1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename gccinstall.info
5 @settitle Installing GCC
10 @c Specify title for specific html page
12 @settitle Installing GCC
15 @settitle Host/Target specific installation notes for GCC
17 @ifset prerequisiteshtml
18 @settitle Prerequisites for GCC
21 @settitle Downloading GCC
24 @settitle Installing GCC: Configuration
27 @settitle Installing GCC: Building
30 @settitle Installing GCC: Testing
32 @ifset finalinstallhtml
33 @settitle Installing GCC: Final installation
36 @settitle Installing GCC: Binaries
39 @settitle Installing GCC: Old documentation
42 @settitle Installing GCC: GNU Free Documentation License
45 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
46 @c 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
47 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
49 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
50 @c test the generation of HTML documents for the gcc.gnu.org web pages.
52 @c Do not use @footnote{} in this file as it breaks install.texi2html!
54 @c Include everything if we're not making html
58 @set prerequisiteshtml
69 @c Part 2 Summary Description and Copyright
71 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
72 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
74 Permission is granted to copy, distribute and/or modify this document
75 under the terms of the GNU Free Documentation License, Version 1.2 or
76 any later version published by the Free Software Foundation; with no
77 Invariant Sections, the Front-Cover texts being (a) (see below), and
78 with the Back-Cover Texts being (b) (see below). A copy of the
79 license is included in the section entitled ``@uref{./gfdl.html,,GNU
80 Free Documentation License}''.
82 (a) The FSF's Front-Cover Text is:
86 (b) The FSF's Back-Cover Text is:
88 You have freedom to copy and modify this GNU Manual, like GNU
89 software. Copies published by the Free Software Foundation raise
90 funds for GNU development.
95 @dircategory Software development
97 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
100 @c Part 3 Titlepage and Copyright
103 @comment The title is printed in a large font.
104 @center @titlefont{Installing GCC}
106 @c The following two commands start the copyright page.
108 @vskip 0pt plus 1filll
112 @c Part 4 Top node and Master Menu
115 @comment node-name, next, Previous, up
118 * Installing GCC:: This document describes the generic installation
119 procedure for GCC as well as detailing some target
120 specific installation instructions.
122 * Specific:: Host/target specific installation notes for GCC.
123 * Binaries:: Where to get pre-compiled binaries.
125 * Old:: Old installation documentation.
127 * GNU Free Documentation License:: How you can copy and share this manual.
128 * Concept Index:: This index has two entries.
132 @c Part 5 The Body of the Document
133 @c ***Installing GCC**********************************************************
135 @comment node-name, next, previous, up
136 @node Installing GCC, Binaries, , Top
140 @chapter Installing GCC
143 The latest version of this document is always available at
144 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
146 This document describes the generic installation procedure for GCC as well
147 as detailing some target specific installation instructions.
149 GCC includes several components that previously were separate distributions
150 with their own installation instructions. This document supersedes all
151 package specific installation instructions.
153 @emph{Before} starting the build/install procedure please check the
155 @ref{Specific, host/target specific installation notes}.
158 @uref{specific.html,,host/target specific installation notes}.
160 We recommend you browse the entire generic installation instructions before
163 Lists of successful builds for released versions of GCC are
164 available at @uref{http://gcc.gnu.org/buildstat.html}.
165 These lists are updated as new information becomes available.
167 The installation procedure itself is broken into five steps.
172 * Downloading the source::
175 * Testing:: (optional)
182 @uref{prerequisites.html,,Prerequisites}
184 @uref{download.html,,Downloading the source}
186 @uref{configure.html,,Configuration}
188 @uref{build.html,,Building}
190 @uref{test.html,,Testing} (optional)
192 @uref{finalinstall.html,,Final install}
196 Please note that GCC does not support @samp{make uninstall} and probably
197 won't do so in the near future as this would open a can of worms. Instead,
198 we suggest that you install GCC into a directory of its own and simply
199 remove that directory when you do not need that specific version of GCC
200 any longer, and, if shared libraries are installed there as well, no
201 more binaries exist that use them.
204 There are also some @uref{old.html,,old installation instructions},
205 which are mostly obsolete but still contain some information which has
206 not yet been merged into the main part of this manual.
214 @uref{./index.html,,Return to the GCC Installation page}
220 @c ***Prerequisites**************************************************
222 @comment node-name, next, previous, up
223 @node Prerequisites, Downloading the source, , Installing GCC
225 @ifset prerequisiteshtml
227 @chapter Prerequisites
229 @cindex Prerequisites
231 GCC requires that various tools and packages be available for use in the
232 build procedure. Modifying GCC sources requires additional tools
235 @heading Tools/packages necessary for building GCC
237 @item ISO C90 compiler
238 Necessary to bootstrap GCC, although versions of GCC prior
239 to 3.4 also allow bootstrapping with a traditional (K&R) C compiler.
241 To build all languages in a cross-compiler or other configuration where
242 3-stage bootstrap is not performed, you need to start with an existing
243 GCC binary (version 2.95 or later) because source code for language
244 frontends other than C might use GCC extensions.
248 In order to build the Ada compiler (GNAT) you must already have GNAT
249 installed because portions of the Ada frontend are written in Ada (with
250 GNAT extensions.) Refer to the Ada installation instructions for more
251 specific information.
253 @item A ``working'' POSIX compatible shell, or GNU bash
255 Necessary when running @command{configure} because some
256 @command{/bin/sh} shells have bugs and may crash when configuring the
257 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
258 have disastrous corner-case performance problems. This
259 can cause target @command{configure} runs to literally take days to
260 complete in some cases.
262 So on some platforms @command{/bin/ksh} is sufficient, on others it
263 isn't. See the host/target specific instructions for your platform, or
264 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
265 environment to your ``good'' shell prior to running
266 @command{configure}/@command{make}.
268 @command{zsh} is not a fully compliant POSIX shell and will not
269 work when configuring GCC@.
273 Necessary in some circumstances, optional in others. See the
274 host/target specific instructions for your platform for the exact
277 @item gzip version 1.2.4 (or later) or
278 @itemx bzip2 version 1.0.2 (or later)
280 Necessary to uncompress GCC @command{tar} files when source code is
281 obtained via FTP mirror sites.
283 @item GNU make version 3.79.1 (or later)
285 You must have GNU make installed to build GCC@.
287 @item GNU tar version 1.14 (or later)
289 Necessary (only on some platforms) to untar the source code. Many
290 systems' @command{tar} programs will also work, only try GNU
291 @command{tar} if you have problems.
293 @item GNU Multiple Precision Library (GMP) version 4.1 (or later)
295 Necessary to build the Fortran frontend. If you do not have it
296 installed in your library search path, you will have to configure with
297 the @option{--with-gmp} or @option{--with-gmp-dir} configure option.
299 @item MPFR Library version 2.2 (or later)
301 Necessary to build the Fortran frontend. It can be downloaded from
302 @uref{http://www.mpfr.org/}. The version of MPFR that is bundled with
303 GMP 4.1.x contains numerous bugs. Although gfortran will appear
304 to function with the buggy verions of MPFR, there are few gfortran bugs
305 that will not be fixed to use this version. It is strongly
306 recommended to upgrade to at least MPFR version 2.2.
308 The @option{--with-mpfr} or @option{--with-mpfr-dir} configure option should
309 be used if your MPFR Library is not installed in your library search path.
311 @item @command{jar}, or InfoZIP (@command{zip} and @command{unzip})
313 Necessary to build libgcj, the GCJ runtime.
318 @heading Tools/packages necessary for modifying GCC
320 @item autoconf versions 2.13 and 2.59
321 @itemx GNU m4 version 1.4 (or later)
323 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
324 to regenerate @file{configure} and @file{config.in} files. Most
325 directories require autoconf 2.59 (exactly), but the toplevel
326 still requires autoconf 2.13 (exactly).
328 @item automake versions 1.9.3
330 Necessary when modifying a @file{Makefile.am} file to regenerate its
331 associated @file{Makefile.in}.
333 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
334 file. Specifically this applies to the @file{gcc}, @file{intl},
335 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
336 as any of their subdirectories.
338 For directories that use automake, GCC requires the latest release in
339 the 1.9.x series, which is currently 1.9.3. When regenerating a directory
340 to a newer version, please update all the directories using an older 1.9.x
341 to the latest released version.
343 @item gettext version 0.14.5 (or later)
345 Needed to regenerate @file{gcc.pot}.
347 @item gperf version 2.7.2 (or later)
349 Necessary when modifying @command{gperf} input files, e.g.@:
350 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
351 @file{gcc/cp/cfns.h}.
357 Necessary to run the GCC testsuite; see the section on testing for details.
359 @item autogen version 5.5.4 (or later) and
360 @itemx guile version 1.4.1 (or later)
362 Necessary to regenerate @file{fixinc/fixincl.x} from
363 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
365 Necessary to run @samp{make check} for @file{fixinc}.
367 Necessary to regenerate the top level @file{Makefile.in} file from
368 @file{Makefile.tpl} and @file{Makefile.def}.
370 @item GNU Bison version 1.28 (or later)
371 Berkeley @command{yacc} (@command{byacc}) is also reported to work other
374 Necessary when modifying @file{*.y} files.
376 Necessary to build GCC during development because the generated output
377 files are not included in the SVN repository. They are included in
380 @item Flex version 2.5.4 (or later)
382 Necessary when modifying @file{*.l} files.
384 Necessary to build GCC during development because the generated output
385 files are not included in the SVN repository. They are included in
388 @item Texinfo version 4.4 (or later)
390 Necessary for running @command{makeinfo} when modifying @file{*.texi}
391 files to test your changes.
393 Necessary to build GCC documentation during development because the
394 generated output files are not included in the SVN repository. They are
395 included in releases.
397 @item @TeX{} (any working version)
399 Necessary for running @command{texi2dvi}, used when running
400 @command{make dvi} to create DVI files.
402 @item SVN (any version)
403 @itemx SSH (any version)
405 Necessary to access the SVN repository. Public releases and weekly
406 snapshots of the development sources are also available via FTP@.
408 @item Perl version 5.6.1 (or later)
410 Necessary when regenerating @file{Makefile} dependencies in libiberty.
411 Necessary when regenerating @file{libiberty/functions.texi}.
412 Necessary when generating manpages from Texinfo manuals.
413 Necessary when targetting Darwin, building libstdc++,
414 and not using @option{--disable-symvers}.
415 Used by various scripts to generate some files included in SVN (mainly
416 Unicode-related and rarely changing) from source tables.
418 @item GNU diffutils version 2.7 (or later)
420 Useful when submitting patches for the GCC source code.
422 @item patch version 2.5.4 (or later)
424 Necessary when applying patches, created with @command{diff}, to one's
434 @uref{./index.html,,Return to the GCC Installation page}
438 @c ***Downloading the source**************************************************
440 @comment node-name, next, previous, up
441 @node Downloading the source, Configuration, Prerequisites, Installing GCC
445 @chapter Downloading GCC
447 @cindex Downloading GCC
448 @cindex Downloading the Source
450 GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP
451 tarballs compressed with @command{gzip} or
452 @command{bzip2}. It is possible to download a full distribution or specific
455 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
456 for information on how to obtain GCC@.
458 The full distribution includes the C, C++, Objective-C, Fortran 77, Fortran
459 (in case of GCC 4.0 and later), Java, and Ada (in case of GCC 3.1 and later)
460 compilers. The full distribution also includes runtime libraries for C++,
461 Objective-C, Fortran 77, Fortran, and Java. In GCC 3.0 and later versions,
462 GNU compiler testsuites are also included in the full distribution.
464 If you choose to download specific components, you must download the core
465 GCC distribution plus any language specific distributions you wish to
466 use. The core distribution includes the C language front end as well as the
467 shared components. Each language has a tarball which includes the language
468 front end as well as the language runtime (when appropriate).
470 Unpack the core distribution as well as any language specific
471 distributions in the same directory.
473 If you also intend to build binutils (either to upgrade an existing
474 installation or for use in place of the corresponding tools of your
475 OS), unpack the binutils distribution either in the same directory or
476 a separate one. In the latter case, add symbolic links to any
477 components of the binutils you intend to build alongside the compiler
478 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
479 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
486 @uref{./index.html,,Return to the GCC Installation page}
490 @c ***Configuration***********************************************************
492 @comment node-name, next, previous, up
493 @node Configuration, Building, Downloading the source, Installing GCC
497 @chapter Installing GCC: Configuration
499 @cindex Configuration
500 @cindex Installing GCC: Configuration
502 Like most GNU software, GCC must be configured before it can be built.
503 This document describes the recommended configuration procedure
504 for both native and cross targets.
506 We use @var{srcdir} to refer to the toplevel source directory for
507 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
509 If you obtained the sources via SVN, @var{srcdir} must refer to the top
510 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
511 and not its @file{gcc} subdirectory, otherwise the build will fail.
513 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
514 file system, the shell's built-in @command{pwd} command will return
515 temporary pathnames. Using these can lead to various sorts of build
516 problems. To avoid this issue, set the @env{PWDCMD} environment
517 variable to an automounter-aware @command{pwd} command, e.g.,
518 @command{pawd} or @samp{amq -w}, during the configuration and build
521 First, we @strong{highly} recommend that GCC be built into a
522 separate directory than the sources which does @strong{not} reside
523 within the source tree. This is how we generally build GCC; building
524 where @var{srcdir} == @var{objdir} should still work, but doesn't
525 get extensive testing; building where @var{objdir} is a subdirectory
526 of @var{srcdir} is unsupported.
528 If you have previously built GCC in the same directory for a
529 different target machine, do @samp{make distclean} to delete all files
530 that might be invalid. One of the files this deletes is @file{Makefile};
531 if @samp{make distclean} complains that @file{Makefile} does not exist
532 or issues a message like ``don't know how to make distclean'' it probably
533 means that the directory is already suitably clean. However, with the
534 recommended method of building in a separate @var{objdir}, you should
535 simply use a different @var{objdir} for each target.
537 Second, when configuring a native system, either @command{cc} or
538 @command{gcc} must be in your path or you must set @env{CC} in
539 your environment before running configure. Otherwise the configuration
543 Note that the bootstrap compiler and the resulting GCC must be link
544 compatible, else the bootstrap will fail with linker errors about
545 incompatible object file formats. Several multilibed targets are
546 affected by this requirement, see
548 @ref{Specific, host/target specific installation notes}.
551 @uref{specific.html,,host/target specific installation notes}.
560 % @var{srcdir}/configure [@var{options}] [@var{target}]
564 @heading Target specification
567 GCC has code to correctly determine the correct value for @var{target}
568 for nearly all native systems. Therefore, we highly recommend you not
569 provide a configure target when configuring a native compiler.
572 @var{target} must be specified as @option{--target=@var{target}}
573 when configuring a cross compiler; examples of valid targets would be
574 m68k-coff, sh-elf, etc.
577 Specifying just @var{target} instead of @option{--target=@var{target}}
578 implies that the host defaults to @var{target}.
582 @heading Options specification
584 Use @var{options} to override several configure time options for
585 GCC@. A list of supported @var{options} follows; @samp{configure
586 --help} may list other options, but those not listed below may not
587 work and should not normally be used.
589 Note that each @option{--enable} option has a corresponding
590 @option{--disable} option and that each @option{--with} option has a
591 corresponding @option{--without} option.
594 @item --prefix=@var{dirname}
595 Specify the toplevel installation
596 directory. This is the recommended way to install the tools into a directory
597 other than the default. The toplevel installation directory defaults to
600 We @strong{highly} recommend against @var{dirname} being the same or a
601 subdirectory of @var{objdir} or vice versa. If specifying a directory
602 beneath a user's home directory tree, some shells will not expand
603 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
606 The following standard @command{autoconf} options are supported. Normally you
607 should not need to use these options.
609 @item --exec-prefix=@var{dirname}
610 Specify the toplevel installation directory for architecture-dependent
611 files. The default is @file{@var{prefix}}.
613 @item --bindir=@var{dirname}
614 Specify the installation directory for the executables called by users
615 (such as @command{gcc} and @command{g++}). The default is
616 @file{@var{exec-prefix}/bin}.
618 @item --libdir=@var{dirname}
619 Specify the installation directory for object code libraries and
620 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
622 @item --libexecdir=@var{dirname}
623 Specify the installation directory for internal executables of GCC@.
624 The default is @file{@var{exec-prefix}/libexec}.
626 @item --with-slibdir=@var{dirname}
627 Specify the installation directory for the shared libgcc library. The
628 default is @file{@var{libdir}}.
