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 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 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.
91 @c Part 3 Titlepage and Copyright
94 @comment The title is printed in a large font.
95 @center @titlefont{Installing GCC}
97 @c The following two commands start the copyright page.
99 @vskip 0pt plus 1filll
103 @c Part 4 Top node and Master Menu
106 @comment node-name, next, Previous, up
109 * Installing GCC:: This document describes the generic installation
110 procedure for GCC as well as detailing some target
111 specific installation instructions.
113 * Specific:: Host/target specific installation notes for GCC.
114 * Binaries:: Where to get pre-compiled binaries.
116 * Old:: Old installation documentation.
118 * GNU Free Documentation License:: How you can copy and share this manual.
119 * Concept Index:: This index has two entries.
123 @c Part 5 The Body of the Document
124 @c ***Installing GCC**********************************************************
126 @comment node-name, next, previous, up
127 @node Installing GCC, Binaries, , Top
131 @chapter Installing GCC
134 The latest version of this document is always available at
135 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
137 This document describes the generic installation procedure for GCC as well
138 as detailing some target specific installation instructions.
140 GCC includes several components that previously were separate distributions
141 with their own installation instructions. This document supersedes all
142 package specific installation instructions.
144 @emph{Before} starting the build/install procedure please check the
146 @ref{Specific, host/target specific installation notes}.
149 @uref{specific.html,,host/target specific installation notes}.
151 We recommend you browse the entire generic installation instructions before
154 Lists of successful builds for released versions of GCC are
155 available at @uref{http://gcc.gnu.org/buildstat.html}.
156 These lists are updated as new information becomes available.
158 The installation procedure itself is broken into five steps.
163 * Downloading the source::
166 * Testing:: (optional)
173 @uref{prerequisites.html,,Prerequisites}
175 @uref{download.html,,Downloading the source}
177 @uref{configure.html,,Configuration}
179 @uref{build.html,,Building}
181 @uref{test.html,,Testing} (optional)
183 @uref{finalinstall.html,,Final install}
187 Please note that GCC does not support @samp{make uninstall} and probably
188 won't do so in the near future as this would open a can of worms. Instead,
189 we suggest that you install GCC into a directory of its own and simply
190 remove that directory when you do not need that specific version of GCC
191 any longer, and, if shared libraries are installed there as well, no
192 more binaries exist that use them.
195 There are also some @uref{old.html,,old installation instructions},
196 which are mostly obsolete but still contain some information which has
197 not yet been merged into the main part of this manual.
205 @uref{./index.html,,Return to the GCC Installation page}
211 @c ***Prerequisites**************************************************
213 @comment node-name, next, previous, up
214 @node Prerequisites, Downloading the source, , Installing GCC
216 @ifset prerequisiteshtml
218 @chapter Prerequisites
220 @cindex Prerequisites
222 GCC requires that various tools and packages be available for use in the
223 build procedure. Modifying GCC sources requires additional tools
226 @heading Tools/packages necessary for building GCC
228 @item ISO C90 compiler
229 Necessary to bootstrap the GCC package, although versions of GCC prior
230 to 3.4 also allow bootstrapping with a traditional (K&R) C compiler.
232 To make all languages in a cross-compiler or other configuration where
233 3-stage bootstrap is not performed, you need to start with an existing
234 GCC binary (version 2.95 or later) because source code for language
235 frontends other than C might use GCC extensions.
239 In order to build the Ada compiler (GNAT) you must already have GNAT
240 installed because portions of the Ada frontend are written in Ada (with
241 GNAT extensions.) Refer to the Ada installation instructions for more
242 specific information.
244 @item A ``working'' POSIX compatible shell, or GNU bash
246 Necessary when running @command{configure} because some
247 @command{/bin/sh} shells have bugs and may crash when configuring the
248 target libraries. In other cases, @command{/bin/sh} or even some
249 @command{ksh} have disastrous corner-case performance problems. This
250 can cause target @command{configure} runs to literally take days to
251 complete in some cases.
253 So on some platforms @command{/bin/ksh} is sufficient, on others it
254 isn't. See the host/target specific instructions for your platform, or
255 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
256 environment to your ``good'' shell prior to running
257 @command{configure}/@command{make}.
259 It is known that @command{zsh} is not a fully compliant POSIX shell and will not
260 work when configuring gcc.
264 Necessary in some circumstances, optional in others. See the
265 host/target specific instructions for your platform for the exact
268 @item gzip version 1.2.4 (or later) or
269 @itemx bzip2 version 1.0.2 (or later)
271 Necessary to uncompress GCC @command{tar} files when source code is
272 obtained via FTP mirror sites.
274 @item GNU make version 3.79.1 (or later)
276 You must have GNU make installed to build GCC.
278 @item GNU tar version 1.12 (or later)
280 Necessary (only on some platforms) to untar the source code. Many
281 systems' @command{tar} programs will also work, only try GNU
282 @command{tar} if you have problems.
287 @heading Tools/packages necessary for modifying GCC
290 @item autoconf version 2.13 (NO earlier or later versions) and
291 @itemx GNU m4 version 1.4 (or later)
293 Necessary when modifying @file{configure.in}, @file{aclocal.m4}, etc.@:
294 to regenerate @file{configure} and @file{config.in} files
296 @item automake version ???
298 Necessary when modifying a @file{Makefile.am} file to regenerate its
299 associated @file{Makefile.in}
301 @item gperf version 2.7.2 (or later)
303 Necessary when modifying @command{gperf} input files, e.g.@:
304 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
305 @file{gcc/cp/cfns.h}.
307 @item expect version ???
308 @itemx tcl version ???
309 @itemx dejagnu version ???
311 Necessary to run the GCC testsuite.
313 @item autogen version 5.5.4 (or later) and
314 @itemx guile version 1.4.1 (or later)
316 Necessary to regenerate @file{fixinc/fixincl.x} from
317 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
319 Necessary to run the @file{fixinc} @command{make check}.
321 Necessary to regenerate the top level @file{Makefile.am} files from
322 @file{Makefile.tpl} and @file{Makefile.def}.
324 @item GNU Bison version 1.28 (or later)
325 Berkeley @command{yacc} (@command{byacc}) is also reported to work other
328 Necessary when modifying @file{*.y} files.
330 Necessary to build GCC during development because the generated output
331 files are not included in the CVS repository. They are included in
334 @item Flex version 2.5.4 (or later)
336 Necessary when modifying @file{*.l} files.
338 Necessary to build GCC during development because the generated output
339 files are not included in the CVS repository. They are included in
342 @item Texinfo version 4.2 (or later)
344 Necessary for running @command{makeinfo} when modifying @file{*.texi}
345 files to test your changes.
347 Necessary to build GCC documentation during development because the
348 generated output files are not included in the CVS repository. They are
349 included in releases.
351 @item @TeX{} (any working version)
353 Necessary for running @command{texi2dvi}, used when running
354 @command{make dvi} to create DVI files.
356 @item cvs version 1.10 (or later)
357 @itemx ssh (any version)
359 Necessary to access the CVS repository. Public releases and weekly
360 snapshots of the development sources are also available via FTP.
362 @item perl version 5.6.1 (or later)
364 Necessary when regenerating @file{Makefile} dependencies in libiberty.
365 Necessary when regenerating something with intl??? (pod2man???)
368 @item GNU diffutils version 2.7 (or later)
370 Necessary when creating changes to GCC source code to submit for review.
372 @item patch version 2.5.4 (or later)
374 Necessary when applying patches, created with @command{diff}, to one's
384 @uref{./index.html,,Return to the GCC Installation page}
388 @c ***Downloading the source**************************************************
390 @comment node-name, next, previous, up
391 @node Downloading the source, Configuration, Prerequisites, Installing GCC
395 @chapter Downloading GCC
397 @cindex Downloading GCC
398 @cindex Downloading the Source
400 GCC is distributed via @uref{http://gcc.gnu.org/cvs.html,,CVS} and FTP
401 tarballs compressed with @command{gzip} or
402 @command{bzip2}. It is possible to download a full distribution or specific
405 Please refer to our @uref{http://gcc.gnu.org/releases.html,,releases web page}
406 for information on how to obtain GCC@.
408 The full distribution includes the C, C++, Objective-C, Fortran, Java,
409 and Ada (in case of GCC 3.1 and later) compilers. The full distribution
410 also includes runtime libraries for C++, Objective-C, Fortran, and Java.
411 In GCC 3.0 and later versions, GNU compiler testsuites are also included
412 in the full distribution.
414 If you choose to download specific components, you must download the core
415 GCC distribution plus any language specific distributions you wish to
416 use. The core distribution includes the C language front end as well as the
417 shared components. Each language has a tarball which includes the language
418 front end as well as the language runtime (when appropriate).
420 Unpack the core distribution as well as any language specific
421 distributions in the same directory.
423 If you also intend to build binutils (either to upgrade an existing
424 installation or for use in place of the corresponding tools of your
425 OS), unpack the binutils distribution either in the same directory or
426 a separate one. In the latter case, add symbolic links to any
427 components of the binutils you intend to build alongside the compiler
428 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
429 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
436 @uref{./index.html,,Return to the GCC Installation page}
440 @c ***Configuration***********************************************************
442 @comment node-name, next, previous, up
443 @node Configuration, Building, Downloading the source, Installing GCC
447 @chapter Installing GCC: Configuration
449 @cindex Configuration
450 @cindex Installing GCC: Configuration
452 Like most GNU software, GCC must be configured before it can be built.
453 This document describes the recommended configuration procedure
454 for both native and cross targets.
456 We use @var{srcdir} to refer to the toplevel source directory for
457 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
459 If you obtained the sources via CVS, @var{srcdir} must refer to the top
460 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
461 and not its @file{gcc} subdirectory, otherwise the build will fail.
463 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
464 file system, the shell's built-in @command{pwd} command will return
465 temporary pathnames. Using these can lead to various sorts of build
466 problems. To avoid this issue, set the @env{PWDCMD} environment
467 variable to an automounter-aware @command{pwd} command, e.g.,
468 @command{pawd} or @samp{amq -w}, during the configuration and build
471 First, we @strong{highly} recommend that GCC be built into a
472 separate directory than the sources which does @strong{not} reside
473 within the source tree. This is how we generally build GCC; building
474 where @var{srcdir} == @var{objdir} should still work, but doesn't
475 get extensive testing; building where @var{objdir} is a subdirectory
476 of @var{srcdir} is unsupported.
478 If you have previously built GCC in the same directory for a
479 different target machine, do @samp{make distclean} to delete all files
480 that might be invalid. One of the files this deletes is @file{Makefile};
481 if @samp{make distclean} complains that @file{Makefile} does not exist
482 or issues a message like ``don't know how to make distclean'' it probably
483 means that the directory is already suitably clean. However, with the
484 recommended method of building in a separate @var{objdir}, you should
485 simply use a different @var{objdir} for each target.
487 Second, when configuring a native system, either @command{cc} or
488 @command{gcc} must be in your path or you must set @env{CC} in
489 your environment before running configure. Otherwise the configuration
492 Note that the bootstrap compiler and the resulting GCC must be link
493 compatible, else the bootstrap will fail with linker errors about
494 incompatible object file formats. Several multilibed targets are
495 affected by this requirement, see
497 @ref{Specific, host/target specific installation notes}.
500 @uref{specific.html,,host/target specific installation notes}.
508 % @var{srcdir}/configure [@var{options}] [@var{target}]
512 @heading Target specification
515 GCC has code to correctly determine the correct value for @var{target}
516 for nearly all native systems. Therefore, we highly recommend you not
517 provide a configure target when configuring a native compiler.
520 @var{target} must be specified as @option{--target=@var{target}}
521 when configuring a cross compiler; examples of valid targets would be
522 i960-rtems, m68k-coff, sh-elf, etc.
525 Specifying just @var{target} instead of @option{--target=@var{target}}
526 implies that the host defaults to @var{target}.
530 @heading Options specification
532 Use @var{options} to override several configure time options for
533 GCC@. A list of supported @var{options} follows; @samp{configure
534 --help} may list other options, but those not listed below may not
535 work and should not normally be used.
538 @item --prefix=@var{dirname}
539 Specify the toplevel installation
540 directory. This is the recommended way to install the tools into a directory
541 other than the default. The toplevel installation directory defaults to
544 We @strong{highly} recommend against @var{dirname} being the same or a
545 subdirectory of @var{objdir} or vice versa. If specifying a directory
546 beneath a user's home directory tree, some shells will not expand
547 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
550 The following standard @command{autoconf} options are supported. Normally you
551 should not need to use these options.
553 @item --exec-prefix=@var{dirname}
554 Specify the toplevel installation directory for architecture-dependent
555 files. The default is @file{@var{prefix}}.
557 @item --bindir=@var{dirname}
558 Specify the installation directory for the executables called by users
559 (such as @command{gcc} and @command{g++}). The default is
560 @file{@var{exec-prefix}/bin}.
562 @item --libdir=@var{dirname}
563 Specify the installation directory for object code libraries and
564 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
566 @item --libexecdir=@var{dirname}
567 Specify the installation directory for internal executables of GCC@.
568 The default is @file{@var{exec-prefix}/libexec}.
570 @item --with-slibdir=@var{dirname}
571 Specify the installation directory for the shared libgcc library. The
572 default is @file{@var{libdir}}.
574 @item --infodir=@var{dirname}
575 Specify the installation directory for documentation in info format.
576 The default is @file{@var{prefix}/info}.
578 @item --datadir=@var{dirname}
579 Specify the installation directory for some architecture-independent
580 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
582 @item --mandir=@var{dirname}
583 Specify the installation directory for manual pages. The default is
584 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
585 the full GCC manuals, which are provided in Texinfo format. The manpages
586 are derived by an automatic conversion process from parts of the full
589 @item --with-gxx-include-dir=@var{dirname}
591 the installation directory for G++ header files. The default is
592 @file{@var{prefix}/include/g++-v3}.
