1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename gccinstall.info
5 @settitle Installing GCC
10 @include gcc-common.texi
12 @c Specify title for specific html page
14 @settitle Installing GCC
17 @settitle Host/Target specific installation notes for GCC
19 @ifset prerequisiteshtml
20 @settitle Prerequisites for GCC
23 @settitle Downloading GCC
26 @settitle Installing GCC: Configuration
29 @settitle Installing GCC: Building
32 @settitle Installing GCC: Testing
34 @ifset finalinstallhtml
35 @settitle Installing GCC: Final installation
38 @settitle Installing GCC: Binaries
41 @settitle Installing GCC: Old documentation
44 @settitle Installing GCC: GNU Free Documentation License
47 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
48 @c 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007 Free Software Foundation, Inc.
49 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
51 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
52 @c test the generation of HTML documents for the gcc.gnu.org web pages.
54 @c Do not use @footnote{} in this file as it breaks install.texi2html!
56 @c Include everything if we're not making html
60 @set prerequisiteshtml
71 @c Part 2 Summary Description and Copyright
73 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
74 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
76 Permission is granted to copy, distribute and/or modify this document
77 under the terms of the GNU Free Documentation License, Version 1.2 or
78 any later version published by the Free Software Foundation; with no
79 Invariant Sections, the Front-Cover texts being (a) (see below), and
80 with the Back-Cover Texts being (b) (see below). A copy of the
81 license is included in the section entitled ``@uref{./gfdl.html,,GNU
82 Free Documentation License}''.
84 (a) The FSF's Front-Cover Text is:
88 (b) The FSF's Back-Cover Text is:
90 You have freedom to copy and modify this GNU Manual, like GNU
91 software. Copies published by the Free Software Foundation raise
92 funds for GNU development.
97 @dircategory Software development
99 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
102 @c Part 3 Titlepage and Copyright
104 @title Installing GCC
107 @c The following two commands start the copyright page.
109 @vskip 0pt plus 1filll
113 @c Part 4 Top node, Master Menu, and/or Table of Contents
116 @comment node-name, next, Previous, up
119 * Installing GCC:: This document describes the generic installation
120 procedure for GCC as well as detailing some target
121 specific installation instructions.
123 * Specific:: Host/target specific installation notes for GCC.
124 * Binaries:: Where to get pre-compiled binaries.
126 * Old:: Old installation documentation.
128 * GNU Free Documentation License:: How you can copy and share this manual.
129 * Concept Index:: This index has two entries.
137 @c Part 5 The Body of the Document
138 @c ***Installing GCC**********************************************************
140 @comment node-name, next, previous, up
141 @node Installing GCC, Binaries, , Top
145 @chapter Installing GCC
148 The latest version of this document is always available at
149 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
151 This document describes the generic installation procedure for GCC as well
152 as detailing some target specific installation instructions.
154 GCC includes several components that previously were separate distributions
155 with their own installation instructions. This document supersedes all
156 package specific installation instructions.
158 @emph{Before} starting the build/install procedure please check the
160 @ref{Specific, host/target specific installation notes}.
163 @uref{specific.html,,host/target specific installation notes}.
165 We recommend you browse the entire generic installation instructions before
168 Lists of successful builds for released versions of GCC are
169 available at @uref{http://gcc.gnu.org/buildstat.html}.
170 These lists are updated as new information becomes available.
172 The installation procedure itself is broken into five steps.
177 * Downloading the source::
180 * Testing:: (optional)
187 @uref{prerequisites.html,,Prerequisites}
189 @uref{download.html,,Downloading the source}
191 @uref{configure.html,,Configuration}
193 @uref{build.html,,Building}
195 @uref{test.html,,Testing} (optional)
197 @uref{finalinstall.html,,Final install}
201 Please note that GCC does not support @samp{make uninstall} and probably
202 won't do so in the near future as this would open a can of worms. Instead,
203 we suggest that you install GCC into a directory of its own and simply
204 remove that directory when you do not need that specific version of GCC
205 any longer, and, if shared libraries are installed there as well, no
206 more binaries exist that use them.
209 There are also some @uref{old.html,,old installation instructions},
210 which are mostly obsolete but still contain some information which has
211 not yet been merged into the main part of this manual.
219 @uref{./index.html,,Return to the GCC Installation page}
225 @c ***Prerequisites**************************************************
227 @comment node-name, next, previous, up
228 @node Prerequisites, Downloading the source, , Installing GCC
230 @ifset prerequisiteshtml
232 @chapter Prerequisites
234 @cindex Prerequisites
236 GCC requires that various tools and packages be available for use in the
237 build procedure. Modifying GCC sources requires additional tools
240 @heading Tools/packages necessary for building GCC
242 @item ISO C90 compiler
243 Necessary to bootstrap GCC, although versions of GCC prior
244 to 3.4 also allow bootstrapping with a traditional (K&R) C compiler.
246 To build all languages in a cross-compiler or other configuration where
247 3-stage bootstrap is not performed, you need to start with an existing
248 GCC binary (version 2.95 or later) because source code for language
249 frontends other than C might use GCC extensions.
253 In order to build the Ada compiler (GNAT) you must already have GNAT
254 installed because portions of the Ada frontend are written in Ada (with
255 GNAT extensions.) Refer to the Ada installation instructions for more
256 specific information.
258 @item A ``working'' POSIX compatible shell, or GNU bash
260 Necessary when running @command{configure} because some
261 @command{/bin/sh} shells have bugs and may crash when configuring the
262 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
263 have disastrous corner-case performance problems. This
264 can cause target @command{configure} runs to literally take days to
265 complete in some cases.
267 So on some platforms @command{/bin/ksh} is sufficient, on others it
268 isn't. See the host/target specific instructions for your platform, or
269 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
270 environment to your ``good'' shell prior to running
271 @command{configure}/@command{make}.
273 @command{zsh} is not a fully compliant POSIX shell and will not
274 work when configuring GCC@.
278 Necessary in some circumstances, optional in others. See the
279 host/target specific instructions for your platform for the exact
282 @item gzip version 1.2.4 (or later) or
283 @itemx bzip2 version 1.0.2 (or later)
285 Necessary to uncompress GCC @command{tar} files when source code is
286 obtained via FTP mirror sites.
288 @item GNU make version 3.79.1 (or later)
290 You must have GNU make installed to build GCC@.
292 @item GNU tar version 1.14 (or later)
294 Necessary (only on some platforms) to untar the source code. Many
295 systems' @command{tar} programs will also work, only try GNU
296 @command{tar} if you have problems.
298 @item GNU Multiple Precision Library (GMP) version 4.1 (or later)
300 Necessary to build GCC. If you do not have it installed in your
301 library search path, you will have to configure with the
302 @option{--with-gmp} configure option. See also
303 @option{--with-gmp-lib} and @option{--with-gmp-include}.
305 @item MPFR Library version 2.2.1 (or later)
307 Necessary to build GCC. It can be downloaded from
308 @uref{http://www.mpfr.org/}. The version of MPFR that is bundled with
309 GMP 4.1.x contains numerous bugs. Although GCC may appear to function
310 with the buggy versions of MPFR, there are a few bugs that will not be
311 fixed when using this version. It is strongly recommended to upgrade
312 to the recommended version of MPFR.
314 The @option{--with-mpfr} configure option should be used if your MPFR
315 Library is not installed in your default library search path. See
316 also @option{--with-mpfr-lib} and @option{--with-mpfr-include}.
318 @item @command{jar}, or InfoZIP (@command{zip} and @command{unzip})
320 Necessary to build libgcj, the GCJ runtime.
325 @heading Tools/packages necessary for modifying GCC
327 @item autoconf versions 2.13 and 2.59
328 @itemx GNU m4 version 1.4 (or later)
330 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
331 to regenerate @file{configure} and @file{config.in} files. Most
332 directories require autoconf 2.59 (exactly), but the toplevel
333 still requires autoconf 2.13 (exactly).
335 @item automake version 1.9.6
337 Necessary when modifying a @file{Makefile.am} file to regenerate its
338 associated @file{Makefile.in}.
340 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
341 file. Specifically this applies to the @file{gcc}, @file{intl},
342 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
343 as any of their subdirectories.
345 For directories that use automake, GCC requires the latest release in
346 the 1.9.x series, which is currently 1.9.6. When regenerating a directory
347 to a newer version, please update all the directories using an older 1.9.x
348 to the latest released version.
350 @item gettext version 0.14.5 (or later)
352 Needed to regenerate @file{gcc.pot}.
354 @item gperf version 2.7.2 (or later)
356 Necessary when modifying @command{gperf} input files, e.g.@:
357 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
358 @file{gcc/cp/cfns.h}.
364 Necessary to run the GCC testsuite; see the section on testing for details.
366 @item autogen version 5.5.4 (or later) and
367 @itemx guile version 1.4.1 (or later)
369 Necessary to regenerate @file{fixinc/fixincl.x} from
370 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
372 Necessary to run @samp{make check} for @file{fixinc}.
374 Necessary to regenerate the top level @file{Makefile.in} file from
375 @file{Makefile.tpl} and @file{Makefile.def}.
377 @item GNU Bison version 1.28 (or later)
379 Necessary when modifying @file{*.y} files. Necessary to build the
380 @code{treelang} front end (which is not enabled by default) from a
381 checkout of the SVN repository; the generated files are not in the
382 repository. They are included in releases.
384 Berkeley @command{yacc} (@command{byacc}) has been reported to work
387 @item Flex version 2.5.4 (or later)
389 Necessary when modifying @file{*.l} files.
391 Necessary to build GCC during development because the generated output
392 files are not included in the SVN repository. They are included in
395 @item Texinfo version 4.4 (or later)
397 Necessary for running @command{makeinfo} when modifying @file{*.texi}
398 files to test your changes.
400 Necessary for running @command{make dvi} or @command{make pdf} to
401 create printable documentation in DVI or PDF format. Texinfo version
402 4.8 or later is required for @command{make pdf}.
404 Necessary to build GCC documentation during development because the
405 generated output files are not included in the SVN repository. They are
406 included in releases.
408 @item @TeX{} (any working version)
410 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
411 are used when running @command{make dvi} or @command{make pdf} to create
412 DVI or PDF files, respectively.
414 @item SVN (any version)
415 @itemx SSH (any version)
417 Necessary to access the SVN repository. Public releases and weekly
418 snapshots of the development sources are also available via FTP@.
420 @item Perl version 5.6.1 (or later)
422 Necessary when regenerating @file{Makefile} dependencies in libiberty.
423 Necessary when regenerating @file{libiberty/functions.texi}.
424 Necessary when generating manpages from Texinfo manuals.
425 Necessary when targetting Darwin, building libstdc++,
426 and not using @option{--disable-symvers}.
427 Used by various scripts to generate some files included in SVN (mainly
428 Unicode-related and rarely changing) from source tables.
430 @item GNU diffutils version 2.7 (or later)
432 Useful when submitting patches for the GCC source code.
434 @item patch version 2.5.4 (or later)
436 Necessary when applying patches, created with @command{diff}, to one's
442 If you wish to modify @file{.java} files in libjava, you will need to
443 configure with @option{--enable-java-maintainer-mode}, and you will need
444 to have executables named @command{ecj1} and @command{gjavah} in your path.
445 The @command{ecj1} executable should run the Eclipse Java compiler via
446 the GCC-specific entry point. You can download a suitable jar from
447 @uref{ftp://sourceware.org/pub/java/}, or by running the script
448 @command{contrib/download_ecj}.
457 @uref{./index.html,,Return to the GCC Installation page}
461 @c ***Downloading the source**************************************************
463 @comment node-name, next, previous, up
464 @node Downloading the source, Configuration, Prerequisites, Installing GCC
468 @chapter Downloading GCC
470 @cindex Downloading GCC
471 @cindex Downloading the Source
473 GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP
474 tarballs compressed with @command{gzip} or
475 @command{bzip2}. It is possible to download a full distribution or specific
478 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
479 for information on how to obtain GCC@.
481 The full distribution includes the C, C++, Objective-C, Fortran, Java,
482 and Ada (in the case of GCC 3.1 and later) compilers. The full
483 distribution also includes runtime libraries for C++, Objective-C,
484 Fortran, and Java. In GCC 3.0 and later versions, the GNU compiler
485 testsuites are also included in the full distribution.
487 If you choose to download specific components, you must download the core
488 GCC distribution plus any language specific distributions you wish to
489 use. The core distribution includes the C language front end as well as the
490 shared components. Each language has a tarball which includes the language
491 front end as well as the language runtime (when appropriate).
493 Unpack the core distribution as well as any language specific
494 distributions in the same directory.
496 If you also intend to build binutils (either to upgrade an existing
497 installation or for use in place of the corresponding tools of your
498 OS), unpack the binutils distribution either in the same directory or
499 a separate one. In the latter case, add symbolic links to any
500 components of the binutils you intend to build alongside the compiler
501 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
502 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
509 @uref{./index.html,,Return to the GCC Installation page}
513 @c ***Configuration***********************************************************
515 @comment node-name, next, previous, up
516 @node Configuration, Building, Downloading the source, Installing GCC
520 @chapter Installing GCC: Configuration
522 @cindex Configuration
523 @cindex Installing GCC: Configuration
525 Like most GNU software, GCC must be configured before it can be built.
526 This document describes the recommended configuration procedure
527 for both native and cross targets.
529 We use @var{srcdir} to refer to the toplevel source directory for
530 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
532 If you obtained the sources via SVN, @var{srcdir} must refer to the top
533 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
534 and not its @file{gcc} subdirectory, otherwise the build will fail.
536 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
537 file system, the shell's built-in @command{pwd} command will return
538 temporary pathnames. Using these can lead to various sorts of build
539 problems. To avoid this issue, set the @env{PWDCMD} environment
540 variable to an automounter-aware @command{pwd} command, e.g.,
541 @command{pawd} or @samp{amq -w}, during the configuration and build
544 First, we @strong{highly} recommend that GCC be built into a
545 separate directory than the sources which does @strong{not} reside
546 within the source tree. This is how we generally build GCC; building
547 where @var{srcdir} == @var{objdir} should still work, but doesn't
548 get extensive testing; building where @var{objdir} is a subdirectory
549 of @var{srcdir} is unsupported.
551 If you have previously built GCC in the same directory for a
552 different target machine, do @samp{make distclean} to delete all files
553 that might be invalid. One of the files this deletes is @file{Makefile};
554 if @samp{make distclean} complains that @file{Makefile} does not exist
555 or issues a message like ``don't know how to make distclean'' it probably
556 means that the directory is already suitably clean. However, with the
557 recommended method of building in a separate @var{objdir}, you should
558 simply use a different @var{objdir} for each target.
560 Second, when configuring a native system, either @command{cc} or
561 @command{gcc} must be in your path or you must set @env{CC} in
562 your environment before running configure. Otherwise the configuration
566 Note that the bootstrap compiler and the resulting GCC must be link
567 compatible, else the bootstrap will fail with linker errors about
568 incompatible object file formats. Several multilibed targets are
569 affected by this requirement, see
571 @ref{Specific, host/target specific installation notes}.
574 @uref{specific.html,,host/target specific installation notes}.
583 % @var{srcdir}/configure [@var{options}] [@var{target}]
587 @heading Target specification
590 GCC has code to correctly determine the correct value for @var{target}
591 for nearly all native systems. Therefore, we highly recommend you not
592 provide a configure target when configuring a native compiler.
595 @var{target} must be specified as @option{--target=@var{target}}
596 when configuring a cross compiler; examples of valid targets would be
597 m68k-coff, sh-elf, etc.
600 Specifying just @var{target} instead of @option{--target=@var{target}}
601 implies that the host defaults to @var{target}.
605 @heading Options specification
607 Use @var{options} to override several configure time options for
608 GCC@. A list of supported @var{options} follows; @samp{configure
609 --help} may list other options, but those not listed below may not
610 work and should not normally be used.
612 Note that each @option{--enable} option has a corresponding
613 @option{--disable} option and that each @option{--with} option has a
614 corresponding @option{--without} option.
617 @item --prefix=@var{dirname}
618 Specify the toplevel installation
619 directory. This is the recommended way to install the tools into a directory
620 other than the default. The toplevel installation directory defaults to
623 We @strong{highly} recommend against @var{dirname} being the same or a
624 subdirectory of @var{objdir} or vice versa. If specifying a directory
625 beneath a user's home directory tree, some shells will not expand
626 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
629 The following standard @command{autoconf} options are supported. Normally you
630 should not need to use these options.
632 @item --exec-prefix=@var{dirname}
633 Specify the toplevel installation directory for architecture-dependent
634 files. The default is @file{@var{prefix}}.
636 @item --bindir=@var{dirname}
637 Specify the installation directory for the executables called by users
638 (such as @command{gcc} and @command{g++}). The default is
639 @file{@var{exec-prefix}/bin}.
641 @item --libdir=@var{dirname}
642 Specify the installation directory for object code libraries and
643 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
645 @item --libexecdir=@var{dirname}
646 Specify the installation directory for internal executables of GCC@.
647 The default is @file{@var{exec-prefix}/libexec}.
649 @item --with-slibdir=@var{dirname}
650 Specify the installation directory for the shared libgcc library. The
651 default is @file{@var{libdir}}.
