1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename gccinstall.info
5 @settitle Installing GCC
10 @c Specify title for specific html page
12 @settitle Installing GCC
15 @settitle Host/Target specific installation notes for GCC
17 @ifset prerequisiteshtml
18 @settitle Prerequisites for GCC
21 @settitle Downloading GCC
24 @settitle Installing GCC: Configuration
27 @settitle Installing GCC: Building
30 @settitle Installing GCC: Testing
32 @ifset finalinstallhtml
33 @settitle Installing GCC: Final installation
36 @settitle Installing GCC: Binaries
39 @settitle Installing GCC: Old documentation
42 @settitle Installing GCC: GNU Free Documentation License
45 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
46 @c 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
47 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
49 @c Include everything if we're not making html
53 @set prerequisiteshtml
64 @c Part 2 Summary Description and Copyright
66 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
67 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
69 Permission is granted to copy, distribute and/or modify this document
70 under the terms of the GNU Free Documentation License, Version 1.2 or
71 any later version published by the Free Software Foundation; with no
72 Invariant Sections, the Front-Cover texts being (a) (see below), and
73 with the Back-Cover Texts being (b) (see below). A copy of the
74 license is included in the section entitled ``@uref{./gfdl.html,,GNU
75 Free Documentation License}''.
77 (a) The FSF's Front-Cover Text is:
81 (b) The FSF's Back-Cover Text is:
83 You have freedom to copy and modify this GNU Manual, like GNU
84 software. Copies published by the Free Software Foundation raise
85 funds for GNU development.
90 @dircategory Programming
92 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
95 @c Part 3 Titlepage and Copyright
98 @comment The title is printed in a large font.
99 @center @titlefont{Installing GCC}
101 @c The following two commands start the copyright page.
103 @vskip 0pt plus 1filll
107 @c Part 4 Top node and Master Menu
110 @comment node-name, next, Previous, up
113 * Installing GCC:: This document describes the generic installation
114 procedure for GCC as well as detailing some target
115 specific installation instructions.
117 * Specific:: Host/target specific installation notes for GCC.
118 * Binaries:: Where to get pre-compiled binaries.
120 * Old:: Old installation documentation.
122 * GNU Free Documentation License:: How you can copy and share this manual.
123 * Concept Index:: This index has two entries.
127 @c Part 5 The Body of the Document
128 @c ***Installing GCC**********************************************************
130 @comment node-name, next, previous, up
131 @node Installing GCC, Binaries, , Top
135 @chapter Installing GCC
138 The latest version of this document is always available at
139 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
141 This document describes the generic installation procedure for GCC as well
142 as detailing some target specific installation instructions.
144 GCC includes several components that previously were separate distributions
145 with their own installation instructions. This document supersedes all
146 package specific installation instructions.
148 @emph{Before} starting the build/install procedure please check the
150 @ref{Specific, host/target specific installation notes}.
153 @uref{specific.html,,host/target specific installation notes}.
155 We recommend you browse the entire generic installation instructions before
158 Lists of successful builds for released versions of GCC are
159 available at @uref{http://gcc.gnu.org/buildstat.html}.
160 These lists are updated as new information becomes available.
162 The installation procedure itself is broken into five steps.
167 * Downloading the source::
170 * Testing:: (optional)
177 @uref{prerequisites.html,,Prerequisites}
179 @uref{download.html,,Downloading the source}
181 @uref{configure.html,,Configuration}
183 @uref{build.html,,Building}
185 @uref{test.html,,Testing} (optional)
187 @uref{finalinstall.html,,Final install}
191 Please note that GCC does not support @samp{make uninstall} and probably
192 won't do so in the near future as this would open a can of worms. Instead,
193 we suggest that you install GCC into a directory of its own and simply
194 remove that directory when you do not need that specific version of GCC
195 any longer, and, if shared libraries are installed there as well, no
196 more binaries exist that use them.
199 There are also some @uref{old.html,,old installation instructions},
200 which are mostly obsolete but still contain some information which has
201 not yet been merged into the main part of this manual.
209 @uref{./index.html,,Return to the GCC Installation page}
215 @c ***Prerequisites**************************************************
217 @comment node-name, next, previous, up
218 @node Prerequisites, Downloading the source, , Installing GCC
220 @ifset prerequisiteshtml
222 @chapter Prerequisites
224 @cindex Prerequisites
226 GCC requires that various tools and packages be available for use in the
227 build procedure. Modifying GCC sources requires additional tools
230 @heading Tools/packages necessary for building GCC
232 @item ISO C90 compiler
233 Necessary to bootstrap the GCC package, although versions of GCC prior
234 to 3.4 also allow bootstrapping with a traditional (K&R) C compiler.
236 To make all languages in a cross-compiler or other configuration where
237 3-stage bootstrap is not performed, you need to start with an existing
238 GCC binary (version 2.95 or later) because source code for language
239 frontends other than C might use GCC extensions.
243 In order to build the Ada compiler (GNAT) you must already have GNAT
244 installed because portions of the Ada frontend are written in Ada (with
245 GNAT extensions.) Refer to the Ada installation instructions for more
246 specific information.
248 @item A ``working'' POSIX compatible shell, or GNU bash
250 Necessary when running @command{configure} because some
251 @command{/bin/sh} shells have bugs and may crash when configuring the
252 target libraries. In other cases, @command{/bin/sh} or even some
253 @command{ksh} have disastrous corner-case performance problems. This
254 can cause target @command{configure} runs to literally take days to
255 complete in some cases.
257 So on some platforms @command{/bin/ksh} is sufficient, on others it
258 isn't. See the host/target specific instructions for your platform, or
259 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
260 environment to your ``good'' shell prior to running
261 @command{configure}/@command{make}.
263 @command{zsh} is not a fully compliant POSIX shell and will not
264 work when configuring GCC@.
268 Necessary in some circumstances, optional in others. See the
269 host/target specific instructions for your platform for the exact
272 @item gzip version 1.2.4 (or later) or
273 @itemx bzip2 version 1.0.2 (or later)
275 Necessary to uncompress GCC @command{tar} files when source code is
276 obtained via FTP mirror sites.
278 @item GNU make version 3.79.1 (or later)
280 You must have GNU make installed to build GCC@.
282 @item GNU tar version 1.12 (or later)
284 Necessary (only on some platforms) to untar the source code. Many
285 systems' @command{tar} programs will also work, only try GNU
286 @command{tar} if you have problems.
288 @item GNU Multiple Precision Library (GMP) version 4.0 (or later)
290 Necessary to build the Fortran frontend. If you don't have it
291 installed in your library search path, you will have to configure with
292 the @option{--with-gmp} or @option{--with-gmp-dir} configure option.
296 Necessary to build the Fortran frontend. It can be downloaded from
297 @uref{http://www.mpfr.org/}. It is also included in the current GMP
298 release (4.1.3) when configured with @option{--enable-mpfr}.
300 The @option{--with-mpfr} or @option{--with-mpfr-dir} configure option should
301 be used if your MPFR Library is not installed in your library search path.
306 @heading Tools/packages necessary for modifying GCC
308 @item autoconf versions 2.13 and 2.59
309 @itemx GNU m4 version 1.4 (or later)
311 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
312 to regenerate @file{configure} and @file{config.in} files. Most
313 directories require autoconf 2.59 (exactly), but the toplevel
314 still requires autoconf 2.13 (exactly).
316 @item automake versions 1.9.3
318 Necessary when modifying a @file{Makefile.am} file to regenerate its
319 associated @file{Makefile.in}.
321 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
322 file. Specifically this applies to the @file{gcc}, @file{intl},
323 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
324 as any of their subdirectories.
326 For directories that use automake, GCC requires the latest release in
327 the 1.9.x series, which is currently 1.9.3. When regenerating a directory
328 to a newer version, please update all the directories using an older 1.9.x
329 to the latest released version.
331 @item gettext version 0.12 (or later)
333 Needed to regenerate @file{gcc.pot}.
335 @item gperf version 2.7.2 (or later)
337 Necessary when modifying @command{gperf} input files, e.g.@:
338 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
339 @file{gcc/cp/cfns.h}.
341 @item expect version ???
342 @itemx tcl version ???
343 @itemx dejagnu version 1.4.4 (or later)
345 Necessary to run the GCC testsuite.
347 @item autogen version 5.5.4 (or later) and
348 @itemx guile version 1.4.1 (or later)
350 Necessary to regenerate @file{fixinc/fixincl.x} from
351 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
353 Necessary to run the @file{fixinc} @command{make check}.
355 Necessary to regenerate the top level @file{Makefile.in} file from
356 @file{Makefile.tpl} and @file{Makefile.def}.
358 @item GNU Bison version 1.28 (or later)
359 Berkeley @command{yacc} (@command{byacc}) is also reported to work other
362 Necessary when modifying @file{*.y} files.
364 Necessary to build GCC during development because the generated output
365 files are not included in the CVS repository. They are included in
368 @item Flex version 2.5.4 (or later)
370 Necessary when modifying @file{*.l} files.
372 Necessary to build GCC during development because the generated output
373 files are not included in the CVS repository. They are included in
376 @item Texinfo version 4.2 (or later)
378 Necessary for running @command{makeinfo} when modifying @file{*.texi}
379 files to test your changes.
381 Necessary to build GCC documentation during development because the
382 generated output files are not included in the CVS repository. They are
383 included in releases.
385 @item @TeX{} (any working version)
387 Necessary for running @command{texi2dvi}, used when running
388 @command{make dvi} to create DVI files.
390 @item cvs version 1.10 (or later)
391 @itemx ssh (any version)
393 Necessary to access the CVS repository. Public releases and weekly
394 snapshots of the development sources are also available via FTP@.
396 @item perl version 5.6.1 (or later)
398 Necessary when regenerating @file{Makefile} dependencies in libiberty.
399 Necessary when regenerating @file{libiberty/functions.texi}.
400 Necessary when generating manpages from Texinfo manuals.
401 Used by various scripts to generate some files included in CVS (mainly
402 Unicode-related and rarely changing) from source tables.
404 @item GNU diffutils version 2.7 (or later)
406 Necessary when creating changes to GCC source code to submit for review.
408 @item patch version 2.5.4 (or later)
410 Necessary when applying patches, created with @command{diff}, to one's
420 @uref{./index.html,,Return to the GCC Installation page}
424 @c ***Downloading the source**************************************************
426 @comment node-name, next, previous, up
427 @node Downloading the source, Configuration, Prerequisites, Installing GCC
431 @chapter Downloading GCC
433 @cindex Downloading GCC
434 @cindex Downloading the Source
436 GCC is distributed via @uref{http://gcc.gnu.org/cvs.html,,CVS} and FTP
437 tarballs compressed with @command{gzip} or
438 @command{bzip2}. It is possible to download a full distribution or specific
441 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
442 for information on how to obtain GCC@.
444 The full distribution includes the C, C++, Objective-C, Fortran 77, Fortran
445 (in case of GCC 3.5 and later), Java, and Ada (in case of GCC 3.1 and later)
446 compilers. The full distribution also includes runtime libraries for C++,
447 Objective-C, Fortran 77, Fortran, and Java. In GCC 3.0 and later versions,
448 GNU compiler testsuites are also included in the full distribution.
450 If you choose to download specific components, you must download the core
451 GCC distribution plus any language specific distributions you wish to
452 use. The core distribution includes the C language front end as well as the
453 shared components. Each language has a tarball which includes the language
454 front end as well as the language runtime (when appropriate).
456 Unpack the core distribution as well as any language specific
457 distributions in the same directory.
459 If you also intend to build binutils (either to upgrade an existing
460 installation or for use in place of the corresponding tools of your
461 OS), unpack the binutils distribution either in the same directory or
462 a separate one. In the latter case, add symbolic links to any
463 components of the binutils you intend to build alongside the compiler
464 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
465 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
472 @uref{./index.html,,Return to the GCC Installation page}
476 @c ***Configuration***********************************************************
478 @comment node-name, next, previous, up
479 @node Configuration, Building, Downloading the source, Installing GCC
483 @chapter Installing GCC: Configuration
485 @cindex Configuration
486 @cindex Installing GCC: Configuration
488 Like most GNU software, GCC must be configured before it can be built.
489 This document describes the recommended configuration procedure
490 for both native and cross targets.
492 We use @var{srcdir} to refer to the toplevel source directory for
493 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
495 If you obtained the sources via CVS, @var{srcdir} must refer to the top
496 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
497 and not its @file{gcc} subdirectory, otherwise the build will fail.
499 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
500 file system, the shell's built-in @command{pwd} command will return
501 temporary pathnames. Using these can lead to various sorts of build
502 problems. To avoid this issue, set the @env{PWDCMD} environment
503 variable to an automounter-aware @command{pwd} command, e.g.,
504 @command{pawd} or @samp{amq -w}, during the configuration and build
507 First, we @strong{highly} recommend that GCC be built into a
508 separate directory than the sources which does @strong{not} reside
509 within the source tree. This is how we generally build GCC; building
510 where @var{srcdir} == @var{objdir} should still work, but doesn't
511 get extensive testing; building where @var{objdir} is a subdirectory
512 of @var{srcdir} is unsupported.
514 If you have previously built GCC in the same directory for a
515 different target machine, do @samp{make distclean} to delete all files
516 that might be invalid. One of the files this deletes is @file{Makefile};
517 if @samp{make distclean} complains that @file{Makefile} does not exist
518 or issues a message like ``don't know how to make distclean'' it probably
519 means that the directory is already suitably clean. However, with the
520 recommended method of building in a separate @var{objdir}, you should
521 simply use a different @var{objdir} for each target.
523 Second, when configuring a native system, either @command{cc} or
524 @command{gcc} must be in your path or you must set @env{CC} in
525 your environment before running configure. Otherwise the configuration
528 Note that the bootstrap compiler and the resulting GCC must be link
529 compatible, else the bootstrap will fail with linker errors about
530 incompatible object file formats. Several multilibed targets are
531 affected by this requirement, see
533 @ref{Specific, host/target specific installation notes}.
536 @uref{specific.html,,host/target specific installation notes}.
544 % @var{srcdir}/configure [@var{options}] [@var{target}]
548 @heading Target specification
551 GCC has code to correctly determine the correct value for @var{target}
552 for nearly all native systems. Therefore, we highly recommend you not
553 provide a configure target when configuring a native compiler.
556 @var{target} must be specified as @option{--target=@var{target}}
557 when configuring a cross compiler; examples of valid targets would be
558 m68k-coff, sh-elf, etc.
561 Specifying just @var{target} instead of @option{--target=@var{target}}
562 implies that the host defaults to @var{target}.
566 @heading Options specification
568 Use @var{options} to override several configure time options for
569 GCC@. A list of supported @var{options} follows; @samp{configure
570 --help} may list other options, but those not listed below may not
571 work and should not normally be used.
573 Note that each @option{--enable} option has a corresponding
574 @option{--disable} option and that each @option{--with} option has a
575 corresponding @option{--without} option.
578 @item --prefix=@var{dirname}
579 Specify the toplevel installation
580 directory. This is the recommended way to install the tools into a directory
581 other than the default. The toplevel installation directory defaults to
584 We @strong{highly} recommend against @var{dirname} being the same or a
585 subdirectory of @var{objdir} or vice versa. If specifying a directory
586 beneath a user's home directory tree, some shells will not expand
587 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
590 The following standard @command{autoconf} options are supported. Normally you
591 should not need to use these options.