630 @item --infodir=@var{dirname}
631 Specify the installation directory for documentation in info format.
632 The default is @file{@var{prefix}/info}.
634 @item --datadir=@var{dirname}
635 Specify the installation directory for some architecture-independent
636 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
638 @item --mandir=@var{dirname}
639 Specify the installation directory for manual pages. The default is
640 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
641 the full GCC manuals, which are provided in Texinfo format. The manpages
642 are derived by an automatic conversion process from parts of the full
645 @item --with-gxx-include-dir=@var{dirname}
647 the installation directory for G++ header files. The default is
648 @file{@var{prefix}/include/c++/@var{version}}.
652 @item --program-prefix=@var{prefix}
653 GCC supports some transformations of the names of its programs when
654 installing them. This option prepends @var{prefix} to the names of
655 programs to install in @var{bindir} (see above). For example, specifying
656 @option{--program-prefix=foo-} would result in @samp{gcc}
657 being installed as @file{/usr/local/bin/foo-gcc}.
659 @item --program-suffix=@var{suffix}
660 Appends @var{suffix} to the names of programs to install in @var{bindir}
661 (see above). For example, specifying @option{--program-suffix=-3.1}
662 would result in @samp{gcc} being installed as
663 @file{/usr/local/bin/gcc-3.1}.
665 @item --program-transform-name=@var{pattern}
666 Applies the @samp{sed} script @var{pattern} to be applied to the names
667 of programs to install in @var{bindir} (see above). @var{pattern} has to
668 consist of one or more basic @samp{sed} editing commands, separated by
669 semicolons. For example, if you want the @samp{gcc} program name to be
670 transformed to the installed program @file{/usr/local/bin/myowngcc} and
671 the @samp{g++} program name to be transformed to
672 @file{/usr/local/bin/gspecial++} without changing other program names,
673 you could use the pattern
674 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
675 to achieve this effect.
677 All three options can be combined and used together, resulting in more
678 complex conversion patterns. As a basic rule, @var{prefix} (and
679 @var{suffix}) are prepended (appended) before further transformations
680 can happen with a special transformation script @var{pattern}.
682 As currently implemented, this option only takes effect for native
683 builds; cross compiler binaries' names are not transformed even when a
684 transformation is explicitly asked for by one of these options.
686 For native builds, some of the installed programs are also installed
687 with the target alias in front of their name, as in
688 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
689 before the target alias is prepended to the name---so, specifying
690 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
691 resulting binary would be installed as
692 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
694 As a last shortcoming, none of the installed Ada programs are
695 transformed yet, which will be fixed in some time.
697 @item --with-local-prefix=@var{dirname}
699 installation directory for local include files. The default is
700 @file{/usr/local}. Specify this option if you want the compiler to
701 search directory @file{@var{dirname}/include} for locally installed
702 header files @emph{instead} of @file{/usr/local/include}.
704 You should specify @option{--with-local-prefix} @strong{only} if your
705 site has a different convention (not @file{/usr/local}) for where to put
708 The default value for @option{--with-local-prefix} is @file{/usr/local}
709 regardless of the value of @option{--prefix}. Specifying
710 @option{--prefix} has no effect on which directory GCC searches for
711 local header files. This may seem counterintuitive, but actually it is
714 The purpose of @option{--prefix} is to specify where to @emph{install
715 GCC}. The local header files in @file{/usr/local/include}---if you put
716 any in that directory---are not part of GCC@. They are part of other
717 programs---perhaps many others. (GCC installs its own header files in
718 another directory which is based on the @option{--prefix} value.)
720 Both the local-prefix include directory and the GCC-prefix include
721 directory are part of GCC's ``system include'' directories. Although these
722 two directories are not fixed, they need to be searched in the proper
723 order for the correct processing of the include_next directive. The
724 local-prefix include directory is searched before the GCC-prefix
725 include directory. Another characteristic of system include directories
726 is that pedantic warnings are turned off for headers in these directories.
728 Some autoconf macros add @option{-I @var{directory}} options to the
729 compiler command line, to ensure that directories containing installed
730 packages' headers are searched. When @var{directory} is one of GCC's
731 system include directories, GCC will ignore the option so that system
732 directories continue to be processed in the correct order. This
733 may result in a search order different from what was specified but the
734 directory will still be searched.
736 GCC automatically searches for ordinary libraries using
737 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
738 used for both GCC and packages, GCC will automatically search for
739 both headers and libraries. This provides a configuration that is
740 easy to use. GCC behaves in a manner similar to that when it is
741 installed as a system compiler in @file{/usr}.
743 Sites that need to install multiple versions of GCC may not want to
744 use the above simple configuration. It is possible to use the
745 @option{--program-prefix}, @option{--program-suffix} and
746 @option{--program-transform-name} options to install multiple versions
747 into a single directory, but it may be simpler to use different prefixes
748 and the @option{--with-local-prefix} option to specify the location of the
749 site-specific files for each version. It will then be necessary for
750 users to specify explicitly the location of local site libraries
751 (e.g., with @env{LIBRARY_PATH}).
753 The same value can be used for both @option{--with-local-prefix} and
754 @option{--prefix} provided it is not @file{/usr}. This can be used
755 to avoid the default search of @file{/usr/local/include}.
757 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
758 The directory you use for @option{--with-local-prefix} @strong{must not}
759 contain any of the system's standard header files. If it did contain
760 them, certain programs would be miscompiled (including GNU Emacs, on
761 certain targets), because this would override and nullify the header
762 file corrections made by the @command{fixincludes} script.
764 Indications are that people who use this option use it based on mistaken
765 ideas of what it is for. People use it as if it specified where to
766 install part of GCC@. Perhaps they make this assumption because
767 installing GCC creates the directory.
769 @item --enable-shared[=@var{package}[,@dots{}]]
770 Build shared versions of libraries, if shared libraries are supported on
771 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
772 are enabled by default on all platforms that support shared libraries.
774 If a list of packages is given as an argument, build shared libraries
775 only for the listed packages. For other packages, only static libraries
776 will be built. Package names currently recognized in the GCC tree are
777 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
778 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
779 @samp{ada}, @samp{libada}, @samp{libjava} and @samp{libobjc}.
780 Note @samp{libiberty} does not support shared libraries at all.
782 Use @option{--disable-shared} to build only static libraries. Note that
783 @option{--disable-shared} does not accept a list of package names as
784 argument, only @option{--enable-shared} does.
786 @item @anchor{with-gnu-as}--with-gnu-as
787 Specify that the compiler should assume that the
788 assembler it finds is the GNU assembler. However, this does not modify
789 the rules to find an assembler and will result in confusion if the
790 assembler found is not actually the GNU assembler. (Confusion may also
791 result if the compiler finds the GNU assembler but has not been
792 configured with @option{--with-gnu-as}.) If you have more than one
793 assembler installed on your system, you may want to use this option in
794 connection with @option{--with-as=@var{pathname}} or
795 @option{--with-build-time-tools=@var{pathname}}.
797 The following systems are the only ones where it makes a difference
798 whether you use the GNU assembler. On any other system,
799 @option{--with-gnu-as} has no effect.
802 @item @samp{hppa1.0-@var{any}-@var{any}}
803 @item @samp{hppa1.1-@var{any}-@var{any}}
804 @item @samp{i386-@var{any}-sysv}
805 @item @samp{m68k-bull-sysv}
806 @item @samp{m68k-hp-hpux}
807 @item @samp{m68000-hp-hpux}
808 @item @samp{m68000-att-sysv}
809 @item @samp{sparc-sun-solaris2.@var{any}}
810 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
813 On the systems listed above (except for the HP-PA, the SPARC, for ISC on
814 the 386, if you use the GNU assembler, you should also use the GNU linker
815 (and specify @option{--with-gnu-ld}).
817 @item @anchor{with-as}--with-as=@var{pathname}
818 Specify that the compiler should use the assembler pointed to by
819 @var{pathname}, rather than the one found by the standard rules to find
820 an assembler, which are:
823 Unless GCC is being built with a cross compiler, check the
824 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
825 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
826 @var{exec-prefix} defaults to @var{prefix}, which
827 defaults to @file{/usr/local} unless overridden by the
828 @option{--prefix=@var{pathname}} switch described above. @var{target}
829 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
830 @var{version} denotes the GCC version, such as 3.0.
833 If the target system is the same that you are building on, check
834 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
838 Check in the @env{PATH} for a tool whose name is prefixed by the
839 target system triple.
842 Check in the @env{PATH} for a tool whose name is not prefixed by the
843 target system triple, if the host and target system triple are
844 the same (in other words, we use a host tool if it can be used for
848 You may want to use @option{--with-as} if no assembler
849 is installed in the directories listed above, or if you have multiple
850 assemblers installed and want to choose one that is not found by the
853 @item @anchor{with-gnu-ld}--with-gnu-ld
854 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
857 @item --with-ld=@var{pathname}
858 Same as @uref{#with-as,,@option{--with-as}}
862 Specify that stabs debugging
863 information should be used instead of whatever format the host normally
864 uses. Normally GCC uses the same debug format as the host system.
866 On MIPS based systems and on Alphas, you must specify whether you want
867 GCC to create the normal ECOFF debugging format, or to use BSD-style
868 stabs passed through the ECOFF symbol table. The normal ECOFF debug
869 format cannot fully handle languages other than C@. BSD stabs format can
870 handle other languages, but it only works with the GNU debugger GDB@.
872 Normally, GCC uses the ECOFF debugging format by default; if you
873 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
875 No matter which default you choose when you configure GCC, the user
876 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
877 the debug format for a particular compilation.
879 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
880 @option{--with-gas} is used. It selects use of stabs debugging
881 information embedded in COFF output. This kind of debugging information
882 supports C++ well; ordinary COFF debugging information does not.
884 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
885 selects use of stabs debugging information embedded in ELF output. The
886 C++ compiler currently (2.6.0) does not support the DWARF debugging
887 information normally used on 386 SVR4 platforms; stabs provide a
888 workable alternative. This requires gas and gdb, as the normal SVR4
889 tools can not generate or interpret stabs.
891 @item --disable-multilib
892 Specify that multiple target
893 libraries to support different target variants, calling
894 conventions, etc should not be built. The default is to build a
895 predefined set of them.
897 Some targets provide finer-grained control over which multilibs are built
898 (e.g., @option{--disable-softfloat}):
904 fpu, 26bit, underscore, interwork, biendian, nofmult.
907 softfloat, m68881, m68000, m68020.
910 single-float, biendian, softfloat.
912 @item powerpc*-*-*, rs6000*-*-*
913 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
918 @item --enable-threads
919 Specify that the target
920 supports threads. This affects the Objective-C compiler and runtime
921 library, and exception handling for other languages like C++ and Java.
922 On some systems, this is the default.
924 In general, the best (and, in many cases, the only known) threading
925 model available will be configured for use. Beware that on some
926 systems, GCC has not been taught what threading models are generally
927 available for the system. In this case, @option{--enable-threads} is an
928 alias for @option{--enable-threads=single}.
930 @item --disable-threads
931 Specify that threading support should be disabled for the system.
932 This is an alias for @option{--enable-threads=single}.
934 @item --enable-threads=@var{lib}
936 @var{lib} is the thread support library. This affects the Objective-C
937 compiler and runtime library, and exception handling for other languages
938 like C++ and Java. The possibilities for @var{lib} are:
946 Ada tasking support. For non-Ada programs, this setting is equivalent
947 to @samp{single}. When used in conjunction with the Ada run time, it
948 causes GCC to use the same thread primitives as Ada uses. This option
949 is necessary when using both Ada and the back end exception handling,
950 which is the default for most Ada targets.
952 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
953 that the file needed to support this configuration, @file{gthr-mach.h}, is
954 missing and thus this setting will cause a known bootstrap failure.)
956 This is an alias for @samp{single}.
958 Generic POSIX/Unix98 thread support.
960 Generic POSIX/Unix95 thread support.
962 RTEMS thread support.
964 Disable thread support, should work for all platforms.
966 Sun Solaris 2 thread support.
968 VxWorks thread support.
970 Microsoft Win32 API thread support.
972 Novell Kernel Services thread support.
976 Specify that the target supports TLS (Thread Local Storage). Usually
977 configure can correctly determine if TLS is supported. In cases where
978 it guesses incorrectly, TLS can be explicitly enabled or disabled with
979 @option{--enable-tls} or @option{--disable-tls}. This can happen if
980 the assembler supports TLS but the C library does not, or if the
981 assumptions made by the configure test are incorrect.
984 Specify that the target does not support TLS.
985 This is an alias for @option{--enable-tls=no}.
987 @item --with-cpu=@var{cpu}
988 Specify which cpu variant the compiler should generate code for by default.
989 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
990 This option is only supported on some targets, including ARM, i386, PowerPC,
993 @item --with-schedule=@var{cpu}
994 @itemx --with-arch=@var{cpu}
995 @itemx --with-tune=@var{cpu}
996 @itemx --with-abi=@var{abi}
997 @itemx --with-fpu=@var{type}
998 @itemx --with-float=@var{type}
999 These configure options provide default values for the @option{-mschedule=},
1000 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1001 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1002 @option{--with-cpu}, which switches will be accepted and acceptable values
1003 of the arguments depend on the target.
1005 @item --with-mode=@var{mode}
1006 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1007 This option is only supported on ARM targets.
1009 @item --with-divide=@var{type}
1010 Specify how the compiler should generate code for checking for
1011 division by zero. This option is only supported on the MIPS target.
1012 The possibilities for @var{type} are:
1015 Division by zero checks use conditional traps (this is the default on
1016 systems that support conditional traps).
1018 Division by zero checks use the break instruction.
1021 @item --enable-__cxa_atexit
1022 Define if you want to use __cxa_atexit, rather than atexit, to
1023 register C++ destructors for local statics and global objects.
1024 This is essential for fully standards-compliant handling of
1025 destructors, but requires __cxa_atexit in libc. This option is currently
1026 only available on systems with GNU libc. When enabled, this will cause
1027 @option{-fuse-cxa-exit} to be passed by default.
1029 @item --enable-target-optspace
1031 libraries should be optimized for code space instead of code speed.
1032 This is the default for the m32r platform.
1035 Specify that a user visible @command{cpp} program should not be installed.
1037 @item --with-cpp-install-dir=@var{dirname}
1038 Specify that the user visible @command{cpp} program should be installed
1039 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1041 @item --enable-initfini-array
1042 Force the use of sections @code{.init_array} and @code{.fini_array}
1043 (instead of @code{.init} and @code{.fini}) for constructors and
1044 destructors. Option @option{--disable-initfini-array} has the
1045 opposite effect. If neither option is specified, the configure script
1046 will try to guess whether the @code{.init_array} and
1047 @code{.fini_array} sections are supported and, if they are, use them.
1049 @item --enable-maintainer-mode
1050 The build rules that
1051 regenerate the GCC master message catalog @file{gcc.pot} are normally
1052 disabled. This is because it can only be rebuilt if the complete source
1053 tree is present. If you have changed the sources and want to rebuild the
1054 catalog, configuring with @option{--enable-maintainer-mode} will enable
1055 this. Note that you need a recent version of the @code{gettext} tools
1058 @item --disable-bootstrap
1059 For a native build, the default configuration is to perform
1060 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1061 testing that GCC can compile itself correctly. If you want to disable
1062 this process, you can configure with @option{--disable-bootstrap}.
1064 @item --enable-bootstrap
1065 In special cases, you may want to perform a 3-stage build
1066 even if the target and host triplets are different.
1067 This could happen when the host can run code compiled for
1068 the target (e.g.@: host is i686-linux, target is i486-linux).
1069 Starting from GCC 4.2, to do this you have to configure explicitly
1070 with @option{--enable-bootstrap}.