653 @item --infodir=@var{dirname}
654 Specify the installation directory for documentation in info format.
655 The default is @file{@var{prefix}/info}.
657 @item --datadir=@var{dirname}
658 Specify the installation directory for some architecture-independent
659 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
661 @item --mandir=@var{dirname}
662 Specify the installation directory for manual pages. The default is
663 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
664 the full GCC manuals, which are provided in Texinfo format. The manpages
665 are derived by an automatic conversion process from parts of the full
668 @item --with-gxx-include-dir=@var{dirname}
670 the installation directory for G++ header files. The default is
671 @file{@var{prefix}/include/c++/@var{version}}.
675 @item --program-prefix=@var{prefix}
676 GCC supports some transformations of the names of its programs when
677 installing them. This option prepends @var{prefix} to the names of
678 programs to install in @var{bindir} (see above). For example, specifying
679 @option{--program-prefix=foo-} would result in @samp{gcc}
680 being installed as @file{/usr/local/bin/foo-gcc}.
682 @item --program-suffix=@var{suffix}
683 Appends @var{suffix} to the names of programs to install in @var{bindir}
684 (see above). For example, specifying @option{--program-suffix=-3.1}
685 would result in @samp{gcc} being installed as
686 @file{/usr/local/bin/gcc-3.1}.
688 @item --program-transform-name=@var{pattern}
689 Applies the @samp{sed} script @var{pattern} to be applied to the names
690 of programs to install in @var{bindir} (see above). @var{pattern} has to
691 consist of one or more basic @samp{sed} editing commands, separated by
692 semicolons. For example, if you want the @samp{gcc} program name to be
693 transformed to the installed program @file{/usr/local/bin/myowngcc} and
694 the @samp{g++} program name to be transformed to
695 @file{/usr/local/bin/gspecial++} without changing other program names,
696 you could use the pattern
697 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
698 to achieve this effect.
700 All three options can be combined and used together, resulting in more
701 complex conversion patterns. As a basic rule, @var{prefix} (and
702 @var{suffix}) are prepended (appended) before further transformations
703 can happen with a special transformation script @var{pattern}.
705 As currently implemented, this option only takes effect for native
706 builds; cross compiler binaries' names are not transformed even when a
707 transformation is explicitly asked for by one of these options.
709 For native builds, some of the installed programs are also installed
710 with the target alias in front of their name, as in
711 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
712 before the target alias is prepended to the name---so, specifying
713 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
714 resulting binary would be installed as
715 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
717 As a last shortcoming, none of the installed Ada programs are
718 transformed yet, which will be fixed in some time.
720 @item --with-local-prefix=@var{dirname}
722 installation directory for local include files. The default is
723 @file{/usr/local}. Specify this option if you want the compiler to
724 search directory @file{@var{dirname}/include} for locally installed
725 header files @emph{instead} of @file{/usr/local/include}.
727 You should specify @option{--with-local-prefix} @strong{only} if your
728 site has a different convention (not @file{/usr/local}) for where to put
731 The default value for @option{--with-local-prefix} is @file{/usr/local}
732 regardless of the value of @option{--prefix}. Specifying
733 @option{--prefix} has no effect on which directory GCC searches for
734 local header files. This may seem counterintuitive, but actually it is
737 The purpose of @option{--prefix} is to specify where to @emph{install
738 GCC}. The local header files in @file{/usr/local/include}---if you put
739 any in that directory---are not part of GCC@. They are part of other
740 programs---perhaps many others. (GCC installs its own header files in
741 another directory which is based on the @option{--prefix} value.)
743 Both the local-prefix include directory and the GCC-prefix include
744 directory are part of GCC's ``system include'' directories. Although these
745 two directories are not fixed, they need to be searched in the proper
746 order for the correct processing of the include_next directive. The
747 local-prefix include directory is searched before the GCC-prefix
748 include directory. Another characteristic of system include directories
749 is that pedantic warnings are turned off for headers in these directories.
751 Some autoconf macros add @option{-I @var{directory}} options to the
752 compiler command line, to ensure that directories containing installed
753 packages' headers are searched. When @var{directory} is one of GCC's
754 system include directories, GCC will ignore the option so that system
755 directories continue to be processed in the correct order. This
756 may result in a search order different from what was specified but the
757 directory will still be searched.
759 GCC automatically searches for ordinary libraries using
760 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
761 used for both GCC and packages, GCC will automatically search for
762 both headers and libraries. This provides a configuration that is
763 easy to use. GCC behaves in a manner similar to that when it is
764 installed as a system compiler in @file{/usr}.
766 Sites that need to install multiple versions of GCC may not want to
767 use the above simple configuration. It is possible to use the
768 @option{--program-prefix}, @option{--program-suffix} and
769 @option{--program-transform-name} options to install multiple versions
770 into a single directory, but it may be simpler to use different prefixes
771 and the @option{--with-local-prefix} option to specify the location of the
772 site-specific files for each version. It will then be necessary for
773 users to specify explicitly the location of local site libraries
774 (e.g., with @env{LIBRARY_PATH}).
776 The same value can be used for both @option{--with-local-prefix} and
777 @option{--prefix} provided it is not @file{/usr}. This can be used
778 to avoid the default search of @file{/usr/local/include}.
780 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
781 The directory you use for @option{--with-local-prefix} @strong{must not}
782 contain any of the system's standard header files. If it did contain
783 them, certain programs would be miscompiled (including GNU Emacs, on
784 certain targets), because this would override and nullify the header
785 file corrections made by the @command{fixincludes} script.
787 Indications are that people who use this option use it based on mistaken
788 ideas of what it is for. People use it as if it specified where to
789 install part of GCC@. Perhaps they make this assumption because
790 installing GCC creates the directory.
792 @item --enable-shared[=@var{package}[,@dots{}]]
793 Build shared versions of libraries, if shared libraries are supported on
794 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
795 are enabled by default on all platforms that support shared libraries.
797 If a list of packages is given as an argument, build shared libraries
798 only for the listed packages. For other packages, only static libraries
799 will be built. Package names currently recognized in the GCC tree are
800 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
801 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
802 @samp{ada}, @samp{libada}, @samp{libjava} and @samp{libobjc}.
803 Note @samp{libiberty} does not support shared libraries at all.
805 Use @option{--disable-shared} to build only static libraries. Note that
806 @option{--disable-shared} does not accept a list of package names as
807 argument, only @option{--enable-shared} does.
809 @item @anchor{with-gnu-as}--with-gnu-as
810 Specify that the compiler should assume that the
811 assembler it finds is the GNU assembler. However, this does not modify
812 the rules to find an assembler and will result in confusion if the
813 assembler found is not actually the GNU assembler. (Confusion may also
814 result if the compiler finds the GNU assembler but has not been
815 configured with @option{--with-gnu-as}.) If you have more than one
816 assembler installed on your system, you may want to use this option in
817 connection with @option{--with-as=@var{pathname}} or
818 @option{--with-build-time-tools=@var{pathname}}.
820 The following systems are the only ones where it makes a difference
821 whether you use the GNU assembler. On any other system,
822 @option{--with-gnu-as} has no effect.
825 @item @samp{hppa1.0-@var{any}-@var{any}}
826 @item @samp{hppa1.1-@var{any}-@var{any}}
827 @item @samp{i386-@var{any}-sysv}
828 @item @samp{m68k-bull-sysv}
829 @item @samp{m68k-hp-hpux}
830 @item @samp{m68000-hp-hpux}
831 @item @samp{m68000-att-sysv}
832 @item @samp{sparc-sun-solaris2.@var{any}}
833 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
836 On the systems listed above (except for the HP-PA, the SPARC, for ISC on
837 the 386, if you use the GNU assembler, you should also use the GNU linker
838 (and specify @option{--with-gnu-ld}).
840 @item @anchor{with-as}--with-as=@var{pathname}
841 Specify that the compiler should use the assembler pointed to by
842 @var{pathname}, rather than the one found by the standard rules to find
843 an assembler, which are:
846 Unless GCC is being built with a cross compiler, check the
847 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
848 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
849 @var{exec-prefix} defaults to @var{prefix}, which
850 defaults to @file{/usr/local} unless overridden by the
851 @option{--prefix=@var{pathname}} switch described above. @var{target}
852 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
853 @var{version} denotes the GCC version, such as 3.0.
856 If the target system is the same that you are building on, check
857 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
861 Check in the @env{PATH} for a tool whose name is prefixed by the
862 target system triple.
865 Check in the @env{PATH} for a tool whose name is not prefixed by the
866 target system triple, if the host and target system triple are
867 the same (in other words, we use a host tool if it can be used for
871 You may want to use @option{--with-as} if no assembler
872 is installed in the directories listed above, or if you have multiple
873 assemblers installed and want to choose one that is not found by the
876 @item @anchor{with-gnu-ld}--with-gnu-ld
877 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
880 @item --with-ld=@var{pathname}
881 Same as @uref{#with-as,,@option{--with-as}}
885 Specify that stabs debugging
886 information should be used instead of whatever format the host normally
887 uses. Normally GCC uses the same debug format as the host system.
889 On MIPS based systems and on Alphas, you must specify whether you want
890 GCC to create the normal ECOFF debugging format, or to use BSD-style
891 stabs passed through the ECOFF symbol table. The normal ECOFF debug
892 format cannot fully handle languages other than C@. BSD stabs format can
893 handle other languages, but it only works with the GNU debugger GDB@.
895 Normally, GCC uses the ECOFF debugging format by default; if you
896 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
898 No matter which default you choose when you configure GCC, the user
899 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
900 the debug format for a particular compilation.
902 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
903 @option{--with-gas} is used. It selects use of stabs debugging
904 information embedded in COFF output. This kind of debugging information
905 supports C++ well; ordinary COFF debugging information does not.
907 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
908 selects use of stabs debugging information embedded in ELF output. The
909 C++ compiler currently (2.6.0) does not support the DWARF debugging
910 information normally used on 386 SVR4 platforms; stabs provide a
911 workable alternative. This requires gas and gdb, as the normal SVR4
912 tools can not generate or interpret stabs.
914 @item --disable-multilib
915 Specify that multiple target
916 libraries to support different target variants, calling
917 conventions, etc.@: should not be built. The default is to build a
918 predefined set of them.
920 Some targets provide finer-grained control over which multilibs are built
921 (e.g., @option{--disable-softfloat}):
927 fpu, 26bit, underscore, interwork, biendian, nofmult.
930 softfloat, m68881, m68000, m68020.
933 single-float, biendian, softfloat.
935 @item powerpc*-*-*, rs6000*-*-*
936 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
941 @item --enable-threads
942 Specify that the target
943 supports threads. This affects the Objective-C compiler and runtime
944 library, and exception handling for other languages like C++ and Java.
945 On some systems, this is the default.
947 In general, the best (and, in many cases, the only known) threading
948 model available will be configured for use. Beware that on some
949 systems, GCC has not been taught what threading models are generally
950 available for the system. In this case, @option{--enable-threads} is an
951 alias for @option{--enable-threads=single}.
953 @item --disable-threads
954 Specify that threading support should be disabled for the system.
955 This is an alias for @option{--enable-threads=single}.
957 @item --enable-threads=@var{lib}
959 @var{lib} is the thread support library. This affects the Objective-C
960 compiler and runtime library, and exception handling for other languages
961 like C++ and Java. The possibilities for @var{lib} are:
969 Ada tasking support. For non-Ada programs, this setting is equivalent
970 to @samp{single}. When used in conjunction with the Ada run time, it
971 causes GCC to use the same thread primitives as Ada uses. This option
972 is necessary when using both Ada and the back end exception handling,
973 which is the default for most Ada targets.
975 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
976 that the file needed to support this configuration, @file{gthr-mach.h}, is
977 missing and thus this setting will cause a known bootstrap failure.)
979 This is an alias for @samp{single}.
981 Generic POSIX/Unix98 thread support.
983 Generic POSIX/Unix95 thread support.
985 RTEMS thread support.
987 Disable thread support, should work for all platforms.
989 Sun Solaris 2 thread support.
991 VxWorks thread support.
993 Microsoft Win32 API thread support.
995 Novell Kernel Services thread support.
999 Specify that the target supports TLS (Thread Local Storage). Usually
1000 configure can correctly determine if TLS is supported. In cases where
1001 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1002 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1003 the assembler supports TLS but the C library does not, or if the
1004 assumptions made by the configure test are incorrect.
1007 Specify that the target does not support TLS.
1008 This is an alias for @option{--enable-tls=no}.
1010 @item --with-cpu=@var{cpu}
1011 Specify which cpu variant the compiler should generate code for by default.
1012 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1013 This option is only supported on some targets, including ARM, i386, M68k,
1014 PowerPC, and SPARC@.
1016 @item --with-schedule=@var{cpu}
1017 @itemx --with-arch=@var{cpu}
1018 @itemx --with-tune=@var{cpu}
1019 @itemx --with-abi=@var{abi}
1020 @itemx --with-fpu=@var{type}
1021 @itemx --with-float=@var{type}
1022 These configure options provide default values for the @option{-mschedule=},
1023 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1024 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1025 @option{--with-cpu}, which switches will be accepted and acceptable values
1026 of the arguments depend on the target.
1028 @item --with-mode=@var{mode}
1029 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1030 This option is only supported on ARM targets.
1032 @item --with-divide=@var{type}
1033 Specify how the compiler should generate code for checking for
1034 division by zero. This option is only supported on the MIPS target.
1035 The possibilities for @var{type} are:
1038 Division by zero checks use conditional traps (this is the default on
1039 systems that support conditional traps).
1041 Division by zero checks use the break instruction.
1044 @item --enable-__cxa_atexit
1045 Define if you want to use __cxa_atexit, rather than atexit, to
1046 register C++ destructors for local statics and global objects.
1047 This is essential for fully standards-compliant handling of
1048 destructors, but requires __cxa_atexit in libc. This option is currently
1049 only available on systems with GNU libc. When enabled, this will cause
1050 @option{-fuse-cxa-exit} to be passed by default.
1052 @item --enable-target-optspace
1054 libraries should be optimized for code space instead of code speed.
1055 This is the default for the m32r platform.
1058 Specify that a user visible @command{cpp} program should not be installed.
1060 @item --with-cpp-install-dir=@var{dirname}
1061 Specify that the user visible @command{cpp} program should be installed
1062 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1064 @item --enable-initfini-array
1065 Force the use of sections @code{.init_array} and @code{.fini_array}
1066 (instead of @code{.init} and @code{.fini}) for constructors and
1067 destructors. Option @option{--disable-initfini-array} has the
1068 opposite effect. If neither option is specified, the configure script
1069 will try to guess whether the @code{.init_array} and
1070 @code{.fini_array} sections are supported and, if they are, use them.
1072 @item --enable-maintainer-mode
1073 The build rules that
1074 regenerate the GCC master message catalog @file{gcc.pot} are normally
1075 disabled. This is because it can only be rebuilt if the complete source
1076 tree is present. If you have changed the sources and want to rebuild the
1077 catalog, configuring with @option{--enable-maintainer-mode} will enable
1078 this. Note that you need a recent version of the @code{gettext} tools
1081 @item --disable-bootstrap
1082 For a native build, the default configuration is to perform
1083 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1084 testing that GCC can compile itself correctly. If you want to disable
1085 this process, you can configure with @option{--disable-bootstrap}.
1087 @item --enable-bootstrap
1088 In special cases, you may want to perform a 3-stage build
1089 even if the target and host triplets are different.
1090 This could happen when the host can run code compiled for
1091 the target (e.g.@: host is i686-linux, target is i486-linux).
1092 Starting from GCC 4.2, to do this you have to configure explicitly
1093 with @option{--enable-bootstrap}.
1095 @item --enable-generated-files-in-srcdir
1096 Neither the .c and .h files that are generated from Bison and flex nor the
1097 info manuals and man pages that are built from the .texi files are present
1098 in the SVN development tree. When building GCC from that development tree,
1099 or from one of our snapshots, those generated files are placed in your
1100 build directory, which allows for the source to be in a readonly
1103 If you configure with @option{--enable-generated-files-in-srcdir} then those
1104 generated files will go into the source directory. This is mainly intended
1105 for generating release or prerelease tarballs of the GCC sources, since it
1106 is not a requirement that the users of source releases to have flex, Bison,
1109 @item --enable-version-specific-runtime-libs
1111 that runtime libraries should be installed in the compiler specific
1112 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1113 addition, @samp{libstdc++}'s include files will be installed into
1114 @file{@var{libdir}} unless you overruled it by using
1115 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1116 particularly useful if you intend to use several versions of GCC in
1117 parallel. This is currently supported by @samp{libgfortran},
1118 @samp{libjava}, @samp{libmudflap}, @samp{libstdc++}, and @samp{libobjc}.
1120 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1121 Specify that only a particular subset of compilers and
1122 their runtime libraries should be built. For a list of valid values for
1123 @var{langN} you can issue the following command in the
1124 @file{gcc} directory of your GCC source tree:@*
1126 grep language= */config-lang.in
1128 Currently, you can use any of the following:
1129 @code{all}, @code{ada}, @code{c}, @code{c++}, @code{fortran}, @code{java},
1130 @code{objc}, @code{obj-c++}, @code{treelang}.