593 @item --exec-prefix=@var{dirname}
594 Specify the toplevel installation directory for architecture-dependent
595 files. The default is @file{@var{prefix}}.
597 @item --bindir=@var{dirname}
598 Specify the installation directory for the executables called by users
599 (such as @command{gcc} and @command{g++}). The default is
600 @file{@var{exec-prefix}/bin}.
602 @item --libdir=@var{dirname}
603 Specify the installation directory for object code libraries and
604 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
606 @item --libexecdir=@var{dirname}
607 Specify the installation directory for internal executables of GCC@.
608 The default is @file{@var{exec-prefix}/libexec}.
610 @item --with-slibdir=@var{dirname}
611 Specify the installation directory for the shared libgcc library. The
612 default is @file{@var{libdir}}.
614 @item --infodir=@var{dirname}
615 Specify the installation directory for documentation in info format.
616 The default is @file{@var{prefix}/info}.
618 @item --datadir=@var{dirname}
619 Specify the installation directory for some architecture-independent
620 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
622 @item --mandir=@var{dirname}
623 Specify the installation directory for manual pages. The default is
624 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
625 the full GCC manuals, which are provided in Texinfo format. The manpages
626 are derived by an automatic conversion process from parts of the full
629 @item --with-gxx-include-dir=@var{dirname}
631 the installation directory for G++ header files. The default is
632 @file{@var{prefix}/include/c++/@var{version}}.
636 @item --program-prefix=@var{prefix}
637 GCC supports some transformations of the names of its programs when
638 installing them. This option prepends @var{prefix} to the names of
639 programs to install in @var{bindir} (see above). For example, specifying
640 @option{--program-prefix=foo-} would result in @samp{gcc}
641 being installed as @file{/usr/local/bin/foo-gcc}.
643 @item --program-suffix=@var{suffix}
644 Appends @var{suffix} to the names of programs to install in @var{bindir}
645 (see above). For example, specifying @option{--program-suffix=-3.1}
646 would result in @samp{gcc} being installed as
647 @file{/usr/local/bin/gcc-3.1}.
649 @item --program-transform-name=@var{pattern}
650 Applies the @samp{sed} script @var{pattern} to be applied to the names
651 of programs to install in @var{bindir} (see above). @var{pattern} has to
652 consist of one or more basic @samp{sed} editing commands, separated by
653 semicolons. For example, if you want the @samp{gcc} program name to be
654 transformed to the installed program @file{/usr/local/bin/myowngcc} and
655 the @samp{g++} program name to be transformed to
656 @file{/usr/local/bin/gspecial++} without changing other program names,
657 you could use the pattern
658 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
659 to achieve this effect.
661 All three options can be combined and used together, resulting in more
662 complex conversion patterns. As a basic rule, @var{prefix} (and
663 @var{suffix}) are prepended (appended) before further transformations
664 can happen with a special transformation script @var{pattern}.
666 As currently implemented, this option only takes effect for native
667 builds; cross compiler binaries' names are not transformed even when a
668 transformation is explicitly asked for by one of these options.
670 For native builds, some of the installed programs are also installed
671 with the target alias in front of their name, as in
672 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
673 before the target alias is prepended to the name---so, specifying
674 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
675 resulting binary would be installed as
676 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
678 As a last shortcoming, none of the installed Ada programs are
679 transformed yet, which will be fixed in some time.
681 @item --with-local-prefix=@var{dirname}
683 installation directory for local include files. The default is
684 @file{/usr/local}. Specify this option if you want the compiler to
685 search directory @file{@var{dirname}/include} for locally installed
686 header files @emph{instead} of @file{/usr/local/include}.
688 You should specify @option{--with-local-prefix} @strong{only} if your
689 site has a different convention (not @file{/usr/local}) for where to put
692 The default value for @option{--with-local-prefix} is @file{/usr/local}
693 regardless of the value of @option{--prefix}. Specifying
694 @option{--prefix} has no effect on which directory GCC searches for
695 local header files. This may seem counterintuitive, but actually it is
698 The purpose of @option{--prefix} is to specify where to @emph{install
699 GCC}. The local header files in @file{/usr/local/include}---if you put
700 any in that directory---are not part of GCC@. They are part of other
701 programs---perhaps many others. (GCC installs its own header files in
702 another directory which is based on the @option{--prefix} value.)
704 Both the local-prefix include directory and the GCC-prefix include
705 directory are part of GCC's ``system include'' directories. Although these
706 two directories are not fixed, they need to be searched in the proper
707 order for the correct processing of the include_next directive. The
708 local-prefix include directory is searched before the GCC-prefix
709 include directory. Another characteristic of system include directories
710 is that pedantic warnings are turned off for headers in these directories.
712 Some autoconf macros add @option{-I @var{directory}} options to the
713 compiler command line, to ensure that directories containing installed
714 packages' headers are searched. When @var{directory} is one of GCC's
715 system include directories, GCC will ignore the option so that system
716 directories continue to be processed in the correct order. This
717 may result in a search order different from what was specified but the
718 directory will still be searched.
720 GCC automatically searches for ordinary libraries using
721 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
722 used for both GCC and packages, GCC will automatically search for
723 both headers and libraries. This provides a configuration that is
724 easy to use. GCC behaves in a manner similar to that when it is
725 installed as a system compiler in @file{/usr}.
727 Sites that need to install multiple versions of GCC may not want to
728 use the above simple configuration. It is possible to use the
729 @option{--program-prefix}, @option{--program-suffix} and
730 @option{--program-transform-name} options to install multiple versions
731 into a single directory, but it may be simpler to use different prefixes
732 and the @option{--with-local-prefix} option to specify the location of the
733 site-specific files for each version. It will then be necessary for
734 users to specify explicitly the location of local site libraries
735 (e.g., with @env{LIBRARY_PATH}).
737 The same value can be used for both @option{--with-local-prefix} and
738 @option{--prefix} provided it is not @file{/usr}. This can be used
739 to avoid the default search of @file{/usr/local/include}.
741 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
742 The directory you use for @option{--with-local-prefix} @strong{must not}
743 contain any of the system's standard header files. If it did contain
744 them, certain programs would be miscompiled (including GNU Emacs, on
745 certain targets), because this would override and nullify the header
746 file corrections made by the @command{fixincludes} script.
748 Indications are that people who use this option use it based on mistaken
749 ideas of what it is for. People use it as if it specified where to
750 install part of GCC@. Perhaps they make this assumption because
751 installing GCC creates the directory.
753 @item --enable-shared[=@var{package}[,@dots{}]]
754 Build shared versions of libraries, if shared libraries are supported on
755 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
756 are enabled by default on all platforms that support shared libraries.
758 If a list of packages is given as an argument, build shared libraries
759 only for the listed packages. For other packages, only static libraries
760 will be built. Package names currently recognized in the GCC tree are
761 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
762 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
763 @samp{ada}, @samp{libada}, @samp{libjava} and @samp{libobjc}.
764 Note @samp{libiberty} does not support shared libraries at all.
766 Use @option{--disable-shared} to build only static libraries. Note that
767 @option{--disable-shared} does not accept a list of package names as
768 argument, only @option{--enable-shared} does.
770 @item @anchor{with-gnu-as}--with-gnu-as
771 Specify that the compiler should assume that the
772 assembler it finds is the GNU assembler. However, this does not modify
773 the rules to find an assembler and will result in confusion if the
774 assembler found is not actually the GNU assembler. (Confusion may also
775 result if the compiler finds the GNU assembler but has not been
776 configured with @option{--with-gnu-as}.) If you have more than one
777 assembler installed on your system, you may want to use this option in
778 connection with @option{--with-as=@var{pathname}}.
780 The following systems are the only ones where it makes a difference
781 whether you use the GNU assembler. On any other system,
782 @option{--with-gnu-as} has no effect.
785 @item @samp{hppa1.0-@var{any}-@var{any}}
786 @item @samp{hppa1.1-@var{any}-@var{any}}
787 @item @samp{i386-@var{any}-sysv}
788 @item @samp{m68k-bull-sysv}
789 @item @samp{m68k-hp-hpux}
790 @item @samp{m68000-hp-hpux}
791 @item @samp{m68000-att-sysv}
792 @item @samp{sparc-sun-solaris2.@var{any}}
793 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
796 On the systems listed above (except for the HP-PA, the SPARC, for ISC on
797 the 386, if you use the GNU assembler, you should also use the GNU linker
798 (and specify @option{--with-gnu-ld}).
800 @item @anchor{with-as}--with-as=@var{pathname}
802 compiler should use the assembler pointed to by @var{pathname}, rather
803 than the one found by the standard rules to find an assembler, which
807 Check the @file{@var{libexec}/gcc/@var{target}/@var{version}}
808 directory, where @var{libexec} defaults to
809 @file{@var{exec-prefix}/libexec} and @var{exec-prefix} defaults to
810 @var{prefix} which defaults to @file{/usr/local} unless overridden by
811 the @option{--prefix=@var{pathname}} switch described
812 above. @var{target} is the target system triple, such as
813 @samp{sparc-sun-solaris2.7}, and @var{version} denotes the GCC
814 version, such as 3.0.
816 Check operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
819 Note that these rules do not check for the value of @env{PATH}. You may
820 want to use @option{--with-as} if no assembler is installed in the
821 directories listed above, or if you have multiple assemblers installed
822 and want to choose one that is not found by the above rules.
824 @item @anchor{with-gnu-ld}--with-gnu-ld
825 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
828 @item --with-ld=@var{pathname}
829 Same as @uref{#with-as,,@option{--with-as}}
833 Specify that stabs debugging
834 information should be used instead of whatever format the host normally
835 uses. Normally GCC uses the same debug format as the host system.
837 On MIPS based systems and on Alphas, you must specify whether you want
838 GCC to create the normal ECOFF debugging format, or to use BSD-style
839 stabs passed through the ECOFF symbol table. The normal ECOFF debug
840 format cannot fully handle languages other than C@. BSD stabs format can
841 handle other languages, but it only works with the GNU debugger GDB@.
843 Normally, GCC uses the ECOFF debugging format by default; if you
844 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
846 No matter which default you choose when you configure GCC, the user
847 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
848 the debug format for a particular compilation.
850 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
851 @option{--with-gas} is used. It selects use of stabs debugging
852 information embedded in COFF output. This kind of debugging information
853 supports C++ well; ordinary COFF debugging information does not.
855 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
856 selects use of stabs debugging information embedded in ELF output. The
857 C++ compiler currently (2.6.0) does not support the DWARF debugging
858 information normally used on 386 SVR4 platforms; stabs provide a
859 workable alternative. This requires gas and gdb, as the normal SVR4
860 tools can not generate or interpret stabs.
862 @item --disable-multilib
863 Specify that multiple target
864 libraries to support different target variants, calling
865 conventions, etc should not be built. The default is to build a
866 predefined set of them.
868 Some targets provide finer-grained control over which multilibs are built
869 (e.g., @option{--disable-softfloat}):
875 fpu, 26bit, underscore, interwork, biendian, nofmult.
878 softfloat, m68881, m68000, m68020.
881 single-float, biendian, softfloat.
883 @item powerpc*-*-*, rs6000*-*-*
884 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
889 @item --enable-threads
890 Specify that the target
891 supports threads. This affects the Objective-C compiler and runtime
892 library, and exception handling for other languages like C++ and Java.
893 On some systems, this is the default.
895 In general, the best (and, in many cases, the only known) threading
896 model available will be configured for use. Beware that on some
897 systems, GCC has not been taught what threading models are generally
898 available for the system. In this case, @option{--enable-threads} is an
899 alias for @option{--enable-threads=single}.
901 @item --disable-threads
902 Specify that threading support should be disabled for the system.
903 This is an alias for @option{--enable-threads=single}.
905 @item --enable-threads=@var{lib}
907 @var{lib} is the thread support library. This affects the Objective-C
908 compiler and runtime library, and exception handling for other languages
909 like C++ and Java. The possibilities for @var{lib} are:
917 Ada tasking support. For non-Ada programs, this setting is equivalent
918 to @samp{single}. When used in conjunction with the Ada run time, it
919 causes GCC to use the same thread primitives as Ada uses. This option
920 is necessary when using both Ada and the back end exception handling,
921 which is the default for most Ada targets.
923 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
924 that the file needed to support this configuration, @file{gthr-mach.h}, is
925 missing and thus this setting will cause a known bootstrap failure.)
927 This is an alias for @samp{single}.
929 Generic POSIX/Unix98 thread support.
931 Generic POSIX/Unix95 thread support.
933 RTEMS thread support.
935 Disable thread support, should work for all platforms.
937 Sun Solaris 2 thread support.
939 VxWorks thread support.
941 Microsoft Win32 API thread support.
943 Novell Kernel Services thread support.
946 @item --with-cpu=@var{cpu}
947 Specify which cpu variant the compiler should generate code for by default.
948 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
949 This option is only supported on some targets, including ARM, i386, PowerPC,
952 @item --with-schedule=@var{cpu}
953 @itemx --with-arch=@var{cpu}
954 @itemx --with-tune=@var{cpu}
955 @itemx --with-abi=@var{abi}
956 @itemx --with-fpu=@var{type}
957 @itemx --with-float=@var{type}
958 These configure options provide default values for the @option{-mschedule=},
959 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
960 options and for @option{-mhard-float} or @option{-msoft-float}. As with
961 @option{--with-cpu}, which switches will be accepted and acceptable values
962 of the arguments depend on the target.
964 @item --with-divide=@var{type}
965 Specify how the compiler should generate code for checking for
966 division by zero. This option is only supported on the MIPS target.
967 The possibilities for @var{type} are:
970 Division by zero checks use conditional traps (this is the default on
971 systems that support conditional traps).
973 Division by zero checks use the break instruction.
976 @item --enable-__cxa_atexit
977 Define if you want to use __cxa_atexit, rather than atexit, to
978 register C++ destructors for local statics and global objects.
979 This is essential for fully standards-compliant handling of
980 destructors, but requires __cxa_atexit in libc. This option is currently
981 only available on systems with GNU libc. When enabled, this will cause
982 @option{-fuse-cxa-exit} to be passed by default.
984 @item --enable-target-optspace
986 libraries should be optimized for code space instead of code speed.
987 This is the default for the m32r platform.
990 Specify that a user visible @command{cpp} program should not be installed.
992 @item --with-cpp-install-dir=@var{dirname}
993 Specify that the user visible @command{cpp} program should be installed
994 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
996 @item --enable-initfini-array
997 Force the use of sections @code{.init_array} and @code{.fini_array}
998 (instead of @code{.init} and @code{.fini}) for constructors and
999 destructors. Option @option{--disable-initfini-array} has the
1000 opposite effect. If neither option is specified, the configure script
1001 will try to guess whether the @code{.init_array} and
1002 @code{.fini_array} sections are supported and, if they are, use them.