596 @item --program-prefix=@var{prefix}
597 GCC supports some transformations of the names of its programs when
598 installing them. This option prepends @var{prefix} to the names of
599 programs to install in @var{bindir} (see above). For example, specifying
600 @option{--program-prefix=foo-} would result in @samp{gcc}
601 being installed as @file{/usr/local/bin/foo-gcc}.
603 @item --program-suffix=@var{suffix}
604 Appends @var{suffix} to the names of programs to install in @var{bindir}
605 (see above). For example, specifying @option{--program-suffix=-3.1}
606 would result in @samp{gcc} being installed as
607 @file{/usr/local/bin/gcc-3.1}.
609 @item --program-transform-name=@var{pattern}
610 Applies the @samp{sed} script @var{pattern} to be applied to the names
611 of programs to install in @var{bindir} (see above). @var{pattern} has to
612 consist of one or more basic @samp{sed} editing commands, separated by
613 semicolons. For example, if you want the @samp{gcc} program name to be
614 transformed to the installed program @file{/usr/local/bin/myowngcc} and
615 the @samp{g++} program name to be transformed to
616 @file{/usr/local/bin/gspecial++} without changing other program names,
617 you could use the pattern
618 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
619 to achieve this effect.
621 All three options can be combined and used together, resulting in more
622 complex conversion patterns. As a basic rule, @var{prefix} (and
623 @var{suffix}) are prepended (appended) before further transformations
624 can happen with a special transformation script @var{pattern}.
626 As currently implemented, this option only takes effect for native
627 builds; cross compiler binaries' names are not transformed even when a
628 transformation is explicitly asked for by one of these options.
630 For native builds, some of the installed programs are also installed
631 with the target alias in front of their name, as in
632 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
633 before the target alias is prepended to the name - so, specifying
634 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
635 resulting binary would be installed as
636 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
638 As a last shortcoming, none of the installed Ada programs are
639 transformed yet, which will be fixed in some time.
641 @item --with-local-prefix=@var{dirname}
643 installation directory for local include files. The default is
644 @file{/usr/local}. Specify this option if you want the compiler to
645 search directory @file{@var{dirname}/include} for locally installed
646 header files @emph{instead} of @file{/usr/local/include}.
648 You should specify @option{--with-local-prefix} @strong{only} if your
649 site has a different convention (not @file{/usr/local}) for where to put
652 The default value for @option{--with-local-prefix} is @file{/usr/local}
653 regardless of the value of @option{--prefix}. Specifying
654 @option{--prefix} has no effect on which directory GCC searches for
655 local header files. This may seem counterintuitive, but actually it is
658 The purpose of @option{--prefix} is to specify where to @emph{install
659 GCC}. The local header files in @file{/usr/local/include}---if you put
660 any in that directory---are not part of GCC@. They are part of other
661 programs---perhaps many others. (GCC installs its own header files in
662 another directory which is based on the @option{--prefix} value.)
664 Both the local-prefix include directory and the GCC-prefix include
665 directory are part of GCC's "system include" directories. Although these
666 two directories are not fixed, they need to be searched in the proper
667 order for the correct processing of the include_next directive. The
668 local-prefix include directory is searched before the GCC-prefix
669 include directory. Another characteristic of system include directories
670 is that pedantic warnings are turned off for headers in these directories.
672 Some autoconf macros add @option{-I @var{directory}} options to the
673 compiler command line, to ensure that directories containing installed
674 packages' headers are searched. When @var{directory} is one of GCC's
675 system include directories, GCC will ignore the option so that system
676 directories continue to be processed in the correct order. This
677 may result in a search order different from what was specified but the
678 directory will still be searched.
680 GCC automatically searches for ordinary libraries using
681 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
682 used for both GCC and packages, GCC will automatically search for
683 both headers and libraries. This provides a configuration that is
684 easy to use. GCC behaves in a manner similar to that when it is
685 installed as a system compiler in @file{/usr}.
687 Sites that need to install multiple versions of GCC may not want to
688 use the above simple configuration. It is possible to use the
689 @option{--program-prefix}, @option{--program-suffix} and
690 @option{--program-transform-name} options to install multiple versions
691 into a single directory, but it may be simpler to use different prefixes
692 and the @option{--with-local-prefix} option to specify the location of the
693 site-specific files for each version. It will then be necessary for
694 users to specify explicitly the location of local site libraries
695 (e.g., with @env{LIBRARY_PATH}).
697 The same value can be used for both @option{--with-local-prefix} and
698 @option{--prefix} provided it is not @file{/usr}. This can be used
699 to avoid the default search of @file{/usr/local/include}.
701 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
702 The directory you use for @option{--with-local-prefix} @strong{must not}
703 contain any of the system's standard header files. If it did contain
704 them, certain programs would be miscompiled (including GNU Emacs, on
705 certain targets), because this would override and nullify the header
706 file corrections made by the @command{fixincludes} script.
708 Indications are that people who use this option use it based on mistaken
709 ideas of what it is for. People use it as if it specified where to
710 install part of GCC@. Perhaps they make this assumption because
711 installing GCC creates the directory.
713 @item --enable-shared[=@var{package}[,@dots{}]]
714 Build shared versions of libraries, if shared libraries are supported on
715 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
716 are enabled by default on all platforms that support shared libraries,
717 except for @samp{libobjc} which is built as a static library only by
720 If a list of packages is given as an argument, build shared libraries
721 only for the listed packages. For other packages, only static libraries
722 will be built. Package names currently recognized in the GCC tree are
723 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
724 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc} and
725 @samp{libjava}. Note that @samp{libobjc} does not recognize itself by
726 any name, so, if you list package names in @option{--enable-shared},
727 you will only get static Objective-C libraries. @samp{libf2c} and
728 @samp{libiberty} do not support shared libraries at all.
730 Use @option{--disable-shared} to build only static libraries. Note that
731 @option{--disable-shared} does not accept a list of package names as
732 argument, only @option{--enable-shared} does.
734 @item @anchor{with-gnu-as}--with-gnu-as
735 Specify that the compiler should assume that the
736 assembler it finds is the GNU assembler. However, this does not modify
737 the rules to find an assembler and will result in confusion if the
738 assembler found is not actually the GNU assembler. (Confusion may also
739 result if the compiler finds the GNU assembler but has not been
740 configured with @option{--with-gnu-as}.) If you have more than one
741 assembler installed on your system, you may want to use this option in
742 connection with @option{--with-as=@var{pathname}}.
744 The following systems are the only ones where it makes a difference
745 whether you use the GNU assembler. On any other system,
746 @option{--with-gnu-as} has no effect.
749 @item @samp{hppa1.0-@var{any}-@var{any}}
750 @item @samp{hppa1.1-@var{any}-@var{any}}
751 @item @samp{i386-@var{any}-sysv}
752 @item @samp{m68k-bull-sysv}
753 @item @samp{m68k-hp-hpux}
754 @item @samp{m68000-hp-hpux}
755 @item @samp{m68000-att-sysv}
756 @item @samp{@var{any}-lynx-lynxos}
757 @item @samp{mips-@var{any}}
758 @item @samp{sparc-sun-solaris2.@var{any}}
759 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
762 On the systems listed above (except for the HP-PA, the SPARC, for ISC on
763 the 386, and for @samp{mips-sgi-irix5.*}), if you use the GNU assembler,
764 you should also use the GNU linker (and specify @option{--with-gnu-ld}).
766 @item @anchor{with-as}--with-as=@var{pathname}
768 compiler should use the assembler pointed to by @var{pathname}, rather
769 than the one found by the standard rules to find an assembler, which
773 Check the @file{@var{libexec}/gcc/@var{target}/@var{version}}
774 directory, where @var{libexec} defaults to
775 @file{@var{exec-prefix}/libexec} and @var{exec-prefix} defaults to
776 @var{prefix} which defaults to @file{/usr/local} unless overridden by
777 the @option{--prefix=@var{pathname}} switch described
778 above. @var{target} is the target system triple, such as
779 @samp{sparc-sun-solaris2.7}, and @var{version} denotes the GCC
780 version, such as 3.0.
782 Check operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
785 Note that these rules do not check for the value of @env{PATH}. You may
786 want to use @option{--with-as} if no assembler is installed in the
787 directories listed above, or if you have multiple assemblers installed
788 and want to choose one that is not found by the above rules.
790 @item @anchor{with-gnu-ld}--with-gnu-ld
791 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
794 @item --with-ld=@var{pathname}
795 Same as @uref{#with-as,,@option{--with-as}}
799 Specify that stabs debugging
800 information should be used instead of whatever format the host normally
801 uses. Normally GCC uses the same debug format as the host system.
803 On MIPS based systems and on Alphas, you must specify whether you want
804 GCC to create the normal ECOFF debugging format, or to use BSD-style
805 stabs passed through the ECOFF symbol table. The normal ECOFF debug
806 format cannot fully handle languages other than C@. BSD stabs format can
807 handle other languages, but it only works with the GNU debugger GDB@.
809 Normally, GCC uses the ECOFF debugging format by default; if you
810 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
812 No matter which default you choose when you configure GCC, the user
813 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
814 the debug format for a particular compilation.
816 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
817 @option{--with-gas} is used. It selects use of stabs debugging
818 information embedded in COFF output. This kind of debugging information
819 supports C++ well; ordinary COFF debugging information does not.
821 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
822 selects use of stabs debugging information embedded in ELF output. The
823 C++ compiler currently (2.6.0) does not support the DWARF debugging
824 information normally used on 386 SVR4 platforms; stabs provide a
825 workable alternative. This requires gas and gdb, as the normal SVR4
826 tools can not generate or interpret stabs.
828 @item --disable-multilib
829 Specify that multiple target
830 libraries to support different target variants, calling
831 conventions, etc should not be built. The default is to build a
832 predefined set of them.
834 Some targets provide finer-grained control over which multilibs are built
835 (e.g., @option{--disable-softfloat}):
841 fpu, 26bit, underscore, interwork, biendian, nofmult.
844 softfloat, m68881, m68000, m68020.
847 single-float, biendian, softfloat.
849 @item powerpc*-*-*, rs6000*-*-*
850 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
855 @item --enable-threads
856 Specify that the target
857 supports threads. This affects the Objective-C compiler and runtime
858 library, and exception handling for other languages like C++ and Java.
859 On some systems, this is the default.
861 In general, the best (and, in many cases, the only known) threading
862 model available will be configured for use. Beware that on some
863 systems, gcc has not been taught what threading models are generally
864 available for the system. In this case, @option{--enable-threads} is an
865 alias for @option{--enable-threads=single}.
867 @item --disable-threads
868 Specify that threading support should be disabled for the system.
869 This is an alias for @option{--enable-threads=single}.
871 @item --enable-threads=@var{lib}
873 @var{lib} is the thread support library. This affects the Objective-C
874 compiler and runtime library, and exception handling for other languages
875 like C++ and Java. The possibilities for @var{lib} are:
883 Ada tasking support. For non-Ada programs, this setting is equivalent
884 to @samp{single}. When used in conjunction with the Ada run time, it
885 causes GCC to use the same thread primitives as Ada uses. This option
886 is necessary when using both Ada and the back end exception handling,
887 which is the default for most Ada targets.
889 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
890 that the file needed to support this configuration, @file{gthr-mach.h}, is
891 missing and thus this setting will cause a known bootstrap failure.)
893 This is an alias for @samp{single}.
895 Generic POSIX thread support.
897 RTEMS thread support.
899 Disable thread support, should work for all platforms.
901 Sun Solaris 2 thread support.
903 VxWorks thread support.
905 Microsoft Win32 API thread support.
908 @item --with-cpu=@var{cpu}
909 Specify which cpu variant the compiler should generate code for by default.
910 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
911 This option is only supported on some targets, including ARM, i386, PowerPC,
914 @item --with-schedule=@var{cpu}
915 @itemx --with-arch=@var{cpu}
916 @itemx --with-tune=@var{cpu}
917 @itemx --with-abi=@var{abi}
918 @itemx --with-float=@var{type}
919 These configure options provide default values for the @option{-mschedule=},
920 @option{-march=}, @option{-mtune=}, and @option{-mabi=} options and for
921 @option{-mhard-float} or @option{-msoft-float}. As with @option{--with-cpu},
922 which switches will be accepted and acceptable values of the arguments depend
925 @item --enable-altivec
926 Specify that the target supports AltiVec vector enhancements. This
927 option will adjust the ABI for AltiVec enhancements, as well as generate
928 AltiVec code when appropriate. This option is only available for
931 @item --enable-target-optspace
933 libraries should be optimized for code space instead of code speed.
934 This is the default for the m32r platform.
937 Specify that a user visible @command{cpp} program should not be installed.
939 @item --with-cpp-install-dir=@var{dirname}
940 Specify that the user visible @command{cpp} program should be installed
941 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
943 @item --enable-initfini-array
944 Force the use of sections @code{.init_array} and @code{.fini_array}
945 (instead of @code{.init} and @code{.fini}) for constructors and
946 destructors. Option @option{--disable-initfini-array} has the
947 opposite effect. If neither option is specified, the configure script
948 will try to guess whether the @code{.init_array} and
949 @code{.fini_array} sections are supported and, if they are, use them.
951 @item --enable-maintainer-mode
953 regenerate the GCC master message catalog @file{gcc.pot} are normally
954 disabled. This is because it can only be rebuilt if the complete source
955 tree is present. If you have changed the sources and want to rebuild the
956 catalog, configuring with @option{--enable-maintainer-mode} will enable
957 this. Note that you need a recent version of the @code{gettext} tools
960 @item --enable-version-specific-runtime-libs
962 that runtime libraries should be installed in the compiler specific
963 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
964 addition, @samp{libstdc++}'s include files will be installed into
965 @file{@var{libdir}} unless you overruled it by using
966 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
967 particularly useful if you intend to use several versions of GCC in
968 parallel. This is currently supported by @samp{libf2c} and
969 @samp{libstdc++}, and is the default for @samp{libobjc} which cannot be
970 changed in this case.
