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 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 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://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 1.4-gcj and 1.8.5
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{libiberty}, @file{libobjc} directories as well as any
324 of their subdirectories.
326 The Java directory @file{libjava} requires a modified version of
327 automake 1.4 downloadable from
328 @uref{ftp://gcc.gnu.org/pub/java/automake-gcj-1.4.tar.gz}.
330 Every other directory requires automake 1.8.5.
332 @item gettext version 0.12 (or later)
334 Needed to regenerate @file{gcc.pot}.
336 @item gperf version 2.7.2 (or later)
338 Necessary when modifying @command{gperf} input files, e.g.@:
339 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
340 @file{gcc/cp/cfns.h}.
342 @item expect version ???
343 @itemx tcl version ???
344 @itemx dejagnu version 1.4.4 (or later)
346 Necessary to run the GCC testsuite.
348 @item autogen version 5.5.4 (or later) and
349 @itemx guile version 1.4.1 (or later)
351 Necessary to regenerate @file{fixinc/fixincl.x} from
352 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
354 Necessary to run the @file{fixinc} @command{make check}.
356 Necessary to regenerate the top level @file{Makefile.in} file from
357 @file{Makefile.tpl} and @file{Makefile.def}.
359 @item GNU Bison version 1.28 (or later)
360 Berkeley @command{yacc} (@command{byacc}) is also reported to work other
363 Necessary when modifying @file{*.y} files.
365 Necessary to build GCC during development because the generated output
366 files are not included in the CVS repository. They are included in
369 @item Flex version 2.5.4 (or later)
371 Necessary when modifying @file{*.l} files.
373 Necessary to build GCC during development because the generated output
374 files are not included in the CVS repository. They are included in
377 @item Texinfo version 4.2 (or later)
379 Necessary for running @command{makeinfo} when modifying @file{*.texi}
380 files to test your changes.
382 Necessary to build GCC documentation during development because the
383 generated output files are not included in the CVS repository. They are
384 included in releases.
386 @item @TeX{} (any working version)
388 Necessary for running @command{texi2dvi}, used when running
389 @command{make dvi} to create DVI files.
391 @item cvs version 1.10 (or later)
392 @itemx ssh (any version)
394 Necessary to access the CVS repository. Public releases and weekly
395 snapshots of the development sources are also available via FTP.
397 @item perl version 5.6.1 (or later)
399 Necessary when regenerating @file{Makefile} dependencies in libiberty.
400 Necessary when regenerating @file{libiberty/functions.texi}.
401 Necessary when generating manpages from Texinfo manuals.
402 Used by various scripts to generate some files included in CVS (mainly
403 Unicode-related and rarely changing) from source tables.
405 @item GNU diffutils version 2.7 (or later)
407 Necessary when creating changes to GCC source code to submit for review.
409 @item patch version 2.5.4 (or later)
411 Necessary when applying patches, created with @command{diff}, to one's
421 @uref{./index.html,,Return to the GCC Installation page}
425 @c ***Downloading the source**************************************************
427 @comment node-name, next, previous, up
428 @node Downloading the source, Configuration, Prerequisites, Installing GCC
432 @chapter Downloading GCC
434 @cindex Downloading GCC
435 @cindex Downloading the Source
437 GCC is distributed via @uref{http://gcc.gnu.org/cvs.html,,CVS} and FTP
438 tarballs compressed with @command{gzip} or
439 @command{bzip2}. It is possible to download a full distribution or specific
442 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
443 for information on how to obtain GCC@.
445 The full distribution includes the C, C++, Objective-C, Fortran 77, Fortran
446 (in case of GCC 3.5 and later), Java, and Ada (in case of GCC 3.1 and later)
447 compilers. The full distribution also includes runtime libraries for C++,
448 Objective-C, Fortran 77, Fortran, and Java. In GCC 3.0 and later versions,
449 GNU compiler testsuites are also included in the full distribution.
451 If you choose to download specific components, you must download the core
452 GCC distribution plus any language specific distributions you wish to
453 use. The core distribution includes the C language front end as well as the
454 shared components. Each language has a tarball which includes the language
455 front end as well as the language runtime (when appropriate).
457 Unpack the core distribution as well as any language specific
458 distributions in the same directory.
460 If you also intend to build binutils (either to upgrade an existing
461 installation or for use in place of the corresponding tools of your
462 OS), unpack the binutils distribution either in the same directory or
463 a separate one. In the latter case, add symbolic links to any
464 components of the binutils you intend to build alongside the compiler
465 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
466 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
473 @uref{./index.html,,Return to the GCC Installation page}
477 @c ***Configuration***********************************************************
479 @comment node-name, next, previous, up
480 @node Configuration, Building, Downloading the source, Installing GCC
484 @chapter Installing GCC: Configuration
486 @cindex Configuration
487 @cindex Installing GCC: Configuration
489 Like most GNU software, GCC must be configured before it can be built.
490 This document describes the recommended configuration procedure
491 for both native and cross targets.
493 We use @var{srcdir} to refer to the toplevel source directory for
494 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
496 If you obtained the sources via CVS, @var{srcdir} must refer to the top
497 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
498 and not its @file{gcc} subdirectory, otherwise the build will fail.
500 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
501 file system, the shell's built-in @command{pwd} command will return
502 temporary pathnames. Using these can lead to various sorts of build
503 problems. To avoid this issue, set the @env{PWDCMD} environment
504 variable to an automounter-aware @command{pwd} command, e.g.,
505 @command{pawd} or @samp{amq -w}, during the configuration and build
508 First, we @strong{highly} recommend that GCC be built into a
509 separate directory than the sources which does @strong{not} reside
510 within the source tree. This is how we generally build GCC; building
511 where @var{srcdir} == @var{objdir} should still work, but doesn't
512 get extensive testing; building where @var{objdir} is a subdirectory
513 of @var{srcdir} is unsupported.
515 If you have previously built GCC in the same directory for a
516 different target machine, do @samp{make distclean} to delete all files
517 that might be invalid. One of the files this deletes is @file{Makefile};
518 if @samp{make distclean} complains that @file{Makefile} does not exist
519 or issues a message like ``don't know how to make distclean'' it probably
520 means that the directory is already suitably clean. However, with the
521 recommended method of building in a separate @var{objdir}, you should
522 simply use a different @var{objdir} for each target.
524 Second, when configuring a native system, either @command{cc} or
525 @command{gcc} must be in your path or you must set @env{CC} in
526 your environment before running configure. Otherwise the configuration
529 Note that the bootstrap compiler and the resulting GCC must be link
530 compatible, else the bootstrap will fail with linker errors about
531 incompatible object file formats. Several multilibed targets are
532 affected by this requirement, see
534 @ref{Specific, host/target specific installation notes}.
537 @uref{specific.html,,host/target specific installation notes}.
545 % @var{srcdir}/configure [@var{options}] [@var{target}]
549 @heading Target specification
552 GCC has code to correctly determine the correct value for @var{target}
553 for nearly all native systems. Therefore, we highly recommend you not
554 provide a configure target when configuring a native compiler.
557 @var{target} must be specified as @option{--target=@var{target}}
558 when configuring a cross compiler; examples of valid targets would be
559 m68k-coff, sh-elf, etc.
562 Specifying just @var{target} instead of @option{--target=@var{target}}
563 implies that the host defaults to @var{target}.
567 @heading Options specification
569 Use @var{options} to override several configure time options for
570 GCC@. A list of supported @var{options} follows; @samp{configure
571 --help} may list other options, but those not listed below may not
572 work and should not normally be used.
574 Note that each @option{--enable} option has a corresponding
575 @option{--disable} option and that each @option{--with} option has a
576 corresponding @option{--without} option.
579 @item --prefix=@var{dirname}
580 Specify the toplevel installation
581 directory. This is the recommended way to install the tools into a directory
582 other than the default. The toplevel installation directory defaults to
585 We @strong{highly} recommend against @var{dirname} being the same or a
586 subdirectory of @var{objdir} or vice versa. If specifying a directory
587 beneath a user's home directory tree, some shells will not expand
588 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
591 The following standard @command{autoconf} options are supported. Normally you
592 should not need to use these options.
594 @item --exec-prefix=@var{dirname}
595 Specify the toplevel installation directory for architecture-dependent
596 files. The default is @file{@var{prefix}}.
598 @item --bindir=@var{dirname}
599 Specify the installation directory for the executables called by users
600 (such as @command{gcc} and @command{g++}). The default is
601 @file{@var{exec-prefix}/bin}.
603 @item --libdir=@var{dirname}
604 Specify the installation directory for object code libraries and
605 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
607 @item --libexecdir=@var{dirname}
608 Specify the installation directory for internal executables of GCC@.
609 The default is @file{@var{exec-prefix}/libexec}.
611 @item --with-slibdir=@var{dirname}
612 Specify the installation directory for the shared libgcc library. The
613 default is @file{@var{libdir}}.
615 @item --infodir=@var{dirname}
616 Specify the installation directory for documentation in info format.
617 The default is @file{@var{prefix}/info}.
619 @item --datadir=@var{dirname}
620 Specify the installation directory for some architecture-independent
621 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
623 @item --mandir=@var{dirname}
624 Specify the installation directory for manual pages. The default is
625 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
626 the full GCC manuals, which are provided in Texinfo format. The manpages
627 are derived by an automatic conversion process from parts of the full
630 @item --with-gxx-include-dir=@var{dirname}
632 the installation directory for G++ header files. The default is
633 @file{@var{prefix}/include/c++/@var{version}}.
637 @item --program-prefix=@var{prefix}
638 GCC supports some transformations of the names of its programs when
639 installing them. This option prepends @var{prefix} to the names of
640 programs to install in @var{bindir} (see above). For example, specifying
641 @option{--program-prefix=foo-} would result in @samp{gcc}
642 being installed as @file{/usr/local/bin/foo-gcc}.
644 @item --program-suffix=@var{suffix}
645 Appends @var{suffix} to the names of programs to install in @var{bindir}
646 (see above). For example, specifying @option{--program-suffix=-3.1}
647 would result in @samp{gcc} being installed as
648 @file{/usr/local/bin/gcc-3.1}.
650 @item --program-transform-name=@var{pattern}
651 Applies the @samp{sed} script @var{pattern} to be applied to the names
652 of programs to install in @var{bindir} (see above). @var{pattern} has to
653 consist of one or more basic @samp{sed} editing commands, separated by
654 semicolons. For example, if you want the @samp{gcc} program name to be
655 transformed to the installed program @file{/usr/local/bin/myowngcc} and
656 the @samp{g++} program name to be transformed to
657 @file{/usr/local/bin/gspecial++} without changing other program names,
658 you could use the pattern
659 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
660 to achieve this effect.
662 All three options can be combined and used together, resulting in more
663 complex conversion patterns. As a basic rule, @var{prefix} (and
664 @var{suffix}) are prepended (appended) before further transformations
665 can happen with a special transformation script @var{pattern}.
667 As currently implemented, this option only takes effect for native
668 builds; cross compiler binaries' names are not transformed even when a
669 transformation is explicitly asked for by one of these options.
671 For native builds, some of the installed programs are also installed
672 with the target alias in front of their name, as in
673 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
674 before the target alias is prepended to the name - so, specifying
675 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
676 resulting binary would be installed as
677 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
679 As a last shortcoming, none of the installed Ada programs are
680 transformed yet, which will be fixed in some time.
682 @item --with-local-prefix=@var{dirname}
684 installation directory for local include files. The default is
685 @file{/usr/local}. Specify this option if you want the compiler to
686 search directory @file{@var{dirname}/include} for locally installed
687 header files @emph{instead} of @file{/usr/local/include}.
689 You should specify @option{--with-local-prefix} @strong{only} if your
690 site has a different convention (not @file{/usr/local}) for where to put
693 The default value for @option{--with-local-prefix} is @file{/usr/local}
694 regardless of the value of @option{--prefix}. Specifying
695 @option{--prefix} has no effect on which directory GCC searches for
696 local header files. This may seem counterintuitive, but actually it is
699 The purpose of @option{--prefix} is to specify where to @emph{install
700 GCC}. The local header files in @file{/usr/local/include}---if you put
701 any in that directory---are not part of GCC@. They are part of other
702 programs---perhaps many others. (GCC installs its own header files in
703 another directory which is based on the @option{--prefix} value.)
705 Both the local-prefix include directory and the GCC-prefix include
706 directory are part of GCC's "system include" directories. Although these
707 two directories are not fixed, they need to be searched in the proper
708 order for the correct processing of the include_next directive. The
709 local-prefix include directory is searched before the GCC-prefix
710 include directory. Another characteristic of system include directories
711 is that pedantic warnings are turned off for headers in these directories.
713 Some autoconf macros add @option{-I @var{directory}} options to the
714 compiler command line, to ensure that directories containing installed
715 packages' headers are searched. When @var{directory} is one of GCC's
716 system include directories, GCC will ignore the option so that system
717 directories continue to be processed in the correct order. This
718 may result in a search order different from what was specified but the
719 directory will still be searched.
721 GCC automatically searches for ordinary libraries using
722 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
723 used for both GCC and packages, GCC will automatically search for
724 both headers and libraries. This provides a configuration that is
725 easy to use. GCC behaves in a manner similar to that when it is
726 installed as a system compiler in @file{/usr}.
728 Sites that need to install multiple versions of GCC may not want to
729 use the above simple configuration. It is possible to use the
730 @option{--program-prefix}, @option{--program-suffix} and
731 @option{--program-transform-name} options to install multiple versions
732 into a single directory, but it may be simpler to use different prefixes
733 and the @option{--with-local-prefix} option to specify the location of the
734 site-specific files for each version. It will then be necessary for
735 users to specify explicitly the location of local site libraries
736 (e.g., with @env{LIBRARY_PATH}).
738 The same value can be used for both @option{--with-local-prefix} and
739 @option{--prefix} provided it is not @file{/usr}. This can be used
740 to avoid the default search of @file{/usr/local/include}.