1004 @item --enable-maintainer-mode
1005 The build rules that
1006 regenerate the GCC master message catalog @file{gcc.pot} are normally
1007 disabled. This is because it can only be rebuilt if the complete source
1008 tree is present. If you have changed the sources and want to rebuild the
1009 catalog, configuring with @option{--enable-maintainer-mode} will enable
1010 this. Note that you need a recent version of the @code{gettext} tools
1013 @item --enable-generated-files-in-srcdir
1014 Neither the .c and .h files that are generated from bison and flex nor the
1015 info manuals and man pages that are built from the .texi files are present
1016 in the CVS development tree. When building GCC from that development tree,
1017 or from a snapshot which are created from CVS, then those generated files
1018 are placed in your build directory, which allows for the source to be in a
1021 If you configure with @option{--enable-generated-files-in-srcdir} then those
1022 generated files will go into the source directory. This is mainly intended
1023 for generating release or prerelease tarballs of the GCC sources, since it
1024 is not a requirement that the users of source releases to have flex, bison, or
1027 @item --enable-version-specific-runtime-libs
1029 that runtime libraries should be installed in the compiler specific
1030 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1031 addition, @samp{libstdc++}'s include files will be installed into
1032 @file{@var{libdir}} unless you overruled it by using
1033 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1034 particularly useful if you intend to use several versions of GCC in
1035 parallel. This is currently supported by @samp{libgfortran},
1036 @samp{libjava}, @samp{libmudflap}, @samp{libstdc++}, and @samp{libobjc}.
1039 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1040 Specify that only a particular subset of compilers and
1041 their runtime libraries should be built. For a list of valid values for
1042 @var{langN} you can issue the following command in the
1043 @file{gcc} directory of your GCC source tree:@*
1045 grep language= */config-lang.in
1047 Currently, you can use any of the following:
1048 @code{ada}, @code{c}, @code{c++}, @code{f77}, @code{f95}, @code{java},
1049 @code{objc}, @code{obj-c++}.
1050 Building the Ada compiler has special requirements, see below.@*
1051 If you do not pass this flag, all languages available in the @file{gcc}
1052 sub-tree will be configured. Re-defining @code{LANGUAGES} when calling
1053 @samp{make bootstrap} @strong{does not} work anymore, as those
1054 language sub-directories might not have been configured!
1056 @item --disable-libada
1057 Specify that the run-time libraries and tools used by GNAT should not
1058 be built. This can be useful for debugging, or for compatibility with
1059 previous Ada build procedures, when it was required to explicitly
1060 do a @samp{make -C gcc gnatlib_and_tools}.
1063 Specify that the compiler should
1064 use DWARF 2 debugging information as the default.
1066 @item --enable-win32-registry
1067 @itemx --enable-win32-registry=@var{key}
1068 @itemx --disable-win32-registry
1069 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1070 to look up installations paths in the registry using the following key:
1073 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1076 @var{key} defaults to GCC version number, and can be overridden by the
1077 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1078 who use custom installers are encouraged to provide a different key,
1079 perhaps one comprised of vendor name and GCC version number, to
1080 avoid conflict with existing installations. This feature is enabled
1081 by default, and can be disabled by @option{--disable-win32-registry}
1082 option. This option has no effect on the other hosts.
1085 Specify that the machine does not have a floating point unit. This
1086 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1087 system, @option{--nfp} has no effect.
1089 @item --enable-werror
1090 @itemx --disable-werror
1091 @itemx --enable-werror=yes
1092 @itemx --enable-werror=no
1093 When you specify this option, it controls whether certain files in the
1094 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1095 If you don't specify it, @option{-Werror} is turned on for the main
1096 development trunk. However it defaults to off for release branches and
1097 final releases. The specific files which get @option{-Werror} are
1098 controlled by the Makefiles.
1100 @item --enable-checking
1101 @itemx --enable-checking=@var{list}
1102 When you specify this option, the compiler is built to perform checking
1103 of tree node types when referencing fields of that node, and some other
1104 internal consistency checks. This does not change the generated code,
1105 but adds error checking within the compiler. This will slow down the
1106 compiler and may only work properly if you are building the compiler
1107 with GCC@. This is on by default when building from CVS or snapshots,
1108 but off for releases. More control over the checks may be had by
1109 specifying @var{list}; the categories of checks available are
1110 @samp{release}, @samp{assert}, @samp{misc}, @samp{tree}, @samp{gc},
1111 @samp{rtl}, @samp{rtlflag}, @samp{runtime}, @samp{fold}, @samp{gcac} and
1112 @samp{valgrind}. The @samp{release} category enables only those checks
1113 suitable for release builds, currently this is @samp{assert} and
1114 @samp{runtime}. The check @samp{valgrind} requires the external
1115 @command{valgrind} simulator, available from
1116 @uref{http://valgrind.kde.org/}. The checks @samp{rtl}, @samp{gcac} and
1117 @samp{valgrind} are very expensive. The default when @var{list} is not
1118 specified is @samp{assert,misc,tree,gc,rtlflag,runtime}. That is also
1119 the default for development builds, when @samp{--enable-checking} is not
1120 specified. For release builds the default, when
1121 @samp{--enable-checking} is not given, is @samp{release}. To disable
1122 all checking, @samp{--disable-checking} must be explicitly requested.
1123 Disabling assertions will make the compiler and runtime slightly faster
1124 but increase the risk of undetected internal errors causing wrong code
1127 @item --enable-coverage
1128 @itemx --enable-coverage=@var{level}
1129 With this option, the compiler is built to collect self coverage
1130 information, every time it is run. This is for internal development
1131 purposes, and only works when the compiler is being built with gcc. The
1132 @var{level} argument controls whether the compiler is built optimized or
1133 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1134 want to disable optimization, for performance analysis you want to
1135 enable optimization. When coverage is enabled, the default level is
1136 without optimization.
1138 @item --enable-gather-detailed-mem-stats
1139 When this option is specified more detailed information on memory
1140 allocation is gathered. This information is printed when using
1141 @option{-fmem-report}.
1144 @itemx --with-gc=@var{choice}
1145 With this option you can specify the garbage collector implementation
1146 used during the compilation process. @var{choice} can be one of
1147 @samp{page} and @samp{zone}, where @samp{page} is the default.
1150 @itemx --disable-nls
1151 The @option{--enable-nls} option enables Native Language Support (NLS),
1152 which lets GCC output diagnostics in languages other than American
1153 English. Native Language Support is enabled by default if not doing a
1154 canadian cross build. The @option{--disable-nls} option disables NLS@.
1156 @item --with-included-gettext
1157 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1158 procedure to prefer its copy of GNU @command{gettext}.
1160 @item --with-catgets
1161 If NLS is enabled, and if the host lacks @code{gettext} but has the
1162 inferior @code{catgets} interface, the GCC build procedure normally
1163 ignores @code{catgets} and instead uses GCC's copy of the GNU
1164 @code{gettext} library. The @option{--with-catgets} option causes the
1165 build procedure to use the host's @code{catgets} in this situation.
1167 @item --with-libiconv-prefix=@var{dir}
1168 Search for libiconv header files in @file{@var{dir}/include} and
1169 libiconv library files in @file{@var{dir}/lib}.
1171 @item --enable-obsolete
1172 Enable configuration for an obsoleted system. If you attempt to
1173 configure GCC for a system (build, host, or target) which has been
1174 obsoleted, and you do not specify this flag, configure will halt with an
1177 All support for systems which have been obsoleted in one release of GCC
1178 is removed entirely in the next major release, unless someone steps
1179 forward to maintain the port.
1182 @subheading Cross-Compiler-Specific Options
1183 The following options only apply to building cross compilers.
1185 @item --with-sysroot
1186 @itemx --with-sysroot=@var{dir}
1187 Tells GCC to consider @var{dir} as the root of a tree that contains a
1188 (subset of) the root filesystem of the target operating system.
1189 Target system headers, libraries and run-time object files will be
1190 searched in there. The specified directory is not copied into the
1191 install tree, unlike the options @option{--with-headers} and
1192 @option{--with-libs} that this option obsoletes. The default value,
1193 in case @option{--with-sysroot} is not given an argument, is
1194 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1195 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1196 the GCC binaries if the installation tree is moved.
1198 @item --with-headers
1199 @itemx --with-headers=@var{dir}
1200 Deprecated in favor of @option{--with-sysroot}.
1201 Specifies that target headers are available when building a cross compiler.
1202 The @var{dir} argument specifies a directory which has the target include
1203 files. These include files will be copied into the @file{gcc} install
1204 directory. @emph{This option with the @var{dir} argument is required} when
1205 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1206 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1207 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1208 will be run on these files to make them compatible with GCC@.
1210 @item --without-headers
1211 Tells GCC not use any target headers from a libc when building a cross
1212 compiler. When crossing to GNU/Linux, you need the headers so GCC
1213 can build the exception handling for libgcc.
1214 See @uref{http://www.objsw.com/CrossGCC/,,CrossGCC} for more information
1218 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1219 Deprecated in favor of @option{--with-sysroot}.
1220 Specifies a list of directories which contain the target runtime
1221 libraries. These libraries will be copied into the @file{gcc} install
1222 directory. If the directory list is omitted, this option has no
1225 Specifies that @samp{newlib} is
1226 being used as the target C library. This causes @code{__eprintf} to be
1227 omitted from @file{libgcc.a} on the assumption that it will be provided by
1231 @subheading Fortran-specific Option
1233 The following options apply to the build of the Fortran front end.
1237 @item --with-gmp=@var{pathname}
1238 @itemx --with-mpfr=@var{pathname}
1239 @itemx --with-gmp-dir=@var{pathname}
1240 @itemx --with-mpfr-dir=@var{pathname}
1241 If you don't have GMP (the GNU Multiple Precision library) and the MPFR
1242 Libraries installed in a standard location and you want to build the Fortran
1243 front-end, you can explicitly specify the directory where they are installed
1244 (@samp{--with-gmp=gmpinstalldir}, @samp{--with-mpfr=mpfrinstalldir}) or where
1245 you built them without installing (@samp{--with-gmp-dir=gmpbuilddir},
1246 @samp{--with-mpfr-dir=gmpbuilddir}).
1250 @subheading Java-Specific Options
1252 The following option applies to the build of the Java front end.
1255 @item --disable-libgcj
1256 Specify that the run-time libraries
1257 used by GCJ should not be built. This is useful in case you intend
1258 to use GCJ with some other run-time, or you're going to install it
1259 separately, or it just happens not to build on your particular
1260 machine. In general, if the Java front end is enabled, the GCJ
1261 libraries will be enabled too, unless they're known to not work on
1262 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1263 may need to port it; in this case, before modifying the top-level
1264 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1265 you may use @option{--enable-libgcj} to override the default.
1269 The following options apply to building @samp{libgcj}.
1271 @subsubheading General Options
1274 @item --disable-getenv-properties
1275 Don't set system properties from @env{GCJ_PROPERTIES}.
1277 @item --enable-hash-synchronization
1278 Use a global hash table for monitor locks. Ordinarily,
1279 @samp{libgcj}'s @samp{configure} script automatically makes
1280 the correct choice for this option for your platform. Only use
1281 this if you know you need the library to be configured differently.
1283 @item --enable-interpreter
1284 Enable the Java interpreter. The interpreter is automatically
1285 enabled by default on all platforms that support it. This option
1286 is really only useful if you want to disable the interpreter
1287 (using @option{--disable-interpreter}).
1289 @item --disable-java-net
1290 Disable java.net. This disables the native part of java.net only,
1291 using non-functional stubs for native method implementations.
1293 @item --disable-jvmpi
1294 Disable JVMPI support.
1297 Enable runtime eCos target support.
1299 @item --without-libffi
1300 Don't use @samp{libffi}. This will disable the interpreter and JNI
1301 support as well, as these require @samp{libffi} to work.
1303 @item --enable-libgcj-debug
1304 Enable runtime debugging code.
1306 @item --enable-libgcj-multifile
1307 If specified, causes all @file{.java} source files to be
1308 compiled into @file{.class} files in one invocation of
1309 @samp{gcj}. This can speed up build time, but is more
1310 resource-intensive. If this option is unspecified or
1311 disabled, @samp{gcj} is invoked once for each @file{.java}
1312 file to compile into a @file{.class} file.
1314 @item --with-libiconv-prefix=DIR
1315 Search for libiconv in @file{DIR/include} and @file{DIR/lib}.
1317 @item --enable-sjlj-exceptions
1318 Force use of @code{builtin_setjmp} for exceptions. @samp{configure}
1319 ordinarily picks the correct value based on the platform. Only use
1320 this option if you are sure you need a different setting.
1322 @item --with-system-zlib
1323 Use installed @samp{zlib} rather than that included with GCC@.
1325 @item --with-win32-nlsapi=ansi, unicows or unicode
1326 Indicates how MinGW @samp{libgcj} translates between UNICODE
1327 characters and the Win32 API@.
1330 Use the single-byte @code{char} and the Win32 A functions natively,
1331 translating to and from UNICODE when using these functions. If
1332 unspecified, this is the default.
1335 Use the @code{WCHAR} and Win32 W functions natively. Adds
1336 @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}.
1337 @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines
1338 running built executables. @file{libunicows.a}, an open-source
1339 import library around Microsoft's @code{unicows.dll}, is obtained from
1340 @uref{http://libunicows.sourceforge.net/}, which also gives details
1341 on getting @file{unicows.dll} from Microsoft.
1344 Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not}
1345 add @code{-lunicows} to @file{libgcj.spec}. The built executables will
1346 only run on Microsoft Windows NT and above.
1350 @subsubheading AWT-Specific Options
1354 Use the X Window System.
1356 @item --enable-java-awt=PEER(S)
1357 Specifies the AWT peer library or libraries to build alongside
1358 @samp{libgcj}. If this option is unspecified or disabled, AWT
1359 will be non-functional. Current valid values are @option{gtk} and
1360 @option{xlib}. Multiple libraries should be separated by a
1361 comma (i.e.@: @option{--enable-java-awt=gtk,xlib}).
1363 @item --enable-gtk-cairo
1364 Build the cairo Graphics2D implementation on GTK@.
1366 @item --enable-java-gc=TYPE
1367 Choose garbage collector. Defaults to @option{boehm} if unspecified.
1369 @item --disable-gtktest
1370 Do not try to compile and run a test GTK+ program.
1372 @item --disable-glibtest
1373 Do not try to compile and run a test GLIB program.
1375 @item --with-libart-prefix=PFX
1376 Prefix where libart is installed (optional).
1378 @item --with-libart-exec-prefix=PFX
1379 Exec prefix where libart is installed (optional).
1381 @item --disable-libarttest
1382 Do not try to compile and run a test libart program.
1391 @uref{./index.html,,Return to the GCC Installation page}
1395 @c ***Building****************************************************************
1397 @comment node-name, next, previous, up
1398 @node Building, Testing, Configuration, Installing GCC
1404 @cindex Installing GCC: Building
1406 Now that GCC is configured, you are ready to build the compiler and
1409 Some commands executed when making the compiler may fail (return a
1410 nonzero status) and be ignored by @command{make}. These failures, which
1411 are often due to files that were not found, are expected, and can safely
1414 It is normal to have compiler warnings when compiling certain files.
1415 Unless you are a GCC developer, you can generally ignore these warnings
1416 unless they cause compilation to fail. Developers should attempt to fix
1417 any warnings encountered, however they can temporarily continue past
1418 warnings-as-errors by specifying the configure flag
1419 @option{--disable-werror}.
1421 On certain old systems, defining certain environment variables such as
1422 @env{CC} can interfere with the functioning of @command{make}.
1424 If you encounter seemingly strange errors when trying to build the
1425 compiler in a directory other than the source directory, it could be
1426 because you have previously configured the compiler in the source
1427 directory. Make sure you have done all the necessary preparations.