1072 @item --enable-generated-files-in-srcdir
1073 Neither the .c and .h files that are generated from Bison and flex nor the
1074 info manuals and man pages that are built from the .texi files are present
1075 in the SVN development tree. When building GCC from that development tree,
1076 or from one of our snapshots, those generated files are placed in your
1077 build directory, which allows for the source to be in a readonly
1080 If you configure with @option{--enable-generated-files-in-srcdir} then those
1081 generated files will go into the source directory. This is mainly intended
1082 for generating release or prerelease tarballs of the GCC sources, since it
1083 is not a requirement that the users of source releases to have flex, Bison,
1086 @item --enable-version-specific-runtime-libs
1088 that runtime libraries should be installed in the compiler specific
1089 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1090 addition, @samp{libstdc++}'s include files will be installed into
1091 @file{@var{libdir}} unless you overruled it by using
1092 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1093 particularly useful if you intend to use several versions of GCC in
1094 parallel. This is currently supported by @samp{libgfortran},
1095 @samp{libjava}, @samp{libmudflap}, @samp{libstdc++}, and @samp{libobjc}.
1097 @item --with-java-home=@var{dirname}
1098 This @samp{libjava} option overrides the default value of the
1099 @samp{java.home} system property. It is also used to set
1100 @samp{sun.boot.class.path} to @file{@var{dirname}/lib/rt.jar}. By
1101 default @samp{java.home} is set to @file{@var{prefix}} and
1102 @samp{sun.boot.class.path} to
1103 @file{@var{datadir}/java/libgcj-@var{version}.jar}.
1105 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1106 Specify that only a particular subset of compilers and
1107 their runtime libraries should be built. For a list of valid values for
1108 @var{langN} you can issue the following command in the
1109 @file{gcc} directory of your GCC source tree:@*
1111 grep language= */config-lang.in
1113 Currently, you can use any of the following:
1114 @code{all}, @code{ada}, @code{c}, @code{c++}, @code{fortran}, @code{java},
1115 @code{objc}, @code{obj-c++}, @code{treelang}.
1116 Building the Ada compiler has special requirements, see below.
1117 If you do not pass this flag, or specify the option @code{all}, then all
1118 default languages available in the @file{gcc} sub-tree will be configured.
1119 Ada, Objective-C++, and treelang are not default languages; the rest are.
1120 Re-defining @code{LANGUAGES} when calling @samp{make} @strong{does not}
1121 work anymore, as those language sub-directories might not have been
1124 @item --disable-libada
1125 Specify that the run-time libraries and tools used by GNAT should not
1126 be built. This can be useful for debugging, or for compatibility with
1127 previous Ada build procedures, when it was required to explicitly
1128 do a @samp{make -C gcc gnatlib_and_tools}.
1130 @item --disable-libssp
1131 Specify that the run-time libraries for stack smashing protection
1132 should not be built.
1134 @item --disable-libgomp
1135 Specify that the run-time libraries used by GOMP should not be built.
1138 Specify that the compiler should
1139 use DWARF 2 debugging information as the default.
1141 @item --enable-targets=all
1142 @itemx --enable-targets=@var{target_list}
1143 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1144 These are compilers that are able to generate either 64-bit or 32-bit
1145 code. Typically, the corresponding 32-bit target, e.g.@:
1146 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1147 option enables the 32-bit target to be a bi-arch compiler, which is
1148 useful when you want a bi-arch compiler that defaults to 32-bit, and
1149 you are building a bi-arch or multi-arch binutils in a combined tree.
1150 Currently, this option only affects powerpc-linux.
1152 @item --enable-secureplt
1153 This option enables @option{-msecure-plt} by default for powerpc-linux.
1155 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1156 Using the GNU Compiler Collection (GCC)},
1159 See ``RS/6000 and PowerPC Options'' in the main manual
1162 @item --enable-win32-registry
1163 @itemx --enable-win32-registry=@var{key}
1164 @itemx --disable-win32-registry
1165 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1166 to look up installations paths in the registry using the following key:
1169 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1172 @var{key} defaults to GCC version number, and can be overridden by the
1173 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1174 who use custom installers are encouraged to provide a different key,
1175 perhaps one comprised of vendor name and GCC version number, to
1176 avoid conflict with existing installations. This feature is enabled
1177 by default, and can be disabled by @option{--disable-win32-registry}
1178 option. This option has no effect on the other hosts.
1181 Specify that the machine does not have a floating point unit. This
1182 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1183 system, @option{--nfp} has no effect.
1185 @item --enable-werror
1186 @itemx --disable-werror
1187 @itemx --enable-werror=yes
1188 @itemx --enable-werror=no
1189 When you specify this option, it controls whether certain files in the
1190 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1191 If you don't specify it, @option{-Werror} is turned on for the main
1192 development trunk. However it defaults to off for release branches and
1193 final releases. The specific files which get @option{-Werror} are
1194 controlled by the Makefiles.
1196 @item --enable-checking
1197 @itemx --enable-checking=@var{list}
1198 When you specify this option, the compiler is built to perform internal
1199 consistency checks of the requested complexity. This does not change the
1200 generated code, but adds error checking within the compiler. This will
1201 slow down the compiler and may only work properly if you are building
1202 the compiler with GCC@. This is @samp{yes} by default when building
1203 from SVN or snapshots, but @samp{release} for releases. More control
1204 over the checks may be had by specifying @var{list}. The categories of
1205 checks available are @samp{yes} (most common checks
1206 @samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at
1207 all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest
1208 checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1209 Individual checks can be enabled with these flags @samp{assert},
1210 @samp{fold}, @samp{gc}, @samp{gcac} @samp{misc}, @samp{rtl},
1211 @samp{rtlflag}, @samp{runtime}, @samp{tree}, and @samp{valgrind}.
1213 The @samp{valgrind} check requires the external @command{valgrind}
1214 simulator, available from @uref{http://valgrind.org/}. The
1215 @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive.
1216 To disable all checking, @samp{--disable-checking} or
1217 @samp{--enable-checking=none} must be explicitly requested. Disabling
1218 assertions will make the compiler and runtime slightly faster but
1219 increase the risk of undetected internal errors causing wrong code to be
1222 @item --enable-coverage
1223 @itemx --enable-coverage=@var{level}
1224 With this option, the compiler is built to collect self coverage
1225 information, every time it is run. This is for internal development
1226 purposes, and only works when the compiler is being built with gcc. The
1227 @var{level} argument controls whether the compiler is built optimized or
1228 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1229 want to disable optimization, for performance analysis you want to
1230 enable optimization. When coverage is enabled, the default level is
1231 without optimization.
1233 @item --enable-gather-detailed-mem-stats
1234 When this option is specified more detailed information on memory
1235 allocation is gathered. This information is printed when using
1236 @option{-fmem-report}.
1239 @itemx --with-gc=@var{choice}
1240 With this option you can specify the garbage collector implementation
1241 used during the compilation process. @var{choice} can be one of
1242 @samp{page} and @samp{zone}, where @samp{page} is the default.
1245 @itemx --disable-nls
1246 The @option{--enable-nls} option enables Native Language Support (NLS),
1247 which lets GCC output diagnostics in languages other than American
1248 English. Native Language Support is enabled by default if not doing a
1249 canadian cross build. The @option{--disable-nls} option disables NLS@.
1251 @item --with-included-gettext
1252 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1253 procedure to prefer its copy of GNU @command{gettext}.
1255 @item --with-catgets
1256 If NLS is enabled, and if the host lacks @code{gettext} but has the
1257 inferior @code{catgets} interface, the GCC build procedure normally
1258 ignores @code{catgets} and instead uses GCC's copy of the GNU
1259 @code{gettext} library. The @option{--with-catgets} option causes the
1260 build procedure to use the host's @code{catgets} in this situation.
1262 @item --with-libiconv-prefix=@var{dir}
1263 Search for libiconv header files in @file{@var{dir}/include} and
1264 libiconv library files in @file{@var{dir}/lib}.
1266 @item --enable-obsolete
1267 Enable configuration for an obsoleted system. If you attempt to
1268 configure GCC for a system (build, host, or target) which has been
1269 obsoleted, and you do not specify this flag, configure will halt with an
1272 All support for systems which have been obsoleted in one release of GCC
1273 is removed entirely in the next major release, unless someone steps
1274 forward to maintain the port.
1276 @item --enable-decimal-float
1277 @itemx --disable-decimal-float
1278 Enable (or disable) support for the C decimal floating point
1279 extension. This is enabled by default only on PowerPC GNU/Linux
1280 systems. Other systems may also support it, but require the user to
1281 specifically enable it.
1283 @item --with-long-double-128
1284 Specify if @code{long double} type should be 128-bit by default on selected
1285 GNU/Linux architectures. If using @code{--without-long-double-128},
1286 @code{long double} will be by default 64-bit, the same as @code{double} type.
1287 When neither of these configure options are used, the default will be
1288 128-bit @code{long double} when built against GNU C Library 2.4 and later,
1289 64-bit @code{long double} otherwise.
1293 @subheading Cross-Compiler-Specific Options
1294 The following options only apply to building cross compilers.
1296 @item --with-sysroot
1297 @itemx --with-sysroot=@var{dir}
1298 Tells GCC to consider @var{dir} as the root of a tree that contains a
1299 (subset of) the root filesystem of the target operating system.
1300 Target system headers, libraries and run-time object files will be
1301 searched in there. The specified directory is not copied into the
1302 install tree, unlike the options @option{--with-headers} and
1303 @option{--with-libs} that this option obsoletes. The default value,
1304 in case @option{--with-sysroot} is not given an argument, is
1305 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1306 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1307 the GCC binaries if the installation tree is moved.
1309 @item --with-build-sysroot
1310 @itemx --with-build-sysroot=@var{dir}
1311 Tells GCC to consider @var{dir} as the system root (see
1312 @option{--with-sysroot}) while building target libraries, instead of
1313 the directory specified with @option{--with-sysroot}. This option is
1314 only useful when you are already using @option{--with-sysroot}. You
1315 can use @option{--with-build-sysroot} when you are configuring with
1316 @option{--prefix} set to a directory that is different from the one in
1317 which you are installing GCC and your target libraries.
1319 This option affects the system root for the compiler used to build
1320 target libraries (which runs on the build system); it does not affect
1321 the compiler which is used to build GCC itself.
1323 @item --with-headers
1324 @itemx --with-headers=@var{dir}
1325 Deprecated in favor of @option{--with-sysroot}.
1326 Specifies that target headers are available when building a cross compiler.
1327 The @var{dir} argument specifies a directory which has the target include
1328 files. These include files will be copied into the @file{gcc} install
1329 directory. @emph{This option with the @var{dir} argument is required} when
1330 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1331 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1332 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1333 will be run on these files to make them compatible with GCC@.
1335 @item --without-headers
1336 Tells GCC not use any target headers from a libc when building a cross
1337 compiler. When crossing to GNU/Linux, you need the headers so GCC
1338 can build the exception handling for libgcc.
1341 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1342 Deprecated in favor of @option{--with-sysroot}.
1343 Specifies a list of directories which contain the target runtime
1344 libraries. These libraries will be copied into the @file{gcc} install
1345 directory. If the directory list is omitted, this option has no
1349 Specifies that @samp{newlib} is
1350 being used as the target C library. This causes @code{__eprintf} to be
1351 omitted from @file{libgcc.a} on the assumption that it will be provided by
1354 @item --with-build-time-tools=@var{dir}
1355 Specifies where to find the set of target tools (assembler, linker, etc.)
1356 that will be used while building GCC itself. This option can be useful
1357 if the directory layouts are different between the system you are building
1358 GCC on, and the system where you will deploy it.
1360 For example, on a @option{ia64-hp-hpux} system, you may have the GNU
1361 assembler and linker in @file{/usr/bin}, and the native tools in a
1362 different path, and build a toolchain that expects to find the
1363 native tools in @file{/usr/bin}.
1365 When you use this option, you should ensure that @var{dir} includes
1366 @command{ar}, @command{as}, @command{ld}, @command{nm},
1367 @command{ranlib} and @command{strip} if necessary, and possibly
1368 @command{objdump}. Otherwise, GCC may use an inconsistent set of
1372 @subheading Fortran-Specific Options
1374 The following options apply to the build of the Fortran front end.
1378 @item --with-gmp=@var{pathname}
1379 @itemx --with-mpfr=@var{pathname}
1380 @itemx --with-gmp-dir=@var{pathname}
1381 @itemx --with-mpfr-dir=@var{pathname}
1382 If you don't have GMP (the GNU Multiple Precision library) and the MPFR
1383 Libraries installed in a standard location and you want to build the Fortran
1384 front-end, you can explicitly specify the directory where they are installed
1385 (@samp{--with-gmp=gmpinstalldir}, @samp{--with-mpfr=mpfrinstalldir}) or where
1386 you built them without installing (@samp{--with-gmp-dir=gmpbuilddir},
1387 @samp{--with-mpfr-dir=gmpbuilddir}).
1391 @subheading Java-Specific Options
1393 The following option applies to the build of the Java front end.
1396 @item --disable-libgcj
1397 Specify that the run-time libraries
1398 used by GCJ should not be built. This is useful in case you intend
1399 to use GCJ with some other run-time, or you're going to install it
1400 separately, or it just happens not to build on your particular
1401 machine. In general, if the Java front end is enabled, the GCJ
1402 libraries will be enabled too, unless they're known to not work on
1403 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1404 may need to port it; in this case, before modifying the top-level
1405 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1406 you may use @option{--enable-libgcj} to override the default.
1410 The following options apply to building @samp{libgcj}.
1412 @subsubheading General Options
1415 @item --disable-getenv-properties
1416 Don't set system properties from @env{GCJ_PROPERTIES}.
1418 @item --enable-hash-synchronization
1419 Use a global hash table for monitor locks. Ordinarily,
1420 @samp{libgcj}'s @samp{configure} script automatically makes
1421 the correct choice for this option for your platform. Only use
1422 this if you know you need the library to be configured differently.
1424 @item --enable-interpreter
1425 Enable the Java interpreter. The interpreter is automatically
1426 enabled by default on all platforms that support it. This option
1427 is really only useful if you want to disable the interpreter
1428 (using @option{--disable-interpreter}).
1430 @item --disable-java-net
1431 Disable java.net. This disables the native part of java.net only,
1432 using non-functional stubs for native method implementations.
1434 @item --disable-jvmpi
1435 Disable JVMPI support.
1438 Enable runtime eCos target support.
1440 @item --without-libffi
1441 Don't use @samp{libffi}. This will disable the interpreter and JNI
1442 support as well, as these require @samp{libffi} to work.
1444 @item --enable-libgcj-debug
1445 Enable runtime debugging code.
1447 @item --enable-libgcj-multifile
1448 If specified, causes all @file{.java} source files to be
1449 compiled into @file{.class} files in one invocation of
1450 @samp{gcj}. This can speed up build time, but is more
1451 resource-intensive. If this option is unspecified or
1452 disabled, @samp{gcj} is invoked once for each @file{.java}
1453 file to compile into a @file{.class} file.
1455 @item --with-libiconv-prefix=DIR
1456 Search for libiconv in @file{DIR/include} and @file{DIR/lib}.
1458 @item --enable-sjlj-exceptions
1459 Force use of @code{builtin_setjmp} for exceptions. @samp{configure}
1460 ordinarily picks the correct value based on the platform. Only use
1461 this option if you are sure you need a different setting.
1463 @item --with-system-zlib
1464 Use installed @samp{zlib} rather than that included with GCC@.
1466 @item --with-win32-nlsapi=ansi, unicows or unicode
1467 Indicates how MinGW @samp{libgcj} translates between UNICODE
1468 characters and the Win32 API@.
1471 Use the single-byte @code{char} and the Win32 A functions natively,
1472 translating to and from UNICODE when using these functions. If
1473 unspecified, this is the default.
1476 Use the @code{WCHAR} and Win32 W functions natively. Adds
1477 @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}.
1478 @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines
1479 running built executables. @file{libunicows.a}, an open-source
1480 import library around Microsoft's @code{unicows.dll}, is obtained from
1481 @uref{http://libunicows.sourceforge.net/}, which also gives details
1482 on getting @file{unicows.dll} from Microsoft.
1485 Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not}
1486 add @code{-lunicows} to @file{libgcj.spec}. The built executables will
1487 only run on Microsoft Windows NT and above.
1491 @subsubheading AWT-Specific Options
1495 Use the X Window System.
1497 @item --enable-java-awt=PEER(S)
1498 Specifies the AWT peer library or libraries to build alongside
1499 @samp{libgcj}. If this option is unspecified or disabled, AWT
1500 will be non-functional. Current valid values are @option{gtk} and
1501 @option{xlib}. Multiple libraries should be separated by a
1502 comma (i.e.@: @option{--enable-java-awt=gtk,xlib}).
