1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename gccinstall.info
5 @settitle Installing GCC
10 @c Specify title for specific html page
12 @settitle Installing GCC
15 @settitle Host/Target specific installation notes for GCC
17 @ifset prerequisiteshtml
18 @settitle Prerequisites for GCC
21 @settitle Downloading GCC
24 @settitle Installing GCC: Configuration
27 @settitle Installing GCC: Building
30 @settitle Installing GCC: Testing
32 @ifset finalinstallhtml
33 @settitle Installing GCC: Final installation
36 @settitle Installing GCC: Binaries
39 @settitle Installing GCC: Old documentation
42 @settitle Installing GCC: GNU Free Documentation License
45 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
46 @c 1999, 2000, 2001, 2002, 2003 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 @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 versions 2.13 and 2.57
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. Most
295 directories require autoconf 2.13 (exactly), but @file{libiberty},
296 @file{fastjar}, and @file{libstdc++-v3} require autoconf 2.57 (exactly).
298 @item automake versions 1.4-p? and 1.7.?
300 Necessary when modifying a @file{Makefile.am} file to regenerate its
301 associated @file{Makefile.in}. Most directories require a 1.4 series
302 automake; @file{libstdc++-v3} and @file{fastjar} requires a 1.7 series
305 @item gettext version 0.12 (or later)
307 Needed to regenerate @file{gcc.pot}.
309 @item gperf version 2.7.2 (or later)
311 Necessary when modifying @command{gperf} input files, e.g.@:
312 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
313 @file{gcc/cp/cfns.h}.
315 @item expect version ???
316 @itemx tcl version ???
317 @itemx dejagnu version ???
319 Necessary to run the GCC testsuite.
321 @item autogen version 5.5.4 (or later) and
322 @itemx guile version 1.4.1 (or later)
324 Necessary to regenerate @file{fixinc/fixincl.x} from
325 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
327 Necessary to run the @file{fixinc} @command{make check}.
329 Necessary to regenerate the top level @file{Makefile.am} files from
330 @file{Makefile.tpl} and @file{Makefile.def}.
332 @item GNU Bison version 1.28 (or later)
333 Berkeley @command{yacc} (@command{byacc}) is also reported to work other
336 Necessary when modifying @file{*.y} 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 Flex version 2.5.4 (or later)
344 Necessary when modifying @file{*.l} files.
346 Necessary to build GCC during development because the generated output
347 files are not included in the CVS repository. They are included in
350 @item Texinfo version 4.2 (or later)
352 Necessary for running @command{makeinfo} when modifying @file{*.texi}
353 files to test your changes.
355 Necessary to build GCC documentation during development because the
356 generated output files are not included in the CVS repository. They are
357 included in releases.
359 @item @TeX{} (any working version)
361 Necessary for running @command{texi2dvi}, used when running
362 @command{make dvi} to create DVI files.
364 @item cvs version 1.10 (or later)
365 @itemx ssh (any version)
367 Necessary to access the CVS repository. Public releases and weekly
368 snapshots of the development sources are also available via FTP.
370 @item perl version 5.6.1 (or later)
372 Necessary when regenerating @file{Makefile} dependencies in libiberty.
373 Necessary when regenerating @file{libiberty/functions.texi}.
374 Necessary when generating manpages from Texinfo manuals.
375 Used by various scripts to generate some files included in CVS (mainly
376 Unicode-related and rarely changing) from source tables.
378 @item GNU diffutils version 2.7 (or later)
380 Necessary when creating changes to GCC source code to submit for review.
382 @item patch version 2.5.4 (or later)
384 Necessary when applying patches, created with @command{diff}, to one's
394 @uref{./index.html,,Return to the GCC Installation page}
398 @c ***Downloading the source**************************************************
400 @comment node-name, next, previous, up
401 @node Downloading the source, Configuration, Prerequisites, Installing GCC
405 @chapter Downloading GCC
407 @cindex Downloading GCC
408 @cindex Downloading the Source
410 GCC is distributed via @uref{http://gcc.gnu.org/cvs.html,,CVS} and FTP
411 tarballs compressed with @command{gzip} or
412 @command{bzip2}. It is possible to download a full distribution or specific
415 Please refer to our @uref{http://gcc.gnu.org/releases.html,,releases web page}
416 for information on how to obtain GCC@.
418 The full distribution includes the C, C++, Objective-C, Fortran, Java,
419 and Ada (in case of GCC 3.1 and later) compilers. The full distribution
420 also includes runtime libraries for C++, Objective-C, Fortran, and Java.
421 In GCC 3.0 and later versions, GNU compiler testsuites are also included
422 in the full distribution.
424 If you choose to download specific components, you must download the core
425 GCC distribution plus any language specific distributions you wish to
426 use. The core distribution includes the C language front end as well as the
427 shared components. Each language has a tarball which includes the language
428 front end as well as the language runtime (when appropriate).
430 Unpack the core distribution as well as any language specific
431 distributions in the same directory.
433 If you also intend to build binutils (either to upgrade an existing
434 installation or for use in place of the corresponding tools of your
435 OS), unpack the binutils distribution either in the same directory or
436 a separate one. In the latter case, add symbolic links to any
437 components of the binutils you intend to build alongside the compiler
438 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
439 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
446 @uref{./index.html,,Return to the GCC Installation page}
450 @c ***Configuration***********************************************************
452 @comment node-name, next, previous, up
453 @node Configuration, Building, Downloading the source, Installing GCC
457 @chapter Installing GCC: Configuration
459 @cindex Configuration
460 @cindex Installing GCC: Configuration
462 Like most GNU software, GCC must be configured before it can be built.
463 This document describes the recommended configuration procedure
464 for both native and cross targets.
466 We use @var{srcdir} to refer to the toplevel source directory for
467 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
469 If you obtained the sources via CVS, @var{srcdir} must refer to the top
470 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
471 and not its @file{gcc} subdirectory, otherwise the build will fail.
473 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
474 file system, the shell's built-in @command{pwd} command will return
475 temporary pathnames. Using these can lead to various sorts of build
476 problems. To avoid this issue, set the @env{PWDCMD} environment
477 variable to an automounter-aware @command{pwd} command, e.g.,
478 @command{pawd} or @samp{amq -w}, during the configuration and build
481 First, we @strong{highly} recommend that GCC be built into a
482 separate directory than the sources which does @strong{not} reside
483 within the source tree. This is how we generally build GCC; building
484 where @var{srcdir} == @var{objdir} should still work, but doesn't
485 get extensive testing; building where @var{objdir} is a subdirectory
486 of @var{srcdir} is unsupported.
488 If you have previously built GCC in the same directory for a
489 different target machine, do @samp{make distclean} to delete all files
490 that might be invalid. One of the files this deletes is @file{Makefile};
491 if @samp{make distclean} complains that @file{Makefile} does not exist
492 or issues a message like ``don't know how to make distclean'' it probably
493 means that the directory is already suitably clean. However, with the
494 recommended method of building in a separate @var{objdir}, you should
495 simply use a different @var{objdir} for each target.
497 Second, when configuring a native system, either @command{cc} or
498 @command{gcc} must be in your path or you must set @env{CC} in
499 your environment before running configure. Otherwise the configuration
502 Note that the bootstrap compiler and the resulting GCC must be link
503 compatible, else the bootstrap will fail with linker errors about
504 incompatible object file formats. Several multilibed targets are
505 affected by this requirement, see
507 @ref{Specific, host/target specific installation notes}.
510 @uref{specific.html,,host/target specific installation notes}.
518 % @var{srcdir}/configure [@var{options}] [@var{target}]
522 @heading Target specification
525 GCC has code to correctly determine the correct value for @var{target}
526 for nearly all native systems. Therefore, we highly recommend you not
527 provide a configure target when configuring a native compiler.
530 @var{target} must be specified as @option{--target=@var{target}}
531 when configuring a cross compiler; examples of valid targets would be
532 i960-rtems, m68k-coff, sh-elf, etc.
535 Specifying just @var{target} instead of @option{--target=@var{target}}
536 implies that the host defaults to @var{target}.
540 @heading Options specification
542 Use @var{options} to override several configure time options for
543 GCC@. A list of supported @var{options} follows; @samp{configure
544 --help} may list other options, but those not listed below may not
545 work and should not normally be used.
548 @item --prefix=@var{dirname}
549 Specify the toplevel installation
550 directory. This is the recommended way to install the tools into a directory
551 other than the default. The toplevel installation directory defaults to
554 We @strong{highly} recommend against @var{dirname} being the same or a
555 subdirectory of @var{objdir} or vice versa. If specifying a directory
556 beneath a user's home directory tree, some shells will not expand
557 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
560 The following standard @command{autoconf} options are supported. Normally you
561 should not need to use these options.
563 @item --exec-prefix=@var{dirname}
564 Specify the toplevel installation directory for architecture-dependent
565 files. The default is @file{@var{prefix}}.
567 @item --bindir=@var{dirname}
568 Specify the installation directory for the executables called by users
569 (such as @command{gcc} and @command{g++}). The default is
570 @file{@var{exec-prefix}/bin}.
572 @item --libdir=@var{dirname}
573 Specify the installation directory for object code libraries and
574 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
576 @item --libexecdir=@var{dirname}
577 Specify the installation directory for internal executables of GCC@.
578 The default is @file{@var{exec-prefix}/libexec}.
580 @item --with-slibdir=@var{dirname}
581 Specify the installation directory for the shared libgcc library. The
582 default is @file{@var{libdir}}.
584 @item --infodir=@var{dirname}
585 Specify the installation directory for documentation in info format.
586 The default is @file{@var{prefix}/info}.
588 @item --datadir=@var{dirname}
589 Specify the installation directory for some architecture-independent
590 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
592 @item --mandir=@var{dirname}
593 Specify the installation directory for manual pages. The default is
594 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
595 the full GCC manuals, which are provided in Texinfo format. The manpages
596 are derived by an automatic conversion process from parts of the full
599 @item --with-gxx-include-dir=@var{dirname}
601 the installation directory for G++ header files. The default is
602 @file{@var{prefix}/include/g++-v3}.
606 @item --program-prefix=@var{prefix}
607 GCC supports some transformations of the names of its programs when
608 installing them. This option prepends @var{prefix} to the names of
609 programs to install in @var{bindir} (see above). For example, specifying
610 @option{--program-prefix=foo-} would result in @samp{gcc}
611 being installed as @file{/usr/local/bin/foo-gcc}.
613 @item --program-suffix=@var{suffix}
614 Appends @var{suffix} to the names of programs to install in @var{bindir}
615 (see above). For example, specifying @option{--program-suffix=-3.1}
616 would result in @samp{gcc} being installed as
617 @file{/usr/local/bin/gcc-3.1}.
619 @item --program-transform-name=@var{pattern}
620 Applies the @samp{sed} script @var{pattern} to be applied to the names
621 of programs to install in @var{bindir} (see above). @var{pattern} has to
622 consist of one or more basic @samp{sed} editing commands, separated by
623 semicolons. For example, if you want the @samp{gcc} program name to be
624 transformed to the installed program @file{/usr/local/bin/myowngcc} and
625 the @samp{g++} program name to be transformed to
626 @file{/usr/local/bin/gspecial++} without changing other program names,
627 you could use the pattern
628 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
629 to achieve this effect.
631 All three options can be combined and used together, resulting in more
632 complex conversion patterns. As a basic rule, @var{prefix} (and
633 @var{suffix}) are prepended (appended) before further transformations
634 can happen with a special transformation script @var{pattern}.
636 As currently implemented, this option only takes effect for native
637 builds; cross compiler binaries' names are not transformed even when a
638 transformation is explicitly asked for by one of these options.
640 For native builds, some of the installed programs are also installed
641 with the target alias in front of their name, as in
642 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
643 before the target alias is prepended to the name - so, specifying
644 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
645 resulting binary would be installed as
646 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
648 As a last shortcoming, none of the installed Ada programs are
649 transformed yet, which will be fixed in some time.
651 @item --with-local-prefix=@var{dirname}
653 installation directory for local include files. The default is
654 @file{/usr/local}. Specify this option if you want the compiler to
655 search directory @file{@var{dirname}/include} for locally installed
656 header files @emph{instead} of @file{/usr/local/include}.
658 You should specify @option{--with-local-prefix} @strong{only} if your
659 site has a different convention (not @file{/usr/local}) for where to put
662 The default value for @option{--with-local-prefix} is @file{/usr/local}
663 regardless of the value of @option{--prefix}. Specifying
664 @option{--prefix} has no effect on which directory GCC searches for
665 local header files. This may seem counterintuitive, but actually it is
668 The purpose of @option{--prefix} is to specify where to @emph{install
669 GCC}. The local header files in @file{/usr/local/include}---if you put
670 any in that directory---are not part of GCC@. They are part of other
671 programs---perhaps many others. (GCC installs its own header files in
672 another directory which is based on the @option{--prefix} value.)
674 Both the local-prefix include directory and the GCC-prefix include
675 directory are part of GCC's "system include" directories. Although these
676 two directories are not fixed, they need to be searched in the proper
677 order for the correct processing of the include_next directive. The
678 local-prefix include directory is searched before the GCC-prefix
679 include directory. Another characteristic of system include directories
680 is that pedantic warnings are turned off for headers in these directories.
682 Some autoconf macros add @option{-I @var{directory}} options to the
683 compiler command line, to ensure that directories containing installed
684 packages' headers are searched. When @var{directory} is one of GCC's
685 system include directories, GCC will ignore the option so that system
686 directories continue to be processed in the correct order. This
687 may result in a search order different from what was specified but the
688 directory will still be searched.
