},
1398 @samp{--with-mpfr=@var{mpfrinstalldir}}). The
1399 @option{--with-gmp=@var{gmpinstalldir}} option is shorthand for
1400 @option{--with-gmp-lib=@var{gmpinstalldir}/lib} and
1401 @option{--with-gmp-include=@var{gmpinstalldir}/include}. Likewise the
1402 @option{--with-mpfr=@var{mpfrinstalldir}} option is shorthand for
1403 @option{--with-mpfr-lib=@var{mpfrinstalldir}/lib} and
1404 @option{--with-mpfr-include=@var{mpfrinstalldir}/include}. If these
1405 shorthand assumptions are not correct, you can use the explicit
1406 include and lib options directly.
1408 @item --with-debug-prefix-map=@var{map}
1409 Convert source directory names using @option{-fdebug-prefix-map} when
1410 building runtime libraries. @samp{@var{map}} is a space-separated
1411 list of maps of the form @samp{@var{old}=@var{new}}.
1415 @subheading Cross-Compiler-Specific Options
1416 The following options only apply to building cross compilers.
1418 @item --with-sysroot
1419 @itemx --with-sysroot=@var{dir}
1420 Tells GCC to consider @var{dir} as the root of a tree that contains a
1421 (subset of) the root filesystem of the target operating system.
1422 Target system headers, libraries and run-time object files will be
1423 searched in there. The specified directory is not copied into the
1424 install tree, unlike the options @option{--with-headers} and
1425 @option{--with-libs} that this option obsoletes. The default value,
1426 in case @option{--with-sysroot} is not given an argument, is
1427 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1428 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1429 the GCC binaries if the installation tree is moved.
1431 @item --with-build-sysroot
1432 @itemx --with-build-sysroot=@var{dir}
1433 Tells GCC to consider @var{dir} as the system root (see
1434 @option{--with-sysroot}) while building target libraries, instead of
1435 the directory specified with @option{--with-sysroot}. This option is
1436 only useful when you are already using @option{--with-sysroot}. You
1437 can use @option{--with-build-sysroot} when you are configuring with
1438 @option{--prefix} set to a directory that is different from the one in
1439 which you are installing GCC and your target libraries.
1441 This option affects the system root for the compiler used to build
1442 target libraries (which runs on the build system); it does not affect
1443 the compiler which is used to build GCC itself.
1445 @item --with-headers
1446 @itemx --with-headers=@var{dir}
1447 Deprecated in favor of @option{--with-sysroot}.
1448 Specifies that target headers are available when building a cross compiler.
1449 The @var{dir} argument specifies a directory which has the target include
1450 files. These include files will be copied into the @file{gcc} install
1451 directory. @emph{This option with the @var{dir} argument is required} when
1452 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1453 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1454 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1455 will be run on these files to make them compatible with GCC@.
1457 @item --without-headers
1458 Tells GCC not use any target headers from a libc when building a cross
1459 compiler. When crossing to GNU/Linux, you need the headers so GCC
1460 can build the exception handling for libgcc.
1463 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1464 Deprecated in favor of @option{--with-sysroot}.
1465 Specifies a list of directories which contain the target runtime
1466 libraries. These libraries will be copied into the @file{gcc} install
1467 directory. If the directory list is omitted, this option has no
1471 Specifies that @samp{newlib} is
1472 being used as the target C library. This causes @code{__eprintf} to be
1473 omitted from @file{libgcc.a} on the assumption that it will be provided by
1476 @item --with-build-time-tools=@var{dir}
1477 Specifies where to find the set of target tools (assembler, linker, etc.)
1478 that will be used while building GCC itself. This option can be useful
1479 if the directory layouts are different between the system you are building
1480 GCC on, and the system where you will deploy it.
1482 For example, on a @option{ia64-hp-hpux} system, you may have the GNU
1483 assembler and linker in @file{/usr/bin}, and the native tools in a
1484 different path, and build a toolchain that expects to find the
1485 native tools in @file{/usr/bin}.
1487 When you use this option, you should ensure that @var{dir} includes
1488 @command{ar}, @command{as}, @command{ld}, @command{nm},
1489 @command{ranlib} and @command{strip} if necessary, and possibly
1490 @command{objdump}. Otherwise, GCC may use an inconsistent set of
1494 @subheading Java-Specific Options
1496 The following option applies to the build of the Java front end.
1499 @item --disable-libgcj
1500 Specify that the run-time libraries
1501 used by GCJ should not be built. This is useful in case you intend
1502 to use GCJ with some other run-time, or you're going to install it
1503 separately, or it just happens not to build on your particular
1504 machine. In general, if the Java front end is enabled, the GCJ
1505 libraries will be enabled too, unless they're known to not work on
1506 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1507 may need to port it; in this case, before modifying the top-level
1508 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1509 you may use @option{--enable-libgcj} to override the default.
1513 The following options apply to building @samp{libgcj}.
1515 @subsubheading General Options
1518 @item --enable-java-maintainer-mode
1519 By default the @samp{libjava} build will not attempt to compile the
1520 @file{.java} source files to @file{.class}. Instead, it will use the
1521 @file{.class} files from the source tree. If you use this option you
1522 must have executables named @command{ecj1} and @command{gjavah} in your path
1523 for use by the build. You must use this option if you intend to
1524 modify any @file{.java} files in @file{libjava}.
1526 @item --with-java-home=@var{dirname}
1527 This @samp{libjava} option overrides the default value of the
1528 @samp{java.home} system property. It is also used to set
1529 @samp{sun.boot.class.path} to @file{@var{dirname}/lib/rt.jar}. By
1530 default @samp{java.home} is set to @file{@var{prefix}} and
1531 @samp{sun.boot.class.path} to
1532 @file{@var{datadir}/java/libgcj-@var{version}.jar}.
1534 @item --with-ecj-jar=@var{filename}
1535 This option can be used to specify the location of an external jar
1536 file containing the Eclipse Java compiler. A specially modified
1537 version of this compiler is used by @command{gcj} to parse
1538 @file{.java} source files. If this option is given, the
1539 @samp{libjava} build will create and install an @file{ecj1} executable
1540 which uses this jar file at runtime.
1542 If this option is not given, but an @file{ecj.jar} file is found in
1543 the topmost source tree at configure time, then the @samp{libgcj}
1544 build will create and install @file{ecj1}, and will also install the
1545 discovered @file{ecj.jar} into a suitable place in the install tree.
1547 If @file{ecj1} is not installed, then the user will have to supply one
1548 on his path in order for @command{gcj} to properly parse @file{.java}
1549 source files. A suitable jar is available from
1550 @uref{ftp://sourceware.org/pub/java/}.
1552 @item --disable-getenv-properties
1553 Don't set system properties from @env{GCJ_PROPERTIES}.
1555 @item --enable-hash-synchronization
1556 Use a global hash table for monitor locks. Ordinarily,
1557 @samp{libgcj}'s @samp{configure} script automatically makes
1558 the correct choice for this option for your platform. Only use
1559 this if you know you need the library to be configured differently.
1561 @item --enable-interpreter
1562 Enable the Java interpreter. The interpreter is automatically
1563 enabled by default on all platforms that support it. This option
1564 is really only useful if you want to disable the interpreter
1565 (using @option{--disable-interpreter}).
1567 @item --disable-java-net
1568 Disable java.net. This disables the native part of java.net only,
1569 using non-functional stubs for native method implementations.
1571 @item --disable-jvmpi
1572 Disable JVMPI support.
1574 @item --disable-libgcj-bc
1575 Disable BC ABI compilation of certain parts of libgcj. By default,
1576 some portions of libgcj are compiled with @option{-findirect-dispatch}
1577 and @option{-fno-indirect-classes}, allowing them to be overridden at
1580 If @option{--disable-libgcj-bc} is specified, libgcj is built without
1581 these options. This allows the compile-time linker to resolve
1582 dependencies when statically linking to libgcj. However it makes it
1583 impossible to override the affected portions of libgcj at run-time.
1585 @item --enable-reduced-reflection
1586 Build most of libgcj with @option{-freduced-reflection}. This reduces
1587 the size of libgcj at the expense of not being able to do accurate
1588 reflection on the classes it contains. This option is safe if you
1589 know that code using libgcj will never use reflection on the standard
1590 runtime classes in libgcj (including using serialization, RMI or CORBA).
1593 Enable runtime eCos target support.
1595 @item --without-libffi
1596 Don't use @samp{libffi}. This will disable the interpreter and JNI
1597 support as well, as these require @samp{libffi} to work.
1599 @item --enable-libgcj-debug
1600 Enable runtime debugging code.
1602 @item --enable-libgcj-multifile
1603 If specified, causes all @file{.java} source files to be
1604 compiled into @file{.class} files in one invocation of
1605 @samp{gcj}. This can speed up build time, but is more
1606 resource-intensive. If this option is unspecified or
1607 disabled, @samp{gcj} is invoked once for each @file{.java}
1608 file to compile into a @file{.class} file.
1610 @item --with-libiconv-prefix=DIR
1611 Search for libiconv in @file{DIR/include} and @file{DIR/lib}.
1613 @item --enable-sjlj-exceptions
1614 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1615 @samp{configure} ordinarily picks the correct value based on the platform.
1616 Only use this option if you are sure you need a different setting.
1618 @item --with-system-zlib
1619 Use installed @samp{zlib} rather than that included with GCC@.
1621 @item --with-win32-nlsapi=ansi, unicows or unicode
1622 Indicates how MinGW @samp{libgcj} translates between UNICODE
1623 characters and the Win32 API@.
1625 @item --enable-java-home
1626 If enabled, this creates a JPackage compatible SDK environment during install.
1627 Note that if --enable-java-home is used, --with-arch-directory=ARCH must also
1630 @item --with-arch-directory=ARCH
1631 Specifies the name to use for the @file{jre/lib/ARCH} directory in the SDK
1632 environment created when --enable-java-home is passed. Typical names for this
1633 directory include i386, amd64, ia64, etc.
1635 @item --with-os-directory=DIR
1636 Specifies the OS directory for the SDK include directory. This is set to auto
1637 detect, and is typically 'linux'.
1639 @item --with-origin-name=NAME
1640 Specifies the JPackage origin name. This defaults to the 'gcj' in
1643 @item --with-arch-suffix=SUFFIX
1644 Specifies the suffix for the sdk directory. Defaults to the empty string.
1645 Examples include '.x86_64' in 'java-1.5.0-gcj-1.5.0.0.x86_64'.
1647 @item --with-jvm-root-dir=DIR
1648 Specifies where to install the SDK. Default is $(prefix)/lib/jvm.
1650 @item --with-jvm-jar-dir=DIR
1651 Specifies where to install jars. Default is $(prefix)/lib/jvm-exports.
1653 @item --with-python-dir=DIR
1654 Specifies where to install the Python modules used for aot-compile. DIR should
1655 not include the prefix used in installation. For example, if the Python modules
1656 are to be installed in /usr/lib/python2.5/site-packages, then
1657 --with-python-dir=/lib/python2.5/site-packages should be passed. If this is
1658 not specified, then the Python modules are installed in $(prefix)/share/python.
1660 @item --enable-aot-compile-rpm
1661 Adds aot-compile-rpm to the list of installed scripts.
1665 Use the single-byte @code{char} and the Win32 A functions natively,
1666 translating to and from UNICODE when using these functions. If
1667 unspecified, this is the default.
1670 Use the @code{WCHAR} and Win32 W functions natively. Adds
1671 @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}.