1504 @item --enable-gtk-cairo
1505 Build the cairo Graphics2D implementation on GTK@.
1507 @item --enable-java-gc=TYPE
1508 Choose garbage collector. Defaults to @option{boehm} if unspecified.
1510 @item --disable-gtktest
1511 Do not try to compile and run a test GTK+ program.
1513 @item --disable-glibtest
1514 Do not try to compile and run a test GLIB program.
1516 @item --with-libart-prefix=PFX
1517 Prefix where libart is installed (optional).
1519 @item --with-libart-exec-prefix=PFX
1520 Exec prefix where libart is installed (optional).
1522 @item --disable-libarttest
1523 Do not try to compile and run a test libart program.
1532 @uref{./index.html,,Return to the GCC Installation page}
1536 @c ***Building****************************************************************
1538 @comment node-name, next, previous, up
1539 @node Building, Testing, Configuration, Installing GCC
1545 @cindex Installing GCC: Building
1547 Now that GCC is configured, you are ready to build the compiler and
1550 Some commands executed when making the compiler may fail (return a
1551 nonzero status) and be ignored by @command{make}. These failures, which
1552 are often due to files that were not found, are expected, and can safely
1555 It is normal to have compiler warnings when compiling certain files.
1556 Unless you are a GCC developer, you can generally ignore these warnings
1557 unless they cause compilation to fail. Developers should attempt to fix
1558 any warnings encountered, however they can temporarily continue past
1559 warnings-as-errors by specifying the configure flag
1560 @option{--disable-werror}.
1562 On certain old systems, defining certain environment variables such as
1563 @env{CC} can interfere with the functioning of @command{make}.
1565 If you encounter seemingly strange errors when trying to build the
1566 compiler in a directory other than the source directory, it could be
1567 because you have previously configured the compiler in the source
1568 directory. Make sure you have done all the necessary preparations.
1570 If you build GCC on a BSD system using a directory stored in an old System
1571 V file system, problems may occur in running @command{fixincludes} if the
1572 System V file system doesn't support symbolic links. These problems
1573 result in a failure to fix the declaration of @code{size_t} in
1574 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1575 that type mismatches occur, this could be the cause.
1577 The solution is not to use such a directory for building GCC@.
1579 When building from SVN or snapshots, or if you modify parser sources,
1580 you need the Bison parser generator installed. If you do not modify
1581 parser sources, releases contain the Bison-generated files and you do
1582 not need Bison installed to build them.
1584 When building from SVN or snapshots, or if you modify Texinfo
1585 documentation, you need version 4.4 or later of Texinfo installed if you
1586 want Info documentation to be regenerated. Releases contain Info
1587 documentation pre-built for the unmodified documentation in the release.
1589 @section Building a native compiler
1591 For a native build, the default configuration is to perform
1592 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
1593 This will build the entire GCC system and ensure that it compiles
1594 itself correctly. It can be disabled with the @option{--disable-bootstrap}
1595 parameter to @samp{configure}, but bootstrapping is suggested because
1596 the compiler will be tested more completely and could also have
1599 The bootstrapping process will complete the following steps:
1603 Build tools necessary to build the compiler.
1606 Perform a 3-stage bootstrap of the compiler. This includes building
1607 three times the target tools for use by the compiler such as binutils
1608 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
1609 individually linked or moved into the top level GCC source tree before
1613 Perform a comparison test of the stage2 and stage3 compilers.
1616 Build runtime libraries using the stage3 compiler from the previous step.
1620 If you are short on disk space you might consider @samp{make
1621 bootstrap-lean} instead. The sequence of compilation is the
1622 same described above, but object files from the stage1 and
1623 stage2 of the 3-stage bootstrap of the compiler are deleted as
1624 soon as they are no longer needed.
1626 If you want to save additional space during the bootstrap and in
1627 the final installation as well, you can build the compiler binaries
1628 without debugging information as in the following example. This will save
1629 roughly 40% of disk space both for the bootstrap and the final installation.
1630 (Libraries will still contain debugging information.)
1633 make CFLAGS='-O' LIBCFLAGS='-g -O2' \
1634 LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap
1637 If you wish to use non-default GCC flags when compiling the stage2 and
1638 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
1639 @samp{make}. Non-default optimization flags are less well
1640 tested here than the default of @samp{-g -O2}, but should still work.
1641 In a few cases, you may find that you need to specify special flags such
1642 as @option{-msoft-float} here to complete the bootstrap; or, if the
1643 native compiler miscompiles the stage1 compiler, you may need to work
1644 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
1645 stage1 compiler that were miscompiled, or by using @samp{make
1646 bootstrap4} to increase the number of stages of bootstrap.
1648 Note that using non-standard @code{CFLAGS} can cause bootstrap to fail
1649 if these trigger a warning with the new compiler. For example using
1650 @samp{-O2 -g -mcpu=i686} on @code{i686-pc-linux-gnu} will cause bootstrap
1651 failure as @option{-mcpu=} is deprecated in 3.4.0 and above.
1654 If you used the flag @option{--enable-languages=@dots{}} to restrict
1655 the compilers to be built, only those you've actually enabled will be
1656 built. This will of course only build those runtime libraries, for
1657 which the particular compiler has been built. Please note,
1658 that re-defining @env{LANGUAGES} when calling @samp{make}
1659 @strong{does not} work anymore!
1661 If the comparison of stage2 and stage3 fails, this normally indicates
1662 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1663 a potentially serious bug which you should investigate and report. (On
1664 a few systems, meaningful comparison of object files is impossible; they
1665 always appear ``different''. If you encounter this problem, you will
1666 need to disable comparison in the @file{Makefile}.)
1668 If you do not want to bootstrap your compiler, you can configure with
1669 @option{--disable-bootstrap}. In particular cases, you may want to
1670 bootstrap your compiler even if the target system is not the same as
1671 the one you are building on: for example, you could build a
1672 @code{powerpc-unknown-linux-gnu} toolchain on a
1673 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
1674 @option{--enable-bootstrap} to the configure script.
1677 @section Building a cross compiler
1679 When building a cross compiler, it is not generally possible to do a
1680 3-stage bootstrap of the compiler. This makes for an interesting problem
1681 as parts of GCC can only be built with GCC@.
1683 To build a cross compiler, we first recommend building and installing a
1684 native compiler. You can then use the native GCC compiler to build the
1685 cross compiler. The installed native compiler needs to be GCC version
1688 Assuming you have already installed a native copy of GCC and configured
1689 your cross compiler, issue the command @command{make}, which performs the
1694 Build host tools necessary to build the compiler.
1697 Build target tools for use by the compiler such as binutils (bfd,
1698 binutils, gas, gprof, ld, and opcodes)
1699 if they have been individually linked or moved into the top level GCC source
1700 tree before configuring.
1703 Build the compiler (single stage only).
1706 Build runtime libraries using the compiler from the previous step.
1709 Note that if an error occurs in any step the make process will exit.
1711 If you are not building GNU binutils in the same source tree as GCC,
1712 you will need a cross-assembler and cross-linker installed before
1713 configuring GCC@. Put them in the directory
1714 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
1715 you should put in this directory:
1719 This should be the cross-assembler.
1722 This should be the cross-linker.
1725 This should be the cross-archiver: a program which can manipulate
1726 archive files (linker libraries) in the target machine's format.
1729 This should be a program to construct a symbol table in an archive file.
1732 The installation of GCC will find these programs in that directory,
1733 and copy or link them to the proper place to for the cross-compiler to
1734 find them when run later.
1736 The easiest way to provide these files is to build the Binutils package.
1737 Configure it with the same @option{--host} and @option{--target}
1738 options that you use for configuring GCC, then build and install
1739 them. They install their executables automatically into the proper
1740 directory. Alas, they do not support all the targets that GCC
1743 If you are not building a C library in the same source tree as GCC,
1744 you should also provide the target libraries and headers before
1745 configuring GCC, specifying the directories with
1746 @option{--with-sysroot} or @option{--with-headers} and
1747 @option{--with-libs}. Many targets also require ``start files'' such
1748 as @file{crt0.o} and
1749 @file{crtn.o} which are linked into each executable. There may be several
1750 alternatives for @file{crt0.o}, for use with profiling or other
1751 compilation options. Check your target's definition of
1752 @code{STARTFILE_SPEC} to find out what start files it uses.
1754 @section Building in parallel
1756 GNU Make 3.79 and above, which is necessary to build GCC, support
1757 building in parallel. To activate this, you can use @samp{make -j 2}
1758 instead of @samp{make}. You can also specify a bigger number, and
1759 in most cases using a value greater than the number of processors in
1760 your machine will result in fewer and shorter I/O latency hits, thus
1761 improving overall throughput; this is especially true for slow drives
1762 and network filesystems.
1764 @section Building the Ada compiler
1766 In order to build GNAT, the Ada compiler, you need a working GNAT
1767 compiler (GNAT version 3.14 or later, or GCC version 3.1 or later).
1768 This includes GNAT tools such as @command{gnatmake} and
1769 @command{gnatlink}, since the Ada front end is written in Ada and
1770 uses some GNAT-specific extensions.
1772 In order to build a cross compiler, it is suggested to install
1773 the new compiler as native first, and then use it to build the cross
1776 @command{configure} does not test whether the GNAT installation works
1777 and has a sufficiently recent version; if too old a GNAT version is
1778 installed, the build will fail unless @option{--enable-languages} is
1779 used to disable building the Ada front end.
1781 @section Building with profile feedback
1783 It is possible to use profile feedback to optimize the compiler itself. This
1784 should result in a faster compiler binary. Experiments done on x86 using gcc
1785 3.3 showed approximately 7 percent speedup on compiling C programs. To
1786 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
1788 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
1789 compiler. This compiler is used to build a @code{stageprofile} compiler
1790 instrumented to collect execution counts of instruction and branch
1791 probabilities. Then runtime libraries are compiled with profile collected.
1792 Finally a @code{stagefeedback} compiler is built using the information collected.
1794 Unlike standard bootstrap, several additional restrictions apply. The
1795 compiler used to build @code{stage1} needs to support a 64-bit integral type.
1796 It is recommended to only use GCC for this. Also parallel make is currently
1797 not supported since collisions in profile collecting may occur.
1804 @uref{./index.html,,Return to the GCC Installation page}
1808 @c ***Testing*****************************************************************
1810 @comment node-name, next, previous, up
1811 @node Testing, Final install, Building, Installing GCC
1815 @chapter Installing GCC: Testing
1818 @cindex Installing GCC: Testing
1821 Before you install GCC, we encourage you to run the testsuites and to
1822 compare your results with results from a similar configuration that have
1823 been submitted to the
1824 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
1825 Some of these archived results are linked from the build status lists
1826 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
1827 reports a successful build runs the testsuites and submits the results.
1828 This step is optional and may require you to download additional software,
1829 but it can give you confidence in your new GCC installation or point out
1830 problems before you install and start using your new GCC@.
1832 First, you must have @uref{download.html,,downloaded the testsuites}.
1833 These are part of the full distribution, but if you downloaded the
1834 ``core'' compiler plus any front ends, you must download the testsuites
1837 Second, you must have the testing tools installed. This includes
1838 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
1839 the DejaGnu site has links to these.
1841 If the directories where @command{runtest} and @command{expect} were
1842 installed are not in the @env{PATH}, you may need to set the following
1843 environment variables appropriately, as in the following example (which
1844 assumes that DejaGnu has been installed under @file{/usr/local}):
1847 TCL_LIBRARY = /usr/local/share/tcl8.0
1848 DEJAGNULIBS = /usr/local/share/dejagnu
1851 (On systems such as Cygwin, these paths are required to be actual
1852 paths, not mounts or links; presumably this is due to some lack of
1853 portability in the DejaGnu code.)
1856 Finally, you can run the testsuite (which may take a long time):
1858 cd @var{objdir}; make -k check
1861 This will test various components of GCC, such as compiler
1862 front ends and runtime libraries. While running the testsuite, DejaGnu
1863 might emit some harmless messages resembling
1864 @samp{WARNING: Couldn't find the global config file.} or
1865 @samp{WARNING: Couldn't find tool init file} that can be ignored.
1867 @section How can you run the testsuite on selected tests?
1869 In order to run sets of tests selectively, there are targets
1870 @samp{make check-gcc} and @samp{make check-g++}
1871 in the @file{gcc} subdirectory of the object directory. You can also
1872 just run @samp{make check} in a subdirectory of the object directory.
1875 A more selective way to just run all @command{gcc} execute tests in the
1879 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
1882 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
1883 the testsuite with filenames matching @samp{9805*}, you would use
1886 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1889 The @file{*.exp} files are located in the testsuite directories of the GCC
1890 source, the most important ones being @file{compile.exp},
1891 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1892 To get a list of the possible @file{*.exp} files, pipe the
1893 output of @samp{make check} into a file and look at the
1894 @samp{Running @dots{} .exp} lines.
1896 @section Passing options and running multiple testsuites
1898 You can pass multiple options to the testsuite using the
1899 @samp{--target_board} option of DejaGNU, either passed as part of
1900 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
1901 work outside the makefiles. For example,
1904 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fno-strength-reduce"
1907 will run the standard @command{g++} testsuites (``unix'' is the target name
1908 for a standard native testsuite situation), passing
1909 @samp{-O3 -fno-strength-reduce} to the compiler on every test, i.e.,
1910 slashes separate options.
1912 You can run the testsuites multiple times using combinations of options
1913 with a syntax similar to the brace expansion of popular shells:
1916 @dots{}"--target_board=arm-sim/@{-mhard-float,-msoft-float@}@{-O1,-O2,-O3,@}"
1919 (Note the empty option caused by the trailing comma in the final group.)
1920 The following will run each testsuite eight times using the @samp{arm-sim}
1921 target, as if you had specified all possible combinations yourself:
1924 --target_board=arm-sim/-mhard-float/-O1
1925 --target_board=arm-sim/-mhard-float/-O2
1926 --target_board=arm-sim/-mhard-float/-O3
1927 --target_board=arm-sim/-mhard-float
1928 --target_board=arm-sim/-msoft-float/-O1
1929 --target_board=arm-sim/-msoft-float/-O2
1930 --target_board=arm-sim/-msoft-float/-O3
1931 --target_board=arm-sim/-msoft-float
1934 They can be combined as many times as you wish, in arbitrary ways. This
1938 @dots{}"--target_board=unix/-Wextra@{-O3,-fno-strength-reduce@}@{-fomit-frame-pointer,@}"
1941 will generate four combinations, all involving @samp{-Wextra}.
1943 The disadvantage to this method is that the testsuites are run in serial,
1944 which is a waste on multiprocessor systems. For users with GNU Make and
1945 a shell which performs brace expansion, you can run the testsuites in
1946 parallel by having the shell perform the combinations and @command{make}
1947 do the parallel runs. Instead of using @samp{--target_board}, use a
1948 special makefile target:
1951 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
1957 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
1960 will run three concurrent ``make-gcc'' testsuites, eventually testing all
1961 ten combinations as described above. Note that this is currently only
1962 supported in the @file{gcc} subdirectory. (To see how this works, try
1963 typing @command{echo} before the example given here.)
1966 @section Additional testing for Java Class Libraries
1968 The Java runtime tests can be executed via @samp{make check}
1969 in the @file{@var{target}/libjava/testsuite} directory in
1972 The @uref{http://sources.redhat.com/mauve/,,Mauve Project} provides
1973 a suite of tests for the Java Class Libraries. This suite can be run
1974 as part of libgcj testing by placing the Mauve tree within the libjava
1975 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
1976 specifying the location of that tree when invoking @samp{make}, as in
1977 @samp{make MAUVEDIR=~/mauve check}.
1979 @uref{http://sources.redhat.com/mauve/jacks.html,,Jacks}
1980 is a free testsuite that tests Java compiler front ends. This suite
1981 can be run as part of libgcj testing by placing the Jacks tree within
1982 the libjava testsuite at @file{libjava/testsuite/libjava.jacks/jacks}.
1984 @section How to interpret test results
1986 The result of running the testsuite are various @file{*.sum} and @file{*.log}
1987 files in the testsuite subdirectories. The @file{*.log} files contain a
1988 detailed log of the compiler invocations and the corresponding
1989 results, the @file{*.sum} files summarize the results. These summaries
1990 contain status codes for all tests:
1994 PASS: the test passed as expected
1996 XPASS: the test unexpectedly passed
1998 FAIL: the test unexpectedly failed
2000 XFAIL: the test failed as expected
2002 UNSUPPORTED: the test is not supported on this platform
2004 ERROR: the testsuite detected an error
2006 WARNING: the testsuite detected a possible problem
2009 It is normal for some tests to report unexpected failures. At the
2010 current time the testing harness does not allow fine grained control
2011 over whether or not a test is expected to fail. This problem should
2012 be fixed in future releases.