742 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
743 The directory you use for @option{--with-local-prefix} @strong{must not}
744 contain any of the system's standard header files. If it did contain
745 them, certain programs would be miscompiled (including GNU Emacs, on
746 certain targets), because this would override and nullify the header
747 file corrections made by the @command{fixincludes} script.
749 Indications are that people who use this option use it based on mistaken
750 ideas of what it is for. People use it as if it specified where to
751 install part of GCC@. Perhaps they make this assumption because
752 installing GCC creates the directory.
754 @item --enable-shared[=@var{package}[,@dots{}]]
755 Build shared versions of libraries, if shared libraries are supported on
756 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
757 are enabled by default on all platforms that support shared libraries.
759 If a list of packages is given as an argument, build shared libraries
760 only for the listed packages. For other packages, only static libraries
761 will be built. Package names currently recognized in the GCC tree are
762 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
763 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
764 @samp{ada}, @samp{libada}, @samp{libjava} and @samp{libobjc}.
765 Note @samp{libiberty} does not support shared libraries at all.
767 Use @option{--disable-shared} to build only static libraries. Note that
768 @option{--disable-shared} does not accept a list of package names as
769 argument, only @option{--enable-shared} does.
771 @item @anchor{with-gnu-as}--with-gnu-as
772 Specify that the compiler should assume that the
773 assembler it finds is the GNU assembler. However, this does not modify
774 the rules to find an assembler and will result in confusion if the
775 assembler found is not actually the GNU assembler. (Confusion may also
776 result if the compiler finds the GNU assembler but has not been
777 configured with @option{--with-gnu-as}.) If you have more than one
778 assembler installed on your system, you may want to use this option in
779 connection with @option{--with-as=@var{pathname}}.
781 The following systems are the only ones where it makes a difference
782 whether you use the GNU assembler. On any other system,
783 @option{--with-gnu-as} has no effect.
786 @item @samp{hppa1.0-@var{any}-@var{any}}
787 @item @samp{hppa1.1-@var{any}-@var{any}}
788 @item @samp{i386-@var{any}-sysv}
789 @item @samp{m68k-bull-sysv}
790 @item @samp{m68k-hp-hpux}
791 @item @samp{m68000-hp-hpux}
792 @item @samp{m68000-att-sysv}
793 @item @samp{mips-@var{any}}
794 @item @samp{sparc-sun-solaris2.@var{any}}
795 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
798 On the systems listed above (except for the HP-PA, the SPARC, for ISC on
799 the 386, and for @samp{mips-sgi-irix5.*}), if you use the GNU assembler,
800 you should also use the GNU linker (and specify @option{--with-gnu-ld}).
802 @item @anchor{with-as}--with-as=@var{pathname}
804 compiler should use the assembler pointed to by @var{pathname}, rather
805 than the one found by the standard rules to find an assembler, which
809 Check the @file{@var{libexec}/gcc/@var{target}/@var{version}}
810 directory, where @var{libexec} defaults to
811 @file{@var{exec-prefix}/libexec} and @var{exec-prefix} defaults to
812 @var{prefix} which defaults to @file{/usr/local} unless overridden by
813 the @option{--prefix=@var{pathname}} switch described
814 above. @var{target} is the target system triple, such as
815 @samp{sparc-sun-solaris2.7}, and @var{version} denotes the GCC
816 version, such as 3.0.
818 Check operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
821 Note that these rules do not check for the value of @env{PATH}. You may
822 want to use @option{--with-as} if no assembler is installed in the
823 directories listed above, or if you have multiple assemblers installed
824 and want to choose one that is not found by the above rules.
826 @item @anchor{with-gnu-ld}--with-gnu-ld
827 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
830 @item --with-ld=@var{pathname}
831 Same as @uref{#with-as,,@option{--with-as}}
835 Specify that stabs debugging
836 information should be used instead of whatever format the host normally
837 uses. Normally GCC uses the same debug format as the host system.
839 On MIPS based systems and on Alphas, you must specify whether you want
840 GCC to create the normal ECOFF debugging format, or to use BSD-style
841 stabs passed through the ECOFF symbol table. The normal ECOFF debug
842 format cannot fully handle languages other than C@. BSD stabs format can
843 handle other languages, but it only works with the GNU debugger GDB@.
845 Normally, GCC uses the ECOFF debugging format by default; if you
846 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
848 No matter which default you choose when you configure GCC, the user
849 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
850 the debug format for a particular compilation.
852 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
853 @option{--with-gas} is used. It selects use of stabs debugging
854 information embedded in COFF output. This kind of debugging information
855 supports C++ well; ordinary COFF debugging information does not.
857 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
858 selects use of stabs debugging information embedded in ELF output. The
859 C++ compiler currently (2.6.0) does not support the DWARF debugging
860 information normally used on 386 SVR4 platforms; stabs provide a
861 workable alternative. This requires gas and gdb, as the normal SVR4
862 tools can not generate or interpret stabs.
864 @item --disable-multilib
865 Specify that multiple target
866 libraries to support different target variants, calling
867 conventions, etc should not be built. The default is to build a
868 predefined set of them.
870 Some targets provide finer-grained control over which multilibs are built
871 (e.g., @option{--disable-softfloat}):
877 fpu, 26bit, underscore, interwork, biendian, nofmult.
880 softfloat, m68881, m68000, m68020.
883 single-float, biendian, softfloat.
885 @item powerpc*-*-*, rs6000*-*-*
886 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
891 @item --enable-threads
892 Specify that the target
893 supports threads. This affects the Objective-C compiler and runtime
894 library, and exception handling for other languages like C++ and Java.
895 On some systems, this is the default.
897 In general, the best (and, in many cases, the only known) threading
898 model available will be configured for use. Beware that on some
899 systems, GCC has not been taught what threading models are generally
900 available for the system. In this case, @option{--enable-threads} is an
901 alias for @option{--enable-threads=single}.
903 @item --disable-threads
904 Specify that threading support should be disabled for the system.
905 This is an alias for @option{--enable-threads=single}.
907 @item --enable-threads=@var{lib}
909 @var{lib} is the thread support library. This affects the Objective-C
910 compiler and runtime library, and exception handling for other languages
911 like C++ and Java. The possibilities for @var{lib} are:
919 Ada tasking support. For non-Ada programs, this setting is equivalent
920 to @samp{single}. When used in conjunction with the Ada run time, it
921 causes GCC to use the same thread primitives as Ada uses. This option
922 is necessary when using both Ada and the back end exception handling,
923 which is the default for most Ada targets.
925 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
926 that the file needed to support this configuration, @file{gthr-mach.h}, is
927 missing and thus this setting will cause a known bootstrap failure.)
929 This is an alias for @samp{single}.
931 Generic POSIX 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.
944 @item --with-cpu=@var{cpu}
945 Specify which cpu variant the compiler should generate code for by default.
946 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
947 This option is only supported on some targets, including ARM, i386, PowerPC,
950 @item --with-schedule=@var{cpu}
951 @itemx --with-arch=@var{cpu}
952 @itemx --with-tune=@var{cpu}
953 @itemx --with-abi=@var{abi}
954 @itemx --with-fpu=@var{type}
955 @itemx --with-float=@var{type}
956 These configure options provide default values for the @option{-mschedule=},
957 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
958 options and for @option{-mhard-float} or @option{-msoft-float}. As with
959 @option{--with-cpu}, which switches will be accepted and acceptable values
960 of the arguments depend on the target.
962 @item --enable-altivec
963 Specify that the target supports AltiVec vector enhancements. This
964 option will adjust the ABI for AltiVec enhancements, as well as generate
965 AltiVec code when appropriate. This option is only available for
968 @item --enable-__cxa_atexit
969 Define if you want to use __cxa_atexit, rather than atexit, to
970 register C++ destructors for local statics and global objects.
971 This is essential for fully standards-compliant handling of
972 destructors, but requires __cxa_atexit in libc. This option is currently
973 only available on systems with GNU libc. When enabled, this will cause
974 @option{-fuse-cxa-exit} to be passed by default.
976 @item --enable-target-optspace
978 libraries should be optimized for code space instead of code speed.
979 This is the default for the m32r platform.
982 Specify that a user visible @command{cpp} program should not be installed.
984 @item --with-cpp-install-dir=@var{dirname}
985 Specify that the user visible @command{cpp} program should be installed
986 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
988 @item --enable-initfini-array
989 Force the use of sections @code{.init_array} and @code{.fini_array}
990 (instead of @code{.init} and @code{.fini}) for constructors and
991 destructors. Option @option{--disable-initfini-array} has the
992 opposite effect. If neither option is specified, the configure script
993 will try to guess whether the @code{.init_array} and
994 @code{.fini_array} sections are supported and, if they are, use them.
996 @item --enable-maintainer-mode
998 regenerate the GCC master message catalog @file{gcc.pot} are normally
999 disabled. This is because it can only be rebuilt if the complete source
1000 tree is present. If you have changed the sources and want to rebuild the
1001 catalog, configuring with @option{--enable-maintainer-mode} will enable
1002 this. Note that you need a recent version of the @code{gettext} tools
1005 @item --enable-generated-files-in-srcdir
1006 Neither the .c and .h files that are generated from bison and flex nor the
1007 info manuals and man pages that are built from the .texi files are present
1008 in the CVS development tree. When building GCC from that development tree,
1009 or from a snapshot which are created from CVS, then those generated files
1010 are placed in your build directory, which allows for the source to be in a
1013 If you configure with @option{--enable-generated-files-in-srcdir} then those
1014 generated files will go into the source directory. This is mainly intended
1015 for generating release or prerelease tarballs of the GCC sources, since it
1016 is not a requirement that the users of source releases to have flex, bison, or
1019 @item --enable-version-specific-runtime-libs
1021 that runtime libraries should be installed in the compiler specific
1022 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1023 addition, @samp{libstdc++}'s include files will be installed into
1024 @file{@var{libdir}} unless you overruled it by using
1025 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1026 particularly useful if you intend to use several versions of GCC in
1027 parallel. This is currently supported by @samp{libgfortran},
1028 @samp{libjava}, @samp{libmudflap}, @samp{libstdc++}, and @samp{libobjc}.
1031 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1032 Specify that only a particular subset of compilers and
1033 their runtime libraries should be built. For a list of valid values for
1034 @var{langN} you can issue the following command in the
1035 @file{gcc} directory of your GCC source tree:@*
1037 grep language= */config-lang.in
1039 Currently, you can use any of the following:
1040 @code{ada}, @code{c}, @code{c++}, @code{f77}, @code{f95}, @code{java},
1041 @code{objc}, @code{obj-c++}.
1042 Building the Ada compiler has special requirements, see below.@*
1043 If you do not pass this flag, all languages available in the @file{gcc}
1044 sub-tree will be configured. Re-defining @code{LANGUAGES} when calling
1045 @samp{make bootstrap} @strong{does not} work anymore, as those
1046 language sub-directories might not have been configured!
1048 @item --disable-libada
1049 Specify that the run-time libraries and tools used by GNAT should not
1050 be built. This can be useful for debugging, or for compatibility with
1051 previous Ada build procedures, when it was required to explicitly
1052 do a @samp{make -C gcc gnatlib_and_tools}.
1055 Specify that the compiler should
1056 use DWARF 2 debugging information as the default.
1058 @item --enable-win32-registry
1059 @itemx --enable-win32-registry=@var{key}
1060 @itemx --disable-win32-registry
1061 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1062 to look up installations paths in the registry using the following key:
1065 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1068 @var{key} defaults to GCC version number, and can be overridden by the
1069 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1070 who use custom installers are encouraged to provide a different key,
1071 perhaps one comprised of vendor name and GCC version number, to
1072 avoid conflict with existing installations. This feature is enabled
1073 by default, and can be disabled by @option{--disable-win32-registry}
1074 option. This option has no effect on the other hosts.
1077 Specify that the machine does not have a floating point unit. This
1078 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1079 system, @option{--nfp} has no effect.
1081 @item --enable-werror
1082 @itemx --disable-werror
1083 @itemx --enable-werror=yes
1084 @itemx --enable-werror=no
1085 When you specify this option, it controls whether certain files in the
1086 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1087 If you don't specify it, @option{-Werror} is turned on for the main
1088 development trunk. However it defaults to off for release branches and
1089 final releases. The specific files which get @option{-Werror} are
1090 controlled by the Makefiles.
1092 @item --enable-checking
1093 @itemx --enable-checking=@var{list}
1094 When you specify this option, the compiler is built to perform checking
1095 of tree node types when referencing fields of that node, and some other
1096 internal consistency checks. This does not change the generated code,
1097 but adds error checking within the compiler. This will slow down the
1098 compiler and may only work properly if you are building the compiler
1099 with GCC@. This is on by default when building from CVS or snapshots,
1100 but off for releases. More control over the checks may be had by
1101 specifying @var{list}; the categories of checks available are
1102 @samp{misc}, @samp{tree}, @samp{gc}, @samp{rtl}, @samp{rtlflag},
1103 @samp{fold}, @samp{gcac} and @samp{valgrind}. The check @samp{valgrind}
1104 requires the external @command{valgrind} simulator, available from
1105 @uref{http://valgrind.kde.org/}. The default when @var{list} is
1106 not specified is @samp{misc,tree,gc,rtlflag}; the checks @samp{rtl},
1107 @samp{gcac} and @samp{valgrind} are very expensive.
1109 @item --enable-coverage
1110 @itemx --enable-coverage=@var{level}
1111 With this option, the compiler is built to collect self coverage
1112 information, every time it is run. This is for internal development
1113 purposes, and only works when the compiler is being built with gcc. The
1114 @var{level} argument controls whether the compiler is built optimized or
1115 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1116 want to disable optimization, for performance analysis you want to
1117 enable optimization. When coverage is enabled, the default level is
1118 without optimization.
1120 @item --enable-gather-detailed-mem-stats
1121 When this option is specified more detailed information on memory
1122 allocation is gathered. This information is printed when using
1123 @option{-fmem-report}.