690 GCC automatically searches for ordinary libraries using
691 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
692 used for both GCC and packages, GCC will automatically search for
693 both headers and libraries. This provides a configuration that is
694 easy to use. GCC behaves in a manner similar to that when it is
695 installed as a system compiler in @file{/usr}.
697 Sites that need to install multiple versions of GCC may not want to
698 use the above simple configuration. It is possible to use the
699 @option{--program-prefix}, @option{--program-suffix} and
700 @option{--program-transform-name} options to install multiple versions
701 into a single directory, but it may be simpler to use different prefixes
702 and the @option{--with-local-prefix} option to specify the location of the
703 site-specific files for each version. It will then be necessary for
704 users to specify explicitly the location of local site libraries
705 (e.g., with @env{LIBRARY_PATH}).
707 The same value can be used for both @option{--with-local-prefix} and
708 @option{--prefix} provided it is not @file{/usr}. This can be used
709 to avoid the default search of @file{/usr/local/include}.
711 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
712 The directory you use for @option{--with-local-prefix} @strong{must not}
713 contain any of the system's standard header files. If it did contain
714 them, certain programs would be miscompiled (including GNU Emacs, on
715 certain targets), because this would override and nullify the header
716 file corrections made by the @command{fixincludes} script.
718 Indications are that people who use this option use it based on mistaken
719 ideas of what it is for. People use it as if it specified where to
720 install part of GCC@. Perhaps they make this assumption because
721 installing GCC creates the directory.
723 @item --enable-shared[=@var{package}[,@dots{}]]
724 Build shared versions of libraries, if shared libraries are supported on
725 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
726 are enabled by default on all platforms that support shared libraries,
727 except for @samp{libobjc} which is built as a static library only by
730 If a list of packages is given as an argument, build shared libraries
731 only for the listed packages. For other packages, only static libraries
732 will be built. Package names currently recognized in the GCC tree are
733 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
734 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc} and
735 @samp{libjava}. Note that @samp{libobjc} does not recognize itself by
736 any name, so, if you list package names in @option{--enable-shared},
737 you will only get static Objective-C libraries. @samp{libf2c} and
738 @samp{libiberty} do not support shared libraries at all.
740 Use @option{--disable-shared} to build only static libraries. Note that
741 @option{--disable-shared} does not accept a list of package names as
742 argument, only @option{--enable-shared} does.
744 @item @anchor{with-gnu-as}--with-gnu-as
745 Specify that the compiler should assume that the
746 assembler it finds is the GNU assembler. However, this does not modify
747 the rules to find an assembler and will result in confusion if the
748 assembler found is not actually the GNU assembler. (Confusion may also
749 result if the compiler finds the GNU assembler but has not been
750 configured with @option{--with-gnu-as}.) If you have more than one
751 assembler installed on your system, you may want to use this option in
752 connection with @option{--with-as=@var{pathname}}.
754 The following systems are the only ones where it makes a difference
755 whether you use the GNU assembler. On any other system,
756 @option{--with-gnu-as} has no effect.
759 @item @samp{hppa1.0-@var{any}-@var{any}}
760 @item @samp{hppa1.1-@var{any}-@var{any}}
761 @item @samp{i386-@var{any}-sysv}
762 @item @samp{m68k-bull-sysv}
763 @item @samp{m68k-hp-hpux}
764 @item @samp{m68000-hp-hpux}
765 @item @samp{m68000-att-sysv}
766 @item @samp{@var{any}-lynx-lynxos}
767 @item @samp{mips-@var{any}}
768 @item @samp{sparc-sun-solaris2.@var{any}}
769 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
772 On the systems listed above (except for the HP-PA, the SPARC, for ISC on
773 the 386, and for @samp{mips-sgi-irix5.*}), if you use the GNU assembler,
774 you should also use the GNU linker (and specify @option{--with-gnu-ld}).
776 @item @anchor{with-as}--with-as=@var{pathname}
778 compiler should use the assembler pointed to by @var{pathname}, rather
779 than the one found by the standard rules to find an assembler, which
783 Check the @file{@var{libexec}/gcc/@var{target}/@var{version}}
784 directory, where @var{libexec} defaults to
785 @file{@var{exec-prefix}/libexec} and @var{exec-prefix} defaults to
786 @var{prefix} which defaults to @file{/usr/local} unless overridden by
787 the @option{--prefix=@var{pathname}} switch described
788 above. @var{target} is the target system triple, such as
789 @samp{sparc-sun-solaris2.7}, and @var{version} denotes the GCC
790 version, such as 3.0.
792 Check operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
795 Note that these rules do not check for the value of @env{PATH}. You may
796 want to use @option{--with-as} if no assembler is installed in the
797 directories listed above, or if you have multiple assemblers installed
798 and want to choose one that is not found by the above rules.
800 @item @anchor{with-gnu-ld}--with-gnu-ld
801 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
804 @item --with-ld=@var{pathname}
805 Same as @uref{#with-as,,@option{--with-as}}
809 Specify that stabs debugging
810 information should be used instead of whatever format the host normally
811 uses. Normally GCC uses the same debug format as the host system.
813 On MIPS based systems and on Alphas, you must specify whether you want
814 GCC to create the normal ECOFF debugging format, or to use BSD-style
815 stabs passed through the ECOFF symbol table. The normal ECOFF debug
816 format cannot fully handle languages other than C@. BSD stabs format can
817 handle other languages, but it only works with the GNU debugger GDB@.
819 Normally, GCC uses the ECOFF debugging format by default; if you
820 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
822 No matter which default you choose when you configure GCC, the user
823 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
824 the debug format for a particular compilation.
826 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
827 @option{--with-gas} is used. It selects use of stabs debugging
828 information embedded in COFF output. This kind of debugging information
829 supports C++ well; ordinary COFF debugging information does not.
831 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
832 selects use of stabs debugging information embedded in ELF output. The
833 C++ compiler currently (2.6.0) does not support the DWARF debugging
834 information normally used on 386 SVR4 platforms; stabs provide a
835 workable alternative. This requires gas and gdb, as the normal SVR4
836 tools can not generate or interpret stabs.
838 @item --disable-multilib
839 Specify that multiple target
840 libraries to support different target variants, calling
841 conventions, etc should not be built. The default is to build a
842 predefined set of them.
844 Some targets provide finer-grained control over which multilibs are built
845 (e.g., @option{--disable-softfloat}):
851 fpu, 26bit, underscore, interwork, biendian, nofmult.
854 softfloat, m68881, m68000, m68020.
857 single-float, biendian, softfloat.
859 @item powerpc*-*-*, rs6000*-*-*
860 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
865 @item --enable-threads
866 Specify that the target
867 supports threads. This affects the Objective-C compiler and runtime
868 library, and exception handling for other languages like C++ and Java.
869 On some systems, this is the default.
871 In general, the best (and, in many cases, the only known) threading
872 model available will be configured for use. Beware that on some
873 systems, gcc has not been taught what threading models are generally
874 available for the system. In this case, @option{--enable-threads} is an
875 alias for @option{--enable-threads=single}.
877 @item --disable-threads
878 Specify that threading support should be disabled for the system.
879 This is an alias for @option{--enable-threads=single}.
881 @item --enable-threads=@var{lib}
883 @var{lib} is the thread support library. This affects the Objective-C
884 compiler and runtime library, and exception handling for other languages
885 like C++ and Java. The possibilities for @var{lib} are:
893 Ada tasking support. For non-Ada programs, this setting is equivalent
894 to @samp{single}. When used in conjunction with the Ada run time, it
895 causes GCC to use the same thread primitives as Ada uses. This option
896 is necessary when using both Ada and the back end exception handling,
897 which is the default for most Ada targets.
899 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
900 that the file needed to support this configuration, @file{gthr-mach.h}, is
901 missing and thus this setting will cause a known bootstrap failure.)
903 This is an alias for @samp{single}.
905 Generic POSIX thread support.
907 RTEMS thread support.
909 Disable thread support, should work for all platforms.
911 Sun Solaris 2 thread support.
913 VxWorks thread support.
915 Microsoft Win32 API thread support.
918 @item --with-cpu=@var{cpu}
919 Specify which cpu variant the compiler should generate code for by default.
920 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
921 This option is only supported on some targets, including ARM, i386, PowerPC,
924 @item --with-schedule=@var{cpu}
925 @itemx --with-arch=@var{cpu}
926 @itemx --with-tune=@var{cpu}
927 @itemx --with-abi=@var{abi}
928 @itemx --with-float=@var{type}
929 These configure options provide default values for the @option{-mschedule=},
930 @option{-march=}, @option{-mtune=}, and @option{-mabi=} options and for
931 @option{-mhard-float} or @option{-msoft-float}. As with @option{--with-cpu},
932 which switches will be accepted and acceptable values of the arguments depend
935 @item --enable-altivec
936 Specify that the target supports AltiVec vector enhancements. This
937 option will adjust the ABI for AltiVec enhancements, as well as generate
938 AltiVec code when appropriate. This option is only available for
941 @item --enable-target-optspace
943 libraries should be optimized for code space instead of code speed.
944 This is the default for the m32r platform.
947 Specify that a user visible @command{cpp} program should not be installed.
949 @item --with-cpp-install-dir=@var{dirname}
950 Specify that the user visible @command{cpp} program should be installed
951 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
953 @item --enable-initfini-array
954 Force the use of sections @code{.init_array} and @code{.fini_array}
955 (instead of @code{.init} and @code{.fini}) for constructors and
956 destructors. Option @option{--disable-initfini-array} has the
957 opposite effect. If neither option is specified, the configure script
958 will try to guess whether the @code{.init_array} and
959 @code{.fini_array} sections are supported and, if they are, use them.
961 @item --enable-maintainer-mode
963 regenerate the GCC master message catalog @file{gcc.pot} are normally
964 disabled. This is because it can only be rebuilt if the complete source
965 tree is present. If you have changed the sources and want to rebuild the
966 catalog, configuring with @option{--enable-maintainer-mode} will enable
967 this. Note that you need a recent version of the @code{gettext} tools
970 @item --enable-generated-files-in-srcdir
971 Neither the .c and .h files that are generated from bison and flex nor the
972 info manuals and man pages that are built from the .texi files are present
973 in the CVS development tree. When building GCC from that development tree,
974 or from a snapshot which are created from CVS, then those generated files
975 are placed in your build directory, which allows for the source to be in a
978 If you configure with @option{--enable-generated-files-in-srcdir} then those
979 generated files will go into the source directory. This is mainly intended
980 for generating release or prerelease tarballs of the GCC sources, since it
981 is not a requirement that the users of source releases to have flex, bison, or
984 @item --enable-version-specific-runtime-libs
986 that runtime libraries should be installed in the compiler specific
987 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
988 addition, @samp{libstdc++}'s include files will be installed into
989 @file{@var{libdir}} unless you overruled it by using
990 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
991 particularly useful if you intend to use several versions of GCC in
992 parallel. This is currently supported by @samp{libf2c} and
993 @samp{libstdc++}, and is the default for @samp{libobjc} which cannot be
994 changed in this case.
996 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
997 Specify that only a particular subset of compilers and
998 their runtime libraries should be built. For a list of valid values for
999 @var{langN} you can issue the following command in the
1000 @file{gcc} directory of your GCC source tree:@*
1002 grep language= */config-lang.in
1004 Currently, you can use any of the following:
1005 @code{ada}, @code{c}, @code{c++}, @code{f77}, @code{java}, @code{objc}.
1006 Building the Ada compiler has special requirements, see below.@*
1007 If you do not pass this flag, all languages available in the @file{gcc}
1008 sub-tree will be configured. Re-defining @code{LANGUAGES} when calling
1009 @samp{make bootstrap} @strong{does not} work anymore, as those
1010 language sub-directories might not have been configured!
1012 @item --disable-libgcj
1013 Specify that the run-time libraries
1014 used by GCJ should not be built. This is useful in case you intend
1015 to use GCJ with some other run-time, or you're going to install it
1016 separately, or it just happens not to build on your particular
1017 machine. In general, if the Java front end is enabled, the GCJ
1018 libraries will be enabled too, unless they're known to not work on
1019 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
1020 may need to port it; in this case, before modifying the top-level
1021 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
1022 you may use @option{--enable-libgcj} to override the default.
1025 Specify that the compiler should
1026 use DWARF 2 debugging information as the default.
1028 @item --enable-win32-registry
1029 @itemx --enable-win32-registry=@var{key}
1030 @itemx --disable-win32-registry
1031 The @option{--enable-win32-registry} option enables Windows-hosted GCC
1032 to look up installations paths in the registry using the following key:
1035 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1038 @var{key} defaults to GCC version number, and can be overridden by the
1039 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1040 who use custom installers are encouraged to provide a different key,
1041 perhaps one comprised of vendor name and GCC version number, to
1042 avoid conflict with existing installations. This feature is enabled
1043 by default, and can be disabled by @option{--disable-win32-registry}
1044 option. This option has no effect on the other hosts.
1047 Specify that the machine does not have a floating point unit. This
1048 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1049 system, @option{--nfp} has no effect.
1051 @item --enable-werror
1052 @itemx --disable-werror
1053 @itemx --enable-werror=yes
1054 @itemx --enable-werror=no
1055 When you specify this option, it controls whether certain files in the
1056 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1057 If you don't specify it, @option{-Werror} is turned on for the main
1058 development trunk. However it defaults to off for release branches and
1059 final releases. The specific files which get @option{-Werror} are
1060 controlled by the Makefiles.