1131 Building the Ada compiler has special requirements, see below.
1132 If you do not pass this flag, or specify the option @code{all}, then all
1133 default languages available in the @file{gcc} sub-tree will be configured.
1134 Ada, Objective-C++, and treelang are not default languages; the rest are.
1135 Re-defining @code{LANGUAGES} when calling @samp{make} @strong{does not}
1136 work anymore, as those language sub-directories might not have been
1139 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1140 Specify that a particular subset of compilers and their runtime
1141 libraries should be built with the system C compiler during stage 1 of
1142 the bootstrap process, rather than only in later stages with the
1143 bootstrapped C compiler. The list of valid values is the same as for
1144 @option{--enable-languages}, and the option @code{all} will select all
1145 of the languages enabled by @option{--enable-languages}. This option is
1146 primarily useful for GCC development; for instance, when a development
1147 version of the compiler cannot bootstrap due to compiler bugs, or when
1148 one is debugging front ends other than the C front end. When this
1149 option is used, one can then build the target libraries for the
1150 specified languages with the stage-1 compiler by using @command{make
1151 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1152 for the specified languages using @command{make stage1-start check-gcc}.
1154 @item --disable-libada
1155 Specify that the run-time libraries and tools used by GNAT should not
1156 be built. This can be useful for debugging, or for compatibility with
1157 previous Ada build procedures, when it was required to explicitly
1158 do a @samp{make -C gcc gnatlib_and_tools}.
1160 @item --disable-libssp
1161 Specify that the run-time libraries for stack smashing protection
1162 should not be built.
1164 @item --disable-libgomp
1165 Specify that the run-time libraries used by GOMP should not be built.
1168 Specify that the compiler should
1169 use DWARF 2 debugging information as the default.
1171 @item --enable-targets=all
1172 @itemx --enable-targets=@var{target_list}
1173 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1174 These are compilers that are able to generate either 64-bit or 32-bit
1175 code. Typically, the corresponding 32-bit target, e.g.@:
1176 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1177 option enables the 32-bit target to be a bi-arch compiler, which is
1178 useful when you want a bi-arch compiler that defaults to 32-bit, and
1179 you are building a bi-arch or multi-arch binutils in a combined tree.
1180 Currently, this option only affects powerpc-linux and x86-linux.
1182 @item --enable-secureplt
1183 This option enables @option{-msecure-plt} by default for powerpc-linux.
1185 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1186 Using the GNU Compiler Collection (GCC)},
1189 See ``RS/6000 and PowerPC Options'' in the main manual
1192 @item --enable-win32-registry
1193 @itemx --enable-win32-registry=@var{key}
1194 @itemx --disable-win32-registry
1195 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1196 to look up installations paths in the registry using the following key:
1199 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1202 @var{key} defaults to GCC version number, and can be overridden by the
1203 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1204 who use custom installers are encouraged to provide a different key,
1205 perhaps one comprised of vendor name and GCC version number, to
1206 avoid conflict with existing installations. This feature is enabled
1207 by default, and can be disabled by @option{--disable-win32-registry}
1208 option. This option has no effect on the other hosts.
1211 Specify that the machine does not have a floating point unit. This
1212 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1213 system, @option{--nfp} has no effect.
1215 @item --enable-werror
1216 @itemx --disable-werror
1217 @itemx --enable-werror=yes
1218 @itemx --enable-werror=no
1219 When you specify this option, it controls whether certain files in the
1220 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1221 If you don't specify it, @option{-Werror} is turned on for the main
1222 development trunk. However it defaults to off for release branches and
1223 final releases. The specific files which get @option{-Werror} are
1224 controlled by the Makefiles.
1226 @item --enable-checking
1227 @itemx --enable-checking=@var{list}
1228 When you specify this option, the compiler is built to perform internal
1229 consistency checks of the requested complexity. This does not change the
1230 generated code, but adds error checking within the compiler. This will
1231 slow down the compiler and may only work properly if you are building
1232 the compiler with GCC@. This is @samp{yes} by default when building
1233 from SVN or snapshots, but @samp{release} for releases. More control
1234 over the checks may be had by specifying @var{list}. The categories of
1235 checks available are @samp{yes} (most common checks
1236 @samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at
1237 all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest
1238 checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1239 Individual checks can be enabled with these flags @samp{assert},
1240 @samp{fold}, @samp{gc}, @samp{gcac} @samp{misc}, @samp{rtl},
1241 @samp{rtlflag}, @samp{runtime}, @samp{tree}, and @samp{valgrind}.
1243 The @samp{valgrind} check requires the external @command{valgrind}
1244 simulator, available from @uref{http://valgrind.org/}. The
1245 @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive.
1246 To disable all checking, @samp{--disable-checking} or
1247 @samp{--enable-checking=none} must be explicitly requested. Disabling
1248 assertions will make the compiler and runtime slightly faster but
1249 increase the risk of undetected internal errors causing wrong code to be
1252 @item --enable-coverage
1253 @itemx --enable-coverage=@var{level}
1254 With this option, the compiler is built to collect self coverage
1255 information, every time it is run. This is for internal development
1256 purposes, and only works when the compiler is being built with gcc. The
1257 @var{level} argument controls whether the compiler is built optimized or
1258 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1259 want to disable optimization, for performance analysis you want to
1260 enable optimization. When coverage is enabled, the default level is
1261 without optimization.
1263 @item --enable-gather-detailed-mem-stats
1264 When this option is specified more detailed information on memory
1265 allocation is gathered. This information is printed when using
1266 @option{-fmem-report}.
1269 @itemx --with-gc=@var{choice}
1270 With this option you can specify the garbage collector implementation
1271 used during the compilation process. @var{choice} can be one of
1272 @samp{page} and @samp{zone}, where @samp{page} is the default.
1275 @itemx --disable-nls
1276 The @option{--enable-nls} option enables Native Language Support (NLS),
1277 which lets GCC output diagnostics in languages other than American
1278 English. Native Language Support is enabled by default if not doing a
1279 canadian cross build. The @option{--disable-nls} option disables NLS@.
1281 @item --with-included-gettext
1282 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1283 procedure to prefer its copy of GNU @command{gettext}.
1285 @item --with-catgets
1286 If NLS is enabled, and if the host lacks @code{gettext} but has the
1287 inferior @code{catgets} interface, the GCC build procedure normally
1288 ignores @code{catgets} and instead uses GCC's copy of the GNU
1289 @code{gettext} library. The @option{--with-catgets} option causes the
1290 build procedure to use the host's @code{catgets} in this situation.
1292 @item --with-libiconv-prefix=@var{dir}
1293 Search for libiconv header files in @file{@var{dir}/include} and
1294 libiconv library files in @file{@var{dir}/lib}.
1296 @item --enable-obsolete
1297 Enable configuration for an obsoleted system. If you attempt to
1298 configure GCC for a system (build, host, or target) which has been
1299 obsoleted, and you do not specify this flag, configure will halt with an
1302 All support for systems which have been obsoleted in one release of GCC
1303 is removed entirely in the next major release, unless someone steps
1304 forward to maintain the port.
1306 @item --enable-decimal-float
1307 @itemx --enable-decimal-float=yes
1308 @itemx --enable-decimal-float=no
1309 @itemx --enable-decimal-float=bid
1310 @itemx --enable-decimal-float=dpd
1311 @itemx --disable-decimal-float
1312 Enable (or disable) support for the C decimal floating point extension
1313 that is in the IEEE 754R extension to the IEEE754 floating point
1314 standard. This is enabled by default only on PowerPC, i386, and
1315 x86_64 GNU/Linux systems. Other systems may also support it, but
1316 require the user to specifically enable it. You can optionally
1317 control which decimal floating point format is used (either @samp{bid}
1318 or @samp{dpd}). The @samp{bid} (binary integer decimal) format is
1319 default on i386 and x86_64 systems, and the @samp{dpd} (densely packed
1320 decimal) format is default on PowerPC systems.
1322 @item --with-long-double-128
1323 Specify if @code{long double} type should be 128-bit by default on selected
1324 GNU/Linux architectures. If using @code{--without-long-double-128},
1325 @code{long double} will be by default 64-bit, the same as @code{double} type.
1326 When neither of these configure options are used, the default will be
1327 128-bit @code{long double} when built against GNU C Library 2.4 and later,
1328 64-bit @code{long double} otherwise.
1330 @item --with-gmp=@var{pathname}
1331 @itemx --with-gmp-include=@var{pathname}
1332 @itemx --with-gmp-lib=@var{pathname}
1333 @itemx --with-mpfr=@var{pathname}
1334 @itemx --with-mpfr-include=@var{pathname}
1335 @itemx --with-mpfr-lib=@var{pathname}
1336 If you do not have GMP (the GNU Multiple Precision library) and the
1337 MPFR Libraries installed in a standard location and you want to build
1338 GCC, you can explicitly specify the directory where they are installed
1339 (@samp{--with-gmp=@var{gmpinstalldir}},
1340 @samp{--with-mpfr=@var{mpfrinstalldir}}). The
1341 @option{--with-gmp=@var{gmpinstalldir}} option is shorthand for
1342 @option{--with-gmp-lib=@var{gmpinstalldir}/lib} and
1343 @option{--with-gmp-include=@var{gmpinstalldir}/include}. Likewise the
1344 @option{--with-mpfr=@var{mpfrinstalldir}} option is shorthand for
1345 @option{--with-mpfr-lib=@var{mpfrinstalldir}/lib} and
1346 @option{--with-mpfr-include=@var{mpfrinstalldir}/include}. If these
1347 shorthand assumptions are not correct, you can use the explicit
1348 include and lib options directly.
1352 @subheading Cross-Compiler-Specific Options
1353 The following options only apply to building cross compilers.
1355 @item --with-sysroot
1356 @itemx --with-sysroot=@var{dir}
1357 Tells GCC to consider @var{dir} as the root of a tree that contains a
1358 (subset of) the root filesystem of the target operating system.
1359 Target system headers, libraries and run-time object files will be
1360 searched in there. The specified directory is not copied into the
1361 install tree, unlike the options @option{--with-headers} and
1362 @option{--with-libs} that this option obsoletes. The default value,
1363 in case @option{--with-sysroot} is not given an argument, is
1364 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1365 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1366 the GCC binaries if the installation tree is moved.
1368 @item --with-build-sysroot
1369 @itemx --with-build-sysroot=@var{dir}
1370 Tells GCC to consider @var{dir} as the system root (see
1371 @option{--with-sysroot}) while building target libraries, instead of
1372 the directory specified with @option{--with-sysroot}. This option is
1373 only useful when you are already using @option{--with-sysroot}. You
1374 can use @option{--with-build-sysroot} when you are configuring with
1375 @option{--prefix} set to a directory that is different from the one in
1376 which you are installing GCC and your target libraries.
1378 This option affects the system root for the compiler used to build
1379 target libraries (which runs on the build system); it does not affect
1380 the compiler which is used to build GCC itself.
1382 @item --with-headers
1383 @itemx --with-headers=@var{dir}
1384 Deprecated in favor of @option{--with-sysroot}.
1385 Specifies that target headers are available when building a cross compiler.
1386 The @var{dir} argument specifies a directory which has the target include
1387 files. These include files will be copied into the @file{gcc} install
1388 directory. @emph{This option with the @var{dir} argument is required} when
1389 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1390 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1391 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1392 will be run on these files to make them compatible with GCC@.
1394 @item --without-headers
1395 Tells GCC not use any target headers from a libc when building a cross
1396 compiler. When crossing to GNU/Linux, you need the headers so GCC
1397 can build the exception handling for libgcc.
1400 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1401 Deprecated in favor of @option{--with-sysroot}.
1402 Specifies a list of directories which contain the target runtime
1403 libraries. These libraries will be copied into the @file{gcc} install
1404 directory. If the directory list is omitted, this option has no
1408 Specifies that @samp{newlib} is
1409 being used as the target C library. This causes @code{__eprintf} to be
1410 omitted from @file{libgcc.a} on the assumption that it will be provided by
1413 @item --with-build-time-tools=@var{dir}
1414 Specifies where to find the set of target tools (assembler, linker, etc.)
1415 that will be used while building GCC itself. This option can be useful
1416 if the directory layouts are different between the system you are building
1417 GCC on, and the system where you will deploy it.
1419 For example, on a @option{ia64-hp-hpux} system, you may have the GNU
1420 assembler and linker in @file{/usr/bin}, and the native tools in a
1421 different path, and build a toolchain that expects to find the
1422 native tools in @file{/usr/bin}.
1424 When you use this option, you should ensure that @var{dir} includes
1425 @command{ar}, @command{as}, @command{ld}, @command{nm},
1426 @command{ranlib} and @command{strip} if necessary, and possibly
1427 @command{objdump}. Otherwise, GCC may use an inconsistent set of
1431 @subheading Java-Specific Options
1433 The following option applies to the build of the Java front end.
1436 @item --disable-libgcj
1437 Specify that the run-time libraries
1438 used by GCJ should not be built. This is useful in case you intend
1439 to use GCJ with some other run-time, or you're going to install it
1440 separately, or it just happens not to build on your particular
1441 machine. In general, if the Java front end is enabled, the GCJ
1442 libraries will be enabled too, unless they're known to not work on
1443 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1444 may need to port it; in this case, before modifying the top-level
1445 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1446 you may use @option{--enable-libgcj} to override the default.
1450 The following options apply to building @samp{libgcj}.
1452 @subsubheading General Options
1455 @item --enable-java-maintainer-mode
1456 By default the @samp{libjava} build will not attempt to compile the
1457 @file{.java} source files to @file{.class}. Instead, it will use the
1458 @file{.class} files from the source tree. If you use this option you
1459 must have executables named @command{ecj1} and @command{gjavah} in your path
1460 for use by the build. You must use this option if you intend to
1461 modify any @file{.java} files in @file{libjava}.
1463 @item --with-java-home=@var{dirname}
1464 This @samp{libjava} option overrides the default value of the
1465 @samp{java.home} system property. It is also used to set
1466 @samp{sun.boot.class.path} to @file{@var{dirname}/lib/rt.jar}. By
1467 default @samp{java.home} is set to @file{@var{prefix}} and
1468 @samp{sun.boot.class.path} to
1469 @file{@var{datadir}/java/libgcj-@var{version}.jar}.
1471 @item --with-ecj-jar=@var{filename}
1472 This option can be used to specify the location of an external jar
1473 file containing the Eclipse Java compiler. A specially modified
1474 version of this compiler is used by @command{gcj} to parse
1475 @file{.java} source files. If this option is given, the
1476 @samp{libjava} build will create and install an @file{ecj1} executable
1477 which uses this jar file at runtime.
1479 If this option is not given, but an @file{ecj.jar} file is found in
1480 the topmost source tree at configure time, then the @samp{libgcj}
1481 build will create and install @file{ecj1}, and will also install the
1482 discovered @file{ecj.jar} into a suitable place in the install tree.
1484 If @file{ecj1} is not installed, then the user will have to supply one
1485 on his path in order for @command{gcj} to properly parse @file{.java}
1486 source files. A suitable jar is available from
1487 @uref{ftp://sourceware.org/pub/java/}.
1489 @item --disable-getenv-properties
1490 Don't set system properties from @env{GCJ_PROPERTIES}.
1492 @item --enable-hash-synchronization
1493 Use a global hash table for monitor locks. Ordinarily,
1494 @samp{libgcj}'s @samp{configure} script automatically makes
1495 the correct choice for this option for your platform. Only use
1496 this if you know you need the library to be configured differently.
1498 @item --enable-interpreter
1499 Enable the Java interpreter. The interpreter is automatically
1500 enabled by default on all platforms that support it. This option
1501 is really only useful if you want to disable the interpreter
1502 (using @option{--disable-interpreter}).
1504 @item --disable-java-net
1505 Disable java.net. This disables the native part of java.net only,
1506 using non-functional stubs for native method implementations.
1508 @item --disable-jvmpi
1509 Disable JVMPI support.
1512 Enable runtime eCos target support.
1514 @item --without-libffi
1515 Don't use @samp{libffi}. This will disable the interpreter and JNI
1516 support as well, as these require @samp{libffi} to work.
1518 @item --enable-libgcj-debug
1519 Enable runtime debugging code.
1521 @item --enable-libgcj-multifile
1522 If specified, causes all @file{.java} source files to be
1523 compiled into @file{.class} files in one invocation of
1524 @samp{gcj}. This can speed up build time, but is more
1525 resource-intensive. If this option is unspecified or
1526 disabled, @samp{gcj} is invoked once for each @file{.java}
1527 file to compile into a @file{.class} file.
1529 @item --with-libiconv-prefix=DIR
1530 Search for libiconv in @file{DIR/include} and @file{DIR/lib}.
1532 @item --enable-sjlj-exceptions
1533 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1534 @samp{configure} ordinarily picks the correct value based on the platform.
1535 Only use this option if you are sure you need a different setting.
1537 @item --with-system-zlib
1538 Use installed @samp{zlib} rather than that included with GCC@.
1540 @item --with-win32-nlsapi=ansi, unicows or unicode
1541 Indicates how MinGW @samp{libgcj} translates between UNICODE
1542 characters and the Win32 API@.