1126 @itemx --disable-nls
1127 The @option{--enable-nls} option enables Native Language Support (NLS),
1128 which lets GCC output diagnostics in languages other than American
1129 English. Native Language Support is enabled by default if not doing a
1130 canadian cross build. The @option{--disable-nls} option disables NLS@.
1132 @item --with-included-gettext
1133 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1134 procedure to prefer its copy of GNU @command{gettext}.
1136 @item --with-catgets
1137 If NLS is enabled, and if the host lacks @code{gettext} but has the
1138 inferior @code{catgets} interface, the GCC build procedure normally
1139 ignores @code{catgets} and instead uses GCC's copy of the GNU
1140 @code{gettext} library. The @option{--with-catgets} option causes the
1141 build procedure to use the host's @code{catgets} in this situation.
1143 @item --with-libiconv-prefix=@var{dir}
1144 Search for libiconv header files in @file{@var{dir}/include} and
1145 libiconv library files in @file{@var{dir}/lib}.
1147 @item --enable-obsolete
1148 Enable configuration for an obsoleted system. If you attempt to
1149 configure GCC for a system (build, host, or target) which has been
1150 obsoleted, and you do not specify this flag, configure will halt with an
1153 All support for systems which have been obsoleted in one release of GCC
1154 is removed entirely in the next major release, unless someone steps
1155 forward to maintain the port.
1158 @subheading Cross-Compiler-Specific Options
1159 The following options only apply to building cross compilers.
1161 @item --with-sysroot
1162 @itemx --with-sysroot=@var{dir}
1163 Tells GCC to consider @var{dir} as the root of a tree that contains a
1164 (subset of) the root filesystem of the target operating system.
1165 Target system headers, libraries and run-time object files will be
1166 searched in there. The specified directory is not copied into the
1167 install tree, unlike the options @option{--with-headers} and
1168 @option{--with-libs} that this option obsoletes. The default value,
1169 in case @option{--with-sysroot} is not given an argument, is
1170 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1171 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1172 the GCC binaries if the installation tree is moved.
1174 @item --with-headers
1175 @itemx --with-headers=@var{dir}
1176 Deprecated in favor of @option{--with-sysroot}.
1177 Specifies that target headers are available when building a cross compiler.
1178 The @var{dir} argument specifies a directory which has the target include
1179 files. These include files will be copied into the @file{gcc} install
1180 directory. @emph{This option with the @var{dir} argument is required} when
1181 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1182 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1183 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1184 will be run on these files to make them compatible with GCC.
1186 @item --without-headers
1187 Tells GCC not use any target headers from a libc when building a cross
1188 compiler. When crossing to GNU/Linux, you need the headers so GCC
1189 can build the exception handling for libgcc.
1190 See @uref{http://www.objsw.com/CrossGCC/,,CrossGCC} for more information
1194 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1195 Deprecated in favor of @option{--with-sysroot}.
1196 Specifies a list of directories which contain the target runtime
1197 libraries. These libraries will be copied into the @file{gcc} install
1198 directory. If the directory list is omitted, this option has no
1201 Specifies that @samp{newlib} is
1202 being used as the target C library. This causes @code{__eprintf} to be
1203 omitted from @file{libgcc.a} on the assumption that it will be provided by
1207 @subheading Fortran-specific Option
1209 The following options apply to the build of the Fortran front end.
1213 @item --with-gmp=@var{pathname}
1214 @itemx --with-mpfr=@var{pathname}
1215 @itemx --with-gmp-dir=@var{pathname}
1216 @itemx --with-mpfr-dir=@var{pathname}
1217 If you don't have GMP (the GNU Multiple Precision library) and the MPFR
1218 Libraries installed in a standard location and you want to build the Fortran
1219 front-end, you can explicitly specify the directory where they are installed
1220 (@samp{--with-gmp=gmpinstalldir}, @samp{--with-mpfr=mpfrinstalldir}) or where
1221 you built them without installing (@samp{--with-gmp-dir=gmpbuilddir},
1222 @samp{--with-mpfr-dir=gmpbuilddir}).
1226 @subheading Java-Specific Options
1228 The following option applies to the build of the Java front end.
1231 @item --disable-libgcj
1232 Specify that the run-time libraries
1233 used by GCJ should not be built. This is useful in case you intend
1234 to use GCJ with some other run-time, or you're going to install it
1235 separately, or it just happens not to build on your particular
1236 machine. In general, if the Java front end is enabled, the GCJ
1237 libraries will be enabled too, unless they're known to not work on
1238 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1239 may need to port it; in this case, before modifying the top-level
1240 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1241 you may use @option{--enable-libgcj} to override the default.
1245 The following options apply to building @samp{libgcj}.
1247 @subsubheading General Options
1250 @item --disable-getenv-properties
1251 Don't set system properties from @env{GCJ_PROPERTIES}.
1253 @item --enable-hash-synchronization
1254 Use a global hash table for monitor locks. Ordinarily,
1255 @samp{libgcj}'s @samp{configure} script automatically makes
1256 the correct choice for this option for your platform. Only use
1257 this if you know you need the library to be configured differently.
1259 @item --enable-interpreter
1260 Enable the Java interpreter. The interpreter is automatically
1261 enabled by default on all platforms that support it. This option
1262 is really only useful if you want to disable the interpreter
1263 (using @option{--disable-interpreter}).
1265 @item --disable-java-net
1266 Disable java.net. This disables the native part of java.net only,
1267 using non-functional stubs for native method implementations.
1269 @item --disable-jvmpi
1270 Disable JVMPI support.
1273 Enable runtime eCos target support.
1275 @item --without-libffi
1276 Don't use @samp{libffi}. This will disable the interpreter and JNI
1277 support as well, as these require @samp{libffi} to work.
1279 @item --enable-libgcj-debug
1280 Enable runtime debugging code.
1282 @item --enable-libgcj-multifile
1283 If specified, causes all @file{.java} source files to be
1284 compiled into @file{.class} files in one invocation of
1285 @samp{gcj}. This can speed up build time, but is more
1286 resource-intensive. If this option is unspecified or
1287 disabled, @samp{gcj} is invoked once for each @file{.java}
1288 file to compile into a @file{.class} file.
1290 @item --with-libiconv-prefix=DIR
1291 Search for libiconv in @file{DIR/include} and @file{DIR/lib}.
1293 @item --enable-sjlj-exceptions
1294 Force use of @code{builtin_setjmp} for exceptions. @samp{configure}
1295 ordinarily picks the correct value based on the platform. Only use
1296 this option if you are sure you need a different setting.
1298 @item --with-system-zlib
1299 Use installed @samp{zlib} rather than that included with GCC@.
1301 @item --with-win32-nlsapi=ansi, unicows or unicode
1302 Indicates how MinGW @samp{libgcj} translates between UNICODE
1303 characters and the Win32 API.
1306 Use the single-byte @code{char} and the Win32 A functions natively,
1307 translating to and from UNICODE when using these functions. If
1308 unspecified, this is the default.
1311 Use the @code{WCHAR} and Win32 W functions natively. Adds
1312 @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}.
1313 @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines
1314 running built executables. @file{libunicows.a}, an open-source
1315 import library around Microsoft's @code{unicows.dll}, is obtained from
1316 @uref{http://libunicows.sourceforge.net/}, which also gives details
1317 on getting @file{unicows.dll} from Microsoft.
1320 Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not}
1321 add @code{-lunicows} to @file{libgcj.spec}. The built executables will
1322 only run on Microsoft Windows NT and above.
1326 @subsubheading AWT-Specific Options
1330 Use the X Window System.
1332 @item --enable-java-awt=PEER(S)
1333 Specifies the AWT peer library or libraries to build alongside
1334 @samp{libgcj}. If this option is unspecified or disabled, AWT
1335 will be non-functional. Current valid values are @option{gtk} and
1336 @option{xlib}. Multiple libraries should be separated by a
1337 comma (i.e. @option{--enable-java-awt=gtk,xlib}).
1339 @item --enable-gtk-cairo
1340 Build the cairo Graphics2D implementation on GTK.
1342 @item --enable-java-gc=TYPE
1343 Choose garbage collector. Defaults to @option{boehm} if unspecified.
1345 @item --disable-gtktest
1346 Do not try to compile and run a test GTK+ program.
1348 @item --disable-glibtest
1349 Do not try to compile and run a test GLIB program.
1351 @item --with-libart-prefix=PFX
1352 Prefix where libart is installed (optional).
1354 @item --with-libart-exec-prefix=PFX
1355 Exec prefix where libart is installed (optional).
1357 @item --disable-libarttest
1358 Do not try to compile and run a test libart program.
1367 @uref{./index.html,,Return to the GCC Installation page}
1371 @c ***Building****************************************************************
1373 @comment node-name, next, previous, up
1374 @node Building, Testing, Configuration, Installing GCC
1380 @cindex Installing GCC: Building
1382 Now that GCC is configured, you are ready to build the compiler and
1385 Some commands executed when making the compiler may fail (return a
1386 nonzero status) and be ignored by @command{make}. These failures, which
1387 are often due to files that were not found, are expected, and can safely
1390 It is normal to have compiler warnings when compiling certain files.
1391 Unless you are a GCC developer, you can generally ignore these warnings
1392 unless they cause compilation to fail. Developers should attempt to fix
1393 any warnings encountered, however they can temporarily continue past
1394 warnings-as-errors by specifying the configure flag
1395 @option{--disable-werror}.
1397 On certain old systems, defining certain environment variables such as
1398 @env{CC} can interfere with the functioning of @command{make}.
1400 If you encounter seemingly strange errors when trying to build the
1401 compiler in a directory other than the source directory, it could be
1402 because you have previously configured the compiler in the source
1403 directory. Make sure you have done all the necessary preparations.
1405 If you build GCC on a BSD system using a directory stored in an old System
1406 V file system, problems may occur in running @command{fixincludes} if the
1407 System V file system doesn't support symbolic links. These problems
1408 result in a failure to fix the declaration of @code{size_t} in
1409 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1410 that type mismatches occur, this could be the cause.
1412 The solution is not to use such a directory for building GCC@.
1414 When building from CVS or snapshots, or if you modify parser sources,
1415 you need the Bison parser generator installed. Any version 1.25 or
1416 later should work; older versions may also work. If you do not modify
1417 parser sources, releases contain the Bison-generated files and you do
1418 not need Bison installed to build them.
1420 When building from CVS or snapshots, or if you modify Texinfo
1421 documentation, you need version 4.2 or later of Texinfo installed if you
1422 want Info documentation to be regenerated. Releases contain Info
1423 documentation pre-built for the unmodified documentation in the release.
1425 @section Building a native compiler
1427 For a native build issue the command @samp{make bootstrap}. This
1428 will build the entire GCC system, which includes the following steps:
1432 Build host tools necessary to build the compiler such as texinfo, bison,
1436 Build target tools for use by the compiler such as binutils (bfd,
1437 binutils, gas, gprof, ld, and opcodes)
1438 if they have been individually linked
1439 or moved into the top level GCC source tree before configuring.
1442 Perform a 3-stage bootstrap of the compiler.
1445 Perform a comparison test of the stage2 and stage3 compilers.
1448 Build runtime libraries using the stage3 compiler from the previous step.
1452 If you are short on disk space you might consider @samp{make
1453 bootstrap-lean} instead. This is identical to @samp{make
1454 bootstrap} except that object files from the stage1 and
1455 stage2 of the 3-stage bootstrap of the compiler are deleted as
1456 soon as they are no longer needed.
1458 If you want to save additional space during the bootstrap and in
1459 the final installation as well, you can build the compiler binaries
1460 without debugging information as in the following example. This will save
1461 roughly 40% of disk space both for the bootstrap and the final installation.
1462 (Libraries will still contain debugging information.)
1465 make CFLAGS='-O' LIBCFLAGS='-g -O2' \
1466 LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap
1469 If you wish to use non-default GCC flags when compiling the stage2 and
1470 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
1471 @samp{make bootstrap}. Non-default optimization flags are less well
1472 tested here than the default of @samp{-g -O2}, but should still work.
1473 In a few cases, you may find that you need to specify special flags such
1474 as @option{-msoft-float} here to complete the bootstrap; or, if the
1475 native compiler miscompiles the stage1 compiler, you may need to work
1476 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
1477 stage1 compiler that were miscompiled, or by using @samp{make
1478 bootstrap4} to increase the number of stages of bootstrap.
1480 Note that using non-standard @code{CFLAGS} can cause bootstrap to fail in
1481 @file{libiberty}, if these trigger a warning with the new compiler. For
1482 example using @samp{-O2 -g -mcpu=i686} on @code{i686-pc-linux-gnu} will
1483 cause bootstrap failure as @code{-mcpu=} is deprecated in 3.4.0 and above.
1486 If you used the flag @option{--enable-languages=@dots{}} to restrict
1487 the compilers to be built, only those you've actually enabled will be
1488 built. This will of course only build those runtime libraries, for
1489 which the particular compiler has been built. Please note,
1490 that re-defining @env{LANGUAGES} when calling @samp{make bootstrap}
1491 @strong{does not} work anymore!
1493 If the comparison of stage2 and stage3 fails, this normally indicates
1494 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1495 a potentially serious bug which you should investigate and report. (On
1496 a few systems, meaningful comparison of object files is impossible; they
1497 always appear ``different''. If you encounter this problem, you will
1498 need to disable comparison in the @file{Makefile}.)
1500 @section Building a cross compiler
1502 We recommend reading the
1503 @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ}
1504 for information about building cross compilers.
1506 When building a cross compiler, it is not generally possible to do a
1507 3-stage bootstrap of the compiler. This makes for an interesting problem
1508 as parts of GCC can only be built with GCC@.
1510 To build a cross compiler, we first recommend building and installing a
1511 native compiler. You can then use the native GCC compiler to build the
1512 cross compiler. The installed native compiler needs to be GCC version
1515 Assuming you have already installed a native copy of GCC and configured
1516 your cross compiler, issue the command @command{make}, which performs the
1521 Build host tools necessary to build the compiler such as texinfo, bison,
1525 Build target tools for use by the compiler such as binutils (bfd,
1526 binutils, gas, gprof, ld, and opcodes)
1527 if they have been individually linked or moved into the top level GCC source
1528 tree before configuring.