972 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
973 Specify that only a particular subset of compilers and
974 their runtime libraries should be built. For a list of valid values for
975 @var{langN} you can issue the following command in the
976 @file{gcc} directory of your GCC source tree:@*
978 grep language= */config-lang.in
980 Currently, you can use any of the following:
981 @code{ada}, @code{c}, @code{c++}, @code{f77}, @code{java}, @code{objc}.
982 Building the Ada compiler has special requirements, see below.@*
983 If you do not pass this flag, all languages available in the @file{gcc}
984 sub-tree will be configured. Re-defining @code{LANGUAGES} when calling
985 @samp{make bootstrap} @strong{does not} work anymore, as those
986 language sub-directories might not have been configured!
988 @item --disable-libgcj
989 Specify that the run-time libraries
990 used by GCJ should not be built. This is useful in case you intend
991 to use GCJ with some other run-time, or you're going to install it
992 separately, or it just happens not to build on your particular
993 machine. In general, if the Java front end is enabled, the GCJ
994 libraries will be enabled too, unless they're known to not work on
995 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
996 may need to port it; in this case, before modifying the top-level
997 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
998 you may use @option{--enable-libgcj} to override the default.
1001 Specify that the compiler should
1002 use DWARF 2 debugging information as the default.
1004 @item --enable-win32-registry
1005 @itemx --enable-win32-registry=@var{key}
1006 @itemx --disable-win32-registry
1007 The @option{--enable-win32-registry} option enables Windows-hosted GCC
1008 to look up installations paths in the registry using the following key:
1011 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1014 @var{key} defaults to GCC version number, and can be overridden by the
1015 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1016 who use custom installers are encouraged to provide a different key,
1017 perhaps one comprised of vendor name and GCC version number, to
1018 avoid conflict with existing installations. This feature is enabled
1019 by default, and can be disabled by @option{--disable-win32-registry}
1020 option. This option has no effect on the other hosts.
1023 Specify that the machine does not have a floating point unit. This
1024 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1025 system, @option{--nfp} has no effect.
1027 @item --enable-werror
1028 @itemx --disable-werror
1029 @itemx --enable-werror=yes
1030 @itemx --enable-werror=no
1031 When you specify this option, it controls whether certain files in the
1032 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1033 If you don't specify it, @option{-Werror} is turned on for the main
1034 development trunk. However it defaults to off for release branches and
1035 final releases. The specific files which get @option{-Werror} are
1036 controlled by the Makefiles.
1038 @item --enable-checking
1039 @itemx --enable-checking=@var{list}
1040 When you specify this option, the compiler is built to perform checking
1041 of tree node types when referencing fields of that node, and some other
1042 internal consistency checks. This does not change the generated code,
1043 but adds error checking within the compiler. This will slow down the
1044 compiler and may only work properly if you are building the compiler
1045 with GCC@. This is on by default when building from CVS or snapshots,
1046 but off for releases. More control over the checks may be had by
1047 specifying @var{list}; the categories of checks available are
1048 @samp{misc}, @samp{tree}, @samp{gc}, @samp{rtl}, @samp{rtlflag},
1049 @samp{fold}, @samp{gcac} and @samp{valgrind}. The check @samp{valgrind}
1050 requires the external @command{valgrind} simulator, available from
1051 @uref{http://developer.kde.org/~sewardj/}. The default when @var{list} is
1052 not specified is @samp{misc,tree,gc,rtlflag}; the checks @samp{rtl},
1053 @samp{gcac} and @samp{valgrind} are very expensive.
1055 @item --enable-coverage
1056 @item --enable-coverage=@var{level}
1057 With this option, the compiler is built to collect self coverage
1058 information, every time it is run. This is for internal development
1059 purposes, and only works when the compiler is being built with gcc. The
1060 @var{level} argument controls whether the compiler is built optimized or
1061 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1062 want to disable optimization, for performance analysis you want to
1063 enable optimization. When coverage is enabled, the default level is
1064 without optimization.
1067 @itemx --disable-nls
1068 The @option{--enable-nls} option enables Native Language Support (NLS),
1069 which lets GCC output diagnostics in languages other than American
1070 English. Native Language Support is enabled by default if not doing a
1071 canadian cross build. The @option{--disable-nls} option disables NLS@.
1073 @item --with-included-gettext
1074 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1075 procedure to prefer its copy of GNU @command{gettext}.
1077 @item --with-catgets
1078 If NLS is enabled, and if the host lacks @code{gettext} but has the
1079 inferior @code{catgets} interface, the GCC build procedure normally
1080 ignores @code{catgets} and instead uses GCC's copy of the GNU
1081 @code{gettext} library. The @option{--with-catgets} option causes the
1082 build procedure to use the host's @code{catgets} in this situation.
1084 @item --with-libiconv-prefix=@var{dir}
1085 Search for libiconv header files in @file{@var{dir}/include} and
1086 libiconv library files in @file{@var{dir}/lib}.
1088 @item --with-system-zlib
1089 Use installed zlib rather than that included with GCC@. This option
1090 only applies if the Java front end is being built.
1092 @item --enable-obsolete
1093 Enable configuration for an obsoleted system. If you attempt to
1094 configure GCC for a system (build, host, or target) which has been
1095 obsoleted, and you do not specify this flag, configure will halt with an
1098 All support for systems which have been obsoleted in one release of GCC
1099 is removed entirely in the next major release, unless someone steps
1100 forward to maintain the port.
1103 Some options which only apply to building cross compilers:
1105 @item --with-sysroot
1106 @itemx --with-sysroot=@var{dir}
1107 Tells GCC to consider @var{dir} as the root of a tree that contains a
1108 (subset of) the root filesystem of the target operating system.
1109 Target system headers, libraries and run-time object files will be
1110 searched in there. The specified directory is not copied into the
1111 install tree, unlike the options @option{--with-headers} and
1112 @option{--with-libs} that this option obsoletes. The default value,
1113 in case @option{--with-sysroot} is not given an argument, is
1114 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1115 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1116 the GCC binaries if the installation tree is moved.
1118 @item --with-headers
1119 @itemx --with-headers=@var{dir}
1120 Deprecated in favor of @option{--with-sysroot}.
1121 Specifies that target headers are available when building a cross compiler.
1122 The @var{dir} argument specifies a directory which has the target include
1123 files. These include files will be copied into the @file{gcc} install
1124 directory. @emph{This option with the @var{dir} argument is required} when
1125 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1126 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1127 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1128 will be run on these files to make them compatible with GCC.
1130 @item --without-headers
1131 Tells GCC not use any target headers from a libc when building a cross
1132 compiler. When crossing to GNU/Linux, you need the headers so gcc
1133 can build the exception handling for libgcc.
1134 See @uref{http://www.objsw.com/CrossGCC/,,CrossGCC} for more information
1138 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1139 Deprecated in favor of @option{--with-sysroot}.
1140 Specifies a list of directories which contain the target runtime
1141 libraries. These libraries will be copied into the @file{gcc} install
1142 directory. If the directory list is omitted, this option has no
1145 Specifies that @samp{newlib} is
1146 being used as the target C library. This causes @code{__eprintf} to be
1147 omitted from @file{libgcc.a} on the assumption that it will be provided by
1151 Note that each @option{--enable} option has a corresponding
1152 @option{--disable} option and that each @option{--with} option has a
1153 corresponding @option{--without} option.
1160 @uref{./index.html,,Return to the GCC Installation page}
1164 @c ***Building****************************************************************
1166 @comment node-name, next, previous, up
1167 @node Building, Testing, Configuration, Installing GCC
1173 @cindex Installing GCC: Building
1175 Now that GCC is configured, you are ready to build the compiler and
1178 We @strong{highly} recommend that GCC be built using GNU make;
1179 other versions may work, then again they might not.
1180 GNU make is required for compiling GNAT (the Ada compiler) and the Java
1183 (For example, many broken versions of make will fail if you use the
1184 recommended setup where @var{objdir} is different from @var{srcdir}.
1185 Other broken versions may recompile parts of the compiler when
1186 installing the compiler.)
1188 Some commands executed when making the compiler may fail (return a
1189 nonzero status) and be ignored by @command{make}. These failures, which
1190 are often due to files that were not found, are expected, and can safely
1193 It is normal to have compiler warnings when compiling certain files.
1194 Unless you are a GCC developer, you can generally ignore these warnings
1195 unless they cause compilation to fail. Developers should attempt to fix
1196 any warnings encountered, however they can temporarily continue past
1197 warnings-as-errors by specifying the configure flag
1198 @option{--disable-werror}.
1200 On certain old systems, defining certain environment variables such as
1201 @env{CC} can interfere with the functioning of @command{make}.
1203 If you encounter seemingly strange errors when trying to build the
1204 compiler in a directory other than the source directory, it could be
1205 because you have previously configured the compiler in the source
1206 directory. Make sure you have done all the necessary preparations.
1208 If you build GCC on a BSD system using a directory stored in an old System
1209 V file system, problems may occur in running @command{fixincludes} if the
1210 System V file system doesn't support symbolic links. These problems
1211 result in a failure to fix the declaration of @code{size_t} in
1212 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1213 that type mismatches occur, this could be the cause.
1215 The solution is not to use such a directory for building GCC@.
1217 When building from CVS or snapshots, or if you modify parser sources,
1218 you need the Bison parser generator installed. Any version 1.25 or
1219 later should work; older versions may also work. If you do not modify
1220 parser sources, releases contain the Bison-generated files and you do
1221 not need Bison installed to build them.
1223 When building from CVS or snapshots, or if you modify Texinfo
1224 documentation, you need version 4.2 or later of Texinfo installed if you
1225 want Info documentation to be regenerated. Releases contain Info
1226 documentation pre-built for the unmodified documentation in the release.
1228 @section Building a native compiler
1230 For a native build issue the command @samp{make bootstrap}. This
1231 will build the entire GCC system, which includes the following steps:
1235 Build host tools necessary to build the compiler such as texinfo, bison,
1239 Build target tools for use by the compiler such as binutils (bfd,
1240 binutils, gas, gprof, ld, and opcodes)
1241 if they have been individually linked
1242 or moved into the top level GCC source tree before configuring.
1245 Perform a 3-stage bootstrap of the compiler.
1248 Perform a comparison test of the stage2 and stage3 compilers.
1251 Build runtime libraries using the stage3 compiler from the previous step.
1255 If you are short on disk space you might consider @samp{make
1256 bootstrap-lean} instead. This is identical to @samp{make
1257 bootstrap} except that object files from the stage1 and
1258 stage2 of the 3-stage bootstrap of the compiler are deleted as
1259 soon as they are no longer needed.
1261 If you want to save additional space during the bootstrap and in
1262 the final installation as well, you can build the compiler binaries
1263 without debugging information as in the following example. This will save
1264 roughly 40% of disk space both for the bootstrap and the final installation.
1265 (Libraries will still contain debugging information.)
1268 make CFLAGS='-O' LIBCFLAGS='-g -O2' \
1269 LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap
1272 If you wish to use non-default GCC flags when compiling the stage2 and
1273 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
1274 @samp{make bootstrap}. Non-default optimization flags are less well
1275 tested here than the default of @samp{-g -O2}, but should still work.
1276 In a few cases, you may find that you need to specify special flags such
1277 as @option{-msoft-float} here to complete the bootstrap; or, if the
1278 native compiler miscompiles the stage1 compiler, you may need to work
1279 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
1280 stage1 compiler that were miscompiled, or by using @samp{make
1281 bootstrap4} to increase the number of stages of bootstrap.
1283 If you used the flag @option{--enable-languages=@dots{}} to restrict
1284 the compilers to be built, only those you've actually enabled will be
1285 built. This will of course only build those runtime libraries, for
1286 which the particular compiler has been built. Please note,
1287 that re-defining @env{LANGUAGES} when calling @samp{make bootstrap}
1288 @strong{does not} work anymore!
1290 If the comparison of stage2 and stage3 fails, this normally indicates
1291 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1292 a potentially serious bug which you should investigate and report. (On
1293 a few systems, meaningful comparison of object files is impossible; they
1294 always appear ``different''. If you encounter this problem, you will
1295 need to disable comparison in the @file{Makefile}.)
1297 @section Building a cross compiler
1299 We recommend reading the
1300 @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ}
1301 for information about building cross compilers.
1303 When building a cross compiler, it is not generally possible to do a
1304 3-stage bootstrap of the compiler. This makes for an interesting problem
1305 as parts of GCC can only be built with GCC@.
1307 To build a cross compiler, we first recommend building and installing a
1308 native compiler. You can then use the native GCC compiler to build the
1309 cross compiler. The installed native compiler needs to be GCC version
1312 Assuming you have already installed a native copy of GCC and configured
1313 your cross compiler, issue the command @command{make}, which performs the
1318 Build host tools necessary to build the compiler such as texinfo, bison,
1322 Build target tools for use by the compiler such as binutils (bfd,
1323 binutils, gas, gprof, ld, and opcodes)
1324 if they have been individually linked or moved into the top level GCC source
1325 tree before configuring.
1328 Build the compiler (single stage only).
1331 Build runtime libraries using the compiler from the previous step.
1334 Note that if an error occurs in any step the make process will exit.
1336 @section Building in parallel
1338 You can use @samp{make bootstrap MAKE="make -j 2" -j 2}, or just
1339 @samp{make -j 2 bootstrap} for GNU Make 3.79 and above, instead of
1340 @samp{make bootstrap} to build GCC in parallel.
1341 You can also specify a bigger number, and in most cases using a value
1342 greater than the number of processors in your machine will result in
1343 fewer and shorter I/O latency hits, thus improving overall throughput;
1344 this is especially true for slow drives and network filesystems.