1062 @item --enable-checking
1063 @itemx --enable-checking=@var{list}
1064 When you specify this option, the compiler is built to perform checking
1065 of tree node types when referencing fields of that node, and some other
1066 internal consistency checks. This does not change the generated code,
1067 but adds error checking within the compiler. This will slow down the
1068 compiler and may only work properly if you are building the compiler
1069 with GCC@. This is on by default when building from CVS or snapshots,
1070 but off for releases. More control over the checks may be had by
1071 specifying @var{list}; the categories of checks available are
1072 @samp{misc}, @samp{tree}, @samp{gc}, @samp{rtl}, @samp{rtlflag},
1073 @samp{fold}, @samp{gcac} and @samp{valgrind}. The check @samp{valgrind}
1074 requires the external @command{valgrind} simulator, available from
1075 @uref{http://valgrind.kde.org/}. The default when @var{list} is
1076 not specified is @samp{misc,tree,gc,rtlflag}; the checks @samp{rtl},
1077 @samp{gcac} and @samp{valgrind} are very expensive.
1079 @item --enable-coverage
1080 @item --enable-coverage=@var{level}
1081 With this option, the compiler is built to collect self coverage
1082 information, every time it is run. This is for internal development
1083 purposes, and only works when the compiler is being built with gcc. The
1084 @var{level} argument controls whether the compiler is built optimized or
1085 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1086 want to disable optimization, for performance analysis you want to
1087 enable optimization. When coverage is enabled, the default level is
1088 without optimization.
1091 @itemx --disable-nls
1092 The @option{--enable-nls} option enables Native Language Support (NLS),
1093 which lets GCC output diagnostics in languages other than American
1094 English. Native Language Support is enabled by default if not doing a
1095 canadian cross build. The @option{--disable-nls} option disables NLS@.
1097 @item --with-included-gettext
1098 If NLS is enabled, the @option{--with-included-gettext} option causes the build
1099 procedure to prefer its copy of GNU @command{gettext}.
1101 @item --with-catgets
1102 If NLS is enabled, and if the host lacks @code{gettext} but has the
1103 inferior @code{catgets} interface, the GCC build procedure normally
1104 ignores @code{catgets} and instead uses GCC's copy of the GNU
1105 @code{gettext} library. The @option{--with-catgets} option causes the
1106 build procedure to use the host's @code{catgets} in this situation.
1108 @item --with-libiconv-prefix=@var{dir}
1109 Search for libiconv header files in @file{@var{dir}/include} and
1110 libiconv library files in @file{@var{dir}/lib}.
1112 @item --with-system-zlib
1113 Use installed zlib rather than that included with GCC@. This option
1114 only applies if the Java front end is being built.
1116 @item --enable-obsolete
1117 Enable configuration for an obsoleted system. If you attempt to
1118 configure GCC for a system (build, host, or target) which has been
1119 obsoleted, and you do not specify this flag, configure will halt with an
1122 All support for systems which have been obsoleted in one release of GCC
1123 is removed entirely in the next major release, unless someone steps
1124 forward to maintain the port.
1127 Some options which only apply to building cross compilers:
1129 @item --with-sysroot
1130 @itemx --with-sysroot=@var{dir}
1131 Tells GCC to consider @var{dir} as the root of a tree that contains a
1132 (subset of) the root filesystem of the target operating system.
1133 Target system headers, libraries and run-time object files will be
1134 searched in there. The specified directory is not copied into the
1135 install tree, unlike the options @option{--with-headers} and
1136 @option{--with-libs} that this option obsoletes. The default value,
1137 in case @option{--with-sysroot} is not given an argument, is
1138 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
1139 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
1140 the GCC binaries if the installation tree is moved.
1142 @item --with-headers
1143 @itemx --with-headers=@var{dir}
1144 Deprecated in favor of @option{--with-sysroot}.
1145 Specifies that target headers are available when building a cross compiler.
1146 The @var{dir} argument specifies a directory which has the target include
1147 files. These include files will be copied into the @file{gcc} install
1148 directory. @emph{This option with the @var{dir} argument is required} when
1149 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
1150 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
1151 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
1152 will be run on these files to make them compatible with GCC.
1154 @item --without-headers
1155 Tells GCC not use any target headers from a libc when building a cross
1156 compiler. When crossing to GNU/Linux, you need the headers so gcc
1157 can build the exception handling for libgcc.
1158 See @uref{http://www.objsw.com/CrossGCC/,,CrossGCC} for more information
1162 @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
1163 Deprecated in favor of @option{--with-sysroot}.
1164 Specifies a list of directories which contain the target runtime
1165 libraries. These libraries will be copied into the @file{gcc} install
1166 directory. If the directory list is omitted, this option has no
1169 Specifies that @samp{newlib} is
1170 being used as the target C library. This causes @code{__eprintf} to be
1171 omitted from @file{libgcc.a} on the assumption that it will be provided by
1175 Note that each @option{--enable} option has a corresponding
1176 @option{--disable} option and that each @option{--with} option has a
1177 corresponding @option{--without} option.
1184 @uref{./index.html,,Return to the GCC Installation page}
1188 @c ***Building****************************************************************
1190 @comment node-name, next, previous, up
1191 @node Building, Testing, Configuration, Installing GCC
1197 @cindex Installing GCC: Building
1199 Now that GCC is configured, you are ready to build the compiler and
1202 We @strong{highly} recommend that GCC be built using GNU make;
1203 other versions may work, then again they might not.
1204 GNU make is required for compiling GNAT (the Ada compiler) and the Java
1207 (For example, many broken versions of make will fail if you use the
1208 recommended setup where @var{objdir} is different from @var{srcdir}.
1209 Other broken versions may recompile parts of the compiler when
1210 installing the compiler.)
1212 Some commands executed when making the compiler may fail (return a
1213 nonzero status) and be ignored by @command{make}. These failures, which
1214 are often due to files that were not found, are expected, and can safely
1217 It is normal to have compiler warnings when compiling certain files.
1218 Unless you are a GCC developer, you can generally ignore these warnings
1219 unless they cause compilation to fail. Developers should attempt to fix
1220 any warnings encountered, however they can temporarily continue past
1221 warnings-as-errors by specifying the configure flag
1222 @option{--disable-werror}.
1224 On certain old systems, defining certain environment variables such as
1225 @env{CC} can interfere with the functioning of @command{make}.
1227 If you encounter seemingly strange errors when trying to build the
1228 compiler in a directory other than the source directory, it could be
1229 because you have previously configured the compiler in the source
1230 directory. Make sure you have done all the necessary preparations.
1232 If you build GCC on a BSD system using a directory stored in an old System
1233 V file system, problems may occur in running @command{fixincludes} if the
1234 System V file system doesn't support symbolic links. These problems
1235 result in a failure to fix the declaration of @code{size_t} in
1236 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
1237 that type mismatches occur, this could be the cause.
1239 The solution is not to use such a directory for building GCC@.
1241 When building from CVS or snapshots, or if you modify parser sources,
1242 you need the Bison parser generator installed. Any version 1.25 or
1243 later should work; older versions may also work. If you do not modify
1244 parser sources, releases contain the Bison-generated files and you do
1245 not need Bison installed to build them.
1247 When building from CVS or snapshots, or if you modify Texinfo
1248 documentation, you need version 4.2 or later of Texinfo installed if you
1249 want Info documentation to be regenerated. Releases contain Info
1250 documentation pre-built for the unmodified documentation in the release.
1252 @section Building a native compiler
1254 For a native build issue the command @samp{make bootstrap}. This
1255 will build the entire GCC system, which includes the following steps:
1259 Build host tools necessary to build the compiler such as texinfo, bison,
1263 Build target tools for use by the compiler such as binutils (bfd,
1264 binutils, gas, gprof, ld, and opcodes)
1265 if they have been individually linked
1266 or moved into the top level GCC source tree before configuring.
1269 Perform a 3-stage bootstrap of the compiler.
1272 Perform a comparison test of the stage2 and stage3 compilers.
1275 Build runtime libraries using the stage3 compiler from the previous step.
1279 If you are short on disk space you might consider @samp{make
1280 bootstrap-lean} instead. This is identical to @samp{make
1281 bootstrap} except that object files from the stage1 and
1282 stage2 of the 3-stage bootstrap of the compiler are deleted as
1283 soon as they are no longer needed.
1285 If you want to save additional space during the bootstrap and in
1286 the final installation as well, you can build the compiler binaries
1287 without debugging information as in the following example. This will save
1288 roughly 40% of disk space both for the bootstrap and the final installation.
1289 (Libraries will still contain debugging information.)
1292 make CFLAGS='-O' LIBCFLAGS='-g -O2' \
1293 LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap
1296 If you wish to use non-default GCC flags when compiling the stage2 and
1297 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
1298 @samp{make bootstrap}. Non-default optimization flags are less well
1299 tested here than the default of @samp{-g -O2}, but should still work.
1300 In a few cases, you may find that you need to specify special flags such
1301 as @option{-msoft-float} here to complete the bootstrap; or, if the
1302 native compiler miscompiles the stage1 compiler, you may need to work
1303 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
1304 stage1 compiler that were miscompiled, or by using @samp{make
1305 bootstrap4} to increase the number of stages of bootstrap.
1307 If you used the flag @option{--enable-languages=@dots{}} to restrict
1308 the compilers to be built, only those you've actually enabled will be
1309 built. This will of course only build those runtime libraries, for
1310 which the particular compiler has been built. Please note,
1311 that re-defining @env{LANGUAGES} when calling @samp{make bootstrap}
1312 @strong{does not} work anymore!
1314 If the comparison of stage2 and stage3 fails, this normally indicates
1315 that the stage2 compiler has compiled GCC incorrectly, and is therefore
1316 a potentially serious bug which you should investigate and report. (On
1317 a few systems, meaningful comparison of object files is impossible; they
1318 always appear ``different''. If you encounter this problem, you will
1319 need to disable comparison in the @file{Makefile}.)
1321 @section Building a cross compiler
1323 We recommend reading the
1324 @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ}
1325 for information about building cross compilers.
1327 When building a cross compiler, it is not generally possible to do a
1328 3-stage bootstrap of the compiler. This makes for an interesting problem
1329 as parts of GCC can only be built with GCC@.
1331 To build a cross compiler, we first recommend building and installing a
1332 native compiler. You can then use the native GCC compiler to build the
1333 cross compiler. The installed native compiler needs to be GCC version
1336 Assuming you have already installed a native copy of GCC and configured
1337 your cross compiler, issue the command @command{make}, which performs the
1342 Build host tools necessary to build the compiler such as texinfo, bison,
1346 Build target tools for use by the compiler such as binutils (bfd,
1347 binutils, gas, gprof, ld, and opcodes)
1348 if they have been individually linked or moved into the top level GCC source
1349 tree before configuring.
1352 Build the compiler (single stage only).
1355 Build runtime libraries using the compiler from the previous step.
1358 Note that if an error occurs in any step the make process will exit.
1360 If you are not building GNU binutils in the same source tree as GCC,
1361 you will need a cross-assembler and cross-linker installed before
1362 configuring GCC@. Put them in the directory
1363 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
1364 you should put in this directory:
1368 This should be the cross-assembler.
1371 This should be the cross-linker.
1374 This should be the cross-archiver: a program which can manipulate
1375 archive files (linker libraries) in the target machine's format.
1378 This should be a program to construct a symbol table in an archive file.
1381 The installation of GCC will find these programs in that directory,
1382 and copy or link them to the proper place to for the cross-compiler to
1383 find them when run later.
1385 The easiest way to provide these files is to build the Binutils package.
1386 Configure it with the same @option{--host} and @option{--target}
1387 options that you use for configuring GCC, then build and install
1388 them. They install their executables automatically into the proper
1389 directory. Alas, they do not support all the targets that GCC
1392 If you are not building a C library in the same source tree as GCC,
1393 you should also provide the target libraries and headers before
1394 configuring GCC, specifying the directories with
1395 @option{--with-sysroot} or @option{--with-headers} and
1396 @option{--with-libs}. Many targets also require ``start files'' such
1397 as @file{crt0.o} and
1398 @file{crtn.o} which are linked into each executable. There may be several
1399 alternatives for @file{crt0.o}, for use with profiling or other
1400 compilation options. Check your target's definition of
1401 @code{STARTFILE_SPEC} to find out what start files it uses.
1403 @section Building in parallel
1405 You can use @samp{make bootstrap MAKE="make -j 2" -j 2}, or just
1406 @samp{make -j 2 bootstrap} for GNU Make 3.79 and above, instead of
1407 @samp{make bootstrap} to build GCC in parallel.
1408 You can also specify a bigger number, and in most cases using a value
1409 greater than the number of processors in your machine will result in
1410 fewer and shorter I/O latency hits, thus improving overall throughput;
1411 this is especially true for slow drives and network filesystems.
1413 @section Building the Ada compiler
1415 In order to build GNAT, the Ada compiler, you need a working GNAT
1416 compiler (GNAT version 3.14 or later, or GCC version 3.1 or later),
1417 including GNAT tools such as @command{gnatmake} and @command{gnatlink},
1418 since the Ada front end is written in Ada (with some
1419 GNAT-specific extensions), and GNU make.
1421 @command{configure} does not test whether the GNAT installation works
1422 and has a sufficiently recent version; if too old a GNAT version is
1423 installed, the build will fail unless @option{--enable-languages} is
1424 used to disable building the Ada front end.
1426 At the moment, the GNAT library and several tools for GNAT are not built
1427 by @samp{make bootstrap}. For a native build, you have to invoke
1428 @samp{make gnatlib_and_tools} in the @file{@var{objdir}/gcc}
1429 subdirectory before proceeding with the next steps.
1430 For a cross build, you need to invoke
1431 @samp{make gnatlib cross-gnattools ada.all.cross}. For a canadian
1432 cross you only need to invoke @samp{make cross-gnattools}; the GNAT
1433 library would be the same as the one built for the cross compiler.
