1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename install.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
18 @settitle Downloading GCC
21 @settitle Installing GCC: Configuration
24 @settitle Installing GCC: Building
27 @settitle Installing GCC: Testing
29 @ifset finalinstallhtml
30 @settitle Installing GCC: Final installation
33 @settitle Installing GCC: Binaries
36 @c Copyright (C) 2001 Free Software Foundation, Inc.
37 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
39 @c Include everything if we're not making html
51 @c Part 2 Summary Description and Copyright
54 Copyright @copyright{} 2001 Free Software Foundation, Inc.
57 @c Part 3 Titlepage and Copyright
60 @comment The title is printed in a large font.
61 @center @titlefont{Installing GCC}
63 @c The following two commands start the copyright page.
65 @vskip 0pt plus 1filll
66 Copyright @copyright{} 2001 Free Software Foundation, Inc.
69 @c Part 4 Top node and Master Menu
72 @comment node-name, next, Previous, up
75 * Installing GCC:: This document describes the generic installation
76 procedure for GCC as well as detailing some target
77 specific installation instructions.
79 * Specific:: Host/target specific installation notes for GCC.
80 * Binaries:: Where to get pre-compiled binaries.
82 * Concept Index:: This index has two entries.
86 @c Part 5 The Body of the Document
87 @c ***Installing GCC**********************************************************
89 @comment node-name, next, previous, up
90 @node Installing GCC, Binaries, , Top
94 <h1 align="center">Installing GCC</h1>
97 @chapter Installing GCC
100 The latest version of this document is always available at
101 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
103 This document describes the generic installation procedure for GCC as well
104 as detailing some target specific installation instructions.
106 GCC includes several components that previously were separate distributions
107 with their own installation instructions. This document supersedes all
108 package specific installation instructions.
110 @emph{Before} starting the build/install procedure please check the
112 @ref{Specific, host/target specific installation notes}.
115 @uref{specific.html,,host/target specific installation notes}.
117 We recommend you browse the entire generic installation instructions before
120 The installation procedure itself is broken into five steps.
124 * Downloading the source::
127 * Testing:: (optional)
134 @uref{download.html,,Downloading the source}
136 @uref{configure.html,,Configuration}
138 @uref{build.html,,Building}
140 @uref{test.html,,Testing} (optional)
142 @uref{finalinstall.html,,Final install}
146 Please note that GCC does not support @samp{make uninstall} and probably
147 won't do so in the near future as this would open a can of worms. Instead,
148 we suggest that you install GCC into a directory of its own and simply
149 remove that directory when you do not need that specific version of GCC
150 any longer, and, if shared libraries are installed there as well, no
151 more binaries exist that use them.
158 @uref{./index.html,,Return to the GCC Installation page}
162 @c ***Downloading the source**************************************************
164 @comment node-name, next, previous, up
165 @node Downloading the source, Configuration, , Installing GCC
169 <h1 align="center">Downloading GCC</h1>
172 @chapter Downloading GCC
174 @cindex Downloading GCC
175 @cindex Downloading the Source
177 GCC is distributed via @uref{http://gcc.gnu.org/cvs.html,,CVS} and FTP
178 tarballs compressed with @command{gzip} or
179 @command{bzip2}. It is possible to download a full distribution or specific
182 Please refer to our @uref{http://gcc.gnu.org/releases.html,,releases web page}
183 for information on how to obtain GCC@.
185 The full distribution includes the C, C++, Objective-C, Fortran, Java,
186 and CHILL compilers. The full distribution also includes runtime libraries
187 for C++, Objective-C, Fortran, Java and CHILL. (GCC 3.0 does not
188 include CHILL.) In GCC 3.0 and later versions, GNU compiler testsuites
189 are also included in the full distribution.
191 If you choose to download specific components, you must download the core
192 GCC distribution plus any language specific distributions you wish to
193 use. The core distribution includes the C language front end as well as the
194 shared components. Each language has a tarball which includes the language
195 front end as well as the language runtime (when appropriate).
197 Unpack the core distribution as well as any language specific
198 distributions in the same directory.
200 If you also intend to build binutils (either to upgrade an existing
201 installation or for use in place of the corresponding tools of your
202 OS), unpack the binutils distribution either in the same directory or
203 a separate one. In the latter case, add symbolic links to any
204 components of the binutils you intend to build alongside the compiler
205 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
206 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
213 @uref{./index.html,,Return to the GCC Installation page}
217 @c ***Configuration***********************************************************
219 @comment node-name, next, previous, up
220 @node Configuration, Building, Downloading the source, Installing GCC
224 <h1 align="center">Installing GCC: Configuration</h1>
227 @chapter Installing GCC: Configuration
229 @cindex Configuration
230 @cindex Installing GCC: Configuration
232 Like most GNU software, GCC must be configured before it can be built.
233 This document describes the recommended configuration procedure
234 for both native and cross targets.
236 We use @var{srcdir} to refer to the toplevel source directory for
237 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
239 If you obtained the sources via CVS, @var{srcdir} must refer to the top
240 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
241 and not its @file{gcc} subdirectory, otherwise the build will fail.
243 First, we @strong{highly} recommend that GCC be built into a
244 separate directory than the sources which does @strong{not} reside
245 within the source tree. This is how we generally build GCC; building
246 where @var{srcdir} == @var{objdir} should still work, but doesn't
247 get extensive testing; building where @var{objdir} is a subdirectory
248 of @var{srcdir} is unsupported.
250 If you have previously built GCC in the same directory for a
251 different target machine, do @samp{make distclean} to delete all files
252 that might be invalid. One of the files this deletes is
253 @file{Makefile}; if @samp{make distclean} complains that @file{Makefile}
254 does not exist, it probably means that the directory is already suitably
255 clean. However, with the recommended method of building in a separate
256 @var{objdir}, you should simply use a different @var{objdir} for each
259 Second, when configuring a native system, either @command{cc} or
260 @command{gcc} must be in your path or you must set @env{CC} in
261 your environment before running configure. Otherwise the configuration
264 Note that the bootstrap compiler and the resulting GCC must be link
265 compatible, else the bootstrap will fail with linker errors about
266 incompatible object file formats. Several multilibed targets are
267 affected by this requirement, see
269 @ref{Specific, host/target specific installation notes}.
272 @uref{specific.html,,host/target specific installation notes}.
280 % @var{srcdir}/configure [@var{options}] [@var{target}]
284 @heading Target specification
287 GCC has code to correctly determine the correct value for @var{target}
288 for nearly all native systems. Therefore, we highly recommend you not
289 provide a configure target when configuring a native compiler.
292 @var{target} must be specified as @option{--target=@var{target}}
293 when configuring a cross compiler; examples of valid targets would be
294 i960-rtems, m68k-coff, sh-elf, etc.
297 Specifying just @var{target} instead of @option{--target=@var{target}}
298 implies that the host defaults to @var{target}.
302 @heading Options specification
304 Use @var{options} to override several configure time options for
305 GCC@. A partial list of supported @var{options}:
308 @item --prefix=@var{dirname}
309 Specify the toplevel installation
310 directory. This is the recommended way to install the tools into a directory
311 other than the default. The toplevel installation directory defaults to
314 We @strong{highly} recommend against @var{dirname} being the same or a
315 subdirectory of @var{objdir} or vice versa.
317 These additional options control where certain parts of the distribution
318 are installed. Normally you should not need to use these options.
321 @item --exec-prefix=@var{dirname}
322 Specify the toplevel installation directory for architecture-dependent
323 files. The default is @file{@var{prefix}}.
325 @item --bindir=@var{dirname}
326 Specify the installation directory for the executables called by users
327 (such as @command{gcc} and @command{g++}). The default is
328 @file{@var{exec-prefix}/bin}.
330 @item --libdir=@var{dirname}
331 Specify the installation directory for object code libraries and
332 internal parts of GCC@. The default is @file{@var{exec-prefix}/lib}.
334 @item --with-slibdir=@var{dirname}
335 Specify the installation directory for the shared libgcc library. The
336 default is @file{@var{libdir}}.
338 @item --infodir=@var{dirname}
339 Specify the installation directory for documentation in info format.
340 The default is @file{@var{prefix}/info}.
342 @item --mandir=@var{dirname}
343 Specify the installation directory for manual pages. The default is
344 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
345 the full GCC manuals, which are provided in Texinfo format. The
346 @command{g77} manpage is unmaintained and may be out of date; the others
347 are derived by an automatic conversion process from parts of the full
350 @item --with-gxx-include-dir=@var{dirname}
352 the installation directory for G++ header files. The default is
353 @file{@var{prefix}/include/g++-v3}.
357 @item --with-local-prefix=@var{dirname}
359 installation directory for local include files. The default is
360 @file{/usr/local}. Specify this option if you want the compiler to
361 search directory @file{@var{dirname}/include} for locally installed
362 header files @emph{instead} of @file{/usr/local/include}.