1672 @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines
1673 running built executables. @file{libunicows.a}, an open-source
1674 import library around Microsoft's @code{unicows.dll}, is obtained from
1675 @uref{http://libunicows.sourceforge.net/}, which also gives details
1676 on getting @file{unicows.dll} from Microsoft.
1679 Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not}
1680 add @code{-lunicows} to @file{libgcj.spec}. The built executables will
1681 only run on Microsoft Windows NT and above.
1685 @subsubheading AWT-Specific Options
1689 Use the X Window System.
1691 @item --enable-java-awt=PEER(S)
1692 Specifies the AWT peer library or libraries to build alongside
1693 @samp{libgcj}. If this option is unspecified or disabled, AWT
1694 will be non-functional. Current valid values are @option{gtk} and
1695 @option{xlib}. Multiple libraries should be separated by a
1696 comma (i.e.@: @option{--enable-java-awt=gtk,xlib}).
1698 @item --enable-gtk-cairo
1699 Build the cairo Graphics2D implementation on GTK@.
1701 @item --enable-java-gc=TYPE
1702 Choose garbage collector. Defaults to @option{boehm} if unspecified.
1704 @item --disable-gtktest
1705 Do not try to compile and run a test GTK+ program.
1707 @item --disable-glibtest
1708 Do not try to compile and run a test GLIB program.
1710 @item --with-libart-prefix=PFX
1711 Prefix where libart is installed (optional).
1713 @item --with-libart-exec-prefix=PFX
1714 Exec prefix where libart is installed (optional).
1716 @item --disable-libarttest
1717 Do not try to compile and run a test libart program.
1726 @uref{./index.html,,Return to the GCC Installation page}
1730 @c ***Building****************************************************************
1732 @comment node-name, next, previous, up
1733 @node Building, Testing, Configuration, Installing GCC
1739 @cindex Installing GCC: Building
1741 Now that GCC is configured, you are ready to build the compiler and
1744 Some commands executed when making the compiler may fail (return a
1745 nonzero status) and be ignored by @command{make}. These failures, which
1746 are often due to files that were not found, are expected, and can safely
1749 It is normal to have compiler warnings when compiling certain files.
1750 Unless you are a GCC developer, you can generally ignore these warnings
1751 unless they cause compilation to fail. Developers should attempt to fix
1752 any warnings encountered, however they can temporarily continue past
1753 warnings-as-errors by specifying the configure flag
1754 @option{--disable-werror}.
1756 On certain old systems, defining certain environment variables such as
1757 @env{CC} can interfere with the functioning of @command{make}.
1759 If you encounter seemingly strange errors when trying to build the
1760 compiler in a directory other than the source directory, it could be
1761 because you have previously configured the compiler in the source
1762 directory. Make sure you have done all the necessary preparations.
1764 If you build GCC on a BSD system using a directory stored in an old System
1765 V file system, problems may occur in running @command{fixincludes} if the
1766 System V file system doesn't support symbolic links. These problems
1767 result in a failure to fix the declaration of @code{size_t} in
1768 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1769 that type mismatches occur, this could be the cause.
1771 The solution is not to use such a directory for building GCC@.
1773 Similarly, when building from SVN or snapshots, or if you modify
1774 @file{*.l} files, you need the Flex lexical analyzer generator
1775 installed. If you do not modify @file{*.l} files, releases contain
1776 the Flex-generated files and you do not need Flex installed to build
1777 them. There is still one Flex-based lexical analyzer (part of the
1778 build machinery, not of GCC itself) that is used even if you only
1779 build the C front end.
1781 When building from SVN or snapshots, or if you modify Texinfo
1782 documentation, you need version 4.7 or later of Texinfo installed if you
1783 want Info documentation to be regenerated. Releases contain Info
1784 documentation pre-built for the unmodified documentation in the release.
1786 @section Building a native compiler
1788 For a native build, the default configuration is to perform
1789 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
1790 This will build the entire GCC system and ensure that it compiles
1791 itself correctly. It can be disabled with the @option{--disable-bootstrap}
1792 parameter to @samp{configure}, but bootstrapping is suggested because
1793 the compiler will be tested more completely and could also have
1796 The bootstrapping process will complete the following steps:
1800 Build tools necessary to build the compiler.
1803 Perform a 3-stage bootstrap of the compiler. This includes building
1804 three times the target tools for use by the compiler such as binutils
1805 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
1806 individually linked or moved into the top level GCC source tree before
1810 Perform a comparison test of the stage2 and stage3 compilers.
1813 Build runtime libraries using the stage3 compiler from the previous step.
1817 If you are short on disk space you might consider @samp{make
1818 bootstrap-lean} instead. The sequence of compilation is the
1819 same described above, but object files from the stage1 and
1820 stage2 of the 3-stage bootstrap of the compiler are deleted as
1821 soon as they are no longer needed.
1823 If you wish to use non-default GCC flags when compiling the stage2
1824 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
1825 doing @samp{make}. For example, if you want to save additional space
1826 during the bootstrap and in the final installation as well, you can
1827 build the compiler binaries without debugging information as in the
1828 following example. This will save roughly 40% of disk space both for
1829 the bootstrap and the final installation. (Libraries will still contain
1830 debugging information.)
1833 make BOOT_CFLAGS='-O' bootstrap
1836 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
1837 are less well tested here than the default of @samp{-g -O2}, but should
1838 still work. In a few cases, you may find that you need to specify special
1839 flags such as @option{-msoft-float} here to complete the bootstrap; or,
1840 if the native compiler miscompiles the stage1 compiler, you may need
1841 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
1842 of the stage1 compiler that were miscompiled, or by using @samp{make
1843 bootstrap4} to increase the number of stages of bootstrap.
1845 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
1846 Since these are always compiled with the compiler currently being
1847 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
1848 compilation flags, as for non-bootstrapped target libraries.
1849 Again, if the native compiler miscompiles the stage1 compiler, you may
1850 need to work around this by avoiding non-working parts of the stage1
1851 compiler. Use @code{STAGE1_LIBCFLAGS} to this end.
1853 If you used the flag @option{--enable-languages=@dots{}} to restrict
1854 the compilers to be built, only those you've actually enabled will be
1855 built. This will of course only build those runtime libraries, for
1856 which the particular compiler has been built. Please note,
1857 that re-defining @env{LANGUAGES} when calling @samp{make}
1858 @strong{does not} work anymore!
1860 If the comparison of stage2 and stage3 fails, this normally indicates
1861 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1862 a potentially serious bug which you should investigate and report. (On
1863 a few systems, meaningful comparison of object files is impossible; they
1864 always appear ``different''. If you encounter this problem, you will
1865 need to disable comparison in the @file{Makefile}.)
1867 If you do not want to bootstrap your compiler, you can configure with
1868 @option{--disable-bootstrap}. In particular cases, you may want to
1869 bootstrap your compiler even if the target system is not the same as
1870 the one you are building on: for example, you could build a
1871 @code{powerpc-unknown-linux-gnu} toolchain on a
1872 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
1873 @option{--enable-bootstrap} to the configure script.
1876 @section Building a cross compiler
1878 When building a cross compiler, it is not generally possible to do a
1879 3-stage bootstrap of the compiler. This makes for an interesting problem
1880 as parts of GCC can only be built with GCC@.
1882 To build a cross compiler, we first recommend building and installing a
1883 native compiler. You can then use the native GCC compiler to build the
1884 cross compiler. The installed native compiler needs to be GCC version
1887 If the cross compiler is to be built with support for the Java
1888 programming language and the ability to compile .java source files is
1889 desired, the installed native compiler used to build the cross
1890 compiler needs to be the same GCC version as the cross compiler. In
1891 addition the cross compiler needs to be configured with
1892 @option{--with-ecj-jar=@dots{}}.
1894 Assuming you have already installed a native copy of GCC and configured
1895 your cross compiler, issue the command @command{make}, which performs the
1900 Build host tools necessary to build the compiler.
1903 Build target tools for use by the compiler such as binutils (bfd,
1904 binutils, gas, gprof, ld, and opcodes)
1905 if they have been individually linked or moved into the top level GCC source
1906 tree before configuring.
1909 Build the compiler (single stage only).
1912 Build runtime libraries using the compiler from the previous step.
1915 Note that if an error occurs in any step the make process will exit.
1917 If you are not building GNU binutils in the same source tree as GCC,
1918 you will need a cross-assembler and cross-linker installed before
1919 configuring GCC@. Put them in the directory
1920 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
1921 you should put in this directory:
1925 This should be the cross-assembler.
1928 This should be the cross-linker.
1931 This should be the cross-archiver: a program which can manipulate
1932 archive files (linker libraries) in the target machine's format.
1935 This should be a program to construct a symbol table in an archive file.
1938 The installation of GCC will find these programs in that directory,
1939 and copy or link them to the proper place to for the cross-compiler to
1940 find them when run later.
1942 The easiest way to provide these files is to build the Binutils package.
1943 Configure it with the same @option{--host} and @option{--target}
1944 options that you use for configuring GCC, then build and install
1945 them. They install their executables automatically into the proper
1946 directory. Alas, they do not support all the targets that GCC
1949 If you are not building a C library in the same source tree as GCC,
1950 you should also provide the target libraries and headers before
1951 configuring GCC, specifying the directories with
1952 @option{--with-sysroot} or @option{--with-headers} and
1953 @option{--with-libs}. Many targets also require ``start files'' such
1954 as @file{crt0.o} and
1955 @file{crtn.o} which are linked into each executable. There may be several
1956 alternatives for @file{crt0.o}, for use with profiling or other
1957 compilation options. Check your target's definition of
1958 @code{STARTFILE_SPEC} to find out what start files it uses.
1960 @section Building in parallel
1962 GNU Make 3.79 and above, which is necessary to build GCC, support
1963 building in parallel. To activate this, you can use @samp{make -j 2}
1964 instead of @samp{make}. You can also specify a bigger number, and
1965 in most cases using a value greater than the number of processors in
1966 your machine will result in fewer and shorter I/O latency hits, thus
1967 improving overall throughput; this is especially true for slow drives
1968 and network filesystems.
1970 @section Building the Ada compiler
1972 In order to build GNAT, the Ada compiler, you need a working GNAT
1973 compiler (GCC version 3.4 or later).
1974 This includes GNAT tools such as @command{gnatmake} and
1975 @command{gnatlink}, since the Ada front end is written in Ada and
1976 uses some GNAT-specific extensions.
1978 In order to build a cross compiler, it is suggested to install
1979 the new compiler as native first, and then use it to build the cross
1982 @command{configure} does not test whether the GNAT installation works
1983 and has a sufficiently recent version; if too old a GNAT version is
1984 installed, the build will fail unless @option{--enable-languages} is
1985 used to disable building the Ada front end.
1987 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
1988 must not be set when building the Ada compiler, the Ada tools, or the
1989 Ada runtime libraries. You can check that your build environment is clean
1990 by verifying that @samp{gnatls -v} lists only one explicit path in each
1993 @section Building with profile feedback
1995 It is possible to use profile feedback to optimize the compiler itself. This
1996 should result in a faster compiler binary. Experiments done on x86 using gcc
1997 3.3 showed approximately 7 percent speedup on compiling C programs. To
1998 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
2000 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
2001 compiler. This compiler is used to build a @code{stageprofile} compiler
2002 instrumented to collect execution counts of instruction and branch
2003 probabilities. Then runtime libraries are compiled with profile collected.
2004 Finally a @code{stagefeedback} compiler is built using the information collected.
2006 Unlike standard bootstrap, several additional restrictions apply. The
2007 compiler used to build @code{stage1} needs to support a 64-bit integral type.