1435 For example, you can build a native Ada compiler by issuing the
1436 following commands (assuming @command{make} is GNU make):
1440 @var{srcdir}/configure --enable-languages=c,ada
1444 make gnatlib_and_tools
1448 Currently, when compiling the Ada front end, you cannot use the parallel
1449 build feature described in the previous section.
1451 @section Building with profile feedback
1453 It is possible to use profile feedback to optimize the compiler itself. This
1454 should result in a faster compiler binary. Experiments done on x86 using gcc
1455 3.3 showed approximately 7 percent speedup on compiling C programs. To
1456 bootstrap compiler with profile feedback, use @code{make profiledbootstrap}.
1458 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
1459 compiler. This compiler is used to build a @code{stageprofile} compiler
1460 instrumented to collect execution counts of instruction and branch
1461 probabilities. Then runtime libraries are compiled with profile collected.
1462 Finally a @code{stagefeedback} compiler is built using the information collected.
1464 Unlike @samp{make bootstrap} several additional restrictions apply. The
1465 compiler used to build @code{stage1} needs to support a 64-bit integral type.
1466 It is recommended to only use GCC for this. Also parallel make is currently
1467 not supported since collisions in profile collecting may occur.
1474 @uref{./index.html,,Return to the GCC Installation page}
1478 @c ***Testing*****************************************************************
1480 @comment node-name, next, previous, up
1481 @node Testing, Final install, Building, Installing GCC
1485 @chapter Installing GCC: Testing
1488 @cindex Installing GCC: Testing
1491 Before you install GCC, we encourage you to run the testsuites and to
1492 compare your results with results from a similar configuration that have
1493 been submitted to the
1494 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
1495 Some of these archived results are linked from the build status lists
1496 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
1497 reports a successful build runs the testsuites and submits the results.
1498 This step is optional and may require you to download additional software,
1499 but it can give you confidence in your new GCC installation or point out
1500 problems before you install and start using your new GCC.
1502 First, you must have @uref{download.html,,downloaded the testsuites}.
1503 These are part of the full distribution, but if you downloaded the
1504 ``core'' compiler plus any front ends, you must download the testsuites
1507 Second, you must have the testing tools installed. This includes
1508 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu} 1.4.1 or 1.4.3
1509 and later, Tcl, and Expect; the DejaGnu site has links to these.
1511 If the directories where @command{runtest} and @command{expect} were
1512 installed are not in the @env{PATH}, you may need to set the following
1513 environment variables appropriately, as in the following example (which
1514 assumes that DejaGnu has been installed under @file{/usr/local}):
1517 TCL_LIBRARY = /usr/local/share/tcl8.0
1518 DEJAGNULIBS = /usr/local/share/dejagnu
1521 (On systems such as Cygwin, these paths are required to be actual
1522 paths, not mounts or links; presumably this is due to some lack of
1523 portability in the DejaGnu code.)
1526 Finally, you can run the testsuite (which may take a long time):
1528 cd @var{objdir}; make -k check
1531 This will test various components of GCC, such as compiler
1532 front ends and runtime libraries. While running the testsuite, DejaGnu
1533 might emit some harmless messages resembling
1534 @samp{WARNING: Couldn't find the global config file.} or
1535 @samp{WARNING: Couldn't find tool init file} that can be ignored.
1537 @section How can I run the test suite on selected tests?
1539 In order to run sets of tests selectively, there are targets
1540 @samp{make check-gcc} and @samp{make check-g++}
1541 in the @file{gcc} subdirectory of the object directory. You can also
1542 just run @samp{make check} in a subdirectory of the object directory.
1545 A more selective way to just run all @command{gcc} execute tests in the
1549 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
1552 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
1553 the testsuite with filenames matching @samp{9805*}, you would use
1556 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1559 The @file{*.exp} files are located in the testsuite directories of the GCC
1560 source, the most important ones being @file{compile.exp},
1561 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1562 To get a list of the possible @file{*.exp} files, pipe the
1563 output of @samp{make check} into a file and look at the
1564 @samp{Running @dots{} .exp} lines.
1566 @section Passing options and running multiple testsuites
1568 You can pass multiple options to the testsuite using the
1569 @samp{--target_board} option of DejaGNU, either passed as part of
1570 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
1571 work outside the makefiles. For example,
1574 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fno-strength-reduce"
1577 will run the standard @command{g++} testsuites (``unix'' is the target name
1578 for a standard native testsuite situation), passing
1579 @samp{-O3 -fno-strength-reduce} to the compiler on every test, i.e.,
1580 slashes separate options.
1582 You can run the testsuites multiple times using combinations of options
1583 with a syntax similar to the brace expansion of popular shells:
1586 @dots{}"--target_board=arm-sim@{-mhard-float,-msoft-float@}@{-O1,-O2,-O3,@}"
1589 (Note the empty option caused by the trailing comma in the final group.)
1590 The following will run each testsuite eight times using the @samp{arm-sim}
1591 target, as if you had specified all possible combinations yourself:
1594 --target_board=arm-sim/-mhard-float/-O1
1595 --target_board=arm-sim/-mhard-float/-O2
1596 --target_board=arm-sim/-mhard-float/-O3
1597 --target_board=arm-sim/-mhard-float
1598 --target_board=arm-sim/-msoft-float/-O1
1599 --target_board=arm-sim/-msoft-float/-O2
1600 --target_board=arm-sim/-msoft-float/-O3
1601 --target_board=arm-sim/-msoft-float
1604 They can be combined as many times as you wish, in arbitrary ways. This
1608 @dots{}"--target_board=unix/-Wextra@{-O3,-fno-strength-reduce@}@{-fomit-frame-pointer,@}"
1611 will generate four combinations, all involving @samp{-Wextra}.
1613 The disadvantage to this method is that the testsuites are run in serial,
1614 which is a waste on multiprocessor systems. For users with GNU Make and
1615 a shell which performs brace expansion, you can run the testsuites in
1616 parallel by having the shell perform the combinations and @command{make}
1617 do the parallel runs. Instead of using @samp{--target_board}, use a
1618 special makefile target:
1621 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
1627 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
1630 will run three concurrent ``make-gcc'' testsuites, eventually testing all
1631 ten combinations as described above. Note that this is currently only
1632 supported in the @file{gcc} subdirectory. (To see how this works, try
1633 typing @command{echo} before the example given here.)
1636 @section Additional testing for Java Class Libraries
1638 The Java runtime tests can be executed via @samp{make check}
1639 in the @file{@var{target}/libjava/testsuite} directory in
1642 The @uref{http://sources.redhat.com/mauve/,,Mauve Project} provides
1643 a suite of tests for the Java Class Libraries. This suite can be run
1644 as part of libgcj testing by placing the Mauve tree within the libjava
1645 testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by
1646 specifying the location of that tree when invoking @samp{make}, as in
1647 @samp{make MAUVEDIR=~/mauve check}.
1649 @uref{http://www-124.ibm.com/developerworks/oss/cvs/jikes/~checkout~/jacks/jacks.html,,Jacks}
1650 is a free test suite that tests Java compiler front ends. This suite
1651 can be run as part of libgcj testing by placing the Jacks tree within
1652 the libjava testsuite at @file{libjava/testsuite/libjava.jacks/jacks}.
1654 @section How to interpret test results
1656 The result of running the testsuite are various @file{*.sum} and @file{*.log}
1657 files in the testsuite subdirectories. The @file{*.log} files contain a
1658 detailed log of the compiler invocations and the corresponding
1659 results, the @file{*.sum} files summarize the results. These summaries
1660 contain status codes for all tests:
1664 PASS: the test passed as expected
1666 XPASS: the test unexpectedly passed
1668 FAIL: the test unexpectedly failed
1670 XFAIL: the test failed as expected
1672 UNSUPPORTED: the test is not supported on this platform
1674 ERROR: the testsuite detected an error
1676 WARNING: the testsuite detected a possible problem
1679 It is normal for some tests to report unexpected failures. At the
1680 current time our testing harness does not allow fine grained control
1681 over whether or not a test is expected to fail. We expect to fix this
1682 problem in future releases.
1685 @section Submitting test results
1687 If you want to report the results to the GCC project, use the
1688 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
1691 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
1692 -m gcc-testresults@@gcc.gnu.org |sh
1695 This script uses the @command{Mail} program to send the results, so
1696 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
1697 prepended to the testsuite summary and should contain any special
1698 remarks you have on your results or your build environment. Please
1699 do not edit the testsuite result block or the subject line, as these
1700 messages may be automatically processed.
1707 @uref{./index.html,,Return to the GCC Installation page}
1711 @c ***Final install***********************************************************
1713 @comment node-name, next, previous, up
1714 @node Final install, , Testing, Installing GCC
1716 @ifset finalinstallhtml
1718 @chapter Installing GCC: Final installation
1721 Now that GCC has been built (and optionally tested), you can install it with
1723 cd @var{objdir}; make install
1726 We strongly recommend to install into a target directory where there is
1727 no previous version of GCC present.
1729 That step completes the installation of GCC; user level binaries can
1730 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
1731 you specified with the @option{--prefix} to configure (or
1732 @file{/usr/local} by default). (If you specified @option{--bindir},
1733 that directory will be used instead; otherwise, if you specified
1734 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
1735 Headers for the C++ and Java libraries are installed in
1736 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
1737 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
1738 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
1739 in info format in @file{@var{infodir}} (normally
1740 @file{@var{prefix}/info}).
1742 When installing cross-compilers, GCC's executables
1743 are not only installed into @file{@var{bindir}}, that
1744 is, @file{@var{exec-prefix}/bin}, but additionally into
1745 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
1746 exists. Typically, such @dfn{tooldirs} hold target-specific
1747 binutils, including assembler and linker.
1749 Installation into a temporary staging area or into a @command{chroot}
1750 jail can be achieved with the command
1753 make DESTDIR=@var{path-to-rootdir} install
1756 @noindent where @var{path-to-rootdir} is the absolute path of
1757 a directory relative to which all installation paths will be
1758 interpreted. Note that the directory specified by @code{DESTDIR}
1759 need not exist yet; it will be created if necessary.
1761 There is a subtle point with tooldirs and @code{DESTDIR}:
1762 If you relocate a cross-compiler installation with
1763 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
1764 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
1765 be filled with duplicated GCC executables only if it already exists,
1766 it will not be created otherwise. This is regarded as a feature,
1767 not as a bug, because it gives slightly more control to the packagers
1768 using the @code{DESTDIR} feature.
1770 If you built a released version of GCC using @samp{make bootstrap} then please
1771 quickly review the build status page for your release, available from
1772 @uref{http://gcc.gnu.org/buildstat.html}.
1773 If your system is not listed for the version of GCC that you built,
1775 @email{gcc@@gcc.gnu.org} indicating
1776 that you successfully built and installed GCC.
1777 Include the following information:
1781 Output from running @file{@var{srcdir}/config.guess}. Do not send us
1782 that file itself, just the one-line output from running it.
1785 The output of @samp{gcc -v} for your newly installed gcc.
1786 This tells us which version of GCC you built and the options you passed to
1790 Whether you enabled all languages or a subset of them. If you used a
1791 full distribution then this information is part of the configure
1792 options in the output of @samp{gcc -v}, but if you downloaded the
1793 ``core'' compiler plus additional front ends then it isn't apparent
1794 which ones you built unless you tell us about it.
1797 If the build was for GNU/Linux, also include:
1800 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
1801 this information should be available from @file{/etc/issue}.
1804 The version of the Linux kernel, available from @samp{uname --version}
1808 The version of glibc you used; for RPM-based systems like Red Hat,
1809 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
1810 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
1812 For other systems, you can include similar information if you think it is
1816 Any other information that you think would be useful to people building
1817 GCC on the same configuration. The new entry in the build status list
1818 will include a link to the archived copy of your message.
1821 We'd also like to know if the
1823 @ref{Specific, host/target specific installation notes}
1826 @uref{specific.html,,host/target specific installation notes}
1828 didn't include your host/target information or if that information is
1829 incomplete or out of date. Send a note to
1830 @email{gcc@@gcc.gnu.org} telling us how the information should be changed.
1832 If you find a bug, please report it following our
1833 @uref{../bugs.html,,bug reporting guidelines}.
1835 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
1836 dvi}. You will need to have @command{texi2dvi} (version at least 4.2)
1837 and @TeX{} installed. This creates a number of @file{.dvi} files in
1838 subdirectories of @file{@var{objdir}}; these may be converted for
1839 printing with programs such as @command{dvips}. You can also
1840 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
1841 Free Software Foundation}, though such manuals may not be for the most
1842 recent version of GCC@.
1849 @uref{./index.html,,Return to the GCC Installation page}
1853 @c ***Binaries****************************************************************
1855 @comment node-name, next, previous, up
1856 @node Binaries, Specific, Installing GCC, Top
1860 @chapter Installing GCC: Binaries
1863 @cindex Installing GCC: Binaries
1865 We are often asked about pre-compiled versions of GCC@. While we cannot
1866 provide these for all platforms, below you'll find links to binaries for
1867 various platforms where creating them by yourself is not easy due to various
1870 Please note that we did not create these binaries, nor do we
1871 support them. If you have any problems installing them, please
1872 contact their makers.
1879 @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX};
1882 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX}.
1886 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
1889 Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
1890 Development Tools for the Renesas H8/300[HS] Series}.
1896 @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center};
1899 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
1903 Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU
1904 Development Tools for the Motorola 68HC11/68HC12}.
1907 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
1908 OpenServer/Unixware}.
1911 Sinix/Reliant Unix---@uref{ftp://ftp.fujitsu-siemens.com/pub/pd/gnu/gcc/,,Siemens}.
1914 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}.