1429 If you build GCC on a BSD system using a directory stored in an old System
1430 V file system, problems may occur in running @command{fixincludes} if the
1431 System V file system doesn't support symbolic links. These problems
1432 result in a failure to fix the declaration of @code{size_t} in
1433 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1434 that type mismatches occur, this could be the cause.
1436 The solution is not to use such a directory for building GCC@.
1438 When building from CVS or snapshots, or if you modify parser sources,
1439 you need the Bison parser generator installed. Any version 1.25 or
1440 later should work; older versions may also work. If you do not modify
1441 parser sources, releases contain the Bison-generated files and you do
1442 not need Bison installed to build them.
1444 When building from CVS or snapshots, or if you modify Texinfo
1445 documentation, you need version 4.2 or later of Texinfo installed if you
1446 want Info documentation to be regenerated. Releases contain Info
1447 documentation pre-built for the unmodified documentation in the release.
1449 @section Building a native compiler
1451 For a native build issue the command @samp{make bootstrap}. This
1452 will build the entire GCC system, which includes the following steps:
1456 Build host tools necessary to build the compiler such as texinfo, bison,
1460 Build target tools for use by the compiler such as binutils (bfd,
1461 binutils, gas, gprof, ld, and opcodes)
1462 if they have been individually linked
1463 or moved into the top level GCC source tree before configuring.
1466 Perform a 3-stage bootstrap of the compiler.
1469 Perform a comparison test of the stage2 and stage3 compilers.
1472 Build runtime libraries using the stage3 compiler from the previous step.
1476 If you are short on disk space you might consider @samp{make
1477 bootstrap-lean} instead. This is identical to @samp{make
1478 bootstrap} except that object files from the stage1 and
1479 stage2 of the 3-stage bootstrap of the compiler are deleted as
1480 soon as they are no longer needed.
1482 If you want to save additional space during the bootstrap and in
1483 the final installation as well, you can build the compiler binaries
1484 without debugging information as in the following example. This will save
1485 roughly 40% of disk space both for the bootstrap and the final installation.
1486 (Libraries will still contain debugging information.)
1489 make CFLAGS='-O' LIBCFLAGS='-g -O2' \
1490 LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap
1493 If you wish to use non-default GCC flags when compiling the stage2 and
1494 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
1495 @samp{make bootstrap}. Non-default optimization flags are less well
1496 tested here than the default of @samp{-g -O2}, but should still work.
1497 In a few cases, you may find that you need to specify special flags such
1498 as @option{-msoft-float} here to complete the bootstrap; or, if the
1499 native compiler miscompiles the stage1 compiler, you may need to work
1500 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
1501 stage1 compiler that were miscompiled, or by using @samp{make
1502 bootstrap4} to increase the number of stages of bootstrap.
1504 Note that using non-standard @code{CFLAGS} can cause bootstrap to fail in
1505 @file{libiberty}, if these trigger a warning with the new compiler. For
1506 example using @samp{-O2 -g -mcpu=i686} on @code{i686-pc-linux-gnu} will
1507 cause bootstrap failure as @option{-mcpu=} is deprecated in 3.4.0 and above.
1510 If you used the flag @option{--enable-languages=@dots{}} to restrict
1511 the compilers to be built, only those you've actually enabled will be
1512 built. This will of course only build those runtime libraries, for
1513 which the particular compiler has been built. Please note,
1514 that re-defining @env{LANGUAGES} when calling @samp{make bootstrap}
1515 @strong{does not} work anymore!
1517 If the comparison of stage2 and stage3 fails, this normally indicates
1518 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1519 a potentially serious bug which you should investigate and report. (On
1520 a few systems, meaningful comparison of object files is impossible; they
1521 always appear ``different''. If you encounter this problem, you will
1522 need to disable comparison in the @file{Makefile}.)
1524 @section Building a cross compiler
1526 We recommend reading the
1527 @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ}
1528 for information about building cross compilers.
1530 When building a cross compiler, it is not generally possible to do a
1531 3-stage bootstrap of the compiler. This makes for an interesting problem
1532 as parts of GCC can only be built with GCC@.
1534 To build a cross compiler, we first recommend building and installing a
1535 native compiler. You can then use the native GCC compiler to build the
1536 cross compiler. The installed native compiler needs to be GCC version
1539 Assuming you have already installed a native copy of GCC and configured
1540 your cross compiler, issue the command @command{make}, which performs the
1545 Build host tools necessary to build the compiler such as texinfo, bison,
1549 Build target tools for use by the compiler such as binutils (bfd,
1550 binutils, gas, gprof, ld, and opcodes)
1551 if they have been individually linked or moved into the top level GCC source
1552 tree before configuring.
1555 Build the compiler (single stage only).
1558 Build runtime libraries using the compiler from the previous step.
1561 Note that if an error occurs in any step the make process will exit.
1563 If you are not building GNU binutils in the same source tree as GCC,
1564 you will need a cross-assembler and cross-linker installed before
1565 configuring GCC@. Put them in the directory
1566 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
1567 you should put in this directory:
1571 This should be the cross-assembler.
1574 This should be the cross-linker.
1577 This should be the cross-archiver: a program which can manipulate
1578 archive files (linker libraries) in the target machine's format.
1581 This should be a program to construct a symbol table in an archive file.
1584 The installation of GCC will find these programs in that directory,
1585 and copy or link them to the proper place to for the cross-compiler to
1586 find them when run later.
1588 The easiest way to provide these files is to build the Binutils package.
1589 Configure it with the same @option{--host} and @option{--target}
1590 options that you use for configuring GCC, then build and install
1591 them. They install their executables automatically into the proper
1592 directory. Alas, they do not support all the targets that GCC
1595 If you are not building a C library in the same source tree as GCC,
1596 you should also provide the target libraries and headers before
1597 configuring GCC, specifying the directories with
1598 @option{--with-sysroot} or @option{--with-headers} and
1599 @option{--with-libs}. Many targets also require ``start files'' such
1600 as @file{crt0.o} and
1601 @file{crtn.o} which are linked into each executable. There may be several
1602 alternatives for @file{crt0.o}, for use with profiling or other
1603 compilation options. Check your target's definition of
1604 @code{STARTFILE_SPEC} to find out what start files it uses.
1606 @section Building in parallel
1608 You can use @samp{make bootstrap MAKE="make -j 2" -j 2}, or just
1609 @samp{make -j 2 bootstrap} for GNU Make 3.79 and above, instead of
1610 @samp{make bootstrap} to build GCC in parallel.
1611 You can also specify a bigger number, and in most cases using a value
1612 greater than the number of processors in your machine will result in
1613 fewer and shorter I/O latency hits, thus improving overall throughput;
1614 this is especially true for slow drives and network filesystems.
1616 @section Building the Ada compiler
1618 In order to build GNAT, the Ada compiler, you need a working GNAT
1619 compiler (GNAT version 3.14 or later, or GCC version 3.1 or later),
1620 including GNAT tools such as @command{gnatmake} and @command{gnatlink},
1621 since the Ada front end is written in Ada (with some
1622 GNAT-specific extensions), and GNU make.
1624 @command{configure} does not test whether the GNAT installation works
1625 and has a sufficiently recent version; if too old a GNAT version is
1626 installed, the build will fail unless @option{--enable-languages} is
1627 used to disable building the Ada front end.
1629 @section Building with profile feedback
1631 It is possible to use profile feedback to optimize the compiler itself. This
1632 should result in a faster compiler binary. Experiments done on x86 using gcc
1633 3.3 showed approximately 7 percent speedup on compiling C programs. To
1634 bootstrap compiler with profile feedback, use @code{make profiledbootstrap}.
1636 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
1637 compiler. This compiler is used to build a @code{stageprofile} compiler
1638 instrumented to collect execution counts of instruction and branch
1639 probabilities. Then runtime libraries are compiled with profile collected.
1640 Finally a @code{stagefeedback} compiler is built using the information collected.
1642 Unlike @samp{make bootstrap} several additional restrictions apply. The
1643 compiler used to build @code{stage1} needs to support a 64-bit integral type.
1644 It is recommended to only use GCC for this. Also parallel make is currently
1645 not supported since collisions in profile collecting may occur.
1652 @uref{./index.html,,Return to the GCC Installation page}
1656 @c ***Testing*****************************************************************
1658 @comment node-name, next, previous, up
1659 @node Testing, Final install, Building, Installing GCC
1663 @chapter Installing GCC: Testing
1666 @cindex Installing GCC: Testing
1669 Before you install GCC, we encourage you to run the testsuites and to
1670 compare your results with results from a similar configuration that have
1671 been submitted to the
1672 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
1673 Some of these archived results are linked from the build status lists
1674 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
1675 reports a successful build runs the testsuites and submits the results.
1676 This step is optional and may require you to download additional software,
1677 but it can give you confidence in your new GCC installation or point out
1678 problems before you install and start using your new GCC@.
1680 First, you must have @uref{download.html,,downloaded the testsuites}.
1681 These are part of the full distribution, but if you downloaded the
1682 ``core'' compiler plus any front ends, you must download the testsuites
1685 Second, you must have the testing tools installed. This includes
1686 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu} 1.4.4 and later,
1687 Tcl, and Expect; the DejaGnu site has links to these.
1689 If the directories where @command{runtest} and @command{expect} were
1690 installed are not in the @env{PATH}, you may need to set the following
1691 environment variables appropriately, as in the following example (which
1692 assumes that DejaGnu has been installed under @file{/usr/local}):
1695 TCL_LIBRARY = /usr/local/share/tcl8.0
1696 DEJAGNULIBS = /usr/local/share/dejagnu
1699 (On systems such as Cygwin, these paths are required to be actual
1700 paths, not mounts or links; presumably this is due to some lack of
1701 portability in the DejaGnu code.)
1704 Finally, you can run the testsuite (which may take a long time):
1706 cd @var{objdir}; make -k check
1709 This will test various components of GCC, such as compiler
1710 front ends and runtime libraries. While running the testsuite, DejaGnu
1711 might emit some harmless messages resembling
1712 @samp{WARNING: Couldn't find the global config file.} or
1713 @samp{WARNING: Couldn't find tool init file} that can be ignored.
1715 @section How can you run the testsuite on selected tests?
1717 In order to run sets of tests selectively, there are targets
1718 @samp{make check-gcc} and @samp{make check-g++}
1719 in the @file{gcc} subdirectory of the object directory. You can also
1720 just run @samp{make check} in a subdirectory of the object directory.
1723 A more selective way to just run all @command{gcc} execute tests in the
1727 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
1730 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
1731 the testsuite with filenames matching @samp{9805*}, you would use
1734 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1737 The @file{*.exp} files are located in the testsuite directories of the GCC
1738 source, the most important ones being @file{compile.exp},
1739 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1740 To get a list of the possible @file{*.exp} files, pipe the
1741 output of @samp{make check} into a file and look at the
1742 @samp{Running @dots{} .exp} lines.
1744 @section Passing options and running multiple testsuites
1746 You can pass multiple options to the testsuite using the
1747 @samp{--target_board} option of DejaGNU, either passed as part of
1748 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
1749 work outside the makefiles. For example,
1752 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fno-strength-reduce"
1755 will run the standard @command{g++} testsuites (``unix'' is the target name
1756 for a standard native testsuite situation), passing
1757 @samp{-O3 -fno-strength-reduce} to the compiler on every test, i.e.,
1758 slashes separate options.
1760 You can run the testsuites multiple times using combinations of options
1761 with a syntax similar to the brace expansion of popular shells:
1764 @dots{}"--target_board=arm-sim/@{-mhard-float,-msoft-float@}@{-O1,-O2,-O3,@}"
1767 (Note the empty option caused by the trailing comma in the final group.)
1768 The following will run each testsuite eight times using the @samp{arm-sim}
1769 target, as if you had specified all possible combinations yourself:
1772 --target_board=arm-sim/-mhard-float/-O1
1773 --target_board=arm-sim/-mhard-float/-O2
1774 --target_board=arm-sim/-mhard-float/-O3
1775 --target_board=arm-sim/-mhard-float
1776 --target_board=arm-sim/-msoft-float/-O1
1777 --target_board=arm-sim/-msoft-float/-O2
1778 --target_board=arm-sim/-msoft-float/-O3
1779 --target_board=arm-sim/-msoft-float
1782 They can be combined as many times as you wish, in arbitrary ways. This
1786 @dots{}"--target_board=unix/-Wextra@{-O3,-fno-strength-reduce@}@{-fomit-frame-pointer,@}"
1789 will generate four combinations, all involving @samp{-Wextra}.
1791 The disadvantage to this method is that the testsuites are run in serial,
1792 which is a waste on multiprocessor systems. For users with GNU Make and
1793 a shell which performs brace expansion, you can run the testsuites in
1794 parallel by having the shell perform the combinations and @command{make}
1795 do the parallel runs. Instead of using @samp{--target_board}, use a
1796 special makefile target:
1799 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
1805 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
1808 will run three concurrent ``make-gcc'' testsuites, eventually testing all
1809 ten combinations as described above. Note that this is currently only
1810 supported in the @file{gcc} subdirectory. (To see how this works, try
1811 typing @command{echo} before the example given here.)
1814 @section Additional testing for Java Class Libraries
1816 The Java runtime tests can be executed via @samp{make check}
1817 in the @file{@var{target}/libjava/testsuite} directory in
1820 The @uref{http://sources.redhat.com/mauve/,,Mauve Project} provides
1821 a suite of tests for the Java Class Libraries. This suite can be run
1822 as part of libgcj testing by placing the Mauve tree within the libjava
1823 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
1824 specifying the location of that tree when invoking @samp{make}, as in
1825 @samp{make MAUVEDIR=~/mauve check}.
1827 @uref{http://www-124.ibm.com/developerworks/oss/cvs/jikes/~checkout~/jacks/jacks.html,,Jacks}
1828 is a free testsuite that tests Java compiler front ends. This suite
1829 can be run as part of libgcj testing by placing the Jacks tree within
1830 the libjava testsuite at @file{libjava/testsuite/libjava.jacks/jacks}.
1832 @section How to interpret test results
1834 The result of running the testsuite are various @file{*.sum} and @file{*.log}
1835 files in the testsuite subdirectories. The @file{*.log} files contain a
1836 detailed log of the compiler invocations and the corresponding
1837 results, the @file{*.sum} files summarize the results. These summaries
1838 contain status codes for all tests:
1842 PASS: the test passed as expected
1844 XPASS: the test unexpectedly passed
1846 FAIL: the test unexpectedly failed
1848 XFAIL: the test failed as expected
1850 UNSUPPORTED: the test is not supported on this platform
1852 ERROR: the testsuite detected an error
1854 WARNING: the testsuite detected a possible problem
1857 It is normal for some tests to report unexpected failures. At the
1858 current time the testing harness does not allow fine grained control
1859 over whether or not a test is expected to fail. This problem should
1860 be fixed in future releases.
1863 @section Submitting test results
1865 If you want to report the results to the GCC project, use the
1866 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
1869 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
1870 -m gcc-testresults@@gcc.gnu.org |sh
1873 This script uses the @command{Mail} program to send the results, so
1874 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
1875 prepended to the testsuite summary and should contain any special
1876 remarks you have on your results or your build environment. Please
1877 do not edit the testsuite result block or the subject line, as these
1878 messages may be automatically processed.