2008 It is recommended to only use GCC for this. Also parallel make is currently
2009 not supported since collisions in profile collecting may occur.
2016 @uref{./index.html,,Return to the GCC Installation page}
2020 @c ***Testing*****************************************************************
2022 @comment node-name, next, previous, up
2023 @node Testing, Final install, Building, Installing GCC
2027 @chapter Installing GCC: Testing
2030 @cindex Installing GCC: Testing
2033 Before you install GCC, we encourage you to run the testsuites and to
2034 compare your results with results from a similar configuration that have
2035 been submitted to the
2036 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
2037 Some of these archived results are linked from the build status lists
2038 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
2039 reports a successful build runs the testsuites and submits the results.
2040 This step is optional and may require you to download additional software,
2041 but it can give you confidence in your new GCC installation or point out
2042 problems before you install and start using your new GCC@.
2044 First, you must have @uref{download.html,,downloaded the testsuites}.
2045 These are part of the full distribution, but if you downloaded the
2046 ``core'' compiler plus any front ends, you must download the testsuites
2049 Second, you must have the testing tools installed. This includes
2050 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
2051 the DejaGnu site has links to these.
2053 If the directories where @command{runtest} and @command{expect} were
2054 installed are not in the @env{PATH}, you may need to set the following
2055 environment variables appropriately, as in the following example (which
2056 assumes that DejaGnu has been installed under @file{/usr/local}):
2059 TCL_LIBRARY = /usr/local/share/tcl8.0
2060 DEJAGNULIBS = /usr/local/share/dejagnu
2063 (On systems such as Cygwin, these paths are required to be actual
2064 paths, not mounts or links; presumably this is due to some lack of
2065 portability in the DejaGnu code.)
2068 Finally, you can run the testsuite (which may take a long time):
2070 cd @var{objdir}; make -k check
2073 This will test various components of GCC, such as compiler
2074 front ends and runtime libraries. While running the testsuite, DejaGnu
2075 might emit some harmless messages resembling
2076 @samp{WARNING: Couldn't find the global config file.} or
2077 @samp{WARNING: Couldn't find tool init file} that can be ignored.
2079 If you are testing a cross-compiler, you may want to run the testsuite
2080 on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}.
2082 @section How can you run the testsuite on selected tests?
2084 In order to run sets of tests selectively, there are targets
2085 @samp{make check-gcc} and @samp{make check-g++}
2086 in the @file{gcc} subdirectory of the object directory. You can also
2087 just run @samp{make check} in a subdirectory of the object directory.
2090 A more selective way to just run all @command{gcc} execute tests in the
2094 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
2097 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
2098 the testsuite with filenames matching @samp{9805*}, you would use
2101 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
2104 The @file{*.exp} files are located in the testsuite directories of the GCC
2105 source, the most important ones being @file{compile.exp},
2106 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
2107 To get a list of the possible @file{*.exp} files, pipe the
2108 output of @samp{make check} into a file and look at the
2109 @samp{Running @dots{} .exp} lines.
2111 @section Passing options and running multiple testsuites
2113 You can pass multiple options to the testsuite using the
2114 @samp{--target_board} option of DejaGNU, either passed as part of
2115 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
2116 work outside the makefiles. For example,
2119 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
2122 will run the standard @command{g++} testsuites (``unix'' is the target name
2123 for a standard native testsuite situation), passing
2124 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
2125 slashes separate options.
2127 You can run the testsuites multiple times using combinations of options
2128 with a syntax similar to the brace expansion of popular shells:
2131 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
2134 (Note the empty option caused by the trailing comma in the final group.)
2135 The following will run each testsuite eight times using the @samp{arm-sim}
2136 target, as if you had specified all possible combinations yourself:
2139 --target_board=arm-sim/-mhard-float/-O1
2140 --target_board=arm-sim/-mhard-float/-O2
2141 --target_board=arm-sim/-mhard-float/-O3
2142 --target_board=arm-sim/-mhard-float
2143 --target_board=arm-sim/-msoft-float/-O1
2144 --target_board=arm-sim/-msoft-float/-O2
2145 --target_board=arm-sim/-msoft-float/-O3
2146 --target_board=arm-sim/-msoft-float
2149 They can be combined as many times as you wish, in arbitrary ways. This
2153 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
2156 will generate four combinations, all involving @samp{-Wextra}.
2158 The disadvantage to this method is that the testsuites are run in serial,
2159 which is a waste on multiprocessor systems. For users with GNU Make and
2160 a shell which performs brace expansion, you can run the testsuites in
2161 parallel by having the shell perform the combinations and @command{make}
2162 do the parallel runs. Instead of using @samp{--target_board}, use a
2163 special makefile target:
2166 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
2172 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
2175 will run three concurrent ``make-gcc'' testsuites, eventually testing all
2176 ten combinations as described above. Note that this is currently only
2177 supported in the @file{gcc} subdirectory. (To see how this works, try
2178 typing @command{echo} before the example given here.)
2181 @section Additional testing for Java Class Libraries
2183 The Java runtime tests can be executed via @samp{make check}
2184 in the @file{@var{target}/libjava/testsuite} directory in
2187 The @uref{http://sourceware.org/mauve/,,Mauve Project} provides
2188 a suite of tests for the Java Class Libraries. This suite can be run
2189 as part of libgcj testing by placing the Mauve tree within the libjava
2190 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
2191 specifying the location of that tree when invoking @samp{make}, as in
2192 @samp{make MAUVEDIR=~/mauve check}.
2194 @section How to interpret test results
2196 The result of running the testsuite are various @file{*.sum} and @file{*.log}
2197 files in the testsuite subdirectories. The @file{*.log} files contain a
2198 detailed log of the compiler invocations and the corresponding
2199 results, the @file{*.sum} files summarize the results. These summaries
2200 contain status codes for all tests:
2204 PASS: the test passed as expected
2206 XPASS: the test unexpectedly passed
2208 FAIL: the test unexpectedly failed
2210 XFAIL: the test failed as expected
2212 UNSUPPORTED: the test is not supported on this platform
2214 ERROR: the testsuite detected an error
2216 WARNING: the testsuite detected a possible problem
2219 It is normal for some tests to report unexpected failures. At the
2220 current time the testing harness does not allow fine grained control
2221 over whether or not a test is expected to fail. This problem should
2222 be fixed in future releases.
2225 @section Submitting test results
2227 If you want to report the results to the GCC project, use the
2228 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
2231 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
2232 -m gcc-testresults@@gcc.gnu.org |sh
2235 This script uses the @command{Mail} program to send the results, so
2236 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
2237 prepended to the testsuite summary and should contain any special
2238 remarks you have on your results or your build environment. Please
2239 do not edit the testsuite result block or the subject line, as these
2240 messages may be automatically processed.
2247 @uref{./index.html,,Return to the GCC Installation page}
2251 @c ***Final install***********************************************************
2253 @comment node-name, next, previous, up
2254 @node Final install, , Testing, Installing GCC
2256 @ifset finalinstallhtml
2258 @chapter Installing GCC: Final installation
2261 Now that GCC has been built (and optionally tested), you can install it with
2263 cd @var{objdir}; make install
2266 We strongly recommend to install into a target directory where there is
2267 no previous version of GCC present. Also, the GNAT runtime should not
2268 be stripped, as this would break certain features of the debugger that
2269 depend on this debugging information (catching Ada exceptions for
2272 That step completes the installation of GCC; user level binaries can
2273 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
2274 you specified with the @option{--prefix} to configure (or
2275 @file{/usr/local} by default). (If you specified @option{--bindir},
2276 that directory will be used instead; otherwise, if you specified
2277 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
2278 Headers for the C++ and Java libraries are installed in
2279 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
2280 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
2281 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
2282 in info format in @file{@var{infodir}} (normally
2283 @file{@var{prefix}/info}).
2285 When installing cross-compilers, GCC's executables
2286 are not only installed into @file{@var{bindir}}, that
2287 is, @file{@var{exec-prefix}/bin}, but additionally into
2288 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
2289 exists. Typically, such @dfn{tooldirs} hold target-specific
2290 binutils, including assembler and linker.
2292 Installation into a temporary staging area or into a @command{chroot}
2293 jail can be achieved with the command
2296 make DESTDIR=@var{path-to-rootdir} install
2299 @noindent where @var{path-to-rootdir} is the absolute path of
2300 a directory relative to which all installation paths will be
2301 interpreted. Note that the directory specified by @code{DESTDIR}
2302 need not exist yet; it will be created if necessary.
2304 There is a subtle point with tooldirs and @code{DESTDIR}:
2305 If you relocate a cross-compiler installation with
2306 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
2307 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
2308 be filled with duplicated GCC executables only if it already exists,
2309 it will not be created otherwise. This is regarded as a feature,
2310 not as a bug, because it gives slightly more control to the packagers
2311 using the @code{DESTDIR} feature.
2313 If you are bootstrapping a released version of GCC then please
2314 quickly review the build status page for your release, available from
2315 @uref{http://gcc.gnu.org/buildstat.html}.
2316 If your system is not listed for the version of GCC that you built,
2318 @email{gcc@@gcc.gnu.org} indicating
2319 that you successfully built and installed GCC@.
2320 Include the following information:
2324 Output from running @file{@var{srcdir}/config.guess}. Do not send
2325 that file itself, just the one-line output from running it.
2328 The output of @samp{gcc -v} for your newly installed @command{gcc}.
2329 This tells us which version of GCC you built and the options you passed to
2333 Whether you enabled all languages or a subset of them. If you used a
2334 full distribution then this information is part of the configure
2335 options in the output of @samp{gcc -v}, but if you downloaded the
2336 ``core'' compiler plus additional front ends then it isn't apparent
2337 which ones you built unless you tell us about it.
2340 If the build was for GNU/Linux, also include:
2343 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
2344 this information should be available from @file{/etc/issue}.
2347 The version of the Linux kernel, available from @samp{uname --version}
2351 The version of glibc you used; for RPM-based systems like Red Hat,
2352 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
2353 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
2355 For other systems, you can include similar information if you think it is
2359 Any other information that you think would be useful to people building
2360 GCC on the same configuration. The new entry in the build status list
2361 will include a link to the archived copy of your message.
2364 We'd also like to know if the
2366 @ref{Specific, host/target specific installation notes}
2369 @uref{specific.html,,host/target specific installation notes}
2371 didn't include your host/target information or if that information is
2372 incomplete or out of date. Send a note to
2373 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
2375 If you find a bug, please report it following the
2376 @uref{../bugs.html,,bug reporting guidelines}.
2378 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
2379 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
2380 and @TeX{} installed. This creates a number of @file{.dvi} files in
2381 subdirectories of @file{@var{objdir}}; these may be converted for
2382 printing with programs such as @command{dvips}. Alternately, by using
2383 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
2384 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
2385 is included with Texinfo version 4.8 and later. You can also
2386 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
2387 Free Software Foundation}, though such manuals may not be for the most
2388 recent version of GCC@.
2390 If you would like to generate online HTML documentation, do @samp{cd
2391 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
2392 @file{@var{objdir}/gcc/HTML}.
2399 @uref{./index.html,,Return to the GCC Installation page}
2403 @c ***Binaries****************************************************************
2405 @comment node-name, next, previous, up
2406 @node Binaries, Specific, Installing GCC, Top
2410 @chapter Installing GCC: Binaries
2413 @cindex Installing GCC: Binaries
2415 We are often asked about pre-compiled versions of GCC@. While we cannot
2416 provide these for all platforms, below you'll find links to binaries for
2417 various platforms where creating them by yourself is not easy due to various
2420 Please note that we did not create these binaries, nor do we
2421 support them. If you have any problems installing them, please
2422 contact their makers.