1531 Build the compiler (single stage only).
1534 Build runtime libraries using the compiler from the previous step.
1537 Note that if an error occurs in any step the make process will exit.
1539 If you are not building GNU binutils in the same source tree as GCC,
1540 you will need a cross-assembler and cross-linker installed before
1541 configuring GCC@. Put them in the directory
1542 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
1543 you should put in this directory:
1547 This should be the cross-assembler.
1550 This should be the cross-linker.
1553 This should be the cross-archiver: a program which can manipulate
1554 archive files (linker libraries) in the target machine's format.
1557 This should be a program to construct a symbol table in an archive file.
1560 The installation of GCC will find these programs in that directory,
1561 and copy or link them to the proper place to for the cross-compiler to
1562 find them when run later.
1564 The easiest way to provide these files is to build the Binutils package.
1565 Configure it with the same @option{--host} and @option{--target}
1566 options that you use for configuring GCC, then build and install
1567 them. They install their executables automatically into the proper
1568 directory. Alas, they do not support all the targets that GCC
1571 If you are not building a C library in the same source tree as GCC,
1572 you should also provide the target libraries and headers before
1573 configuring GCC, specifying the directories with
1574 @option{--with-sysroot} or @option{--with-headers} and
1575 @option{--with-libs}. Many targets also require ``start files'' such
1576 as @file{crt0.o} and
1577 @file{crtn.o} which are linked into each executable. There may be several
1578 alternatives for @file{crt0.o}, for use with profiling or other
1579 compilation options. Check your target's definition of
1580 @code{STARTFILE_SPEC} to find out what start files it uses.
1582 @section Building in parallel
1584 You can use @samp{make bootstrap MAKE="make -j 2" -j 2}, or just
1585 @samp{make -j 2 bootstrap} for GNU Make 3.79 and above, instead of
1586 @samp{make bootstrap} to build GCC in parallel.
1587 You can also specify a bigger number, and in most cases using a value
1588 greater than the number of processors in your machine will result in
1589 fewer and shorter I/O latency hits, thus improving overall throughput;
1590 this is especially true for slow drives and network filesystems.
1592 @section Building the Ada compiler
1594 In order to build GNAT, the Ada compiler, you need a working GNAT
1595 compiler (GNAT version 3.14 or later, or GCC version 3.1 or later),
1596 including GNAT tools such as @command{gnatmake} and @command{gnatlink},
1597 since the Ada front end is written in Ada (with some
1598 GNAT-specific extensions), and GNU make.
1600 @command{configure} does not test whether the GNAT installation works
1601 and has a sufficiently recent version; if too old a GNAT version is
1602 installed, the build will fail unless @option{--enable-languages} is
1603 used to disable building the Ada front end.
1605 @section Building with profile feedback
1607 It is possible to use profile feedback to optimize the compiler itself. This
1608 should result in a faster compiler binary. Experiments done on x86 using gcc
1609 3.3 showed approximately 7 percent speedup on compiling C programs. To
1610 bootstrap compiler with profile feedback, use @code{make profiledbootstrap}.
1612 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
1613 compiler. This compiler is used to build a @code{stageprofile} compiler
1614 instrumented to collect execution counts of instruction and branch
1615 probabilities. Then runtime libraries are compiled with profile collected.
1616 Finally a @code{stagefeedback} compiler is built using the information collected.
1618 Unlike @samp{make bootstrap} several additional restrictions apply. The
1619 compiler used to build @code{stage1} needs to support a 64-bit integral type.
1620 It is recommended to only use GCC for this. Also parallel make is currently
1621 not supported since collisions in profile collecting may occur.
1628 @uref{./index.html,,Return to the GCC Installation page}
1632 @c ***Testing*****************************************************************
1634 @comment node-name, next, previous, up
1635 @node Testing, Final install, Building, Installing GCC
1639 @chapter Installing GCC: Testing
1642 @cindex Installing GCC: Testing
1645 Before you install GCC, we encourage you to run the testsuites and to
1646 compare your results with results from a similar configuration that have
1647 been submitted to the
1648 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
1649 Some of these archived results are linked from the build status lists
1650 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
1651 reports a successful build runs the testsuites and submits the results.
1652 This step is optional and may require you to download additional software,
1653 but it can give you confidence in your new GCC installation or point out
1654 problems before you install and start using your new GCC.
1656 First, you must have @uref{download.html,,downloaded the testsuites}.
1657 These are part of the full distribution, but if you downloaded the
1658 ``core'' compiler plus any front ends, you must download the testsuites
1661 Second, you must have the testing tools installed. This includes
1662 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu} 1.4.4 and later,
1663 Tcl, and Expect; the DejaGnu site has links to these.
1665 If the directories where @command{runtest} and @command{expect} were
1666 installed are not in the @env{PATH}, you may need to set the following
1667 environment variables appropriately, as in the following example (which
1668 assumes that DejaGnu has been installed under @file{/usr/local}):
1671 TCL_LIBRARY = /usr/local/share/tcl8.0
1672 DEJAGNULIBS = /usr/local/share/dejagnu
1675 (On systems such as Cygwin, these paths are required to be actual
1676 paths, not mounts or links; presumably this is due to some lack of
1677 portability in the DejaGnu code.)
1680 Finally, you can run the testsuite (which may take a long time):
1682 cd @var{objdir}; make -k check
1685 This will test various components of GCC, such as compiler
1686 front ends and runtime libraries. While running the testsuite, DejaGnu
1687 might emit some harmless messages resembling
1688 @samp{WARNING: Couldn't find the global config file.} or
1689 @samp{WARNING: Couldn't find tool init file} that can be ignored.
1691 @section How can you run the testsuite on selected tests?
1693 In order to run sets of tests selectively, there are targets
1694 @samp{make check-gcc} and @samp{make check-g++}
1695 in the @file{gcc} subdirectory of the object directory. You can also
1696 just run @samp{make check} in a subdirectory of the object directory.
1699 A more selective way to just run all @command{gcc} execute tests in the
1703 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
1706 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
1707 the testsuite with filenames matching @samp{9805*}, you would use
1710 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1713 The @file{*.exp} files are located in the testsuite directories of the GCC
1714 source, the most important ones being @file{compile.exp},
1715 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1716 To get a list of the possible @file{*.exp} files, pipe the
1717 output of @samp{make check} into a file and look at the
1718 @samp{Running @dots{} .exp} lines.
1720 @section Passing options and running multiple testsuites
1722 You can pass multiple options to the testsuite using the
1723 @samp{--target_board} option of DejaGNU, either passed as part of
1724 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
1725 work outside the makefiles. For example,
1728 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fno-strength-reduce"
1731 will run the standard @command{g++} testsuites (``unix'' is the target name
1732 for a standard native testsuite situation), passing
1733 @samp{-O3 -fno-strength-reduce} to the compiler on every test, i.e.,
1734 slashes separate options.
1736 You can run the testsuites multiple times using combinations of options
1737 with a syntax similar to the brace expansion of popular shells:
1740 @dots{}"--target_board=arm-sim/@{-mhard-float,-msoft-float@}@{-O1,-O2,-O3,@}"
1743 (Note the empty option caused by the trailing comma in the final group.)
1744 The following will run each testsuite eight times using the @samp{arm-sim}
1745 target, as if you had specified all possible combinations yourself:
1748 --target_board=arm-sim/-mhard-float/-O1
1749 --target_board=arm-sim/-mhard-float/-O2
1750 --target_board=arm-sim/-mhard-float/-O3
1751 --target_board=arm-sim/-mhard-float
1752 --target_board=arm-sim/-msoft-float/-O1
1753 --target_board=arm-sim/-msoft-float/-O2
1754 --target_board=arm-sim/-msoft-float/-O3
1755 --target_board=arm-sim/-msoft-float
1758 They can be combined as many times as you wish, in arbitrary ways. This
1762 @dots{}"--target_board=unix/-Wextra@{-O3,-fno-strength-reduce@}@{-fomit-frame-pointer,@}"
1765 will generate four combinations, all involving @samp{-Wextra}.
1767 The disadvantage to this method is that the testsuites are run in serial,
1768 which is a waste on multiprocessor systems. For users with GNU Make and
1769 a shell which performs brace expansion, you can run the testsuites in
1770 parallel by having the shell perform the combinations and @command{make}
1771 do the parallel runs. Instead of using @samp{--target_board}, use a
1772 special makefile target:
1775 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
1781 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
1784 will run three concurrent ``make-gcc'' testsuites, eventually testing all
1785 ten combinations as described above. Note that this is currently only
1786 supported in the @file{gcc} subdirectory. (To see how this works, try
1787 typing @command{echo} before the example given here.)
1790 @section Additional testing for Java Class Libraries
1792 The Java runtime tests can be executed via @samp{make check}
1793 in the @file{@var{target}/libjava/testsuite} directory in
1796 The @uref{http://sources.redhat.com/mauve/,,Mauve Project} provides
1797 a suite of tests for the Java Class Libraries. This suite can be run
1798 as part of libgcj testing by placing the Mauve tree within the libjava
1799 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
1800 specifying the location of that tree when invoking @samp{make}, as in
1801 @samp{make MAUVEDIR=~/mauve check}.
1803 @uref{http://www-124.ibm.com/developerworks/oss/cvs/jikes/~checkout~/jacks/jacks.html,,Jacks}
1804 is a free testsuite that tests Java compiler front ends. This suite
1805 can be run as part of libgcj testing by placing the Jacks tree within
1806 the libjava testsuite at @file{libjava/testsuite/libjava.jacks/jacks}.
1808 @section How to interpret test results
1810 The result of running the testsuite are various @file{*.sum} and @file{*.log}
1811 files in the testsuite subdirectories. The @file{*.log} files contain a
1812 detailed log of the compiler invocations and the corresponding
1813 results, the @file{*.sum} files summarize the results. These summaries
1814 contain status codes for all tests:
1818 PASS: the test passed as expected
1820 XPASS: the test unexpectedly passed
1822 FAIL: the test unexpectedly failed
1824 XFAIL: the test failed as expected
1826 UNSUPPORTED: the test is not supported on this platform
1828 ERROR: the testsuite detected an error
1830 WARNING: the testsuite detected a possible problem
1833 It is normal for some tests to report unexpected failures. At the
1834 current time the testing harness does not allow fine grained control
1835 over whether or not a test is expected to fail. This problem should
1836 be fixed in future releases.
1839 @section Submitting test results
1841 If you want to report the results to the GCC project, use the
1842 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
1845 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
1846 -m gcc-testresults@@gcc.gnu.org |sh
1849 This script uses the @command{Mail} program to send the results, so
1850 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
1851 prepended to the testsuite summary and should contain any special
1852 remarks you have on your results or your build environment. Please
1853 do not edit the testsuite result block or the subject line, as these
1854 messages may be automatically processed.
1861 @uref{./index.html,,Return to the GCC Installation page}
1865 @c ***Final install***********************************************************
1867 @comment node-name, next, previous, up
1868 @node Final install, , Testing, Installing GCC
1870 @ifset finalinstallhtml
1872 @chapter Installing GCC: Final installation
1875 Now that GCC has been built (and optionally tested), you can install it with
1877 cd @var{objdir}; make install
1880 We strongly recommend to install into a target directory where there is
1881 no previous version of GCC present.
1883 That step completes the installation of GCC; user level binaries can
1884 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
1885 you specified with the @option{--prefix} to configure (or
1886 @file{/usr/local} by default). (If you specified @option{--bindir},
1887 that directory will be used instead; otherwise, if you specified
1888 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
1889 Headers for the C++ and Java libraries are installed in
1890 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
1891 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
1892 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
1893 in info format in @file{@var{infodir}} (normally
1894 @file{@var{prefix}/info}).
1896 When installing cross-compilers, GCC's executables
1897 are not only installed into @file{@var{bindir}}, that
1898 is, @file{@var{exec-prefix}/bin}, but additionally into
1899 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
1900 exists. Typically, such @dfn{tooldirs} hold target-specific
1901 binutils, including assembler and linker.
1903 Installation into a temporary staging area or into a @command{chroot}
1904 jail can be achieved with the command
1907 make DESTDIR=@var{path-to-rootdir} install
1910 @noindent where @var{path-to-rootdir} is the absolute path of
1911 a directory relative to which all installation paths will be
1912 interpreted. Note that the directory specified by @code{DESTDIR}
1913 need not exist yet; it will be created if necessary.
1915 There is a subtle point with tooldirs and @code{DESTDIR}:
1916 If you relocate a cross-compiler installation with
1917 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
1918 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
1919 be filled with duplicated GCC executables only if it already exists,
1920 it will not be created otherwise. This is regarded as a feature,
1921 not as a bug, because it gives slightly more control to the packagers
1922 using the @code{DESTDIR} feature.
1924 If you built a released version of GCC using @samp{make bootstrap} then please
1925 quickly review the build status page for your release, available from
1926 @uref{http://gcc.gnu.org/buildstat.html}.
1927 If your system is not listed for the version of GCC that you built,
1929 @email{gcc@@gcc.gnu.org} indicating
1930 that you successfully built and installed GCC.
1931 Include the following information:
1935 Output from running @file{@var{srcdir}/config.guess}. Do not send
1936 that file itself, just the one-line output from running it.
1939 The output of @samp{gcc -v} for your newly installed @command{gcc}.
1940 This tells us which version of GCC you built and the options you passed to
1944 Whether you enabled all languages or a subset of them. If you used a
1945 full distribution then this information is part of the configure
1946 options in the output of @samp{gcc -v}, but if you downloaded the
1947 ``core'' compiler plus additional front ends then it isn't apparent
1948 which ones you built unless you tell us about it.
1951 If the build was for GNU/Linux, also include:
1954 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
1955 this information should be available from @file{/etc/issue}.