1885 @uref{./index.html,,Return to the GCC Installation page}
1889 @c ***Final install***********************************************************
1891 @comment node-name, next, previous, up
1892 @node Final install, , Testing, Installing GCC
1894 @ifset finalinstallhtml
1896 @chapter Installing GCC: Final installation
1899 Now that GCC has been built (and optionally tested), you can install it with
1901 cd @var{objdir}; make install
1904 We strongly recommend to install into a target directory where there is
1905 no previous version of GCC present.
1907 That step completes the installation of GCC; user level binaries can
1908 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
1909 you specified with the @option{--prefix} to configure (or
1910 @file{/usr/local} by default). (If you specified @option{--bindir},
1911 that directory will be used instead; otherwise, if you specified
1912 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
1913 Headers for the C++ and Java libraries are installed in
1914 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
1915 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
1916 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
1917 in info format in @file{@var{infodir}} (normally
1918 @file{@var{prefix}/info}).
1920 When installing cross-compilers, GCC's executables
1921 are not only installed into @file{@var{bindir}}, that
1922 is, @file{@var{exec-prefix}/bin}, but additionally into
1923 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
1924 exists. Typically, such @dfn{tooldirs} hold target-specific
1925 binutils, including assembler and linker.
1927 Installation into a temporary staging area or into a @command{chroot}
1928 jail can be achieved with the command
1931 make DESTDIR=@var{path-to-rootdir} install
1934 @noindent where @var{path-to-rootdir} is the absolute path of
1935 a directory relative to which all installation paths will be
1936 interpreted. Note that the directory specified by @code{DESTDIR}
1937 need not exist yet; it will be created if necessary.
1939 There is a subtle point with tooldirs and @code{DESTDIR}:
1940 If you relocate a cross-compiler installation with
1941 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
1942 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
1943 be filled with duplicated GCC executables only if it already exists,
1944 it will not be created otherwise. This is regarded as a feature,
1945 not as a bug, because it gives slightly more control to the packagers
1946 using the @code{DESTDIR} feature.
1948 If you built a released version of GCC using @samp{make bootstrap} then please
1949 quickly review the build status page for your release, available from
1950 @uref{http://gcc.gnu.org/buildstat.html}.
1951 If your system is not listed for the version of GCC that you built,
1953 @email{gcc@@gcc.gnu.org} indicating
1954 that you successfully built and installed GCC@.
1955 Include the following information:
1959 Output from running @file{@var{srcdir}/config.guess}. Do not send
1960 that file itself, just the one-line output from running it.
1963 The output of @samp{gcc -v} for your newly installed @command{gcc}.
1964 This tells us which version of GCC you built and the options you passed to
1968 Whether you enabled all languages or a subset of them. If you used a
1969 full distribution then this information is part of the configure
1970 options in the output of @samp{gcc -v}, but if you downloaded the
1971 ``core'' compiler plus additional front ends then it isn't apparent
1972 which ones you built unless you tell us about it.
1975 If the build was for GNU/Linux, also include:
1978 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
1979 this information should be available from @file{/etc/issue}.
1982 The version of the Linux kernel, available from @samp{uname --version}
1986 The version of glibc you used; for RPM-based systems like Red Hat,
1987 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
1988 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
1990 For other systems, you can include similar information if you think it is
1994 Any other information that you think would be useful to people building
1995 GCC on the same configuration. The new entry in the build status list
1996 will include a link to the archived copy of your message.
1999 We'd also like to know if the
2001 @ref{Specific, host/target specific installation notes}
2004 @uref{specific.html,,host/target specific installation notes}
2006 didn't include your host/target information or if that information is
2007 incomplete or out of date. Send a note to
2008 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
2010 If you find a bug, please report it following the
2011 @uref{../bugs.html,,bug reporting guidelines}.
2013 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
2014 dvi}. You will need to have @command{texi2dvi} (version at least 4.2)
2015 and @TeX{} installed. This creates a number of @file{.dvi} files in
2016 subdirectories of @file{@var{objdir}}; these may be converted for
2017 printing with programs such as @command{dvips}. You can also
2018 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
2019 Free Software Foundation}, though such manuals may not be for the most
2020 recent version of GCC@.
2022 If you would like to generate online HTML documentation, do @samp{cd
2023 @var{objdir}; make html} and HTML will be generated for the manuals in
2024 @file{@var{objdir}/HTML}.
2031 @uref{./index.html,,Return to the GCC Installation page}
2035 @c ***Binaries****************************************************************
2037 @comment node-name, next, previous, up
2038 @node Binaries, Specific, Installing GCC, Top
2042 @chapter Installing GCC: Binaries
2045 @cindex Installing GCC: Binaries
2047 We are often asked about pre-compiled versions of GCC@. While we cannot
2048 provide these for all platforms, below you'll find links to binaries for
2049 various platforms where creating them by yourself is not easy due to various
2052 Please note that we did not create these binaries, nor do we
2053 support them. If you have any problems installing them, please
2054 contact their makers.
2061 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
2064 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX}.
2068 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
2071 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
2072 Development Tools for the Renesas H8/300[HS] Series}.
2078 @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center};
2081 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
2085 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
2086 Development Tools for the Motorola 68HC11/68HC12}.
2089 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
2090 OpenServer/Unixware}.
2093 Sinix/Reliant Unix---@uref{ftp://ftp.fujitsu-siemens.com/pub/pd/gnu/gcc/,,Siemens}.
2096 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
2099 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
2105 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
2107 The @uref{http://www.mingw.org/,,MinGW} project.
2111 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
2112 Written Word} offers binaries for
2115 Digital UNIX 4.0D and 5.1,
2117 HP-UX 10.20, 11.00, and 11.11, and
2118 Solaris/SPARC 2.5.1, 2.6, 7, 8, and 9.
2121 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
2122 number of platforms.
2125 In addition to those specific offerings, you can get a binary
2126 distribution CD-ROM from the
2127 @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}.
2128 It contains binaries for a number of platforms, and
2129 includes not only GCC, but other stuff as well. The current CD does
2130 not contain the latest version of GCC, but it should allow
2131 bootstrapping the compiler. An updated version of that disk is in the
2139 @uref{./index.html,,Return to the GCC Installation page}
2143 @c ***Specific****************************************************************
2145 @comment node-name, next, previous, up
2146 @node Specific, Old, Binaries, Top
2150 @chapter Host/target specific installation notes for GCC
2153 @cindex Specific installation notes
2154 @cindex Target specific installation
2155 @cindex Host specific installation
2156 @cindex Target specific installation notes
2158 Please read this document carefully @emph{before} installing the
2159 GNU Compiler Collection on your machine.
2164 @uref{#alpha*-*-*,,alpha*-*-*}
2166 @uref{#alpha*-dec-osf*,,alpha*-dec-osf*}
2168 @uref{#alphaev5-cray-unicosmk*,,alphaev5-cray-unicosmk*}
2170 @uref{#arc-*-elf,,arc-*-elf}
2172 @uref{#arm-*-elf,,arm-*-elf}
2173 @uref{#arm-*-coff,,arm-*-coff}
2174 @uref{#arm-*-aout,,arm-*-aout}
2176 @uref{#xscale-*-*,,xscale-*-*}
2184 @uref{#*-*-freebsd*,,*-*-freebsd*}
2186 @uref{#h8300-hms,,h8300-hms}
2188 @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*}
2190 @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10}
2192 @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11}
2194 @uref{#*-*-linux-gnu,,*-*-linux-gnu}
2196 @uref{#ix86-*-linux*aout,,i?86-*-linux*aout}
2198 @uref{#ix86-*-linux*,,i?86-*-linux*}
2200 @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*}
2202 @uref{#ix86-*-udk,,i?86-*-udk}
2204 @uref{#ix86-*-esix,,i?86-*-esix}
2206 @uref{#ia64-*-linux,,ia64-*-linux}
2208 @uref{#ia64-*-hpux*,,ia64-*-hpux*}
2210 @uref{#*-ibm-aix*,,*-ibm-aix*}
2212 @uref{#ip2k-*-elf,,ip2k-*-elf}
2214 @uref{#iq2000-*-elf,,iq2000-*-elf}
2216 @uref{#m32r-*-elf,,m32r-*-elf}
2218 @uref{#m6811-elf,,m6811-elf}
2220 @uref{#m6812-elf,,m6812-elf}
2222 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
2224 @uref{#mips-*-*,,mips-*-*}
2226 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2228 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2230 @uref{#powerpc*-*-*,,powerpc*-*-*, powerpc-*-sysv4}
2232 @uref{#powerpc-*-darwin*,,powerpc-*-darwin*}
2234 @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4}
2236 @uref{#powerpc*-*-linux-gnu*,,powerpc*-*-linux-gnu*}
2238 @uref{#powerpc-*-netbsd*,,powerpc-*-netbsd*}
2240 @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix}
2242 @uref{#powerpc-*-eabisim,,powerpc-*-eabisim}
2244 @uref{#powerpc-*-eabi,,powerpc-*-eabi}
2246 @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
2248 @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim}
2250 @uref{#powerpcle-*-eabi,,powerpcle-*-eabi}
2252 @uref{#s390-*-linux*,,s390-*-linux*}
2254 @uref{#s390x-*-linux*,,s390x-*-linux*}
2256 @uref{#s390x-ibm-tpf*,,s390x-ibm-tpf*}
2258 @uref{#*-*-solaris2*,,*-*-solaris2*}
2260 @uref{#sparc-sun-solaris2*,,sparc-sun-solaris2*}
2262 @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7}
2264 @uref{#sparc-*-linux*,,sparc-*-linux*}
2266 @uref{#sparc64-*-solaris2*,,sparc64-*-solaris2*}
2268 @uref{#sparcv9-*-solaris2*,,sparcv9-*-solaris2*}
2270 @uref{#*-*-sysv*,,*-*-sysv*}
2272 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
2274 @uref{#*-*-vxworks*,,*-*-vxworks*}
2276 @uref{#x86_64-*-*,,x86_64-*-*, amd64-*-*}
2278 @uref{#xtensa-*-elf,,xtensa-*-elf}
2280 @uref{#xtensa-*-linux*,,xtensa-*-linux*}
2282 @uref{#windows,,Microsoft Windows}
2286 @uref{#older,,Older systems}
2291 @uref{#elf_targets,,all ELF targets} (SVR4, Solaris 2, etc.)
2297 <!-- -------- host/target specific issues start here ---------------- -->
2300 @heading @anchor{alpha*-*-*}alpha*-*-*
2302 This section contains general configuration information for all
2303 alpha-based platforms using ELF (in particular, ignore this section for
2304 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2305 section, please read all other sections that match your target.
2307 We require binutils 2.11.2 or newer.
2308 Previous binutils releases had a number of problems with DWARF 2
2309 debugging information, not the least of which is incorrect linking of
2315 @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf*
2316 Systems using processors that implement the DEC Alpha architecture and
2317 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2318 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2320 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2321 supported. (These are the versions which identify themselves as DEC
2324 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2325 may be fixed by configuring with @option{--with-gc=simple},
2326 reconfiguring Kernel Virtual Memory and Swap parameters
2327 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2328 or applying the patch in
2329 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2331 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2332 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2333 we need to use the old assembler, invoked via the barely documented
2334 @option{-oldas} option. To bootstrap GCC, you either need to use the
2338 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2341 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2344 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2347 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2348 are supported on Tru64 UNIX, so you must not configure GCC with
2349 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2351 GCC writes a @samp{.verstamp} directive to the assembler output file
2352 unless it is built as a cross-compiler. It gets the version to use from
2353 the system header file @file{/usr/include/stamp.h}. If you install a
2354 new version of DEC Unix, you should rebuild GCC to pick up the new version
2357 Note that since the Alpha is a 64-bit architecture, cross-compilers from
2358 32-bit machines will not generate code as efficient as that generated
2359 when the compiler is running on a 64-bit machine because many
2360 optimizations that depend on being able to represent a word on the
2361 target in an integral value on the host cannot be performed. Building
2362 cross-compilers on the Alpha for 32-bit machines has only been tested in
2363 a few cases and may not work properly.
2365 @samp{make compare} may fail on old versions of DEC Unix unless you add
2366 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2367 assembler input file is stored in the object file, and that makes
2368 comparison fail if it differs between the @code{stage1} and
2369 @code{stage2} compilations. The option @option{-save-temps} forces a
2370 fixed name to be used for the assembler input file, instead of a
2371 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2372 unless the comparisons fail without that option. If you add
2373 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2374 @samp{.s} files after each series of compilations.
2376 GCC now supports both the native (ECOFF) debugging format used by DBX
2377 and GDB and an encapsulated STABS format for use only with GDB@. See the
2378 discussion of the @option{--with-stabs} option of @file{configure} above
2379 for more information on these formats and how to select them.
2381 There is a bug in DEC's assembler that produces incorrect line numbers
2382 for ECOFF format when the @samp{.align} directive is used. To work
2383 around this problem, GCC will not emit such alignment directives
2384 while writing ECOFF format debugging information even if optimization is
2385 being performed. Unfortunately, this has the very undesirable
2386 side-effect that code addresses when @option{-O} is specified are
2387 different depending on whether or not @option{-g} is also specified.
2389 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2390 DBX@. DEC is now aware of this problem with the assembler and hopes to
2391 provide a fix shortly.
2396 @heading @anchor{alphaev5-cray-unicosmk*}alphaev5-cray-unicosmk*
2397 Cray T3E systems running Unicos/Mk.
2399 This port is incomplete and has many known bugs. We hope to improve the
2400 support for this target soon. Currently, only the C front end is supported,
2401 and it is not possible to build parallel applications. Cray modules are not
2402 supported; in particular, Craylibs are assumed to be in
2403 @file{/opt/ctl/craylibs/craylibs}.
2405 You absolutely @strong{must} use GNU make on this platform. Also, you
2406 need to tell GCC where to find the assembler and the linker. The
2407 simplest way to do so is by providing @option{--with-as} and
2408 @option{--with-ld} to @file{configure}, e.g.@:
2411 configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld \
2412 --enable-languages=c
2415 The comparison test during @samp{make bootstrap} fails on Unicos/Mk
2416 because the assembler inserts timestamps into object files. You should
2417 be able to work around this by doing @samp{make all} after getting this
2423 @heading @anchor{arc-*-elf}arc-*-elf
2424 Argonaut ARC processor.
2425 This configuration is intended for embedded systems.
2430 @heading @anchor{arm-*-elf}arm-*-elf
2431 @heading @anchor{xscale-*-*}xscale-*-*
2432 ARM-family processors. Subtargets that use the ELF object format
2433 require GNU binutils 2.13 or newer. Such subtargets include:
2434 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux},
2435 @code{arm-*-rtems} and @code{arm-*-kaos}.
2440 @heading @anchor{arm-*-coff}arm-*-coff
2441 ARM-family processors. Note that there are two different varieties
2442 of PE format subtarget supported: @code{arm-wince-pe} and
2443 @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}.
2448 @heading @anchor{arm-*-aout}arm-*-aout
2449 ARM-family processors. These targets support the AOUT file format:
2450 @code{arm-*-aout}, @code{arm-*-netbsd}.
2455 @heading @anchor{avr}avr
2457 ATMEL AVR-family micro controllers. These are used in embedded
2458 applications. There are no standard Unix configurations.
2460 @xref{AVR Options,, AVR Options, gcc, Using and Porting the GNU Compiler
2464 See ``AVR Options'' in the main manual
2466 for the list of supported MCU types.