2015 @section Submitting test results
2017 If you want to report the results to the GCC project, use the
2018 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
2021 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
2022 -m gcc-testresults@@gcc.gnu.org |sh
2025 This script uses the @command{Mail} program to send the results, so
2026 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
2027 prepended to the testsuite summary and should contain any special
2028 remarks you have on your results or your build environment. Please
2029 do not edit the testsuite result block or the subject line, as these
2030 messages may be automatically processed.
2037 @uref{./index.html,,Return to the GCC Installation page}
2041 @c ***Final install***********************************************************
2043 @comment node-name, next, previous, up
2044 @node Final install, , Testing, Installing GCC
2046 @ifset finalinstallhtml
2048 @chapter Installing GCC: Final installation
2051 Now that GCC has been built (and optionally tested), you can install it with
2053 cd @var{objdir}; make install
2056 We strongly recommend to install into a target directory where there is
2057 no previous version of GCC present.
2059 That step completes the installation of GCC; user level binaries can
2060 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
2061 you specified with the @option{--prefix} to configure (or
2062 @file{/usr/local} by default). (If you specified @option{--bindir},
2063 that directory will be used instead; otherwise, if you specified
2064 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
2065 Headers for the C++ and Java libraries are installed in
2066 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
2067 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
2068 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
2069 in info format in @file{@var{infodir}} (normally
2070 @file{@var{prefix}/info}).
2072 When installing cross-compilers, GCC's executables
2073 are not only installed into @file{@var{bindir}}, that
2074 is, @file{@var{exec-prefix}/bin}, but additionally into
2075 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
2076 exists. Typically, such @dfn{tooldirs} hold target-specific
2077 binutils, including assembler and linker.
2079 Installation into a temporary staging area or into a @command{chroot}
2080 jail can be achieved with the command
2083 make DESTDIR=@var{path-to-rootdir} install
2086 @noindent where @var{path-to-rootdir} is the absolute path of
2087 a directory relative to which all installation paths will be
2088 interpreted. Note that the directory specified by @code{DESTDIR}
2089 need not exist yet; it will be created if necessary.
2091 There is a subtle point with tooldirs and @code{DESTDIR}:
2092 If you relocate a cross-compiler installation with
2093 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
2094 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
2095 be filled with duplicated GCC executables only if it already exists,
2096 it will not be created otherwise. This is regarded as a feature,
2097 not as a bug, because it gives slightly more control to the packagers
2098 using the @code{DESTDIR} feature.
2100 If you are bootstrapping a released version of GCC then please
2101 quickly review the build status page for your release, available from
2102 @uref{http://gcc.gnu.org/buildstat.html}.
2103 If your system is not listed for the version of GCC that you built,
2105 @email{gcc@@gcc.gnu.org} indicating
2106 that you successfully built and installed GCC@.
2107 Include the following information:
2111 Output from running @file{@var{srcdir}/config.guess}. Do not send
2112 that file itself, just the one-line output from running it.
2115 The output of @samp{gcc -v} for your newly installed @command{gcc}.
2116 This tells us which version of GCC you built and the options you passed to
2120 Whether you enabled all languages or a subset of them. If you used a
2121 full distribution then this information is part of the configure
2122 options in the output of @samp{gcc -v}, but if you downloaded the
2123 ``core'' compiler plus additional front ends then it isn't apparent
2124 which ones you built unless you tell us about it.
2127 If the build was for GNU/Linux, also include:
2130 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
2131 this information should be available from @file{/etc/issue}.
2134 The version of the Linux kernel, available from @samp{uname --version}
2138 The version of glibc you used; for RPM-based systems like Red Hat,
2139 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
2140 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
2142 For other systems, you can include similar information if you think it is
2146 Any other information that you think would be useful to people building
2147 GCC on the same configuration. The new entry in the build status list
2148 will include a link to the archived copy of your message.
2151 We'd also like to know if the
2153 @ref{Specific, host/target specific installation notes}
2156 @uref{specific.html,,host/target specific installation notes}
2158 didn't include your host/target information or if that information is
2159 incomplete or out of date. Send a note to
2160 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
2162 If you find a bug, please report it following the
2163 @uref{../bugs.html,,bug reporting guidelines}.
2165 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
2166 dvi}. You will need to have @command{texi2dvi} (version at least 4.4)
2167 and @TeX{} installed. This creates a number of @file{.dvi} files in
2168 subdirectories of @file{@var{objdir}}; these may be converted for
2169 printing with programs such as @command{dvips}. You can also
2170 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
2171 Free Software Foundation}, though such manuals may not be for the most
2172 recent version of GCC@.
2174 If you would like to generate online HTML documentation, do @samp{cd
2175 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
2176 @file{@var{objdir}/gcc/HTML}.
2183 @uref{./index.html,,Return to the GCC Installation page}
2187 @c ***Binaries****************************************************************
2189 @comment node-name, next, previous, up
2190 @node Binaries, Specific, Installing GCC, Top
2194 @chapter Installing GCC: Binaries
2197 @cindex Installing GCC: Binaries
2199 We are often asked about pre-compiled versions of GCC@. While we cannot
2200 provide these for all platforms, below you'll find links to binaries for
2201 various platforms where creating them by yourself is not easy due to various
2204 Please note that we did not create these binaries, nor do we
2205 support them. If you have any problems installing them, please
2206 contact their makers.
2213 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
2216 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX}.
2220 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
2223 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
2224 Development Tools for the Renesas H8/300[HS] Series}.
2230 @uref{http://hpux.cs.utah.edu/,,HP-UX Porting Center};
2233 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
2237 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
2238 Development Tools for the Motorola 68HC11/68HC12}.
2241 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
2242 OpenServer/Unixware}.
2245 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
2248 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
2254 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
2256 The @uref{http://www.mingw.org/,,MinGW} project.
2260 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
2261 Written Word} offers binaries for
2264 Digital UNIX 4.0D and 5.1,
2266 HP-UX 10.20, 11.00, and 11.11, and
2267 Solaris/SPARC 2.5.1, 2.6, 7, 8, and 9.
2270 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
2271 number of platforms.
2274 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
2275 links to gfortran binaries for several platforms.
2278 In addition to those specific offerings, you can get a binary
2279 distribution CD-ROM from the
2280 @uref{http://www.gnu.org/order/order.html,,Free Software Foundation}.
2281 It contains binaries for a number of platforms, and
2282 includes not only GCC, but other stuff as well. The current CD does
2283 not contain the latest version of GCC, but it should allow
2284 bootstrapping the compiler. An updated version of that disk is in the
2292 @uref{./index.html,,Return to the GCC Installation page}
2296 @c ***Specific****************************************************************
2298 @comment node-name, next, previous, up
2299 @node Specific, Old, Binaries, Top
2303 @chapter Host/target specific installation notes for GCC
2306 @cindex Specific installation notes
2307 @cindex Target specific installation
2308 @cindex Host specific installation
2309 @cindex Target specific installation notes
2311 Please read this document carefully @emph{before} installing the
2312 GNU Compiler Collection on your machine.
2314 Note that this list of install notes is @emph{not} a list of supported
2315 hosts or targets. Not all supported hosts and targets are listed
2316 here, only the ones that require host-specific or target-specific
2322 @uref{#alpha-x-x,,alpha*-*-*}
2324 @uref{#alpha-dec-osf,,alpha*-dec-osf*}
2326 @uref{#alphaev5-cray-unicosmk,,alphaev5-cray-unicosmk*}
2328 @uref{#arc-x-elf,,arc-*-elf}
2330 @uref{#arm-x-elf,,arm-*-elf}
2331 @uref{#arm-x-coff,,arm-*-coff}
2332 @uref{#arm-x-aout,,arm-*-aout}
2334 @uref{#xscale-x-x,,xscale-*-*}
2338 @uref{#bfin,,Blackfin}
2344 @uref{#x-x-freebsd,,*-*-freebsd*}
2346 @uref{#h8300-hms,,h8300-hms}
2348 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
2350 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
2352 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
2354 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
2356 @uref{#ix86-x-linuxaout,,i?86-*-linux*aout}
2358 @uref{#ix86-x-linux,,i?86-*-linux*}
2360 @uref{#ix86-x-sco32v5,,i?86-*-sco3.2v5*}
2362 @uref{#ix86-x-solaris210,,i?86-*-solaris2.10}
2364 @uref{#ix86-x-udk,,i?86-*-udk}
2366 @uref{#ia64-x-linux,,ia64-*-linux}
2368 @uref{#ia64-x-hpux,,ia64-*-hpux*}
2370 @uref{#x-ibm-aix,,*-ibm-aix*}
2372 @uref{#iq2000-x-elf,,iq2000-*-elf}
2374 @uref{#m32c-x-elf,,m32c-*-elf}
2376 @uref{#m32r-x-elf,,m32r-*-elf}
2378 @uref{#m6811-elf,,m6811-elf}
2380 @uref{#m6812-elf,,m6812-elf}
2382 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
2384 @uref{#mips-x-x,,mips-*-*}
2386 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2388 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2390 @uref{#powerpc-x-x,,powerpc*-*-*, powerpc-*-sysv4}
2392 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
2394 @uref{#powerpc-x-elf,,powerpc-*-elf, powerpc-*-sysv4}
2396 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
2398 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
2400 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
2402 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
2404 @uref{#powerpcle-x-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
2406 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
2408 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
2410 @uref{#s390-x-linux,,s390-*-linux*}
2412 @uref{#s390x-x-linux,,s390x-*-linux*}
2414 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
2416 @uref{#x-x-solaris2,,*-*-solaris2*}
2418 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
2420 @uref{#sparc-sun-solaris27,,sparc-sun-solaris2.7}
2422 @uref{#sparc-x-linux,,sparc-*-linux*}
2424 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
2426 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
2428 @uref{#x-x-sysv,,*-*-sysv*}
2430 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
2432 @uref{#x-x-vxworks,,*-*-vxworks*}
2434 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
2436 @uref{#xtensa-x-elf,,xtensa-*-elf}
2438 @uref{#xtensa-x-linux,,xtensa-*-linux*}
2440 @uref{#windows,,Microsoft Windows}
2444 @uref{#older,,Older systems}
2449 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
2455 <!-- -------- host/target specific issues start here ---------------- -->
2458 @heading @anchor{alpha-x-x}alpha*-*-*
2460 This section contains general configuration information for all
2461 alpha-based platforms using ELF (in particular, ignore this section for
2462 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2463 section, please read all other sections that match your target.
2465 We require binutils 2.11.2 or newer.
2466 Previous binutils releases had a number of problems with DWARF 2
2467 debugging information, not the least of which is incorrect linking of
2473 @heading @anchor{alpha-dec-osf}alpha*-dec-osf*
2474 Systems using processors that implement the DEC Alpha architecture and
2475 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2476 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2478 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2479 supported. (These are the versions which identify themselves as DEC
2482 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2483 may be fixed by configuring with @option{--with-gc=simple},
2484 reconfiguring Kernel Virtual Memory and Swap parameters
2485 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2486 or applying the patch in
2487 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2489 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2490 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2491 we need to use the old assembler, invoked via the barely documented
2492 @option{-oldas} option. To bootstrap GCC, you either need to use the
2496 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2499 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2502 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2505 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2506 are supported on Tru64 UNIX, so you must not configure GCC with
2507 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2509 GCC writes a @samp{.verstamp} directive to the assembler output file
2510 unless it is built as a cross-compiler. It gets the version to use from
2511 the system header file @file{/usr/include/stamp.h}. If you install a
2512 new version of DEC Unix, you should rebuild GCC to pick up the new version
2515 Note that since the Alpha is a 64-bit architecture, cross-compilers from
2516 32-bit machines will not generate code as efficient as that generated
2517 when the compiler is running on a 64-bit machine because many
2518 optimizations that depend on being able to represent a word on the
2519 target in an integral value on the host cannot be performed. Building
2520 cross-compilers on the Alpha for 32-bit machines has only been tested in
2521 a few cases and may not work properly.
2523 @samp{make compare} may fail on old versions of DEC Unix unless you add
2524 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2525 assembler input file is stored in the object file, and that makes
2526 comparison fail if it differs between the @code{stage1} and
2527 @code{stage2} compilations. The option @option{-save-temps} forces a
2528 fixed name to be used for the assembler input file, instead of a
2529 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2530 unless the comparisons fail without that option. If you add
2531 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2532 @samp{.s} files after each series of compilations.
2534 GCC now supports both the native (ECOFF) debugging format used by DBX
2535 and GDB and an encapsulated STABS format for use only with GDB@. See the
2536 discussion of the @option{--with-stabs} option of @file{configure} above
2537 for more information on these formats and how to select them.
2539 There is a bug in DEC's assembler that produces incorrect line numbers
2540 for ECOFF format when the @samp{.align} directive is used. To work
2541 around this problem, GCC will not emit such alignment directives
2542 while writing ECOFF format debugging information even if optimization is
2543 being performed. Unfortunately, this has the very undesirable
2544 side-effect that code addresses when @option{-O} is specified are
2545 different depending on whether or not @option{-g} is also specified.
2547 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2548 DBX@. DEC is now aware of this problem with the assembler and hopes to
2549 provide a fix shortly.
2554 @heading @anchor{alphaev5-cray-unicosmk}alphaev5-cray-unicosmk*
2555 Cray T3E systems running Unicos/Mk.
2557 This port is incomplete and has many known bugs. We hope to improve the
2558 support for this target soon. Currently, only the C front end is supported,
2559 and it is not possible to build parallel applications. Cray modules are not
2560 supported; in particular, Craylibs are assumed to be in
2561 @file{/opt/ctl/craylibs/craylibs}.
2563 On this platform, you need to tell GCC where to find the assembler and
2564 the linker. The simplest way to do so is by providing @option{--with-as}
2565 and @option{--with-ld} to @file{configure}, e.g.@:
2568 configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld \
2569 --enable-languages=c
2572 The comparison test at the end of the bootstrapping process fails on Unicos/Mk
2573 because the assembler inserts timestamps into object files. You should
2574 be able to work around this by doing @samp{make all} after getting this
2580 @heading @anchor{arc-x-elf}arc-*-elf
2581 Argonaut ARC processor.
2582 This configuration is intended for embedded systems.
2587 @heading @anchor{arm-x-elf}arm-*-elf
2588 @heading @anchor{xscale-x-x}xscale-*-*
2589 ARM-family processors. Subtargets that use the ELF object format
2590 require GNU binutils 2.13 or newer. Such subtargets include:
2591 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux},
2592 @code{arm-*-rtems} and @code{arm-*-kaos}.
2597 @heading @anchor{arm-x-coff}arm-*-coff
2598 ARM-family processors. Note that there are two different varieties
2599 of PE format subtarget supported: @code{arm-wince-pe} and
2600 @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}.
2605 @heading @anchor{arm-x-aout}arm-*-aout
2606 ARM-family processors. These targets support the AOUT file format:
2607 @code{arm-*-aout}, @code{arm-*-netbsd}.
2612 @heading @anchor{avr}avr
2614 ATMEL AVR-family micro controllers. These are used in embedded
2615 applications. There are no standard Unix configurations.
2617 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
2621 See ``AVR Options'' in the main manual
2623 for the list of supported MCU types.
2625 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2627 Further installation notes and other useful information about AVR tools
2628 can also be obtained from:
2632 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
2634 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2636 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2639 We @emph{strongly} recommend using binutils 2.13 or newer.
2641 The following error:
2643 Error: register required
2646 indicates that you should upgrade to a newer version of the binutils.
2651 @heading @anchor{bfin}Blackfin
2653 The Blackfin processor, an Analog Devices DSP.
2655 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
2659 See ``Blackfin Options'' in the main manual
2662 More information, and a version of binutils with support for this processor,
2663 is available at @uref{http://blackfin.uclinux.org}
2668 @heading @anchor{c4x}c4x
2670 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
2671 Processors. These are used in embedded applications. There are no
2672 standard Unix configurations.
2674 @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using the
2675 GNU Compiler Collection (GCC)},
2678 See ``TMS320C3x/C4x Options'' in the main manual
2680 for the list of supported MCU types.