364 You should specify @option{--with-local-prefix} @strong{only} if your
365 site has a different convention (not @file{/usr/local}) for where to put
368 The default value for @option{--with-local-prefix} is @file{/usr/local}
369 regardless of the value of @option{--prefix}. Specifying
370 @option{--prefix} has no effect on which directory GCC searches for
371 local header files. This may seem counterintuitive, but actually it is
374 The purpose of @option{--prefix} is to specify where to @emph{install
375 GCC}. The local header files in @file{/usr/local/include}---if you put
376 any in that directory---are not part of GCC@. They are part of other
377 programs---perhaps many others. (GCC installs its own header files in
378 another directory which is based on the @option{--prefix} value.)
380 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
381 The directory you use for @option{--with-local-prefix} @strong{must not}
382 contain any of the system's standard header files. If it did contain
383 them, certain programs would be miscompiled (including GNU Emacs, on
384 certain targets), because this would override and nullify the header
385 file corrections made by the @code{fixincludes} script.
387 Indications are that people who use this option use it based on mistaken
388 ideas of what it is for. People use it as if it specified where to
389 install part of GCC@. Perhaps they make this assumption because
390 installing GCC creates the directory.
392 @item --enable-shared[=@var{package}[,@dots{}]]
393 Build shared versions of libraries, if shared libraries are supported on
394 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
395 are enabled by default on all platforms that support shared libraries,
396 except for @samp{libobjc} which is built as a static library only by
399 If a list of packages is given as an argument, build shared libraries
400 only for the listed packages. For other packages, only static libraries
401 will be built. Package names currently recognized in the GCC tree are
402 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
403 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc} and
404 @samp{libjava}. Note that @samp{libobjc} does not recognize itself by
405 any name, so, if you list package names in @option{--enable-shared},
406 you will only get static Objective-C libraries. @samp{libf2c} and
407 @samp{libiberty} do not support shared libraries at all.
409 Use @option{--disable-shared} to build only static libraries. Note that
410 @option{--disable-shared} does not accept a list of package names as
411 argument, only @option{--enable-shared} does.
413 @item @anchor{with-gnu-as}--with-gnu-as
414 Specify that the compiler should assume that the
415 assembler it finds is the GNU assembler. However, this does not modify
416 the rules to find an assembler and will result in confusion if found
417 assembler is not actually the GNU assembler. (Confusion will also
418 result if the compiler finds the GNU assembler but has not been
419 configured with @option{--with-gnu-as}.) If you have more than one
420 assembler installed on your system, you may want to use this option in
421 connection with @option{--with-as=@var{pathname}}.
423 The systems where it makes a difference whether you use the GNU assembler are
424 @samp{hppa1.0-@var{any}-@var{any}}, @samp{hppa1.1-@var{any}-@var{any}},
425 @samp{i386-@var{any}-sysv}, @samp{i386-@var{any}-isc},
426 @samp{i860-@var{any}-bsd}, @samp{m68k-bull-sysv},
427 @samp{m68k-hp-hpux}, @samp{m68k-sony-bsd},
428 @samp{m68k-altos-sysv}, @samp{m68000-hp-hpux},
429 @samp{m68000-att-sysv}, @samp{@var{any}-lynx-lynxos},
430 and @samp{mips-@var{any}}.
431 On any other system, @option{--with-gnu-as} has no effect.
433 On the systems listed above (except for the HP-PA, for ISC on the
434 386, and for @samp{mips-sgi-irix5.*}), if you use the GNU assembler,
435 you should also use the GNU linker (and specify @option{--with-gnu-ld}).
437 @item --with-as=@var{pathname}
439 compiler should use the assembler pointed to by @var{pathname}, rather
440 than the one found by the standard rules to find an assembler, which
445 @file{@var{exec_prefix}/lib/gcc-lib/@var{target}/@var{version}}
446 directory, where @var{exec_prefix} defaults to @var{prefix} which
447 defaults to @file{/usr/local} unless overridden by the
448 @option{--prefix=@var{pathname}} switch described above. @var{target} is the
449 target system triple, such as @samp{sparc-sun-solaris2.7}, and
450 @var{version} denotes the GCC version, such as 3.0.
452 Check operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
455 Note that these rules do not check for the value of @env{PATH}. You may
456 want to use @option{--with-as} if no assembler is installed in the
457 directories listed above, or if you have multiple assemblers installed
458 and want to choose one that is not found by the above rules.
460 @item @anchor{with-gnu-ld}--with-gnu-ld
461 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
465 @item --with-ld=@var{pathname}
467 @option{--with-as}, but for the linker.
470 Specify that stabs debugging
471 information should be used instead of whatever format the host normally
472 uses. Normally GCC uses the same debug format as the host system.
474 On MIPS based systems and on Alphas, you must specify whether you want
475 GCC to create the normal ECOFF debugging format, or to use BSD-style
476 stabs passed through the ECOFF symbol table. The normal ECOFF debug
477 format cannot fully handle languages other than C@. BSD stabs format can
478 handle other languages, but it only works with the GNU debugger GDB@.
480 Normally, GCC uses the ECOFF debugging format by default; if you
481 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
483 No matter which default you choose when you configure GCC, the user
484 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
485 the debug format for a particular compilation.
487 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
488 @option{--with-gas} is used. It selects use of stabs debugging
489 information embedded in COFF output. This kind of debugging information
490 supports C++ well; ordinary COFF debugging information does not.
492 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
493 selects use of stabs debugging information embedded in ELF output. The
494 C++ compiler currently (2.6.0) does not support the DWARF debugging
495 information normally used on 386 SVR4 platforms; stabs provide a
496 workable alternative. This requires gas and gdb, as the normal SVR4
497 tools can not generate or interpret stabs.
499 @item --disable-multilib
500 Specify that multiple target
501 libraries to support different target variants, calling
502 conventions, etc should not be built. The default is to build a
503 predefined set of them.
505 Some targets provide finer-grained control over which multilibs are built
506 (e.g., @option{--disable-softfloat}):
513 fpu, 26bit, underscore, interwork, biendian, nofmult.
516 softfloat, m68881, m68000, m68020.
519 single-float, biendian, softfloat.
521 @item powerpc*-*-*, rs6000*-*-*
522 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
527 @item --enable-threads
528 Specify that the target
529 supports threads. This affects the Objective-C compiler and runtime
530 library, and exception handling for other languages like C++ and Java.
531 On some systems, this is the default.
533 In general, the best (and, in many cases, the only known) threading
534 model available will be configured for use. Beware that on some
535 systems, gcc has not been taught what threading models are generally
536 available for the system. In this case, @option{--enable-threads} is an
537 alias for @option{--enable-threads=single}.
539 @item --disable-threads
540 Specify that threading support should be disabled for the system.
541 This is an alias for @option{--enable-threads=single}.
543 @item --enable-threads=@var{lib}
545 @var{lib} is the thread support library. This affects the Objective-C
546 compiler and runtime library, and exception handling for other languages
547 like C++ and Java. The possibilities for @var{lib} are:
555 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
556 that the file needed to support this configuration, @file{gthr-mach.h}, is
557 missing and thus this setting will cause a known bootstrap failure.)
559 This is an alias for @samp{single}.
561 Generic POSIX thread support.
563 Same as @samp{posix} on arm*-*-linux*, *-*-chorusos* and *-*-freebsd*
564 only. A future release of gcc might remove this alias or extend it
567 RTEMS thread support.
569 Disable thread support, should work for all platforms.
571 Sun Solaris 2 thread support.
573 VxWorks thread support.
575 Microsoft Win32 API thread support.
578 @item --with-cpu=@var{cpu}
579 Specify which cpu variant the
580 compiler should generate code for by default. This is currently
581 only supported on the some ports, specifically arm, powerpc, and
582 SPARC@. If configure does not recognize the model name (e.g.@: arm700,
583 603e, or ultrasparc) you provide, please check the configure script
584 for a complete list of supported models.
586 @item --enable-target-optspace
588 libraries should be optimized for code space instead of code speed.
589 This is the default for the m32r platform.
592 Specify that a user visible @command{cpp} program should not be installed.
594 @item --with-cpp-install-dir=@var{dirname}
595 Specify that the user visible @command{cpp} program should be installed
596 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
598 @item --enable-maintainer-mode
600 regenerate the GCC master message catalog @file{gcc.pot} are normally
601 disabled. This is because it can only be rebuilt if the complete source
602 tree is present. If you have changed the sources and want to rebuild the
603 catalog, configuring with @option{--enable-maintainer-mode} will enable
604 this. Note that you need a recent version of the @code{gettext} tools
607 @item --enable-version-specific-runtime-libs
609 that runtime libraries should be installed in the compiler specific
610 subdirectory (@file{@var{libsubdir}}) rather than the usual places. In
611 addition, @samp{libstdc++}'s include files will be installed in
612 @file{@var{libsubdir}/include/g++} unless you overruled it by using
613 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
614 particularly useful if you intend to use several versions of GCC in
615 parallel. This is currently supported by @samp{libf2c} and
616 @samp{libstdc++}, and is the default for @samp{libobjc} which cannot be
617 changed in this case.