1917 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}.
1923 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
1925 The @uref{http://www.mingw.org/,,MinGW} project.
1929 @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The
1930 Written Word} offers binaries for
1933 Digital UNIX 4.0D and 5.1,
1935 HP-UX 10.20, 11.00, and 11.11, and
1936 Solaris/SPARC 2.5.1, 2.6, 2.7, 8, and 9,
1939 In addition to those specific offerings, you can get a binary
1940 distribution CD-ROM from the
1941 @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}.
1942 It contains binaries for a number of platforms, and
1943 includes not only GCC, but other stuff as well. The current CD does
1944 not contain the latest version of GCC, but it should allow
1945 bootstrapping the compiler. An updated version of that disk is in the
1953 @uref{./index.html,,Return to the GCC Installation page}
1957 @c ***Specific****************************************************************
1959 @comment node-name, next, previous, up
1960 @node Specific, Old, Binaries, Top
1964 @chapter Host/target specific installation notes for GCC
1967 @cindex Specific installation notes
1968 @cindex Target specific installation
1969 @cindex Host specific installation
1970 @cindex Target specific installation notes
1972 Please read this document carefully @emph{before} installing the
1973 GNU Compiler Collection on your machine.
1978 @uref{#alpha*-*-*,,alpha*-*-*}
1980 @uref{#alpha*-dec-osf*,,alpha*-dec-osf*}
1982 @uref{#alphaev5-cray-unicosmk*,,alphaev5-cray-unicosmk*}
1984 @uref{#arc-*-elf,,arc-*-elf}
1986 @uref{#arm-*-elf,,arm-*-elf}
1987 @uref{#arm-*-coff,,arm-*-coff}
1988 @uref{#arm-*-aout,,arm-*-aout}
1990 @uref{#xscale-*-*,,xscale-*-*}
1998 @uref{#dsp16xx,,dsp16xx}
2000 @uref{#*-*-freebsd*,,*-*-freebsd*}
2002 @uref{#h8300-hms,,h8300-hms}
2004 @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*}
2006 @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10}
2008 @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11}
2010 @uref{#i370-*-*,,i370-*-*}
2012 @uref{#*-*-linux-gnu,,*-*-linux-gnu}
2014 @uref{#ix86-*-linux*aout,,i?86-*-linux*aout}
2016 @uref{#ix86-*-linux*,,i?86-*-linux*}
2018 @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*}
2020 @uref{#ix86-*-udk,,i?86-*-udk}
2022 @uref{#ix86-*-esix,,i?86-*-esix}
2024 @uref{#ia64-*-linux,,ia64-*-linux}
2026 @uref{#ia64-*-hpux*,,ia64-*-hpux*}
2028 @uref{#*-ibm-aix*,,*-ibm-aix*}
2030 @uref{#ip2k-*-elf,,ip2k-*-elf}
2032 @uref{#iq2000-*-elf,,iq2000-*-elf}
2034 @uref{#m32r-*-elf,,m32r-*-elf}
2036 @uref{#m6811-elf,,m6811-elf}
2038 @uref{#m6812-elf,,m6812-elf}
2040 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
2042 @uref{#mips-*-*,,mips-*-*}
2044 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
2046 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
2048 @uref{#powerpc*-*-*,,powerpc*-*-*, powerpc-*-sysv4}
2050 @uref{#powerpc-*-darwin*,,powerpc-*-darwin*}
2052 @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4}
2054 @uref{#powerpc-*-linux-gnu*,,powerpc-*-linux-gnu*}
2056 @uref{#powerpc-*-netbsd*,,powerpc-*-netbsd*}
2058 @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix}
2060 @uref{#powerpc-*-eabisim,,powerpc-*-eabisim}
2062 @uref{#powerpc-*-eabi,,powerpc-*-eabi}
2064 @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
2066 @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim}
2068 @uref{#powerpcle-*-eabi,,powerpcle-*-eabi}
2070 @uref{#s390-*-linux*,,s390-*-linux*}
2072 @uref{#s390x-*-linux*,,s390x-*-linux*}
2074 @uref{#s390x-ibm-tpf*,,s390x-ibm-tpf*}
2076 @uref{#*-*-solaris2*,,*-*-solaris2*}
2078 @uref{#sparc-sun-solaris2*,,sparc-sun-solaris2*}
2080 @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7}
2082 @uref{#sparc-*-linux*,,sparc-*-linux*}
2084 @uref{#sparc64-*-solaris2*,,sparc64-*-solaris2*}
2086 @uref{#sparcv9-*-solaris2*,,sparcv9-*-solaris2*}
2088 @uref{#*-*-sysv*,,*-*-sysv*}
2090 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
2092 @uref{#*-*-vxworks*,,*-*-vxworks*}
2094 @uref{#xtensa-*-elf,,xtensa-*-elf}
2096 @uref{#xtensa-*-linux*,,xtensa-*-linux*}
2098 @uref{#windows,,Microsoft Windows}
2102 @uref{#older,,Older systems}
2107 @uref{#elf_targets,,all ELF targets} (SVR4, Solaris 2, etc.)
2113 <!-- -------- host/target specific issues start here ---------------- -->
2116 @heading @anchor{alpha*-*-*}alpha*-*-*
2118 This section contains general configuration information for all
2119 alpha-based platforms using ELF (in particular, ignore this section for
2120 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
2121 section, please read all other sections that match your target.
2123 We require binutils 2.11.2 or newer.
2124 Previous binutils releases had a number of problems with DWARF 2
2125 debugging information, not the least of which is incorrect linking of
2131 @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf*
2132 Systems using processors that implement the DEC Alpha architecture and
2133 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
2134 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
2136 As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer
2137 supported. (These are the versions which identify themselves as DEC
2140 In Digital Unix V4.0, virtual memory exhausted bootstrap failures
2141 may be fixed by configuring with @option{--with-gc=simple},
2142 reconfiguring Kernel Virtual Memory and Swap parameters
2143 per the @command{/usr/sbin/sys_check} Tuning Suggestions,
2144 or applying the patch in
2145 @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}.
2147 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
2148 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
2149 we need to use the old assembler, invoked via the barely documented
2150 @option{-oldas} option. To bootstrap GCC, you either need to use the
2154 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
2157 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
2160 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
2163 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
2164 are supported on Tru64 UNIX, so you must not configure GCC with
2165 @option{--with-gnu-as} or @option{--with-gnu-ld}.
2167 GCC writes a @samp{.verstamp} directive to the assembler output file
2168 unless it is built as a cross-compiler. It gets the version to use from
2169 the system header file @file{/usr/include/stamp.h}. If you install a
2170 new version of DEC Unix, you should rebuild GCC to pick up the new version
2173 Note that since the Alpha is a 64-bit architecture, cross-compilers from
2174 32-bit machines will not generate code as efficient as that generated
2175 when the compiler is running on a 64-bit machine because many
2176 optimizations that depend on being able to represent a word on the
2177 target in an integral value on the host cannot be performed. Building
2178 cross-compilers on the Alpha for 32-bit machines has only been tested in
2179 a few cases and may not work properly.
2181 @samp{make compare} may fail on old versions of DEC Unix unless you add
2182 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2183 assembler input file is stored in the object file, and that makes
2184 comparison fail if it differs between the @code{stage1} and
2185 @code{stage2} compilations. The option @option{-save-temps} forces a
2186 fixed name to be used for the assembler input file, instead of a
2187 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2188 unless the comparisons fail without that option. If you add
2189 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2190 @samp{.s} files after each series of compilations.
2192 GCC now supports both the native (ECOFF) debugging format used by DBX
2193 and GDB and an encapsulated STABS format for use only with GDB@. See the
2194 discussion of the @option{--with-stabs} option of @file{configure} above
2195 for more information on these formats and how to select them.
2197 There is a bug in DEC's assembler that produces incorrect line numbers
2198 for ECOFF format when the @samp{.align} directive is used. To work
2199 around this problem, GCC will not emit such alignment directives
2200 while writing ECOFF format debugging information even if optimization is
2201 being performed. Unfortunately, this has the very undesirable
2202 side-effect that code addresses when @option{-O} is specified are
2203 different depending on whether or not @option{-g} is also specified.
2205 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
2206 DBX@. DEC is now aware of this problem with the assembler and hopes to
2207 provide a fix shortly.
2212 @heading @anchor{alphaev5-cray-unicosmk*}alphaev5-cray-unicosmk*
2213 Cray T3E systems running Unicos/Mk.
2215 This port is incomplete and has many known bugs. We hope to improve the
2216 support for this target soon. Currently, only the C front end is supported,
2217 and it is not possible to build parallel applications. Cray modules are not
2218 supported; in particular, Craylibs are assumed to be in
2219 @file{/opt/ctl/craylibs/craylibs}.
2221 You absolutely @strong{must} use GNU make on this platform. Also, you
2222 need to tell GCC where to find the assembler and the linker. The
2223 simplest way to do so is by providing @option{--with-as} and
2224 @option{--with-ld} to @file{configure}, e.g.@:
2227 configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld \
2228 --enable-languages=c
2231 The comparison test during @samp{make bootstrap} fails on Unicos/Mk
2232 because the assembler inserts timestamps into object files. You should
2233 be able to work around this by doing @samp{make all} after getting this
2239 @heading @anchor{arc-*-elf}arc-*-elf
2240 Argonaut ARC processor.
2241 This configuration is intended for embedded systems.
2246 @heading @anchor{arm-*-elf}arm-*-elf
2247 @heading @anchor{xscale-*-*}xscale-*-*
2248 ARM-family processors. Subtargets that use the ELF object format
2249 require GNU binutils 2.13 or newer. Such subtargets include:
2250 @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux},
2251 @code{arm-*-rtems} and @code{arm-*-kaos}.
2256 @heading @anchor{arm-*-coff}arm-*-coff
2257 ARM-family processors. Note that there are two different varieties
2258 of PE format subtarget supported: @code{arm-wince-pe} and
2259 @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}.
2264 @heading @anchor{arm-*-aout}arm-*-aout
2265 ARM-family processors. These targets support the AOUT file format:
2266 @code{arm-*-aout}, @code{arm-*-netbsd}.
2271 @heading @anchor{avr}avr
2273 ATMEL AVR-family micro controllers. These are used in embedded
2274 applications. There are no standard Unix configurations.
2276 @xref{AVR Options,, AVR Options, gcc, Using and Porting the GNU Compiler
2280 See ``AVR Options'' in the main manual
2282 for the list of supported MCU types.
2284 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
2286 Further installation notes and other useful information about AVR tools
2287 can also be obtained from:
2291 @uref{http://www.openavr.org,,http://www.openavr.org}
2293 @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/}
2295 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
2298 We @emph{strongly} recommend using binutils 2.13 or newer.
2300 The following error:
2302 Error: register required
2305 indicates that you should upgrade to a newer version of the binutils.
2310 @heading @anchor{c4x}c4x
2312 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
2313 Processors. These are used in embedded applications. There are no
2314 standard Unix configurations.
2316 @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using and
2317 Porting the GNU Compiler Collection (GCC)},
2320 See ``TMS320C3x/C4x Options'' in the main manual
2322 for the list of supported MCU types.
2324 GCC can be configured as a cross compiler for both the C3x and C4x
2325 architectures on the same system. Use @samp{configure --target=c4x
2326 --enable-languages="c,c++"} to configure.
2329 Further installation notes and other useful information about C4x tools
2330 can also be obtained from:
2334 @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/}
2340 @heading @anchor{cris}CRIS
2342 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
2343 series. These are used in embedded applications.
2346 @xref{CRIS Options,, CRIS Options, gcc, Using and Porting the GNU Compiler
2350 See ``CRIS Options'' in the main manual
2352 for a list of CRIS-specific options.
2354 There are a few different CRIS targets:
2356 @item cris-axis-aout
2357 Old target. Includes a multilib for the @samp{elinux} a.out-based
2358 target. No multilibs for newer architecture variants.
2360 Mainly for monolithic embedded systems. Includes a multilib for the
2361 @samp{v10} core used in @samp{ETRAX 100 LX}.
2362 @item cris-axis-linux-gnu
2363 A GNU/Linux port for the CRIS architecture, currently targeting
2364 @samp{ETRAX 100 LX} by default.
2367 For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11
2368 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
2370 Pre-packaged tools can be obtained from
2371 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
2372 information about this platform is available at
2373 @uref{http://developer.axis.com/}.
2378 @heading @anchor{dos}DOS
2380 Please have a look at our @uref{binaries.html,,binaries page}.
2382 You cannot install GCC by itself on MSDOS; it will not compile under
2383 any MSDOS compiler except itself. You need to get the complete
2384 compilation package DJGPP, which includes binaries as well as sources,
2385 and includes all the necessary compilation tools and libraries.
2390 @heading @anchor{dsp16xx}dsp16xx
2391 A port to the AT&T DSP1610 family of processors.
2396 @heading @anchor{*-*-freebsd*}*-*-freebsd*
2398 The version of binutils installed in @file{/usr/bin} is known to work unless
2399 otherwise specified in any per-architecture notes. However, binutils
2400 2.12.1 or greater is known to improve overall testsuite results.
2402 Support for FreeBSD 1 was discontinued in GCC 3.2.
2404 For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
2405 configuration support and files as shipped with GCC 2.95 are still in
2406 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
2407 it is unknown which version of binutils was used (it is assumed that it
2408 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
2410 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
2411 default for all CPU architectures. It had been the default on
2412 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
2413 of @option{-g}, if you really want the old debugging format. There are
2414 no known issues with mixing object files and libraries with different
2415 debugging formats. Otherwise, this release of GCC should now match more
2416 of the configuration used in the stock FreeBSD configuration of GCC. In
2417 particular, @option{--enable-threads} is now configured by default.