1545 Use the single-byte @code{char} and the Win32 A functions natively,
1546 translating to and from UNICODE when using these functions. If
1547 unspecified, this is the default.
1550 Use the @code{WCHAR} and Win32 W functions natively. Adds
1551 @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}.
1552 @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines
1553 running built executables. @file{libunicows.a}, an open-source
1554 import library around Microsoft's @code{unicows.dll}, is obtained from
1555 @uref{http://libunicows.sourceforge.net/}, which also gives details
1556 on getting @file{unicows.dll} from Microsoft.
1559 Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not}
1560 add @code{-lunicows} to @file{libgcj.spec}. The built executables will
1561 only run on Microsoft Windows NT and above.
1565 @subsubheading AWT-Specific Options
1569 Use the X Window System.
1571 @item --enable-java-awt=PEER(S)
1572 Specifies the AWT peer library or libraries to build alongside
1573 @samp{libgcj}. If this option is unspecified or disabled, AWT
1574 will be non-functional. Current valid values are @option{gtk} and
1575 @option{xlib}. Multiple libraries should be separated by a
1576 comma (i.e.@: @option{--enable-java-awt=gtk,xlib}).
1578 @item --enable-gtk-cairo
1579 Build the cairo Graphics2D implementation on GTK@.
1581 @item --enable-java-gc=TYPE
1582 Choose garbage collector. Defaults to @option{boehm} if unspecified.
1584 @item --disable-gtktest
1585 Do not try to compile and run a test GTK+ program.
1587 @item --disable-glibtest
1588 Do not try to compile and run a test GLIB program.
1590 @item --with-libart-prefix=PFX
1591 Prefix where libart is installed (optional).
1593 @item --with-libart-exec-prefix=PFX
1594 Exec prefix where libart is installed (optional).
1596 @item --disable-libarttest
1597 Do not try to compile and run a test libart program.
1606 @uref{./index.html,,Return to the GCC Installation page}
1610 @c ***Building****************************************************************
1612 @comment node-name, next, previous, up
1613 @node Building, Testing, Configuration, Installing GCC
1619 @cindex Installing GCC: Building
1621 Now that GCC is configured, you are ready to build the compiler and
1624 Some commands executed when making the compiler may fail (return a
1625 nonzero status) and be ignored by @command{make}. These failures, which
1626 are often due to files that were not found, are expected, and can safely
1629 It is normal to have compiler warnings when compiling certain files.
1630 Unless you are a GCC developer, you can generally ignore these warnings
1631 unless they cause compilation to fail. Developers should attempt to fix
1632 any warnings encountered, however they can temporarily continue past
1633 warnings-as-errors by specifying the configure flag
1634 @option{--disable-werror}.
1636 On certain old systems, defining certain environment variables such as
1637 @env{CC} can interfere with the functioning of @command{make}.
1639 If you encounter seemingly strange errors when trying to build the
1640 compiler in a directory other than the source directory, it could be
1641 because you have previously configured the compiler in the source
1642 directory. Make sure you have done all the necessary preparations.
1644 If you build GCC on a BSD system using a directory stored in an old System
1645 V file system, problems may occur in running @command{fixincludes} if the
1646 System V file system doesn't support symbolic links. These problems
1647 result in a failure to fix the declaration of @code{size_t} in
1648 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1649 that type mismatches occur, this could be the cause.
1651 The solution is not to use such a directory for building GCC@.
1653 When building from SVN or snapshots and enabling the @code{treelang}
1654 front end, or if you modify @file{*.y} files, you need the Bison parser
1655 generator installed. If you do not modify @file{*.y} files, releases
1656 contain the Bison-generated files and you do not need Bison installed
1657 to build them. Note that most front ends now use hand-written parsers,
1658 which can be modified with no need for Bison.
1660 Similarly, when building from SVN or snapshots, or if you modify
1661 @file{*.l} files, you need the Flex lexical analyzer generator installed.
1662 There is still one Flex-based lexical analyzer (part of the build
1663 machinery, not of GCC itself) that is used even if you only build the
1666 When building from SVN or snapshots, or if you modify Texinfo
1667 documentation, you need version 4.4 or later of Texinfo installed if you
1668 want Info documentation to be regenerated. Releases contain Info
1669 documentation pre-built for the unmodified documentation in the release.
1671 @section Building a native compiler
1673 For a native build, the default configuration is to perform
1674 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
1675 This will build the entire GCC system and ensure that it compiles
1676 itself correctly. It can be disabled with the @option{--disable-bootstrap}
1677 parameter to @samp{configure}, but bootstrapping is suggested because
1678 the compiler will be tested more completely and could also have
1681 The bootstrapping process will complete the following steps:
1685 Build tools necessary to build the compiler.
1688 Perform a 3-stage bootstrap of the compiler. This includes building
1689 three times the target tools for use by the compiler such as binutils
1690 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
1691 individually linked or moved into the top level GCC source tree before
1695 Perform a comparison test of the stage2 and stage3 compilers.
1698 Build runtime libraries using the stage3 compiler from the previous step.
1702 If you are short on disk space you might consider @samp{make
1703 bootstrap-lean} instead. The sequence of compilation is the
1704 same described above, but object files from the stage1 and
1705 stage2 of the 3-stage bootstrap of the compiler are deleted as
1706 soon as they are no longer needed.
1708 If you want to save additional space during the bootstrap and in
1709 the final installation as well, you can build the compiler binaries
1710 without debugging information as in the following example. This will save
1711 roughly 40% of disk space both for the bootstrap and the final installation.
1712 (Libraries will still contain debugging information.)
1715 make CFLAGS='-O' LIBCFLAGS='-g -O2' \
1716 LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap
1719 If you wish to use non-default GCC flags when compiling the stage2 and
1720 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
1721 @samp{make}. Non-default optimization flags are less well
1722 tested here than the default of @samp{-g -O2}, but should still work.
1723 In a few cases, you may find that you need to specify special flags such
1724 as @option{-msoft-float} here to complete the bootstrap; or, if the
1725 native compiler miscompiles the stage1 compiler, you may need to work
1726 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
1727 stage1 compiler that were miscompiled, or by using @samp{make
1728 bootstrap4} to increase the number of stages of bootstrap.
1730 Note that using non-standard @code{CFLAGS} can cause bootstrap to fail
1731 if these trigger a warning with the new compiler. For example using
1732 @samp{-O2 -g -mcpu=i686} on @code{i686-pc-linux-gnu} will cause bootstrap
1733 failure as @option{-mcpu=} is deprecated in 3.4.0 and above.
1736 If you used the flag @option{--enable-languages=@dots{}} to restrict
1737 the compilers to be built, only those you've actually enabled will be
1738 built. This will of course only build those runtime libraries, for
1739 which the particular compiler has been built. Please note,
1740 that re-defining @env{LANGUAGES} when calling @samp{make}
1741 @strong{does not} work anymore!
1743 If the comparison of stage2 and stage3 fails, this normally indicates
1744 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1745 a potentially serious bug which you should investigate and report. (On
1746 a few systems, meaningful comparison of object files is impossible; they
1747 always appear ``different''. If you encounter this problem, you will
1748 need to disable comparison in the @file{Makefile}.)
1750 If you do not want to bootstrap your compiler, you can configure with
1751 @option{--disable-bootstrap}. In particular cases, you may want to
1752 bootstrap your compiler even if the target system is not the same as
1753 the one you are building on: for example, you could build a
1754 @code{powerpc-unknown-linux-gnu} toolchain on a
1755 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
1756 @option{--enable-bootstrap} to the configure script.
1759 @section Building a cross compiler
1761 When building a cross compiler, it is not generally possible to do a
1762 3-stage bootstrap of the compiler. This makes for an interesting problem
1763 as parts of GCC can only be built with GCC@.
1765 To build a cross compiler, we first recommend building and installing a
1766 native compiler. You can then use the native GCC compiler to build the
1767 cross compiler. The installed native compiler needs to be GCC version
1770 Assuming you have already installed a native copy of GCC and configured
1771 your cross compiler, issue the command @command{make}, which performs the
1776 Build host tools necessary to build the compiler.
1779 Build target tools for use by the compiler such as binutils (bfd,
1780 binutils, gas, gprof, ld, and opcodes)
1781 if they have been individually linked or moved into the top level GCC source
1782 tree before configuring.
1785 Build the compiler (single stage only).
1788 Build runtime libraries using the compiler from the previous step.
1791 Note that if an error occurs in any step the make process will exit.
1793 If you are not building GNU binutils in the same source tree as GCC,
1794 you will need a cross-assembler and cross-linker installed before
1795 configuring GCC@. Put them in the directory
1796 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
1797 you should put in this directory:
1801 This should be the cross-assembler.
1804 This should be the cross-linker.
1807 This should be the cross-archiver: a program which can manipulate
1808 archive files (linker libraries) in the target machine's format.
1811 This should be a program to construct a symbol table in an archive file.
1814 The installation of GCC will find these programs in that directory,
1815 and copy or link them to the proper place to for the cross-compiler to
1816 find them when run later.
1818 The easiest way to provide these files is to build the Binutils package.
1819 Configure it with the same @option{--host} and @option{--target}
1820 options that you use for configuring GCC, then build and install
1821 them. They install their executables automatically into the proper
1822 directory. Alas, they do not support all the targets that GCC
1825 If you are not building a C library in the same source tree as GCC,
1826 you should also provide the target libraries and headers before
1827 configuring GCC, specifying the directories with
1828 @option{--with-sysroot} or @option{--with-headers} and
1829 @option{--with-libs}. Many targets also require ``start files'' such
1830 as @file{crt0.o} and
1831 @file{crtn.o} which are linked into each executable. There may be several
1832 alternatives for @file{crt0.o}, for use with profiling or other
1833 compilation options. Check your target's definition of
1834 @code{STARTFILE_SPEC} to find out what start files it uses.
1836 @section Building in parallel
1838 GNU Make 3.79 and above, which is necessary to build GCC, support
1839 building in parallel. To activate this, you can use @samp{make -j 2}
1840 instead of @samp{make}. You can also specify a bigger number, and
1841 in most cases using a value greater than the number of processors in
1842 your machine will result in fewer and shorter I/O latency hits, thus
1843 improving overall throughput; this is especially true for slow drives
1844 and network filesystems.
1846 @section Building the Ada compiler
1848 In order to build GNAT, the Ada compiler, you need a working GNAT
1849 compiler (GNAT version 3.14 or later, or GCC version 3.1 or later).
1850 This includes GNAT tools such as @command{gnatmake} and
1851 @command{gnatlink}, since the Ada front end is written in Ada and
1852 uses some GNAT-specific extensions.
1854 In order to build a cross compiler, it is suggested to install
1855 the new compiler as native first, and then use it to build the cross
1858 @command{configure} does not test whether the GNAT installation works
1859 and has a sufficiently recent version; if too old a GNAT version is
1860 installed, the build will fail unless @option{--enable-languages} is
1861 used to disable building the Ada front end.
1863 @section Building with profile feedback
1865 It is possible to use profile feedback to optimize the compiler itself. This
1866 should result in a faster compiler binary. Experiments done on x86 using gcc
1867 3.3 showed approximately 7 percent speedup on compiling C programs. To
1868 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
1870 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
1871 compiler. This compiler is used to build a @code{stageprofile} compiler
1872 instrumented to collect execution counts of instruction and branch
1873 probabilities. Then runtime libraries are compiled with profile collected.
1874 Finally a @code{stagefeedback} compiler is built using the information collected.
1876 Unlike standard bootstrap, several additional restrictions apply. The
1877 compiler used to build @code{stage1} needs to support a 64-bit integral type.
1878 It is recommended to only use GCC for this. Also parallel make is currently
1879 not supported since collisions in profile collecting may occur.
1886 @uref{./index.html,,Return to the GCC Installation page}
1890 @c ***Testing*****************************************************************
1892 @comment node-name, next, previous, up
1893 @node Testing, Final install, Building, Installing GCC
1897 @chapter Installing GCC: Testing
1900 @cindex Installing GCC: Testing
1903 Before you install GCC, we encourage you to run the testsuites and to
1904 compare your results with results from a similar configuration that have
1905 been submitted to the
1906 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
1907 Some of these archived results are linked from the build status lists
1908 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
1909 reports a successful build runs the testsuites and submits the results.
1910 This step is optional and may require you to download additional software,
1911 but it can give you confidence in your new GCC installation or point out
1912 problems before you install and start using your new GCC@.
1914 First, you must have @uref{download.html,,downloaded the testsuites}.
1915 These are part of the full distribution, but if you downloaded the
1916 ``core'' compiler plus any front ends, you must download the testsuites
1919 Second, you must have the testing tools installed. This includes
1920 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
1921 the DejaGnu site has links to these.
1923 If the directories where @command{runtest} and @command{expect} were
1924 installed are not in the @env{PATH}, you may need to set the following
1925 environment variables appropriately, as in the following example (which
1926 assumes that DejaGnu has been installed under @file{/usr/local}):
1929 TCL_LIBRARY = /usr/local/share/tcl8.0
1930 DEJAGNULIBS = /usr/local/share/dejagnu
1933 (On systems such as Cygwin, these paths are required to be actual
1934 paths, not mounts or links; presumably this is due to some lack of
1935 portability in the DejaGnu code.)
1938 Finally, you can run the testsuite (which may take a long time):
1940 cd @var{objdir}; make -k check
1943 This will test various components of GCC, such as compiler
1944 front ends and runtime libraries. While running the testsuite, DejaGnu
1945 might emit some harmless messages resembling
1946 @samp{WARNING: Couldn't find the global config file.} or
1947 @samp{WARNING: Couldn't find tool init file} that can be ignored.
1949 @section How can you run the testsuite on selected tests?
1951 In order to run sets of tests selectively, there are targets
1952 @samp{make check-gcc} and @samp{make check-g++}
1953 in the @file{gcc} subdirectory of the object directory. You can also
1954 just run @samp{make check} in a subdirectory of the object directory.
1957 A more selective way to just run all @command{gcc} execute tests in the
1961 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
1964 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
1965 the testsuite with filenames matching @samp{9805*}, you would use
1968 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1971 The @file{*.exp} files are located in the testsuite directories of the GCC
1972 source, the most important ones being @file{compile.exp},
1973 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1974 To get a list of the possible @file{*.exp} files, pipe the
1975 output of @samp{make check} into a file and look at the
1976 @samp{Running @dots{} .exp} lines.
1978 @section Passing options and running multiple testsuites
1980 You can pass multiple options to the testsuite using the
1981 @samp{--target_board} option of DejaGNU, either passed as part of
1982 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
1983 work outside the makefiles. For example,
1986 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
1989 will run the standard @command{g++} testsuites (``unix'' is the target name
1990 for a standard native testsuite situation), passing
1991 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
1992 slashes separate options.
1994 You can run the testsuites multiple times using combinations of options
1995 with a syntax similar to the brace expansion of popular shells:
1998 @dots{}"--target_board=arm-sim/@{-mhard-float,-msoft-float@}@{-O1,-O2,-O3,@}"
2001 (Note the empty option caused by the trailing comma in the final group.)
2002 The following will run each testsuite eight times using the @samp{arm-sim}
2003 target, as if you had specified all possible combinations yourself:
2006 --target_board=arm-sim/-mhard-float/-O1
2007 --target_board=arm-sim/-mhard-float/-O2
2008 --target_board=arm-sim/-mhard-float/-O3
2009 --target_board=arm-sim/-mhard-float
2010 --target_board=arm-sim/-msoft-float/-O1
2011 --target_board=arm-sim/-msoft-float/-O2
2012 --target_board=arm-sim/-msoft-float/-O3
2013 --target_board=arm-sim/-msoft-float
2016 They can be combined as many times as you wish, in arbitrary ways. This
2020 @dots{}"--target_board=unix/-Wextra@{-O3,-fno-strength-reduce@}@{-fomit-frame-pointer,@}"
2023 will generate four combinations, all involving @samp{-Wextra}.
2025 The disadvantage to this method is that the testsuites are run in serial,
2026 which is a waste on multiprocessor systems. For users with GNU Make and
2027 a shell which performs brace expansion, you can run the testsuites in
2028 parallel by having the shell perform the combinations and @command{make}
2029 do the parallel runs. Instead of using @samp{--target_board}, use a
2030 special makefile target:
2033 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
2039 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
2042 will run three concurrent ``make-gcc'' testsuites, eventually testing all
2043 ten combinations as described above. Note that this is currently only
2044 supported in the @file{gcc} subdirectory. (To see how this works, try
2045 typing @command{echo} before the example given here.)
2048 @section Additional testing for Java Class Libraries
2050 The Java runtime tests can be executed via @samp{make check}
2051 in the @file{@var{target}/libjava/testsuite} directory in
2054 The @uref{http://sourceware.org/mauve/,,Mauve Project} provides
2055 a suite of tests for the Java Class Libraries. This suite can be run
2056 as part of libgcj testing by placing the Mauve tree within the libjava
2057 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
2058 specifying the location of that tree when invoking @samp{make}, as in
2059 @samp{make MAUVEDIR=~/mauve check}.