619 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
620 Specify that only a particular subset of compilers and
621 their runtime libraries should be built. For a list of valid values for
622 @var{langN} you can issue the following command in the
623 @file{gcc} directory of your GCC source tree:@*
625 grep language= */config-lang.in
627 Currently, you can use any of the following:
628 @code{c}, @code{c++}, @code{f77}, @code{java} and @code{objc}.
629 @code{CHILL} is not currently maintained, and will almost
630 certainly fail to compile.@*
631 If you do not pass this flag, all languages available in the @file{gcc}
632 sub-tree will be configured. Re-defining @code{LANGUAGES} when calling
633 @samp{make bootstrap} @strong{does not} work anymore, as those
634 language sub-directories might not have been configured!
636 @item --disable-libgcj
637 Specify that the run-time libraries
638 used by GCJ should not be built. This is useful in case you intend
639 to use GCJ with some other run-time, or you're going to install it
640 separately, or it just happens not to build on your particular
641 machine. In general, if the Java front end is enabled, the GCJ
642 libraries will be enabled too, unless they're known to not work on
643 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
644 may need to port it; in this case, before modifying the top-level
645 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
646 you may use @option{--enable-libgcj} to override the default.
649 Specify that the compiler should
650 use DWARF 2 debugging information as the default.
652 @item --enable-win32-registry
653 @itemx --enable-win32-registry=@var{key}
654 @itemx --disable-win32-registry
655 The @option{--enable-win32-registry} option enables Windows-hosted GCC
656 to look up installations paths in the registry using the following key:
659 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
662 @var{key} defaults to GCC version number, and can be overridden by the
663 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
664 who use custom installers are encouraged to provide a different key,
665 perhaps one comprised of vendor name and GCC version number, to
666 avoid conflict with existing installations. This feature is enabled
667 by default, and can be disabled by @option{--disable-win32-registry}
668 option. This option has no effect on the other hosts.
671 Specify that the machine does not have a floating point unit. This
672 option only applies to @samp{m68k-sun-sunos@var{n}} and
673 @samp{m68k-isi-bsd}. On any other system, @option{--nfp} has no effect.
675 @item --enable-checking
676 @itemx --enable-checking=@var{list}
677 When you specify this option, the compiler is built to perform checking
678 of tree node types when referencing fields of that node, and some other
679 internal consistency checks. This does not change the generated code,
680 but adds error checking within the compiler. This will slow down the
681 compiler and may only work properly if you are building the compiler
682 with GCC@. This is on by default when building from CVS or snapshots,
683 but off for releases. More control over the checks may be had by
684 specifying @var{list}; the categories of checks available are
685 @samp{misc}, @samp{tree}, @samp{gc}, @samp{rtl} and @samp{gcac}. The
686 default when @var{list} is not specified is @samp{misc,tree,gc}; the
687 checks @samp{rtl} and @samp{gcac} are very expensive.
691 The @option{--enable-nls} option enables Native Language Support (NLS),
692 which lets GCC output diagnostics in languages other than American
693 English. Native Language Support is enabled by default if not doing a
694 canadian cross build. The @option{--disable-nls} option disables NLS@.
696 @item --with-included-gettext
697 If NLS is enabled, the @option{--with-included-gettext} option causes the build
698 procedure to prefer its copy of GNU @command{gettext}.
701 If NLS is enabled, and if the host lacks @code{gettext} but has the
702 inferior @code{catgets} interface, the GCC build procedure normally
703 ignores @code{catgets} and instead uses GCC's copy of the GNU
704 @code{gettext} library. The @option{--with-catgets} option causes the
705 build procedure to use the host's @code{catgets} in this situation.
708 Some options which only apply to building cross compilers:
710 @item --with-headers=@var{dir}
711 Specifies a directory
712 which has target include files.
713 @emph{This options is required} when building a cross
714 compiler, if @file{@var{prefix}/@var{target}/sys-include} doesn't pre-exist.
715 These include files will be copied into the @file{gcc} install directory.
716 Fixincludes will be run on these files to make them compatible with
718 @item --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
719 Specifies a list of directories which contain the target runtime
720 libraries. These libraries will be copied into the @file{gcc} install
723 Specifies that @samp{newlib} is
724 being used as the target C library. This causes @code{__eprintf} to be
725 omitted from @file{libgcc.a} on the assumption that it will be provided by
729 Note that each @option{--enable} option has a corresponding
730 @option{--disable} option and that each @option{--with} option has a
731 corresponding @option{--without} option.
738 @uref{./index.html,,Return to the GCC Installation page}
742 @c ***Building****************************************************************
744 @comment node-name, next, previous, up
745 @node Building, Testing, Configuration, Installing GCC
749 <h1 align="center">Installing GCC: Building</h1>
754 @cindex Installing GCC: Building
756 Now that GCC is configured, you are ready to build the compiler and
759 We @strong{highly} recommend that GCC be built using GNU make;
760 other versions may work, then again they might not.
762 (For example, many broken versions of make will fail if you use the
763 recommended setup where @var{objdir} is different from @var{srcdir}.
764 Other broken versions may recompile parts of the compiler when
765 installing the compiler.)
767 Some commands executed when making the compiler may fail (return a
768 non-zero status) and be ignored by @code{make}. These failures, which
769 are often due to files that were not found, are expected, and can safely
772 It is normal to have compiler warnings when compiling certain files.
773 Unless you are a GCC developer, you can generally ignore these warnings
774 unless they cause compilation to fail.
776 On certain old systems, defining certain environment variables such as
777 @env{CC} can interfere with the functioning of @command{make}.
779 If you encounter seemingly strange errors when trying to build the
780 compiler in a directory other than the source directory, it could be
781 because you have previously configured the compiler in the source
782 directory. Make sure you have done all the necessary preparations.
784 If you build GCC on a BSD system using a directory stored in an old System
785 V file system, problems may occur in running @code{fixincludes} if the
786 System V file system doesn't support symbolic links. These problems
787 result in a failure to fix the declaration of @code{size_t} in
788 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
789 that type mismatches occur, this could be the cause.
791 The solution is not to use such a directory for building GCC@.
793 When building from CVS or snapshots, or if you modify parser sources,
794 you need the Bison parser generator installed. Any version 1.25 or
795 later should work; older versions may also work. If you do not modify
796 parser sources, releases contain the Bison-generated files and you do
797 not need Bison installed to build them.
799 When building from CVS or snapshots, or if you modify Texinfo
800 documentation, you need version 4.0 or later of Texinfo installed if you
801 want Info documentation to be regenerated. Releases contain Info
802 documentation pre-built for the unmodified documentation in the release.
804 @section Building a native compiler
806 For a native build issue the command @samp{make bootstrap}. This
807 will build the entire GCC system, which includes the following steps:
811 Build host tools necessary to build the compiler such as texinfo, bison,
815 Build target tools for use by the compiler such as binutils (bfd,
816 binutils, gas, gprof, ld, and opcodes)
817 if they have been individually linked
818 or moved into the top level GCC source tree before configuring.
821 Perform a 3-stage bootstrap of the compiler.
824 Perform a comparison test of the stage2 and stage3 compilers.
827 Build runtime libraries using the stage3 compiler from the previous step.
831 If you are short on disk space you might consider @samp{make
832 bootstrap-lean} instead. This is identical to @samp{make
833 bootstrap} except that object files from the stage1 and
834 stage2 of the 3-stage bootstrap of the compiler are deleted as
835 soon as they are no longer needed.
838 If you want to save additional space during the bootstrap and in
839 the final installation as well, you can build the compiler binaries
840 without debugging information with @samp{make CFLAGS='-O' LIBCFLAGS='-g
841 -O2' LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap}. This will save
842 roughly 40% of disk space both for the bootstrap and the final installation.
843 (Libraries will still contain debugging information.)
845 If you wish to use non-default GCC flags when compiling the stage2 and
846 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
847 @samp{make bootstrap}. Non-default optimization flags are less well
848 tested here than the default of @samp{-g -O2}, but should still work.
849 In a few cases, you may find that you need to specify special flags such
850 as @option{-msoft-float} here to complete the bootstrap; or, if the
851 native compiler miscompiles the stage1 compiler, you may need to work
852 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
853 stage1 compiler that were miscompiled, or by using @samp{make
854 bootstrap4} to increase the number of stages of bootstrap.