2468 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2470 Further installation notes and other useful information about AVR tools
2471 can also be obtained from:
2475 @uref{http://www.openavr.org,,http://www.openavr.org}
2477 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2479 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2482 We @emph{strongly} recommend using binutils 2.13 or newer.
2484 The following error:
2486 Error: register required
2489 indicates that you should upgrade to a newer version of the binutils.
2494 @heading @anchor{c4x}c4x
2496 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
2497 Processors. These are used in embedded applications. There are no
2498 standard Unix configurations.
2500 @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using and
2501 Porting the GNU Compiler Collection (GCC)},
2504 See ``TMS320C3x/C4x Options'' in the main manual
2506 for the list of supported MCU types.
2508 GCC can be configured as a cross compiler for both the C3x and C4x
2509 architectures on the same system. Use @samp{configure --target=c4x
2510 --enable-languages="c,c++"} to configure.
2513 Further installation notes and other useful information about C4x tools
2514 can also be obtained from:
2518 @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/}
2524 @heading @anchor{cris}CRIS
2526 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2527 series. These are used in embedded applications.
2530 @xref{CRIS Options,, CRIS Options, gcc, Using and Porting the GNU Compiler
2534 See ``CRIS Options'' in the main manual
2536 for a list of CRIS-specific options.
2538 There are a few different CRIS targets:
2540 @item cris-axis-aout
2541 Old target. Includes a multilib for the @samp{elinux} a.out-based
2542 target. No multilibs for newer architecture variants.
2544 Mainly for monolithic embedded systems. Includes a multilib for the
2545 @samp{v10} core used in @samp{ETRAX 100 LX}.
2546 @item cris-axis-linux-gnu
2547 A GNU/Linux port for the CRIS architecture, currently targeting
2548 @samp{ETRAX 100 LX} by default.
2551 For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11
2552 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2554 Pre-packaged tools can be obtained from
2555 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2556 information about this platform is available at
2557 @uref{http://developer.axis.com/}.
2562 @heading @anchor{dos}DOS
2564 Please have a look at the @uref{binaries.html,,binaries page}.
2566 You cannot install GCC by itself on MSDOS; it will not compile under
2567 any MSDOS compiler except itself. You need to get the complete
2568 compilation package DJGPP, which includes binaries as well as sources,
2569 and includes all the necessary compilation tools and libraries.
2574 @heading @anchor{*-*-freebsd*}*-*-freebsd*
2576 The version of binutils installed in @file{/usr/bin} probably works with
2577 this release of GCC@. However, on FreeBSD 4, bootstrapping against the
2578 latest FSF binutils is known to improve overall testsuite results; and,
2579 on FreeBSD/alpha, using binutils 2.14 or later is required to build libjava.
2581 Support for FreeBSD 1 was discontinued in GCC 3.2.
2583 Support for FreeBSD 2 will be discontinued after GCC 3.4. The
2584 following was true for GCC 3.1 but the current status is unknown.
2585 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2586 configuration support and files as shipped with GCC 2.95 are still in
2587 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2588 it is unknown which version of binutils was used (it is assumed that it
2589 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2591 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2592 default for all CPU architectures. It had been the default on
2593 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2594 of @option{-g}, if you really want the old debugging format. There are
2595 no known issues with mixing object files and libraries with different
2596 debugging formats. Otherwise, this release of GCC should now match more
2597 of the configuration used in the stock FreeBSD configuration of GCC@. In
2598 particular, @option{--enable-threads} is now configured by default.
2599 However, as a general user, do not attempt to replace the system
2600 compiler with this release. Known to bootstrap and check with good
2601 results on FreeBSD 4.9-STABLE and 5-CURRENT@. In the past, known to
2602 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
2603 4.3, 4.4, 4.5, 4.8-STABLE@.
2605 In principle, @option{--enable-threads} is now compatible with
2606 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
2607 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
2609 library may be incorrectly built (symbols are missing at link time).
2610 There is a rare timing-based startup hang (probably involves an
2611 assumption about the thread library). Multi-threaded boehm-gc (required for
2612 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
2613 4.5-RELEASE@. Other CPU architectures
2614 supported by FreeBSD will require additional configuration tuning in, at
2615 the very least, both boehm-gc and libffi.
2617 Shared @file{libgcc_s.so} is now built and installed by default.
2622 @heading @anchor{h8300-hms}h8300-hms
2623 Renesas H8/300 series of processors.
2625 Please have a look at the @uref{binaries.html,,binaries page}.
2627 The calling convention and structure layout has changed in release 2.6.
2628 All code must be recompiled. The calling convention now passes the
2629 first three arguments in function calls in registers. Structures are no
2630 longer a multiple of 2 bytes.
2635 @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux*
2636 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
2638 We @emph{highly} recommend using gas/binutils on all hppa platforms;
2639 you may encounter a variety of problems when using the HP assembler.
2641 Specifically, @option{-g} does not work on HP-UX (since that system
2642 uses a peculiar debugging format which GCC does not know about), unless
2643 you use GAS and GDB@. It may be helpful to configure GCC with the
2644 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
2645 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
2647 If you wish to use the pa-risc 2.0 architecture support with a 32-bit
2648 runtime, you must use either the HP assembler, or gas/binutils 2.11
2651 There are two default scheduling models for instructions. These are
2652 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
2653 architecture specified for the target machine when configuring.
2654 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
2655 the target is a @samp{hppa1*} machine.
2657 The PROCESSOR_8000 model is not well suited to older processors. Thus,
2658 it is important to completely specify the machine architecture when
2659 configuring if you want a model other than PROCESSOR_8000. The macro
2660 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
2661 default scheduling model is desired.
2663 As of GCC 3.5, GCC uses the UNIX 95 namespace for HP-UX 10.10
2664 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
2665 This namespace change might cause problems when bootstrapping with
2666 an earlier version of GCC or the HP compiler as essentially the same
2667 namespace is required for an entire build. This problem can be avoided
2668 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
2669 or @samp{98}. Another way is to add an appropriate set of predefines
2670 to @env{CC}. The description for the @option{munix=} option contains
2671 a list of the predefines used with each standard.
2673 More specific information to @samp{hppa*-hp-hpux*} targets follows.
2678 @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10
2680 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
2681 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
2687 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
2691 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
2695 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
2698 The HP assembler on these systems has some problems. Most notably the
2699 assembler inserts timestamps into each object file it creates, causing
2700 the 3-stage comparison test to fail during a @samp{make bootstrap}.
2701 You should be able to continue by saying @samp{make all} after getting
2702 the failure from @samp{make bootstrap}.
2704 GCC 3.5 requires CVS binutils as of April 28, 2004 or later. Earlier
2705 versions require binutils 2.8 or later.
2707 The C++ ABI has changed incompatibly in GCC 3.5. COMDAT subspaces are
2708 used for one-only code and data. This resolves many of the previous
2709 problems in using C++ on this target. However, the ABI is not compatible
2710 with the one implemented under HP-UX 11 using secondary definitions.
2715 @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11
2717 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
2718 be used to compile GCC 3.0 and up.
2720 Refer to @uref{binaries.html,,binaries} for information about obtaining
2721 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
2722 to build the Ada language as it can't be bootstrapped using C@. Ada is
2723 only available for the 32-bit PA-RISC runtime. The libffi and libjava
2724 haven't been ported to HP-UX and don't build.
2726 It is possible to build GCC 3.3 starting with the bundled HP compiler,
2727 but the process requires several steps. GCC 3.3 can then be used to
2728 build later versions. The fastjar program contains ISO C code and
2729 can't be built with the HP bundled compiler. This problem can be
2730 avoided by not building the Java language. For example, use the
2731 @option{--enable-languages="c,c++,f77,objc"} option in your configure
2734 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
2735 bundled compiler supports only traditional C; you will need either HP's
2736 unbundled compiler, or a binary distribution of GCC@.
2738 There are several possible approaches to building the distribution.
2739 Binutils can be built first using the HP tools. Then, the GCC
2740 distribution can be built. The second approach is to build GCC
2741 first using the HP tools, then build binutils, then rebuild GCC@.
2742 There have been problems with various binary distributions, so it
2743 is best not to start from a binary distribution.
2745 On 64-bit capable systems, there are two distinct targets. Different
2746 installation prefixes must be used if both are to be installed on
2747 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
2748 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
2749 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
2750 PA-RISC 2.0 architecture. The HP and GNU linkers are both supported
2753 The script config.guess now selects the target type based on the compiler
2754 detected during configuration. You must define @env{PATH} or @env{CC} so
2755 that configure finds an appropriate compiler for the initial bootstrap.
2756 When @env{CC} is used, the definition should contain the options that are
2757 needed whenever @env{CC} is used.
2759 Specifically, options that determine the runtime architecture must be
2760 in @env{CC} to correctly select the target for the build. It is also
2761 convenient to place many other compiler options in @env{CC}. For example,
2762 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
2763 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
2764 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
2765 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
2766 macro definition table of cpp needs to be increased for a successful
2767 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
2768 be defined when building with the bundled compiler, or when using the
2769 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
2771 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
2772 with the @option{--with-ld=@dots{}} option. This overrides the standard
2773 search for ld. The two linkers supported on this target require different
2774 commands. The default linker is determined during configuration. As a
2775 result, it's not possible to switch linkers in the middle of a GCC build.
2776 This has been been reported to sometimes occur in unified builds of
2779 GCC 3.0 through 3.2 require binutils 2.11 or above. GCC 3.3 through
2780 GCC 3.5 require binutils 2.14 or later.
2782 Although the HP assembler can be used for an initial build, it shouldn't
2783 be used with any languages other than C and perhaps Fortran due to its
2784 many limitations. For example, it does not support weak symbols or alias
2785 definitions. As a result, explicit template instantiations are required
2786 when using C++. This makes it difficult if not impossible to build many
2787 C++ applications. You can't generate debugging information when using
2788 the HP assembler. Finally, @samp{make bootstrap} fails in the final
2789 comparison of object modules due to the time stamps that it inserts into
2790 the modules. The bootstrap can be continued from this point with
2793 A recent linker patch must be installed for the correct operation of
2794 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
2795 oldest linker patches that are known to work. They are for HP-UX
2796 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
2797 @code{PHSS_24304}, might be usable but it hasn't been tested. These
2798 patches have been superseded. Consult the HP patch database to obtain
2799 the currently recommended linker patch for your system.
2801 The patches are necessary for the support of weak symbols on the
2802 32-bit port, and for the running of initializers and finalizers. Weak
2803 symbols are implemented using SOM secondary definition symbols. Prior
2804 to HP-UX 11, there are bugs in the linker support for secondary symbols.
2805 The patches correct a problem of linker core dumps creating shared
2806 libraries containing secondary symbols, as well as various other
2807 linking issues involving secondary symbols.
2809 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
2810 run initializers and finalizers on the 64-bit port. The 32-bit port
2811 uses the linker @option{+init} and @option{+fini} options for the same
2812 purpose. The patches correct various problems with the +init/+fini
2813 options, including program core dumps. Binutils 2.14 corrects a
2814 problem on the 64-bit port resulting from HP's non-standard use of
2815 the .init and .fini sections for array initializers and finalizers.
2817 There are a number of issues to consider in selecting which linker to
2818 use with the 64-bit port. The GNU 64-bit linker can only create dynamic
2819 binaries. The @option{-static} option causes linking with archive
2820 libraries but doesn't produce a truly static binary. Dynamic binaries
2821 still require final binding by the dynamic loader to resolve a set of
2822 dynamic-loader-defined symbols. The default behavior of the HP linker
2823 is the same as the GNU linker. However, it can generate true 64-bit
2824 static binaries using the @option{+compat} option.
2826 The HP 64-bit linker doesn't support linkonce semantics. As a
2827 result, C++ programs have many more sections than they should.
2829 The GNU 64-bit linker has some issues with shared library support
2830 and exceptions. As a result, we only support libgcc in archive
2831 format. For similar reasons, dwarf2 unwind and exception support
2832 are disabled. The GNU linker also has problems creating binaries
2833 with @option{-static}. It doesn't provide stubs for internal
2834 calls to global functions in shared libraries, so these calls
2835 can't be overloaded.
2837 Thread support is not implemented in GCC 3.0 through 3.2, so the
2838 @option{--enable-threads} configure option does not work. In 3.3
2839 and later, POSIX threads are supported. The optional DCE thread
2840 library is not supported.
2842 This port still is undergoing significant development.
2847 @heading @anchor{*-*-linux-gnu}*-*-linux-gnu
2849 Versions of libstdc++-v3 starting with 3.2.1 require bugfixes present
2850 in glibc 2.2.5 and later. More information is available in the
2851 libstdc++-v3 documentation.
2856 @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout
2857 Use this configuration to generate @file{a.out} binaries on Linux-based
2858 GNU systems. This configuration is being superseded.
2863 @heading @anchor{ix86-*-linux*}i?86-*-linux*
2865 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
2866 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
2868 If you receive Signal 11 errors when building on GNU/Linux, then it is
2869 possible you have a hardware problem. Further information on this can be
2870 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
2875 @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5*
2876 Use this for the SCO OpenServer Release 5 family of operating systems.
2878 Unlike earlier versions of GCC, the ability to generate COFF with this
2879 target is no longer provided.
2881 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
2882 the system debugger to be used. That support was too burdensome to
2883 maintain. GCC now emits only DWARF 2 for this target. This means you
2884 may use either the UDK debugger or GDB to debug programs built by this
2887 GCC is now only supported on releases 5.0.4 and later, and requires that
2888 you install Support Level Supplement OSS646B or later, and Support Level
2889 Supplement OSS631C or later. If you are using release 5.0.7 of
2890 OpenServer, you must have at least the first maintenance pack installed
2891 (this includes the relevant portions of OSS646). OSS646, also known as
2892 the ``Execution Environment Update'', provides updated link editors and
2893 assemblers, as well as updated standard C and math libraries. The C
2894 startup modules are also updated to support the System V gABI draft, and
2895 GCC relies on that behavior. OSS631 provides a collection of commonly
2896 used open source libraries, some of which GCC depends on (such as GNU
2897 gettext and zlib). SCO OpenServer Release 5.0.7 has all of this built
2898 in by default, but OSS631C and later also apply to that release. Please
2900 @uref{ftp://ftp.sco.com/pub/openserver5,,ftp://ftp.sco.com/pub/openserver5}
2901 for the latest versions of these (and other potentially useful)
2904 Although there is support for using the native assembler, it is
2905 recommended that you configure GCC to use the GNU assembler. You do
2906 this by using the flags
2907 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You should
2908 use a modern version of GNU binutils. Version 2.13.2.1 was used for all
2909 testing. In general, only the @option{--with-gnu-as} option is tested.
2910 A modern bintuils (as well as a plethora of other development related
2911 GNU utilities) can be found in Support Level Supplement OSS658A, the
2912 ``GNU Development Tools'' package. See the SCO web and ftp sites for details.
2913 That package also contains the currently ``officially supported'' version of
2914 GCC, version 2.95.3. It is useful for bootstrapping this version.
2919 @heading @anchor{ix86-*-udk}i?86-*-udk
2921 This target emulates the SCO Universal Development Kit and requires that
2922 package be installed. (If it is installed, you will have a
2923 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
2924 @samp{i?86-*-unixware7*} target
2925 but is meant to be used when hosting on a system where UDK isn't the
2926 default compiler such as OpenServer 5 or Unixware 2. This target will
2927 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
2928 with the same warnings and caveats as the SCO UDK@.