2061 @section How to interpret test results
2063 The result of running the testsuite are various @file{*.sum} and @file{*.log}
2064 files in the testsuite subdirectories. The @file{*.log} files contain a
2065 detailed log of the compiler invocations and the corresponding
2066 results, the @file{*.sum} files summarize the results. These summaries
2067 contain status codes for all tests:
2071 PASS: the test passed as expected
2073 XPASS: the test unexpectedly passed
2075 FAIL: the test unexpectedly failed
2077 XFAIL: the test failed as expected
2079 UNSUPPORTED: the test is not supported on this platform
2081 ERROR: the testsuite detected an error
2083 WARNING: the testsuite detected a possible problem
2086 It is normal for some tests to report unexpected failures. At the
2087 current time the testing harness does not allow fine grained control
2088 over whether or not a test is expected to fail. This problem should
2089 be fixed in future releases.
2092 @section Submitting test results
2094 If you want to report the results to the GCC project, use the
2095 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
2098 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
2099 -m gcc-testresults@@gcc.gnu.org |sh
2102 This script uses the @command{Mail} program to send the results, so
2103 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
2104 prepended to the testsuite summary and should contain any special
2105 remarks you have on your results or your build environment. Please
2106 do not edit the testsuite result block or the subject line, as these
2107 messages may be automatically processed.
2114 @uref{./index.html,,Return to the GCC Installation page}
2118 @c ***Final install***********************************************************
2120 @comment node-name, next, previous, up
2121 @node Final install, , Testing, Installing GCC
2123 @ifset finalinstallhtml
2125 @chapter Installing GCC: Final installation
2128 Now that GCC has been built (and optionally tested), you can install it with
2130 cd @var{objdir}; make install
2133 We strongly recommend to install into a target directory where there is
2134 no previous version of GCC present. Also, the GNAT runtime should not
2135 be stripped, as this would break certain features of the debugger that
2136 depend on this debugging information (catching Ada exceptions for
2139 That step completes the installation of GCC; user level binaries can
2140 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
2141 you specified with the @option{--prefix} to configure (or
2142 @file{/usr/local} by default). (If you specified @option{--bindir},
2143 that directory will be used instead; otherwise, if you specified
2144 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
2145 Headers for the C++ and Java libraries are installed in
2146 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
2147 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
2148 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
2149 in info format in @file{@var{infodir}} (normally
2150 @file{@var{prefix}/info}).
2152 When installing cross-compilers, GCC's executables
2153 are not only installed into @file{@var{bindir}}, that
2154 is, @file{@var{exec-prefix}/bin}, but additionally into
2155 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
2156 exists. Typically, such @dfn{tooldirs} hold target-specific
2157 binutils, including assembler and linker.
2159 Installation into a temporary staging area or into a @command{chroot}
2160 jail can be achieved with the command
2163 make DESTDIR=@var{path-to-rootdir} install
2166 @noindent where @var{path-to-rootdir} is the absolute path of
2167 a directory relative to which all installation paths will be
2168 interpreted. Note that the directory specified by @code{DESTDIR}
2169 need not exist yet; it will be created if necessary.
2171 There is a subtle point with tooldirs and @code{DESTDIR}:
2172 If you relocate a cross-compiler installation with
2173 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
2174 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
2175 be filled with duplicated GCC executables only if it already exists,
2176 it will not be created otherwise. This is regarded as a feature,
2177 not as a bug, because it gives slightly more control to the packagers
2178 using the @code{DESTDIR} feature.
2180 If you are bootstrapping a released version of GCC then please
2181 quickly review the build status page for your release, available from
2182 @uref{http://gcc.gnu.org/buildstat.html}.
2183 If your system is not listed for the version of GCC that you built,
2185 @email{gcc@@gcc.gnu.org} indicating
2186 that you successfully built and installed GCC@.
2187 Include the following information:
2191 Output from running @file{@var{srcdir}/config.guess}. Do not send
2192 that file itself, just the one-line output from running it.
2195 The output of @samp{gcc -v} for your newly installed @command{gcc}.
2196 This tells us which version of GCC you built and the options you passed to
2200 Whether you enabled all languages or a subset of them. If you used a
2201 full distribution then this information is part of the configure
2202 options in the output of @samp{gcc -v}, but if you downloaded the
2203 ``core'' compiler plus additional front ends then it isn't apparent
2204 which ones you built unless you tell us about it.
2207 If the build was for GNU/Linux, also include:
2210 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
2211 this information should be available from @file{/etc/issue}.
2214 The version of the Linux kernel, available from @samp{uname --version}
2218 The version of glibc you used; for RPM-based systems like Red Hat,
2219 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
2220 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
2222 For other systems, you can include similar information if you think it is
2226 Any other information that you think would be useful to people building
2227 GCC on the same configuration. The new entry in the build status list
2228 will include a link to the archived copy of your message.
2231 We'd also like to know if the
2233 @ref{Specific, host/target specific installation notes}
2236 @uref{specific.html,,host/target specific installation notes}
2238 didn't include your host/target information or if that information is
2239 incomplete or out of date. Send a note to
2240 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
2242 If you find a bug, please report it following the
2243 @uref{../bugs.html,,bug reporting guidelines}.
2245 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
2246 dvi}. You will need to have @command{texi2dvi} (version at least 4.4)
2247 and @TeX{} installed. This creates a number of @file{.dvi} files in
2248 subdirectories of @file{@var{objdir}}; these may be converted for
2249 printing with programs such as @command{dvips}. Alternately, by using
2250 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
2251 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
2252 is included with Texinfo version 4.8 and later. You can also
2253 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
2254 Free Software Foundation}, though such manuals may not be for the most
2255 recent version of GCC@.
2257 If you would like to generate online HTML documentation, do @samp{cd
2258 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
2259 @file{@var{objdir}/gcc/HTML}.
2266 @uref{./index.html,,Return to the GCC Installation page}
2270 @c ***Binaries****************************************************************
2272 @comment node-name, next, previous, up
2273 @node Binaries, Specific, Installing GCC, Top
2277 @chapter Installing GCC: Binaries
2280 @cindex Installing GCC: Binaries
2282 We are often asked about pre-compiled versions of GCC@. While we cannot
2283 provide these for all platforms, below you'll find links to binaries for
2284 various platforms where creating them by yourself is not easy due to various
2287 Please note that we did not create these binaries, nor do we
2288 support them. If you have any problems installing them, please
2289 contact their makers.
2296 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
2299 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX}.
2303 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
2306 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
2307 Development Tools for the Renesas H8/300[HS] Series}.
2313 @uref{http://hpux.cs.utah.edu/,,HP-UX Porting Center};
2316 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
2320 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
2321 Development Tools for the Motorola 68HC11/68HC12}.
2324 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
2325 OpenServer/Unixware}.
2328 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
2331 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
2337 The @uref{http://sourceware.org/cygwin/,,Cygwin} project;
2339 The @uref{http://www.mingw.org/,,MinGW} project.
2343 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
2344 Written Word} offers binaries for
2347 Digital UNIX 4.0D and 5.1,
2349 HP-UX 10.20, 11.00, and 11.11, and
2350 Solaris/SPARC 2.5.1, 2.6, 7, 8, and 9.
2353 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
2354 number of platforms.
2357 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
2358 links to GNU Fortran binaries for several platforms.
2361 In addition to those specific offerings, you can get a binary
2362 distribution CD-ROM from the
2363 @uref{http://www.gnu.org/order/order.html,,Free Software Foundation}.
2364 It contains binaries for a number of platforms, and
2365 includes not only GCC, but other stuff as well. The current CD does
2366 not contain the latest version of GCC, but it should allow
2367 bootstrapping the compiler. An updated version of that disk is in the
2375 @uref{./index.html,,Return to the GCC Installation page}
2379 @c ***Specific****************************************************************
2381 @comment node-name, next, previous, up
2382 @node Specific, Old, Binaries, Top
2386 @chapter Host/target specific installation notes for GCC
2389 @cindex Specific installation notes
2390 @cindex Target specific installation
2391 @cindex Host specific installation
2392 @cindex Target specific installation notes
2394 Please read this document carefully @emph{before} installing the
2395 GNU Compiler Collection on your machine.
2397 Note that this list of install notes is @emph{not} a list of supported
2398 hosts or targets. Not all supported hosts and targets are listed
2399 here, only the ones that require host-specific or target-specific
2405 @uref{#alpha-x-x,,alpha*-*-*}
2407 @uref{#alpha-dec-osf,,alpha*-dec-osf*}
2409 @uref{#alphaev5-cray-unicosmk,,alphaev5-cray-unicosmk*}
2411 @uref{#arc-x-elf,,arc-*-elf}
2413 @uref{#arm-x-elf,,arm-*-elf}
2414 @uref{#arm-x-coff,,arm-*-coff}
2415 @uref{#arm-x-aout,,arm-*-aout}
2417 @uref{#xscale-x-x,,xscale-*-*}
2421 @uref{#bfin,,Blackfin}
2427 @uref{#x-x-freebsd,,*-*-freebsd*}
2429 @uref{#h8300-hms,,h8300-hms}
2431 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
2433 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
2435 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
2437 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
2439 @uref{#ix86-x-linuxaout,,i?86-*-linux*aout}
2441 @uref{#ix86-x-linux,,i?86-*-linux*}
2443 @uref{#ix86-x-sco32v5,,i?86-*-sco3.2v5*}
2445 @uref{#ix86-x-solaris210,,i?86-*-solaris2.10}
2447 @uref{#ix86-x-udk,,i?86-*-udk}
2449 @uref{#ia64-x-linux,,ia64-*-linux}
2451 @uref{#ia64-x-hpux,,ia64-*-hpux*}
2453 @uref{#x-ibm-aix,,*-ibm-aix*}
2455 @uref{#iq2000-x-elf,,iq2000-*-elf}
2457 @uref{#m32c-x-elf,,m32c-*-elf}
2459 @uref{#m32r-x-elf,,m32r-*-elf}
2461 @uref{#m6811-elf,,m6811-elf}
2463 @uref{#m6812-elf,,m6812-elf}
2465 @uref{#m68k-x-x,,m68k-*-*}
2467 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
2469 @uref{#m68k-uclinux,,m68k-uclinux}
2471 @uref{#mips-x-x,,mips-*-*}
2473 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2475 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2477 @uref{#powerpc-x-x,,powerpc*-*-*, powerpc-*-sysv4}
2479 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
2481 @uref{#powerpc-x-elf,,powerpc-*-elf, powerpc-*-sysv4}
2483 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
2485 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
2487 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
2489 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
2491 @uref{#powerpcle-x-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
2493 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
2495 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
2497 @uref{#s390-x-linux,,s390-*-linux*}
2499 @uref{#s390x-x-linux,,s390x-*-linux*}
2501 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
2503 @uref{#x-x-solaris2,,*-*-solaris2*}
2505 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
2507 @uref{#sparc-sun-solaris27,,sparc-sun-solaris2.7}
2509 @uref{#sparc-x-linux,,sparc-*-linux*}
2511 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
2513 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
2515 @uref{#x-x-sysv,,*-*-sysv*}
2517 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
2519 @uref{#x-x-vxworks,,*-*-vxworks*}
2521 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
2523 @uref{#xtensa-x-elf,,xtensa-*-elf}
2525 @uref{#xtensa-x-linux,,xtensa-*-linux*}
2527 @uref{#windows,,Microsoft Windows}
2531 @uref{#older,,Older systems}
2536 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
2542 <!-- -------- host/target specific issues start here ---------------- -->
2545 @heading @anchor{alpha-x-x}alpha*-*-*
2547 This section contains general configuration information for all
2548 alpha-based platforms using ELF (in particular, ignore this section for
2549 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2550 section, please read all other sections that match your target.
2552 We require binutils 2.11.2 or newer.
2553 Previous binutils releases had a number of problems with DWARF 2
2554 debugging information, not the least of which is incorrect linking of
2560 @heading @anchor{alpha-dec-osf}alpha*-dec-osf*
2561 Systems using processors that implement the DEC Alpha architecture and
2562 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2563 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2565 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2566 supported. (These are the versions which identify themselves as DEC
2569 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2570 may be fixed by configuring with @option{--with-gc=simple},
2571 reconfiguring Kernel Virtual Memory and Swap parameters
2572 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2573 or applying the patch in
2574 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2576 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2577 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2578 we need to use the old assembler, invoked via the barely documented
2579 @option{-oldas} option. To bootstrap GCC, you either need to use the
2583 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2586 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2589 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2592 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2593 are supported on Tru64 UNIX, so you must not configure GCC with
2594 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2596 GCC writes a @samp{.verstamp} directive to the assembler output file
2597 unless it is built as a cross-compiler. It gets the version to use from
2598 the system header file @file{/usr/include/stamp.h}. If you install a
2599 new version of DEC Unix, you should rebuild GCC to pick up the new version
2602 Note that since the Alpha is a 64-bit architecture, cross-compilers from
2603 32-bit machines will not generate code as efficient as that generated
2604 when the compiler is running on a 64-bit machine because many
2605 optimizations that depend on being able to represent a word on the
2606 target in an integral value on the host cannot be performed. Building
2607 cross-compilers on the Alpha for 32-bit machines has only been tested in
2608 a few cases and may not work properly.
2610 @samp{make compare} may fail on old versions of DEC Unix unless you add
2611 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2612 assembler input file is stored in the object file, and that makes
2613 comparison fail if it differs between the @code{stage1} and
2614 @code{stage2} compilations. The option @option{-save-temps} forces a
2615 fixed name to be used for the assembler input file, instead of a
2616 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2617 unless the comparisons fail without that option. If you add
2618 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2619 @samp{.s} files after each series of compilations.
2621 GCC now supports both the native (ECOFF) debugging format used by DBX
2622 and GDB and an encapsulated STABS format for use only with GDB@. See the
2623 discussion of the @option{--with-stabs} option of @file{configure} above
2624 for more information on these formats and how to select them.
2626 There is a bug in DEC's assembler that produces incorrect line numbers
2627 for ECOFF format when the @samp{.align} directive is used. To work
2628 around this problem, GCC will not emit such alignment directives
2629 while writing ECOFF format debugging information even if optimization is
2630 being performed. Unfortunately, this has the very undesirable
2631 side-effect that code addresses when @option{-O} is specified are
2632 different depending on whether or not @option{-g} is also specified.
2634 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2635 DBX@. DEC is now aware of this problem with the assembler and hopes to
2636 provide a fix shortly.
2641 @heading @anchor{alphaev5-cray-unicosmk}alphaev5-cray-unicosmk*
2642 Cray T3E systems running Unicos/Mk.
2644 This port is incomplete and has many known bugs. We hope to improve the
2645 support for this target soon. Currently, only the C front end is supported,
2646 and it is not possible to build parallel applications. Cray modules are not
2647 supported; in particular, Craylibs are assumed to be in
2648 @file{/opt/ctl/craylibs/craylibs}.
2650 On this platform, you need to tell GCC where to find the assembler and
2651 the linker. The simplest way to do so is by providing @option{--with-as}
2652 and @option{--with-ld} to @file{configure}, e.g.@:
2655 configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld \
2656 --enable-languages=c
2659 The comparison test at the end of the bootstrapping process fails on Unicos/Mk
2660 because the assembler inserts timestamps into object files. You should
2661 be able to work around this by doing @samp{make all} after getting this
2667 @heading @anchor{arc-x-elf}arc-*-elf
2668 Argonaut ARC processor.
2669 This configuration is intended for embedded systems.
2674 @heading @anchor{arm-x-elf}arm-*-elf
2675 @heading @anchor{xscale-x-x}xscale-*-*
2676 ARM-family processors. Subtargets that use the ELF object format
2677 require GNU binutils 2.13 or newer. Such subtargets include:
2678 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux},
2679 @code{arm-*-rtems} and @code{arm-*-kaos}.
2684 @heading @anchor{arm-x-coff}arm-*-coff
2685 ARM-family processors. Note that there are two different varieties
2686 of PE format subtarget supported: @code{arm-wince-pe} and
2687 @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}.
2692 @heading @anchor{arm-x-aout}arm-*-aout
2693 ARM-family processors. These targets support the AOUT file format:
2694 @code{arm-*-aout}, @code{arm-*-netbsd}.
2699 @heading @anchor{avr}avr
2701 ATMEL AVR-family micro controllers. These are used in embedded
2702 applications. There are no standard Unix configurations.
2704 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
2708 See ``AVR Options'' in the main manual
2710 for the list of supported MCU types.
2712 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2714 Further installation notes and other useful information about AVR tools
2715 can also be obtained from:
2719 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
2721 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2723 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2726 We @emph{strongly} recommend using binutils 2.13 or newer.
2728 The following error:
2730 Error: register required
2733 indicates that you should upgrade to a newer version of the binutils.
2738 @heading @anchor{bfin}Blackfin
2740 The Blackfin processor, an Analog Devices DSP.
2742 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
2746 See ``Blackfin Options'' in the main manual
2749 More information, and a version of binutils with support for this processor,
2750 is available at @uref{http://blackfin.uclinux.org}
2755 @heading @anchor{c4x}c4x
2757 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
2758 Processors. These are used in embedded applications. There are no
2759 standard Unix configurations.