2418 However, as a general user, do not attempt to replace the system
2419 compiler with this release. Known to bootstrap and check with good
2420 results on FreeBSD 4.8-STABLE and 5-CURRENT@. In the past, known to
2421 bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2,
2422 4.3, 4.4, 4.5-STABLE@.
2424 In principle, @option{--enable-threads} is now compatible with
2425 @option{--enable-libgcj} on FreeBSD@. However, it has only been built
2426 and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}.
2428 library may be incorrectly built (symbols are missing at link time).
2429 There is a rare timing-based startup hang (probably involves an
2430 assumption about the thread library). Multi-threaded boehm-gc (required for
2431 libjava) exposes severe threaded signal-handling bugs on FreeBSD before
2432 4.5-RELEASE@. Other CPU architectures
2433 supported by FreeBSD will require additional configuration tuning in, at
2434 the very least, both boehm-gc and libffi.
2436 Shared @file{libgcc_s.so} is now built and installed by default.
2441 @heading @anchor{h8300-hms}h8300-hms
2442 Renesas H8/300 series of processors.
2444 Please have a look at our @uref{binaries.html,,binaries page}.
2446 The calling convention and structure layout has changed in release 2.6.
2447 All code must be recompiled. The calling convention now passes the
2448 first three arguments in function calls in registers. Structures are no
2449 longer a multiple of 2 bytes.
2454 @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux*
2455 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
2457 We @emph{highly} recommend using gas/binutils 2.8 or newer on all hppa
2458 platforms; you may encounter a variety of problems when using the HP
2461 Specifically, @option{-g} does not work on HP-UX (since that system
2462 uses a peculiar debugging format which GCC does not know about), unless you
2463 use GAS and GDB and configure GCC with the
2464 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
2465 @option{--with-as=@dots{}} options.
2467 If you wish to use the pa-risc 2.0 architecture support with a 32-bit
2468 runtime, you must use either the HP assembler, gas/binutils 2.11 or newer,
2470 @uref{ftp://sources.redhat.com/pub/binutils/snapshots,,snapshot of gas}.
2472 There are two default scheduling models for instructions. These are
2473 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
2474 architecture specified for the target machine when configuring.
2475 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
2476 the target is a @samp{hppa1*} machine.
2478 The PROCESSOR_8000 model is not well suited to older processors. Thus,
2479 it is important to completely specify the machine architecture when
2480 configuring if you want a model other than PROCESSOR_8000. The macro
2481 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
2482 default scheduling model is desired.
2484 More specific information to @samp{hppa*-hp-hpux*} targets follows.
2489 @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10
2491 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
2492 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
2498 <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and
2502 @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific,
2506 @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe.
2509 The HP assembler on these systems has some problems. Most notably the
2510 assembler inserts timestamps into each object file it creates, causing
2511 the 3-stage comparison test to fail during a @samp{make bootstrap}.
2512 You should be able to continue by saying @samp{make all} after getting
2513 the failure from @samp{make bootstrap}.
2519 @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11
2521 GCC 3.0 and up support HP-UX 11. On 64-bit capable systems, there
2522 are two distinct ports. The @samp{hppa2.0w-hp-hpux11*} port generates
2523 code for the 32-bit pa-risc runtime architecture. It uses the HP
2524 linker. The @samp{hppa64-hp-hpux11*} port generates 64-bit code for the
2525 pa-risc 2.0 architecture. The script config.guess now selects the port
2526 type based on the type compiler detected during configuration. You must
2527 set your @env{PATH} or define @env{CC} so that configure finds an appropriate
2528 compiler for the initial bootstrap. Different prefixes must be used if
2529 both ports are to be installed on the same system.
2531 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
2532 with the @option{--with-ld=@dots{}} option. We support both the HP
2533 and GNU linkers for this target. The two linkers require different
2534 link commands. Thus, it's not possible to switch linkers during a
2535 GCC build. This has been been reported to occur in a unified build
2536 of binutils and GCC.
2538 GCC 2.95.x is not supported under HP-UX 11 and cannot be used to
2539 compile GCC 3.0 and up. Refer to @uref{binaries.html,,binaries} for
2540 information about obtaining precompiled GCC binaries for HP-UX.
2542 You must use GNU binutils 2.11 or above with the 32-bit port. Thread
2543 support is not currently implemented, so @option{--enable-threads} does
2547 @item @uref{http://gcc.gnu.org/ml/gcc-prs/2002-01/msg00551.html}
2548 @item @uref{http://gcc.gnu.org/ml/gcc-bugs/2002-01/msg00663.html}
2551 GCC 3.3 and later support weak symbols on the 32-bit port using SOM
2552 secondary definition symbols. This feature is not enabled for earlier
2553 versions of HP-UX since there have been bugs in the linker support for
2554 secondary symbols. The HP linker patches @code{PHSS_26559} and
2555 @code{PHSS_24304} for HP-UX 11.00 and 11.11, respectively, correct the
2556 problem of linker core dumps creating C++ libraries. Earlier patches
2557 may work but they have not been tested.
2559 GCC 3.3 nows uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capability
2560 to run initializers and finalizers on the 64-bit port. The feature
2561 requires CVS binutils as of January 2, 2003, or a subsequent release
2562 to correct a problem arising from HP's non-standard use of the .init
2563 and .fini sections. The 32-bit port uses the linker @option{+init}
2564 and @option{+fini} options. As with the support for secondary symbols,
2565 there have been bugs in the order in which these options are executed
2566 by the HP linker. So, again a recent linker patch is recommended.
2568 The HP assembler has many limitations and is not recommended for either
2569 the 32 or 64-bit ports. For example, it does not support weak symbols
2570 or alias definitions. As a result, explicit template instantiations
2571 are required when using C++. This will make it difficult if not
2572 impossible to build many C++ applications. You also can't generate
2573 debugging information when using the HP assembler with GCC.
2575 There are a number of issues to consider in selecting which linker to
2576 use with the 64-bit port. The GNU 64-bit linker can only create dynamic
2577 binaries. The @option{-static} option causes linking with archive
2578 libraries but doesn't produce a truly static binary. Dynamic binaries
2579 still require final binding by the dynamic loader to resolve a set of
2580 dynamic-loader-defined symbols. The default behavior of the HP linker
2581 is the same as the GNU linker. However, it can generate true 64-bit
2582 static binaries using the @option{+compat} option.
2584 The HP 64-bit linker doesn't support linkonce semantics. As a
2585 result, C++ programs have many more sections than they should.
2587 The GNU 64-bit linker has some issues with shared library support
2588 and exceptions. As a result, we only support libgcc in archive
2589 format. For similar reasons, dwarf2 unwind and exception support
2590 are disabled. The GNU linker also has problems creating binaries
2591 with @option{-static}. It doesn't provide stubs for internal
2592 calls to global functions in shared libraries, so these calls
2593 can't be overloaded.
2595 There are several possible approaches to building the distribution.
2596 Binutils can be built first using the HP tools. Then, the GCC
2597 distribution can be built. The second approach is to build GCC
2598 first using the HP tools, then build binutils, then rebuild GCC.
2599 There have been problems with various binary distributions, so
2600 it is best not to start from a binary distribution.
2602 Starting with GCC 3.4 an ISO C compiler is required to bootstrap.
2603 The bundled compiler supports only traditional C; you will need
2604 either HP's unbundled compiler, or a binary distribution of GCC@.
2606 This port still is undergoing significant development.
2611 @heading @anchor{i370-*-*}i370-*-*
2612 This port is very preliminary and has many known bugs. We hope to
2613 have a higher-quality port for this machine soon.
2618 @heading @anchor{*-*-linux-gnu}*-*-linux-gnu
2620 Versions of libstdc++-v3 starting with 3.2.1 require bugfixes present
2621 in glibc 2.2.5 and later. More information is available in the
2622 libstdc++-v3 documentation.
2627 @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout
2628 Use this configuration to generate @file{a.out} binaries on Linux-based
2629 GNU systems. This configuration is being superseded.
2634 @heading @anchor{ix86-*-linux*}i?86-*-linux*
2636 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
2637 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
2639 If you receive Signal 11 errors when building on GNU/Linux, then it is
2640 possible you have a hardware problem. Further information on this can be
2641 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
2646 @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5*
2647 Use this for the SCO OpenServer Release 5 family of operating systems.
2649 Unlike earlier versions of GCC, the ability to generate COFF with this
2650 target is no longer provided.
2652 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
2653 the system debugger to be used. That support was too burdensome to
2654 maintain. GCC now emits only DWARF 2 for this target. This means you
2655 may use either the UDK debugger or GDB to debug programs built by this
2658 GCC is now only supported on releases 5.0.4 and later, and requires that
2659 you install Support Level Supplement OSS646B or later, and Support Level
2660 Supplement OSS631C or later. If you are using release 5.0.7 of
2661 OpenServer, you must have at least the first maintenance pack installed
2662 (this includes the relevant portions of OSS646). OSS646, also known as
2663 the "Execution Environment Update", provides updated link editors and
2664 assemblers, as well as updated standard C and math libraries. The C
2665 startup modules are also updated to support the System V gABI draft, and
2666 GCC relies on that behavior. OSS631 provides a collection of commonly
2667 used open source libraries, some of which GCC depends on (such as GNU
2668 gettext and zlib). SCO OpenServer Release 5.0.7 has all of this built
2669 in by default, but OSS631C and later also apply to that release. Please
2671 @uref{ftp://ftp.sco.com/pub/openserver5,,ftp://ftp.sco.com/pub/openserver5}
2672 for the latest versions of these (and other potentially useful)
2675 Although there is support for using the native assembler, it is
2676 recommended that you configure GCC to use the GNU assembler. You do
2677 this by using the flags
2678 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You should
2679 use a modern version of GNU binutils. Version 2.13.2.1 was used for all
2680 testing. In general, only the @option{--with-gnu-as} option is tested.
2681 A modern bintuils (as well as a plethora of other development related
2682 GNU utilities) can be found in Support Level Supplement OSS658A, the
2683 "GNU Development Tools" package. See the SCO web and ftp sites for details.
2684 That package also contains the currently "officially supported" version of
2685 GCC, version 2.95.3. It is useful for bootstrapping this version.
2690 @heading @anchor{ix86-*-udk}i?86-*-udk
2692 This target emulates the SCO Universal Development Kit and requires that
2693 package be installed. (If it is installed, you will have a
2694 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
2695 @samp{i?86-*-unixware7*} target
2696 but is meant to be used when hosting on a system where UDK isn't the
2697 default compiler such as OpenServer 5 or Unixware 2. This target will
2698 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
2699 with the same warnings and caveats as the SCO UDK@.
2701 This target is a little tricky to build because we have to distinguish
2702 it from the native tools (so it gets headers, startups, and libraries
2703 from the right place) while making the tools not think we're actually
2704 building a cross compiler. The easiest way to do this is with a configure
2708 CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure \
2709 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-
2712 @emph{You should substitute @samp{i686} in the above command with the appropriate
2713 processor for your host.}
2715 After the usual @samp{make bootstrap} and
2716 @samp{make install}, you can then access the UDK-targeted GCC
2717 tools by adding @command{udk-} before the commonly known name. For
2718 example, to invoke the C compiler, you would use @command{udk-gcc}.
2719 They will coexist peacefully with any native-target GCC tools you may
2726 @heading @anchor{ia64-*-linux}ia64-*-linux
2727 IA-64 processor (also known as IPF, or Itanium Processor Family)
2730 If you are using the optional libunwind library, then you must use
2731 libunwind 0.96 or later.
2733 None of the following versions of GCC has an ABI that is compatible
2734 with any of the other versions in this list, with the exception that
2735 Red Hat 2.96 and Trillian 000171 are compatible with each other:
2736 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
2737 This primarily affects C++ programs and programs that create shared libraries.
2738 GCC 3.1 or later is recommended for compiling linux, the kernel.
2739 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
2740 more major ABI changes are expected.
2745 @heading @anchor{ia64-*-hpux*}ia64-*-hpux*
2746 Building GCC on this target requires the GNU Assembler. The bundled HP
2747 assembler will not work. To prevent GCC from using the wrong assembler,
2748 the option @option{--with-gnu-as} may be necessary.
2750 The GCC libunwind library has not been ported to HPUX. This means that for
2751 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
2752 is required to build GCC. For GCC 3.3 and later, this is the default.
2756 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
2758 @heading @anchor{*-ibm-aix*}*-ibm-aix*
2759 Support for AIX version 3 and older was discontinued in GCC 3.4.
2761 AIX Make frequently has problems with GCC makefiles. GNU Make 3.79.1 or
2762 newer is recommended to build on this platform.
2764 Errors involving @code{alloca} when building GCC generally are due
2765 to an incorrect definition of @code{CC} in the Makefile or mixing files
2766 compiled with the native C compiler and GCC@. During the stage1 phase of
2767 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
2768 (not @command{xlc}). Once @command{configure} has been informed of
2769 @command{xlc}, one needs to use @samp{make distclean} to remove the
2770 configure cache files and ensure that @env{CC} environment variable
2771 does not provide a definition that will confuse @command{configure}.
2772 If this error occurs during stage2 or later, then the problem most likely
2773 is the version of Make (see above).
2775 The native @command{as} and @command{ld} are recommended for bootstrapping
2776 on AIX 4 and required for bootstrapping on AIX 5L. The GNU Assembler
2777 reports that it supports WEAK symbols on AIX 4, which causes GCC to try to
2778 utilize weak symbol functionality although it is not supported. The GNU
2779 Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC.