856 If you used the flag @option{--enable-languages=@dots{}} to restrict
857 the compilers to be built, only those you've actually enabled will be
858 built. This will of course only build those runtime libraries, for
859 which the particular compiler has been built. Please note,
860 that re-defining @env{LANGUAGES} when calling @samp{make bootstrap}
861 @strong{does not} work anymore!
863 If the comparison of stage2 and stage3 fails, this normally indicates
864 that the stage2 compiler has compiled GCC incorrectly, and is therefore
865 a potentially serious bug which you should investigate and report. (On
866 a few systems, meaningful comparison of object files is impossible; they
867 always appear ``different''. If you encounter this problem, you will
868 need to disable comparison in the @file{Makefile}.)
870 @section Building a cross compiler
872 We recommend reading the
873 @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ}
874 for information about building cross compilers.
876 When building a cross compiler, it is not generally possible to do a
877 3-stage bootstrap of the compiler. This makes for an interesting problem
878 as parts of GCC can only be built with GCC@.
880 To build a cross compiler, we first recommend building and installing a
881 native compiler. You can then use the native GCC compiler to build the
884 Assuming you have already installed a native copy of GCC and configured
885 your cross compiler, issue the command @command{make}, which performs the
890 Build host tools necessary to build the compiler such as texinfo, bison,
894 Build target tools for use by the compiler such as binutils (bfd,
895 binutils, gas, gprof, ld, and opcodes)
896 if they have been individually linked or moved into the top level GCC source
897 tree before configuring.
900 Build the compiler (single stage only).
903 Build runtime libraries using the compiler from the previous step.
906 Note that if an error occurs in any step the make process will exit.
908 @section Building in parallel
910 If you have a multiprocessor system you can use @samp{make bootstrap
911 MAKE="make -j 2" -j 2} or just @samp{make -j 2 bootstrap}
912 for GNU Make 3.79 and above instead of just @samp{make bootstrap}
913 when building GCC@. You can use a bigger number instead of two if
914 you like. In most cases, it won't help to use a number bigger than
915 the number of processors in your machine.
922 @uref{./index.html,,Return to the GCC Installation page}
926 @c ***Testing*****************************************************************
928 @comment node-name, next, previous, up
929 @node Testing, Final install, Building, Installing GCC
933 <h1 align="center">Installing GCC: Testing</h1>
936 @chapter Installing GCC: Testing
939 @cindex Installing GCC: Testing
942 Before you install GCC, you might wish to run the testsuite. This
943 step is optional and may require you to download additional software.
945 First, you must have @uref{download.html,,downloaded the testsuites}.
946 The full distribution contains testsuites; only if you downloaded the
947 ``core'' compiler plus any front ends, you do not have the testsuites.
949 Second, you must have a @uref{http://www.gnu.org/software/dejagnu/,,current version of DejaGnu} installed;
950 dejagnu 1.3 is not sufficient.
952 Now you may need specific preparations:
957 The following environment variables may need to be set appropriately, as in
958 the following example (which assumes that DejaGnu has been installed
959 under @file{/usr/local}):
962 TCL_LIBRARY = /usr/local/share/tcl8.0
963 DEJAGNULIBS = /usr/local/share/dejagnu
966 On systems such as Cygwin, these paths are required to be actual
967 paths, not mounts or links; presumably this is due to some lack of
968 portability in the DejaGnu code.
970 If the directories where @command{runtest} and @command{expect} were
971 installed are in the @env{PATH}, it should not be necessary to set these
972 environment variables.
976 Finally, you can run the testsuite (which may take a long time):
978 cd @var{objdir}; make -k check
981 The testing process will try to test as many components in the GCC
982 distribution as possible, including the C, C++, Objective-C and Fortran
983 compilers as well as the C++ and Java runtime libraries.
985 @section How can I run the test suite on selected tests?
987 As a first possibility to cut down the number of tests that are run it is
988 possible to use @samp{make check-gcc} or @samp{make check-g++}
989 in the @file{gcc} subdirectory of the object directory. To further cut down the
990 tests the following is possible:
993 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
996 This will run all @command{gcc} execute tests in the testsuite.
999 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1002 This will run the @command{g++} ``old-deja'' tests in the testsuite where the filename
1003 matches @samp{9805*}.
1005 The @file{*.exp} files are located in the testsuite directories of the GCC
1006 source, the most important ones being @file{compile.exp},
1007 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1008 To get a list of the possible @file{*.exp} files, pipe the
1009 output of @samp{make check} into a file and look at the
1010 @samp{Running @dots{} .exp} lines.
1012 @section How to interpret test results
1014 After the testsuite has run you'll find various @file{*.sum} and @file{*.log}
1015 files in the testsuite subdirectories. The @file{*.log} files contain a
1016 detailed log of the compiler invocations and the corresponding
1017 results, the @file{*.sum} files summarize the results. These summaries list
1018 all the tests that have been run with a corresponding status code:
1022 PASS: the test passed as expected
1024 XPASS: the test unexpectedly passed
1026 FAIL: the test unexpectedly failed
1028 XFAIL: the test failed as expected
1030 UNSUPPORTED: the test is not supported on this platform
1032 ERROR: the testsuite detected an error
1034 WARNING: the testsuite detected a possible problem
1037 It is normal for some tests to report unexpected failures. At the
1038 current time our testing harness does not allow fine grained control
1039 over whether or not a test is expected to fail. We expect to fix this
1040 problem in future releases.
1043 @section Submitting test results
1045 If you want to report the results to the GCC project, use the
1046 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
1049 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
1050 -m gcc-testresults@@gcc.gnu.org |sh
1053 This script uses the @command{Mail} program to send the results, so
1054 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
1055 prepended to the testsuite summary and should contain any special
1056 remarks you have on your results or your build environment. Please
1057 do not edit the testsuite result block or the subject line, as these
1058 messages are automatically parsed and presented at the
1059 @uref{http://gcc.gnu.org/testresults/,,GCC testresults} web
1060 page. Here you can also gather information on how specific tests
1061 behave on different platforms and compare them with your results. A
1062 few failing testcases are possible even on released versions and you
1063 should look here first if you think your results are unreasonable.
1067 @c ***Final install***********************************************************
1069 @comment node-name, next, previous, up
1070 @node Final install, , Testing, Installing GCC
1072 @ifset finalinstallhtml
1074 <h1 align="center">Installing GCC: Final installation</h1>
1077 @chapter Installing GCC: Final installation
1080 Now that GCC has been built (and optionally tested), you can install it with
1082 cd @var{objdir}; make install
1085 That step completes the installation of GCC; user level binaries can
1086 be found in @file{@var{prefix}/bin} where @var{prefix} is the value you
1087 specified with the @option{--prefix} to configure (or @file{/usr/local}
1088 by default). (If you specified @option{--bindir}, that directory will
1089 be used instead; otherwise, if you specified @option{--exec-prefix},
1090 @file{@var{exec-prefix}/bin} will be used.) Headers for the C++ and
1091 Java libraries are installed in @file{@var{prefix}/include}; libraries
1092 in @file{@var{libdir}} (normally @file{@var{prefix}/lib}); internal
1093 parts of the compiler in @file{@var{libdir}/gcc-lib}; documentation in
1094 info format in @file{@var{infodir}} (normally @file{@var{prefix}/info}).
1096 If you don't mind, please quickly review the
1097 @uref{http://gcc.gnu.org/gcc-3.0/buildstat.html,,build status page}.
1098 If your system is not listed, send a note to
1099 @email{gcc@@gcc.gnu.org} indicating
1100 that you successfully built and installed GCC.
1102 Include the output from running @file{@var{srcdir}/config.guess}. (Do
1103 not send us the @file{config.guess} file itself, just the one-line output from
1106 If you find a bug, please report it following our
1107 @uref{../bugs.html,,bug reporting guidelines}.
1109 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
1110 dvi}. You will need to have @command{texi2dvi} (version at least 4.0)
1111 and @TeX{} installed. This creates a number of @file{.dvi} files in
1112 subdirectories of @file{@var{objdir}}; these may be converted for
1113 printing with programs such as @command{dvips}. You can also
1114 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
1115 Free Software Foundation}, though such manuals may not be for the most
1116 recent version of GCC@.
1123 @uref{./index.html,,Return to the GCC Installation page}
1127 @c ***Binaries****************************************************************
1129 @comment node-name, next, previous, up
1130 @node Binaries, Specific, Installing GCC, Top
1134 <h1 align="center">Installing GCC: Binaries</h1>
1137 @chapter Installing GCC: Binaries
1140 @cindex Installing GCC: Binaries
1142 We are often asked about pre-compiled versions of GCC@. While we cannot
1143 provide these for all platforms, below you'll find links to binaries for
1144 various platforms where creating them by yourself is not easy due to various
1147 Please note that we did not create these binaries, nor do we
1148 support them. If you have any problems installing them, please
1149 contact their makers.