2761 @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using the
2762 GNU Compiler Collection (GCC)},
2765 See ``TMS320C3x/C4x Options'' in the main manual
2767 for the list of supported MCU types.
2769 GCC can be configured as a cross compiler for both the C3x and C4x
2770 architectures on the same system. Use @samp{configure --target=c4x
2771 --enable-languages="c,c++"} to configure.
2774 Further installation notes and other useful information about C4x tools
2775 can also be obtained from:
2779 @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/}
2785 @heading @anchor{cris}CRIS
2787 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2788 series. These are used in embedded applications.
2791 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
2795 See ``CRIS Options'' in the main manual
2797 for a list of CRIS-specific options.
2799 There are a few different CRIS targets:
2801 @item cris-axis-aout
2802 Old target. Includes a multilib for the @samp{elinux} a.out-based
2803 target. No multilibs for newer architecture variants.
2805 Mainly for monolithic embedded systems. Includes a multilib for the
2806 @samp{v10} core used in @samp{ETRAX 100 LX}.
2807 @item cris-axis-linux-gnu
2808 A GNU/Linux port for the CRIS architecture, currently targeting
2809 @samp{ETRAX 100 LX} by default.
2812 For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11
2813 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2815 Pre-packaged tools can be obtained from
2816 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2817 information about this platform is available at
2818 @uref{http://developer.axis.com/}.
2823 @heading @anchor{crx}CRX
2825 The CRX CompactRISC architecture is a low-power 32-bit architecture with
2826 fast context switching and architectural extensibility features.
2829 @xref{CRX Options,, CRX Options, gcc, Using and Porting the GNU Compiler
2834 See ``CRX Options'' in the main manual for a list of CRX-specific options.
2837 Use @samp{configure --target=crx-elf --enable-languages=c,c++} to configure
2838 GCC@ for building a CRX cross-compiler. The option @samp{--target=crx-elf}
2839 is also used to build the @samp{newlib} C library for CRX.
2841 It is also possible to build libstdc++-v3 for the CRX architecture. This
2842 needs to be done in a separate step with the following configure settings:
2843 @samp{gcc/libstdc++-v3/configure --host=crx-elf --with-newlib
2844 --enable-sjlj-exceptions --enable-cxx-flags='-fexceptions -frtti'}
2849 @heading @anchor{dos}DOS
2851 Please have a look at the @uref{binaries.html,,binaries page}.
2853 You cannot install GCC by itself on MSDOS; it will not compile under
2854 any MSDOS compiler except itself. You need to get the complete
2855 compilation package DJGPP, which includes binaries as well as sources,
2856 and includes all the necessary compilation tools and libraries.
2861 @heading @anchor{x-x-freebsd}*-*-freebsd*
2863 The version of binutils installed in @file{/usr/bin} probably works with
2864 this release of GCC@. However, on FreeBSD 4, bootstrapping against the
2865 latest FSF binutils is known to improve overall testsuite results; and,
2866 on FreeBSD/alpha, using binutils 2.14 or later is required to build libjava.
2868 Support for FreeBSD 1 was discontinued in GCC 3.2.
2870 Support for FreeBSD 2 will be discontinued after GCC 3.4. The
2871 following was true for GCC 3.1 but the current status is unknown.
2872 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2873 configuration support and files as shipped with GCC 2.95 are still in
2874 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2875 it is unknown which version of binutils was used (it is assumed that it
2876 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2878 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2879 default for all CPU architectures. It had been the default on
2880 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2881 of @option{-g}, if you really want the old debugging format. There are
2882 no known issues with mixing object files and libraries with different
2883 debugging formats. Otherwise, this release of GCC should now match more
2884 of the configuration used in the stock FreeBSD configuration of GCC@. In
2885 particular, @option{--enable-threads} is now configured by default.
2886 However, as a general user, do not attempt to replace the system
2887 compiler with this release. Known to bootstrap and check with good
2888 results on FreeBSD 4.9-STABLE and 5-CURRENT@. In the past, known to
2889 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
2890 4.3, 4.4, 4.5, 4.8-STABLE@.
2892 In principle, @option{--enable-threads} is now compatible with
2893 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
2894 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
2896 library may be incorrectly built (symbols are missing at link time).
2897 There is a rare timing-based startup hang (probably involves an
2898 assumption about the thread library). Multi-threaded boehm-gc (required for
2899 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
2900 4.5-RELEASE@. Other CPU architectures
2901 supported by FreeBSD will require additional configuration tuning in, at
2902 the very least, both boehm-gc and libffi.
2904 Shared @file{libgcc_s.so} is now built and installed by default.
2909 @heading @anchor{h8300-hms}h8300-hms
2910 Renesas H8/300 series of processors.
2912 Please have a look at the @uref{binaries.html,,binaries page}.
2914 The calling convention and structure layout has changed in release 2.6.
2915 All code must be recompiled. The calling convention now passes the
2916 first three arguments in function calls in registers. Structures are no
2917 longer a multiple of 2 bytes.
2922 @heading @anchor{hppa-hp-hpux}hppa*-hp-hpux*
2923 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
2925 We require using gas/binutils on all hppa platforms;
2926 you may encounter a variety of problems if you try to use the HP assembler.
2928 Specifically, @option{-g} does not work on HP-UX (since that system
2929 uses a peculiar debugging format which GCC does not know about), unless
2930 you use GAS and GDB@. It may be helpful to configure GCC with the
2931 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
2932 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
2934 If you wish to use the pa-risc 2.0 architecture support with a 32-bit
2935 runtime, you must use gas/binutils 2.11 or newer.
2937 There are two default scheduling models for instructions. These are
2938 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
2939 architecture specified for the target machine when configuring.
2940 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
2941 the target is a @samp{hppa1*} machine.
2943 The PROCESSOR_8000 model is not well suited to older processors. Thus,
2944 it is important to completely specify the machine architecture when
2945 configuring if you want a model other than PROCESSOR_8000. The macro
2946 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
2947 default scheduling model is desired.
2949 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
2950 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
2951 This namespace change might cause problems when bootstrapping with
2952 an earlier version of GCC or the HP compiler as essentially the same
2953 namespace is required for an entire build. This problem can be avoided
2954 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
2955 or @samp{98}. Another way is to add an appropriate set of predefines
2956 to @env{CC}. The description for the @option{munix=} option contains
2957 a list of the predefines used with each standard.
2959 As of GCC 4.1, @env{DWARF2} exception handling is available on HP-UX.
2960 It is now the default. This exposed a bug in the handling of data
2961 relocations in the GAS assembler. The handling of 64-bit data relocations
2962 was seriously broken, affecting debugging and exception support on all
2963 @samp{hppa64-*-*} targets. Under some circumstances, 32-bit data relocations
2964 could also be handled incorrectly. This problem is fixed in GAS version
2967 GCC versions prior to 4.1 incorrectly passed and returned complex
2968 values. They are now passed in the same manner as aggregates.
2970 More specific information to @samp{hppa*-hp-hpux*} targets follows.
2975 @heading @anchor{hppa-hp-hpux10}hppa*-hp-hpux10
2977 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
2978 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
2984 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
2988 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
2992 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
2995 The HP assembler on these systems has some problems. Most notably the
2996 assembler inserts timestamps into each object file it creates, causing
2997 the 3-stage comparison test to fail during a bootstrap.
2998 You should be able to continue by saying @samp{make all-host all-target}
2999 after getting the failure from @samp{make}.
3001 GCC 4.0 requires CVS binutils as of April 28, 2004 or later. Earlier
3002 versions require binutils 2.8 or later.
3004 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
3005 used for one-only code and data. This resolves many of the previous
3006 problems in using C++ on this target. However, the ABI is not compatible
3007 with the one implemented under HP-UX 11 using secondary definitions.
3012 @heading @anchor{hppa-hp-hpux11}hppa*-hp-hpux11
3014 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
3015 be used to compile GCC 3.0 and up.
3017 Refer to @uref{binaries.html,,binaries} for information about obtaining
3018 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
3019 to build the Ada language as it can't be bootstrapped using C@. Ada is
3020 only available for the 32-bit PA-RISC runtime. The libffi and libjava
3021 haven't been ported to HP-UX and don't build.
3023 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
3024 bundled compiler supports only traditional C; you will need either HP's
3025 unbundled compiler, or a binary distribution of GCC@.
3027 It is possible to build GCC 3.3 starting with the bundled HP compiler,
3028 but the process requires several steps. GCC 3.3 can then be used to
3029 build later versions. The fastjar program contains ISO C code and
3030 can't be built with the HP bundled compiler. This problem can be
3031 avoided by not building the Java language. For example, use the
3032 @option{--enable-languages="c,c++,f77,objc"} option in your configure
3035 There are several possible approaches to building the distribution.
3036 Binutils can be built first using the HP tools. Then, the GCC
3037 distribution can be built. The second approach is to build GCC
3038 first using the HP tools, then build binutils, then rebuild GCC@.
3039 There have been problems with various binary distributions, so it
3040 is best not to start from a binary distribution.
3042 On 64-bit capable systems, there are two distinct targets. Different
3043 installation prefixes must be used if both are to be installed on
3044 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
3045 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
3046 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
3047 PA-RISC 2.0 architecture. The HP and GNU linkers are both supported
3050 The script config.guess now selects the target type based on the compiler
3051 detected during configuration. You must define @env{PATH} or @env{CC} so
3052 that configure finds an appropriate compiler for the initial bootstrap.
3053 When @env{CC} is used, the definition should contain the options that are
3054 needed whenever @env{CC} is used.
3056 Specifically, options that determine the runtime architecture must be
3057 in @env{CC} to correctly select the target for the build. It is also
3058 convenient to place many other compiler options in @env{CC}. For example,
3059 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
3060 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
3061 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
3062 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
3063 macro definition table of cpp needs to be increased for a successful
3064 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
3065 be defined when building with the bundled compiler, or when using the
3066 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
3068 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
3069 with the @option{--with-ld=@dots{}} option. This overrides the standard
3070 search for ld. The two linkers supported on this target require different
3071 commands. The default linker is determined during configuration. As a
3072 result, it's not possible to switch linkers in the middle of a GCC build.
3073 This has been been reported to sometimes occur in unified builds of
3076 GCC 3.0 through 3.2 require binutils 2.11 or above. GCC 3.3 through
3077 GCC 4.0 require binutils 2.14 or later.
3079 Although the HP assembler can be used for an initial build, it shouldn't
3080 be used with any languages other than C and perhaps Fortran due to its
3081 many limitations. For example, it does not support weak symbols or alias
3082 definitions. As a result, explicit template instantiations are required
3083 when using C++. This makes it difficult if not impossible to build many
3084 C++ applications. You can't generate debugging information when using
3085 the HP assembler. Finally, bootstrapping fails in the final
3086 comparison of object modules due to the time stamps that it inserts into
3087 the modules. The bootstrap can be continued from this point with
3088 @samp{make all-host all-target}.
3090 A recent linker patch must be installed for the correct operation of
3091 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
3092 oldest linker patches that are known to work. They are for HP-UX
3093 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
3094 @code{PHSS_24304}, might be usable but it hasn't been tested. These
3095 patches have been superseded. Consult the HP patch database to obtain
3096 the currently recommended linker patch for your system.
3098 The patches are necessary for the support of weak symbols on the
3099 32-bit port, and for the running of initializers and finalizers. Weak
3100 symbols are implemented using SOM secondary definition symbols. Prior
3101 to HP-UX 11, there are bugs in the linker support for secondary symbols.
3102 The patches correct a problem of linker core dumps creating shared
3103 libraries containing secondary symbols, as well as various other
3104 linking issues involving secondary symbols.
3106 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
3107 run initializers and finalizers on the 64-bit port. The 32-bit port
3108 uses the linker @option{+init} and @option{+fini} options for the same
3109 purpose. The patches correct various problems with the +init/+fini
3110 options, including program core dumps. Binutils 2.14 corrects a
3111 problem on the 64-bit port resulting from HP's non-standard use of
3112 the .init and .fini sections for array initializers and finalizers.
3114 There are a number of issues to consider in selecting which linker to
3115 use with the 64-bit port. The GNU 64-bit linker can only create dynamic
3116 binaries. The @option{-static} option causes linking with archive
3117 libraries but doesn't produce a truly static binary. Dynamic binaries
3118 still require final binding by the dynamic loader to resolve a set of
3119 dynamic-loader-defined symbols. The default behavior of the HP linker
3120 is the same as the GNU linker. However, it can generate true 64-bit
3121 static binaries using the @option{+compat} option.
3123 The HP 64-bit linker doesn't support linkonce semantics. As a
3124 result, C++ programs have many more sections than they should.
3126 The GNU 64-bit linker has some issues with shared library support
3127 and exceptions. As a result, we only support libgcc in archive
3128 format. For similar reasons, dwarf2 unwind and exception support
3129 are disabled. The GNU linker also has problems creating binaries
3130 with @option{-static}. It doesn't provide stubs for internal
3131 calls to global functions in shared libraries, so these calls
3132 can't be overloaded.
3134 Thread support is not implemented in GCC 3.0 through 3.2, so the
3135 @option{--enable-threads} configure option does not work. In 3.3
3136 and later, POSIX threads are supported. The optional DCE thread
3137 library is not supported.
3139 This port still is undergoing significant development.
3144 @heading @anchor{x-x-linux-gnu}*-*-linux-gnu
3146 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
3147 in glibc 2.2.5 and later. More information is available in the
3148 libstdc++-v3 documentation.
3153 @heading @anchor{ix86-x-linuxaout}i?86-*-linux*aout
3154 Use this configuration to generate @file{a.out} binaries on Linux-based
3155 GNU systems. This configuration is being superseded.
3160 @heading @anchor{ix86-x-linux}i?86-*-linux*
3162 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
3163 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
3165 If you receive Signal 11 errors when building on GNU/Linux, then it is
3166 possible you have a hardware problem. Further information on this can be
3167 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
3172 @heading @anchor{ix86-x-sco32v5}i?86-*-sco3.2v5*
3173 Use this for the SCO OpenServer Release 5 family of operating systems.
3175 Unlike earlier versions of GCC, the ability to generate COFF with this
3176 target is no longer provided.
3178 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
3179 the system debugger to be used. That support was too burdensome to
3180 maintain. GCC now emits only DWARF 2 for this target. This means you
3181 may use either the UDK debugger or GDB to debug programs built by this
3184 GCC is now only supported on releases 5.0.4 and later, and requires that
3185 you install Support Level Supplement OSS646B or later, and Support Level
3186 Supplement OSS631C or later. If you are using release 5.0.7 of
3187 OpenServer, you must have at least the first maintenance pack installed
3188 (this includes the relevant portions of OSS646). OSS646, also known as
3189 the ``Execution Environment Update'', provides updated link editors and
3190 assemblers, as well as updated standard C and math libraries. The C
3191 startup modules are also updated to support the System V gABI draft, and
3192 GCC relies on that behavior. OSS631 provides a collection of commonly
3193 used open source libraries, some of which GCC depends on (such as GNU
3194 gettext and zlib). SCO OpenServer Release 5.0.7 has all of this built
3195 in by default, but OSS631C and later also apply to that release. Please
3197 @uref{ftp://ftp.sco.com/pub/openserver5,,ftp://ftp.sco.com/pub/openserver5}
3198 for the latest versions of these (and other potentially useful)
3201 Although there is support for using the native assembler, it is
3202 recommended that you configure GCC to use the GNU assembler. You do
3203 this by using the flags
3204 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You should
3205 use a modern version of GNU binutils. Version 2.13.2.1 was used for all
3206 testing. In general, only the @option{--with-gnu-as} option is tested.
3207 A modern bintuils (as well as a plethora of other development related
3208 GNU utilities) can be found in Support Level Supplement OSS658A, the
3209 ``GNU Development Tools'' package. See the SCO web and ftp sites for details.
3210 That package also contains the currently ``officially supported'' version of
3211 GCC, version 2.95.3. It is useful for bootstrapping this version.
3216 @heading @anchor{ix86-x-solaris210}i?86-*-solaris2.10
3217 Use this for Solaris 10 or later on x86 and x86-64 systems. This
3218 configuration is supported by GCC 4.0 and later versions only.
3220 It is recommended that you configure GCC to use the GNU assembler in
3221 @file{/usr/sfw/bin/gas} but the Sun linker, using the options
3222 @option{--with-gnu-as --with-as=/usr/sfw/bin/gas --without-gnu-ld
3223 --with-ld=/usr/ccs/bin/ld}.
3228 @heading @anchor{ix86-x-udk}i?86-*-udk
3230 This target emulates the SCO Universal Development Kit and requires that
3231 package be installed. (If it is installed, you will have a
3232 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
3233 @samp{i?86-*-unixware7*} target
3234 but is meant to be used when hosting on a system where UDK isn't the
3235 default compiler such as OpenServer 5 or Unixware 2. This target will
3236 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
3237 with the same warnings and caveats as the SCO UDK@.