1958 The version of the Linux kernel, available from @samp{uname --version}
1962 The version of glibc you used; for RPM-based systems like Red Hat,
1963 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
1964 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
1966 For other systems, you can include similar information if you think it is
1970 Any other information that you think would be useful to people building
1971 GCC on the same configuration. The new entry in the build status list
1972 will include a link to the archived copy of your message.
1975 We'd also like to know if the
1977 @ref{Specific, host/target specific installation notes}
1980 @uref{specific.html,,host/target specific installation notes}
1982 didn't include your host/target information or if that information is
1983 incomplete or out of date. Send a note to
1984 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
1986 If you find a bug, please report it following the
1987 @uref{../bugs.html,,bug reporting guidelines}.
1989 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
1990 dvi}. You will need to have @command{texi2dvi} (version at least 4.2)
1991 and @TeX{} installed. This creates a number of @file{.dvi} files in
1992 subdirectories of @file{@var{objdir}}; these may be converted for
1993 printing with programs such as @command{dvips}. You can also
1994 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
1995 Free Software Foundation}, though such manuals may not be for the most
1996 recent version of GCC@.
2003 @uref{./index.html,,Return to the GCC Installation page}
2007 @c ***Binaries****************************************************************
2009 @comment node-name, next, previous, up
2010 @node Binaries, Specific, Installing GCC, Top
2014 @chapter Installing GCC: Binaries
2017 @cindex Installing GCC: Binaries
2019 We are often asked about pre-compiled versions of GCC@. While we cannot
2020 provide these for all platforms, below you'll find links to binaries for
2021 various platforms where creating them by yourself is not easy due to various
2024 Please note that we did not create these binaries, nor do we
2025 support them. If you have any problems installing them, please
2026 contact their makers.
2033 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
2036 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX}.
2040 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
2043 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
2044 Development Tools for the Renesas H8/300[HS] Series}.
2050 @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center};
2053 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
2057 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
2058 Development Tools for the Motorola 68HC11/68HC12}.
2061 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
2062 OpenServer/Unixware}.
2065 Sinix/Reliant Unix---@uref{ftp://ftp.fujitsu-siemens.com/pub/pd/gnu/gcc/,,Siemens}.
2068 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
2071 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
2077 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
2079 The @uref{http://www.mingw.org/,,MinGW} project.
2083 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
2084 Written Word} offers binaries for
2087 Digital UNIX 4.0D and 5.1,
2089 HP-UX 10.20, 11.00, and 11.11, and
2090 Solaris/SPARC 2.5.1, 2.6, 2.7, 8, and 9,
2093 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
2094 number of platforms.
2097 In addition to those specific offerings, you can get a binary
2098 distribution CD-ROM from the
2099 @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}.
2100 It contains binaries for a number of platforms, and
2101 includes not only GCC, but other stuff as well. The current CD does
2102 not contain the latest version of GCC, but it should allow
2103 bootstrapping the compiler. An updated version of that disk is in the
2111 @uref{./index.html,,Return to the GCC Installation page}
2115 @c ***Specific****************************************************************
2117 @comment node-name, next, previous, up
2118 @node Specific, Old, Binaries, Top
2122 @chapter Host/target specific installation notes for GCC
2125 @cindex Specific installation notes
2126 @cindex Target specific installation
2127 @cindex Host specific installation
2128 @cindex Target specific installation notes
2130 Please read this document carefully @emph{before} installing the
2131 GNU Compiler Collection on your machine.
2136 @uref{#alpha*-*-*,,alpha*-*-*}
2138 @uref{#alpha*-dec-osf*,,alpha*-dec-osf*}
2140 @uref{#alphaev5-cray-unicosmk*,,alphaev5-cray-unicosmk*}
2142 @uref{#arc-*-elf,,arc-*-elf}
2144 @uref{#arm-*-elf,,arm-*-elf}
2145 @uref{#arm-*-coff,,arm-*-coff}
2146 @uref{#arm-*-aout,,arm-*-aout}
2148 @uref{#xscale-*-*,,xscale-*-*}
2156 @uref{#*-*-freebsd*,,*-*-freebsd*}
2158 @uref{#h8300-hms,,h8300-hms}
2160 @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*}
2162 @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10}
2164 @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11}
2166 @uref{#*-*-linux-gnu,,*-*-linux-gnu}
2168 @uref{#ix86-*-linux*aout,,i?86-*-linux*aout}
2170 @uref{#ix86-*-linux*,,i?86-*-linux*}
2172 @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*}
2174 @uref{#ix86-*-udk,,i?86-*-udk}
2176 @uref{#ix86-*-esix,,i?86-*-esix}
2178 @uref{#ia64-*-linux,,ia64-*-linux}
2180 @uref{#ia64-*-hpux*,,ia64-*-hpux*}
2182 @uref{#*-ibm-aix*,,*-ibm-aix*}
2184 @uref{#ip2k-*-elf,,ip2k-*-elf}
2186 @uref{#iq2000-*-elf,,iq2000-*-elf}
2188 @uref{#m32r-*-elf,,m32r-*-elf}
2190 @uref{#m6811-elf,,m6811-elf}
2192 @uref{#m6812-elf,,m6812-elf}
2194 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
2196 @uref{#mips-*-*,,mips-*-*}
2198 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2200 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2202 @uref{#powerpc*-*-*,,powerpc*-*-*, powerpc-*-sysv4}
2204 @uref{#powerpc-*-darwin*,,powerpc-*-darwin*}
2206 @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4}
2208 @uref{#powerpc-*-linux-gnu*,,powerpc-*-linux-gnu*}
2210 @uref{#powerpc-*-netbsd*,,powerpc-*-netbsd*}
2212 @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix}
2214 @uref{#powerpc-*-eabisim,,powerpc-*-eabisim}
2216 @uref{#powerpc-*-eabi,,powerpc-*-eabi}
2218 @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
2220 @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim}
2222 @uref{#powerpcle-*-eabi,,powerpcle-*-eabi}
2224 @uref{#s390-*-linux*,,s390-*-linux*}
2226 @uref{#s390x-*-linux*,,s390x-*-linux*}
2228 @uref{#s390x-ibm-tpf*,,s390x-ibm-tpf*}
2230 @uref{#*-*-solaris2*,,*-*-solaris2*}
2232 @uref{#sparc-sun-solaris2*,,sparc-sun-solaris2*}
2234 @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7}
2236 @uref{#sparc-*-linux*,,sparc-*-linux*}
2238 @uref{#sparc64-*-solaris2*,,sparc64-*-solaris2*}
2240 @uref{#sparcv9-*-solaris2*,,sparcv9-*-solaris2*}
2242 @uref{#*-*-sysv*,,*-*-sysv*}
2244 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
2246 @uref{#*-*-vxworks*,,*-*-vxworks*}
2248 @uref{#x86_64-*-*,,x86_64-*-*, amd64-*-*}
2250 @uref{#xtensa-*-elf,,xtensa-*-elf}
2252 @uref{#xtensa-*-linux*,,xtensa-*-linux*}
2254 @uref{#windows,,Microsoft Windows}
2258 @uref{#older,,Older systems}
2263 @uref{#elf_targets,,all ELF targets} (SVR4, Solaris 2, etc.)
2269 <!-- -------- host/target specific issues start here ---------------- -->
2272 @heading @anchor{alpha*-*-*}alpha*-*-*
2274 This section contains general configuration information for all
2275 alpha-based platforms using ELF (in particular, ignore this section for
2276 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2277 section, please read all other sections that match your target.
2279 We require binutils 2.11.2 or newer.
2280 Previous binutils releases had a number of problems with DWARF 2
2281 debugging information, not the least of which is incorrect linking of
2287 @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf*
2288 Systems using processors that implement the DEC Alpha architecture and
2289 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2290 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2292 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2293 supported. (These are the versions which identify themselves as DEC
2296 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2297 may be fixed by configuring with @option{--with-gc=simple},
2298 reconfiguring Kernel Virtual Memory and Swap parameters
2299 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2300 or applying the patch in
2301 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2303 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2304 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2305 we need to use the old assembler, invoked via the barely documented
2306 @option{-oldas} option. To bootstrap GCC, you either need to use the
2310 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2313 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2316 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2319 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2320 are supported on Tru64 UNIX, so you must not configure GCC with
2321 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2323 GCC writes a @samp{.verstamp} directive to the assembler output file
2324 unless it is built as a cross-compiler. It gets the version to use from
2325 the system header file @file{/usr/include/stamp.h}. If you install a
2326 new version of DEC Unix, you should rebuild GCC to pick up the new version
2329 Note that since the Alpha is a 64-bit architecture, cross-compilers from
2330 32-bit machines will not generate code as efficient as that generated
2331 when the compiler is running on a 64-bit machine because many
2332 optimizations that depend on being able to represent a word on the
2333 target in an integral value on the host cannot be performed. Building
2334 cross-compilers on the Alpha for 32-bit machines has only been tested in
2335 a few cases and may not work properly.
2337 @samp{make compare} may fail on old versions of DEC Unix unless you add
2338 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2339 assembler input file is stored in the object file, and that makes
2340 comparison fail if it differs between the @code{stage1} and
2341 @code{stage2} compilations. The option @option{-save-temps} forces a
2342 fixed name to be used for the assembler input file, instead of a
2343 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2344 unless the comparisons fail without that option. If you add
2345 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2346 @samp{.s} files after each series of compilations.
2348 GCC now supports both the native (ECOFF) debugging format used by DBX
2349 and GDB and an encapsulated STABS format for use only with GDB@. See the
2350 discussion of the @option{--with-stabs} option of @file{configure} above
2351 for more information on these formats and how to select them.
2353 There is a bug in DEC's assembler that produces incorrect line numbers
2354 for ECOFF format when the @samp{.align} directive is used. To work
2355 around this problem, GCC will not emit such alignment directives
2356 while writing ECOFF format debugging information even if optimization is
2357 being performed. Unfortunately, this has the very undesirable
2358 side-effect that code addresses when @option{-O} is specified are
2359 different depending on whether or not @option{-g} is also specified.
2361 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2362 DBX@. DEC is now aware of this problem with the assembler and hopes to
2363 provide a fix shortly.
2368 @heading @anchor{alphaev5-cray-unicosmk*}alphaev5-cray-unicosmk*
2369 Cray T3E systems running Unicos/Mk.
2371 This port is incomplete and has many known bugs. We hope to improve the
2372 support for this target soon. Currently, only the C front end is supported,
2373 and it is not possible to build parallel applications. Cray modules are not
2374 supported; in particular, Craylibs are assumed to be in
2375 @file{/opt/ctl/craylibs/craylibs}.
2377 You absolutely @strong{must} use GNU make on this platform. Also, you
2378 need to tell GCC where to find the assembler and the linker. The
2379 simplest way to do so is by providing @option{--with-as} and
2380 @option{--with-ld} to @file{configure}, e.g.@:
2383 configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld \
2384 --enable-languages=c
2387 The comparison test during @samp{make bootstrap} fails on Unicos/Mk
2388 because the assembler inserts timestamps into object files. You should
2389 be able to work around this by doing @samp{make all} after getting this
2395 @heading @anchor{arc-*-elf}arc-*-elf
2396 Argonaut ARC processor.
2397 This configuration is intended for embedded systems.
2402 @heading @anchor{arm-*-elf}arm-*-elf
2403 @heading @anchor{xscale-*-*}xscale-*-*
2404 ARM-family processors. Subtargets that use the ELF object format
2405 require GNU binutils 2.13 or newer. Such subtargets include:
2406 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux},
2407 @code{arm-*-rtems} and @code{arm-*-kaos}.
2412 @heading @anchor{arm-*-coff}arm-*-coff
2413 ARM-family processors. Note that there are two different varieties
2414 of PE format subtarget supported: @code{arm-wince-pe} and
2415 @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}.
2420 @heading @anchor{arm-*-aout}arm-*-aout
2421 ARM-family processors. These targets support the AOUT file format:
2422 @code{arm-*-aout}, @code{arm-*-netbsd}.
2427 @heading @anchor{avr}avr
2429 ATMEL AVR-family micro controllers. These are used in embedded
2430 applications. There are no standard Unix configurations.
2432 @xref{AVR Options,, AVR Options, gcc, Using and Porting the GNU Compiler
2436 See ``AVR Options'' in the main manual
2438 for the list of supported MCU types.
2440 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2442 Further installation notes and other useful information about AVR tools
2443 can also be obtained from:
2447 @uref{http://www.openavr.org,,http://www.openavr.org}
2449 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2451 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2454 We @emph{strongly} recommend using binutils 2.13 or newer.
2456 The following error:
2458 Error: register required
2461 indicates that you should upgrade to a newer version of the binutils.
2466 @heading @anchor{c4x}c4x
2468 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
2469 Processors. These are used in embedded applications. There are no
2470 standard Unix configurations.
2472 @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using and
2473 Porting the GNU Compiler Collection (GCC)},
2476 See ``TMS320C3x/C4x Options'' in the main manual
2478 for the list of supported MCU types.
2480 GCC can be configured as a cross compiler for both the C3x and C4x
2481 architectures on the same system. Use @samp{configure --target=c4x
2482 --enable-languages="c,c++"} to configure.
2485 Further installation notes and other useful information about C4x tools
2486 can also be obtained from:
2490 @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/}
2496 @heading @anchor{cris}CRIS
2498 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2499 series. These are used in embedded applications.
2502 @xref{CRIS Options,, CRIS Options, gcc, Using and Porting the GNU Compiler
2506 See ``CRIS Options'' in the main manual
2508 for a list of CRIS-specific options.
2510 There are a few different CRIS targets:
2512 @item cris-axis-aout
2513 Old target. Includes a multilib for the @samp{elinux} a.out-based
2514 target. No multilibs for newer architecture variants.
2516 Mainly for monolithic embedded systems. Includes a multilib for the
2517 @samp{v10} core used in @samp{ETRAX 100 LX}.
2518 @item cris-axis-linux-gnu
2519 A GNU/Linux port for the CRIS architecture, currently targeting
2520 @samp{ETRAX 100 LX} by default.