2682 GCC can be configured as a cross compiler for both the C3x and C4x
2683 architectures on the same system. Use @samp{configure --target=c4x
2684 --enable-languages="c,c++"} to configure.
2687 Further installation notes and other useful information about C4x tools
2688 can also be obtained from:
2692 @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/}
2698 @heading @anchor{cris}CRIS
2700 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2701 series. These are used in embedded applications.
2704 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
2708 See ``CRIS Options'' in the main manual
2710 for a list of CRIS-specific options.
2712 There are a few different CRIS targets:
2714 @item cris-axis-aout
2715 Old target. Includes a multilib for the @samp{elinux} a.out-based
2716 target. No multilibs for newer architecture variants.
2718 Mainly for monolithic embedded systems. Includes a multilib for the
2719 @samp{v10} core used in @samp{ETRAX 100 LX}.
2720 @item cris-axis-linux-gnu
2721 A GNU/Linux port for the CRIS architecture, currently targeting
2722 @samp{ETRAX 100 LX} by default.
2725 For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11
2726 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2728 Pre-packaged tools can be obtained from
2729 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2730 information about this platform is available at
2731 @uref{http://developer.axis.com/}.
2736 @heading @anchor{crx}CRX
2738 The CRX CompactRISC architecture is a low-power 32-bit architecture with
2739 fast context switching and architectural extensibility features.
2742 @xref{CRX Options,, CRX Options, gcc, Using and Porting the GNU Compiler
2747 See ``CRX Options'' in the main manual for a list of CRX-specific options.
2750 Use @samp{configure --target=crx-elf --enable-languages=c,c++} to configure
2751 GCC@ for building a CRX cross-compiler. The option @samp{--target=crx-elf}
2752 is also used to build the @samp{newlib} C library for CRX.
2754 It is also possible to build libstdc++-v3 for the CRX architecture. This
2755 needs to be done in a separate step with the following configure settings:
2756 @samp{gcc/libstdc++-v3/configure --host=crx-elf --with-newlib
2757 --enable-sjlj-exceptions --enable-cxx-flags='-fexceptions -frtti'}
2762 @heading @anchor{dos}DOS
2764 Please have a look at the @uref{binaries.html,,binaries page}.
2766 You cannot install GCC by itself on MSDOS; it will not compile under
2767 any MSDOS compiler except itself. You need to get the complete
2768 compilation package DJGPP, which includes binaries as well as sources,
2769 and includes all the necessary compilation tools and libraries.
2774 @heading @anchor{x-x-freebsd}*-*-freebsd*
2776 The version of binutils installed in @file{/usr/bin} probably works with
2777 this release of GCC@. However, on FreeBSD 4, bootstrapping against the
2778 latest FSF binutils is known to improve overall testsuite results; and,
2779 on FreeBSD/alpha, using binutils 2.14 or later is required to build libjava.
2781 Support for FreeBSD 1 was discontinued in GCC 3.2.
2783 Support for FreeBSD 2 will be discontinued after GCC 3.4. The
2784 following was true for GCC 3.1 but the current status is unknown.
2785 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2786 configuration support and files as shipped with GCC 2.95 are still in
2787 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2788 it is unknown which version of binutils was used (it is assumed that it
2789 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2791 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2792 default for all CPU architectures. It had been the default on
2793 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2794 of @option{-g}, if you really want the old debugging format. There are
2795 no known issues with mixing object files and libraries with different
2796 debugging formats. Otherwise, this release of GCC should now match more
2797 of the configuration used in the stock FreeBSD configuration of GCC@. In
2798 particular, @option{--enable-threads} is now configured by default.
2799 However, as a general user, do not attempt to replace the system
2800 compiler with this release. Known to bootstrap and check with good
2801 results on FreeBSD 4.9-STABLE and 5-CURRENT@. In the past, known to
2802 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
2803 4.3, 4.4, 4.5, 4.8-STABLE@.
2805 In principle, @option{--enable-threads} is now compatible with
2806 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
2807 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
2809 library may be incorrectly built (symbols are missing at link time).
2810 There is a rare timing-based startup hang (probably involves an
2811 assumption about the thread library). Multi-threaded boehm-gc (required for
2812 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
2813 4.5-RELEASE@. Other CPU architectures
2814 supported by FreeBSD will require additional configuration tuning in, at
2815 the very least, both boehm-gc and libffi.
2817 Shared @file{libgcc_s.so} is now built and installed by default.
2822 @heading @anchor{h8300-hms}h8300-hms
2823 Renesas H8/300 series of processors.
2825 Please have a look at the @uref{binaries.html,,binaries page}.
2827 The calling convention and structure layout has changed in release 2.6.
2828 All code must be recompiled. The calling convention now passes the
2829 first three arguments in function calls in registers. Structures are no
2830 longer a multiple of 2 bytes.
2835 @heading @anchor{hppa-hp-hpux}hppa*-hp-hpux*
2836 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
2838 We require using gas/binutils on all hppa platforms;
2839 you may encounter a variety of problems if you try to use the HP assembler.
2841 Specifically, @option{-g} does not work on HP-UX (since that system
2842 uses a peculiar debugging format which GCC does not know about), unless
2843 you use GAS and GDB@. It may be helpful to configure GCC with the
2844 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
2845 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
2847 If you wish to use the pa-risc 2.0 architecture support with a 32-bit
2848 runtime, you must use gas/binutils 2.11 or newer.
2850 There are two default scheduling models for instructions. These are
2851 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
2852 architecture specified for the target machine when configuring.
2853 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
2854 the target is a @samp{hppa1*} machine.
2856 The PROCESSOR_8000 model is not well suited to older processors. Thus,
2857 it is important to completely specify the machine architecture when
2858 configuring if you want a model other than PROCESSOR_8000. The macro
2859 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
2860 default scheduling model is desired.
2862 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
2863 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
2864 This namespace change might cause problems when bootstrapping with
2865 an earlier version of GCC or the HP compiler as essentially the same
2866 namespace is required for an entire build. This problem can be avoided
2867 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
2868 or @samp{98}. Another way is to add an appropriate set of predefines
2869 to @env{CC}. The description for the @option{munix=} option contains
2870 a list of the predefines used with each standard.
2872 As of GCC 4.1, @env{DWARF2} exception handling is available on HP-UX.
2873 It is now the default. This exposed a bug in the handling of data
2874 relocations in the GAS assembler. The handling of 64-bit data relocations
2875 was seriously broken, affecting debugging and exception support on all
2876 @samp{hppa64-*-*} targets. Under some circumstances, 32-bit data relocations
2877 could also be handled incorrectly. This problem is fixed in GAS version
2880 GCC versions prior to 4.1 incorrectly passed and returned complex
2881 values. They are now passed in the same manner as aggregates.
2883 More specific information to @samp{hppa*-hp-hpux*} targets follows.
2888 @heading @anchor{hppa-hp-hpux10}hppa*-hp-hpux10
2890 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
2891 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
2897 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
2901 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
2905 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
2908 The HP assembler on these systems has some problems. Most notably the
2909 assembler inserts timestamps into each object file it creates, causing
2910 the 3-stage comparison test to fail during a bootstrap.
2911 You should be able to continue by saying @samp{make all-host all-target}
2912 after getting the failure from @samp{make}.
2914 GCC 4.0 requires CVS binutils as of April 28, 2004 or later. Earlier
2915 versions require binutils 2.8 or later.
2917 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
2918 used for one-only code and data. This resolves many of the previous
2919 problems in using C++ on this target. However, the ABI is not compatible
2920 with the one implemented under HP-UX 11 using secondary definitions.
2925 @heading @anchor{hppa-hp-hpux11}hppa*-hp-hpux11
2927 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
2928 be used to compile GCC 3.0 and up.
2930 Refer to @uref{binaries.html,,binaries} for information about obtaining
2931 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
2932 to build the Ada language as it can't be bootstrapped using C@. Ada is
2933 only available for the 32-bit PA-RISC runtime. The libffi and libjava
2934 haven't been ported to HP-UX and don't build.
2936 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
2937 bundled compiler supports only traditional C; you will need either HP's
2938 unbundled compiler, or a binary distribution of GCC@.
2940 It is possible to build GCC 3.3 starting with the bundled HP compiler,
2941 but the process requires several steps. GCC 3.3 can then be used to
2942 build later versions. The fastjar program contains ISO C code and
2943 can't be built with the HP bundled compiler. This problem can be
2944 avoided by not building the Java language. For example, use the
2945 @option{--enable-languages="c,c++,f77,objc"} option in your configure
2948 There are several possible approaches to building the distribution.
2949 Binutils can be built first using the HP tools. Then, the GCC
2950 distribution can be built. The second approach is to build GCC
2951 first using the HP tools, then build binutils, then rebuild GCC@.
2952 There have been problems with various binary distributions, so it
2953 is best not to start from a binary distribution.
2955 On 64-bit capable systems, there are two distinct targets. Different
2956 installation prefixes must be used if both are to be installed on
2957 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
2958 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
2959 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
2960 PA-RISC 2.0 architecture. The HP and GNU linkers are both supported
2963 The script config.guess now selects the target type based on the compiler
2964 detected during configuration. You must define @env{PATH} or @env{CC} so
2965 that configure finds an appropriate compiler for the initial bootstrap.
2966 When @env{CC} is used, the definition should contain the options that are
2967 needed whenever @env{CC} is used.
2969 Specifically, options that determine the runtime architecture must be
2970 in @env{CC} to correctly select the target for the build. It is also
2971 convenient to place many other compiler options in @env{CC}. For example,
2972 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
2973 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
2974 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
2975 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
2976 macro definition table of cpp needs to be increased for a successful
2977 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
2978 be defined when building with the bundled compiler, or when using the
2979 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
2981 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
2982 with the @option{--with-ld=@dots{}} option. This overrides the standard
2983 search for ld. The two linkers supported on this target require different
2984 commands. The default linker is determined during configuration. As a
2985 result, it's not possible to switch linkers in the middle of a GCC build.
2986 This has been been reported to sometimes occur in unified builds of
2989 GCC 3.0 through 3.2 require binutils 2.11 or above. GCC 3.3 through
2990 GCC 4.0 require binutils 2.14 or later.
2992 Although the HP assembler can be used for an initial build, it shouldn't
2993 be used with any languages other than C and perhaps Fortran due to its
2994 many limitations. For example, it does not support weak symbols or alias
2995 definitions. As a result, explicit template instantiations are required
2996 when using C++. This makes it difficult if not impossible to build many
2997 C++ applications. You can't generate debugging information when using
2998 the HP assembler. Finally, bootstrapping fails in the final
2999 comparison of object modules due to the time stamps that it inserts into
3000 the modules. The bootstrap can be continued from this point with
3001 @samp{make all-host all-target}.
3003 A recent linker patch must be installed for the correct operation of
3004 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
3005 oldest linker patches that are known to work. They are for HP-UX
3006 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
3007 @code{PHSS_24304}, might be usable but it hasn't been tested. These
3008 patches have been superseded. Consult the HP patch database to obtain
3009 the currently recommended linker patch for your system.
3011 The patches are necessary for the support of weak symbols on the
3012 32-bit port, and for the running of initializers and finalizers. Weak
3013 symbols are implemented using SOM secondary definition symbols. Prior
3014 to HP-UX 11, there are bugs in the linker support for secondary symbols.
3015 The patches correct a problem of linker core dumps creating shared
3016 libraries containing secondary symbols, as well as various other
3017 linking issues involving secondary symbols.
3019 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
3020 run initializers and finalizers on the 64-bit port. The 32-bit port
3021 uses the linker @option{+init} and @option{+fini} options for the same
3022 purpose. The patches correct various problems with the +init/+fini
3023 options, including program core dumps. Binutils 2.14 corrects a
3024 problem on the 64-bit port resulting from HP's non-standard use of
3025 the .init and .fini sections for array initializers and finalizers.
3027 There are a number of issues to consider in selecting which linker to
3028 use with the 64-bit port. The GNU 64-bit linker can only create dynamic
3029 binaries. The @option{-static} option causes linking with archive
3030 libraries but doesn't produce a truly static binary. Dynamic binaries
3031 still require final binding by the dynamic loader to resolve a set of
3032 dynamic-loader-defined symbols. The default behavior of the HP linker
3033 is the same as the GNU linker. However, it can generate true 64-bit
3034 static binaries using the @option{+compat} option.
3036 The HP 64-bit linker doesn't support linkonce semantics. As a
3037 result, C++ programs have many more sections than they should.
3039 The GNU 64-bit linker has some issues with shared library support
3040 and exceptions. As a result, we only support libgcc in archive
3041 format. For similar reasons, dwarf2 unwind and exception support
3042 are disabled. The GNU linker also has problems creating binaries
3043 with @option{-static}. It doesn't provide stubs for internal
3044 calls to global functions in shared libraries, so these calls
3045 can't be overloaded.
3047 Thread support is not implemented in GCC 3.0 through 3.2, so the
3048 @option{--enable-threads} configure option does not work. In 3.3
3049 and later, POSIX threads are supported. The optional DCE thread
3050 library is not supported.
3052 This port still is undergoing significant development.
3057 @heading @anchor{x-x-linux-gnu}*-*-linux-gnu
3059 Versions of libstdc++-v3 starting with 3.2.1 require bugfixes present
3060 in glibc 2.2.5 and later. More information is available in the
3061 libstdc++-v3 documentation.
3066 @heading @anchor{ix86-x-linuxaout}i?86-*-linux*aout
3067 Use this configuration to generate @file{a.out} binaries on Linux-based
3068 GNU systems. This configuration is being superseded.
3073 @heading @anchor{ix86-x-linux}i?86-*-linux*
3075 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
3076 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
3078 If you receive Signal 11 errors when building on GNU/Linux, then it is
3079 possible you have a hardware problem. Further information on this can be
3080 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
3085 @heading @anchor{ix86-x-sco32v5}i?86-*-sco3.2v5*
3086 Use this for the SCO OpenServer Release 5 family of operating systems.
3088 Unlike earlier versions of GCC, the ability to generate COFF with this
3089 target is no longer provided.
3091 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
3092 the system debugger to be used. That support was too burdensome to
3093 maintain. GCC now emits only DWARF 2 for this target. This means you
3094 may use either the UDK debugger or GDB to debug programs built by this
3097 GCC is now only supported on releases 5.0.4 and later, and requires that
3098 you install Support Level Supplement OSS646B or later, and Support Level
3099 Supplement OSS631C or later. If you are using release 5.0.7 of
3100 OpenServer, you must have at least the first maintenance pack installed
3101 (this includes the relevant portions of OSS646). OSS646, also known as
3102 the ``Execution Environment Update'', provides updated link editors and
3103 assemblers, as well as updated standard C and math libraries. The C
3104 startup modules are also updated to support the System V gABI draft, and
3105 GCC relies on that behavior. OSS631 provides a collection of commonly
3106 used open source libraries, some of which GCC depends on (such as GNU
3107 gettext and zlib). SCO OpenServer Release 5.0.7 has all of this built
3108 in by default, but OSS631C and later also apply to that release. Please
3110 @uref{ftp://ftp.sco.com/pub/openserver5,,ftp://ftp.sco.com/pub/openserver5}
3111 for the latest versions of these (and other potentially useful)
3114 Although there is support for using the native assembler, it is
3115 recommended that you configure GCC to use the GNU assembler. You do
3116 this by using the flags
3117 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You should
3118 use a modern version of GNU binutils. Version 2.13.2.1 was used for all
3119 testing. In general, only the @option{--with-gnu-as} option is tested.
3120 A modern bintuils (as well as a plethora of other development related
3121 GNU utilities) can be found in Support Level Supplement OSS658A, the
3122 ``GNU Development Tools'' package. See the SCO web and ftp sites for details.
3123 That package also contains the currently ``officially supported'' version of
3124 GCC, version 2.95.3. It is useful for bootstrapping this version.
3129 @heading @anchor{ix86-x-solaris210}i?86-*-solaris2.10
3130 Use this for Solaris 10 or later on x86 and x86-64 systems. This
3131 configuration is supported by GCC 4.0 and later versions only.
3133 It is recommended that you configure GCC to use the GNU assembler in
3134 @file{/usr/sfw/bin/gas} but the Sun linker, using the options
3135 @option{--with-gnu-as --with-as=/usr/sfw/bin/gas --without-gnu-ld
3136 --with-ld=/usr/ccs/bin/ld}.