1346 @section Building the Ada compiler
1348 In order to build GNAT, the Ada compiler, you need a working GNAT
1349 compiler (GNAT version 3.14 or later, or GCC version 3.1 or later),
1350 since the Ada front end is written in Ada (with some
1351 GNAT-specific extensions), and GNU make.
1353 However, you do not need a full installation of GNAT, just the GNAT
1354 binary @file{gnat1}, a copy of @file{gnatbind}, and a compiler driver
1355 which can deal with Ada input (by invoking the @file{gnat1} binary).
1356 You can specify this compiler driver by setting the @env{ADAC}
1357 environment variable at the configure step. @command{configure} can
1358 detect the driver automatically if it has got a common name such as
1359 @command{gcc} or @command{gnatgcc}. Of course, you still need a working
1360 C compiler (the compiler driver can be different or not).
1361 @command{configure} does not test whether the GNAT installation works
1362 and has a sufficiently recent version; if too old a GNAT version is
1363 installed, the build will fail unless @option{--enable-languages} is
1364 used to disable building the Ada front end.
1366 Additional build tools (such as @command{gnatmake}) or a working GNAT
1367 run-time library installation are usually @emph{not} required. However,
1368 if you want to bootstrap the compiler using a minimal version of GNAT,
1369 you have to issue the following commands before invoking @samp{make
1370 bootstrap} (this assumes that you start with an unmodified and consistent
1371 source distribution):
1374 cd @var{srcdir}/gcc/ada
1375 touch treeprs.ads [es]info.h nmake.ad[bs]
1378 At the moment, the GNAT library and several tools for GNAT are not built
1379 by @samp{make bootstrap}. You have to invoke
1380 @samp{make gnatlib_and_tools} in the @file{@var{objdir}/gcc}
1381 subdirectory before proceeding with the next steps.
1383 For example, you can build a native Ada compiler by issuing the
1384 following commands (assuming @command{make} is GNU make):
1388 @var{srcdir}/configure --enable-languages=c,ada
1389 cd @var{srcdir}/gcc/ada
1390 touch treeprs.ads [es]info.h nmake.ad[bs]
1394 make gnatlib_and_tools
1398 Currently, when compiling the Ada front end, you cannot use the parallel
1399 build feature described in the previous section.
1401 @section Building with profile feedback
1403 It is possible to use profile feedback to optimize the compiler itself. This
1404 should result in a faster compiler binary. Experiments done on x86 using gcc
1405 3.3 showed approximately 7 percent speedup on compiling C programs. To
1406 bootstrap compiler with profile feedback, use @code{make profiledbootstrap}.
1408 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
1409 compiler. This compiler is used to build a @code{stageprofile} compiler
1410 instrumented to collect execution counts of instruction and branch
1411 probabilities. Then runtime libraries are compiled with profile collected.
1412 Finally a @code{stagefeedback} compiler is built using the information collected.
1414 Unlike @samp{make bootstrap} several additional restrictions apply. The
1415 compiler used to build @code{stage1} needs to support a 64-bit integral type.
1416 It is recommended to only use GCC for this. Also parallel make is currently
1417 not supported since collisions in profile collecting may occur.
1424 @uref{./index.html,,Return to the GCC Installation page}
1428 @c ***Testing*****************************************************************
1430 @comment node-name, next, previous, up
1431 @node Testing, Final install, Building, Installing GCC
1435 @chapter Installing GCC: Testing
1438 @cindex Installing GCC: Testing
1441 Before you install GCC, we encourage you to run the testsuites and to
1442 compare your results with results from a similar configuration that have
1443 been submitted to the
1444 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
1445 Some of these archived results are linked from the build status lists
1446 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
1447 reports a successful build runs the testsuites and submits the results.
1448 This step is optional and may require you to download additional software,
1449 but it can give you confidence in your new GCC installation or point out
1450 problems before you install and start using your new GCC.
1452 First, you must have @uref{download.html,,downloaded the testsuites}.
1453 These are part of the full distribution, but if you downloaded the
1454 ``core'' compiler plus any front ends, you must download the testsuites
1457 Second, you must have the testing tools installed. This includes
1458 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu} 1.4.1 or 1.4.3
1459 and later, Tcl, and Expect; the DejaGnu site has links to these.
1461 If the directories where @command{runtest} and @command{expect} were
1462 installed are not in the @env{PATH}, you may need to set the following
1463 environment variables appropriately, as in the following example (which
1464 assumes that DejaGnu has been installed under @file{/usr/local}):
1467 TCL_LIBRARY = /usr/local/share/tcl8.0
1468 DEJAGNULIBS = /usr/local/share/dejagnu
1471 (On systems such as Cygwin, these paths are required to be actual
1472 paths, not mounts or links; presumably this is due to some lack of
1473 portability in the DejaGnu code.)
1476 Finally, you can run the testsuite (which may take a long time):
1478 cd @var{objdir}; make -k check
1481 This will test various components of GCC, such as compiler
1482 front ends and runtime libraries. While running the testsuite, DejaGnu
1483 might emit some harmless messages resembling
1484 @samp{WARNING: Couldn't find the global config file.} or
1485 @samp{WARNING: Couldn't find tool init file} that can be ignored.
1487 @section How can I run the test suite on selected tests?
1489 In order to run sets of tests selectively, there are targets
1490 @samp{make check-gcc} and @samp{make check-g++}
1491 in the @file{gcc} subdirectory of the object directory. You can also
1492 just run @samp{make check} in a subdirectory of the object directory.
1495 A more selective way to just run all @command{gcc} execute tests in the
1499 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
1502 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
1503 the testsuite with filenames matching @samp{9805*}, you would use
1506 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1509 The @file{*.exp} files are located in the testsuite directories of the GCC
1510 source, the most important ones being @file{compile.exp},
1511 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1512 To get a list of the possible @file{*.exp} files, pipe the
1513 output of @samp{make check} into a file and look at the
1514 @samp{Running @dots{} .exp} lines.
1516 @section Passing options and running multiple testsuites
1518 You can pass multiple options to the testsuite using the
1519 @samp{--target_board} option of DejaGNU, either passed as part of
1520 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
1521 work outside the makefiles. For example,
1524 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fno-strength-reduce"
1527 will run the standard @command{g++} testsuites (``unix'' is the target name
1528 for a standard native testsuite situation), passing
1529 @samp{-O3 -fno-strength-reduce} to the compiler on every test, i.e.,
1530 slashes separate options.
1532 You can run the testsuites multiple times using combinations of options
1533 with a syntax similar to the brace expansion of popular shells:
1536 @dots{}"--target_board=arm-sim@{-mhard-float,-msoft-float@}@{-O1,-O2,-O3,@}"
1539 (Note the empty option caused by the trailing comma in the final group.)
1540 The following will run each testsuite eight times using the @samp{arm-sim}
1541 target, as if you had specified all possible combinations yourself:
1544 --target_board=arm-sim/-mhard-float/-O1
1545 --target_board=arm-sim/-mhard-float/-O2
1546 --target_board=arm-sim/-mhard-float/-O3
1547 --target_board=arm-sim/-mhard-float
1548 --target_board=arm-sim/-msoft-float/-O1
1549 --target_board=arm-sim/-msoft-float/-O2
1550 --target_board=arm-sim/-msoft-float/-O3
1551 --target_board=arm-sim/-msoft-float
1554 They can be combined as many times as you wish, in arbitrary ways. This
1558 @dots{}"--target_board=unix/-Wextra@{-O3,-fno-strength-reduce@}@{-fomit-frame-pointer,@}"
1561 will generate four combinations, all involving @samp{-Wextra}.
1563 The disadvantage to this method is that the testsuites are run in serial,
1564 which is a waste on multiprocessor systems. For users with GNU Make and
1565 a shell which performs brace expansion, you can run the testsuites in
1566 parallel by having the shell perform the combinations and @command{make}
1567 do the parallel runs. Instead of using @samp{--target_board}, use a
1568 special makefile target:
1571 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
1577 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
1580 will run three concurrent ``make-gcc'' testsuites, eventually testing all
1581 ten combinations as described above. Note that this is currently only
1582 supported in the @file{gcc} subdirectory. (To see how this works, try
1583 typing @command{echo} before the example given here.)
1586 @section Additional testing for Java Class Libraries
1588 The Java runtime tests can be executed via @samp{make check}
1589 in the @file{@var{target}/libjava/testsuite} directory in
1592 The @uref{http://sources.redhat.com/mauve/,,Mauve Project} provides
1593 a suite of tests for the Java Class Libraries. This suite can be run
1594 as part of libgcj testing by placing the Mauve tree within the libjava
1595 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
1596 specifying the location of that tree when invoking @samp{make}, as in
1597 @samp{make MAUVEDIR=~/mauve check}.
1599 @uref{http://www-124.ibm.com/developerworks/oss/cvs/jikes/~checkout~/jacks/jacks.html,,Jacks}
1600 is a free test suite that tests Java compiler front ends. This suite
1601 can be run as part of libgcj testing by placing the Jacks tree within
1602 the libjava testsuite at @file{libjava/testsuite/libjava.jacks/jacks}.
1604 @section How to interpret test results
1606 The result of running the testsuite are various @file{*.sum} and @file{*.log}
1607 files in the testsuite subdirectories. The @file{*.log} files contain a
1608 detailed log of the compiler invocations and the corresponding
1609 results, the @file{*.sum} files summarize the results. These summaries
1610 contain status codes for all tests:
1614 PASS: the test passed as expected
1616 XPASS: the test unexpectedly passed
1618 FAIL: the test unexpectedly failed
1620 XFAIL: the test failed as expected
1622 UNSUPPORTED: the test is not supported on this platform
1624 ERROR: the testsuite detected an error
1626 WARNING: the testsuite detected a possible problem
1629 It is normal for some tests to report unexpected failures. At the
1630 current time our testing harness does not allow fine grained control
1631 over whether or not a test is expected to fail. We expect to fix this
1632 problem in future releases.
1635 @section Submitting test results
1637 If you want to report the results to the GCC project, use the
1638 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
1641 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
1642 -m gcc-testresults@@gcc.gnu.org |sh
1645 This script uses the @command{Mail} program to send the results, so
1646 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
1647 prepended to the testsuite summary and should contain any special
1648 remarks you have on your results or your build environment. Please
1649 do not edit the testsuite result block or the subject line, as these
1650 messages may be automatically processed.
1657 @uref{./index.html,,Return to the GCC Installation page}
1661 @c ***Final install***********************************************************
1663 @comment node-name, next, previous, up
1664 @node Final install, , Testing, Installing GCC
1666 @ifset finalinstallhtml
1668 @chapter Installing GCC: Final installation
1671 Now that GCC has been built (and optionally tested), you can install it with
1673 cd @var{objdir}; make install
1676 We strongly recommend to install into a target directory where there is
1677 no previous version of GCC present.
1679 That step completes the installation of GCC; user level binaries can
1680 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
1681 you specified with the @option{--prefix} to configure (or
1682 @file{/usr/local} by default). (If you specified @option{--bindir},
1683 that directory will be used instead; otherwise, if you specified
1684 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
1685 Headers for the C++ and Java libraries are installed in
1686 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
1687 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
1688 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
1689 in info format in @file{@var{infodir}} (normally
1690 @file{@var{prefix}/info}).
1692 When installing cross-compilers, GCC's executables
1693 are not only installed into @file{@var{bindir}}, that
1694 is, @file{@var{exec-prefix}/bin}, but additionally into
1695 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
1696 exists. Typically, such @dfn{tooldirs} hold target-specific
1697 binutils, including assembler and linker.
1699 Installation into a temporary staging area or into a @command{chroot}
1700 jail can be achieved with the command
1703 make DESTDIR=@var{path-to-rootdir} install
1706 @noindent where @var{path-to-rootdir} is the absolute path of
1707 a directory relative to which all installation paths will be
1708 interpreted. Note that the directory specified by @code{DESTDIR}
1709 need not exist yet; it will be created if necessary.
1711 There is a subtle point with tooldirs and @code{DESTDIR}:
1712 If you relocate a cross-compiler installation with
1713 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
1714 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
1715 be filled with duplicated GCC executables only if it already exists,
1716 it will not be created otherwise. This is regarded as a feature,
1717 not as a bug, because it gives slightly more control to the packagers
1718 using the @code{DESTDIR} feature.
1720 If you built a released version of GCC using @samp{make bootstrap} then please
1721 quickly review the build status page for your release, available from
1722 @uref{http://gcc.gnu.org/buildstat.html}.
1723 If your system is not listed for the version of GCC that you built,
1725 @email{gcc@@gcc.gnu.org} indicating
1726 that you successfully built and installed GCC.
1727 Include the following information:
1731 Output from running @file{@var{srcdir}/config.guess}. Do not send us
1732 that file itself, just the one-line output from running it.
1735 The output of @samp{gcc -v} for your newly installed gcc.
1736 This tells us which version of GCC you built and the options you passed to
1740 Whether you enabled all languages or a subset of them. If you used a
1741 full distribution then this information is part of the configure
1742 options in the output of @samp{gcc -v}, but if you downloaded the
1743 ``core'' compiler plus additional front ends then it isn't apparent
1744 which ones you built unless you tell us about it.
1747 If the build was for GNU/Linux, also include:
1750 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
1751 this information should be available from @file{/etc/issue}.
1754 The version of the Linux kernel, available from @samp{uname --version}
1758 The version of glibc you used; for RPM-based systems like Red Hat,
1759 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
1760 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
1762 For other systems, you can include similar information if you think it is
1766 Any other information that you think would be useful to people building
1767 GCC on the same configuration. The new entry in the build status list
1768 will include a link to the archived copy of your message.
1771 We'd also like to know if the
1773 @ref{Specific, host/target specific installation notes}
1776 @uref{specific.html,,host/target specific installation notes}
1778 didn't include your host/target information or if that information is
1779 incomplete or out of date. Send a note to
1780 @email{gcc@@gcc.gnu.org} telling us how the information should be changed.
1782 If you find a bug, please report it following our
1783 @uref{../bugs.html,,bug reporting guidelines}.