2429 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
2432 @uref{http://pware.hvcc.edu,,Hudson Valley Community College Open Source Software for IBM System p};
2435 @uref{http://www.perzl.org/aix,,AIX 5L and 6 Open Source Packages}.
2439 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
2442 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
2443 Development Tools for the Renesas H8/300[HS] Series}.
2449 @uref{http://hpux.cs.utah.edu/,,HP-UX Porting Center};
2452 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
2456 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
2457 Development Tools for the Motorola 68HC11/68HC12}.
2460 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
2461 OpenServer/Unixware}.
2464 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
2467 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
2473 The @uref{http://sourceware.org/cygwin/,,Cygwin} project;
2475 The @uref{http://www.mingw.org/,,MinGW} project.
2479 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
2480 Written Word} offers binaries for
2481 AIX 4.3.3, 5.1 and 5.2,
2483 Tru64 UNIX 4.0D and 5.1,
2485 HP-UX 10.20, 11.00, and 11.11, and
2486 Solaris/SPARC 2.5.1, 2.6, 7, 8, 9 and 10.
2489 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
2490 number of platforms.
2493 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
2494 links to GNU Fortran binaries for several platforms.
2497 In addition to those specific offerings, you can get a binary
2498 distribution CD-ROM from the
2499 @uref{http://www.gnu.org/order/order.html,,Free Software Foundation}.
2500 It contains binaries for a number of platforms, and
2501 includes not only GCC, but other stuff as well. The current CD does
2502 not contain the latest version of GCC, but it should allow
2503 bootstrapping the compiler. An updated version of that disk is in the
2511 @uref{./index.html,,Return to the GCC Installation page}
2515 @c ***Specific****************************************************************
2517 @comment node-name, next, previous, up
2518 @node Specific, Old, Binaries, Top
2522 @chapter Host/target specific installation notes for GCC
2525 @cindex Specific installation notes
2526 @cindex Target specific installation
2527 @cindex Host specific installation
2528 @cindex Target specific installation notes
2530 Please read this document carefully @emph{before} installing the
2531 GNU Compiler Collection on your machine.
2533 Note that this list of install notes is @emph{not} a list of supported
2534 hosts or targets. Not all supported hosts and targets are listed
2535 here, only the ones that require host-specific or target-specific
2541 @uref{#alpha-x-x,,alpha*-*-*}
2543 @uref{#alpha-dec-osf,,alpha*-dec-osf*}
2545 @uref{#arc-x-elf,,arc-*-elf}
2547 @uref{#arm-x-elf,,arm-*-elf}
2548 @uref{#arm-x-coff,,arm-*-coff}
2549 @uref{#arm-x-aout,,arm-*-aout}
2553 @uref{#bfin,,Blackfin}
2557 @uref{#x-x-freebsd,,*-*-freebsd*}
2559 @uref{#h8300-hms,,h8300-hms}
2561 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
2563 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
2565 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
2567 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
2569 @uref{#ix86-x-linux,,i?86-*-linux*}
2571 @uref{#ix86-x-solaris210,,i?86-*-solaris2.10}
2573 @uref{#ia64-x-linux,,ia64-*-linux}
2575 @uref{#ia64-x-hpux,,ia64-*-hpux*}
2577 @uref{#x-ibm-aix,,*-ibm-aix*}
2579 @uref{#iq2000-x-elf,,iq2000-*-elf}
2581 @uref{#m32c-x-elf,,m32c-*-elf}
2583 @uref{#m32r-x-elf,,m32r-*-elf}
2585 @uref{#m6811-elf,,m6811-elf}
2587 @uref{#m6812-elf,,m6812-elf}
2589 @uref{#m68k-x-x,,m68k-*-*}
2591 @uref{#m68k-uclinux,,m68k-uclinux}
2593 @uref{#mips-x-x,,mips-*-*}
2595 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2597 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2599 @uref{#powerpc-x-x,,powerpc*-*-*}
2601 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
2603 @uref{#powerpc-x-elf,,powerpc-*-elf}
2605 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
2607 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
2609 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
2611 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
2613 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
2615 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
2617 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
2619 @uref{#s390-x-linux,,s390-*-linux*}
2621 @uref{#s390x-x-linux,,s390x-*-linux*}
2623 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
2625 @uref{#x-x-solaris2,,*-*-solaris2*}
2627 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
2629 @uref{#sparc-sun-solaris27,,sparc-sun-solaris2.7}
2631 @uref{#sparc-x-linux,,sparc-*-linux*}
2633 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
2635 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
2637 @uref{#x-x-vxworks,,*-*-vxworks*}
2639 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
2641 @uref{#xtensa-x-elf,,xtensa*-*-elf}
2643 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
2645 @uref{#windows,,Microsoft Windows}
2647 @uref{#x-x-cygwin,,*-*-cygwin}
2649 @uref{#x-x-interix,,*-*-interix}
2651 @uref{#x-x-mingw,,*-*-mingw}
2655 @uref{#older,,Older systems}
2660 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
2666 <!-- -------- host/target specific issues start here ---------------- -->
2669 @heading @anchor{alpha-x-x}alpha*-*-*
2671 This section contains general configuration information for all
2672 alpha-based platforms using ELF (in particular, ignore this section for
2673 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2674 section, please read all other sections that match your target.
2676 We require binutils 2.11.2 or newer.
2677 Previous binutils releases had a number of problems with DWARF 2
2678 debugging information, not the least of which is incorrect linking of
2684 @heading @anchor{alpha-dec-osf}alpha*-dec-osf*
2685 Systems using processors that implement the DEC Alpha architecture and
2686 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2687 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2689 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2690 supported. (These are the versions which identify themselves as DEC
2693 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2694 may be fixed by configuring with @option{--with-gc=simple},
2695 reconfiguring Kernel Virtual Memory and Swap parameters
2696 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2697 or applying the patch in
2698 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2700 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2701 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2702 we need to use the old assembler, invoked via the barely documented
2703 @option{-oldas} option. To bootstrap GCC, you either need to use the
2707 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2710 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2713 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2716 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2717 are supported on Tru64 UNIX, so you must not configure GCC with
2718 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2720 GCC writes a @samp{.verstamp} directive to the assembler output file
2721 unless it is built as a cross-compiler. It gets the version to use from
2722 the system header file @file{/usr/include/stamp.h}. If you install a
2723 new version of DEC Unix, you should rebuild GCC to pick up the new version
2726 Note that since the Alpha is a 64-bit architecture, cross-compilers from
2727 32-bit machines will not generate code as efficient as that generated
2728 when the compiler is running on a 64-bit machine because many
2729 optimizations that depend on being able to represent a word on the
2730 target in an integral value on the host cannot be performed. Building
2731 cross-compilers on the Alpha for 32-bit machines has only been tested in
2732 a few cases and may not work properly.
2734 @samp{make compare} may fail on old versions of DEC Unix unless you add
2735 @option{-save-temps} to @code{BOOT_CFLAGS}. On these systems, the name
2736 of the assembler input file is stored in the object file, and that makes
2737 comparison fail if it differs between the @code{stage1} and
2738 @code{stage2} compilations. The option @option{-save-temps} forces a
2739 fixed name to be used for the assembler input file, instead of a
2740 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2741 unless the comparisons fail without that option. If you add
2742 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2743 @samp{.s} files after each series of compilations.
2745 GCC now supports both the native (ECOFF) debugging format used by DBX
2746 and GDB and an encapsulated STABS format for use only with GDB@. See the
2747 discussion of the @option{--with-stabs} option of @file{configure} above
2748 for more information on these formats and how to select them.
2750 There is a bug in DEC's assembler that produces incorrect line numbers
2751 for ECOFF format when the @samp{.align} directive is used. To work
2752 around this problem, GCC will not emit such alignment directives
2753 while writing ECOFF format debugging information even if optimization is
2754 being performed. Unfortunately, this has the very undesirable
2755 side-effect that code addresses when @option{-O} is specified are
2756 different depending on whether or not @option{-g} is also specified.
2758 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2759 DBX@. DEC is now aware of this problem with the assembler and hopes to
2760 provide a fix shortly.
2765 @heading @anchor{arc-x-elf}arc-*-elf
2766 Argonaut ARC processor.
2767 This configuration is intended for embedded systems.
2772 @heading @anchor{arm-x-elf}arm-*-elf
2773 ARM-family processors. Subtargets that use the ELF object format
2774 require GNU binutils 2.13 or newer. Such subtargets include:
2775 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux}
2776 and @code{arm-*-rtems}.
2781 @heading @anchor{arm-x-coff}arm-*-coff
2782 ARM-family processors. Note that there are two different varieties
2783 of PE format subtarget supported: @code{arm-wince-pe} and
2784 @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}.
2789 @heading @anchor{arm-x-aout}arm-*-aout
2790 ARM-family processors. These targets support the AOUT file format:
2791 @code{arm-*-aout}, @code{arm-*-netbsd}.
2796 @heading @anchor{avr}avr
2798 ATMEL AVR-family micro controllers. These are used in embedded
2799 applications. There are no standard Unix configurations.
2801 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
2805 See ``AVR Options'' in the main manual
2807 for the list of supported MCU types.
2809 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2811 Further installation notes and other useful information about AVR tools
2812 can also be obtained from:
2816 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
2818 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2820 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2823 We @emph{strongly} recommend using binutils 2.13 or newer.
2825 The following error:
2827 Error: register required
2830 indicates that you should upgrade to a newer version of the binutils.
2835 @heading @anchor{bfin}Blackfin
2837 The Blackfin processor, an Analog Devices DSP.
2839 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
2843 See ``Blackfin Options'' in the main manual
2846 More information, and a version of binutils with support for this processor,
2847 is available at @uref{http://blackfin.uclinux.org}
2852 @heading @anchor{cris}CRIS
2854 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2855 series. These are used in embedded applications.
2858 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
2862 See ``CRIS Options'' in the main manual
2864 for a list of CRIS-specific options.
2866 There are a few different CRIS targets:
2869 Mainly for monolithic embedded systems. Includes a multilib for the
2870 @samp{v10} core used in @samp{ETRAX 100 LX}.
2871 @item cris-axis-linux-gnu
2872 A GNU/Linux port for the CRIS architecture, currently targeting
2873 @samp{ETRAX 100 LX} by default.
2876 For @code{cris-axis-elf} you need binutils 2.11
2877 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2879 Pre-packaged tools can be obtained from
2880 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2881 information about this platform is available at
2882 @uref{http://developer.axis.com/}.
2887 @heading @anchor{crx}CRX
2889 The CRX CompactRISC architecture is a low-power 32-bit architecture with
2890 fast context switching and architectural extensibility features.
2893 @xref{CRX Options,, CRX Options, gcc, Using and Porting the GNU Compiler
2898 See ``CRX Options'' in the main manual for a list of CRX-specific options.
2901 Use @samp{configure --target=crx-elf --enable-languages=c,c++} to configure
2902 GCC@ for building a CRX cross-compiler. The option @samp{--target=crx-elf}
2903 is also used to build the @samp{newlib} C library for CRX.