1156 @uref{http://freeware.bull.net,,Bull's Freeware and Shareware Archive for AIX};
1159 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX};
1163 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP};
1166 @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center};
1169 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
1170 OpenServer/Unixware};
1173 Solaris (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware};
1176 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware};
1179 Windows 95, 98, and NT:
1182 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
1184 @uref{http://www.xraylith.wisc.edu/~khan/software/gnu-win32/,,GNU Win32}
1185 related projects by Mumit Khan.
1189 @uref{ftp://ftp.thewrittenword.com/packages/free/by-name/gcc-2.95.2/,,The
1190 Written Word} offers binaries for Solaris 2.5.1, 2.6, 2.7/SPARC, 2.7/Intel,
1191 IRIX 6.2, 6.5, Digital UNIX 4.0D, HP-UX 10.20, and HP-UX 11.00.
1194 Hitachi H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
1195 Development Tools for the Hitachi H8/300[HS] Series}
1199 In addition to those specific offerings, you can get a binary
1200 distribution CD-ROM from the
1201 @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}.
1202 It contains binaries for a number of platforms, and
1203 includes not only GCC, but other stuff as well. The current CD does
1204 not contain the latest version of GCC, but it should allow
1205 bootstrapping the compiler. An updated version of that disk is in the
1213 @uref{./index.html,,Return to the GCC Installation page}
1217 @c ***Specific****************************************************************
1219 @comment node-name, next, previous, up
1220 @node Specific, Concept Index, Binaries, Top
1224 <h1 align="center">Host/target specific installation notes for GCC</h1>
1227 @chapter Host/target specific installation notes for GCC
1230 @cindex Specific installation notes
1231 @cindex Target specific installation
1232 @cindex Host specific installation
1233 @cindex Target specific installation notes
1235 Please read this document carefully @emph{before} installing the
1236 GNU Compiler Collection on your machine.
1241 @uref{#1750a-*-*,,1750a-*-*}
1245 @uref{#a29k-*-bsd,,a29k-*-bsd}
1247 @uref{#alpha*-*-*,,alpha*-*-*}
1249 @uref{#alpha*-dec-osf*,,alpha*-dec-osf*}
1251 @uref{#arc-*-elf,,arc-*-elf}
1253 @uref{#arm-*-aout,,arm-*-aout}
1255 @uref{#arm-*-elf,,arm-*-elf}
1257 @uref{#arm*-*-linux-gnu,,arm*-*-linux-gnu}
1259 @uref{#arm-*-riscix,,arm-*-riscix}
1267 @uref{#dsp16xx,,dsp16xx}
1269 @uref{#elxsi-elxsi-bsd,,elxsi-elxsi-bsd}
1271 @uref{#*-*-freebsd*,,*-*-freebsd*}
1273 @uref{#h8300-hms,,h8300-hms}
1275 @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*}
1277 @uref{#hppa*-hp-hpux9,,hppa*-hp-hpux9}
1279 @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10}
1281 @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11}
1283 @uref{#i370-*-*,,i370-*-*}
1285 @uref{#*-*-linux-gnu,,*-*-linux-gnu}
1287 @uref{#ix86-*-linux*oldld,,i?86-*-linux*oldld}
1289 @uref{#ix86-*-linux*aout,,i?86-*-linux*aout}
1291 @uref{#ix86-*-linux*,,i?86-*-linux*}
1293 @uref{#ix86-*-sco,,i?86-*-sco}
1295 @uref{#ix86-*-sco3.2v4,,i?86-*-sco3.2v4}
1297 @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*}
1299 @uref{#ix86-*-udk,,i?86-*-udk}
1301 @uref{#ix86-*-isc,,i?86-*-isc}
1303 @uref{#ix86-*-esix,,i?86-*-esix}
1305 @uref{#ix86-ibm-aix,,i?86-ibm-aix}
1307 @uref{#ix86-sequent-bsd,,i?86-sequent-bsd}
1309 @uref{#ix86-sequent-ptx1*,,i?86-sequent-ptx1*, i?86-sequent-ptx2*}
1311 @uref{#ix86-*-sysv3*,,i?86-*-sysv3*}
1313 @uref{#i860-intel-osf*,,i860-intel-osf*}
1315 @uref{#*-lynx-lynxos,,*-lynx-lynxos}
1317 @uref{#*-ibm-aix*,,*-ibm-aix*}
1319 @uref{#m32r-*-elf,,m32r-*-elf}
1321 @uref{#m68000-hp-bsd,,m68000-hp-bsd}
1323 @uref{#m6811-elf,,m6811-elf}
1325 @uref{#m6812-elf,,m6812-elf}
1327 @uref{#m68k-altos,,m68k-altos}
1329 @uref{#m68k-apple-aux,,m68k-apple-aux}
1331 @uref{#m68k-att-sysv,,m68k-att-sysv}
1333 @uref{#m68k-bull-sysv,,m68k-bull-sysv}
1335 @uref{#m68k-crds-unox,,m68k-crds-unox}
1337 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
1339 @uref{#m68k-*-nextstep*,,m68k-*-nextstep*}
1341 @uref{#m68k-ncr-*,,m68k-ncr-*}
1343 @uref{#m68k-sun,,m68k-sun}
1345 @uref{#m68k-sun-sunos4.1.1,,m68k-sun-sunos4.1.1}
1347 @uref{#m88k-*-svr3,,m88k-*-svr3}
1349 @uref{#m88k-*-dgux,,m88k-*-dgux}
1351 @uref{#m88k-tektronix-sysv3,,m88k-tektronix-sysv3}
1353 @uref{#mips-*-*,,mips-*-*}
1355 @uref{#mips-dec-*,,mips-dec-*}
1357 @uref{#mips-mips-bsd,,mips-mips-bsd}
1359 @uref{#mips-mips-riscos*,,mips-mips-riscos*}
1361 @uref{#mips-sgi-irix4,,mips-sgi-irix4}
1363 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
1365 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
1367 @uref{#mips-sony-sysv,,mips-sony-sysv}
1369 @uref{#ns32k-encore,,ns32k-encore}
1371 @uref{#ns32k-*-genix,,ns32k-*-genix}
1373 @uref{#ns32k-sequent,,ns32k-sequent}
1375 @uref{#ns32k-utek,,ns32k-utek}
1377 @uref{#powerpc*-*-*,,powerpc*-*-*, powerpc-*-sysv4}
1379 @uref{#powerpc-*-darwin*,,powerpc-*-darwin*}
1381 @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4}
1383 @uref{#powerpc-*-linux-gnu*,,powerpc-*-linux-gnu*}
1385 @uref{#powerpc-*-netbsd*,,powerpc-*-netbsd*}
1387 @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix}
1389 @uref{#powerpc-*-eabisim,,powerpc-*-eabisim}
1391 @uref{#powerpc-*-eabi,,powerpc-*-eabi}
1393 @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
1395 @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim}
1397 @uref{#powerpcle-*-eabi,,powerpcle-*-eabi}
1399 @uref{#powerpcle-*-winnt,,powerpcle-*-winnt, powerpcle-*-pe}
1401 @uref{#romp-*-aos,,romp-*-aos, romp-*-mach}
1403 @uref{#s390-*-linux*}
1405 @uref{#s390x-*-linux*}
1407 @uref{#*-*-solaris*,,*-*-solaris*}
1409 @uref{#sparc-sun-solaris*,,sparc-sun-solaris*}
1411 @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7}
1413 @uref{#*-sun-solaris2.8,,*-sun-solaris2.8}
1415 @uref{#sparc-sun-sunos*,,sparc-sun-sunos*}
1417 @uref{#sparc-unknown-linux-gnulibc1,,sparc-unknown-linux-gnulibc1}
1419 @uref{#sparc64-*-*,,sparc64-*-*}
1421 @uref{#*-*-sysv*,,*-*-sysv*}
1423 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
1425 @uref{#we32k-*-*,,we32k-*-*}
1427 @uref{#windows,,Microsoft Windows}
1431 @uref{#older,,Older systems}
1436 @uref{#elf_targets,,all ELF targets} (SVR4, Solaris, etc.)
1442 <!-- -------- host/target specific issues start here ---------------- -->
1445 @heading @anchor{1750a-*-*}1750a-*-*
1446 MIL-STD-1750A processors.
1448 The MIL-STD-1750A cross configuration produces output for
1449 @code{as1750}, an assembler/linker available under the GNU General Public
1450 License for the 1750A@. @code{as1750} can be obtained at
1451 @uref{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}.
1452 A similarly licensed simulator for
1453 the 1750A is available from same address.