2780 The native AIX tools do interoperate with GCC@.
2782 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
2783 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1).
2785 @samp{libstdc++} in GCC 3.2 increments the major version number of the
2786 shared object and GCC installation places the @file{libstdc++.a}
2787 shared library in a common location which will overwrite the GCC 3.1
2788 version of the shared library. Applications either need to be
2789 re-linked against the new shared library or the GCC 3.1 version of the
2790 @samp{libstdc++} shared object needs to be available to the AIX
2791 runtime loader. The GCC 3.1 @samp{libstdc++.so.4} shared object can
2792 be installed for runtime dynamic loading using the following steps to
2793 set the @samp{F_LOADONLY} flag in the shared object for @emph{each}
2794 multilib @file{libstdc++.a} installed:
2796 Extract the shared object from each the GCC 3.1 @file{libstdc++.a}
2799 % ar -x libstdc++.a libstdc++.so.4
2802 Enable the @samp{F_LOADONLY} flag so that the shared object will be
2803 available for runtime dynamic loading, but not linking:
2805 % strip -e libstdc++.so.4
2808 Archive the runtime-only shared object in the GCC 3.2
2809 @file{libstdc++.a} archive:
2811 % ar -q libstdc++.a libstdc++.so.4
2814 Linking executables and shared libraries may produce warnings of
2815 duplicate symbols. The assembly files generated by GCC for AIX always
2816 have included multiple symbol definitions for certain global variable
2817 and function declarations in the original program. The warnings should
2818 not prevent the linker from producing a correct library or runnable
2821 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
2822 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
2823 to parse archive libraries did not handle the new format correctly.
2824 These routines are used by GCC and result in error messages during
2825 linking such as ``not a COFF file''. The version of the routines shipped
2826 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
2827 option of the archive command may be used to create archives of 32-bit
2828 objects using the original ``small format''. A correct version of the
2829 routines is shipped with AIX 4.3.2 and above.
2831 Some versions of the AIX binder (linker) can fail with a relocation
2832 overflow severe error when the @option{-bbigtoc} option is used to link
2833 GCC-produced object files into an executable that overflows the TOC@. A fix
2834 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
2835 available from IBM Customer Support and from its
2836 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2837 website as PTF U455193.
2839 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
2840 with a segmentation fault when invoked by any version of GCC@. A fix for
2841 APAR IX87327 is available from IBM Customer Support and from its
2842 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2843 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
2845 The initial assembler shipped with AIX 4.3.0 generates incorrect object
2846 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
2847 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
2848 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2849 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
2851 AIX provides National Language Support (NLS)@. Compilers and assemblers
2852 use NLS to support locale-specific representations of various data
2853 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
2854 separating decimal fractions). There have been problems reported where
2855 GCC does not produce the same floating-point formats that the assembler
2856 expects. If one encounters this problem, set the @env{LANG}
2857 environment variable to @samp{C} or @samp{En_US}.
2859 By default, GCC for AIX 4.1 and above produces code that can be used on
2860 both Power or PowerPC processors.
2862 A default can be specified with the @option{-mcpu=@var{cpu_type}}
2863 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
2868 @heading @anchor{ip2k-*-elf}ip2k-*-elf
2869 Ubicom IP2022 micro controller.
2870 This configuration is intended for embedded systems.
2871 There are no standard Unix configurations.
2873 Use @samp{configure --target=ip2k-elf --enable-languages=c} to configure GCC@.
2878 @heading @anchor{iq2000-*-elf}iq2000-*-elf
2879 Vitesse IQ2000 processors. These are used in embedded
2880 applications. There are no standard Unix configurations.
2885 @heading @anchor{m32r-*-elf}m32r-*-elf
2886 Renesas M32R processor.
2887 This configuration is intended for embedded systems.
2892 @heading @anchor{m6811-elf}m6811-elf
2893 Motorola 68HC11 family micro controllers. These are used in embedded
2894 applications. There are no standard Unix configurations.
2899 @heading @anchor{m6812-elf}m6812-elf
2900 Motorola 68HC12 family micro controllers. These are used in embedded
2901 applications. There are no standard Unix configurations.
2906 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
2907 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
2908 the assembler that prevents compilation of GCC@. This
2909 bug manifests itself during the first stage of compilation, while
2910 building @file{libgcc2.a}:
2914 cc1: warning: `-g' option not supported on this version of GCC
2915 cc1: warning: `-g1' option not supported on this version of GCC
2916 ./xgcc: Internal compiler error: program as got fatal signal 11
2919 A patched version of the assembler is available as the file
2920 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
2921 have HP software support, the patch can also be obtained directly from
2922 HP, as described in the following note:
2925 This is the patched assembler, to patch SR#1653-010439, where the
2926 assembler aborts on floating point constants.
2928 The bug is not really in the assembler, but in the shared library
2929 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
2930 SR#4701-078451. Anyway, the attached assembler uses the archive
2931 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
2934 This patch is also known as PHCO_4484.
2936 In addition gdb does not understand that native HP-UX format, so
2937 you must use gas if you wish to use gdb.
2939 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
2940 @command{fixproto} shell script triggers a bug in the system shell. If you
2941 encounter this problem, upgrade your operating system or use BASH (the
2942 GNU shell) to run @command{fixproto}. This bug will cause the fixproto
2943 program to report an error of the form:
2946 ./fixproto: sh internal 1K buffer overflow
2949 To fix this, you can also change the first line of the fixproto script
2959 @heading @anchor{mips-*-*}mips-*-*
2960 If on a MIPS system you get an error message saying ``does not have gp
2961 sections for all it's [sic] sectons [sic]'', don't worry about it. This
2962 happens whenever you use GAS with the MIPS linker, but there is not
2963 really anything wrong, and it is okay to use the output file. You can
2964 stop such warnings by installing the GNU linker.
2966 It would be nice to extend GAS to produce the gp tables, but they are
2967 optional, and there should not be a warning about their absence.
2969 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
2970 and later. A patch went in just after the GCC 3.3 release to
2971 make @samp{mips*-*-*} use the generic implementation instead. You can also
2972 configure for @samp{mipsel-elf} as a workaround. The
2973 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
2974 work on this is expected in future releases.
2976 Cross-compilers for the Mips as target using the Mips assembler
2977 currently do not work, because the auxiliary programs
2978 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
2979 anything but a Mips. It does work to cross compile for a Mips
2980 if you use the GNU assembler and linker.
2985 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
2987 This configuration has considerable problems, which will be fixed in a
2990 In order to compile GCC on an SGI running IRIX 5, the ``compiler_dev.hdr''
2991 subsystem must be installed from the IDO CD-ROM supplied by Silicon
2992 Graphics. It is also available for download from
2993 @uref{http://www.sgi.com/developers/devtools/apis/ido.html,,http://www.sgi.com/developers/devtools/apis/ido.html}.
2995 @samp{make compare} may fail on version 5 of IRIX unless you add
2996 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2997 assembler input file is stored in the object file, and that makes
2998 comparison fail if it differs between the @code{stage1} and
2999 @code{stage2} compilations. The option @option{-save-temps} forces a
3000 fixed name to be used for the assembler input file, instead of a
3001 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
3002 unless the comparisons fail without that option. If you do you
3003 @option{-save-temps}, you will have to manually delete the @samp{.i} and
3004 @samp{.s} files after each series of compilations.
3006 If you use the MIPS C compiler to bootstrap, it may be necessary
3007 to increase its table size for switch statements with the
3008 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
3009 optimization option, you also need to use @option{-Olimit 3000}.
3011 To enable debugging under IRIX 5, you must use GNU @command{as} 2.11.2
3013 and use the @option{--with-gnu-as} configure option when configuring GCC.
3014 GNU @command{as} is distributed as part of the binutils package.
3015 When using release 2.11.2, you need to apply a patch
3016 @uref{http://sources.redhat.com/ml/binutils/2001-07/msg00352.html,,http://sources.redhat.com/ml/binutils/2001-07/msg00352.html}
3017 which will be included in the next release of binutils.
3019 When building GCC, the build process loops rebuilding @command{cc1} over
3020 and over again. This happens on @samp{mips-sgi-irix5.2}, and possibly
3021 other platforms. It has been reported that this is a known bug in the
3022 @command{make} shipped with IRIX 5.2. We recommend you use GNU
3023 @command{make} instead of the vendor supplied @command{make} program;
3024 however, you may have success with @command{smake} on IRIX 5.2 if you do
3025 not have GNU @command{make} available.
3030 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
3032 If you are using IRIX @command{cc} as your bootstrap compiler, you must
3033 ensure that the N32 ABI is in use. To test this, compile a simple C
3034 file with @command{cc} and then run @command{file} on the
3035 resulting object file. The output should look like:
3038 test.o: ELF N32 MSB @dots{}
3044 test.o: ELF 32-bit MSB @dots{}
3050 test.o: ELF 64-bit MSB @dots{}
3053 then your version of @command{cc} uses the O32 or N64 ABI by default. You
3054 should set the environment variable @env{CC} to @samp{cc -n32}
3055 before configuring GCC@.
3057 If you want the resulting @command{gcc} to run on old 32-bit systems
3058 with the MIPS R4400 CPU, you need to ensure that only code for the mips3
3059 instruction set architecture (ISA) is generated. While GCC 3.x does
3060 this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change
3061 the ISA depending on the machine where GCC is built. Using one of them
3062 as the bootstrap compiler may result in mips4 code, which won't run at
3063 all on mips3-only systems. For the test program above, you should see:
3066 test.o: ELF N32 MSB mips-3 @dots{}
3072 test.o: ELF N32 MSB mips-4 @dots{}
3075 instead, you should set the environment variable @env{CC} to @samp{cc
3076 -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@.
3078 GCC on IRIX 6 is usually built to support both the N32 and N64 ABIs. If
3079 you build GCC on a system that doesn't have the N64 libraries installed,
3080 you need to configure with @option{--disable-multilib} so GCC doesn't
3081 try to use them. Look for @file{/usr/lib64/libc.so.1} to see if you
3082 have the 64-bit libraries installed.
3084 You must @emph{not} use GNU @command{as} (which isn't built anyway as of
3085 binutils 2.11.2) on IRIX 6 platforms; doing so will only cause problems.
3087 GCC does not currently support generating O32 ABI binaries in the
3088 @samp{mips-sgi-irix6} configurations. It is possible to create a GCC
3089 with O32 ABI only support by configuring it for the @samp{mips-sgi-irix5}
3090 target and using a patched GNU @command{as} 2.11.2 as documented in the
3091 @uref{#mips-sgi-irix5,,@samp{mips-sgi-irix5}} section above. Using the
3092 native assembler requires patches to GCC which will be included in a
3093 future release. It is
3094 expected that O32 ABI support will be available again in a future release.
3096 The @option{--enable-threads} option doesn't currently work, a patch is
3097 in preparation for a future release. The @option{--enable-libgcj}
3098 option is disabled by default: IRIX 6 uses a very low default limit
3099 (20480) for the command line length. Although libtool contains a
3100 workaround for this problem, at least the N64 @samp{libgcj} is known not
3101 to build despite this, running into an internal error of the native
3102 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
3103 its maximum of 262144 bytes. If you have root access, you can use the
3104 @command{systune} command to do this.
3106 GCC does not correctly pass/return structures which are
3107 smaller than 16 bytes and which are not 8 bytes. The problem is very
3108 involved and difficult to fix. It affects a number of other targets also,
3109 but IRIX 6 is affected the most, because it is a 64-bit target, and 4 byte
3110 structures are common. The exact problem is that structures are being padded
3111 at the wrong end, e.g.@: a 4 byte structure is loaded into the lower 4 bytes
3112 of the register when it should be loaded into the upper 4 bytes of the
3115 GCC is consistent with itself, but not consistent with the SGI C compiler
3116 (and the SGI supplied runtime libraries), so the only failures that can
3117 happen are when there are library functions that take/return such
3118 structures. There are very few such library functions. Currently this
3119 is known to affect @code{inet_ntoa}, @code{inet_lnaof},
3120 @code{inet_netof}, @code{inet_makeaddr}, and @code{semctl}. Until the
3121 bug is fixed, GCC contains workarounds for the known affected functions.
3123 See @uref{http://freeware.sgi.com/,,http://freeware.sgi.com/} for more
3124 information about using GCC on IRIX platforms.
3129 @heading @anchor{powerpc*-*-*}powerpc-*-*
3131 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3132 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3137 @heading @anchor{powerpc-*-darwin*}powerpc-*-darwin*
3138 PowerPC running Darwin (Mac OS X kernel).
3140 Pre-installed versions of Mac OS X may not include any developer tools,
3141 meaning that you will not be able to build GCC from source. Tool
3142 binaries are available at
3143 @uref{http://developer.apple.com/tools/compilers.html} (free
3144 registration required).
3146 The default stack limit of 512K is too small, which may cause compiles
3147 to fail with 'Bus error'. Set the stack larger, for instance
3148 by doing @samp{limit stack 800}. It's a good idea to use the GNU
3149 preprocessor instead of Apple's @file{cpp-precomp} during the first stage of
3150 bootstrapping; this is automatic when doing @samp{make bootstrap}, but
3151 to do it from the toplevel objdir you will need to say @samp{make
3152 CC='cc -no-cpp-precomp' bootstrap}.
3154 The version of GCC shipped by Apple typically includes a number of
3155 extensions not available in a standard GCC release. These extensions
3156 are generally specific to Mac programming.