3239 This target is a little tricky to build because we have to distinguish
3240 it from the native tools (so it gets headers, startups, and libraries
3241 from the right place) while making the tools not think we're actually
3242 building a cross compiler. The easiest way to do this is with a configure
3246 CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure \
3247 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-
3250 @emph{You should substitute @samp{i686} in the above command with the appropriate
3251 processor for your host.}
3253 After the usual @samp{make} and
3254 @samp{make install}, you can then access the UDK-targeted GCC
3255 tools by adding @command{udk-} before the commonly known name. For
3256 example, to invoke the C compiler, you would use @command{udk-gcc}.
3257 They will coexist peacefully with any native-target GCC tools you may
3264 @heading @anchor{ia64-x-linux}ia64-*-linux
3265 IA-64 processor (also known as IPF, or Itanium Processor Family)
3268 If you are using the installed system libunwind library with
3269 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
3272 None of the following versions of GCC has an ABI that is compatible
3273 with any of the other versions in this list, with the exception that
3274 Red Hat 2.96 and Trillian 000171 are compatible with each other:
3275 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
3276 This primarily affects C++ programs and programs that create shared libraries.
3277 GCC 3.1 or later is recommended for compiling linux, the kernel.
3278 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
3279 more major ABI changes are expected.
3284 @heading @anchor{ia64-x-hpux}ia64-*-hpux*
3285 Building GCC on this target requires the GNU Assembler. The bundled HP
3286 assembler will not work. To prevent GCC from using the wrong assembler,
3287 the option @option{--with-gnu-as} may be necessary.
3289 The GCC libunwind library has not been ported to HPUX@. This means that for
3290 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
3291 is required to build GCC@. For GCC 3.3 and later, this is the default.
3292 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
3293 removed and the system libunwind library will always be used.
3297 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
3299 @heading @anchor{x-ibm-aix}*-ibm-aix*
3300 Support for AIX version 3 and older was discontinued in GCC 3.4.
3302 ``out of memory'' bootstrap failures may indicate a problem with
3303 process resource limits (ulimit). Hard limits are configured in the
3304 @file{/etc/security/limits} system configuration file.
3306 To speed up the configuration phases of bootstrapping and installing GCC,
3307 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
3310 % CONFIG_SHELL=/opt/freeware/bin/bash
3311 % export CONFIG_SHELL
3314 and then proceed as described in @uref{build.html,,the build
3315 instructions}, where we strongly recommend specifying an absolute path
3316 to invoke @var{srcdir}/configure.
3318 Because GCC on AIX is built as a 32-bit executable by default,
3319 (although it can generate 64-bit programs) the GMP and MPFR libraries
3320 required by gfortran must be 32-bit libraries. Building GMP and MPFR
3321 as static archive libraries works better than shared libraries.
3323 Errors involving @code{alloca} when building GCC generally are due
3324 to an incorrect definition of @code{CC} in the Makefile or mixing files
3325 compiled with the native C compiler and GCC@. During the stage1 phase of
3326 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
3327 (not @command{xlc}). Once @command{configure} has been informed of
3328 @command{xlc}, one needs to use @samp{make distclean} to remove the
3329 configure cache files and ensure that @env{CC} environment variable
3330 does not provide a definition that will confuse @command{configure}.
3331 If this error occurs during stage2 or later, then the problem most likely
3332 is the version of Make (see above).
3334 The native @command{as} and @command{ld} are recommended for bootstrapping
3335 on AIX 4 and required for bootstrapping on AIX 5L@. The GNU Assembler
3336 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
3337 utilize weak symbol functionality although it is not supported. The GNU
3338 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC@.
3339 The native AIX tools do interoperate with GCC@.
3341 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
3342 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
3343 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
3344 referenced as APAR IY53606 (AIX 5.2) or a APAR IY54774 (AIX 5.1)
3346 @samp{libstdc++} in GCC 3.4 increments the major version number of the
3347 shared object and GCC installation places the @file{libstdc++.a}
3348 shared library in a common location which will overwrite the and GCC
3349 3.3 version of the shared library. Applications either need to be
3350 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
3351 versions of the @samp{libstdc++} shared object needs to be available
3352 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
3353 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
3354 installed for runtime dynamic loading using the following steps to set
3355 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3356 multilib @file{libstdc++.a} installed:
3358 Extract the shared objects from the currently installed
3359 @file{libstdc++.a} archive:
3361 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3364 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3365 available for runtime dynamic loading, but not linking:
3367 % strip -e libstdc++.so.4 libstdc++.so.5
3370 Archive the runtime-only shared object in the GCC 3.4
3371 @file{libstdc++.a} archive:
3373 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
3376 Linking executables and shared libraries may produce warnings of
3377 duplicate symbols. The assembly files generated by GCC for AIX always
3378 have included multiple symbol definitions for certain global variable
3379 and function declarations in the original program. The warnings should
3380 not prevent the linker from producing a correct library or runnable
3383 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
3384 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
3385 to parse archive libraries did not handle the new format correctly.
3386 These routines are used by GCC and result in error messages during
3387 linking such as ``not a COFF file''. The version of the routines shipped
3388 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
3389 option of the archive command may be used to create archives of 32-bit
3390 objects using the original ``small format''. A correct version of the
3391 routines is shipped with AIX 4.3.2 and above.
3393 Some versions of the AIX binder (linker) can fail with a relocation
3394 overflow severe error when the @option{-bbigtoc} option is used to link
3395 GCC-produced object files into an executable that overflows the TOC@. A fix
3396 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
3397 available from IBM Customer Support and from its
3398 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3399 website as PTF U455193.
3401 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
3402 with a segmentation fault when invoked by any version of GCC@. A fix for
3403 APAR IX87327 is available from IBM Customer Support and from its
3404 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3405 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
3407 The initial assembler shipped with AIX 4.3.0 generates incorrect object
3408 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
3409 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
3410 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3411 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
3413 AIX provides National Language Support (NLS)@. Compilers and assemblers
3414 use NLS to support locale-specific representations of various data
3415 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
3416 separating decimal fractions). There have been problems reported where
3417 GCC does not produce the same floating-point formats that the assembler
3418 expects. If one encounters this problem, set the @env{LANG}
3419 environment variable to @samp{C} or @samp{En_US}.
3421 By default, GCC for AIX 4.1 and above produces code that can be used on
3422 both Power or PowerPC processors.
3424 A default can be specified with the @option{-mcpu=@var{cpu_type}}
3425 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
3430 @heading @anchor{iq2000-x-elf}iq2000-*-elf
3431 Vitesse IQ2000 processors. These are used in embedded
3432 applications. There are no standard Unix configurations.
3437 @heading @anchor{m32c-x-elf}m32c-*-elf
3438 Renesas M32C processor.
3439 This configuration is intended for embedded systems.
3444 @heading @anchor{m32r-x-elf}m32r-*-elf
3445 Renesas M32R processor.
3446 This configuration is intended for embedded systems.
3451 @heading @anchor{m6811-elf}m6811-elf
3452 Motorola 68HC11 family micro controllers. These are used in embedded
3453 applications. There are no standard Unix configurations.
3458 @heading @anchor{m6812-elf}m6812-elf
3459 Motorola 68HC12 family micro controllers. These are used in embedded
3460 applications. There are no standard Unix configurations.
3465 @heading @anchor{m68k-x-x}m68k-*-*
3466 By default, @samp{m68k-*-aout}, @samp{m68k-*-coff*},
3467 @samp{m68k-*-elf*}, @samp{m68k-*-rtems} and @samp{m68k-*-uclinux}
3468 build libraries for both M680x0 and ColdFire processors. If you only
3469 need the M680x0 libraries, you can omit the ColdFire ones by passing
3470 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
3471 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
3472 @command{configure}. These targets default to 5206 code when
3473 configured with @option{--with-arch=cf} and 68020 code otherwise.
3475 The @samp{m68k-*-linux-gnu}, @samp{m68k-*-netbsd} and
3476 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
3477 option. They will generate ColdFire CFV4e code when configured with
3478 @option{--with-arch=cf} and 68020 code otherwise.
3480 You can override the default processors listed above by configuring
3481 with @option{--with-cpu=@var{target}}. This @var{target} can either
3482 be a @option{-mcpu} argument or one of the following values:
3483 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
3484 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
3489 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
3490 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
3491 the assembler that prevents compilation of GCC@. This
3492 bug manifests itself during the first stage of compilation, while
3493 building @file{libgcc2.a}:
3497 cc1: warning: `-g' option not supported on this version of GCC
3498 cc1: warning: `-g1' option not supported on this version of GCC
3499 ./xgcc: Internal compiler error: program as got fatal signal 11
3502 A patched version of the assembler is available as the file
3503 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
3504 have HP software support, the patch can also be obtained directly from
3505 HP, as described in the following note:
3508 This is the patched assembler, to patch SR#1653-010439, where the
3509 assembler aborts on floating point constants.
3511 The bug is not really in the assembler, but in the shared library
3512 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
3513 SR#4701-078451. Anyway, the attached assembler uses the archive
3514 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
3517 This patch is also known as PHCO_4484.
3519 In addition gdb does not understand that native HP-UX format, so
3520 you must use gas if you wish to use gdb.
3522 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
3523 @command{fixproto} shell script triggers a bug in the system shell. If you
3524 encounter this problem, upgrade your operating system or use BASH (the
3525 GNU shell) to run @command{fixproto}. This bug will cause the fixproto
3526 program to report an error of the form:
3529 ./fixproto: sh internal 1K buffer overflow
3532 To fix this, you can also change the first line of the fixproto script
3542 @heading @anchor{m68k-x-uclinux}m68k-*-uclinux
3543 GCC 4.3 changed the uClinux configuration so that it uses the
3544 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
3545 It also added improved support for C++ and flat shared libraries,
3546 both of which were ABI changes. However, you can still use the
3547 original ABI by configuring for @samp{m68k-uclinuxoldabi} or
3548 @samp{m68k-@var{vendor}-uclinuxoldabi}.
3553 @heading @anchor{mips-x-x}mips-*-*
3554 If on a MIPS system you get an error message saying ``does not have gp
3555 sections for all it's [sic] sectons [sic]'', don't worry about it. This
3556 happens whenever you use GAS with the MIPS linker, but there is not
3557 really anything wrong, and it is okay to use the output file. You can
3558 stop such warnings by installing the GNU linker.
3560 It would be nice to extend GAS to produce the gp tables, but they are
3561 optional, and there should not be a warning about their absence.
3563 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
3564 and later. A patch went in just after the GCC 3.3 release to
3565 make @samp{mips*-*-*} use the generic implementation instead. You can also
3566 configure for @samp{mipsel-elf} as a workaround. The
3567 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
3568 work on this is expected in future releases.
3570 MIPS systems check for division by zero (unless
3571 @option{-mno-check-zero-division} is passed to the compiler) by
3572 generating either a conditional trap or a break instruction. Using
3573 trap results in smaller code, but is only supported on MIPS II and
3574 later. Also, some versions of the Linux kernel have a bug that
3575 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
3576 the use of break, use the @option{--with-divide=breaks}
3577 @command{configure} option when configuring GCC@. The default is to
3578 use traps on systems that support them.
3580 Cross-compilers for the MIPS as target using the MIPS assembler
3581 currently do not work, because the auxiliary programs
3582 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
3583 anything but a MIPS. It does work to cross compile for a MIPS
3584 if you use the GNU assembler and linker.
3586 The assembler from GNU binutils 2.17 and earlier has a bug in the way
3587 it sorts relocations for REL targets (o32, o64, EABI). This can cause
3588 bad code to be generated for simple C++ programs. Also the linker
3589 from GNU binutils versions prior to 2.17 has a bug which causes the
3590 runtime linker stubs in very large programs, like @file{libgcj.so}, to
3591 be incorrectly generated. Binutils CVS snapshots and releases made
3592 after Nov. 9, 2006 are thought to be free from both of these problems.
3597 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
3599 In order to compile GCC on an SGI running IRIX 5, the @samp{compiler_dev.hdr}
3600 subsystem must be installed from the IDO CD-ROM supplied by SGI@.
3601 It is also available for download from
3602 @uref{ftp://ftp.sgi.com/sgi/IRIX5.3/iris-development-option-5.3.tardist}.
3604 If you use the MIPS C compiler to bootstrap, it may be necessary
3605 to increase its table size for switch statements with the
3606 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
3607 optimization option, you also need to use @option{-Olimit 3000}.
3609 To enable debugging under IRIX 5, you must use GNU binutils 2.15 or
3610 later, and use the @option{--with-gnu-ld} @command{configure} option
3611 when configuring GCC@. You need to use GNU @command{ar} and @command{nm},
3612 also distributed with GNU binutils.
3614 Some users have reported that @command{/bin/sh} will hang during bootstrap.
3615 This problem can be avoided by running the commands:
3618 % CONFIG_SHELL=/bin/ksh
3619 % export CONFIG_SHELL
3622 before starting the build.
3627 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3629 If you are using SGI's MIPSpro @command{cc} as your bootstrap compiler, you must
3630 ensure that the N32 ABI is in use. To test this, compile a simple C
3631 file with @command{cc} and then run @command{file} on the
3632 resulting object file. The output should look like:
3635 test.o: ELF N32 MSB @dots{}
3641 test.o: ELF 32-bit MSB @dots{}
3647 test.o: ELF 64-bit MSB @dots{}
3650 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3651 should set the environment variable @env{CC} to @samp{cc -n32}
3652 before configuring GCC@.
3654 If you want the resulting @command{gcc} to run on old 32-bit systems
3655 with the MIPS R4400 CPU, you need to ensure that only code for the @samp{mips3}
3656 instruction set architecture (ISA) is generated. While GCC 3.x does
3657 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3658 the ISA depending on the machine where GCC is built. Using one of them
3659 as the bootstrap compiler may result in @samp{mips4} code, which won't run at
3660 all on @samp{mips3}-only systems. For the test program above, you should see:
3663 test.o: ELF N32 MSB mips-3 @dots{}
3669 test.o: ELF N32 MSB mips-4 @dots{}
3672 instead, you should set the environment variable @env{CC} to @samp{cc
3673 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3675 MIPSpro C 7.4 may cause bootstrap failures, due to a bug when inlining
3676 @code{memcmp}. Either add @code{-U__INLINE_INTRINSICS} to the @env{CC}
3677 environment variable as a workaround or upgrade to MIPSpro C 7.4.1m.
3679 GCC on IRIX 6 is usually built to support the N32, O32 and N64 ABIs. If
3680 you build GCC on a system that doesn't have the N64 libraries installed
3681 or cannot run 64-bit binaries,
3682 you need to configure with @option{--disable-multilib} so GCC doesn't
3683 try to use them. This will disable building the O32 libraries, too.
3684 Look for @file{/usr/lib64/libc.so.1} to see if you
3685 have the 64-bit libraries installed.
3687 To enable debugging for the O32 ABI, you must use GNU @command{as} from
3688 GNU binutils 2.15 or later. You may also use GNU @command{ld}, but
3689 this is not required and currently causes some problems with Ada.
3691 The @option{--enable-threads} option doesn't currently work, a patch is
3692 in preparation for a future release. The @option{--enable-libgcj}
3693 option is disabled by default: IRIX 6 uses a very low default limit
3694 (20480) for the command line length. Although @command{libtool} contains a
3695 workaround for this problem, at least the N64 @samp{libgcj} is known not
3696 to build despite this, running into an internal error of the native
3697 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3698 its maximum of 262144 bytes. If you have root access, you can use the
3699 @command{systune} command to do this.
3701 @code{wchar_t} support in @samp{libstdc++} is not available for old
3702 IRIX 6.5.x releases, @math{x < 19}. The problem cannot be autodetected
3703 and in order to build GCC for such targets you need to configure with
3704 @option{--disable-wchar_t}.
3706 See @uref{http://freeware.sgi.com/} for more
3707 information about using GCC on IRIX platforms.
3712 @heading @anchor{powerpc-x-x}powerpc-*-*
3714 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3715 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3720 @heading @anchor{powerpc-x-darwin}powerpc-*-darwin*
3721 PowerPC running Darwin (Mac OS X kernel).
3723 Pre-installed versions of Mac OS X may not include any developer tools,
3724 meaning that you will not be able to build GCC from source. Tool
3725 binaries are available at
3726 @uref{http://developer.apple.com/darwin/projects/compiler/} (free
3727 registration required).
3729 This version of GCC requires at least cctools-590.36. The
3730 cctools-590.36 package referenced from
3731 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
3732 on systems older than 10.3.9 (aka darwin7.9.0).
3737 @heading @anchor{powerpc-x-elf}powerpc-*-elf, powerpc-*-sysv4
3738 PowerPC system in big endian mode, running System V.4.
3743 @heading @anchor{powerpc-x-linux-gnu}powerpc*-*-linux-gnu*
3746 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.15}
3747 or newer for a working GCC@.