2930 This target is a little tricky to build because we have to distinguish
2931 it from the native tools (so it gets headers, startups, and libraries
2932 from the right place) while making the tools not think we're actually
2933 building a cross compiler. The easiest way to do this is with a configure
2937 CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure \
2938 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-
2941 @emph{You should substitute @samp{i686} in the above command with the appropriate
2942 processor for your host.}
2944 After the usual @samp{make bootstrap} and
2945 @samp{make install}, you can then access the UDK-targeted GCC
2946 tools by adding @command{udk-} before the commonly known name. For
2947 example, to invoke the C compiler, you would use @command{udk-gcc}.
2948 They will coexist peacefully with any native-target GCC tools you may
2955 @heading @anchor{ia64-*-linux}ia64-*-linux
2956 IA-64 processor (also known as IPF, or Itanium Processor Family)
2959 If you are using the installed system libunwind library with
2960 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
2963 None of the following versions of GCC has an ABI that is compatible
2964 with any of the other versions in this list, with the exception that
2965 Red Hat 2.96 and Trillian 000171 are compatible with each other:
2966 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
2967 This primarily affects C++ programs and programs that create shared libraries.
2968 GCC 3.1 or later is recommended for compiling linux, the kernel.
2969 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
2970 more major ABI changes are expected.
2975 @heading @anchor{ia64-*-hpux*}ia64-*-hpux*
2976 Building GCC on this target requires the GNU Assembler. The bundled HP
2977 assembler will not work. To prevent GCC from using the wrong assembler,
2978 the option @option{--with-gnu-as} may be necessary.
2980 The GCC libunwind library has not been ported to HPUX@. This means that for
2981 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
2982 is required to build GCC@. For GCC 3.3 and later, this is the default.
2983 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
2984 removed and the system libunwind library will always be used.
2988 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
2990 @heading @anchor{*-ibm-aix*}*-ibm-aix*
2991 Support for AIX version 3 and older was discontinued in GCC 3.4.
2993 AIX Make frequently has problems with GCC makefiles. GNU Make 3.79.1 or
2994 newer is recommended to build on this platform.
2996 To speed up the configuration phases of bootstrapping and installing GCC,
2997 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
3000 % CONFIG_SHELL=/opt/freeware/bin/bash
3001 % export CONFIG_SHELL
3004 and then proceed as described in @uref{build.html,,the build instructions},
3005 where we strongly recommend using GNU make and specifying an absolute path
3006 to invoke @var{srcdir}/configure.
3008 Errors involving @code{alloca} when building GCC generally are due
3009 to an incorrect definition of @code{CC} in the Makefile or mixing files
3010 compiled with the native C compiler and GCC@. During the stage1 phase of
3011 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
3012 (not @command{xlc}). Once @command{configure} has been informed of
3013 @command{xlc}, one needs to use @samp{make distclean} to remove the
3014 configure cache files and ensure that @env{CC} environment variable
3015 does not provide a definition that will confuse @command{configure}.
3016 If this error occurs during stage2 or later, then the problem most likely
3017 is the version of Make (see above).
3019 The native @command{as} and @command{ld} are recommended for bootstrapping
3020 on AIX 4 and required for bootstrapping on AIX 5L@. The GNU Assembler
3021 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
3022 utilize weak symbol functionality although it is not supported. The GNU
3023 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC@.
3024 The native AIX tools do interoperate with GCC@.
3026 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
3027 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
3028 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
3029 referenced as APAR IY53606 (AIX 5.2) or a APAR IY54774 (AIX 5.1)
3031 @samp{libstdc++} in GCC 3.4 increments the major version number of the
3032 shared object and GCC installation places the @file{libstdc++.a}
3033 shared library in a common location which will overwrite the and GCC
3034 3.3 version of the shared library. Applications either need to be
3035 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
3036 versions of the @samp{libstdc++} shared object needs to be available
3037 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
3038 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
3039 installed for runtime dynamic loading using the following steps to set
3040 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3041 multilib @file{libstdc++.a} installed:
3043 Extract the shared objects from the currently installed
3044 @file{libstdc++.a} archive:
3046 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3049 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3050 available for runtime dynamic loading, but not linking:
3052 % strip -e libstdc++.so.4 libstdc++.so.5
3055 Archive the runtime-only shared object in the GCC 3.4
3056 @file{libstdc++.a} archive:
3058 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
3061 Linking executables and shared libraries may produce warnings of
3062 duplicate symbols. The assembly files generated by GCC for AIX always
3063 have included multiple symbol definitions for certain global variable
3064 and function declarations in the original program. The warnings should
3065 not prevent the linker from producing a correct library or runnable
3068 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
3069 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
3070 to parse archive libraries did not handle the new format correctly.
3071 These routines are used by GCC and result in error messages during
3072 linking such as ``not a COFF file''. The version of the routines shipped
3073 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
3074 option of the archive command may be used to create archives of 32-bit
3075 objects using the original ``small format''. A correct version of the
3076 routines is shipped with AIX 4.3.2 and above.
3078 Some versions of the AIX binder (linker) can fail with a relocation
3079 overflow severe error when the @option{-bbigtoc} option is used to link
3080 GCC-produced object files into an executable that overflows the TOC@. A fix
3081 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
3082 available from IBM Customer Support and from its
3083 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3084 website as PTF U455193.
3086 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
3087 with a segmentation fault when invoked by any version of GCC@. A fix for
3088 APAR IX87327 is available from IBM Customer Support and from its
3089 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3090 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
3092 The initial assembler shipped with AIX 4.3.0 generates incorrect object
3093 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
3094 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
3095 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3096 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
3098 AIX provides National Language Support (NLS)@. Compilers and assemblers
3099 use NLS to support locale-specific representations of various data
3100 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
3101 separating decimal fractions). There have been problems reported where
3102 GCC does not produce the same floating-point formats that the assembler
3103 expects. If one encounters this problem, set the @env{LANG}
3104 environment variable to @samp{C} or @samp{En_US}.
3106 By default, GCC for AIX 4.1 and above produces code that can be used on
3107 both Power or PowerPC processors.
3109 A default can be specified with the @option{-mcpu=@var{cpu_type}}
3110 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
3115 @heading @anchor{ip2k-*-elf}ip2k-*-elf
3116 Ubicom IP2022 micro controller.
3117 This configuration is intended for embedded systems.
3118 There are no standard Unix configurations.
3120 Use @samp{configure --target=ip2k-elf --enable-languages=c} to configure GCC@.
3125 @heading @anchor{iq2000-*-elf}iq2000-*-elf
3126 Vitesse IQ2000 processors. These are used in embedded
3127 applications. There are no standard Unix configurations.
3132 @heading @anchor{m32r-*-elf}m32r-*-elf
3133 Renesas M32R processor.
3134 This configuration is intended for embedded systems.
3139 @heading @anchor{m6811-elf}m6811-elf
3140 Motorola 68HC11 family micro controllers. These are used in embedded
3141 applications. There are no standard Unix configurations.
3146 @heading @anchor{m6812-elf}m6812-elf
3147 Motorola 68HC12 family micro controllers. These are used in embedded
3148 applications. There are no standard Unix configurations.
3153 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
3154 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
3155 the assembler that prevents compilation of GCC@. This
3156 bug manifests itself during the first stage of compilation, while
3157 building @file{libgcc2.a}:
3161 cc1: warning: `-g' option not supported on this version of GCC
3162 cc1: warning: `-g1' option not supported on this version of GCC
3163 ./xgcc: Internal compiler error: program as got fatal signal 11
3166 A patched version of the assembler is available as the file
3167 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
3168 have HP software support, the patch can also be obtained directly from
3169 HP, as described in the following note:
3172 This is the patched assembler, to patch SR#1653-010439, where the
3173 assembler aborts on floating point constants.
3175 The bug is not really in the assembler, but in the shared library
3176 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
3177 SR#4701-078451. Anyway, the attached assembler uses the archive
3178 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
3181 This patch is also known as PHCO_4484.
3183 In addition gdb does not understand that native HP-UX format, so
3184 you must use gas if you wish to use gdb.
3186 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
3187 @command{fixproto} shell script triggers a bug in the system shell. If you
3188 encounter this problem, upgrade your operating system or use BASH (the
3189 GNU shell) to run @command{fixproto}. This bug will cause the fixproto
3190 program to report an error of the form:
3193 ./fixproto: sh internal 1K buffer overflow
3196 To fix this, you can also change the first line of the fixproto script
3206 @heading @anchor{mips-*-*}mips-*-*
3207 If on a MIPS system you get an error message saying ``does not have gp
3208 sections for all it's [sic] sectons [sic]'', don't worry about it. This
3209 happens whenever you use GAS with the MIPS linker, but there is not
3210 really anything wrong, and it is okay to use the output file. You can
3211 stop such warnings by installing the GNU linker.
3213 It would be nice to extend GAS to produce the gp tables, but they are
3214 optional, and there should not be a warning about their absence.
3216 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
3217 and later. A patch went in just after the GCC 3.3 release to
3218 make @samp{mips*-*-*} use the generic implementation instead. You can also
3219 configure for @samp{mipsel-elf} as a workaround. The
3220 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
3221 work on this is expected in future releases.
3223 MIPS systems check for division by zero (unless
3224 @option{-mno-check-zero-division} is passed to the compiler) by
3225 generating either a conditional trap or a break instruction. Using
3226 trap results in smaller code, but is only supported on MIPS II and
3227 later. Also, some versions of the Linux kernel have a bug that
3228 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
3229 the use of break, use the @option{--with-divide=breaks}
3230 @command{configure} option when configuring GCC@. The default is to
3231 use traps on systems that support them.
3233 Cross-compilers for the Mips as target using the Mips assembler
3234 currently do not work, because the auxiliary programs
3235 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
3236 anything but a Mips. It does work to cross compile for a Mips
3237 if you use the GNU assembler and linker.
3242 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
3244 In order to compile GCC on an SGI running IRIX 5, the @samp{compiler_dev.hdr}
3245 subsystem must be installed from the IDO CD-ROM supplied by SGI@.
3246 It is also available for download from
3247 @uref{ftp://ftp.sgi.com/sgi/IRIX5.3/iris-development-option-5.3.tardist}.
3249 If you use the MIPS C compiler to bootstrap, it may be necessary
3250 to increase its table size for switch statements with the
3251 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
3252 optimization option, you also need to use @option{-Olimit 3000}.
3254 To enable debugging under IRIX 5, you must use GNU binutils 2.15 or
3255 later, and use the @option{--with-gnu-ld} @command{configure} option
3256 when configuring GCC@. You need to use GNU @command{ar} and @command{nm},
3257 also distributed with GNU binutils.
3259 Some users have reported that @command{/bin/sh} will hang during bootstrap.
3260 This problem can be avoided by running the commands:
3263 % CONFIG_SHELL=/bin/ksh
3264 % export CONFIG_SHELL
3267 before starting the build.
3272 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3274 If you are using SGI's MIPSpro @command{cc} as your bootstrap compiler, you must
3275 ensure that the N32 ABI is in use. To test this, compile a simple C
3276 file with @command{cc} and then run @command{file} on the
3277 resulting object file. The output should look like:
3280 test.o: ELF N32 MSB @dots{}
3286 test.o: ELF 32-bit MSB @dots{}
3292 test.o: ELF 64-bit MSB @dots{}
3295 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3296 should set the environment variable @env{CC} to @samp{cc -n32}
3297 before configuring GCC@.
3299 If you want the resulting @command{gcc} to run on old 32-bit systems
3300 with the MIPS R4400 CPU, you need to ensure that only code for the @samp{mips3}
3301 instruction set architecture (ISA) is generated. While GCC 3.x does
3302 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3303 the ISA depending on the machine where GCC is built. Using one of them
3304 as the bootstrap compiler may result in @samp{mips4} code, which won't run at
3305 all on @samp{mips3}-only systems. For the test program above, you should see:
3308 test.o: ELF N32 MSB mips-3 @dots{}
3314 test.o: ELF N32 MSB mips-4 @dots{}
3317 instead, you should set the environment variable @env{CC} to @samp{cc
3318 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3320 MIPSpro C 7.4 may cause bootstrap failures, due to a bug when inlining
3321 @code{memcmp}. Either add @code{-U__INLINE_INTRINSICS} to the @env{CC}
3322 environment variable as a workaround or upgrade to MIPSpro C 7.4.1m.
3324 GCC on IRIX 6 is usually built to support the N32, O32 and N64 ABIs. If
3325 you build GCC on a system that doesn't have the N64 libraries installed
3326 or cannot run 64-bit binaries,
3327 you need to configure with @option{--disable-multilib} so GCC doesn't
3328 try to use them. This will disable building the O32 libraries, too.
3329 Look for @file{/usr/lib64/libc.so.1} to see if you
3330 have the 64-bit libraries installed.
3332 To enable debugging for the O32 ABI, you must use GNU @command{as} from
3333 GNU binutils 2.15 or later. You may also use GNU @command{ld}, but
3334 this is not required and currently causes some problems with Ada.
3336 The @option{--enable-threads} option doesn't currently work, a patch is
3337 in preparation for a future release. The @option{--enable-libgcj}
3338 option is disabled by default: IRIX 6 uses a very low default limit
3339 (20480) for the command line length. Although @command{libtool} contains a
3340 workaround for this problem, at least the N64 @samp{libgcj} is known not
3341 to build despite this, running into an internal error of the native
3342 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3343 its maximum of 262144 bytes. If you have root access, you can use the
3344 @command{systune} command to do this.
3346 See @uref{http://freeware.sgi.com/} for more
3347 information about using GCC on IRIX platforms.
3352 @heading @anchor{powerpc*-*-*}powerpc-*-*
3354 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3355 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3360 @heading @anchor{powerpc-*-darwin*}powerpc-*-darwin*
3361 PowerPC running Darwin (Mac OS X kernel).
3363 Pre-installed versions of Mac OS X may not include any developer tools,
3364 meaning that you will not be able to build GCC from source. Tool
3365 binaries are available at
3366 @uref{http://developer.apple.com/darwin/projects/compiler/} (free
3367 registration required).
3369 This version of GCC requires at least cctools-528.
3371 The version of GCC shipped by Apple typically includes a number of
3372 extensions not available in a standard GCC release. These extensions
3373 are generally for backwards compatibility and best avoided.
3378 @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4
3379 PowerPC system in big endian mode, running System V.4.
3384 @heading @anchor{powerpc*-*-linux-gnu*}powerpc*-*-linux-gnu*
3387 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.15}
3388 or newer for a working GCC@.
3393 @heading @anchor{powerpc-*-netbsd*}powerpc-*-netbsd*
3394 PowerPC system in big endian mode running NetBSD@. To build the
3395 documentation you will need Texinfo version 4.2 (NetBSD 1.5.1 included
3396 Texinfo version 3.12).
3401 @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim
3402 Embedded PowerPC system in big endian mode for use in running under the
3408 @heading @anchor{powerpc-*-eabi}powerpc-*-eabi
3409 Embedded PowerPC system in big endian mode.
3414 @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4
3415 PowerPC system in little endian mode, running System V.4.
3420 @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim
3421 Embedded PowerPC system in little endian mode for use in running under
3427 @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi
3428 Embedded PowerPC system in little endian mode.
3433 @heading @anchor{s390-*-linux*}s390-*-linux*
3434 S/390 system running GNU/Linux for S/390@.
3439 @heading @anchor{s390x-*-linux*}s390x-*-linux*
3440 zSeries system (64-bit) running GNU/Linux for zSeries@.