2905 It is also possible to build libstdc++-v3 for the CRX architecture. This
2906 needs to be done in a separate step with the following configure settings:
2907 @samp{gcc/libstdc++-v3/configure --host=crx-elf --with-newlib
2908 --enable-sjlj-exceptions --enable-cxx-flags='-fexceptions -frtti'}
2913 @heading @anchor{dos}DOS
2915 Please have a look at the @uref{binaries.html,,binaries page}.
2917 You cannot install GCC by itself on MSDOS; it will not compile under
2918 any MSDOS compiler except itself. You need to get the complete
2919 compilation package DJGPP, which includes binaries as well as sources,
2920 and includes all the necessary compilation tools and libraries.
2925 @heading @anchor{x-x-freebsd}*-*-freebsd*
2927 The version of binutils installed in @file{/usr/bin} probably works with
2928 this release of GCC@. However, on FreeBSD 4, bootstrapping against the
2929 latest FSF binutils is known to improve overall testsuite results; and,
2930 on FreeBSD/alpha, using binutils 2.14 or later is required to build libjava.
2932 Support for FreeBSD 1 was discontinued in GCC 3.2.
2934 Support for FreeBSD 2 will be discontinued after GCC 3.4. The
2935 following was true for GCC 3.1 but the current status is unknown.
2936 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2937 configuration support and files as shipped with GCC 2.95 are still in
2938 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2939 it is unknown which version of binutils was used (it is assumed that it
2940 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2942 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2943 default for all CPU architectures. It had been the default on
2944 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2945 of @option{-g}, if you really want the old debugging format. There are
2946 no known issues with mixing object files and libraries with different
2947 debugging formats. Otherwise, this release of GCC should now match more
2948 of the configuration used in the stock FreeBSD configuration of GCC@. In
2949 particular, @option{--enable-threads} is now configured by default.
2950 However, as a general user, do not attempt to replace the system
2951 compiler with this release. Known to bootstrap and check with good
2952 results on FreeBSD 4.9-STABLE and 5-CURRENT@. In the past, known to
2953 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
2954 4.3, 4.4, 4.5, 4.8-STABLE@.
2956 In principle, @option{--enable-threads} is now compatible with
2957 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
2958 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
2960 library may be incorrectly built (symbols are missing at link time).
2961 There is a rare timing-based startup hang (probably involves an
2962 assumption about the thread library). Multi-threaded boehm-gc (required for
2963 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
2964 4.5-RELEASE@. Other CPU architectures
2965 supported by FreeBSD will require additional configuration tuning in, at
2966 the very least, both boehm-gc and libffi.
2968 Shared @file{libgcc_s.so} is now built and installed by default.
2973 @heading @anchor{h8300-hms}h8300-hms
2974 Renesas H8/300 series of processors.
2976 Please have a look at the @uref{binaries.html,,binaries page}.
2978 The calling convention and structure layout has changed in release 2.6.
2979 All code must be recompiled. The calling convention now passes the
2980 first three arguments in function calls in registers. Structures are no
2981 longer a multiple of 2 bytes.
2986 @heading @anchor{hppa-hp-hpux}hppa*-hp-hpux*
2987 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
2989 We require using gas/binutils on all hppa platforms;
2990 you may encounter a variety of problems if you try to use the HP assembler.
2992 Specifically, @option{-g} does not work on HP-UX (since that system
2993 uses a peculiar debugging format which GCC does not know about), unless
2994 you use GAS and GDB@. It may be helpful to configure GCC with the
2995 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
2996 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
2998 If you wish to use the pa-risc 2.0 architecture support with a 32-bit
2999 runtime, you must use gas/binutils 2.11 or newer.
3001 There are two default scheduling models for instructions. These are
3002 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
3003 architecture specified for the target machine when configuring.
3004 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
3005 the target is a @samp{hppa1*} machine.
3007 The PROCESSOR_8000 model is not well suited to older processors. Thus,
3008 it is important to completely specify the machine architecture when
3009 configuring if you want a model other than PROCESSOR_8000. The macro
3010 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
3011 default scheduling model is desired.
3013 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
3014 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
3015 This namespace change might cause problems when bootstrapping with
3016 an earlier version of GCC or the HP compiler as essentially the same
3017 namespace is required for an entire build. This problem can be avoided
3018 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
3019 or @samp{98}. Another way is to add an appropriate set of predefines
3020 to @env{CC}. The description for the @option{munix=} option contains
3021 a list of the predefines used with each standard.
3023 As of GCC 4.1, @env{DWARF2} exception handling is available on HP-UX@.
3024 It is now the default. This exposed a bug in the handling of data
3025 relocations in the GAS assembler. The handling of 64-bit data relocations
3026 was seriously broken, affecting debugging and exception support on all
3027 @samp{hppa64-*-*} targets. Under some circumstances, 32-bit data relocations
3028 could also be handled incorrectly. This problem is fixed in GAS version
3031 GCC versions prior to 4.1 incorrectly passed and returned complex
3032 values. They are now passed in the same manner as aggregates.
3034 More specific information to @samp{hppa*-hp-hpux*} targets follows.
3039 @heading @anchor{hppa-hp-hpux10}hppa*-hp-hpux10
3041 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
3042 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
3048 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
3052 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
3056 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
3059 The HP assembler on these systems has some problems. Most notably the
3060 assembler inserts timestamps into each object file it creates, causing
3061 the 3-stage comparison test to fail during a bootstrap.
3062 You should be able to continue by saying @samp{make all-host all-target}
3063 after getting the failure from @samp{make}.
3065 GCC 4.0 requires CVS binutils as of April 28, 2004 or later. Earlier
3066 versions require binutils 2.8 or later.
3068 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
3069 used for one-only code and data. This resolves many of the previous
3070 problems in using C++ on this target. However, the ABI is not compatible
3071 with the one implemented under HP-UX 11 using secondary definitions.
3076 @heading @anchor{hppa-hp-hpux11}hppa*-hp-hpux11
3078 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
3079 be used to compile GCC 3.0 and up.
3081 Refer to @uref{binaries.html,,binaries} for information about obtaining
3082 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
3083 to build the Ada language as it can't be bootstrapped using C@. Ada is
3084 only available for the 32-bit PA-RISC runtime. The libffi and libjava
3085 haven't been ported to HP-UX and don't build.
3087 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
3088 bundled compiler supports only traditional C; you will need either HP's
3089 unbundled compiler, or a binary distribution of GCC@.
3091 It is possible to build GCC 3.3 starting with the bundled HP compiler,
3092 but the process requires several steps. GCC 3.3 can then be used to
3093 build later versions. The fastjar program contains ISO C code and
3094 can't be built with the HP bundled compiler. This problem can be
3095 avoided by not building the Java language. For example, use the
3096 @option{--enable-languages="c,c++,f77,objc"} option in your configure
3099 There are several possible approaches to building the distribution.
3100 Binutils can be built first using the HP tools. Then, the GCC
3101 distribution can be built. The second approach is to build GCC
3102 first using the HP tools, then build binutils, then rebuild GCC@.
3103 There have been problems with various binary distributions, so it
3104 is best not to start from a binary distribution.
3106 On 64-bit capable systems, there are two distinct targets. Different
3107 installation prefixes must be used if both are to be installed on
3108 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
3109 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
3110 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
3111 PA-RISC 2.0 architecture. The HP and GNU linkers are both supported
3114 The script config.guess now selects the target type based on the compiler
3115 detected during configuration. You must define @env{PATH} or @env{CC} so
3116 that configure finds an appropriate compiler for the initial bootstrap.
3117 When @env{CC} is used, the definition should contain the options that are
3118 needed whenever @env{CC} is used.
3120 Specifically, options that determine the runtime architecture must be
3121 in @env{CC} to correctly select the target for the build. It is also
3122 convenient to place many other compiler options in @env{CC}. For example,
3123 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
3124 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
3125 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
3126 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
3127 macro definition table of cpp needs to be increased for a successful
3128 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
3129 be defined when building with the bundled compiler, or when using the
3130 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
3132 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
3133 with the @option{--with-ld=@dots{}} option. This overrides the standard
3134 search for ld. The two linkers supported on this target require different
3135 commands. The default linker is determined during configuration. As a
3136 result, it's not possible to switch linkers in the middle of a GCC build.
3137 This has been been reported to sometimes occur in unified builds of
3140 GCC 3.0 through 3.2 require binutils 2.11 or above. GCC 3.3 through
3141 GCC 4.0 require binutils 2.14 or later.
3143 Although the HP assembler can be used for an initial build, it shouldn't
3144 be used with any languages other than C and perhaps Fortran due to its
3145 many limitations. For example, it does not support weak symbols or alias
3146 definitions. As a result, explicit template instantiations are required
3147 when using C++. This makes it difficult if not impossible to build many
3148 C++ applications. You can't generate debugging information when using
3149 the HP assembler. Finally, bootstrapping fails in the final
3150 comparison of object modules due to the time stamps that it inserts into
3151 the modules. The bootstrap can be continued from this point with
3152 @samp{make all-host all-target}.
3154 A recent linker patch must be installed for the correct operation of
3155 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
3156 oldest linker patches that are known to work. They are for HP-UX
3157 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
3158 @code{PHSS_24304}, might be usable but it hasn't been tested. These
3159 patches have been superseded. Consult the HP patch database to obtain
3160 the currently recommended linker patch for your system.
3162 The patches are necessary for the support of weak symbols on the
3163 32-bit port, and for the running of initializers and finalizers. Weak
3164 symbols are implemented using SOM secondary definition symbols. Prior
3165 to HP-UX 11, there are bugs in the linker support for secondary symbols.
3166 The patches correct a problem of linker core dumps creating shared
3167 libraries containing secondary symbols, as well as various other
3168 linking issues involving secondary symbols.
3170 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
3171 run initializers and finalizers on the 64-bit port. The 32-bit port
3172 uses the linker @option{+init} and @option{+fini} options for the same
3173 purpose. The patches correct various problems with the +init/+fini
3174 options, including program core dumps. Binutils 2.14 corrects a
3175 problem on the 64-bit port resulting from HP's non-standard use of
3176 the .init and .fini sections for array initializers and finalizers.
3178 There are a number of issues to consider in selecting which linker to
3179 use with the 64-bit port. The GNU 64-bit linker can only create dynamic
3180 binaries. The @option{-static} option causes linking with archive
3181 libraries but doesn't produce a truly static binary. Dynamic binaries
3182 still require final binding by the dynamic loader to resolve a set of
3183 dynamic-loader-defined symbols. The default behavior of the HP linker
3184 is the same as the GNU linker. However, it can generate true 64-bit
3185 static binaries using the @option{+compat} option.
3187 The HP 64-bit linker doesn't support linkonce semantics. As a
3188 result, C++ programs have many more sections than they should.
3190 The GNU 64-bit linker has some issues with shared library support
3191 and exceptions. As a result, we only support libgcc in archive
3192 format. For similar reasons, dwarf2 unwind and exception support
3193 are disabled. The GNU linker also has problems creating binaries
3194 with @option{-static}. It doesn't provide stubs for internal
3195 calls to global functions in shared libraries, so these calls
3196 can't be overloaded.
3198 Thread support is not implemented in GCC 3.0 through 3.2, so the
3199 @option{--enable-threads} configure option does not work. In 3.3
3200 and later, POSIX threads are supported. The optional DCE thread
3201 library is not supported.
3203 This port still is undergoing significant development.
3208 @heading @anchor{x-x-linux-gnu}*-*-linux-gnu
3210 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
3211 in glibc 2.2.5 and later. More information is available in the
3212 libstdc++-v3 documentation.