2523 For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11
2524 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2526 Pre-packaged tools can be obtained from
2527 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2528 information about this platform is available at
2529 @uref{http://developer.axis.com/}.
2534 @heading @anchor{dos}DOS
2536 Please have a look at the @uref{binaries.html,,binaries page}.
2538 You cannot install GCC by itself on MSDOS; it will not compile under
2539 any MSDOS compiler except itself. You need to get the complete
2540 compilation package DJGPP, which includes binaries as well as sources,
2541 and includes all the necessary compilation tools and libraries.
2546 @heading @anchor{*-*-freebsd*}*-*-freebsd*
2548 The version of binutils installed in @file{/usr/bin} probably works with
2549 this release of GCC. However, on FreeBSD 4, bootstrapping against the
2550 latest FSF binutils is known to improve overall testsuite results; and,
2551 on FreeBSD/alpha, using binutils 2.14 or later is required to build libjava.
2553 Support for FreeBSD 1 was discontinued in GCC 3.2.
2555 Support for FreeBSD 2 will be discontinued after GCC 3.4. The
2556 following was true for GCC 3.1 but the current status is unknown.
2557 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2558 configuration support and files as shipped with GCC 2.95 are still in
2559 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2560 it is unknown which version of binutils was used (it is assumed that it
2561 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2563 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2564 default for all CPU architectures. It had been the default on
2565 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2566 of @option{-g}, if you really want the old debugging format. There are
2567 no known issues with mixing object files and libraries with different
2568 debugging formats. Otherwise, this release of GCC should now match more
2569 of the configuration used in the stock FreeBSD configuration of GCC. In
2570 particular, @option{--enable-threads} is now configured by default.
2571 However, as a general user, do not attempt to replace the system
2572 compiler with this release. Known to bootstrap and check with good
2573 results on FreeBSD 4.9-STABLE and 5-CURRENT@. In the past, known to
2574 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
2575 4.3, 4.4, 4.5, 4.8-STABLE@.
2577 In principle, @option{--enable-threads} is now compatible with
2578 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
2579 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
2581 library may be incorrectly built (symbols are missing at link time).
2582 There is a rare timing-based startup hang (probably involves an
2583 assumption about the thread library). Multi-threaded boehm-gc (required for
2584 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
2585 4.5-RELEASE@. Other CPU architectures
2586 supported by FreeBSD will require additional configuration tuning in, at
2587 the very least, both boehm-gc and libffi.
2589 Shared @file{libgcc_s.so} is now built and installed by default.
2594 @heading @anchor{h8300-hms}h8300-hms
2595 Renesas H8/300 series of processors.
2597 Please have a look at the @uref{binaries.html,,binaries page}.
2599 The calling convention and structure layout has changed in release 2.6.
2600 All code must be recompiled. The calling convention now passes the
2601 first three arguments in function calls in registers. Structures are no
2602 longer a multiple of 2 bytes.
2607 @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux*
2608 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
2610 We @emph{highly} recommend using gas/binutils on all hppa platforms;
2611 you may encounter a variety of problems when using the HP assembler.
2613 Specifically, @option{-g} does not work on HP-UX (since that system
2614 uses a peculiar debugging format which GCC does not know about), unless you
2615 use GAS and GDB and configure GCC with the
2616 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
2617 @option{--with-as=@dots{}} options.
2619 If you wish to use the pa-risc 2.0 architecture support with a 32-bit
2620 runtime, you must use either the HP assembler, or gas/binutils 2.11
2623 There are two default scheduling models for instructions. These are
2624 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
2625 architecture specified for the target machine when configuring.
2626 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
2627 the target is a @samp{hppa1*} machine.
2629 The PROCESSOR_8000 model is not well suited to older processors. Thus,
2630 it is important to completely specify the machine architecture when
2631 configuring if you want a model other than PROCESSOR_8000. The macro
2632 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
2633 default scheduling model is desired.
2635 More specific information to @samp{hppa*-hp-hpux*} targets follows.
2640 @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10
2642 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
2643 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
2649 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
2653 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
2657 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
2660 The HP assembler on these systems has some problems. Most notably the
2661 assembler inserts timestamps into each object file it creates, causing
2662 the 3-stage comparison test to fail during a @samp{make bootstrap}.
2663 You should be able to continue by saying @samp{make all} after getting
2664 the failure from @samp{make bootstrap}.
2666 GCC 3.5 requires CVS binutils as of April 28, 2004 or later. Earlier
2667 versions require binutils 2.8 or later.
2669 The C++ ABI has changed incompatibly in GCC 3.5. COMDAT subspaces are
2670 used for one-only code and data. This resolves many of the previous
2671 problems in using C++ on this target. However, the ABI is not compatible
2672 with the one implemented under HP-UX 11 using secondary definitions.
2677 @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11
2679 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
2680 be used to compile GCC 3.0 and up.
2682 Refer to @uref{binaries.html,,binaries} for information about obtaining
2683 precompiled GCC binaries for HP-UX. Precompiled binaries must be obtained
2684 to build the Ada language as it can't be bootstrapped using C. Ada is
2685 only available for the 32-bit PA-RISC runtime. The libffi and libjava
2686 haven't been ported to HP-UX and don't build.
2688 It is possible to build GCC 3.3 starting with the bundled HP compiler,
2689 but the process requires several steps. GCC 3.3 can then be used to
2690 build later versions. The fastjar program contains ISO C code and
2691 can't be built with the HP bundled compiler. This problem can be
2692 avoided by not building the Java language. For example, use the
2693 @option{--enable-languages="c,c++,f77,objc"} option in your configure
2696 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
2697 bundled compiler supports only traditional C; you will need either HP's
2698 unbundled compiler, or a binary distribution of GCC@.
2700 There are several possible approaches to building the distribution.
2701 Binutils can be built first using the HP tools. Then, the GCC
2702 distribution can be built. The second approach is to build GCC
2703 first using the HP tools, then build binutils, then rebuild GCC.
2704 There have been problems with various binary distributions, so it
2705 is best not to start from a binary distribution.
2707 On 64-bit capable systems, there are two distinct targets. Different
2708 installation prefixes must be used if both are to be installed on
2709 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
2710 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
2711 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
2712 PA-RISC 2.0 architecture. The HP and GNU linkers are both supported
2715 The script config.guess now selects the target type based on the compiler
2716 detected during configuration. You must define @env{PATH} or @env{CC} so
2717 that configure finds an appropriate compiler for the initial bootstrap.
2718 When @env{CC} is used, the definition should contain the options that are
2719 needed whenever @env{CC} is used.
2721 Specifically, options that determine the runtime architecture must be
2722 in @env{CC} to correctly select the target for the build. It is also
2723 convenient to place many other compiler options in @env{CC}. For example,
2724 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
2725 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
2726 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
2727 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
2728 macro definition table of cpp needs to be increased for a successful
2729 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
2730 be defined when building with the bundled compiler, or when using the
2731 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
2733 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
2734 with the @option{--with-ld=@dots{}} option. This overrides the standard
2735 search for ld. The two linkers supported on this target require different
2736 commands. The default linker is determined during configuration. As a
2737 result, it's not possible to switch linkers in the middle of a GCC build.
2738 This has been been reported to sometimes occur in unified builds of
2741 GCC 3.0 through 3.2 require binutils 2.11 or above. GCC 3.3 through
2742 GCC 3.5 require binutils 2.14 or later.
2744 Although the HP assembler can be used for an initial build, it shouldn't
2745 be used with any languages other than C and perhaps Fortran due to its
2746 many limitations. For example, it does not support weak symbols or alias
2747 definitions. As a result, explicit template instantiations are required
2748 when using C++. This makes it difficult if not impossible to build many
2749 C++ applications. You can't generate debugging information when using
2750 the HP assembler. Finally, @samp{make bootstrap} fails in the final
2751 comparison of object modules due to the time stamps that it inserts into
2752 the modules. The bootstrap can be continued from this point with
2755 A recent linker patch must be installed for the correct operation of
2756 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
2757 oldest linker patches that are known to work. They are for HP-UX
2758 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
2759 @code{PHSS_24304}, might be usable but it hasn't been tested. These
2760 patches have been superseded. Consult the HP patch database to obtain
2761 the currently recommended linker patch for your system.
2763 The patches are necessary for the support of weak symbols on the
2764 32-bit port, and for the running of initializers and finalizers. Weak
2765 symbols are implemented using SOM secondary definition symbols. Prior
2766 to HP-UX 11, there are bugs in the linker support for secondary symbols.
2767 The patches correct a problem of linker core dumps creating shared
2768 libraries containing secondary symbols, as well as various other
2769 linking issues involving secondary symbols.
2771 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
2772 run initializers and finalizers on the 64-bit port. The 32-bit port
2773 uses the linker @option{+init} and @option{+fini} options for the same
2774 purpose. The patches correct various problems with the +init/+fini
2775 options, including program core dumps. Binutils 2.14 corrects a
2776 problem on the 64-bit port resulting from HP's non-standard use of
2777 the .init and .fini sections for array initializers and finalizers.
2779 There are a number of issues to consider in selecting which linker to
2780 use with the 64-bit port. The GNU 64-bit linker can only create dynamic
2781 binaries. The @option{-static} option causes linking with archive
2782 libraries but doesn't produce a truly static binary. Dynamic binaries
2783 still require final binding by the dynamic loader to resolve a set of
2784 dynamic-loader-defined symbols. The default behavior of the HP linker
2785 is the same as the GNU linker. However, it can generate true 64-bit
2786 static binaries using the @option{+compat} option.
2788 The HP 64-bit linker doesn't support linkonce semantics. As a
2789 result, C++ programs have many more sections than they should.
2791 The GNU 64-bit linker has some issues with shared library support
2792 and exceptions. As a result, we only support libgcc in archive
2793 format. For similar reasons, dwarf2 unwind and exception support
2794 are disabled. The GNU linker also has problems creating binaries
2795 with @option{-static}. It doesn't provide stubs for internal
2796 calls to global functions in shared libraries, so these calls
2797 can't be overloaded.
2799 Thread support is not implemented in GCC 3.0 through 3.2, so the
2800 @option{--enable-threads} configure option does not work. In 3.3
2801 and later, POSIX threads are supported. The optional DCE thread
2802 library is not supported.
2804 This port still is undergoing significant development.
2809 @heading @anchor{*-*-linux-gnu}*-*-linux-gnu
2811 Versions of libstdc++-v3 starting with 3.2.1 require bugfixes present
2812 in glibc 2.2.5 and later. More information is available in the
2813 libstdc++-v3 documentation.
2818 @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout
2819 Use this configuration to generate @file{a.out} binaries on Linux-based
2820 GNU systems. This configuration is being superseded.
2825 @heading @anchor{ix86-*-linux*}i?86-*-linux*
2827 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
2828 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
2830 If you receive Signal 11 errors when building on GNU/Linux, then it is
2831 possible you have a hardware problem. Further information on this can be
2832 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
2837 @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5*
2838 Use this for the SCO OpenServer Release 5 family of operating systems.
2840 Unlike earlier versions of GCC, the ability to generate COFF with this
2841 target is no longer provided.
2843 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
2844 the system debugger to be used. That support was too burdensome to
2845 maintain. GCC now emits only DWARF 2 for this target. This means you
2846 may use either the UDK debugger or GDB to debug programs built by this
2849 GCC is now only supported on releases 5.0.4 and later, and requires that
2850 you install Support Level Supplement OSS646B or later, and Support Level
2851 Supplement OSS631C or later. If you are using release 5.0.7 of
2852 OpenServer, you must have at least the first maintenance pack installed
2853 (this includes the relevant portions of OSS646). OSS646, also known as
2854 the "Execution Environment Update", provides updated link editors and
2855 assemblers, as well as updated standard C and math libraries. The C
2856 startup modules are also updated to support the System V gABI draft, and
2857 GCC relies on that behavior. OSS631 provides a collection of commonly
2858 used open source libraries, some of which GCC depends on (such as GNU
2859 gettext and zlib). SCO OpenServer Release 5.0.7 has all of this built
2860 in by default, but OSS631C and later also apply to that release. Please
2862 @uref{ftp://ftp.sco.com/pub/openserver5,,ftp://ftp.sco.com/pub/openserver5}
2863 for the latest versions of these (and other potentially useful)
2866 Although there is support for using the native assembler, it is
2867 recommended that you configure GCC to use the GNU assembler. You do
2868 this by using the flags
2869 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You should
2870 use a modern version of GNU binutils. Version 2.13.2.1 was used for all
2871 testing. In general, only the @option{--with-gnu-as} option is tested.
2872 A modern bintuils (as well as a plethora of other development related
2873 GNU utilities) can be found in Support Level Supplement OSS658A, the
2874 "GNU Development Tools" package. See the SCO web and ftp sites for details.
2875 That package also contains the currently "officially supported" version of
2876 GCC, version 2.95.3. It is useful for bootstrapping this version.
2881 @heading @anchor{ix86-*-udk}i?86-*-udk
2883 This target emulates the SCO Universal Development Kit and requires that
2884 package be installed. (If it is installed, you will have a
2885 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
2886 @samp{i?86-*-unixware7*} target
2887 but is meant to be used when hosting on a system where UDK isn't the
2888 default compiler such as OpenServer 5 or Unixware 2. This target will
2889 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
2890 with the same warnings and caveats as the SCO UDK@.
2892 This target is a little tricky to build because we have to distinguish
2893 it from the native tools (so it gets headers, startups, and libraries
2894 from the right place) while making the tools not think we're actually
2895 building a cross compiler. The easiest way to do this is with a configure
2899 CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure \
2900 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-
2903 @emph{You should substitute @samp{i686} in the above command with the appropriate
2904 processor for your host.}
2906 After the usual @samp{make bootstrap} and
2907 @samp{make install}, you can then access the UDK-targeted GCC
2908 tools by adding @command{udk-} before the commonly known name. For
2909 example, to invoke the C compiler, you would use @command{udk-gcc}.