3141 @heading @anchor{ix86-x-udk}i?86-*-udk
3143 This target emulates the SCO Universal Development Kit and requires that
3144 package be installed. (If it is installed, you will have a
3145 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
3146 @samp{i?86-*-unixware7*} target
3147 but is meant to be used when hosting on a system where UDK isn't the
3148 default compiler such as OpenServer 5 or Unixware 2. This target will
3149 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
3150 with the same warnings and caveats as the SCO UDK@.
3152 This target is a little tricky to build because we have to distinguish
3153 it from the native tools (so it gets headers, startups, and libraries
3154 from the right place) while making the tools not think we're actually
3155 building a cross compiler. The easiest way to do this is with a configure
3159 CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure \
3160 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-
3163 @emph{You should substitute @samp{i686} in the above command with the appropriate
3164 processor for your host.}
3166 After the usual @samp{make} and
3167 @samp{make install}, you can then access the UDK-targeted GCC
3168 tools by adding @command{udk-} before the commonly known name. For
3169 example, to invoke the C compiler, you would use @command{udk-gcc}.
3170 They will coexist peacefully with any native-target GCC tools you may
3177 @heading @anchor{ia64-x-linux}ia64-*-linux
3178 IA-64 processor (also known as IPF, or Itanium Processor Family)
3181 If you are using the installed system libunwind library with
3182 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
3185 None of the following versions of GCC has an ABI that is compatible
3186 with any of the other versions in this list, with the exception that
3187 Red Hat 2.96 and Trillian 000171 are compatible with each other:
3188 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
3189 This primarily affects C++ programs and programs that create shared libraries.
3190 GCC 3.1 or later is recommended for compiling linux, the kernel.
3191 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
3192 more major ABI changes are expected.
3197 @heading @anchor{ia64-x-hpux}ia64-*-hpux*
3198 Building GCC on this target requires the GNU Assembler. The bundled HP
3199 assembler will not work. To prevent GCC from using the wrong assembler,
3200 the option @option{--with-gnu-as} may be necessary.
3202 The GCC libunwind library has not been ported to HPUX@. This means that for
3203 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
3204 is required to build GCC@. For GCC 3.3 and later, this is the default.
3205 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
3206 removed and the system libunwind library will always be used.
3210 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
3212 @heading @anchor{x-ibm-aix}*-ibm-aix*
3213 Support for AIX version 3 and older was discontinued in GCC 3.4.
3215 ``out of memory'' bootstrap failures may indicate a problem with
3216 process resource limits (ulimit). Hard limits are configured in the
3217 @file{/etc/security/limits} system configuration file.
3219 To speed up the configuration phases of bootstrapping and installing GCC,
3220 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
3223 % CONFIG_SHELL=/opt/freeware/bin/bash
3224 % export CONFIG_SHELL
3227 and then proceed as described in @uref{build.html,,the build
3228 instructions}, where we strongly recommend specifying an absolute path
3229 to invoke @var{srcdir}/configure.
3231 Errors involving @code{alloca} when building GCC generally are due
3232 to an incorrect definition of @code{CC} in the Makefile or mixing files
3233 compiled with the native C compiler and GCC@. During the stage1 phase of
3234 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
3235 (not @command{xlc}). Once @command{configure} has been informed of
3236 @command{xlc}, one needs to use @samp{make distclean} to remove the
3237 configure cache files and ensure that @env{CC} environment variable
3238 does not provide a definition that will confuse @command{configure}.
3239 If this error occurs during stage2 or later, then the problem most likely
3240 is the version of Make (see above).
3242 The native @command{as} and @command{ld} are recommended for bootstrapping
3243 on AIX 4 and required for bootstrapping on AIX 5L@. The GNU Assembler
3244 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
3245 utilize weak symbol functionality although it is not supported. The GNU
3246 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC@.
3247 The native AIX tools do interoperate with GCC@.
3249 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
3250 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
3251 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
3252 referenced as APAR IY53606 (AIX 5.2) or a APAR IY54774 (AIX 5.1)
3254 @samp{libstdc++} in GCC 3.4 increments the major version number of the
3255 shared object and GCC installation places the @file{libstdc++.a}
3256 shared library in a common location which will overwrite the and GCC
3257 3.3 version of the shared library. Applications either need to be
3258 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
3259 versions of the @samp{libstdc++} shared object needs to be available
3260 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
3261 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
3262 installed for runtime dynamic loading using the following steps to set
3263 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3264 multilib @file{libstdc++.a} installed:
3266 Extract the shared objects from the currently installed
3267 @file{libstdc++.a} archive:
3269 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3272 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3273 available for runtime dynamic loading, but not linking:
3275 % strip -e libstdc++.so.4 libstdc++.so.5
3278 Archive the runtime-only shared object in the GCC 3.4
3279 @file{libstdc++.a} archive:
3281 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
3284 Linking executables and shared libraries may produce warnings of
3285 duplicate symbols. The assembly files generated by GCC for AIX always
3286 have included multiple symbol definitions for certain global variable
3287 and function declarations in the original program. The warnings should
3288 not prevent the linker from producing a correct library or runnable
3291 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
3292 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
3293 to parse archive libraries did not handle the new format correctly.
3294 These routines are used by GCC and result in error messages during
3295 linking such as ``not a COFF file''. The version of the routines shipped
3296 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
3297 option of the archive command may be used to create archives of 32-bit
3298 objects using the original ``small format''. A correct version of the
3299 routines is shipped with AIX 4.3.2 and above.
3301 Some versions of the AIX binder (linker) can fail with a relocation
3302 overflow severe error when the @option{-bbigtoc} option is used to link
3303 GCC-produced object files into an executable that overflows the TOC@. A fix
3304 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
3305 available from IBM Customer Support and from its
3306 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3307 website as PTF U455193.
3309 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
3310 with a segmentation fault when invoked by any version of GCC@. A fix for
3311 APAR IX87327 is available from IBM Customer Support and from its
3312 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3313 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
3315 The initial assembler shipped with AIX 4.3.0 generates incorrect object
3316 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
3317 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
3318 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3319 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
3321 AIX provides National Language Support (NLS)@. Compilers and assemblers
3322 use NLS to support locale-specific representations of various data
3323 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
3324 separating decimal fractions). There have been problems reported where
3325 GCC does not produce the same floating-point formats that the assembler
3326 expects. If one encounters this problem, set the @env{LANG}
3327 environment variable to @samp{C} or @samp{En_US}.
3329 By default, GCC for AIX 4.1 and above produces code that can be used on
3330 both Power or PowerPC processors.
3332 A default can be specified with the @option{-mcpu=@var{cpu_type}}
3333 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
3338 @heading @anchor{iq2000-x-elf}iq2000-*-elf
3339 Vitesse IQ2000 processors. These are used in embedded
3340 applications. There are no standard Unix configurations.
3345 @heading @anchor{m32c-x-elf}m32c-*-elf
3346 Renesas M32C processor.
3347 This configuration is intended for embedded systems.
3352 @heading @anchor{m32r-x-elf}m32r-*-elf
3353 Renesas M32R processor.
3354 This configuration is intended for embedded systems.
3359 @heading @anchor{m6811-elf}m6811-elf
3360 Motorola 68HC11 family micro controllers. These are used in embedded
3361 applications. There are no standard Unix configurations.
3366 @heading @anchor{m6812-elf}m6812-elf
3367 Motorola 68HC12 family micro controllers. These are used in embedded
3368 applications. There are no standard Unix configurations.
3373 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
3374 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
3375 the assembler that prevents compilation of GCC@. This
3376 bug manifests itself during the first stage of compilation, while
3377 building @file{libgcc2.a}:
3381 cc1: warning: `-g' option not supported on this version of GCC
3382 cc1: warning: `-g1' option not supported on this version of GCC
3383 ./xgcc: Internal compiler error: program as got fatal signal 11
3386 A patched version of the assembler is available as the file
3387 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
3388 have HP software support, the patch can also be obtained directly from
3389 HP, as described in the following note:
3392 This is the patched assembler, to patch SR#1653-010439, where the
3393 assembler aborts on floating point constants.
3395 The bug is not really in the assembler, but in the shared library
3396 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
3397 SR#4701-078451. Anyway, the attached assembler uses the archive
3398 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
3401 This patch is also known as PHCO_4484.
3403 In addition gdb does not understand that native HP-UX format, so
3404 you must use gas if you wish to use gdb.
3406 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
3407 @command{fixproto} shell script triggers a bug in the system shell. If you
3408 encounter this problem, upgrade your operating system or use BASH (the
3409 GNU shell) to run @command{fixproto}. This bug will cause the fixproto
3410 program to report an error of the form:
3413 ./fixproto: sh internal 1K buffer overflow
3416 To fix this, you can also change the first line of the fixproto script
3426 @heading @anchor{mips-x-x}mips-*-*
3427 If on a MIPS system you get an error message saying ``does not have gp
3428 sections for all it's [sic] sectons [sic]'', don't worry about it. This
3429 happens whenever you use GAS with the MIPS linker, but there is not
3430 really anything wrong, and it is okay to use the output file. You can
3431 stop such warnings by installing the GNU linker.
3433 It would be nice to extend GAS to produce the gp tables, but they are
3434 optional, and there should not be a warning about their absence.
3436 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
3437 and later. A patch went in just after the GCC 3.3 release to
3438 make @samp{mips*-*-*} use the generic implementation instead. You can also
3439 configure for @samp{mipsel-elf} as a workaround. The
3440 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
3441 work on this is expected in future releases.
3443 MIPS systems check for division by zero (unless
3444 @option{-mno-check-zero-division} is passed to the compiler) by
3445 generating either a conditional trap or a break instruction. Using
3446 trap results in smaller code, but is only supported on MIPS II and
3447 later. Also, some versions of the Linux kernel have a bug that
3448 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
3449 the use of break, use the @option{--with-divide=breaks}
3450 @command{configure} option when configuring GCC@. The default is to
3451 use traps on systems that support them.
3453 Cross-compilers for the MIPS as target using the MIPS assembler
3454 currently do not work, because the auxiliary programs
3455 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
3456 anything but a MIPS. It does work to cross compile for a MIPS
3457 if you use the GNU assembler and linker.
3459 The linker from GNU binutils versions prior to 2.17 has a bug which
3460 causes the runtime linker stubs in @file{libgcj.so} to be incorrectly
3461 generated. If you want to use libgcj, either use binutils 2.17 or
3462 later to build it or export @samp{LD_BIND_NOW=1} in your runtime environment.
3467 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
3469 In order to compile GCC on an SGI running IRIX 5, the @samp{compiler_dev.hdr}
3470 subsystem must be installed from the IDO CD-ROM supplied by SGI@.
3471 It is also available for download from
3472 @uref{ftp://ftp.sgi.com/sgi/IRIX5.3/iris-development-option-5.3.tardist}.
3474 If you use the MIPS C compiler to bootstrap, it may be necessary
3475 to increase its table size for switch statements with the
3476 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
3477 optimization option, you also need to use @option{-Olimit 3000}.
3479 To enable debugging under IRIX 5, you must use GNU binutils 2.15 or
3480 later, and use the @option{--with-gnu-ld} @command{configure} option
3481 when configuring GCC@. You need to use GNU @command{ar} and @command{nm},
3482 also distributed with GNU binutils.
3484 Some users have reported that @command{/bin/sh} will hang during bootstrap.
3485 This problem can be avoided by running the commands:
3488 % CONFIG_SHELL=/bin/ksh
3489 % export CONFIG_SHELL
3492 before starting the build.
3497 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3499 If you are using SGI's MIPSpro @command{cc} as your bootstrap compiler, you must
3500 ensure that the N32 ABI is in use. To test this, compile a simple C
3501 file with @command{cc} and then run @command{file} on the
3502 resulting object file. The output should look like:
3505 test.o: ELF N32 MSB @dots{}
3511 test.o: ELF 32-bit MSB @dots{}
3517 test.o: ELF 64-bit MSB @dots{}
3520 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3521 should set the environment variable @env{CC} to @samp{cc -n32}
3522 before configuring GCC@.
3524 If you want the resulting @command{gcc} to run on old 32-bit systems
3525 with the MIPS R4400 CPU, you need to ensure that only code for the @samp{mips3}
3526 instruction set architecture (ISA) is generated. While GCC 3.x does
3527 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3528 the ISA depending on the machine where GCC is built. Using one of them
3529 as the bootstrap compiler may result in @samp{mips4} code, which won't run at
3530 all on @samp{mips3}-only systems. For the test program above, you should see:
3533 test.o: ELF N32 MSB mips-3 @dots{}
3539 test.o: ELF N32 MSB mips-4 @dots{}
3542 instead, you should set the environment variable @env{CC} to @samp{cc
3543 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3545 MIPSpro C 7.4 may cause bootstrap failures, due to a bug when inlining
3546 @code{memcmp}. Either add @code{-U__INLINE_INTRINSICS} to the @env{CC}
3547 environment variable as a workaround or upgrade to MIPSpro C 7.4.1m.
3549 GCC on IRIX 6 is usually built to support the N32, O32 and N64 ABIs. If
3550 you build GCC on a system that doesn't have the N64 libraries installed
3551 or cannot run 64-bit binaries,
3552 you need to configure with @option{--disable-multilib} so GCC doesn't
3553 try to use them. This will disable building the O32 libraries, too.
3554 Look for @file{/usr/lib64/libc.so.1} to see if you
3555 have the 64-bit libraries installed.
3557 To enable debugging for the O32 ABI, you must use GNU @command{as} from
3558 GNU binutils 2.15 or later. You may also use GNU @command{ld}, but
3559 this is not required and currently causes some problems with Ada.
3561 The @option{--enable-threads} option doesn't currently work, a patch is
3562 in preparation for a future release. The @option{--enable-libgcj}
3563 option is disabled by default: IRIX 6 uses a very low default limit
3564 (20480) for the command line length. Although @command{libtool} contains a
3565 workaround for this problem, at least the N64 @samp{libgcj} is known not
3566 to build despite this, running into an internal error of the native
3567 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3568 its maximum of 262144 bytes. If you have root access, you can use the
3569 @command{systune} command to do this.
3571 @code{wchar_t} support in @samp{libstdc++} is not available for old
3572 IRIX 6.5.x releases, @math{x < 19}. The problem cannot be autodetected
3573 and in order to build GCC for such targets you need to configure with
3574 @option{--disable-wchar_t}.
3576 See @uref{http://freeware.sgi.com/} for more
3577 information about using GCC on IRIX platforms.
3582 @heading @anchor{powerpc-x-x}powerpc-*-*
3584 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3585 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3590 @heading @anchor{powerpc-x-darwin}powerpc-*-darwin*
3591 PowerPC running Darwin (Mac OS X kernel).
3593 Pre-installed versions of Mac OS X may not include any developer tools,
3594 meaning that you will not be able to build GCC from source. Tool
3595 binaries are available at
3596 @uref{http://developer.apple.com/darwin/projects/compiler/} (free
3597 registration required).
3599 This version of GCC requires at least cctools-590.7.
3601 The version of GCC shipped by Apple typically includes a number of
3602 extensions not available in a standard GCC release. These extensions
3603 are generally for backwards compatibility and best avoided.
3608 @heading @anchor{powerpc-x-elf}powerpc-*-elf, powerpc-*-sysv4
3609 PowerPC system in big endian mode, running System V.4.
3614 @heading @anchor{powerpc-x-linux-gnu}powerpc*-*-linux-gnu*
3617 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.15}
3618 or newer for a working GCC@.
3623 @heading @anchor{powerpc-x-netbsd}powerpc-*-netbsd*
3624 PowerPC system in big endian mode running NetBSD@. To build the
3625 documentation you will need Texinfo version 4.4 (NetBSD 1.5.1 included
3626 Texinfo version 3.12).
3631 @heading @anchor{powerpc-x-eabisim}powerpc-*-eabisim
3632 Embedded PowerPC system in big endian mode for use in running under the
3638 @heading @anchor{powerpc-x-eabi}powerpc-*-eabi
3639 Embedded PowerPC system in big endian mode.
3644 @heading @anchor{powerpcle-x-elf}powerpcle-*-elf, powerpcle-*-sysv4
3645 PowerPC system in little endian mode, running System V.4.
3650 @heading @anchor{powerpcle-x-eabisim}powerpcle-*-eabisim
3651 Embedded PowerPC system in little endian mode for use in running under
3657 @heading @anchor{powerpcle-x-eabi}powerpcle-*-eabi
3658 Embedded PowerPC system in little endian mode.
3663 @heading @anchor{s390-x-linux}s390-*-linux*
3664 S/390 system running GNU/Linux for S/390@.
3669 @heading @anchor{s390x-x-linux}s390x-*-linux*
3670 zSeries system (64-bit) running GNU/Linux for zSeries@.