1455 You should ignore a fatal error during the building of @samp{libgcc}
1456 (@samp{libgcc} is not yet implemented for the 1750A@.)
1458 The @code{as1750} assembler requires the file @file{ms1750.inc}, which is
1459 found in the directory @file{gcc/config/1750a}.
1461 GCC produced the same sections as the Fairchild F9450 C Compiler,
1466 The program code section.
1469 The read/write (RAM) data section.
1472 The read-only (ROM) constants section.
1475 Initialization section (code to copy KREL to SREL)@.
1478 The smallest addressable unit is 16 bits (@code{BITS_PER_UNIT} is 16). This
1479 means that type @code{char} is represented with a 16-bit word per character.
1480 The 1750A's ``Load/Store Upper/Lower Byte'' instructions are not used by
1487 @heading @anchor{a29k}a29k
1488 AMD Am29k-family processors. These are normally used in embedded
1489 applications. There are no standard Unix configurations.
1491 corresponds to AMD's standard calling sequence and binary interface
1492 and is compatible with other 29k tools.
1494 You may need to make a variant of the file @file{a29k.h} for your
1495 particular configuration.
1501 @heading @anchor{a29k-*-bsd}a29k-*-bsd
1502 AMD Am29050 used in a system running a variant of BSD Unix.
1508 @heading @anchor{alpha*-*-*}alpha*-*-*
1510 This section contains general configuration information for all
1511 alpha-based platforms using ELF (in particular, ignore this section for
1512 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
1513 section, please read all other sections that match your target.
1515 We require binutils 2.11.2 or newer.
1516 Previous binutils releases had a number of problems with DWARF 2
1517 debugging information, not the least of which is incorrect linking of
1524 @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf*
1525 Systems using processors that implement the DEC Alpha architecture and
1526 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
1527 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
1529 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
1530 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
1531 we need to use the old assembler, invoked via the barely documented
1532 @option{-oldas} option. To bootstrap GCC, you either need to use the
1536 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
1539 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
1542 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
1545 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
1546 are supported on Tru64 UNIX, so you must not configure GCC with
1547 @option{--with-gnu-as} or @option{--with-gnu-ld}.
1549 The @option{--enable-threads} options isn't supported yet. A patch is
1550 in preparation for a future release. The Java runtime library has been
1551 reported to work on Tru64 UNIX V4.0F, V5.0, and V5.1, so you may try
1552 @option{--enable-libgcj} and report your results.
1554 GCC writes a @samp{.verstamp} directive to the assembler output file
1555 unless it is built as a cross-compiler. It gets the version to use from
1556 the system header file @file{/usr/include/stamp.h}. If you install a
1557 new version of DEC Unix, you should rebuild GCC to pick up the new version
1560 Note that since the Alpha is a 64-bit architecture, cross-compilers from
1561 32-bit machines will not generate code as efficient as that generated
1562 when the compiler is running on a 64-bit machine because many
1563 optimizations that depend on being able to represent a word on the
1564 target in an integral value on the host cannot be performed. Building
1565 cross-compilers on the Alpha for 32-bit machines has only been tested in
1566 a few cases and may not work properly.
1568 @code{make compare} may fail on old versions of DEC Unix unless you add
1569 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
1570 assembler input file is stored in the object file, and that makes
1571 comparison fail if it differs between the @code{stage1} and
1572 @code{stage2} compilations. The option @option{-save-temps} forces a
1573 fixed name to be used for the assembler input file, instead of a
1574 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
1575 unless the comparisons fail without that option. If you add
1576 @option{-save-temps}, you will have to manually delete the @samp{.i} and
1577 @samp{.s} files after each series of compilations.
1579 GCC now supports both the native (ECOFF) debugging format used by DBX
1580 and GDB and an encapsulated STABS format for use only with GDB@. See the
1581 discussion of the @option{--with-stabs} option of @file{configure} above
1582 for more information on these formats and how to select them.
1584 There is a bug in DEC's assembler that produces incorrect line numbers
1585 for ECOFF format when the @samp{.align} directive is used. To work
1586 around this problem, GCC will not emit such alignment directives
1587 while writing ECOFF format debugging information even if optimization is
1588 being performed. Unfortunately, this has the very undesirable
1589 side-effect that code addresses when @option{-O} is specified are
1590 different depending on whether or not @option{-g} is also specified.
1592 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
1593 DBX@. DEC is now aware of this problem with the assembler and hopes to
1594 provide a fix shortly.
1600 @heading @anchor{arc-*-elf}arc-*-elf
1601 Argonaut ARC processor.
1602 This configuration is intended for embedded systems.
1608 @heading @anchor{arm-*-aout}arm-*-aout
1609 Advanced RISC Machines ARM-family processors. These are often used in
1610 embedded applications. There are no standard Unix configurations.
1611 This configuration corresponds to the basic instruction sequences and will
1612 produce @file{a.out} format object modules.
1614 You may need to make a variant of the file @file{arm.h} for your particular
1621 @heading @anchor{arm-*-elf}arm-*-elf
1622 This configuration is intended for embedded systems.
1628 @heading @anchor{arm*-*-linux-gnu}arm*-*-linux-gnu
1630 We require GNU binutils 2.10 or newer.
1636 @heading @anchor{arm-*-riscix}arm-*-riscix
1637 The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix.
1638 If you are running a version of RISC iX prior to 1.2 then you must
1639 specify the version number during configuration. Note that the
1640 assembler shipped with RISC iX does not support stabs debugging
1641 information; a new version of the assembler, with stabs support
1642 included, is now available from Acorn and via ftp
1643 @uref{ftp://ftp.acorn.com/pub/riscix/as+xterm.tar.Z}. To enable stabs
1644 debugging, pass @option{--with-gnu-as} to configure.
1646 You will need to install GNU @command{sed} before you can run configure.
1652 @heading @anchor{avr}avr
1654 ATMEL AVR-family micro controllers. These are used in embedded
1655 applications. There are no standard Unix configurations.
1657 @xref{AVR Options,, AVR Options, gcc, Using and Porting the GNU Compiler
1661 See ``AVR Options'' in the main manual
1663 for the list of supported MCU types.
1665 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
1667 Further installation notes and other useful information about AVR tools
1668 can also be obtained from:
1672 @uref{http://home.overta.ru/users/denisc,,http://home.overta.ru/users/denisc}
1674 @uref{http://www.itnet.pl/amelektr/avr,,http://www.itnet.pl/amelektr/avr}
1677 We @emph{strongly} recommend using binutils 2.11 or newer.
1679 The following error:
1681 Error: register required
1684 indicates that you should upgrade to a newer version of the binutils.
1690 @heading @anchor{c4x}c4x
1692 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
1693 Processors. These are used in embedded applications. There are no
1694 standard Unix configurations. @xref{C4x Options,, C4x Options, gcc,
1695 Using and Porting the GNU Compiler Collection (GCC)}, for the list of
1696 supported MCU types.
1698 GCC can be configured as a cross compiler for both the C3x and C4x
1699 architectures on the same system. Use @samp{configure --target=c4x
1700 --enable-languages="c,c++"} to configure.
1703 Further installation notes and other useful information about C4x tools
1704 can also be obtained from:
1708 @uref{http://www.elec.canterbury.ac.nz/c4x,,http://www.elec.canterbury.ac.nz/c4x}
1715 @heading @anchor{dos}DOS
1717 Please have a look at our @uref{binaries.html,,binaries page}.
1719 You cannot install GCC by itself on MSDOS; it will not compile under
1720 any MSDOS compiler except itself. You need to get the complete
1721 compilation package DJGPP, which includes binaries as well as sources,
1722 and includes all the necessary compilation tools and libraries.
1728 @heading @anchor{dsp16xx}dsp16xx
1729 A port to the AT&T DSP1610 family of processors.
1735 @heading @anchor{*-*-freebsd*}*-*-freebsd*
1737 The version of binutils installed in @file{/usr/bin} is known to work unless
1738 otherwise specified in any per-architecture notes. However, binutils
1739 2.11 is known to improve overall testsuite results.
1741 For FreeBSD 1, FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
1742 configuration support and files as shipped with GCC 2.95 are still in
1743 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
1744 it is unknown which version of binutils was used (it is assumed that it
1745 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
1747 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
1748 default for all CPU architectures. It had been the default on
1749 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
1750 of @option{-g}, if you really want the old debugging format. There are
1751 no known issues with mixing object files and libraries with different
1752 debugging formats. Otherwise, this release of GCC should now match more
1753 of the configuration used in the stock FreeBSD configuration of GCC. In
1754 particular, @option{--enable-threads} is now configured by default.
1755 However, as a general user, do not attempt to replace the system
1756 compiler with this release. Known to bootstrap and check with good
1757 results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3 and 5-CURRENT@.