3217 @heading @anchor{ix86-x-linux}i?86-*-linux*
3219 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
3220 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
3222 If you receive Signal 11 errors when building on GNU/Linux, then it is
3223 possible you have a hardware problem. Further information on this can be
3224 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
3229 @heading @anchor{ix86-x-solaris210}i?86-*-solaris2.10
3230 Use this for Solaris 10 or later on x86 and x86-64 systems. This
3231 configuration is supported by GCC 4.0 and later versions only.
3233 It is recommended that you configure GCC to use the GNU assembler in
3234 @file{/usr/sfw/bin/gas} but the Sun linker, using the options
3235 @option{--with-gnu-as --with-as=/usr/sfw/bin/gas --without-gnu-ld
3236 --with-ld=/usr/ccs/bin/ld}.
3241 @heading @anchor{ia64-x-linux}ia64-*-linux
3242 IA-64 processor (also known as IPF, or Itanium Processor Family)
3245 If you are using the installed system libunwind library with
3246 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
3249 None of the following versions of GCC has an ABI that is compatible
3250 with any of the other versions in this list, with the exception that
3251 Red Hat 2.96 and Trillian 000171 are compatible with each other:
3252 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
3253 This primarily affects C++ programs and programs that create shared libraries.
3254 GCC 3.1 or later is recommended for compiling linux, the kernel.
3255 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
3256 more major ABI changes are expected.
3261 @heading @anchor{ia64-x-hpux}ia64-*-hpux*
3262 Building GCC on this target requires the GNU Assembler. The bundled HP
3263 assembler will not work. To prevent GCC from using the wrong assembler,
3264 the option @option{--with-gnu-as} may be necessary.
3266 The GCC libunwind library has not been ported to HPUX@. This means that for
3267 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
3268 is required to build GCC@. For GCC 3.3 and later, this is the default.
3269 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
3270 removed and the system libunwind library will always be used.
3274 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
3276 @heading @anchor{x-ibm-aix}*-ibm-aix*
3277 Support for AIX version 3 and older was discontinued in GCC 3.4.
3279 ``out of memory'' bootstrap failures may indicate a problem with
3280 process resource limits (ulimit). Hard limits are configured in the
3281 @file{/etc/security/limits} system configuration file.
3283 To speed up the configuration phases of bootstrapping and installing GCC,
3284 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
3287 % CONFIG_SHELL=/opt/freeware/bin/bash
3288 % export CONFIG_SHELL
3291 and then proceed as described in @uref{build.html,,the build
3292 instructions}, where we strongly recommend specifying an absolute path
3293 to invoke @var{srcdir}/configure.
3295 Because GCC on AIX is built as a 32-bit executable by default,
3296 (although it can generate 64-bit programs) the GMP and MPFR libraries
3297 required by gfortran must be 32-bit libraries. Building GMP and MPFR
3298 as static archive libraries works better than shared libraries.
3300 Errors involving @code{alloca} when building GCC generally are due
3301 to an incorrect definition of @code{CC} in the Makefile or mixing files
3302 compiled with the native C compiler and GCC@. During the stage1 phase of
3303 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
3304 (not @command{xlc}). Once @command{configure} has been informed of
3305 @command{xlc}, one needs to use @samp{make distclean} to remove the
3306 configure cache files and ensure that @env{CC} environment variable
3307 does not provide a definition that will confuse @command{configure}.
3308 If this error occurs during stage2 or later, then the problem most likely
3309 is the version of Make (see above).
3311 The native @command{as} and @command{ld} are recommended for bootstrapping
3312 on AIX 4 and required for bootstrapping on AIX 5L@. The GNU Assembler
3313 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
3314 utilize weak symbol functionality although it is not supported. The GNU
3315 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC@.
3316 The native AIX tools do interoperate with GCC@.
3318 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
3319 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
3320 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
3321 referenced as APAR IY53606 (AIX 5.2) or a APAR IY54774 (AIX 5.1)
3323 @samp{libstdc++} in GCC 3.4 increments the major version number of the
3324 shared object and GCC installation places the @file{libstdc++.a}
3325 shared library in a common location which will overwrite the and GCC
3326 3.3 version of the shared library. Applications either need to be
3327 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
3328 versions of the @samp{libstdc++} shared object needs to be available
3329 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
3330 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
3331 installed for runtime dynamic loading using the following steps to set
3332 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3333 multilib @file{libstdc++.a} installed:
3335 Extract the shared objects from the currently installed
3336 @file{libstdc++.a} archive:
3338 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3341 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3342 available for runtime dynamic loading, but not linking:
3344 % strip -e libstdc++.so.4 libstdc++.so.5
3347 Archive the runtime-only shared object in the GCC 3.4
3348 @file{libstdc++.a} archive:
3350 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
3353 Linking executables and shared libraries may produce warnings of
3354 duplicate symbols. The assembly files generated by GCC for AIX always
3355 have included multiple symbol definitions for certain global variable
3356 and function declarations in the original program. The warnings should
3357 not prevent the linker from producing a correct library or runnable
3360 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
3361 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
3362 to parse archive libraries did not handle the new format correctly.
3363 These routines are used by GCC and result in error messages during
3364 linking such as ``not a COFF file''. The version of the routines shipped
3365 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
3366 option of the archive command may be used to create archives of 32-bit
3367 objects using the original ``small format''. A correct version of the
3368 routines is shipped with AIX 4.3.2 and above.
3370 Some versions of the AIX binder (linker) can fail with a relocation
3371 overflow severe error when the @option{-bbigtoc} option is used to link
3372 GCC-produced object files into an executable that overflows the TOC@. A fix
3373 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
3374 available from IBM Customer Support and from its
3375 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3376 website as PTF U455193.
3378 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
3379 with a segmentation fault when invoked by any version of GCC@. A fix for
3380 APAR IX87327 is available from IBM Customer Support and from its
3381 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3382 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
3384 The initial assembler shipped with AIX 4.3.0 generates incorrect object
3385 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
3386 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
3387 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3388 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
3390 AIX provides National Language Support (NLS)@. Compilers and assemblers
3391 use NLS to support locale-specific representations of various data
3392 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
3393 separating decimal fractions). There have been problems reported where
3394 GCC does not produce the same floating-point formats that the assembler
3395 expects. If one encounters this problem, set the @env{LANG}
3396 environment variable to @samp{C} or @samp{En_US}.
3398 By default, GCC for AIX 4.1 and above produces code that can be used on
3399 both Power or PowerPC processors.
3401 A default can be specified with the @option{-mcpu=@var{cpu_type}}
3402 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
3407 @heading @anchor{iq2000-x-elf}iq2000-*-elf
3408 Vitesse IQ2000 processors. These are used in embedded
3409 applications. There are no standard Unix configurations.
3414 @heading @anchor{m32c-x-elf}m32c-*-elf
3415 Renesas M32C processor.
3416 This configuration is intended for embedded systems.
3421 @heading @anchor{m32r-x-elf}m32r-*-elf
3422 Renesas M32R processor.
3423 This configuration is intended for embedded systems.
3428 @heading @anchor{m6811-elf}m6811-elf
3429 Motorola 68HC11 family micro controllers. These are used in embedded
3430 applications. There are no standard Unix configurations.
3435 @heading @anchor{m6812-elf}m6812-elf
3436 Motorola 68HC12 family micro controllers. These are used in embedded
3437 applications. There are no standard Unix configurations.
3442 @heading @anchor{m68k-x-x}m68k-*-*
3443 By default, @samp{m68k-*-aout}, @samp{m68k-*-coff*},
3444 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
3446 build libraries for both M680x0 and ColdFire processors. If you only
3447 need the M680x0 libraries, you can omit the ColdFire ones by passing
3448 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
3449 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
3450 @command{configure}. These targets default to 5206 or 5475 code as
3451 appropriate for the target system when
3452 configured with @option{--with-arch=cf} and 68020 code otherwise.
3454 The @samp{m68k-*-netbsd} and
3455 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
3456 option. They will generate ColdFire CFV4e code when configured with
3457 @option{--with-arch=cf} and 68020 code otherwise.
3459 You can override the default processors listed above by configuring
3460 with @option{--with-cpu=@var{target}}. This @var{target} can either
3461 be a @option{-mcpu} argument or one of the following values:
3462 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
3463 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
3468 @heading @anchor{m68k-x-uclinux}m68k-*-uclinux
3469 GCC 4.3 changed the uClinux configuration so that it uses the
3470 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
3471 It also added improved support for C++ and flat shared libraries,
3472 both of which were ABI changes. However, you can still use the
3473 original ABI by configuring for @samp{m68k-uclinuxoldabi} or
3474 @samp{m68k-@var{vendor}-uclinuxoldabi}.
3479 @heading @anchor{mips-x-x}mips-*-*
3480 If on a MIPS system you get an error message saying ``does not have gp
3481 sections for all it's [sic] sectons [sic]'', don't worry about it. This
3482 happens whenever you use GAS with the MIPS linker, but there is not
3483 really anything wrong, and it is okay to use the output file. You can
3484 stop such warnings by installing the GNU linker.
3486 It would be nice to extend GAS to produce the gp tables, but they are
3487 optional, and there should not be a warning about their absence.
3489 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
3490 and later. A patch went in just after the GCC 3.3 release to
3491 make @samp{mips*-*-*} use the generic implementation instead. You can also
3492 configure for @samp{mipsel-elf} as a workaround. The
3493 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
3494 work on this is expected in future releases.
3496 @c If you make --with-llsc the default for another target, please also
3497 @c update the description of the --with-llsc option.
3499 The built-in @code{__sync_*} functions are available on MIPS II and
3500 later systems and others that support the @samp{ll}, @samp{sc} and
3501 @samp{sync} instructions. This can be overridden by passing
3502 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
3503 Since the Linux kernel emulates these instructions if they are
3504 missing, the default for @samp{mips*-*-linux*} targets is
3505 @option{--with-llsc}. The @option{--with-llsc} and
3506 @option{--without-llsc} configure options may be overridden at compile
3507 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
3510 MIPS systems check for division by zero (unless
3511 @option{-mno-check-zero-division} is passed to the compiler) by
3512 generating either a conditional trap or a break instruction. Using
3513 trap results in smaller code, but is only supported on MIPS II and
3514 later. Also, some versions of the Linux kernel have a bug that
3515 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
3516 the use of break, use the @option{--with-divide=breaks}
3517 @command{configure} option when configuring GCC@. The default is to
3518 use traps on systems that support them.
3520 Cross-compilers for the MIPS as target using the MIPS assembler
3521 currently do not work, because the auxiliary programs
3522 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
3523 anything but a MIPS@. It does work to cross compile for a MIPS
3524 if you use the GNU assembler and linker.
3526 The assembler from GNU binutils 2.17 and earlier has a bug in the way
3527 it sorts relocations for REL targets (o32, o64, EABI). This can cause
3528 bad code to be generated for simple C++ programs. Also the linker
3529 from GNU binutils versions prior to 2.17 has a bug which causes the
3530 runtime linker stubs in very large programs, like @file{libgcj.so}, to
3531 be incorrectly generated. GNU Binutils 2.18 and later (and snapshots
3532 made after Nov. 9, 2006) should be free from both of these problems.
3537 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
3539 In order to compile GCC on an SGI running IRIX 5, the @samp{compiler_dev.hdr}
3540 subsystem must be installed from the IDO CD-ROM supplied by SGI@.
3541 It is also available for download from
3542 @uref{ftp://ftp.sgi.com/sgi/IRIX5.3/iris-development-option-5.3.tardist}.