1785 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
1786 dvi}. You will need to have @command{texi2dvi} (version at least 4.2)
1787 and @TeX{} installed. This creates a number of @file{.dvi} files in
1788 subdirectories of @file{@var{objdir}}; these may be converted for
1789 printing with programs such as @command{dvips}. You can also
1790 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
1791 Free Software Foundation}, though such manuals may not be for the most
1792 recent version of GCC@.
1799 @uref{./index.html,,Return to the GCC Installation page}
1803 @c ***Binaries****************************************************************
1805 @comment node-name, next, previous, up
1806 @node Binaries, Specific, Installing GCC, Top
1810 @chapter Installing GCC: Binaries
1813 @cindex Installing GCC: Binaries
1815 We are often asked about pre-compiled versions of GCC@. While we cannot
1816 provide these for all platforms, below you'll find links to binaries for
1817 various platforms where creating them by yourself is not easy due to various
1820 Please note that we did not create these binaries, nor do we
1821 support them. If you have any problems installing them, please
1822 contact their makers.
1829 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
1832 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX}.
1836 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
1839 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
1840 Development Tools for the Renesas H8/300[HS] Series}.
1846 @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center};
1849 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
1853 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
1854 Development Tools for the Motorola 68HC11/68HC12}.
1857 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
1858 OpenServer/Unixware}.
1861 Sinix/Reliant Unix---@uref{ftp://ftp.fujitsu-siemens.com/pub/pd/gnu/gcc/,,Siemens}.
1864 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
1867 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
1873 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
1875 The @uref{http://www.mingw.org/,,MinGW} project.
1879 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
1880 Written Word} offers binaries for
1883 Digital UNIX 4.0D and 5.1,
1885 HP-UX 10.20, 11.00, and 11.11, and
1886 Solaris/SPARC 2.5.1, 2.6, 2.7, 8, and 9,
1889 In addition to those specific offerings, you can get a binary
1890 distribution CD-ROM from the
1891 @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}.
1892 It contains binaries for a number of platforms, and
1893 includes not only GCC, but other stuff as well. The current CD does
1894 not contain the latest version of GCC, but it should allow
1895 bootstrapping the compiler. An updated version of that disk is in the
1903 @uref{./index.html,,Return to the GCC Installation page}
1907 @c ***Specific****************************************************************
1909 @comment node-name, next, previous, up
1910 @node Specific, Old, Binaries, Top
1914 @chapter Host/target specific installation notes for GCC
1917 @cindex Specific installation notes
1918 @cindex Target specific installation
1919 @cindex Host specific installation
1920 @cindex Target specific installation notes
1922 Please read this document carefully @emph{before} installing the
1923 GNU Compiler Collection on your machine.
1928 @uref{#alpha*-*-*,,alpha*-*-*}
1930 @uref{#alpha*-dec-osf*,,alpha*-dec-osf*}
1932 @uref{#alphaev5-cray-unicosmk*,,alphaev5-cray-unicosmk*}
1934 @uref{#arc-*-elf,,arc-*-elf}
1936 @uref{#arm-*-elf,,arm-*-elf}
1937 @uref{#arm-*-coff,,arm-*-coff}
1938 @uref{#arm-*-aout,,arm-*-aout}
1940 @uref{#xscale-*-*,,xscale-*-*}
1948 @uref{#dsp16xx,,dsp16xx}
1950 @uref{#*-*-freebsd*,,*-*-freebsd*}
1952 @uref{#h8300-hms,,h8300-hms}
1954 @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*}
1956 @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10}
1958 @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11}
1960 @uref{#i370-*-*,,i370-*-*}
1962 @uref{#*-*-linux-gnu,,*-*-linux-gnu}
1964 @uref{#ix86-*-linux*aout,,i?86-*-linux*aout}
1966 @uref{#ix86-*-linux*,,i?86-*-linux*}
1968 @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*}
1970 @uref{#ix86-*-udk,,i?86-*-udk}
1972 @uref{#ix86-*-esix,,i?86-*-esix}
1974 @uref{#ia64-*-linux,,ia64-*-linux}
1976 @uref{#ia64-*-hpux*,,ia64-*-hpux*}
1978 @uref{#*-ibm-aix*,,*-ibm-aix*}
1980 @uref{#ip2k-*-elf,,ip2k-*-elf}
1982 @uref{#iq2000-*-elf,,iq2000-*-elf}
1984 @uref{#m32r-*-elf,,m32r-*-elf}
1986 @uref{#m6811-elf,,m6811-elf}
1988 @uref{#m6812-elf,,m6812-elf}
1990 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
1992 @uref{#mips-*-*,,mips-*-*}
1994 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
1996 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
1998 @uref{#powerpc*-*-*,,powerpc*-*-*, powerpc-*-sysv4}
2000 @uref{#powerpc-*-darwin*,,powerpc-*-darwin*}
2002 @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4}
2004 @uref{#powerpc-*-linux-gnu*,,powerpc-*-linux-gnu*}
2006 @uref{#powerpc-*-netbsd*,,powerpc-*-netbsd*}
2008 @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix}
2010 @uref{#powerpc-*-eabisim,,powerpc-*-eabisim}
2012 @uref{#powerpc-*-eabi,,powerpc-*-eabi}
2014 @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
2016 @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim}
2018 @uref{#powerpcle-*-eabi,,powerpcle-*-eabi}
2020 @uref{#s390-*-linux*,,s390-*-linux*}
2022 @uref{#s390x-*-linux*,,s390x-*-linux*}
2024 @uref{#s390x-ibm-tpf*,,s390x-ibm-tpf*}
2026 @uref{#*-*-solaris2*,,*-*-solaris2*}
2028 @uref{#sparc-sun-solaris2*,,sparc-sun-solaris2*}
2030 @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7}
2032 @uref{#sparc-*-linux*,,sparc-*-linux*}
2034 @uref{#sparc64-*-solaris2*,,sparc64-*-solaris2*}
2036 @uref{#sparcv9-*-solaris2*,,sparcv9-*-solaris2*}
2038 @uref{#*-*-sysv*,,*-*-sysv*}
2040 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
2042 @uref{#*-*-vxworks*,,*-*-vxworks*}
2044 @uref{#xtensa-*-elf,,xtensa-*-elf}
2046 @uref{#xtensa-*-linux*,,xtensa-*-linux*}
2048 @uref{#windows,,Microsoft Windows}
2052 @uref{#older,,Older systems}
2057 @uref{#elf_targets,,all ELF targets} (SVR4, Solaris 2, etc.)
2063 <!-- -------- host/target specific issues start here ---------------- -->
2066 @heading @anchor{alpha*-*-*}alpha*-*-*
2068 This section contains general configuration information for all
2069 alpha-based platforms using ELF (in particular, ignore this section for
2070 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2071 section, please read all other sections that match your target.
2073 We require binutils 2.11.2 or newer.
2074 Previous binutils releases had a number of problems with DWARF 2
2075 debugging information, not the least of which is incorrect linking of
2081 @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf*
2082 Systems using processors that implement the DEC Alpha architecture and
2083 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2084 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2086 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2087 supported. (These are the versions which identify themselves as DEC
2090 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2091 may be fixed by configuring with @option{--with-gc=simple},
2092 reconfiguring Kernel Virtual Memory and Swap parameters
2093 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2094 or applying the patch in
2095 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2097 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2098 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2099 we need to use the old assembler, invoked via the barely documented
2100 @option{-oldas} option. To bootstrap GCC, you either need to use the
2104 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2107 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2110 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2113 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2114 are supported on Tru64 UNIX, so you must not configure GCC with
2115 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2117 GCC writes a @samp{.verstamp} directive to the assembler output file
2118 unless it is built as a cross-compiler. It gets the version to use from
2119 the system header file @file{/usr/include/stamp.h}. If you install a
2120 new version of DEC Unix, you should rebuild GCC to pick up the new version
2123 Note that since the Alpha is a 64-bit architecture, cross-compilers from
2124 32-bit machines will not generate code as efficient as that generated
2125 when the compiler is running on a 64-bit machine because many
2126 optimizations that depend on being able to represent a word on the
2127 target in an integral value on the host cannot be performed. Building
2128 cross-compilers on the Alpha for 32-bit machines has only been tested in
2129 a few cases and may not work properly.
2131 @samp{make compare} may fail on old versions of DEC Unix unless you add
2132 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2133 assembler input file is stored in the object file, and that makes
2134 comparison fail if it differs between the @code{stage1} and
2135 @code{stage2} compilations. The option @option{-save-temps} forces a
2136 fixed name to be used for the assembler input file, instead of a
2137 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2138 unless the comparisons fail without that option. If you add
2139 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2140 @samp{.s} files after each series of compilations.
2142 GCC now supports both the native (ECOFF) debugging format used by DBX
2143 and GDB and an encapsulated STABS format for use only with GDB@. See the
2144 discussion of the @option{--with-stabs} option of @file{configure} above
2145 for more information on these formats and how to select them.
2147 There is a bug in DEC's assembler that produces incorrect line numbers
2148 for ECOFF format when the @samp{.align} directive is used. To work
2149 around this problem, GCC will not emit such alignment directives
2150 while writing ECOFF format debugging information even if optimization is
2151 being performed. Unfortunately, this has the very undesirable
2152 side-effect that code addresses when @option{-O} is specified are
2153 different depending on whether or not @option{-g} is also specified.
2155 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2156 DBX@. DEC is now aware of this problem with the assembler and hopes to
2157 provide a fix shortly.
2162 @heading @anchor{alphaev5-cray-unicosmk*}alphaev5-cray-unicosmk*
2163 Cray T3E systems running Unicos/Mk.
2165 This port is incomplete and has many known bugs. We hope to improve the
2166 support for this target soon. Currently, only the C front end is supported,
2167 and it is not possible to build parallel applications. Cray modules are not
2168 supported; in particular, Craylibs are assumed to be in
2169 @file{/opt/ctl/craylibs/craylibs}.
2171 You absolutely @strong{must} use GNU make on this platform. Also, you
2172 need to tell GCC where to find the assembler and the linker. The
2173 simplest way to do so is by providing @option{--with-as} and
2174 @option{--with-ld} to @file{configure}, e.g.@:
2177 configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld \
2178 --enable-languages=c
2181 The comparison test during @samp{make bootstrap} fails on Unicos/Mk
2182 because the assembler inserts timestamps into object files. You should
2183 be able to work around this by doing @samp{make all} after getting this
2189 @heading @anchor{arc-*-elf}arc-*-elf
2190 Argonaut ARC processor.
2191 This configuration is intended for embedded systems.
2196 @heading @anchor{arm-*-elf}arm-*-elf
2197 @heading @anchor{xscale-*-*}xscale-*-*
2198 ARM-family processors. Subtargets that use the ELF object format
2199 require GNU binutils 2.13 or newer. Such subtargets include:
2200 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux},
2201 @code{arm-*-rtems} and @code{arm-*-kaos}.
2206 @heading @anchor{arm-*-coff}arm-*-coff
2207 ARM-family processors. Note that there are two different varieties
2208 of PE format subtarget supported: @code{arm-wince-pe} and
2209 @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}.
2214 @heading @anchor{arm-*-aout}arm-*-aout
2215 ARM-family processors. These targets support the AOUT file format:
2216 @code{arm-*-aout}, @code{arm-*-netbsd}.
2221 @heading @anchor{avr}avr
2223 ATMEL AVR-family micro controllers. These are used in embedded
2224 applications. There are no standard Unix configurations.
2226 @xref{AVR Options,, AVR Options, gcc, Using and Porting the GNU Compiler
2230 See ``AVR Options'' in the main manual
2232 for the list of supported MCU types.
2234 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2236 Further installation notes and other useful information about AVR tools
2237 can also be obtained from:
2241 @uref{http://www.openavr.org,,http://www.openavr.org}
2243 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2245 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2248 We @emph{strongly} recommend using binutils 2.13 or newer.
2250 The following error:
2252 Error: register required
2255 indicates that you should upgrade to a newer version of the binutils.
2260 @heading @anchor{c4x}c4x
2262 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
2263 Processors. These are used in embedded applications. There are no
2264 standard Unix configurations.
2266 @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using and
2267 Porting the GNU Compiler Collection (GCC)},
2270 See ``TMS320C3x/C4x Options'' in the main manual
2272 for the list of supported MCU types.
2274 GCC can be configured as a cross compiler for both the C3x and C4x
2275 architectures on the same system. Use @samp{configure --target=c4x
2276 --enable-languages="c,c++"} to configure.
2279 Further installation notes and other useful information about C4x tools
2280 can also be obtained from:
2284 @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/}
2290 @heading @anchor{cris}CRIS
2292 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2293 series. These are used in embedded applications.
2296 @xref{CRIS Options,, CRIS Options, gcc, Using and Porting the GNU Compiler
2300 See ``CRIS Options'' in the main manual
2302 for a list of CRIS-specific options.
2304 There are a few different CRIS targets:
2306 @item cris-axis-aout
2307 Old target. Includes a multilib for the @samp{elinux} a.out-based
2308 target. No multilibs for newer architecture variants.
2310 Mainly for monolithic embedded systems. Includes a multilib for the
2311 @samp{v10} core used in @samp{ETRAX 100 LX}.
2312 @item cris-axis-linux-gnu
2313 A GNU/Linux port for the CRIS architecture, currently targeting
2314 @samp{ETRAX 100 LX} by default.
2317 For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11
2318 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2320 Pre-packaged tools can be obtained from
2321 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2322 information about this platform is available at
2323 @uref{http://developer.axis.com/}.
2328 @heading @anchor{dos}DOS
2330 Please have a look at our @uref{binaries.html,,binaries page}.
2332 You cannot install GCC by itself on MSDOS; it will not compile under
2333 any MSDOS compiler except itself. You need to get the complete
2334 compilation package DJGPP, which includes binaries as well as sources,
2335 and includes all the necessary compilation tools and libraries.