1759 At this time, @option{--enable-threads} is not compatible with
1760 @option{--enable-libgcj} on FreeBSD@.
1766 @heading @anchor{elxsi-elxsi-bsd}elxsi-elxsi-bsd
1767 The Elxsi's C compiler has known limitations that prevent it from
1768 compiling GCC@. Please contact @email{mrs@@wrs.com} for more details.
1774 @heading @anchor{h8300-hms}h8300-hms
1775 Hitachi H8/300 series of processors.
1777 Please have a look at our @uref{binaries.html,,binaries page}.
1779 The calling convention and structure layout has changed in release 2.6.
1780 All code must be recompiled. The calling convention now passes the
1781 first three arguments in function calls in registers. Structures are no
1782 longer a multiple of 2 bytes.
1788 @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux*
1790 We @emph{highly} recommend using gas/binutils 2.8 or newer on all hppa
1791 platforms; you may encounter a variety of problems when using the HP
1794 Specifically, @option{-g} does not work on HP-UX (since that system
1795 uses a peculiar debugging format which GCC does not know about), unless you
1796 use GAS and GDB and configure GCC with the
1797 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
1798 @option{--with-as=@dots{}} options.
1800 If you wish to use pa-risc 2.0 architecture support, you must use either
1801 the HP assembler, gas/binutils 2.11 or a recent
1802 @uref{ftp://sources.redhat.com/pub/binutils/snapshots,,snapshot of gas}.
1804 More specific information to @samp{hppa*-hp-hpux*} targets follows.
1810 @heading @anchor{hppa*-hp-hpux9}hppa*-hp-hpux9
1812 The HP assembler has major problems on this platform. We've tried to work
1813 around the worst of the problems. However, those workarounds may be causing
1814 linker crashes in some circumstances; the workarounds also probably prevent
1815 shared libraries from working. Use the GNU assembler to avoid these problems.
1818 The configuration scripts for GCC will also trigger a bug in the hpux9
1819 shell. To avoid this problem set @env{CONFIG_SHELL} to @file{/bin/ksh}
1820 and @env{SHELL} to @file{/bin/ksh} in your environment.
1827 @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10
1829 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
1830 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
1836 <a href="http://us-support.external.hp.com">US, Canada, Asia-Pacific, and
1840 @uref{http://us-support.external.hp.com,,}US, Canada, Asia-Pacific, and
1844 @uref{http://europe-support.external.hp.com,,Europe}
1847 The HP assembler on these systems is much better than the hpux9 assembler,
1848 but still has some problems. Most notably the assembler inserts timestamps
1849 into each object file it creates, causing the 3-stage comparison test to fail
1850 during a @samp{make bootstrap}. You should be able to continue by
1851 saying @samp{make all} after getting the failure from @samp{make
1859 @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11
1861 GCC 3.0 supports HP-UX 11. You must use GNU binutils 2.11 or above on
1868 @heading @anchor{i370-*-*}i370-*-*
1869 This port is very preliminary and has many known bugs. We hope to
1870 have a higher-quality port for this machine soon.
1876 @heading @anchor{*-*-linux-gnu}*-*-linux-gnu
1878 If you use glibc 2.2 (or 2.1.9x), GCC 2.95.2 won't install
1879 out-of-the-box. You'll get compile errors while building @samp{libstdc++}.
1880 The patch @uref{glibc-2.2.patch,,glibc-2.2.patch}, that is to be
1881 applied in the GCC source tree, fixes the compatibility problems.
1891 Currently Glibc 2.2.3 (and older releases) and GCC 3.0 are out of sync
1892 since the latest exception handling changes for GCC@. Compiling glibc
1893 with GCC 3.0 will give a binary incompatible glibc and therefore cause
1894 lots of problems and might make your system completly unusable. This
1895 will definitly need fixes in glibc but might also need fixes in GCC@. We
1896 strongly advise to wait for glibc 2.2.4 and to read the release notes of
1897 glibc 2.2.4 whether patches for GCC 3.0 are needed. You can use glibc
1898 2.2.3 with GCC 3.0, just do not try to recompile it.
1904 @heading @anchor{ix86-*-linux*oldld}i?86-*-linux*oldld
1905 Use this configuration to generate @file{a.out} binaries on Linux-based
1906 GNU systems if you do not have gas/binutils version 2.5.2 or later
1907 installed. This is an obsolete configuration.
1913 @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout
1914 Use this configuration to generate @file{a.out} binaries on Linux-based
1915 GNU systems. This configuration is being superseded. You must use
1916 gas/binutils version 2.5.2 or later.
1922 @heading @anchor{ix86-*-linux*}i?86-*-linux*
1924 You will need binutils 2.9.1.0.15 or newer for exception handling to work.
1926 If you receive Signal 11 errors when building on GNU/Linux, then it is
1927 possible you have a hardware problem. Further information on this can be
1928 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
1934 @heading @anchor{ix86-*-sco}i?86-*-sco
1935 Compilation with RCC is recommended. Also, it may be a good idea to
1936 link with GNU malloc instead of the malloc that comes with the system.
1942 @heading @anchor{ix86-*-sco3.2v4}i?86-*-sco3.2v4
1943 Use this configuration for SCO release 3.2 version 4.
1949 @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5*
1950 Use this for the SCO OpenServer Release 5 family of operating systems.
1952 Unlike earlier versions of GCC, the ability to generate COFF with this
1953 target is no longer provided.
1955 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
1956 the system debugger to be used. That support was too burdensome to
1957 maintain. GCC now emits only DWARF 2 for this target. This means you
1958 may use either the UDK debugger or GDB to debug programs built by this
1961 Use of the @option{-march=pentiumpro} flag can result in
1962 unrecognized opcodes when using the native assembler on OS versions before
1963 5.0.6. (Support for P6 opcodes was added to the native ELF assembler in
1964 that version.) While it's rather rare to see these emitted by GCC yet,
1965 errors of the basic form:
1968 /usr/tmp/ccaNlqBc.s:22:unknown instruction: fcomip
1969 /usr/tmp/ccaNlqBc.s:50:unknown instruction: fucomip
1972 are symptoms of this problem. You may work around this by not
1973 building affected files with that flag, by using the GNU assembler, or
1974 by using the assembler provided with the current version of the OS@.
1975 Users of GNU assembler should see the note below for hazards on doing
1978 The native SCO assembler that is provided with the OS at no
1979 charge is normally required. If, however, you must be able to use
1980 the GNU assembler (perhaps you're compiling code with asms that
1981 require GAS syntax) you may configure this package using the flags
1982 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You must
1983 use a recent version of GNU binutils; versions past 2.9.1 seem to work
1986 In general, the @option{--with-gnu-as} option isn't as well tested
1987 as the native assembler.
1989 Look in @file{gcc/config/i386/sco5.h} (search for ``messy'') for
1990 additional OpenServer-specific flags.
1992 Systems based on OpenServer before 5.0.4 (@samp{uname -X}
1993 will tell you what you're running) require TLS597 from
1994 @uref{ftp://ftp.sco.com/TLS/,,ftp://ftp.sco.com/TLS/}
1995 for C++ constructors and destructors to work right.
1997 The system linker in (at least) 5.0.4 and 5.0.5 will sometimes
1998 do the wrong thing for a construct that GCC will emit for PIC
1999 code. This can be seen as execution testsuite failures when using
2000 @option{-fPIC} on @file{921215-1.c}, @file{931002-1.c}, @file{nestfunc-1.c}, and @file{gcov-1.c}.
2001 For 5.0.5, an updated linker that will cure this problem is
2002 available. You must install both
2003 @uref{ftp://ftp.sco.com/Supplements/rs505a/,,ftp://ftp.sco.com/Supplements/rs505a/}
2004 and @uref{ftp://ftp.sco.com/SLS/,,OSS499A}.
2006 The dynamic linker in OpenServer 5.0.5 (earlier versions may show
2007 the same problem) aborts on certain G77-compiled programs. It's particularly
2008 likely to be triggered by building Fortran code with the @option{-fPIC} flag.
2009 Although it's conceivable that the error could be triggered by other
2010 code, only G77-compiled code has been observed to cause this abort.
2011 If you are getting core dumps immediately upon execution of your
2012 G77 program---and especially if it's compiled with @option{-fPIC}---try applying
2013 @uref{sco_osr5_g77.patch,,@file{sco_osr5_g77.patch}} to your @samp{libf2c} and
2015 Affected faults, when analyzed in a debugger, will show a stack
2016 backtrace with a fault occurring in @code{rtld()} and the program
2017 running as @file{/usr/lib/ld.so.1}. This problem has been reported to SCO
2018 engineering and will hopefully be addressed in later releases.