3544 If you use the MIPS C compiler to bootstrap, it may be necessary
3545 to increase its table size for switch statements with the
3546 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
3547 optimization option, you also need to use @option{-Olimit 3000}.
3549 To enable debugging under IRIX 5, you must use GNU binutils 2.15 or
3550 later, and use the @option{--with-gnu-ld} @command{configure} option
3551 when configuring GCC@. You need to use GNU @command{ar} and @command{nm},
3552 also distributed with GNU binutils.
3554 Some users have reported that @command{/bin/sh} will hang during bootstrap.
3555 This problem can be avoided by running the commands:
3558 % CONFIG_SHELL=/bin/ksh
3559 % export CONFIG_SHELL
3562 before starting the build.
3567 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3569 If you are using SGI's MIPSpro @command{cc} as your bootstrap compiler, you must
3570 ensure that the N32 ABI is in use. To test this, compile a simple C
3571 file with @command{cc} and then run @command{file} on the
3572 resulting object file. The output should look like:
3575 test.o: ELF N32 MSB @dots{}
3581 test.o: ELF 32-bit MSB @dots{}
3587 test.o: ELF 64-bit MSB @dots{}
3590 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3591 should set the environment variable @env{CC} to @samp{cc -n32}
3592 before configuring GCC@.
3594 If you want the resulting @command{gcc} to run on old 32-bit systems
3595 with the MIPS R4400 CPU, you need to ensure that only code for the @samp{mips3}
3596 instruction set architecture (ISA) is generated. While GCC 3.x does
3597 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3598 the ISA depending on the machine where GCC is built. Using one of them
3599 as the bootstrap compiler may result in @samp{mips4} code, which won't run at
3600 all on @samp{mips3}-only systems. For the test program above, you should see:
3603 test.o: ELF N32 MSB mips-3 @dots{}
3609 test.o: ELF N32 MSB mips-4 @dots{}
3612 instead, you should set the environment variable @env{CC} to @samp{cc
3613 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3615 MIPSpro C 7.4 may cause bootstrap failures, due to a bug when inlining
3616 @code{memcmp}. Either add @code{-U__INLINE_INTRINSICS} to the @env{CC}
3617 environment variable as a workaround or upgrade to MIPSpro C 7.4.1m.
3619 GCC on IRIX 6 is usually built to support the N32, O32 and N64 ABIs. If
3620 you build GCC on a system that doesn't have the N64 libraries installed
3621 or cannot run 64-bit binaries,
3622 you need to configure with @option{--disable-multilib} so GCC doesn't
3623 try to use them. This will disable building the O32 libraries, too.
3624 Look for @file{/usr/lib64/libc.so.1} to see if you
3625 have the 64-bit libraries installed.
3627 To enable debugging for the O32 ABI, you must use GNU @command{as} from
3628 GNU binutils 2.15 or later. You may also use GNU @command{ld}, but
3629 this is not required and currently causes some problems with Ada.
3631 The @option{--enable-libgcj}
3632 option is disabled by default: IRIX 6 uses a very low default limit
3633 (20480) for the command line length. Although @command{libtool} contains a
3634 workaround for this problem, at least the N64 @samp{libgcj} is known not
3635 to build despite this, running into an internal error of the native
3636 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3637 its maximum of 262144 bytes. If you have root access, you can use the
3638 @command{systune} command to do this.
3640 @code{wchar_t} support in @samp{libstdc++} is not available for old
3641 IRIX 6.5.x releases, @math{x < 19}. The problem cannot be autodetected
3642 and in order to build GCC for such targets you need to configure with
3643 @option{--disable-wchar_t}.
3645 See @uref{http://freeware.sgi.com/} for more
3646 information about using GCC on IRIX platforms.
3651 @heading @anchor{powerpc-x-x}powerpc-*-*
3653 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3654 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3659 @heading @anchor{powerpc-x-darwin}powerpc-*-darwin*
3660 PowerPC running Darwin (Mac OS X kernel).
3662 Pre-installed versions of Mac OS X may not include any developer tools,
3663 meaning that you will not be able to build GCC from source. Tool
3664 binaries are available at
3665 @uref{http://developer.apple.com/darwin/projects/compiler/} (free
3666 registration required).
3668 This version of GCC requires at least cctools-590.36. The
3669 cctools-590.36 package referenced from
3670 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
3671 on systems older than 10.3.9 (aka darwin7.9.0).
3676 @heading @anchor{powerpc-x-elf}powerpc-*-elf
3677 PowerPC system in big endian mode, running System V.4.
3682 @heading @anchor{powerpc-x-linux-gnu}powerpc*-*-linux-gnu*
3685 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.15}
3686 or newer for a working GCC@.
3691 @heading @anchor{powerpc-x-netbsd}powerpc-*-netbsd*
3692 PowerPC system in big endian mode running NetBSD@.
3697 @heading @anchor{powerpc-x-eabisim}powerpc-*-eabisim
3698 Embedded PowerPC system in big endian mode for use in running under the
3704 @heading @anchor{powerpc-x-eabi}powerpc-*-eabi
3705 Embedded PowerPC system in big endian mode.
3710 @heading @anchor{powerpcle-x-elf}powerpcle-*-elf
3711 PowerPC system in little endian mode, running System V.4.
3716 @heading @anchor{powerpcle-x-eabisim}powerpcle-*-eabisim
3717 Embedded PowerPC system in little endian mode for use in running under
3723 @heading @anchor{powerpcle-x-eabi}powerpcle-*-eabi
3724 Embedded PowerPC system in little endian mode.
3729 @heading @anchor{s390-x-linux}s390-*-linux*
3730 S/390 system running GNU/Linux for S/390@.
3735 @heading @anchor{s390x-x-linux}s390x-*-linux*
3736 zSeries system (64-bit) running GNU/Linux for zSeries@.
3741 @heading @anchor{s390x-ibm-tpf}s390x-ibm-tpf*
3742 zSeries system (64-bit) running TPF@. This platform is
3743 supported as cross-compilation target only.
3748 @c Please use Solaris 2 to refer to all release of Solaris, starting
3749 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
3750 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3751 @c alone is too unspecific and must be avoided.
3752 @heading @anchor{x-x-solaris2}*-*-solaris2*
3754 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3755 GCC you first have to install a pre-built compiler, see the
3756 @uref{binaries.html,,binaries page} for details.
3758 The Solaris 2 @command{/bin/sh} will often fail to configure
3759 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3760 recommend using the following initial sequence of commands
3763 % CONFIG_SHELL=/bin/ksh
3764 % export CONFIG_SHELL
3767 and proceed as described in @uref{configure.html,,the configure instructions}.
3768 In addition we strongly recommend specifying an absolute path to invoke
3769 @var{srcdir}/configure.
3771 Solaris 2 comes with a number of optional OS packages. Some of these
3772 are needed to use GCC fully, namely @code{SUNWarc},
3773 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3774 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3775 optional packages when installing Solaris 2, you will need to verify that
3776 the packages that GCC needs are installed.
3778 To check whether an optional package is installed, use
3779 the @command{pkginfo} command. To add an optional package, use the
3780 @command{pkgadd} command. For further details, see the Solaris 2
3783 Trying to use the linker and other tools in
3784 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3785 For example, the linker may hang indefinitely. The fix is to remove
3786 @file{/usr/ucb} from your @env{PATH}.
3788 The build process works more smoothly with the legacy Sun tools so, if you
3789 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3790 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3792 We recommend the use of GNU binutils 2.14 or later, or the vendor tools
3793 (Sun @command{as}, Sun @command{ld}). Note that your mileage may vary
3794 if you use a combination of the GNU tools and the Sun tools: while the
3795 combination GNU @command{as} + Sun @command{ld} should reasonably work,
3796 the reverse combination Sun @command{as} + GNU @command{ld} is known to
3797 cause memory corruption at runtime in some cases for C++ programs.
3799 The stock GNU binutils 2.15 release is broken on this platform because of a
3800 single bug. It has been fixed on the 2.15 branch in the CVS repository.
3801 You can obtain a working version by checking out the binutils-2_15-branch
3802 from the CVS repository or applying the patch
3803 @uref{http://sourceware.org/ml/binutils-cvs/2004-09/msg00036.html} to the
3806 We recommend the use of GNU binutils 2.16 or later in conjunction with GCC
3807 4.x, or the vendor tools (Sun @command{as}, Sun @command{ld}). However,
3808 for Solaris 10 and above, an additional patch is required in order for the
3809 GNU linker to be able to cope with a new flavor of shared libraries. You
3810 can obtain a working version by checking out the binutils-2_16-branch from
3811 the CVS repository or applying the patch
3812 @uref{http://sourceware.org/ml/binutils-cvs/2005-07/msg00122.html} to the
3815 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3816 newer: @command{g++} will complain that types are missing. These headers
3817 assume that omitting the type means @code{int}; this assumption worked for
3818 C89 but is wrong for C++, and is now wrong for C99 also.
3820 @command{g++} accepts such (invalid) constructs with the option
3821 @option{-fpermissive}; it will assume that any missing type is @code{int}
3822 (as defined by C89).
3824 There are patches for Solaris 7 (108376-21 or newer for SPARC,
3825 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3826 108653-22 for Intel) that fix this bug.
3828 Sun bug 4927647 sometimes causes random spurious testsuite failures
3829 related to missing diagnostic output. This bug doesn't affect GCC
3830 itself, rather it is a kernel bug triggered by the @command{expect}
3831 program which is used only by the GCC testsuite driver. When the bug
3832 causes the @command{expect} program to miss anticipated output, extra
3833 testsuite failures appear.
3835 There are patches for Solaris 8 (117350-12 or newer for SPARC,
3836 117351-12 or newer for Intel) and Solaris 9 (117171-11 or newer for
3837 SPARC, 117172-11 or newer for Intel) that address this problem.
3842 @heading @anchor{sparc-sun-solaris2}sparc-sun-solaris2*
3844 When GCC is configured to use binutils 2.14 or later the binaries
3845 produced are smaller than the ones produced using Sun's native tools;
3846 this difference is quite significant for binaries containing debugging
3849 Starting with Solaris 7, the operating system is capable of executing
3850 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3851 this; the @option{-m64} option enables 64-bit code generation.
3852 However, if all you want is code tuned for the UltraSPARC CPU, you
3853 should try the @option{-mtune=ultrasparc} option instead, which produces
3854 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3857 When configuring on a Solaris 7 or later system that is running a kernel
3858 that supports only 32-bit binaries, one must configure with
3859 @option{--disable-multilib}, since we will not be able to build the
3860 64-bit target libraries.
3862 GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
3863 the GNU compiler (especially GCC 3.0.x versions), which lead to the
3864 miscompilation of the stage1 compiler and the subsequent failure of the
3865 bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
3866 stage, i.e.@: to bootstrap that compiler with the base compiler and then
3867 use it to bootstrap the final compiler.
3869 GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
3870 and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
3871 failure in form of a miscompilation of the stage1 compiler by the Sun
3872 compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
3874 GCC 3.4 changed the default debugging format from STABS to DWARF-2 for
3875 32-bit code on Solaris 7 and later. If you use the Sun assembler, this
3876 change apparently runs afoul of Sun bug 4910101 (which is referenced as
3877 a x86-only problem by Sun, probably because they do not use DWARF-2).
3878 A symptom of the problem is that you cannot compile C++ programs like
3879 @command{groff} 1.19.1 without getting messages similar to the following:
3882 ld: warning: relocation error: R_SPARC_UA32: @dots{}
3883 external symbolic relocation against non-allocatable section
3884 .debug_info cannot be processed at runtime: relocation ignored.