3752 @heading @anchor{powerpc-x-netbsd}powerpc-*-netbsd*
3753 PowerPC system in big endian mode running NetBSD@. To build the
3754 documentation you will need Texinfo version 4.4 (NetBSD 1.5.1 included
3755 Texinfo version 3.12).
3760 @heading @anchor{powerpc-x-eabisim}powerpc-*-eabisim
3761 Embedded PowerPC system in big endian mode for use in running under the
3767 @heading @anchor{powerpc-x-eabi}powerpc-*-eabi
3768 Embedded PowerPC system in big endian mode.
3773 @heading @anchor{powerpcle-x-elf}powerpcle-*-elf, powerpcle-*-sysv4
3774 PowerPC system in little endian mode, running System V.4.
3779 @heading @anchor{powerpcle-x-eabisim}powerpcle-*-eabisim
3780 Embedded PowerPC system in little endian mode for use in running under
3786 @heading @anchor{powerpcle-x-eabi}powerpcle-*-eabi
3787 Embedded PowerPC system in little endian mode.
3792 @heading @anchor{s390-x-linux}s390-*-linux*
3793 S/390 system running GNU/Linux for S/390@.
3798 @heading @anchor{s390x-x-linux}s390x-*-linux*
3799 zSeries system (64-bit) running GNU/Linux for zSeries@.
3804 @heading @anchor{s390x-ibm-tpf}s390x-ibm-tpf*
3805 zSeries system (64-bit) running TPF@. This platform is
3806 supported as cross-compilation target only.
3811 @c Please use Solaris 2 to refer to all release of Solaris, starting
3812 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
3813 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3814 @c alone is too unspecific and must be avoided.
3815 @heading @anchor{x-x-solaris2}*-*-solaris2*
3817 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3818 GCC you first have to install a pre-built compiler, see the
3819 @uref{binaries.html,,binaries page} for details.
3821 The Solaris 2 @command{/bin/sh} will often fail to configure
3822 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3823 recommend using the following initial sequence of commands
3826 % CONFIG_SHELL=/bin/ksh
3827 % export CONFIG_SHELL
3830 and proceed as described in @uref{configure.html,,the configure instructions}.
3831 In addition we strongly recommend specifying an absolute path to invoke
3832 @var{srcdir}/configure.
3834 Solaris 2 comes with a number of optional OS packages. Some of these
3835 are needed to use GCC fully, namely @code{SUNWarc},
3836 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3837 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3838 optional packages when installing Solaris 2, you will need to verify that
3839 the packages that GCC needs are installed.
3841 To check whether an optional package is installed, use
3842 the @command{pkginfo} command. To add an optional package, use the
3843 @command{pkgadd} command. For further details, see the Solaris 2
3846 Trying to use the linker and other tools in
3847 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3848 For example, the linker may hang indefinitely. The fix is to remove
3849 @file{/usr/ucb} from your @env{PATH}.
3851 The build process works more smoothly with the legacy Sun tools so, if you
3852 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3853 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3855 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3856 platform. We recommend the use of GNU binutils 2.11.2 or later, or the
3857 vendor tools (Sun @command{as}, Sun @command{ld}). Note that your mileage
3858 may vary if you use a combination of the GNU tools and the Sun tools: while
3859 the combination GNU @command{as} + Sun @command{ld} should reasonably work,
3860 the reverse combination Sun @command{as} + GNU @command{ld} is known to
3861 cause memory corruption at runtime in some cases for C++ programs.
3863 The stock GNU binutils 2.15 release is broken on this platform because of a
3864 single bug. It has been fixed on the 2.15 branch in the CVS repository.
3865 You can obtain a working version by checking out the binutils-2_15-branch
3866 from the CVS repository or applying the patch
3867 @uref{http://sourceware.org/ml/binutils-cvs/2004-09/msg00036.html} to the
3870 We recommend using GNU binutils 2.16 or later in conjunction with GCC 4.x,
3871 or the vendor tools (Sun @command{as}, Sun @command{ld}). However, for
3872 Solaris 10 and above, an additional patch is required in order for the GNU
3873 linker to be able to cope with a new flavor of shared libraries. You
3874 can obtain a working version by checking out the binutils-2_16-branch from
3875 the CVS repository or applying the patch
3876 @uref{http://sourceware.org/ml/binutils-cvs/2005-07/msg00122.html} to the
3879 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3880 newer: @command{g++} will complain that types are missing. These headers assume
3881 that omitting the type means @code{int}; this assumption worked for C89 but
3882 is wrong for C++, and is now wrong for C99 also.
3884 @command{g++} accepts such (invalid) constructs with the option
3885 @option{-fpermissive}; it
3886 will assume that any missing type is @code{int} (as defined by C89).
3888 There are patches for Solaris 2.6 (105633-56 or newer for SPARC,
3889 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC,
3890 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3891 108653-22 for Intel) that fix this bug.
3893 Sun bug 4927647 sometimes causes random spurious testsuite failures
3894 related to missing diagnostic output. This bug doesn't affect GCC
3895 itself, rather it is a kernel bug triggered by the @command{expect}
3896 program which is used only by the GCC testsuite driver. When the bug
3897 causes the @command{expect} program to miss anticipated output, extra
3898 testsuite failures appear.
3900 There are patches for Solaris 8 (117350-12 or newer for SPARC,
3901 117351-12 or newer for Intel) and Solaris 9 (117171-11 or newer for
3902 SPARC, 117172-11 or newer for Intel) that address this problem.
3907 @heading @anchor{sparc-sun-solaris2}sparc-sun-solaris2*
3909 When GCC is configured to use binutils 2.11.2 or later the binaries
3910 produced are smaller than the ones produced using Sun's native tools;
3911 this difference is quite significant for binaries containing debugging
3914 Sun @command{as} 4.x is broken in that it cannot cope with long symbol names.
3915 A typical error message might look similar to the following:
3918 /usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041: error:
3919 can't compute value of an expression involving an external symbol.
3922 This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris
3923 2.6 and has been fixed in later (5.x) versions of the assembler,
3924 starting with Solaris 7.
3926 Starting with Solaris 7, the operating system is capable of executing
3927 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3928 this; the @option{-m64} option enables 64-bit code generation.
3929 However, if all you want is code tuned for the UltraSPARC CPU, you
3930 should try the @option{-mtune=ultrasparc} option instead, which produces
3931 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3934 When configuring on a Solaris 7 or later system that is running a kernel
3935 that supports only 32-bit binaries, one must configure with
3936 @option{--disable-multilib}, since we will not be able to build the
3937 64-bit target libraries.
3939 GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
3940 the GNU compiler (especially GCC 3.0.x versions), which lead to the
3941 miscompilation of the stage1 compiler and the subsequent failure of the
3942 bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
3943 stage, i.e.@: to bootstrap that compiler with the base compiler and then
3944 use it to bootstrap the final compiler.
3946 GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
3947 and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
3948 failure in form of a miscompilation of the stage1 compiler by the Sun
3949 compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
3951 GCC 3.4 changed the default debugging format from STABS to DWARF-2 for
3952 32-bit code on Solaris 7 and later. If you use the Sun assembler, this
3953 change apparently runs afoul of Sun bug 4910101 (which is referenced as
3954 a x86-only problem by Sun, probably because they do not use DWARF-2).
3955 A symptom of the problem is that you cannot compile C++ programs like
3956 @command{groff} 1.19.1 without getting messages similar to the following:
3959 ld: warning: relocation error: R_SPARC_UA32: @dots{}
3960 external symbolic relocation against non-allocatable section
3961 .debug_info cannot be processed at runtime: relocation ignored.
3964 To work around this problem, compile with @option{-gstabs+} instead of
3967 When configuring the GNU Multiple Precision Library (GMP) or the MPFR
3968 library on a Solaris 7 or later system, the canonical target triplet
3969 must be specified as the @command{build} parameter on the configure
3970 line. This triplet can be obtained by invoking ./config.guess in
3971 the toplevel source directory of GCC (and not that of GMP or MPFR).
3972 For example on a Solaris 7 system:
3975 % ./configure --build=sparc-sun-solaris2.7 --prefix=xxx
3981 @heading @anchor{sparc-sun-solaris27}sparc-sun-solaris2.7
3983 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3984 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3985 and later, including all EGCS releases. Sun formerly recommended
3986 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3987 recommend it only for people who use Sun's compilers.
3989 Here are some workarounds to this problem:
3992 Do not install Sun patch 107058-01 until after Sun releases a
3993 complete patch for bug 4210064. This is the simplest course to take,
3994 unless you must also use Sun's C compiler. Unfortunately 107058-01
3995 is preinstalled on some new Solaris 7-based hosts, so you may have to
3999 Copy the original, unpatched Solaris 7
4000 @command{/usr/ccs/bin/as} into
4001 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
4002 adjusting the latter name to fit your local conventions and software
4006 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
4007 both 107058-01 and 106950-03 installed has reported the bug with GCC
4008 and Sun's dynamic linker. This last course of action is riskiest,
4009 for two reasons. First, you must install 106950 on all hosts that
4010 run code generated by GCC; it doesn't suffice to install it only on
4011 the hosts that run GCC itself. Second, Sun says that 106950-03 is
4012 only a partial fix for bug 4210064, but Sun doesn't know whether the
4013 partial fix is adequate for GCC@. Revision -08 or later should fix
4014 the bug. The current (as of 2004-05-23) revision is -24, and is included in
4015 the Solaris 7 Recommended Patch Cluster.
4018 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
4019 which causes a bootstrap failure when linking the 64-bit shared version of
4020 libgcc. A typical error message is:
4023 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
4024 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
4027 This bug has been fixed in the final 5.0 version of the assembler.
4029 A similar problem was reported for version Sun WorkShop 6 99/08/18 of the
4030 Sun assembler, which causes a bootstrap failure with GCC 4.0.0:
4033 ld: fatal: relocation error: R_SPARC_DISP32:
4034 file .libs/libstdc++.lax/libsupc++convenience.a/vterminate.o:
4035 symbol <unknown>: offset 0xfccd33ad is non-aligned
4038 This bug has been fixed in more recent revisions of the assembler.
4043 @heading @anchor{sparc-x-linux}sparc-*-linux*
4045 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
4046 or newer on this platform. All earlier binutils and glibc
4047 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
4053 @heading @anchor{sparc64-x-solaris2}sparc64-*-solaris2*
4055 When configuring the GNU Multiple Precision Library (GMP) or the
4056 MPFR library, the canonical target triplet must be specified as
4057 the @command{build} parameter on the configure line. For example
4058 on a Solaris 7 system:
4061 % ./configure --build=sparc64-sun-solaris2.7 --prefix=xxx
4064 The following compiler flags must be specified in the configure
4065 step in order to bootstrap this target with the Sun compiler:
4068 % CC="cc -xarch=v9 -xildoff" @var{srcdir}/configure [@var{options}] [@var{target}]
4071 @option{-xarch=v9} specifies the SPARC-V9 architecture to the Sun toolchain
4072 and @option{-xildoff} turns off the incremental linker.
4077 @heading @anchor{sparcv9-x-solaris2}sparcv9-*-solaris2*
4079 This is a synonym for sparc64-*-solaris2*.
4084 @heading @anchor{x-x-sysv}*-*-sysv*
4085 On System V release 3, you may get this error message
4089 ld fatal: failed to write symbol name @var{something}
4090 in strings table for file @var{whatever}
4093 This probably indicates that the disk is full or your ulimit won't allow
4094 the file to be as large as it needs to be.
4096 This problem can also result because the kernel parameter @code{MAXUMEM}
4097 is too small. If so, you must regenerate the kernel and make the value
4098 much larger. The default value is reported to be 1024; a value of 32768
4099 is said to work. Smaller values may also work.
4101 On System V, if you get an error like this,
4104 /usr/local/lib/bison.simple: In function `yyparse':
4105 /usr/local/lib/bison.simple:625: virtual memory exhausted
4109 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
4111 On a System V release 4 system, make sure @file{/usr/bin} precedes
4112 @file{/usr/ucb} in @code{PATH}. The @command{cc} command in
4113 @file{/usr/ucb} uses libraries which have bugs.
4118 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
4119 Don't try compiling with VAX C (@command{vcc}). It produces incorrect code
4120 in some cases (for example, when @code{alloca} is used).
4125 @heading @anchor{x-x-vxworks}*-*-vxworks*
4126 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4127 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4128 We welcome patches for other architectures supported by VxWorks 5.5.
4129 Support for VxWorks AE would also be welcome; we believe this is merely
4130 a matter of writing an appropriate ``configlette'' (see below). We are
4131 not interested in supporting older, a.out or COFF-based, versions of
4134 VxWorks comes with an older version of GCC installed in
4135 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4136 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4137 Before running @command{configure}, create the directories @file{@var{prefix}}
4138 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4139 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4140 include that directory while running both @command{configure} and
4143 You must give @command{configure} the
4144 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4145 find the VxWorks system headers. Since VxWorks is a cross compilation
4146 target only, you must also specify @option{--target=@var{target}}.
4147 @command{configure} will attempt to create the directory
4148 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4149 make sure the user running @command{configure} has sufficient privilege
4152 GCC's exception handling runtime requires a special ``configlette''
4153 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4154 that file to add the module to your kernel build. (Future versions of
4155 VxWorks will incorporate this module.)
4160 @heading @anchor{x86-64-x-x}x86_64-*-*, amd64-*-*
4162 GCC supports the x86-64 architecture implemented by the AMD64 processor
4163 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4164 On GNU/Linux the default is a bi-arch compiler which is able to generate
4165 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4170 @heading @anchor{xtensa-x-elf}xtensa-*-elf
4172 This target is intended for embedded Xtensa systems using the
4173 @samp{newlib} C library. It uses ELF but does not support shared
4174 objects. Designed-defined instructions specified via the
4175 Tensilica Instruction Extension (TIE) language are only supported
4176 through inline assembly.
4178 The Xtensa configuration information must be specified prior to
4179 building GCC@. The @file{include/xtensa-config.h} header
4180 file contains the configuration information. If you created your
4181 own Xtensa configuration with the Xtensa Processor Generator, the
4182 downloaded files include a customized copy of this header file,
4183 which you can use to replace the default header file.
4188 @heading @anchor{xtensa-x-linux}xtensa-*-linux*
4190 This target is for Xtensa systems running GNU/Linux. It supports ELF
4191 shared objects and the GNU C library (glibc). It also generates
4192 position-independent code (PIC) regardless of whether the
4193 @option{-fpic} or @option{-fPIC} options are used. In other
4194 respects, this target is the same as the
4195 @uref{#xtensa-*-elf,,@samp{xtensa-*-elf}} target.
4200 @heading @anchor{windows}Microsoft Windows (32-bit)
4202 Ports of GCC are included with the
4203 @uref{http://www.cygwin.com/,,Cygwin environment}.
4205 GCC will build under Cygwin without modification; it does not build
4206 with Microsoft's C++ compiler and there are no plans to make it do so.
4211 @heading @anchor{os2}OS/2
4213 GCC does not currently support OS/2. However, Andrew Zabolotny has been
4214 working on a generic OS/2 port with pgcc. The current code can be found
4215 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
4220 @heading @anchor{older}Older systems
4222 GCC contains support files for many older (1980s and early
4223 1990s) Unix variants. For the most part, support for these systems
4224 has not been deliberately removed, but it has not been maintained for
4225 several years and may suffer from bitrot.
4227 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
4228 Support for these systems is still present in that release, but
4229 @command{configure} will fail unless the @option{--enable-obsolete}
4230 option is given. Unless a maintainer steps forward, support for these
4231 systems will be removed from the next release of GCC@.
4233 Support for old systems as hosts for GCC can cause problems if the
4234 workarounds for compiler, library and operating system bugs affect the
4235 cleanliness or maintainability of the rest of GCC@. In some cases, to
4236 bring GCC up on such a system, if still possible with current GCC, may
4237 require first installing an old version of GCC which did work on that
4238 system, and using it to compile a more recent GCC, to avoid bugs in the
4239 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
4240 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
4241 sites}. Header bugs may generally be avoided using
4242 @command{fixincludes}, but bugs or deficiencies in libraries and the
4243 operating system may still cause problems.
4245 Support for older systems as targets for cross-compilation is less
4246 problematic than support for them as hosts for GCC; if an enthusiast
4247 wishes to make such a target work again (including resurrecting any of
4248 the targets that never worked with GCC 2, starting from the last
4249 version before they were removed), patches
4250 @uref{../contribute.html,,following the usual requirements} would be
4251 likely to be accepted, since they should not affect the support for more
4254 For some systems, old versions of GNU binutils may also be useful,
4255 and are available from @file{pub/binutils/old-releases} on
4256 @uref{http://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
4258 Some of the information on specific systems above relates to
4259 such older systems, but much of the information
4260 about GCC on such systems (which may no longer be applicable to
4261 current GCC) is to be found in the GCC texinfo manual.
4266 @heading @anchor{elf}all ELF targets (SVR4, Solaris 2, etc.)
4268 C++ support is significantly better on ELF targets if you use the
4269 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
4270 inlines, vtables and template instantiations will be discarded
4279 @uref{./index.html,,Return to the GCC Installation page}
4283 @c ***Old documentation******************************************************
4285 @include install-old.texi
4291 @uref{./index.html,,Return to the GCC Installation page}
4295 @c ***GFDL********************************************************************
4303 @uref{./index.html,,Return to the GCC Installation page}
4307 @c ***************************************************************************
4308 @c Part 6 The End of the Document
4310 @comment node-name, next, previous, up
4311 @node Concept Index, , GNU Free Documentation License, Top
4315 @unnumbered Concept Index