2340 @heading @anchor{dsp16xx}dsp16xx
2341 A port to the AT&T DSP1610 family of processors.
2346 @heading @anchor{*-*-freebsd*}*-*-freebsd*
2348 The version of binutils installed in @file{/usr/bin} is known to work unless
2349 otherwise specified in any per-architecture notes. However, binutils
2350 2.12.1 or greater is known to improve overall testsuite results.
2352 Support for FreeBSD 1 was discontinued in GCC 3.2.
2354 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2355 configuration support and files as shipped with GCC 2.95 are still in
2356 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2357 it is unknown which version of binutils was used (it is assumed that it
2358 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2360 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2361 default for all CPU architectures. It had been the default on
2362 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2363 of @option{-g}, if you really want the old debugging format. There are
2364 no known issues with mixing object files and libraries with different
2365 debugging formats. Otherwise, this release of GCC should now match more
2366 of the configuration used in the stock FreeBSD configuration of GCC. In
2367 particular, @option{--enable-threads} is now configured by default.
2368 However, as a general user, do not attempt to replace the system
2369 compiler with this release. Known to bootstrap and check with good
2370 results on FreeBSD 4.8-STABLE and 5-CURRENT@. In the past, known to
2371 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
2372 4.3, 4.4, 4.5-STABLE@.
2374 In principle, @option{--enable-threads} is now compatible with
2375 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
2376 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
2378 library may be incorrectly built (symbols are missing at link time).
2379 There is a rare timing-based startup hang (probably involves an
2380 assumption about the thread library). Multi-threaded boehm-gc (required for
2381 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
2382 4.5-RELEASE@. Other CPU architectures
2383 supported by FreeBSD will require additional configuration tuning in, at
2384 the very least, both boehm-gc and libffi.
2386 Shared @file{libgcc_s.so} is now built and installed by default.
2391 @heading @anchor{h8300-hms}h8300-hms
2392 Renesas H8/300 series of processors.
2394 Please have a look at our @uref{binaries.html,,binaries page}.
2396 The calling convention and structure layout has changed in release 2.6.
2397 All code must be recompiled. The calling convention now passes the
2398 first three arguments in function calls in registers. Structures are no
2399 longer a multiple of 2 bytes.
2404 @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux*
2405 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
2407 We @emph{highly} recommend using gas/binutils 2.8 or newer on all hppa
2408 platforms; you may encounter a variety of problems when using the HP
2411 Specifically, @option{-g} does not work on HP-UX (since that system
2412 uses a peculiar debugging format which GCC does not know about), unless you
2413 use GAS and GDB and configure GCC with the
2414 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
2415 @option{--with-as=@dots{}} options.
2417 If you wish to use the pa-risc 2.0 architecture support with a 32-bit
2418 runtime, you must use either the HP assembler, gas/binutils 2.11 or newer,
2420 @uref{ftp://sources.redhat.com/pub/binutils/snapshots,,snapshot of gas}.
2422 There are two default scheduling models for instructions. These are
2423 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
2424 architecture specified for the target machine when configuring.
2425 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
2426 the target is a @samp{hppa1*} machine.
2428 The PROCESSOR_8000 model is not well suited to older processors. Thus,
2429 it is important to completely specify the machine architecture when
2430 configuring if you want a model other than PROCESSOR_8000. The macro
2431 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
2432 default scheduling model is desired.
2434 More specific information to @samp{hppa*-hp-hpux*} targets follows.
2439 @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10
2441 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
2442 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
2448 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
2452 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
2456 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
2459 The HP assembler on these systems has some problems. Most notably the
2460 assembler inserts timestamps into each object file it creates, causing
2461 the 3-stage comparison test to fail during a @samp{make bootstrap}.
2462 You should be able to continue by saying @samp{make all} after getting
2463 the failure from @samp{make bootstrap}.
2469 @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11
2471 GCC 3.0 and up support HP-UX 11. On 64-bit capable systems, there
2472 are two distinct ports. The @samp{hppa2.0w-hp-hpux11*} port generates
2473 code for the 32-bit pa-risc runtime architecture. It uses the HP
2474 linker. The @samp{hppa64-hp-hpux11*} port generates 64-bit code for the
2475 pa-risc 2.0 architecture. The script config.guess now selects the port
2476 type based on the type compiler detected during configuration. You must
2477 set your @env{PATH} or define @env{CC} so that configure finds an appropriate
2478 compiler for the initial bootstrap. Different prefixes must be used if
2479 both ports are to be installed on the same system.
2481 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
2482 with the @option{--with-ld=@dots{}} option. We support both the HP
2483 and GNU linkers for this target. The two linkers require different
2484 link commands. Thus, it's not possible to switch linkers during a
2485 GCC build. This has been been reported to occur in a unified build
2486 of binutils and GCC.
2488 GCC 2.95.x is not supported under HP-UX 11 and cannot be used to
2489 compile GCC 3.0 and up. Refer to @uref{binaries.html,,binaries} for
2490 information about obtaining precompiled GCC binaries for HP-UX.
2492 You must use GNU binutils 2.11 or above with the 32-bit port. Thread
2493 support is not currently implemented, so @option{--enable-threads} does
2497 @item @uref{http://gcc.gnu.org/ml/gcc-prs/2002-01/msg00551.html}
2498 @item @uref{http://gcc.gnu.org/ml/gcc-bugs/2002-01/msg00663.html}
2501 GCC 3.3 and later support weak symbols on the 32-bit port using SOM
2502 secondary definition symbols. This feature is not enabled for earlier
2503 versions of HP-UX since there have been bugs in the linker support for
2504 secondary symbols. The HP linker patches @code{PHSS_26559} and
2505 @code{PHSS_24304} for HP-UX 11.00 and 11.11, respectively, correct the
2506 problem of linker core dumps creating C++ libraries. Earlier patches
2507 may work but they have not been tested.
2509 GCC 3.3 nows uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capability
2510 to run initializers and finalizers on the 64-bit port. The feature
2511 requires CVS binutils as of January 2, 2003, or a subsequent release
2512 to correct a problem arising from HP's non-standard use of the .init
2513 and .fini sections. The 32-bit port uses the linker @option{+init}
2514 and @option{+fini} options. As with the support for secondary symbols,
2515 there have been bugs in the order in which these options are executed
2516 by the HP linker. So, again a recent linker patch is recommended.
2518 The HP assembler has many limitations and is not recommended for either
2519 the 32 or 64-bit ports. For example, it does not support weak symbols
2520 or alias definitions. As a result, explicit template instantiations
2521 are required when using C++. This will make it difficult if not
2522 impossible to build many C++ applications. You also can't generate
2523 debugging information when using the HP assembler with GCC.
2525 There are a number of issues to consider in selecting which linker to
2526 use with the 64-bit port. The GNU 64-bit linker can only create dynamic
2527 binaries. The @option{-static} option causes linking with archive
2528 libraries but doesn't produce a truly static binary. Dynamic binaries
2529 still require final binding by the dynamic loader to resolve a set of
2530 dynamic-loader-defined symbols. The default behavior of the HP linker
2531 is the same as the GNU linker. However, it can generate true 64-bit
2532 static binaries using the @option{+compat} option.
2534 The HP 64-bit linker doesn't support linkonce semantics. As a
2535 result, C++ programs have many more sections than they should.
2537 The GNU 64-bit linker has some issues with shared library support
2538 and exceptions. As a result, we only support libgcc in archive
2539 format. For similar reasons, dwarf2 unwind and exception support
2540 are disabled. The GNU linker also has problems creating binaries
2541 with @option{-static}. It doesn't provide stubs for internal
2542 calls to global functions in shared libraries, so these calls
2543 can't be overloaded.
2545 There are several possible approaches to building the distribution.
2546 Binutils can be built first using the HP tools. Then, the GCC
2547 distribution can be built. The second approach is to build GCC
2548 first using the HP tools, then build binutils, then rebuild GCC.
2549 There have been problems with various binary distributions, so
2550 it is best not to start from a binary distribution.
2552 Starting with GCC 3.4 an ISO C compiler is required to bootstrap.
2553 The bundled compiler supports only traditional C; you will need
2554 either HP's unbundled compiler, or a binary distribution of GCC@.
2556 This port still is undergoing significant development.
2561 @heading @anchor{i370-*-*}i370-*-*
2562 This port is very preliminary and has many known bugs. We hope to
2563 have a higher-quality port for this machine soon.
2568 @heading @anchor{*-*-linux-gnu}*-*-linux-gnu
2570 Versions of libstdc++-v3 starting with 3.2.1 require bugfixes present
2571 in glibc 2.2.5 and later. More information is available in the
2572 libstdc++-v3 documentation.
2574 If you use glibc 2.2 (or 2.1.9x), GCC 2.95.2 won't install
2575 out-of-the-box. You'll get compile errors while building @samp{libstdc++}.
2576 The patch @uref{glibc-2.2.patch,,glibc-2.2.patch}, that is to be
2577 applied in the GCC source tree, fixes the compatibility problems.
2579 Currently Glibc 2.2.3 (and older releases) and GCC 3.0 are out of sync
2580 since the latest exception handling changes for GCC@. Compiling glibc
2581 with GCC 3.0 will give a binary incompatible glibc and therefore cause
2582 lots of problems and might make your system completely unusable. This
2583 will definitely need fixes in glibc but might also need fixes in GCC@. We
2584 strongly advise to wait for glibc 2.2.4 and to read the release notes of
2585 glibc 2.2.4 whether patches for GCC 3.0 are needed. You can use glibc
2586 2.2.3 with GCC 3.0, just do not try to recompile it.
2591 @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout
2592 Use this configuration to generate @file{a.out} binaries on Linux-based
2593 GNU systems. This configuration is being superseded.
2598 @heading @anchor{ix86-*-linux*}i?86-*-linux*
2600 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
2601 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
2603 If you receive Signal 11 errors when building on GNU/Linux, then it is
2604 possible you have a hardware problem. Further information on this can be
2605 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
2610 @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5*
2611 Use this for the SCO OpenServer Release 5 family of operating systems.
2613 Unlike earlier versions of GCC, the ability to generate COFF with this
2614 target is no longer provided.
2616 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
2617 the system debugger to be used. That support was too burdensome to
2618 maintain. GCC now emits only DWARF 2 for this target. This means you
2619 may use either the UDK debugger or GDB to debug programs built by this
2622 GCC is now only supported on releases 5.0.4 and later, and requires that
2623 you install Support Level Supplement OSS646B or later, and Support Level
2624 Supplement OSS631C or later. If you are using release 5.0.7 of
2625 OpenServer, you must have at least the first maintenance pack installed
2626 (this includes the relevant portions of OSS646). OSS646, also known as
2627 the "Execution Environment Update", provides updated link editors and
2628 assemblers, as well as updated standard C and math libraries. The C
2629 startup modules are also updated to support the System V gABI draft, and
2630 GCC relies on that behavior. OSS631 provides a collection of commonly
2631 used open source libraries, some of which GCC depends on (such as GNU
2632 gettext and zlib). SCO OpenServer Release 5.0.7 has all of this built
2633 in by default, but OSS631C and later also apply to that release. Please
2635 @uref{ftp://ftp.sco.com/pub/openserver5,,ftp://ftp.sco.com/pub/openserver5}
2636 for the latest versions of these (and other potentially useful)
2639 Although there is support for using the native assembler, it is
2640 recommended that you configure GCC to use the GNU assembler. You do
2641 this by using the flags
2642 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You should
2643 use a modern version of GNU binutils. Version 2.13.2.1 was used for all
2644 testing. In general, only the @option{--with-gnu-as} option is tested.
2645 A modern bintuils (as well as a plethora of other development related
2646 GNU utilities) can be found in Support Level Supplement OSS658A, the
2647 "GNU Development Tools" package. See the SCO web and ftp sites for details.
2648 That package also contains the currently "officially supported" version of
2649 GCC, version 2.95.3. It is useful for bootstrapping this version.
2654 @heading @anchor{ix86-*-udk}i?86-*-udk
2656 This target emulates the SCO Universal Development Kit and requires that
2657 package be installed. (If it is installed, you will have a
2658 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
2659 @samp{i?86-*-unixware7*} target
2660 but is meant to be used when hosting on a system where UDK isn't the
2661 default compiler such as OpenServer 5 or Unixware 2. This target will
2662 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
2663 with the same warnings and caveats as the SCO UDK@.
2665 This target is a little tricky to build because we have to distinguish
2666 it from the native tools (so it gets headers, startups, and libraries
2667 from the right place) while making the tools not think we're actually
2668 building a cross compiler. The easiest way to do this is with a configure
2672 CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure \
2673 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-
2676 @emph{You should substitute @samp{i686} in the above command with the appropriate
2677 processor for your host.}
2679 After the usual @samp{make bootstrap} and
2680 @samp{make install}, you can then access the UDK-targeted GCC
2681 tools by adding @command{udk-} before the commonly known name. For
2682 example, to invoke the C compiler, you would use @command{udk-gcc}.
2683 They will coexist peacefully with any native-target GCC tools you may
2690 @heading @anchor{ia64-*-linux}ia64-*-linux
2691 IA-64 processor (also known as IPF, or Itanium Processor Family)
2694 None of the following versions of GCC has an ABI that is compatible
2695 with any of the other versions in this list, with the exception that
2696 Red Hat 2.96 and Trillian 000171 are compatible with each other:
2697 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
2698 This primarily affects C++ programs and programs that create shared libraries.
2699 GCC 3.1 or later is recommended for compiling linux, the kernel.
2700 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
2701 more major ABI changes are expected.
2706 @heading @anchor{ia64-*-hpux*}ia64-*-hpux*
2707 Building GCC on this target requires the GNU Assembler. The bundled HP
2708 assembler will not work. To prevent GCC from using the wrong assembler,
2709 the option @option{--with-gnu-as} may be necessary.