3887 To work around this problem, compile with @option{-gstabs+} instead of
3890 When configuring the GNU Multiple Precision Library (GMP) or the MPFR
3891 library on a Solaris 7 or later system, the canonical target triplet
3892 must be specified as the @command{build} parameter on the configure
3893 line. This triplet can be obtained by invoking ./config.guess in
3894 the toplevel source directory of GCC (and not that of GMP or MPFR).
3895 For example on a Solaris 7 system:
3898 % ./configure --build=sparc-sun-solaris2.7 --prefix=xxx
3904 @heading @anchor{sparc-sun-solaris27}sparc-sun-solaris2.7
3906 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3907 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3908 and later, including all EGCS releases. Sun formerly recommended
3909 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3910 recommend it only for people who use Sun's compilers.
3912 Here are some workarounds to this problem:
3915 Do not install Sun patch 107058-01 until after Sun releases a
3916 complete patch for bug 4210064. This is the simplest course to take,
3917 unless you must also use Sun's C compiler. Unfortunately 107058-01
3918 is preinstalled on some new Solaris 7-based hosts, so you may have to
3922 Copy the original, unpatched Solaris 7
3923 @command{/usr/ccs/bin/as} into
3924 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
3925 adjusting the latter name to fit your local conventions and software
3929 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3930 both 107058-01 and 106950-03 installed has reported the bug with GCC
3931 and Sun's dynamic linker. This last course of action is riskiest,
3932 for two reasons. First, you must install 106950 on all hosts that
3933 run code generated by GCC; it doesn't suffice to install it only on
3934 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3935 only a partial fix for bug 4210064, but Sun doesn't know whether the
3936 partial fix is adequate for GCC@. Revision -08 or later should fix
3937 the bug. The current (as of 2004-05-23) revision is -24, and is included in
3938 the Solaris 7 Recommended Patch Cluster.
3941 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
3942 which causes a bootstrap failure when linking the 64-bit shared version of
3943 libgcc. A typical error message is:
3946 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
3947 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
3950 This bug has been fixed in the final 5.0 version of the assembler.
3952 A similar problem was reported for version Sun WorkShop 6 99/08/18 of the
3953 Sun assembler, which causes a bootstrap failure with GCC 4.0.0:
3956 ld: fatal: relocation error: R_SPARC_DISP32:
3957 file .libs/libstdc++.lax/libsupc++convenience.a/vterminate.o:
3958 symbol <unknown>: offset 0xfccd33ad is non-aligned
3961 This bug has been fixed in more recent revisions of the assembler.
3966 @heading @anchor{sparc-x-linux}sparc-*-linux*
3968 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
3969 or newer on this platform. All earlier binutils and glibc
3970 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
3976 @heading @anchor{sparc64-x-solaris2}sparc64-*-solaris2*
3978 When configuring the GNU Multiple Precision Library (GMP) or the
3979 MPFR library, the canonical target triplet must be specified as
3980 the @command{build} parameter on the configure line. For example
3981 on a Solaris 7 system:
3984 % ./configure --build=sparc64-sun-solaris2.7 --prefix=xxx
3987 The following compiler flags must be specified in the configure
3988 step in order to bootstrap this target with the Sun compiler:
3991 % CC="cc -xarch=v9 -xildoff" @var{srcdir}/configure [@var{options}] [@var{target}]
3994 @option{-xarch=v9} specifies the SPARC-V9 architecture to the Sun toolchain
3995 and @option{-xildoff} turns off the incremental linker.
4000 @heading @anchor{sparcv9-x-solaris2}sparcv9-*-solaris2*
4002 This is a synonym for sparc64-*-solaris2*.
4007 @heading @anchor{x-x-vxworks}*-*-vxworks*
4008 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4009 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4010 We welcome patches for other architectures supported by VxWorks 5.5.
4011 Support for VxWorks AE would also be welcome; we believe this is merely
4012 a matter of writing an appropriate ``configlette'' (see below). We are
4013 not interested in supporting older, a.out or COFF-based, versions of
4016 VxWorks comes with an older version of GCC installed in
4017 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4018 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4019 Before running @command{configure}, create the directories @file{@var{prefix}}
4020 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4021 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4022 include that directory while running both @command{configure} and
4025 You must give @command{configure} the
4026 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4027 find the VxWorks system headers. Since VxWorks is a cross compilation
4028 target only, you must also specify @option{--target=@var{target}}.
4029 @command{configure} will attempt to create the directory
4030 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4031 make sure the user running @command{configure} has sufficient privilege
4034 GCC's exception handling runtime requires a special ``configlette''
4035 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4036 that file to add the module to your kernel build. (Future versions of
4037 VxWorks will incorporate this module.)
4042 @heading @anchor{x86-64-x-x}x86_64-*-*, amd64-*-*
4044 GCC supports the x86-64 architecture implemented by the AMD64 processor
4045 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4046 On GNU/Linux the default is a bi-arch compiler which is able to generate
4047 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4052 @heading @anchor{xtensa-x-elf}xtensa*-*-elf
4054 This target is intended for embedded Xtensa systems using the
4055 @samp{newlib} C library. It uses ELF but does not support shared
4056 objects. Designed-defined instructions specified via the
4057 Tensilica Instruction Extension (TIE) language are only supported
4058 through inline assembly.
4060 The Xtensa configuration information must be specified prior to
4061 building GCC@. The @file{include/xtensa-config.h} header
4062 file contains the configuration information. If you created your
4063 own Xtensa configuration with the Xtensa Processor Generator, the
4064 downloaded files include a customized copy of this header file,
4065 which you can use to replace the default header file.
4070 @heading @anchor{xtensa-x-linux}xtensa*-*-linux*
4072 This target is for Xtensa systems running GNU/Linux. It supports ELF
4073 shared objects and the GNU C library (glibc). It also generates
4074 position-independent code (PIC) regardless of whether the
4075 @option{-fpic} or @option{-fPIC} options are used. In other
4076 respects, this target is the same as the
4077 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
4082 @heading @anchor{windows}Microsoft Windows
4084 @subheading Intel 16-bit versions
4085 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
4088 However, the 32-bit port has limited support for Microsoft
4089 Windows 3.11 in the Win32s environment, as a target only. See below.
4091 @subheading Intel 32-bit versions
4093 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
4094 XP, and Windows Vista, are supported by several different target
4095 platforms. These targets differ in which Windows subsystem they target
4096 and which C libraries are used.
4099 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
4100 Linux API emulation layer in the Win32 subsystem.
4101 @item Interix @uref{#x-x-interix,,*-*-interix}: The Interix subsystem
4102 provides native support for POSIX.
4103 @item MinGW @uref{#x-x-mingw,,*-*-mingw}: MinGW is a native GCC port for
4104 the Win32 subsystem that provides a subset of POSIX.
4105 @item MKS i386-pc-mks: NuTCracker from MKS. See
4106 @uref{http://www.mkssoftware.com/} for more information.
4109 @subheading Intel 64-bit versions
4111 GCC contains support for x86-64 using the mingw-w64
4112 runtime library, available from @uref{http://mingw-w64.sourceforge.net/}.
4113 This library should be used with the target triple x86_64-pc-mingw32.
4115 Presently Windows for Itanium is not supported.
4117 @subheading Windows CE
4119 Windows CE is supported as a target only on ARM (arm-wince-pe), Hitachi
4120 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
4122 @subheading Other Windows Platforms
4124 GCC no longer supports Windows NT on the Alpha or PowerPC.
4126 GCC no longer supports the Windows POSIX subsystem. However, it does
4127 support the Interix subsystem. See above.
4129 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
4131 PW32 (i386-pc-pw32) support was never completed, and the project seems to
4132 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
4134 UWIN support has been removed due to a lack of maintenance.
4139 @heading @anchor{x-x-cygwin}*-*-cygwin
4141 Ports of GCC are included with the
4142 @uref{http://www.cygwin.com/,,Cygwin environment}.
4144 GCC will build under Cygwin without modification; it does not build
4145 with Microsoft's C++ compiler and there are no plans to make it do so.
4147 Cygwin can be compiled with i?86-pc-cygwin.
4152 @heading @anchor{x-x-interix}*-*-interix
4154 The Interix target is used by OpenNT, Interix, Services For UNIX (SFU),
4155 and Subsystem for UNIX-based Applications (SUA). Applications compiled
4156 with this target run in the Interix subsystem, which is separate from
4157 the Win32 subsystem. This target was last known to work in GCC 3.3.
4159 For more information, see @uref{http://www.interix.com/}.
4164 @heading @anchor{x-x-mingw32}*-*-mingw32
4166 GCC will build with and support only MinGW runtime 3.12 and later.
4167 Earlier versions of headers are incompatible with the new default semantics
4168 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
4173 @heading @anchor{os2}OS/2
4175 GCC does not currently support OS/2. However, Andrew Zabolotny has been
4176 working on a generic OS/2 port with pgcc. The current code can be found
4177 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
4182 @heading @anchor{older}Older systems
4184 GCC contains support files for many older (1980s and early
4185 1990s) Unix variants. For the most part, support for these systems
4186 has not been deliberately removed, but it has not been maintained for
4187 several years and may suffer from bitrot.
4189 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
4190 Support for these systems is still present in that release, but
4191 @command{configure} will fail unless the @option{--enable-obsolete}
4192 option is given. Unless a maintainer steps forward, support for these
4193 systems will be removed from the next release of GCC@.
4195 Support for old systems as hosts for GCC can cause problems if the
4196 workarounds for compiler, library and operating system bugs affect the
4197 cleanliness or maintainability of the rest of GCC@. In some cases, to
4198 bring GCC up on such a system, if still possible with current GCC, may
4199 require first installing an old version of GCC which did work on that
4200 system, and using it to compile a more recent GCC, to avoid bugs in the
4201 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
4202 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
4203 sites}. Header bugs may generally be avoided using
4204 @command{fixincludes}, but bugs or deficiencies in libraries and the
4205 operating system may still cause problems.
4207 Support for older systems as targets for cross-compilation is less
4208 problematic than support for them as hosts for GCC; if an enthusiast
4209 wishes to make such a target work again (including resurrecting any of
4210 the targets that never worked with GCC 2, starting from the last
4211 version before they were removed), patches
4212 @uref{../contribute.html,,following the usual requirements} would be
4213 likely to be accepted, since they should not affect the support for more
4216 For some systems, old versions of GNU binutils may also be useful,
4217 and are available from @file{pub/binutils/old-releases} on
4218 @uref{http://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
4220 Some of the information on specific systems above relates to
4221 such older systems, but much of the information
4222 about GCC on such systems (which may no longer be applicable to
4223 current GCC) is to be found in the GCC texinfo manual.
4228 @heading @anchor{elf}all ELF targets (SVR4, Solaris 2, etc.)
4230 C++ support is significantly better on ELF targets if you use the
4231 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
4232 inlines, vtables and template instantiations will be discarded
4241 @uref{./index.html,,Return to the GCC Installation page}
4245 @c ***Old documentation******************************************************
4247 @include install-old.texi
4253 @uref{./index.html,,Return to the GCC Installation page}
4257 @c ***GFDL********************************************************************
4265 @uref{./index.html,,Return to the GCC Installation page}
4269 @c ***************************************************************************
4270 @c Part 6 The End of the Document
4272 @comment node-name, next, previous, up
4273 @node Concept Index, , GNU Free Documentation License, Top
4277 @unnumbered Concept Index