3161 @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4
3162 PowerPC system in big endian mode, running System V.4.
3167 @heading @anchor{powerpc-*-linux-gnu*}powerpc-*-linux-gnu*
3170 @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.13.90.0.10}
3171 or newer for a working GCC@.
3176 @heading @anchor{powerpc-*-netbsd*}powerpc-*-netbsd*
3177 PowerPC system in big endian mode running NetBSD@. To build the
3178 documentation you will need Texinfo version 4.2 (NetBSD 1.5.1 included
3179 Texinfo version 3.12).
3184 @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim
3185 Embedded PowerPC system in big endian mode for use in running under the
3191 @heading @anchor{powerpc-*-eabi}powerpc-*-eabi
3192 Embedded PowerPC system in big endian mode.
3197 @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4
3198 PowerPC system in little endian mode, running System V.4.
3203 @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim
3204 Embedded PowerPC system in little endian mode for use in running under
3210 @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi
3211 Embedded PowerPC system in little endian mode.
3216 @heading @anchor{s390-*-linux*}s390-*-linux*
3217 S/390 system running Linux for S/390@.
3222 @heading @anchor{s390x-*-linux*}s390x-*-linux*
3223 zSeries system (64-bit) running Linux for zSeries@.
3228 @heading @anchor{s390x-ibm-tpf*}s390x-ibm-tpf*
3229 zSeries system (64-bit) running TPF. This platform is
3230 supported as cross-compilation target only.
3235 @c Please use Solaris 2 to refer to all release of Solaris, starting
3236 @c with 2.0 until 2.6, 7, and 8. Solaris 1 was a marketing name for
3237 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3238 @c alone is too unspecific and must be avoided.
3239 @heading @anchor{*-*-solaris2*}*-*-solaris2*
3241 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3242 GCC you first have to install a pre-built compiler, see our
3243 @uref{binaries.html,,binaries page} for details.
3245 The Solaris 2 @command{/bin/sh} will often fail to configure
3246 @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore
3247 recommend to use the following sequence of commands to bootstrap and
3251 % CONFIG_SHELL=/bin/ksh
3252 % export CONFIG_SHELL
3255 and then proceed as described in @uref{build.html,,the build instructions},
3256 where we strongly recommend using GNU make and specifying an absolute path
3257 to invoke @var{srcdir}/configure.
3259 Solaris 2 comes with a number of optional OS packages. Some of these
3260 are needed to use GCC fully, namely @code{SUNWarc},
3261 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3262 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3263 optional packages when installing Solaris 2, you will need to verify that
3264 the packages that GCC needs are installed.
3266 To check whether an optional package is installed, use
3267 the @command{pkginfo} command. To add an optional package, use the
3268 @command{pkgadd} command. For further details, see the Solaris 2
3271 Trying to use the linker and other tools in
3272 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3273 For example, the linker may hang indefinitely. The fix is to remove
3274 @file{/usr/ucb} from your @env{PATH}.
3276 The build process works more smoothly with the legacy Sun tools so, if you
3277 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
3278 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
3280 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3281 platform. We recommend the use of GNU binutils 2.11.2 or the vendor
3282 tools (Sun @command{as}, Sun @command{ld}).
3284 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3285 newer: @command{g++} will complain that types are missing. These headers assume
3286 that omitting the type means @code{int}; this assumption worked for C89 but
3287 is wrong for C++, and is now wrong for C99 also.
3289 @command{g++} accepts such (invalid) constructs with the option
3290 @option{-fpermissive}; it
3291 will assume that any missing type is @code{int} (as defined by C89).
3293 There are patches for Solaris 2.6 (105633-56 or newer for SPARC,
3294 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC,
3295 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3296 108653-22 for Intel) that fix this bug.
3301 @heading @anchor{sparc-sun-solaris2*}sparc-sun-solaris2*
3303 When GCC is configured to use binutils 2.11.2 or later the binaries
3304 produced are smaller than the ones produced using Sun's native tools;
3305 this difference is quite significant for binaries containing debugging
3308 Sun @command{as} 4.x is broken in that it cannot cope with long symbol names.
3309 A typical error message might look similar to the following:
3312 /usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041: error:
3313 can't compute value of an expression involving an external symbol.
3316 This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris
3317 2.6 and has been fixed in later (5.x) versions of the assembler,
3318 starting with Solaris 7.
3320 Starting with Solaris 7, the operating system is capable of executing
3321 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
3322 this; the @option{-m64} option enables 64-bit code generation.
3323 However, if all you want is code tuned for the UltraSPARC CPU, you
3324 should try the @option{-mtune=ultrasparc} option instead, which produces
3325 code that, unlike full 64-bit code, can still run on non-UltraSPARC
3328 When configuring on a Solaris 7 or later system that is running a kernel
3329 that supports only 32-bit binaries, one must configure with
3330 @option{--disable-multilib}, since we will not be able to build the
3331 64-bit target libraries.
3336 @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7
3338 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3339 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3340 and later, including all EGCS releases. Sun formerly recommended
3341 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3342 recommend it only for people who use Sun's compilers.
3344 Here are some workarounds to this problem:
3347 Do not install Sun patch 107058-01 until after Sun releases a
3348 complete patch for bug 4210064. This is the simplest course to take,
3349 unless you must also use Sun's C compiler. Unfortunately 107058-01
3350 is preinstalled on some new Solaris 7-based hosts, so you may have to
3354 Copy the original, unpatched Solaris 7
3355 @command{/usr/ccs/bin/as} into
3356 @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as},
3357 adjusting the latter name to fit your local conventions and software
3361 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3362 both 107058-01 and 106950-03 installed has reported the bug with GCC
3363 and Sun's dynamic linker. This last course of action is riskiest,
3364 for two reasons. First, you must install 106950 on all hosts that
3365 run code generated by GCC; it doesn't suffice to install it only on
3366 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3367 only a partial fix for bug 4210064, but Sun doesn't know whether the
3368 partial fix is adequate for GCC@. Revision -08 or later should fix
3369 the bug. The current (as of 2001-09-24) revision is -14, and is included in
3370 the Solaris 7 Recommended Patch Cluster.
3373 GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
3374 which causes a bootstrap failure when linking the 64-bit shared version of
3375 libgcc. A typical error message is:
3378 ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
3379 symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
3382 This bug has been fixed in the final 5.0 version of the assembler.
3387 @heading @anchor{sparc-*-linux*}sparc-*-linux*
3389 GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
3390 or newer on this platform. All earlier binutils and glibc
3391 releases mishandled unaligned relocations on @code{sparc-*-*} targets.
3397 @heading @anchor{sparc64-*-solaris2*}sparc64-*-solaris2*
3399 The following compiler flags must be specified in the configure
3400 step in order to bootstrap this target with the Sun compiler:
3403 % CC="cc -xildoff -xarch=v9" @var{srcdir}/configure [@var{options}] [@var{target}]
3406 @option{-xildoff} turns off the incremental linker, and @option{-xarch=v9}
3407 specifies the SPARC-V9 architecture to the Sun linker and assembler.
3412 @heading @anchor{sparcv9-*-solaris2*}sparcv9-*-solaris2*
3414 This is a synonym for sparc64-*-solaris2*.
3419 @heading @anchor{#*-*-sysv*}*-*-sysv*
3420 On System V release 3, you may get this error message
3424 ld fatal: failed to write symbol name @var{something}
3425 in strings table for file @var{whatever}
3428 This probably indicates that the disk is full or your ulimit won't allow
3429 the file to be as large as it needs to be.
3431 This problem can also result because the kernel parameter @code{MAXUMEM}
3432 is too small. If so, you must regenerate the kernel and make the value
3433 much larger. The default value is reported to be 1024; a value of 32768
3434 is said to work. Smaller values may also work.
3436 On System V, if you get an error like this,
3439 /usr/local/lib/bison.simple: In function `yyparse':
3440 /usr/local/lib/bison.simple:625: virtual memory exhausted
3444 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
3446 On a System V release 4 system, make sure @file{/usr/bin} precedes
3447 @file{/usr/ucb} in @code{PATH}. The @command{cc} command in
3448 @file{/usr/ucb} uses libraries which have bugs.
3453 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3454 Don't try compiling with VAX C (@command{vcc}). It produces incorrect code
3455 in some cases (for example, when @code{alloca} is used).
3460 @heading @anchor{*-*-vxworks*}*-*-vxworks*
3461 Support for VxWorks is in flux. At present GCC supports @emph{only} the
3462 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC.
3463 We welcome patches for other architectures supported by VxWorks 5.5.
3464 Support for VxWorks AE would also be welcome; we believe this is merely
3465 a matter of writing an appropriate ``configlette'' (see below). We are
3466 not interested in supporting older, a.out or COFF-based, versions of
3469 VxWorks comes with an older version of GCC installed in
3470 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
3471 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
3472 Before running @command{configure}, create the directories @file{@var{prefix}}
3473 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
3474 linker, etc. into @file{@var{prefix}/bin}, and set your @var{PATH} to
3475 include that directory while running both @command{configure} and
3478 You must give @command{configure} the
3479 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
3480 find the VxWorks system headers. Since VxWorks is a cross compilation
3481 target only, you must also specify @option{--target=@var{target}}.
3482 @command{configure} will attempt to create the directory
3483 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
3484 make sure the user running @command{configure} has sufficient privilege
3487 GCC's exception handling runtime requires a special ``configlette''
3488 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
3489 that file to add the module to your kernel build. (Future versions of
3490 VxWorks will incorporate this module.)
3495 @heading @anchor{xtensa-*-elf}xtensa-*-elf
3497 This target is intended for embedded Xtensa systems using the
3498 @samp{newlib} C library. It uses ELF but does not support shared
3499 objects. Designed-defined instructions specified via the
3500 Tensilica Instruction Extension (TIE) language are only supported
3501 through inline assembly.
3503 The Xtensa configuration information must be specified prior to
3504 building GCC@. The @file{include/xtensa-config.h} header
3505 file contains the configuration information. If you created your
3506 own Xtensa configuration with the Xtensa Processor Generator, the
3507 downloaded files include a customized copy of this header file,
3508 which you can use to replace the default header file.
3513 @heading @anchor{xtensa-*-linux*}xtensa-*-linux*
3515 This target is for Xtensa systems running GNU/Linux. It supports ELF
3516 shared objects and the GNU C library (glibc). It also generates
3517 position-independent code (PIC) regardless of whether the
3518 @option{-fpic} or @option{-fPIC} options are used. In other
3519 respects, this target is the same as the
3520 @uref{#xtensa-*-elf,,@samp{xtensa-*-elf}} target.
3525 @heading @anchor{windows}Microsoft Windows (32-bit)
3527 A port of GCC 2.95.2 and 3.x is included with the
3528 @uref{http://www.cygwin.com/,,Cygwin environment}.
3530 Current (as of early 2001) snapshots of GCC will build under Cygwin
3531 without modification.
3533 GCC does not currently build with Microsoft's C++ compiler and there
3534 are no plans to make it do so.
3539 @heading @anchor{os2}OS/2
3541 GCC does not currently support OS/2. However, Andrew Zabolotny has been
3542 working on a generic OS/2 port with pgcc. The current code can be found
3543 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
3545 An older copy of GCC 2.8.1 is included with the EMX tools available at
3546 @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,,
3547 ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}.
3552 @heading @anchor{older}Older systems
3554 GCC contains support files for many older (1980s and early
3555 1990s) Unix variants. For the most part, support for these systems
3556 has not been deliberately removed, but it has not been maintained for
3557 several years and may suffer from bitrot.
3559 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
3560 Support for these systems is still present in that release, but
3561 @command{configure} will fail unless the @option{--enable-obsolete}
3562 option is given. Unless a maintainer steps forward, support for these
3563 systems will be removed from the next release of GCC@.
3565 Support for old systems as hosts for GCC can cause problems if the
3566 workarounds for compiler, library and operating system bugs affect the
3567 cleanliness or maintainability of the rest of GCC@. In some cases, to
3568 bring GCC up on such a system, if still possible with current GCC, may
3569 require first installing an old version of GCC which did work on that
3570 system, and using it to compile a more recent GCC, to avoid bugs in the
3571 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
3572 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
3573 sites}. Header bugs may generally be avoided using
3574 @command{fixincludes}, but bugs or deficiencies in libraries and the
3575 operating system may still cause problems.
3577 Support for older systems as targets for cross-compilation is less
3578 problematic than support for them as hosts for GCC; if an enthusiast
3579 wishes to make such a target work again (including resurrecting any of
3580 the targets that never worked with GCC 2, starting from the last CVS
3581 version before they were removed), patches
3582 @uref{../contribute.html,,following the usual requirements} would be
3583 likely to be accepted, since they should not affect the support for more
3586 For some systems, old versions of GNU binutils may also be useful,
3587 and are available from @file{pub/binutils/old-releases} on
3588 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
3590 Some of the information on specific systems above relates to
3591 such older systems, but much of the information
3592 about GCC on such systems (which may no longer be applicable to
3593 current GCC) is to be found in the GCC texinfo manual.
3598 @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris 2, etc.)
3600 C++ support is significantly better on ELF targets if you use the
3601 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
3602 inlines, vtables and template instantiations will be discarded
3611 @uref{./index.html,,Return to the GCC Installation page}
3615 @c ***Old documentation******************************************************
3617 @include install-old.texi
3623 @uref{./index.html,,Return to the GCC Installation page}
3627 @c ***GFDL********************************************************************
3635 @uref{./index.html,,Return to the GCC Installation page}
3639 @c ***************************************************************************
3640 @c Part 6 The End of the Document
3642 @comment node-name, next, previous, up
3643 @node Concept Index, , GNU Free Documentation License, Top
3647 @unnumbered Concept Index