2711 The GCC libunwind library has not been ported to HPUX. This means that for
2712 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
2713 is required to build GCC. For GCC 3.3 and later, this is the default.
2717 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
2719 @heading @anchor{*-ibm-aix*}*-ibm-aix*
2720 Support for AIX version 3 and older was discontinued in GCC 3.4.
2722 AIX Make frequently has problems with GCC makefiles. GNU Make 3.79.1 or
2723 newer is recommended to build on this platform.
2725 Errors involving @code{alloca} when building GCC generally are due
2726 to an incorrect definition of @code{CC} in the Makefile or mixing files
2727 compiled with the native C compiler and GCC@. During the stage1 phase of
2728 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
2729 (not @command{xlc}). Once @command{configure} has been informed of
2730 @command{xlc}, one needs to use @samp{make distclean} to remove the
2731 configure cache files and ensure that @env{CC} environment variable
2732 does not provide a definition that will confuse @command{configure}.
2733 If this error occurs during stage2 or later, then the problem most likely
2734 is the version of Make (see above).
2736 The native @command{as} and @command{ld} are recommended for bootstrapping
2737 on AIX 4 and required for bootstrapping on AIX 5L. The GNU Assembler
2738 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
2739 utilize weak symbol functionality although it is not supported. The GNU
2740 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC.
2741 The native AIX tools do interoperate with GCC@.
2743 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
2744 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1).
2746 @samp{libstdc++} in GCC 3.2 increments the major version number of the
2747 shared object and GCC installation places the @file{libstdc++.a}
2748 shared library in a common location which will overwrite the GCC 3.1
2749 version of the shared library. Applications either need to be
2750 re-linked against the new shared library or the GCC 3.1 version of the
2751 @samp{libstdc++} shared object needs to be available to the AIX
2752 runtime loader. The GCC 3.1 @samp{libstdc++.so.4} shared object can
2753 be installed for runtime dynamic loading using the following steps to
2754 set the @samp{F_LOADONLY} flag in the shared object for @emph{each}
2755 multilib @file{libstdc++.a} installed:
2757 Extract the shared object from each the GCC 3.1 @file{libstdc++.a}
2760 % ar -x libstdc++.a libstdc++.so.4
2763 Enable the @samp{F_LOADONLY} flag so that the shared object will be
2764 available for runtime dynamic loading, but not linking:
2766 % strip -e libstdc++.so.4
2769 Archive the runtime-only shared object in the GCC 3.2
2770 @file{libstdc++.a} archive:
2772 % ar -q libstdc++.a libstdc++.so.4
2775 Linking executables and shared libraries may produce warnings of
2776 duplicate symbols. The assembly files generated by GCC for AIX always
2777 have included multiple symbol definitions for certain global variable
2778 and function declarations in the original program. The warnings should
2779 not prevent the linker from producing a correct library or runnable
2782 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
2783 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
2784 to parse archive libraries did not handle the new format correctly.
2785 These routines are used by GCC and result in error messages during
2786 linking such as ``not a COFF file''. The version of the routines shipped
2787 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
2788 option of the archive command may be used to create archives of 32-bit
2789 objects using the original ``small format''. A correct version of the
2790 routines is shipped with AIX 4.3.2 and above.
2792 Some versions of the AIX binder (linker) can fail with a relocation
2793 overflow severe error when the @option{-bbigtoc} option is used to link
2794 GCC-produced object files into an executable that overflows the TOC@. A fix
2795 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
2796 available from IBM Customer Support and from its
2797 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2798 website as PTF U455193.
2800 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
2801 with a segmentation fault when invoked by any version of GCC@. A fix for
2802 APAR IX87327 is available from IBM Customer Support and from its
2803 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2804 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
2806 The initial assembler shipped with AIX 4.3.0 generates incorrect object
2807 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
2808 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
2809 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2810 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
2812 AIX provides National Language Support (NLS)@. Compilers and assemblers
2813 use NLS to support locale-specific representations of various data
2814 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
2815 separating decimal fractions). There have been problems reported where
2816 GCC does not produce the same floating-point formats that the assembler
2817 expects. If one encounters this problem, set the @env{LANG}
2818 environment variable to @samp{C} or @samp{En_US}.
2820 By default, GCC for AIX 4.1 and above produces code that can be used on
2821 both Power or PowerPC processors.
2823 A default can be specified with the @option{-mcpu=@var{cpu_type}}
2824 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
2829 @heading @anchor{ip2k-*-elf}ip2k-*-elf
2830 Ubicom IP2022 micro controller.
2831 This configuration is intended for embedded systems.
2832 There are no standard Unix configurations.
2834 Use @samp{configure --target=ip2k-elf --enable-languages=c} to configure GCC@.
2839 @heading @anchor{iq2000-*-elf}iq2000-*-elf
2840 Vitesse IQ2000 processors. These are used in embedded
2841 applications. There are no standard Unix configurations.
2846 @heading @anchor{m32r-*-elf}m32r-*-elf
2847 Renesas M32R processor.
2848 This configuration is intended for embedded systems.
2853 @heading @anchor{m6811-elf}m6811-elf
2854 Motorola 68HC11 family micro controllers. These are used in embedded
2855 applications. There are no standard Unix configurations.
2860 @heading @anchor{m6812-elf}m6812-elf
2861 Motorola 68HC12 family micro controllers. These are used in embedded
2862 applications. There are no standard Unix configurations.
2867 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
2868 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
2869 the assembler that prevents compilation of GCC@. This
2870 bug manifests itself during the first stage of compilation, while
2871 building @file{libgcc2.a}:
2875 cc1: warning: `-g' option not supported on this version of GCC
2876 cc1: warning: `-g1' option not supported on this version of GCC
2877 ./xgcc: Internal compiler error: program as got fatal signal 11
2880 A patched version of the assembler is available as the file
2881 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
2882 have HP software support, the patch can also be obtained directly from
2883 HP, as described in the following note:
2886 This is the patched assembler, to patch SR#1653-010439, where the
2887 assembler aborts on floating point constants.
2889 The bug is not really in the assembler, but in the shared library
2890 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
2891 SR#4701-078451. Anyway, the attached assembler uses the archive
2892 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
2895 This patch is also known as PHCO_4484.
2897 In addition gdb does not understand that native HP-UX format, so
2898 you must use gas if you wish to use gdb.
2900 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
2901 @command{fixproto} shell script triggers a bug in the system shell. If you
2902 encounter this problem, upgrade your operating system or use BASH (the
2903 GNU shell) to run @command{fixproto}. This bug will cause the fixproto
2904 program to report an error of the form:
2907 ./fixproto: sh internal 1K buffer overflow
2910 To fix this, you can also change the first line of the fixproto script
2920 @heading @anchor{mips-*-*}mips-*-*
2921 If on a MIPS system you get an error message saying ``does not have gp
2922 sections for all it's [sic] sectons [sic]'', don't worry about it. This
2923 happens whenever you use GAS with the MIPS linker, but there is not
2924 really anything wrong, and it is okay to use the output file. You can
2925 stop such warnings by installing the GNU linker.
2927 It would be nice to extend GAS to produce the gp tables, but they are
2928 optional, and there should not be a warning about their absence.
2930 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
2931 and later. A patch went in just after the GCC 3.3 release to
2932 make @samp{mips*-*-*} use the generic implementation instead. You can also
2933 configure for @samp{mipsel-elf} as a workaround. The
2934 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
2935 work on this is expected in future releases.
2940 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
2942 This configuration has considerable problems, which will be fixed in a
2945 In order to compile GCC on an SGI running IRIX 5, the ``compiler_dev.hdr''
2946 subsystem must be installed from the IDO CD-ROM supplied by Silicon
2947 Graphics. It is also available for download from
2948 @uref{http://www.sgi.com/developers/devtools/apis/ido.html,,http://www.sgi.com/developers/devtools/apis/ido.html}.
2950 @samp{make compare} may fail on version 5 of IRIX unless you add
2951 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2952 assembler input file is stored in the object file, and that makes
2953 comparison fail if it differs between the @code{stage1} and
2954 @code{stage2} compilations. The option @option{-save-temps} forces a
2955 fixed name to be used for the assembler input file, instead of a
2956 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2957 unless the comparisons fail without that option. If you do you
2958 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2959 @samp{.s} files after each series of compilations.
2961 If you use the MIPS C compiler to bootstrap, it may be necessary
2962 to increase its table size for switch statements with the
2963 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2964 optimization option, you also need to use @option{-Olimit 3000}.
2966 To enable debugging under IRIX 5, you must use GNU @command{as} 2.11.2
2968 and use the @option{--with-gnu-as} configure option when configuring GCC.
2969 GNU @command{as} is distributed as part of the binutils package.
2970 When using release 2.11.2, you need to apply a patch
2971 @uref{http://sources.redhat.com/ml/binutils/2001-07/msg00352.html,,http://sources.redhat.com/ml/binutils/2001-07/msg00352.html}
2972 which will be included in the next release of binutils.
2974 When building GCC, the build process loops rebuilding @command{cc1} over
2975 and over again. This happens on @samp{mips-sgi-irix5.2}, and possibly
2976 other platforms. It has been reported that this is a known bug in the
2977 @command{make} shipped with IRIX 5.2. We recommend you use GNU
2978 @command{make} instead of the vendor supplied @command{make} program;
2979 however, you may have success with @command{smake} on IRIX 5.2 if you do
2980 not have GNU @command{make} available.
2985 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
2987 If you are using IRIX @command{cc} as your bootstrap compiler, you must
2988 ensure that the N32 ABI is in use. To test this, compile a simple C
2989 file with @command{cc} and then run @command{file} on the
2990 resulting object file. The output should look like:
2993 test.o: ELF N32 MSB @dots{}
2999 test.o: ELF 32-bit MSB @dots{}
3005 test.o: ELF 64-bit MSB @dots{}
3008 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3009 should set the environment variable @env{CC} to @samp{cc -n32}
3010 before configuring GCC@.
3012 If you want the resulting @command{gcc} to run on old 32-bit systems
3013 with the MIPS R4400 CPU, you need to ensure that only code for the mips3
3014 instruction set architecture (ISA) is generated. While GCC 3.x does
3015 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3016 the ISA depending on the machine where GCC is built. Using one of them
3017 as the bootstrap compiler may result in mips4 code, which won't run at
3018 all on mips3-only systems. For the test program above, you should see:
3021 test.o: ELF N32 MSB mips-3 @dots{}
3027 test.o: ELF N32 MSB mips-4 @dots{}
3030 instead, you should set the environment variable @env{CC} to @samp{cc
3031 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3033 GCC on IRIX 6 is usually built to support both the N32 and N64 ABIs. If
3034 you build GCC on a system that doesn't have the N64 libraries installed,
3035 you need to configure with @option{--disable-multilib} so GCC doesn't
3036 try to use them. Look for @file{/usr/lib64/libc.so.1} to see if you
3037 have the 64-bit libraries installed.
3039 You must @emph{not} use GNU @command{as} (which isn't built anyway as of
3040 binutils 2.11.2) on IRIX 6 platforms; doing so will only cause problems.
3042 GCC does not currently support generating O32 ABI binaries in the
3043 @samp{mips-sgi-irix6} configurations. It is possible to create a GCC
3044 with O32 ABI only support by configuring it for the @samp{mips-sgi-irix5}
3045 target and using a patched GNU @command{as} 2.11.2 as documented in the
3046 @uref{#mips-sgi-irix5,,@samp{mips-sgi-irix5}} section above. Using the
3047 native assembler requires patches to GCC which will be included in a
3048 future release. It is
3049 expected that O32 ABI support will be available again in a future release.
3051 The @option{--enable-threads} option doesn't currently work, a patch is
3052 in preparation for a future release. The @option{--enable-libgcj}
3053 option is disabled by default: IRIX 6 uses a very low default limit
3054 (20480) for the command line length. Although libtool contains a
3055 workaround for this problem, at least the N64 @samp{libgcj} is known not
3056 to build despite this, running into an internal error of the native
3057 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3058 its maximum of 262144 bytes. If you have root access, you can use the
3059 @command{systune} command to do this.
3061 GCC does not correctly pass/return structures which are
3062 smaller than 16 bytes and which are not 8 bytes. The problem is very
3063 involved and difficult to fix. It affects a number of other targets also,
3064 but IRIX 6 is affected the most, because it is a 64-bit target, and 4 byte
3065 structures are common. The exact problem is that structures are being padded
3066 at the wrong end, e.g.@: a 4 byte structure is loaded into the lower 4 bytes
3067 of the register when it should be loaded into the upper 4 bytes of the
3070 GCC is consistent with itself, but not consistent with the SGI C compiler
3071 (and the SGI supplied runtime libraries), so the only failures that can
3072 happen are when there are library functions that take/return such
3073 structures. There are very few such library functions. Currently this
3074 is known to affect @code{inet_ntoa}, @code{inet_lnaof},
3075 @code{inet_netof}, @code{inet_makeaddr}, and @code{semctl}. Until the
3076 bug is fixed, GCC contains workarounds for the known affected functions.
3078 See @uref{http://freeware.sgi.com/,,http://freeware.sgi.com/} for more
3079 information about using GCC on IRIX platforms.
3084 @heading @anchor{powerpc*-*-*}powerpc-*-*
3086 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3087 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3092 @heading @anchor{powerpc-*-darwin*}powerpc-*-darwin*
3093 PowerPC running Darwin (Mac OS X kernel).
3095 Pre-installed versions of Mac OS X may not include any developer tools,
3096 meaning that you will not be able to build GCC from source. Tool
3097 binaries are available at
3098 @uref{http://developer.apple.com/tools/compilers.html} (free
3099 registration required).