2910 They will coexist peacefully with any native-target GCC tools you may
2917 @heading @anchor{ia64-*-linux}ia64-*-linux
2918 IA-64 processor (also known as IPF, or Itanium Processor Family)
2921 If you are using the optional libunwind library, then you must use
2922 libunwind 0.96 or later.
2924 None of the following versions of GCC has an ABI that is compatible
2925 with any of the other versions in this list, with the exception that
2926 Red Hat 2.96 and Trillian 000171 are compatible with each other:
2927 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
2928 This primarily affects C++ programs and programs that create shared libraries.
2929 GCC 3.1 or later is recommended for compiling linux, the kernel.
2930 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
2931 more major ABI changes are expected.
2936 @heading @anchor{ia64-*-hpux*}ia64-*-hpux*
2937 Building GCC on this target requires the GNU Assembler. The bundled HP
2938 assembler will not work. To prevent GCC from using the wrong assembler,
2939 the option @option{--with-gnu-as} may be necessary.
2941 The GCC libunwind library has not been ported to HPUX. This means that for
2942 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
2943 is required to build GCC. For GCC 3.3 and later, this is the default.
2947 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
2949 @heading @anchor{*-ibm-aix*}*-ibm-aix*
2950 Support for AIX version 3 and older was discontinued in GCC 3.4.
2952 AIX Make frequently has problems with GCC makefiles. GNU Make 3.79.1 or
2953 newer is recommended to build on this platform.
2955 To speed up the configuration phases of bootstrapping and installing GCC,
2956 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
2959 % CONFIG_SHELL=/opt/freeware/bin/bash
2960 % export CONFIG_SHELL
2963 and then proceed as described in @uref{build.html,,the build instructions},
2964 where we strongly recommend using GNU make and specifying an absolute path
2965 to invoke @var{srcdir}/configure.
2967 Errors involving @code{alloca} when building GCC generally are due
2968 to an incorrect definition of @code{CC} in the Makefile or mixing files
2969 compiled with the native C compiler and GCC@. During the stage1 phase of
2970 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
2971 (not @command{xlc}). Once @command{configure} has been informed of
2972 @command{xlc}, one needs to use @samp{make distclean} to remove the
2973 configure cache files and ensure that @env{CC} environment variable
2974 does not provide a definition that will confuse @command{configure}.
2975 If this error occurs during stage2 or later, then the problem most likely
2976 is the version of Make (see above).
2978 The native @command{as} and @command{ld} are recommended for bootstrapping
2979 on AIX 4 and required for bootstrapping on AIX 5L. The GNU Assembler
2980 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
2981 utilize weak symbol functionality although it is not supported. The GNU
2982 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC.
2983 The native AIX tools do interoperate with GCC@.
2985 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
2986 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
2987 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
2988 referenced as APAR IY53606 (AIX 5.2) or a APAR IY54774 (AIX 5.1)
2990 @samp{libstdc++} in GCC 3.4 increments the major version number of the
2991 shared object and GCC installation places the @file{libstdc++.a}
2992 shared library in a common location which will overwrite the and GCC
2993 3.3 version of the shared library. Applications either need to be
2994 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
2995 versions of the @samp{libstdc++} shared object needs to be available
2996 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
2997 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
2998 installed for runtime dynamic loading using the following steps to set
2999 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
3000 multilib @file{libstdc++.a} installed:
3002 Extract the shared objects from the currently installed
3003 @file{libstdc++.a} archive:
3005 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
3008 Enable the @samp{F_LOADONLY} flag so that the shared object will be
3009 available for runtime dynamic loading, but not linking:
3011 % strip -e libstdc++.so.4 libstdc++.so.5
3014 Archive the runtime-only shared object in the GCC 3.4
3015 @file{libstdc++.a} archive:
3017 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
3020 Linking executables and shared libraries may produce warnings of
3021 duplicate symbols. The assembly files generated by GCC for AIX always
3022 have included multiple symbol definitions for certain global variable
3023 and function declarations in the original program. The warnings should
3024 not prevent the linker from producing a correct library or runnable
3027 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
3028 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
3029 to parse archive libraries did not handle the new format correctly.
3030 These routines are used by GCC and result in error messages during
3031 linking such as ``not a COFF file''. The version of the routines shipped
3032 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
3033 option of the archive command may be used to create archives of 32-bit
3034 objects using the original ``small format''. A correct version of the
3035 routines is shipped with AIX 4.3.2 and above.
3037 Some versions of the AIX binder (linker) can fail with a relocation
3038 overflow severe error when the @option{-bbigtoc} option is used to link
3039 GCC-produced object files into an executable that overflows the TOC@. A fix
3040 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
3041 available from IBM Customer Support and from its
3042 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3043 website as PTF U455193.
3045 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
3046 with a segmentation fault when invoked by any version of GCC@. A fix for
3047 APAR IX87327 is available from IBM Customer Support and from its
3048 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3049 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
3051 The initial assembler shipped with AIX 4.3.0 generates incorrect object
3052 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
3053 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
3054 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
3055 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
3057 AIX provides National Language Support (NLS)@. Compilers and assemblers
3058 use NLS to support locale-specific representations of various data
3059 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
3060 separating decimal fractions). There have been problems reported where
3061 GCC does not produce the same floating-point formats that the assembler
3062 expects. If one encounters this problem, set the @env{LANG}
3063 environment variable to @samp{C} or @samp{En_US}.
3065 By default, GCC for AIX 4.1 and above produces code that can be used on
3066 both Power or PowerPC processors.
3068 A default can be specified with the @option{-mcpu=@var{cpu_type}}
3069 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
3074 @heading @anchor{ip2k-*-elf}ip2k-*-elf
3075 Ubicom IP2022 micro controller.
3076 This configuration is intended for embedded systems.
3077 There are no standard Unix configurations.
3079 Use @samp{configure --target=ip2k-elf --enable-languages=c} to configure GCC@.
3084 @heading @anchor{iq2000-*-elf}iq2000-*-elf
3085 Vitesse IQ2000 processors. These are used in embedded
3086 applications. There are no standard Unix configurations.
3091 @heading @anchor{m32r-*-elf}m32r-*-elf
3092 Renesas M32R processor.
3093 This configuration is intended for embedded systems.
3098 @heading @anchor{m6811-elf}m6811-elf
3099 Motorola 68HC11 family micro controllers. These are used in embedded
3100 applications. There are no standard Unix configurations.
3105 @heading @anchor{m6812-elf}m6812-elf
3106 Motorola 68HC12 family micro controllers. These are used in embedded
3107 applications. There are no standard Unix configurations.
3112 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
3113 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
3114 the assembler that prevents compilation of GCC@. This
3115 bug manifests itself during the first stage of compilation, while
3116 building @file{libgcc2.a}:
3120 cc1: warning: `-g' option not supported on this version of GCC
3121 cc1: warning: `-g1' option not supported on this version of GCC
3122 ./xgcc: Internal compiler error: program as got fatal signal 11
3125 A patched version of the assembler is available as the file
3126 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
3127 have HP software support, the patch can also be obtained directly from
3128 HP, as described in the following note:
3131 This is the patched assembler, to patch SR#1653-010439, where the
3132 assembler aborts on floating point constants.
3134 The bug is not really in the assembler, but in the shared library
3135 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
3136 SR#4701-078451. Anyway, the attached assembler uses the archive
3137 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
3140 This patch is also known as PHCO_4484.
3142 In addition gdb does not understand that native HP-UX format, so
3143 you must use gas if you wish to use gdb.
3145 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
3146 @command{fixproto} shell script triggers a bug in the system shell. If you
3147 encounter this problem, upgrade your operating system or use BASH (the
3148 GNU shell) to run @command{fixproto}. This bug will cause the fixproto
3149 program to report an error of the form:
3152 ./fixproto: sh internal 1K buffer overflow
3155 To fix this, you can also change the first line of the fixproto script
3165 @heading @anchor{mips-*-*}mips-*-*
3166 If on a MIPS system you get an error message saying ``does not have gp
3167 sections for all it's [sic] sectons [sic]'', don't worry about it. This
3168 happens whenever you use GAS with the MIPS linker, but there is not
3169 really anything wrong, and it is okay to use the output file. You can
3170 stop such warnings by installing the GNU linker.
3172 It would be nice to extend GAS to produce the gp tables, but they are
3173 optional, and there should not be a warning about their absence.
3175 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
3176 and later. A patch went in just after the GCC 3.3 release to
3177 make @samp{mips*-*-*} use the generic implementation instead. You can also
3178 configure for @samp{mipsel-elf} as a workaround. The
3179 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
3180 work on this is expected in future releases.
3182 Cross-compilers for the Mips as target using the Mips assembler
3183 currently do not work, because the auxiliary programs
3184 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
3185 anything but a Mips. It does work to cross compile for a Mips
3186 if you use the GNU assembler and linker.
3191 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
3193 In order to compile GCC on an SGI running IRIX 5, the @samp{compiler_dev.hdr}
3194 subsystem must be installed from the IDO CD-ROM supplied by SGI@.
3195 It is also available for download from
3196 @uref{ftp://ftp.sgi.com/sgi/IRIX5.3/iris-development-option-5.3.tardist}.
3198 If you use the MIPS C compiler to bootstrap, it may be necessary
3199 to increase its table size for switch statements with the
3200 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
3201 optimization option, you also need to use @option{-Olimit 3000}.
3203 To enable debugging under IRIX 5, you must use GNU binutils 2.15 or
3204 later, and use the @option{--with-gnu-as} and @option{--with-gnu-ld}
3205 @command{configure} options when configuring GCC@. You need to use GNU
3206 @command{ar} and @command{nm}, also distributed with GNU binutils.
3211 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3213 If you are using SGI's MIPSpro @command{cc} as your bootstrap compiler, you must
3214 ensure that the N32 ABI is in use. To test this, compile a simple C
3215 file with @command{cc} and then run @command{file} on the
3216 resulting object file. The output should look like:
3219 test.o: ELF N32 MSB @dots{}
3225 test.o: ELF 32-bit MSB @dots{}
3231 test.o: ELF 64-bit MSB @dots{}
3234 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3235 should set the environment variable @env{CC} to @samp{cc -n32}
3236 before configuring GCC@. SGI's MIPSpro 7.2 assembler may misassemble
3237 parts of the compiler, causing bootstrap failures. MIPSpro 7.3 is
3238 known to work. MIPSpro C 7.4 may cause bootstrap failures, too, due
3239 to a bug when inlining @code{memcmp}. Either add @code{-U__INLINE_INTRINSICS}
3240 to the @env{CC} environment variable as a workaround or upgrade to
3243 If you want the resulting @command{gcc} to run on old 32-bit systems
3244 with the MIPS R4400 CPU, you need to ensure that only code for the @samp{mips3}
3245 instruction set architecture (ISA) is generated. While GCC 3.x does
3246 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3247 the ISA depending on the machine where GCC is built. Using one of them
3248 as the bootstrap compiler may result in @samp{mips4} code, which won't run at
3249 all on @samp{mips3}-only systems. For the test program above, you should see:
3252 test.o: ELF N32 MSB mips-3 @dots{}
3258 test.o: ELF N32 MSB mips-4 @dots{}
3261 instead, you should set the environment variable @env{CC} to @samp{cc
3262 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3264 GCC on IRIX 6 is usually built to support the N32, O32 and N64 ABIs. If
3265 you build GCC on a system that doesn't have the N64 libraries installed
3266 or cannot run 64-bit binaries,
3267 you need to configure with @option{--disable-multilib} so GCC doesn't
3268 try to use them. This will disable building the O32 libraries, too.
3269 Look for @file{/usr/lib64/libc.so.1} to see if you
3270 have the 64-bit libraries installed.
3272 To enable debugging for the O32 ABI, you must use GNU @command{as} from
3273 GNU binutils 2.15 or later. You may also use GNU @command{ld}, but
3274 this is not required and currently causes some problems with Ada.
3276 The @option{--enable-threads} option doesn't currently work, a patch is
3277 in preparation for a future release. The @option{--enable-libgcj}
3278 option is disabled by default: IRIX 6 uses a very low default limit
3279 (20480) for the command line length. Although @command{libtool} contains a
3280 workaround for this problem, at least the N64 @samp{libgcj} is known not
3281 to build despite this, running into an internal error of the native
3282 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3283 its maximum of 262144 bytes. If you have root access, you can use the
3284 @command{systune} command to do this.
3286 See @uref{http://freeware.sgi.com/} for more
3287 information about using GCC on IRIX platforms.
3292 @heading @anchor{powerpc*-*-*}powerpc-*-*
3294 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3295 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3300 @heading @anchor{powerpc-*-darwin*}powerpc-*-darwin*
3301 PowerPC running Darwin (Mac OS X kernel).
3303 Pre-installed versions of Mac OS X may not include any developer tools,
3304 meaning that you will not be able to build GCC from source. Tool
3305 binaries are available at
3306 @uref{http://developer.apple.com/tools/compilers.html} (free
3307 registration required).
3309 This version of GCC requires at least cctools-528.
3311 The version of GCC shipped by Apple typically includes a number of
3312 extensions not available in a standard GCC release. These extensions
3313 are generally for backwards compatibility and best avoided.
3318 @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4
3319 PowerPC system in big endian mode, running System V.4.
3324 @heading @anchor{powerpc-*-linux-gnu*}powerpc-*-linux-gnu*
3327 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.13.90.0.10}
3328 or newer for a working GCC@.
3333 @heading @anchor{powerpc-*-netbsd*}powerpc-*-netbsd*
3334 PowerPC system in big endian mode running NetBSD@. To build the
3335 documentation you will need Texinfo version 4.2 (NetBSD 1.5.1 included
3336 Texinfo version 3.12).
3341 @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim
3342 Embedded PowerPC system in big endian mode for use in running under the
3348 @heading @anchor{powerpc-*-eabi}powerpc-*-eabi
3349 Embedded PowerPC system in big endian mode.
3354 @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4
3355 PowerPC system in little endian mode, running System V.4.
3360 @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim
3361 Embedded PowerPC system in little endian mode for use in running under
3367 @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi
3368 Embedded PowerPC system in little endian mode.
3373 @heading @anchor{s390-*-linux*}s390-*-linux*
3374 S/390 system running GNU/Linux for S/390@.
3379 @heading @anchor{s390x-*-linux*}s390x-*-linux*
3380 zSeries system (64-bit) running GNU/Linux for zSeries@.