3675 @heading @anchor{s390x-ibm-tpf}s390x-ibm-tpf*
3676 zSeries system (64-bit) running TPF@. This platform is
3677 supported as cross-compilation target only.
3682 @c Please use Solaris 2 to refer to all release of Solaris, starting
3683 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
3684 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3685 @c alone is too unspecific and must be avoided.
3686 @heading @anchor{x-x-solaris2}*-*-solaris2*
3688 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3689 GCC you first have to install a pre-built compiler, see the
3690 @uref{binaries.html,,binaries page} for details.
3692 The Solaris 2 @command{/bin/sh} will often fail to configure
3693 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3694 recommend using the following initial sequence of commands
3697 % CONFIG_SHELL=/bin/ksh
3698 % export CONFIG_SHELL
3701 and proceed as described in @uref{configure.html,,the configure instructions}.
3702 In addition we strongly recommend specifying an absolute path to invoke
3703 @var{srcdir}/configure.
3705 Solaris 2 comes with a number of optional OS packages. Some of these
3706 are needed to use GCC fully, namely @code{SUNWarc},
3707 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3708 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3709 optional packages when installing Solaris 2, you will need to verify that
3710 the packages that GCC needs are installed.
3712 To check whether an optional package is installed, use
3713 the @command{pkginfo} command. To add an optional package, use the
3714 @command{pkgadd} command. For further details, see the Solaris 2
3717 Trying to use the linker and other tools in
3718 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3719 For example, the linker may hang indefinitely. The fix is to remove
3720 @file{/usr/ucb} from your @env{PATH}.
3722 The build process works more smoothly with the legacy Sun tools so, if you
3723 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3724 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3726 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3727 platform. We recommend the use of GNU binutils 2.11.2 or later, or the
3728 vendor tools (Sun @command{as}, Sun @command{ld}). Note that your mileage
3729 may vary if you use a combination of the GNU tools and the Sun tools: while
3730 the combination GNU @command{as} + Sun @command{ld} should reasonably work,
3731 the reverse combination Sun @command{as} + GNU @command{ld} is known to
3732 cause memory corruption at runtime in some cases for C++ programs.
3734 The stock GNU binutils 2.15 release is broken on this platform because of a
3735 single bug. It has been fixed on the 2.15 branch in the CVS repository.
3736 You can obtain a working version by checking out the binutils-2_15-branch
3737 from the CVS repository or applying the patch
3738 @uref{http://sources.redhat.com/ml/binutils-cvs/2004-09/msg00036.html} to the
3741 We recommend using GNU binutils 2.16 or later in conjunction with GCC 4.x,
3742 or the vendor tools (Sun @command{as}, Sun @command{ld}). However, for
3743 Solaris 10 and above, an additional patch is required in order for the GNU
3744 linker to be able to cope with a new flavor of shared libraries. You
3745 can obtain a working version by checking out the binutils-2_16-branch from
3746 the CVS repository or applying the patch
3747 @uref{http://sourceware.org/ml/binutils-cvs/2005-07/msg00122.html} to the
3750 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3751 newer: @command{g++} will complain that types are missing. These headers assume
3752 that omitting the type means @code{int}; this assumption worked for C89 but
3753 is wrong for C++, and is now wrong for C99 also.
3755 @command{g++} accepts such (invalid) constructs with the option
3756 @option{-fpermissive}; it
3757 will assume that any missing type is @code{int} (as defined by C89).
3759 There are patches for Solaris 2.6 (105633-56 or newer for SPARC,
3760 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC,
3761 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3762 108653-22 for Intel) that fix this bug.
3764 Sun bug 4927647 sometimes causes random spurious testsuite failures
3765 related to missing diagnostic output. This bug doesn't affect GCC
3766 itself, rather it is a kernel bug triggered by the @command{expect}
3767 program which is used only by the GCC testsuite driver. When the bug
3768 causes the @command{expect} program to miss anticipated output, extra
3769 testsuite failures appear.
3771 There are patches for Solaris 8 (117350-12 or newer for SPARC,
3772 117351-12 or newer for Intel) and Solaris 9 (117171-11 or newer for
3773 SPARC, 117172-11 or newer for Intel) that address this problem.
3778 @heading @anchor{sparc-sun-solaris2}sparc-sun-solaris2*
3780 When GCC is configured to use binutils 2.11.2 or later the binaries
3781 produced are smaller than the ones produced using Sun's native tools;
3782 this difference is quite significant for binaries containing debugging
3785 Sun @command{as} 4.x is broken in that it cannot cope with long symbol names.
3786 A typical error message might look similar to the following:
3789 /usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041: error:
3790 can't compute value of an expression involving an external symbol.
3793 This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris
3794 2.6 and has been fixed in later (5.x) versions of the assembler,
3795 starting with Solaris 7.
3797 Starting with Solaris 7, the operating system is capable of executing
3798 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3799 this; the @option{-m64} option enables 64-bit code generation.
3800 However, if all you want is code tuned for the UltraSPARC CPU, you
3801 should try the @option{-mtune=ultrasparc} option instead, which produces
3802 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3805 When configuring on a Solaris 7 or later system that is running a kernel
3806 that supports only 32-bit binaries, one must configure with
3807 @option{--disable-multilib}, since we will not be able to build the
3808 64-bit target libraries.
3810 GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
3811 the GNU compiler (especially GCC 3.0.x versions), which lead to the
3812 miscompilation of the stage1 compiler and the subsequent failure of the
3813 bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
3814 stage, i.e.@: to bootstrap that compiler with the base compiler and then
3815 use it to bootstrap the final compiler.
3817 GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
3818 and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
3819 failure in form of a miscompilation of the stage1 compiler by the Sun
3820 compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
3822 GCC 3.4 changed the default debugging format from STABS to DWARF-2 for
3823 32-bit code on Solaris 7 and later. If you use the Sun assembler, this
3824 change apparently runs afoul of Sun bug 4910101 (which is referenced as
3825 a x86-only problem by Sun, probably because they do not use DWARF-2).
3826 A symptom of the problem is that you cannot compile C++ programs like
3827 @command{groff} 1.19.1 without getting messages similar to the following:
3830 ld: warning: relocation error: R_SPARC_UA32: @dots{}
3831 external symbolic relocation against non-allocatable section
3832 .debug_info cannot be processed at runtime: relocation ignored.
3835 To work around this problem, compile with @option{-gstabs+} instead of
3838 When configuring the GNU Multiple Precision Library (GMP) version 4.1.x
3839 on a Solaris 7 or later system, the canonical target triplet must be
3840 specified as the @command{build} parameter on the configure line:
3843 ./configure --build=sparc-sun-solaris2.7 --prefix=xxx --enable-mpfr
3849 @heading @anchor{sparc-sun-solaris27}sparc-sun-solaris2.7
3851 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3852 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3853 and later, including all EGCS releases. Sun formerly recommended
3854 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3855 recommend it only for people who use Sun's compilers.
3857 Here are some workarounds to this problem:
3860 Do not install Sun patch 107058-01 until after Sun releases a
3861 complete patch for bug 4210064. This is the simplest course to take,
3862 unless you must also use Sun's C compiler. Unfortunately 107058-01
3863 is preinstalled on some new Solaris 7-based hosts, so you may have to
3867 Copy the original, unpatched Solaris 7
3868 @command{/usr/ccs/bin/as} into
3869 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
3870 adjusting the latter name to fit your local conventions and software
3874 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3875 both 107058-01 and 106950-03 installed has reported the bug with GCC
3876 and Sun's dynamic linker. This last course of action is riskiest,
3877 for two reasons. First, you must install 106950 on all hosts that
3878 run code generated by GCC; it doesn't suffice to install it only on
3879 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3880 only a partial fix for bug 4210064, but Sun doesn't know whether the
3881 partial fix is adequate for GCC@. Revision -08 or later should fix
3882 the bug. The current (as of 2004-05-23) revision is -24, and is included in
3883 the Solaris 7 Recommended Patch Cluster.
3886 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
3887 which causes a bootstrap failure when linking the 64-bit shared version of
3888 libgcc. A typical error message is:
3891 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
3892 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
3895 This bug has been fixed in the final 5.0 version of the assembler.
3897 A similar problem was reported for version Sun WorkShop 6 99/08/18 of the
3898 Sun assembler, which causes a bootstrap failure with GCC 4.0.0:
3901 ld: fatal: relocation error: R_SPARC_DISP32:
3902 file .libs/libstdc++.lax/libsupc++convenience.a/vterminate.o:
3903 symbol <unknown>: offset 0xfccd33ad is non-aligned
3906 This bug has been fixed in more recent revisions of the assembler.
3911 @heading @anchor{sparc-x-linux}sparc-*-linux*
3913 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
3914 or newer on this platform. All earlier binutils and glibc
3915 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
3921 @heading @anchor{sparc64-x-solaris2}sparc64-*-solaris2*
3923 The following compiler flags must be specified in the configure
3924 step in order to bootstrap this target with the Sun compiler:
3927 % CC="cc -xildoff -xarch=v9" @var{srcdir}/configure [@var{options}] [@var{target}]
3930 @option{-xildoff} turns off the incremental linker, and @option{-xarch=v9}
3931 specifies the SPARC-V9 architecture to the Sun linker and assembler.
3936 @heading @anchor{sparcv9-x-solaris2}sparcv9-*-solaris2*
3938 This is a synonym for sparc64-*-solaris2*.
3943 @heading @anchor{x-x-sysv}*-*-sysv*
3944 On System V release 3, you may get this error message
3948 ld fatal: failed to write symbol name @var{something}
3949 in strings table for file @var{whatever}
3952 This probably indicates that the disk is full or your ulimit won't allow
3953 the file to be as large as it needs to be.
3955 This problem can also result because the kernel parameter @code{MAXUMEM}
3956 is too small. If so, you must regenerate the kernel and make the value
3957 much larger. The default value is reported to be 1024; a value of 32768
3958 is said to work. Smaller values may also work.
3960 On System V, if you get an error like this,
3963 /usr/local/lib/bison.simple: In function `yyparse':
3964 /usr/local/lib/bison.simple:625: virtual memory exhausted
3968 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
3970 On a System V release 4 system, make sure @file{/usr/bin} precedes
3971 @file{/usr/ucb} in @code{PATH}. The @command{cc} command in
3972 @file{/usr/ucb} uses libraries which have bugs.
3977 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3978 Don't try compiling with VAX C (@command{vcc}). It produces incorrect code
3979 in some cases (for example, when @code{alloca} is used).
3984 @heading @anchor{x-x-vxworks}*-*-vxworks*
3985 Support for VxWorks is in flux. At present GCC supports @emph{only} the
3986 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
3987 We welcome patches for other architectures supported by VxWorks 5.5.
3988 Support for VxWorks AE would also be welcome; we believe this is merely
3989 a matter of writing an appropriate ``configlette'' (see below). We are
3990 not interested in supporting older, a.out or COFF-based, versions of
3993 VxWorks comes with an older version of GCC installed in
3994 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
3995 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
3996 Before running @command{configure}, create the directories @file{@var{prefix}}
3997 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
3998 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
3999 include that directory while running both @command{configure} and
4002 You must give @command{configure} the
4003 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4004 find the VxWorks system headers. Since VxWorks is a cross compilation
4005 target only, you must also specify @option{--target=@var{target}}.
4006 @command{configure} will attempt to create the directory
4007 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4008 make sure the user running @command{configure} has sufficient privilege
4011 GCC's exception handling runtime requires a special ``configlette''
4012 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4013 that file to add the module to your kernel build. (Future versions of
4014 VxWorks will incorporate this module.)
4019 @heading @anchor{x86-64-x-x}x86_64-*-*, amd64-*-*
4021 GCC supports the x86-64 architecture implemented by the AMD64 processor
4022 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4023 On GNU/Linux the default is a bi-arch compiler which is able to generate
4024 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4029 @heading @anchor{xtensa-x-elf}xtensa-*-elf
4031 This target is intended for embedded Xtensa systems using the
4032 @samp{newlib} C library. It uses ELF but does not support shared
4033 objects. Designed-defined instructions specified via the
4034 Tensilica Instruction Extension (TIE) language are only supported
4035 through inline assembly.
4037 The Xtensa configuration information must be specified prior to
4038 building GCC@. The @file{include/xtensa-config.h} header
4039 file contains the configuration information. If you created your
4040 own Xtensa configuration with the Xtensa Processor Generator, the
4041 downloaded files include a customized copy of this header file,
4042 which you can use to replace the default header file.
4047 @heading @anchor{xtensa-x-linux}xtensa-*-linux*
4049 This target is for Xtensa systems running GNU/Linux. It supports ELF
4050 shared objects and the GNU C library (glibc). It also generates
4051 position-independent code (PIC) regardless of whether the
4052 @option{-fpic} or @option{-fPIC} options are used. In other
4053 respects, this target is the same as the
4054 @uref{#xtensa-*-elf,,@samp{xtensa-*-elf}} target.
4059 @heading @anchor{windows}Microsoft Windows (32-bit)
4061 Ports of GCC are included with the
4062 @uref{http://www.cygwin.com/,,Cygwin environment}.
4064 GCC will build under Cygwin without modification; it does not build
4065 with Microsoft's C++ compiler and there are no plans to make it do so.
4070 @heading @anchor{os2}OS/2
4072 GCC does not currently support OS/2. However, Andrew Zabolotny has been
4073 working on a generic OS/2 port with pgcc. The current code can be found
4074 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
4079 @heading @anchor{older}Older systems
4081 GCC contains support files for many older (1980s and early
4082 1990s) Unix variants. For the most part, support for these systems
4083 has not been deliberately removed, but it has not been maintained for
4084 several years and may suffer from bitrot.
4086 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
4087 Support for these systems is still present in that release, but
4088 @command{configure} will fail unless the @option{--enable-obsolete}
4089 option is given. Unless a maintainer steps forward, support for these
4090 systems will be removed from the next release of GCC@.
4092 Support for old systems as hosts for GCC can cause problems if the
4093 workarounds for compiler, library and operating system bugs affect the
4094 cleanliness or maintainability of the rest of GCC@. In some cases, to
4095 bring GCC up on such a system, if still possible with current GCC, may
4096 require first installing an old version of GCC which did work on that
4097 system, and using it to compile a more recent GCC, to avoid bugs in the
4098 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
4099 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
4100 sites}. Header bugs may generally be avoided using
4101 @command{fixincludes}, but bugs or deficiencies in libraries and the
4102 operating system may still cause problems.
4104 Support for older systems as targets for cross-compilation is less
4105 problematic than support for them as hosts for GCC; if an enthusiast
4106 wishes to make such a target work again (including resurrecting any of
4107 the targets that never worked with GCC 2, starting from the last
4108 version before they were removed), patches
4109 @uref{../contribute.html,,following the usual requirements} would be
4110 likely to be accepted, since they should not affect the support for more
4113 For some systems, old versions of GNU binutils may also be useful,
4114 and are available from @file{pub/binutils/old-releases} on
4115 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
4117 Some of the information on specific systems above relates to
4118 such older systems, but much of the information
4119 about GCC on such systems (which may no longer be applicable to
4120 current GCC) is to be found in the GCC texinfo manual.
4125 @heading @anchor{elf}all ELF targets (SVR4, Solaris 2, etc.)
4127 C++ support is significantly better on ELF targets if you use the
4128 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
4129 inlines, vtables and template instantiations will be discarded
4138 @uref{./index.html,,Return to the GCC Installation page}
4142 @c ***Old documentation******************************************************
4144 @include install-old.texi
4150 @uref{./index.html,,Return to the GCC Installation page}
4154 @c ***GFDL********************************************************************
4162 @uref{./index.html,,Return to the GCC Installation page}
4166 @c ***************************************************************************
4167 @c Part 6 The End of the Document
4169 @comment node-name, next, previous, up
4170 @node Concept Index, , GNU Free Documentation License, Top
4174 @unnumbered Concept Index