3445 @heading @anchor{s390x-ibm-tpf*}s390x-ibm-tpf*
3446 zSeries system (64-bit) running TPF@. This platform is
3447 supported as cross-compilation target only.
3452 @c Please use Solaris 2 to refer to all release of Solaris, starting
3453 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
3454 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3455 @c alone is too unspecific and must be avoided.
3456 @heading @anchor{*-*-solaris2*}*-*-solaris2*
3458 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3459 GCC you first have to install a pre-built compiler, see the
3460 @uref{binaries.html,,binaries page} for details.
3462 The Solaris 2 @command{/bin/sh} will often fail to configure
3463 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3464 recommend to use the following sequence of commands to bootstrap and
3468 % CONFIG_SHELL=/bin/ksh
3469 % export CONFIG_SHELL
3472 and then proceed as described in @uref{build.html,,the build instructions}.
3473 In addition we strongly recommend specifying an absolute path to invoke
3474 @var{srcdir}/configure.
3476 Solaris 2 comes with a number of optional OS packages. Some of these
3477 are needed to use GCC fully, namely @code{SUNWarc},
3478 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3479 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3480 optional packages when installing Solaris 2, you will need to verify that
3481 the packages that GCC needs are installed.
3483 To check whether an optional package is installed, use
3484 the @command{pkginfo} command. To add an optional package, use the
3485 @command{pkgadd} command. For further details, see the Solaris 2
3488 Trying to use the linker and other tools in
3489 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3490 For example, the linker may hang indefinitely. The fix is to remove
3491 @file{/usr/ucb} from your @env{PATH}.
3493 The build process works more smoothly with the legacy Sun tools so, if you
3494 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3495 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3497 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3498 platform. We recommend the use of GNU binutils 2.11.2 or later, or the
3499 vendor tools (Sun @command{as}, Sun @command{ld}). Note that your mileage
3500 may vary if you use a combination of the GNU tools and the Sun tools: while
3501 the combination GNU @command{as} + Sun @command{ld} should reasonably work,
3502 the reverse combination Sun @command{as} + GNU @command{ld} is known to
3503 cause memory corruption at runtime in some cases for C++ programs.
3505 The stock GNU binutils 2.15 release is broken on this platform because of a
3506 single bug. It has been fixed on the 2.15 branch in the CVS repository.
3507 You can obtain a working version by checking out the binutils-2_15-branch
3508 from the CVS repository or applying the patch
3509 @uref{http://sources.redhat.com/ml/binutils-cvs/2004-09/msg00036.html} to the
3512 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3513 newer: @command{g++} will complain that types are missing. These headers assume
3514 that omitting the type means @code{int}; this assumption worked for C89 but
3515 is wrong for C++, and is now wrong for C99 also.
3517 @command{g++} accepts such (invalid) constructs with the option
3518 @option{-fpermissive}; it
3519 will assume that any missing type is @code{int} (as defined by C89).
3521 There are patches for Solaris 2.6 (105633-56 or newer for SPARC,
3522 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC,
3523 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3524 108653-22 for Intel) that fix this bug.
3526 Sun bug 4927647 sometimes causes random spurious testsuite failures
3527 related to missing diagnostic output. This bug doesn't affect GCC
3528 itself, rather it is a kernel bug triggered by the @command{expect}
3529 program which is used only by the GCC testsuite driver. When the bug
3530 causes the @command{expect} program to miss anticipated output, extra
3531 testsuite failures appear.
3533 There are patches for Solaris 8 (117350-12 or newer for SPARC,
3534 117351-12 or newer for Intel) and Solaris 9 (117171-11 or newer for
3535 SPARC, 117172-11 or newer for Intel) that address this problem.
3540 @heading @anchor{sparc-sun-solaris2*}sparc-sun-solaris2*
3542 When GCC is configured to use binutils 2.11.2 or later the binaries
3543 produced are smaller than the ones produced using Sun's native tools;
3544 this difference is quite significant for binaries containing debugging
3547 Sun @command{as} 4.x is broken in that it cannot cope with long symbol names.
3548 A typical error message might look similar to the following:
3551 /usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041: error:
3552 can't compute value of an expression involving an external symbol.
3555 This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris
3556 2.6 and has been fixed in later (5.x) versions of the assembler,
3557 starting with Solaris 7.
3559 Starting with Solaris 7, the operating system is capable of executing
3560 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3561 this; the @option{-m64} option enables 64-bit code generation.
3562 However, if all you want is code tuned for the UltraSPARC CPU, you
3563 should try the @option{-mtune=ultrasparc} option instead, which produces
3564 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3567 When configuring on a Solaris 7 or later system that is running a kernel
3568 that supports only 32-bit binaries, one must configure with
3569 @option{--disable-multilib}, since we will not be able to build the
3570 64-bit target libraries.
3572 GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
3573 the GNU compiler (especially GCC 3.0.x versions), which lead to the
3574 miscompilation of the stage1 compiler and the subsequent failure of the
3575 bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
3576 stage, i.e.@: to bootstrap that compiler with the base compiler and then
3577 use it to bootstrap the final compiler.
3579 GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
3580 and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
3581 failure in form of a miscompilation of the stage1 compiler by the Sun
3582 compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
3584 GCC 3.4 changed the default debugging format from STABS to DWARF-2 for
3585 32-bit code on Solaris 7 and later. If you are using the Sun
3586 assembler, this change apparently runs afoul of Sun bug 4910101, for
3587 which (as of 2004-05-23) there is no fix. A symptom of the problem is
3588 that you cannot compile C++ programs like @command{groff} 1.19.1
3589 without getting messages similar to the following:
3592 ld: warning: relocation error: R_SPARC_UA32: @dots{}
3593 external symbolic relocation against non-allocatable section
3594 .debug_info cannot be processed at runtime: relocation ignored.
3597 To work around this problem, compile with @option{-gstabs+} instead of
3600 When configuring the GNU Multiple Precision Library (GMP) on a Solaris 7
3601 or later system, the canonical target triplet must be specified as the
3602 @command{build} parameter on the configure line:
3605 ./configure --build=sparc-sun-solaris2.7 --prefix=xxx --enable-mpfr
3611 @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7
3613 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3614 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3615 and later, including all EGCS releases. Sun formerly recommended
3616 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3617 recommend it only for people who use Sun's compilers.
3619 Here are some workarounds to this problem:
3622 Do not install Sun patch 107058-01 until after Sun releases a
3623 complete patch for bug 4210064. This is the simplest course to take,
3624 unless you must also use Sun's C compiler. Unfortunately 107058-01
3625 is preinstalled on some new Solaris 7-based hosts, so you may have to
3629 Copy the original, unpatched Solaris 7
3630 @command{/usr/ccs/bin/as} into
3631 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
3632 adjusting the latter name to fit your local conventions and software
3636 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3637 both 107058-01 and 106950-03 installed has reported the bug with GCC
3638 and Sun's dynamic linker. This last course of action is riskiest,
3639 for two reasons. First, you must install 106950 on all hosts that
3640 run code generated by GCC; it doesn't suffice to install it only on
3641 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3642 only a partial fix for bug 4210064, but Sun doesn't know whether the
3643 partial fix is adequate for GCC@. Revision -08 or later should fix
3644 the bug. The current (as of 2004-05-23) revision is -24, and is included in
3645 the Solaris 7 Recommended Patch Cluster.
3648 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
3649 which causes a bootstrap failure when linking the 64-bit shared version of
3650 libgcc. A typical error message is:
3653 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
3654 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
3657 This bug has been fixed in the final 5.0 version of the assembler.
3662 @heading @anchor{sparc-*-linux*}sparc-*-linux*
3664 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
3665 or newer on this platform. All earlier binutils and glibc
3666 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
3672 @heading @anchor{sparc64-*-solaris2*}sparc64-*-solaris2*
3674 The following compiler flags must be specified in the configure
3675 step in order to bootstrap this target with the Sun compiler:
3678 % CC="cc -xildoff -xarch=v9" @var{srcdir}/configure [@var{options}] [@var{target}]
3681 @option{-xildoff} turns off the incremental linker, and @option{-xarch=v9}
3682 specifies the SPARC-V9 architecture to the Sun linker and assembler.
3687 @heading @anchor{sparcv9-*-solaris2*}sparcv9-*-solaris2*
3689 This is a synonym for sparc64-*-solaris2*.
3694 @heading @anchor{#*-*-sysv*}*-*-sysv*
3695 On System V release 3, you may get this error message
3699 ld fatal: failed to write symbol name @var{something}
3700 in strings table for file @var{whatever}
3703 This probably indicates that the disk is full or your ulimit won't allow
3704 the file to be as large as it needs to be.
3706 This problem can also result because the kernel parameter @code{MAXUMEM}
3707 is too small. If so, you must regenerate the kernel and make the value
3708 much larger. The default value is reported to be 1024; a value of 32768
3709 is said to work. Smaller values may also work.
3711 On System V, if you get an error like this,
3714 /usr/local/lib/bison.simple: In function `yyparse':
3715 /usr/local/lib/bison.simple:625: virtual memory exhausted
3719 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
3721 On a System V release 4 system, make sure @file{/usr/bin} precedes
3722 @file{/usr/ucb} in @code{PATH}. The @command{cc} command in
3723 @file{/usr/ucb} uses libraries which have bugs.
3728 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3729 Don't try compiling with VAX C (@command{vcc}). It produces incorrect code
3730 in some cases (for example, when @code{alloca} is used).
3735 @heading @anchor{*-*-vxworks*}*-*-vxworks*
3736 Support for VxWorks is in flux. At present GCC supports @emph{only} the
3737 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
3738 We welcome patches for other architectures supported by VxWorks 5.5.
3739 Support for VxWorks AE would also be welcome; we believe this is merely
3740 a matter of writing an appropriate ``configlette'' (see below). We are
3741 not interested in supporting older, a.out or COFF-based, versions of
3744 VxWorks comes with an older version of GCC installed in
3745 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
3746 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
3747 Before running @command{configure}, create the directories @file{@var{prefix}}
3748 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
3749 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
3750 include that directory while running both @command{configure} and
3753 You must give @command{configure} the
3754 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
3755 find the VxWorks system headers. Since VxWorks is a cross compilation
3756 target only, you must also specify @option{--target=@var{target}}.
3757 @command{configure} will attempt to create the directory
3758 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
3759 make sure the user running @command{configure} has sufficient privilege
3762 GCC's exception handling runtime requires a special ``configlette''
3763 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
3764 that file to add the module to your kernel build. (Future versions of
3765 VxWorks will incorporate this module.)
3770 @heading @anchor{x86_64-*-*}x86_64-*-*, amd64-*-*
3772 GCC supports the x86-64 architecture implemented by the AMD64 processor
3773 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
3774 On GNU/Linux the default is a bi-arch compiler which is able to generate
3775 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
3780 @heading @anchor{xtensa-*-elf}xtensa-*-elf
3782 This target is intended for embedded Xtensa systems using the
3783 @samp{newlib} C library. It uses ELF but does not support shared
3784 objects. Designed-defined instructions specified via the
3785 Tensilica Instruction Extension (TIE) language are only supported
3786 through inline assembly.
3788 The Xtensa configuration information must be specified prior to
3789 building GCC@. The @file{include/xtensa-config.h} header
3790 file contains the configuration information. If you created your
3791 own Xtensa configuration with the Xtensa Processor Generator, the
3792 downloaded files include a customized copy of this header file,
3793 which you can use to replace the default header file.
3798 @heading @anchor{xtensa-*-linux*}xtensa-*-linux*
3800 This target is for Xtensa systems running GNU/Linux. It supports ELF
3801 shared objects and the GNU C library (glibc). It also generates
3802 position-independent code (PIC) regardless of whether the
3803 @option{-fpic} or @option{-fPIC} options are used. In other
3804 respects, this target is the same as the
3805 @uref{#xtensa-*-elf,,@samp{xtensa-*-elf}} target.
3810 @heading @anchor{windows}Microsoft Windows (32-bit)
3812 A port of GCC 2.95.2 and 3.x is included with the
3813 @uref{http://www.cygwin.com/,,Cygwin environment}.
3815 Current (as of early 2001) snapshots of GCC will build under Cygwin
3816 without modification.
3818 GCC does not currently build with Microsoft's C++ compiler and there
3819 are no plans to make it do so.
3824 @heading @anchor{os2}OS/2
3826 GCC does not currently support OS/2. However, Andrew Zabolotny has been
3827 working on a generic OS/2 port with pgcc. The current code can be found
3828 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
3830 An older copy of GCC 2.8.1 is included with the EMX tools available at
3831 @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,,
3832 ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}.
3837 @heading @anchor{older}Older systems
3839 GCC contains support files for many older (1980s and early
3840 1990s) Unix variants. For the most part, support for these systems
3841 has not been deliberately removed, but it has not been maintained for
3842 several years and may suffer from bitrot.
3844 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
3845 Support for these systems is still present in that release, but
3846 @command{configure} will fail unless the @option{--enable-obsolete}
3847 option is given. Unless a maintainer steps forward, support for these
3848 systems will be removed from the next release of GCC@.
3850 Support for old systems as hosts for GCC can cause problems if the
3851 workarounds for compiler, library and operating system bugs affect the
3852 cleanliness or maintainability of the rest of GCC@. In some cases, to
3853 bring GCC up on such a system, if still possible with current GCC, may
3854 require first installing an old version of GCC which did work on that
3855 system, and using it to compile a more recent GCC, to avoid bugs in the
3856 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
3857 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
3858 sites}. Header bugs may generally be avoided using
3859 @command{fixincludes}, but bugs or deficiencies in libraries and the
3860 operating system may still cause problems.
3862 Support for older systems as targets for cross-compilation is less
3863 problematic than support for them as hosts for GCC; if an enthusiast
3864 wishes to make such a target work again (including resurrecting any of
3865 the targets that never worked with GCC 2, starting from the last CVS
3866 version before they were removed), patches
3867 @uref{../contribute.html,,following the usual requirements} would be
3868 likely to be accepted, since they should not affect the support for more
3871 For some systems, old versions of GNU binutils may also be useful,
3872 and are available from @file{pub/binutils/old-releases} on
3873 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
3875 Some of the information on specific systems above relates to
3876 such older systems, but much of the information
3877 about GCC on such systems (which may no longer be applicable to
3878 current GCC) is to be found in the GCC texinfo manual.
3883 @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris 2, etc.)
3885 C++ support is significantly better on ELF targets if you use the
3886 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
3887 inlines, vtables and template instantiations will be discarded
3896 @uref{./index.html,,Return to the GCC Installation page}
3900 @c ***Old documentation******************************************************
3902 @include install-old.texi
3908 @uref{./index.html,,Return to the GCC Installation page}
3912 @c ***GFDL********************************************************************
3920 @uref{./index.html,,Return to the GCC Installation page}
3924 @c ***************************************************************************
3925 @c Part 6 The End of the Document
3927 @comment node-name, next, previous, up
3928 @node Concept Index, , GNU Free Documentation License, Top
3932 @unnumbered Concept Index