3101 The default stack limit of 512K is too small, which may cause compiles
3102 to fail with 'Bus error'. Set the stack larger, for instance
3103 by doing @samp{limit stack 800}. It's a good idea to use the GNU
3104 preprocessor instead of Apple's @file{cpp-precomp} during the first stage of
3105 bootstrapping; this is automatic when doing @samp{make bootstrap}, but
3106 to do it from the toplevel objdir you will need to say @samp{make
3107 CC='cc -no-cpp-precomp' bootstrap}.
3109 The version of GCC shipped by Apple typically includes a number of
3110 extensions not available in a standard GCC release. These extensions
3111 are generally specific to Mac programming.
3116 @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4
3117 PowerPC system in big endian mode, running System V.4.
3122 @heading @anchor{powerpc-*-linux-gnu*}powerpc-*-linux-gnu*
3125 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.13.90.0.10}
3126 or newer for a working GCC@.
3131 @heading @anchor{powerpc-*-netbsd*}powerpc-*-netbsd*
3132 PowerPC system in big endian mode running NetBSD@. To build the
3133 documentation you will need Texinfo version 4.2 (NetBSD 1.5.1 included
3134 Texinfo version 3.12).
3139 @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim
3140 Embedded PowerPC system in big endian mode for use in running under the
3146 @heading @anchor{powerpc-*-eabi}powerpc-*-eabi
3147 Embedded PowerPC system in big endian mode.
3152 @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4
3153 PowerPC system in little endian mode, running System V.4.
3158 @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim
3159 Embedded PowerPC system in little endian mode for use in running under
3165 @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi
3166 Embedded PowerPC system in little endian mode.
3171 @heading @anchor{s390-*-linux*}s390-*-linux*
3172 S/390 system running Linux for S/390@.
3177 @heading @anchor{s390x-*-linux*}s390x-*-linux*
3178 zSeries system (64-bit) running Linux for zSeries@.
3183 @heading @anchor{s390x-ibm-tpf*}s390x-ibm-tpf*
3184 zSeries system (64-bit) running TPF. This platform is
3185 supported as cross-compilation target only.
3190 @c Please use Solaris 2 to refer to all release of Solaris, starting
3191 @c with 2.0 until 2.6, 7, and 8. Solaris 1 was a marketing name for
3192 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3193 @c alone is too unspecific and must be avoided.
3194 @heading @anchor{*-*-solaris2*}*-*-solaris2*
3196 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3197 GCC you first have to install a pre-built compiler, see our
3198 @uref{binaries.html,,binaries page} for details.
3200 The Solaris 2 @command{/bin/sh} will often fail to configure
3201 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3202 recommend to use the following sequence of commands to bootstrap and
3206 % CONFIG_SHELL=/bin/ksh
3207 % export CONFIG_SHELL
3210 and then proceed as described in @uref{build.html,,the build instructions},
3211 where we strongly recommend using GNU make and specifying an absolute path
3212 to invoke @var{srcdir}/configure.
3214 Solaris 2 comes with a number of optional OS packages. Some of these
3215 are needed to use GCC fully, namely @code{SUNWarc},
3216 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3217 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3218 optional packages when installing Solaris 2, you will need to verify that
3219 the packages that GCC needs are installed.
3221 To check whether an optional package is installed, use
3222 the @command{pkginfo} command. To add an optional package, use the
3223 @command{pkgadd} command. For further details, see the Solaris 2
3226 Trying to use the linker and other tools in
3227 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3228 For example, the linker may hang indefinitely. The fix is to remove
3229 @file{/usr/ucb} from your @env{PATH}.
3231 The build process works more smoothly with the legacy Sun tools so, if you
3232 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3233 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3235 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3236 platform. We recommend the use of GNU binutils 2.11.2 or the vendor
3237 tools (Sun @command{as}, Sun @command{ld}).
3239 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3240 newer: @command{g++} will complain that types are missing. These headers assume
3241 that omitting the type means @code{int}; this assumption worked for C89 but
3242 is wrong for C++, and is now wrong for C99 also.
3244 @command{g++} accepts such (invalid) constructs with the option
3245 @option{-fpermissive}; it
3246 will assume that any missing type is @code{int} (as defined by C89).
3248 There are patches for Solaris 2.6 (105633-56 or newer for SPARC,
3249 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC,
3250 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3251 108653-22 for Intel) that fix this bug.
3256 @heading @anchor{sparc-sun-solaris2*}sparc-sun-solaris2*
3258 When GCC is configured to use binutils 2.11.2 or later the binaries
3259 produced are smaller than the ones produced using Sun's native tools;
3260 this difference is quite significant for binaries containing debugging
3263 Sun @command{as} 4.x is broken in that it cannot cope with long symbol names.
3264 A typical error message might look similar to the following:
3267 /usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041: error:
3268 can't compute value of an expression involving an external symbol.
3271 This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris
3272 2.6 and has been fixed in later (5.x) versions of the assembler,
3273 starting with Solaris 7.
3275 Starting with Solaris 7, the operating system is capable of executing
3276 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3277 this; the @option{-m64} option enables 64-bit code generation.
3278 However, if all you want is code tuned for the UltraSPARC CPU, you
3279 should try the @option{-mtune=ultrasparc} option instead, which produces
3280 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3283 When configuring on a Solaris 7 or later system that is running a kernel
3284 that supports only 32-bit binaries, one must configure with
3285 @option{--disable-multilib}, since we will not be able to build the
3286 64-bit target libraries.
3291 @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7
3293 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3294 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3295 and later, including all EGCS releases. Sun formerly recommended
3296 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3297 recommend it only for people who use Sun's compilers.
3299 Here are some workarounds to this problem:
3302 Do not install Sun patch 107058-01 until after Sun releases a
3303 complete patch for bug 4210064. This is the simplest course to take,
3304 unless you must also use Sun's C compiler. Unfortunately 107058-01
3305 is preinstalled on some new Solaris 7-based hosts, so you may have to
3309 Copy the original, unpatched Solaris 7
3310 @command{/usr/ccs/bin/as} into
3311 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
3312 adjusting the latter name to fit your local conventions and software
3316 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3317 both 107058-01 and 106950-03 installed has reported the bug with GCC
3318 and Sun's dynamic linker. This last course of action is riskiest,
3319 for two reasons. First, you must install 106950 on all hosts that
3320 run code generated by GCC; it doesn't suffice to install it only on
3321 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3322 only a partial fix for bug 4210064, but Sun doesn't know whether the
3323 partial fix is adequate for GCC@. Revision -08 or later should fix
3324 the bug. The current (as of 2001-09-24) revision is -14, and is included in
3325 the Solaris 7 Recommended Patch Cluster.
3328 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
3329 which causes a bootstrap failure when linking the 64-bit shared version of
3330 libgcc. A typical error message is:
3333 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
3334 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
3337 This bug has been fixed in the final 5.0 version of the assembler.
3342 @heading @anchor{sparc-*-linux*}sparc-*-linux*
3344 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
3345 or newer on this platform. All earlier binutils and glibc
3346 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
3352 @heading @anchor{sparc64-*-solaris2*}sparc64-*-solaris2*
3354 The following compiler flags must be specified in the configure
3355 step in order to bootstrap this target with the Sun compiler:
3358 % CC="cc -xildoff -xarch=v9" @var{srcdir}/configure [@var{options}] [@var{target}]
3361 @option{-xildoff} turns off the incremental linker, and @option{-xarch=v9}
3362 specifies the SPARC-V9 architecture to the Sun linker and assembler.
3367 @heading @anchor{sparcv9-*-solaris2*}sparcv9-*-solaris2*
3369 This is a synonym for sparc64-*-solaris2*.
3374 @heading @anchor{#*-*-sysv*}*-*-sysv*
3375 On System V release 3, you may get this error message
3379 ld fatal: failed to write symbol name @var{something}
3380 in strings table for file @var{whatever}
3383 This probably indicates that the disk is full or your ulimit won't allow
3384 the file to be as large as it needs to be.
3386 This problem can also result because the kernel parameter @code{MAXUMEM}
3387 is too small. If so, you must regenerate the kernel and make the value
3388 much larger. The default value is reported to be 1024; a value of 32768
3389 is said to work. Smaller values may also work.
3391 On System V, if you get an error like this,
3394 /usr/local/lib/bison.simple: In function `yyparse':
3395 /usr/local/lib/bison.simple:625: virtual memory exhausted
3399 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
3401 On a System V release 4 system, make sure @file{/usr/bin} precedes
3402 @file{/usr/ucb} in @code{PATH}. The @command{cc} command in
3403 @file{/usr/ucb} uses libraries which have bugs.
3408 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3409 Don't try compiling with VAX C (@command{vcc}). It produces incorrect code
3410 in some cases (for example, when @code{alloca} is used).
3415 @heading @anchor{*-*-vxworks*}*-*-vxworks*
3416 Support for VxWorks is in flux. At present GCC supports @emph{only} the
3417 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC.
3418 We welcome patches for other architectures supported by VxWorks 5.5.
3419 Support for VxWorks AE would also be welcome; we believe this is merely
3420 a matter of writing an appropriate ``configlette'' (see below). We are
3421 not interested in supporting older, a.out or COFF-based, versions of
3424 VxWorks comes with an older version of GCC installed in
3425 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
3426 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
3427 Before running @command{configure}, create the directories @file{@var{prefix}}
3428 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
3429 linker, etc. into @file{@var{prefix}/bin}, and set your @var{PATH} to
3430 include that directory while running both @command{configure} and
3433 You must give @command{configure} the
3434 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
3435 find the VxWorks system headers. Since VxWorks is a cross compilation
3436 target only, you must also specify @option{--target=@var{target}}.
3437 @command{configure} will attempt to create the directory
3438 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
3439 make sure the user running @command{configure} has sufficient privilege
3442 GCC's exception handling runtime requires a special ``configlette''
3443 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
3444 that file to add the module to your kernel build. (Future versions of
3445 VxWorks will incorporate this module.)
3450 @heading @anchor{xtensa-*-elf}xtensa-*-elf
3452 This target is intended for embedded Xtensa systems using the
3453 @samp{newlib} C library. It uses ELF but does not support shared
3454 objects. Designed-defined instructions specified via the
3455 Tensilica Instruction Extension (TIE) language are only supported
3456 through inline assembly.
3458 The Xtensa configuration information must be specified prior to
3459 building GCC@. The @file{include/xtensa-config.h} header
3460 file contains the configuration information. If you created your
3461 own Xtensa configuration with the Xtensa Processor Generator, the
3462 downloaded files include a customized copy of this header file,
3463 which you can use to replace the default header file.
3468 @heading @anchor{xtensa-*-linux*}xtensa-*-linux*
3470 This target is for Xtensa systems running GNU/Linux. It supports ELF
3471 shared objects and the GNU C library (glibc). It also generates
3472 position-independent code (PIC) regardless of whether the
3473 @option{-fpic} or @option{-fPIC} options are used. In other
3474 respects, this target is the same as the
3475 @uref{#xtensa-*-elf,,@samp{xtensa-*-elf}} target.
3480 @heading @anchor{windows}Microsoft Windows (32-bit)
3482 A port of GCC 2.95.2 and 3.x is included with the
3483 @uref{http://www.cygwin.com/,,Cygwin environment}.
3485 Current (as of early 2001) snapshots of GCC will build under Cygwin
3486 without modification.
3488 GCC does not currently build with Microsoft's C++ compiler and there
3489 are no plans to make it do so.
3494 @heading @anchor{os2}OS/2
3496 GCC does not currently support OS/2. However, Andrew Zabolotny has been
3497 working on a generic OS/2 port with pgcc. The current code can be found
3498 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
3500 An older copy of GCC 2.8.1 is included with the EMX tools available at
3501 @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,,
3502 ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}.
3507 @heading @anchor{older}Older systems
3509 GCC contains support files for many older (1980s and early
3510 1990s) Unix variants. For the most part, support for these systems
3511 has not been deliberately removed, but it has not been maintained for
3512 several years and may suffer from bitrot.
3514 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
3515 Support for these systems is still present in that release, but
3516 @command{configure} will fail unless the @option{--enable-obsolete}
3517 option is given. Unless a maintainer steps forward, support for these
3518 systems will be removed from the next release of GCC@.
3520 Support for old systems as hosts for GCC can cause problems if the
3521 workarounds for compiler, library and operating system bugs affect the
3522 cleanliness or maintainability of the rest of GCC@. In some cases, to
3523 bring GCC up on such a system, if still possible with current GCC, may
3524 require first installing an old version of GCC which did work on that
3525 system, and using it to compile a more recent GCC, to avoid bugs in the
3526 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
3527 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
3528 sites}. Header bugs may generally be avoided using
3529 @command{fixincludes}, but bugs or deficiencies in libraries and the
3530 operating system may still cause problems.
3532 Support for older systems as targets for cross-compilation is less
3533 problematic than support for them as hosts for GCC; if an enthusiast
3534 wishes to make such a target work again (including resurrecting any of
3535 the targets that never worked with GCC 2, starting from the last CVS
3536 version before they were removed), patches
3537 @uref{../contribute.html,,following the usual requirements} would be
3538 likely to be accepted, since they should not affect the support for more
3541 For some systems, old versions of GNU binutils may also be useful,
3542 and are available from @file{pub/binutils/old-releases} on
3543 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
3545 Some of the information on specific systems above relates to
3546 such older systems, but much of the information
3547 about GCC on such systems (which may no longer be applicable to
3548 current GCC) is to be found in the GCC texinfo manual.
3553 @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris 2, etc.)
3555 C++ support is significantly better on ELF targets if you use the
3556 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
3557 inlines, vtables and template instantiations will be discarded
3566 @uref{./index.html,,Return to the GCC Installation page}
3570 @c ***Old documentation******************************************************
3572 @include install-old.texi
3578 @uref{./index.html,,Return to the GCC Installation page}
3582 @c ***GFDL********************************************************************
3590 @uref{./index.html,,Return to the GCC Installation page}
3594 @c ***************************************************************************
3595 @c Part 6 The End of the Document
3597 @comment node-name, next, previous, up
3598 @node Concept Index, , GNU Free Documentation License, Top
3602 @unnumbered Concept Index