3385 @heading @anchor{s390x-ibm-tpf*}s390x-ibm-tpf*
3386 zSeries system (64-bit) running TPF. This platform is
3387 supported as cross-compilation target only.
3392 @c Please use Solaris 2 to refer to all release of Solaris, starting
3393 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
3394 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3395 @c alone is too unspecific and must be avoided.
3396 @heading @anchor{*-*-solaris2*}*-*-solaris2*
3398 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3399 GCC you first have to install a pre-built compiler, see the
3400 @uref{binaries.html,,binaries page} for details.
3402 The Solaris 2 @command{/bin/sh} will often fail to configure
3403 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3404 recommend to use the following sequence of commands to bootstrap and
3408 % CONFIG_SHELL=/bin/ksh
3409 % export CONFIG_SHELL
3412 and then proceed as described in @uref{build.html,,the build instructions}.
3413 In addition we strongly recommend specifying an absolute path to invoke
3414 @var{srcdir}/configure.
3416 Solaris 2 comes with a number of optional OS packages. Some of these
3417 are needed to use GCC fully, namely @code{SUNWarc},
3418 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3419 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3420 optional packages when installing Solaris 2, you will need to verify that
3421 the packages that GCC needs are installed.
3423 To check whether an optional package is installed, use
3424 the @command{pkginfo} command. To add an optional package, use the
3425 @command{pkgadd} command. For further details, see the Solaris 2
3428 Trying to use the linker and other tools in
3429 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3430 For example, the linker may hang indefinitely. The fix is to remove
3431 @file{/usr/ucb} from your @env{PATH}.
3433 The build process works more smoothly with the legacy Sun tools so, if you
3434 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3435 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3437 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3438 platform. We recommend the use of GNU binutils 2.11.2 or the vendor
3439 tools (Sun @command{as}, Sun @command{ld}).
3441 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3442 newer: @command{g++} will complain that types are missing. These headers assume
3443 that omitting the type means @code{int}; this assumption worked for C89 but
3444 is wrong for C++, and is now wrong for C99 also.
3446 @command{g++} accepts such (invalid) constructs with the option
3447 @option{-fpermissive}; it
3448 will assume that any missing type is @code{int} (as defined by C89).
3450 There are patches for Solaris 2.6 (105633-56 or newer for SPARC,
3451 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC,
3452 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3453 108653-22 for Intel) that fix this bug.
3458 @heading @anchor{sparc-sun-solaris2*}sparc-sun-solaris2*
3460 When GCC is configured to use binutils 2.11.2 or later the binaries
3461 produced are smaller than the ones produced using Sun's native tools;
3462 this difference is quite significant for binaries containing debugging
3465 Sun @command{as} 4.x is broken in that it cannot cope with long symbol names.
3466 A typical error message might look similar to the following:
3469 /usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041: error:
3470 can't compute value of an expression involving an external symbol.
3473 This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris
3474 2.6 and has been fixed in later (5.x) versions of the assembler,
3475 starting with Solaris 7.
3477 Starting with Solaris 7, the operating system is capable of executing
3478 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3479 this; the @option{-m64} option enables 64-bit code generation.
3480 However, if all you want is code tuned for the UltraSPARC CPU, you
3481 should try the @option{-mtune=ultrasparc} option instead, which produces
3482 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3485 When configuring on a Solaris 7 or later system that is running a kernel
3486 that supports only 32-bit binaries, one must configure with
3487 @option{--disable-multilib}, since we will not be able to build the
3488 64-bit target libraries.
3490 GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
3491 the GNU compiler (especially GCC 3.0.x versions), which lead to the
3492 miscompilation of the stage1 compiler and the subsequent failure of the
3493 bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
3494 stage, i.e. to bootstrap that compiler with the base compiler and then
3495 use it to bootstrap the final compiler.
3497 GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
3498 and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
3499 failure in form of a miscompilation of the stage1 compiler by the Sun
3500 compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
3502 GCC 3.4 changed the default debugging format from STABS to DWARF-2 for
3503 32-bit code on Solaris 7 and later. If you are using the Sun
3504 assembler, this change apparently runs afoul of Sun bug 4910101, for
3505 which (as of 2004-05-23) there is no fix. A symptom of the problem is
3506 that you cannot compile C++ programs like @command{groff} 1.19.1
3507 without getting messages like @samp{ld: warning: relocation error:
3508 R_SPARC_UA32 @dots{} external symbolic relocation against
3509 non-allocatable section .debug_info; cannot be processed at runtime:
3510 relocation ignored}. To work around this problem, compile with
3511 @option{-gstabs+} instead of plain @option{-g}.
3516 @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7
3518 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3519 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3520 and later, including all EGCS releases. Sun formerly recommended
3521 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3522 recommend it only for people who use Sun's compilers.
3524 Here are some workarounds to this problem:
3527 Do not install Sun patch 107058-01 until after Sun releases a
3528 complete patch for bug 4210064. This is the simplest course to take,
3529 unless you must also use Sun's C compiler. Unfortunately 107058-01
3530 is preinstalled on some new Solaris 7-based hosts, so you may have to
3534 Copy the original, unpatched Solaris 7
3535 @command{/usr/ccs/bin/as} into
3536 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
3537 adjusting the latter name to fit your local conventions and software
3541 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3542 both 107058-01 and 106950-03 installed has reported the bug with GCC
3543 and Sun's dynamic linker. This last course of action is riskiest,
3544 for two reasons. First, you must install 106950 on all hosts that
3545 run code generated by GCC; it doesn't suffice to install it only on
3546 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3547 only a partial fix for bug 4210064, but Sun doesn't know whether the
3548 partial fix is adequate for GCC@. Revision -08 or later should fix
3549 the bug. The current (as of 2004-05-23) revision is -24, and is included in
3550 the Solaris 7 Recommended Patch Cluster.
3553 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
3554 which causes a bootstrap failure when linking the 64-bit shared version of
3555 libgcc. A typical error message is:
3558 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
3559 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
3562 This bug has been fixed in the final 5.0 version of the assembler.
3567 @heading @anchor{sparc-*-linux*}sparc-*-linux*
3569 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
3570 or newer on this platform. All earlier binutils and glibc
3571 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
3577 @heading @anchor{sparc64-*-solaris2*}sparc64-*-solaris2*
3579 The following compiler flags must be specified in the configure
3580 step in order to bootstrap this target with the Sun compiler:
3583 % CC="cc -xildoff -xarch=v9" @var{srcdir}/configure [@var{options}] [@var{target}]
3586 @option{-xildoff} turns off the incremental linker, and @option{-xarch=v9}
3587 specifies the SPARC-V9 architecture to the Sun linker and assembler.
3592 @heading @anchor{sparcv9-*-solaris2*}sparcv9-*-solaris2*
3594 This is a synonym for sparc64-*-solaris2*.
3599 @heading @anchor{#*-*-sysv*}*-*-sysv*
3600 On System V release 3, you may get this error message
3604 ld fatal: failed to write symbol name @var{something}
3605 in strings table for file @var{whatever}
3608 This probably indicates that the disk is full or your ulimit won't allow
3609 the file to be as large as it needs to be.
3611 This problem can also result because the kernel parameter @code{MAXUMEM}
3612 is too small. If so, you must regenerate the kernel and make the value
3613 much larger. The default value is reported to be 1024; a value of 32768
3614 is said to work. Smaller values may also work.
3616 On System V, if you get an error like this,
3619 /usr/local/lib/bison.simple: In function `yyparse':
3620 /usr/local/lib/bison.simple:625: virtual memory exhausted
3624 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
3626 On a System V release 4 system, make sure @file{/usr/bin} precedes
3627 @file{/usr/ucb} in @code{PATH}. The @command{cc} command in
3628 @file{/usr/ucb} uses libraries which have bugs.
3633 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3634 Don't try compiling with VAX C (@command{vcc}). It produces incorrect code
3635 in some cases (for example, when @code{alloca} is used).
3640 @heading @anchor{*-*-vxworks*}*-*-vxworks*
3641 Support for VxWorks is in flux. At present GCC supports @emph{only} the
3642 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC.
3643 We welcome patches for other architectures supported by VxWorks 5.5.
3644 Support for VxWorks AE would also be welcome; we believe this is merely
3645 a matter of writing an appropriate ``configlette'' (see below). We are
3646 not interested in supporting older, a.out or COFF-based, versions of
3649 VxWorks comes with an older version of GCC installed in
3650 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
3651 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
3652 Before running @command{configure}, create the directories @file{@var{prefix}}
3653 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
3654 linker, etc. into @file{@var{prefix}/bin}, and set your @var{PATH} to
3655 include that directory while running both @command{configure} and
3658 You must give @command{configure} the
3659 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
3660 find the VxWorks system headers. Since VxWorks is a cross compilation
3661 target only, you must also specify @option{--target=@var{target}}.
3662 @command{configure} will attempt to create the directory
3663 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
3664 make sure the user running @command{configure} has sufficient privilege
3667 GCC's exception handling runtime requires a special ``configlette''
3668 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
3669 that file to add the module to your kernel build. (Future versions of
3670 VxWorks will incorporate this module.)
3675 @heading @anchor{x86_64-*-*}x86_64-*-*, amd64-*-*
3677 GCC supports the x86-64 architecture implemented by the AMD64 processor
3678 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD.
3679 On GNU/Linux the default is a bi-arch compiler which is able to generate
3680 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
3685 @heading @anchor{xtensa-*-elf}xtensa-*-elf
3687 This target is intended for embedded Xtensa systems using the
3688 @samp{newlib} C library. It uses ELF but does not support shared
3689 objects. Designed-defined instructions specified via the
3690 Tensilica Instruction Extension (TIE) language are only supported
3691 through inline assembly.
3693 The Xtensa configuration information must be specified prior to
3694 building GCC@. The @file{include/xtensa-config.h} header
3695 file contains the configuration information. If you created your
3696 own Xtensa configuration with the Xtensa Processor Generator, the
3697 downloaded files include a customized copy of this header file,
3698 which you can use to replace the default header file.
3703 @heading @anchor{xtensa-*-linux*}xtensa-*-linux*
3705 This target is for Xtensa systems running GNU/Linux. It supports ELF
3706 shared objects and the GNU C library (glibc). It also generates
3707 position-independent code (PIC) regardless of whether the
3708 @option{-fpic} or @option{-fPIC} options are used. In other
3709 respects, this target is the same as the
3710 @uref{#xtensa-*-elf,,@samp{xtensa-*-elf}} target.
3715 @heading @anchor{windows}Microsoft Windows (32-bit)
3717 A port of GCC 2.95.2 and 3.x is included with the
3718 @uref{http://www.cygwin.com/,,Cygwin environment}.
3720 Current (as of early 2001) snapshots of GCC will build under Cygwin
3721 without modification.
3723 GCC does not currently build with Microsoft's C++ compiler and there
3724 are no plans to make it do so.
3729 @heading @anchor{os2}OS/2
3731 GCC does not currently support OS/2. However, Andrew Zabolotny has been
3732 working on a generic OS/2 port with pgcc. The current code can be found
3733 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
3735 An older copy of GCC 2.8.1 is included with the EMX tools available at
3736 @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,,
3737 ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}.
3742 @heading @anchor{older}Older systems
3744 GCC contains support files for many older (1980s and early
3745 1990s) Unix variants. For the most part, support for these systems
3746 has not been deliberately removed, but it has not been maintained for
3747 several years and may suffer from bitrot.
3749 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
3750 Support for these systems is still present in that release, but
3751 @command{configure} will fail unless the @option{--enable-obsolete}
3752 option is given. Unless a maintainer steps forward, support for these
3753 systems will be removed from the next release of GCC@.
3755 Support for old systems as hosts for GCC can cause problems if the
3756 workarounds for compiler, library and operating system bugs affect the
3757 cleanliness or maintainability of the rest of GCC@. In some cases, to
3758 bring GCC up on such a system, if still possible with current GCC, may
3759 require first installing an old version of GCC which did work on that
3760 system, and using it to compile a more recent GCC, to avoid bugs in the
3761 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
3762 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
3763 sites}. Header bugs may generally be avoided using
3764 @command{fixincludes}, but bugs or deficiencies in libraries and the
3765 operating system may still cause problems.
3767 Support for older systems as targets for cross-compilation is less
3768 problematic than support for them as hosts for GCC; if an enthusiast
3769 wishes to make such a target work again (including resurrecting any of
3770 the targets that never worked with GCC 2, starting from the last CVS
3771 version before they were removed), patches
3772 @uref{../contribute.html,,following the usual requirements} would be
3773 likely to be accepted, since they should not affect the support for more
3776 For some systems, old versions of GNU binutils may also be useful,
3777 and are available from @file{pub/binutils/old-releases} on
3778 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
3780 Some of the information on specific systems above relates to
3781 such older systems, but much of the information
3782 about GCC on such systems (which may no longer be applicable to
3783 current GCC) is to be found in the GCC texinfo manual.
3788 @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris 2, etc.)
3790 C++ support is significantly better on ELF targets if you use the
3791 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
3792 inlines, vtables and template instantiations will be discarded
3801 @uref{./index.html,,Return to the GCC Installation page}
3805 @c ***Old documentation******************************************************
3807 @include install-old.texi
3813 @uref{./index.html,,Return to the GCC Installation page}
3817 @c ***GFDL********************************************************************
3825 @uref{./index.html,,Return to the GCC Installation page}
3829 @c ***************************************************************************
3830 @c Part 6 The End of the Document
3832 @comment node-name, next, previous, up
3833 @node Concept Index, , GNU Free Documentation License, Top
3837 @unnumbered Concept Index