2025 @heading @anchor{ix86-*-udk}i?86-*-udk
2027 This target emulates the SCO Universal Development Kit and requires that
2028 package be installed. (If it is installed, you will have a
2029 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
2030 @samp{i?86-*-unixware7*} target
2031 but is meant to be used when hosting on a system where UDK isn't the
2032 default compiler such as OpenServer 5 or Unixware 2. This target will
2033 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
2034 with the same warnings and caveats as the SCO UDK@.
2036 This target is a little tricky to build because we have to distinguish
2037 it from the native tools (so it gets headers, startups, and libraries
2038 from the right place) while making the tools not think we're actually
2039 building a cross compiler. The easiest way to do this is with a configure
2042 @samp{CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure
2043 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-}
2045 @emph{You should substitute @samp{i686} in the above command with the appropriate
2046 processor for your host.}
2048 After the usual @samp{make bootstrap} and
2049 @samp{make install}, you can then access the UDK-targeted GCC
2050 tools by adding @command{udk-} before the commonly known name. For
2051 example, to invoke the C compiler, you would use @command{udk-gcc}.
2052 They will coexist peacefully with any native-target GCC tools you may
2060 @heading @anchor{ix86-*-isc}i?86-*-isc
2061 It may be a good idea to link with GNU malloc instead of the malloc that
2062 comes with the system.
2064 In ISC version 4.1, @command{sed} core dumps when building
2065 @file{deduced.h}. Use the version of @command{sed} from version 4.0.
2071 @heading @anchor{ix86-*-esix}i?86-*-esix
2072 It may be good idea to link with GNU malloc instead of the malloc that
2073 comes with the system.
2079 @heading @anchor{ix86-ibm-aix}i?86-ibm-aix
2080 You need to use GAS version 2.1 or later, and LD from
2081 GNU binutils version 2.2 or later.
2087 @heading @anchor{ix86-sequent-bsd}i?86-sequent-bsd
2088 Go to the Berkeley universe before compiling.
2094 @heading @anchor{ix86-sequent-ptx1*}i?86-sequent-ptx1*, i?86-sequent-ptx2*
2095 You must install GNU @command{sed} before running @command{configure}.
2101 @heading @anchor{#ix86-*-sysv3*}i?86-*-sysv3*
2102 The @code{fixproto} shell script may trigger a bug in the system shell.
2103 If you encounter this problem, upgrade your operating system or
2104 use @command{bash} (the GNU shell) to run @code{fixproto}.
2111 @heading @anchor{i860-intel-osf*}i860-intel-osf*
2112 On the Intel Paragon (an i860 machine), if you are using operating
2113 system version 1.0, you will get warnings or errors about redefinition
2114 of @code{va_arg} when you build GCC@.
2116 If this happens, then you need to link most programs with the library
2117 @file{iclib.a}. You must also modify @file{stdio.h} as follows: before
2121 #if defined(__i860__) && !defined(_VA_LIST)
2122 #include <va_list.h>
2136 extern int vprintf(const char *, va_list );
2137 extern int vsprintf(char *, const char *, va_list );
2145 #endif /* __PGC__ */
2148 These problems don't exist in operating system version 1.1.
2154 @heading @anchor{*-lynx-lynxos}*-lynx-lynxos
2155 LynxOS 2.2 and earlier comes with GCC 1.x already installed as
2156 @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}.
2157 You can tell GCC to use the GNU assembler and linker, by specifying
2158 @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce
2159 COFF format object files and executables; otherwise GCC will use the
2160 installed tools, which produce @file{a.out} format executables.
2165 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
2167 @heading @anchor{*-ibm-aix*}*-ibm-aix*
2169 AIX Make frequently has problems with GCC makefiles. GNU Make 3.76 or
2170 newer is recommended to build on this platform.
2172 Errors involving @code{alloca} when building GCC generally are due
2173 to an incorrect definition of @code{CC} in the Makefile or mixing files
2174 compiled with the native C compiler and GCC@. During the stage1 phase of
2175 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
2176 (not @command{xlc}). Once @command{configure} has been informed of
2177 @command{xlc}, one needs to use @samp{make distclean} to remove the
2178 configure cache files and ensure that @env{CC} environment variable
2179 does not provide a definition that will confuse @command{configure}.
2180 If this error occurs during stage2 or later, then the problem most likely
2181 is the version of Make (see above).
2183 Binutils 2.10 does not support AIX 4.3. Binutils available from the
2184 @uref{http://www-1.ibm.com/servers/aix/products/aixos/linux/,,AIX
2185 Toolbox for Linux: GNU and Open Source tools for AIX};
2186 website does work. Binutils 2.11 is expected to include AIX 4.3
2187 support. The GNU Assembler is necessary for @samp{libstdc++} to build. The
2188 AIX native @command{ld} still is recommended. The native AIX tools do
2189 interoperate with GCC@.
2191 Linking executables and shared libraries may produce warnings of
2192 duplicate symbols. The assembly files generated by GCC for AIX always
2193 have included multiple symbol definitions for certain global variable
2194 and function declarations in the original program. The warnings should
2195 not prevent the linker from producing a correct library or runnable
2198 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
2199 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
2200 to parse archive libraries did not handle the new format correctly.
2201 These routines are used by GCC and result in error messages during
2202 linking such as ``not a COFF file''. The version of the routines shipped
2203 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
2204 option of the archive command may be used to create archives of 32-bit
2205 objects using the original ``small format''. A correct version of the
2206 routines is shipped with AIX 4.3.2.
2208 Some versions of the AIX binder (linker) can fail with a relocation
2209 overflow severe error when the @option{-bbigtoc} option is used to link
2210 GCC-produced object files into an executable that overflows the TOC@. A fix
2211 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
2212 available from IBM Customer Support and from its
2213 @uref{http://service.boulder.ibm.com/,,service.boulder.ibm.com}
2214 website as PTF U455193.
2216 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
2217 with a segmentation fault when invoked by any version of GCC@. A fix for
2218 APAR IX87327 is available from IBM Customer Support and from its
2219 @uref{http://service.boulder.ibm.com/,,service.boulder.ibm.com}
2220 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
2222 The initial assembler shipped with AIX 4.3.0 generates incorrect object
2223 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
2224 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
2225 @uref{http://service.boulder.ibm.com/,,service.boulder.ibm.com}
2226 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
2228 AIX provides National Language Support (NLS)@. Compilers and assemblers
2229 use NLS to support locale-specific representations of various data
2230 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
2231 separating decimal fractions). There have been problems reported where
2232 GCC does not produce the same floating-point formats that the assembler
2233 expects. If one encounters this problem, set the @env{LANG}
2234 environment variable to @samp{C} or @samp{En_US}.
2236 By default, GCC for AIX 4.1 and above produces code that can be used on
2237 both Power or PowerPC processors.
2239 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2240 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2246 @heading @anchor{m32r-*-elf}m32r-*-elf
2247 Mitsubishi M32R processor.
2248 This configuration is intended for embedded systems.
2254 @heading @anchor{m68000-hp-bsd}m68000-hp-bsd
2255 HP 9000 series 200 running BSD@. Note that the C compiler that comes
2256 with this system cannot compile GCC; contact @email{law@@cygnus.com}
2257 to get binaries of GCC for bootstrapping.
2263 @heading @anchor{m6811-elf}m6811-elf
2264 Motorola 68HC11 family micro controllers. These are used in embedded
2265 applications. There are no standard Unix configurations.
2271 @heading @anchor{m6812-elf}m6812-elf
2272 Motorola 68HC12 family micro controllers. These are used in embedded
2273 applications. There are no standard Unix configurations.
2279 @heading @anchor{m68k-altos}m68k-altos
2280 Altos 3068. You must use the GNU assembler, linker and debugger.
2281 Also, you must fix a kernel bug.
2287 @heading @anchor{m68k-apple-aux}m68k-apple-aux
2288 Apple Macintosh running A/UX@.
2289 You may configure GCC to use either the system assembler and
2290 linker or the GNU assembler and linker. You should use the GNU configuration
2291 if you can, especially if you also want to use G++. You enable
2292 that configuration with the @option{--with-gnu-as} and @option{--with-gnu-ld}
2293 options to @code{configure}.
2295 Note the C compiler that comes
2296 with this system cannot compile GCC@. You can find binaries of GCC
2297 for bootstrapping on @code{jagubox.gsfc.nasa.gov}.
2298 You will also a patched version of @file{/bin/ld} there that
2299 raises some of the arbitrary limits found in the original.
2305 @heading @anchor{m68k-att-sysv}m68k-att-sysv
2306 AT&T 3b1, a.k.a.@: 7300 PC@. This version of GCC cannot
2307 be compiled with the system C compiler, which is too buggy.
2308 You will need to get a previous version of GCC and use it to
2309 bootstrap. Binaries are available from the OSU-CIS archive, at
2310 @uref{ftp://archive.cis.ohio-state.edu/pub/att7300/}.
2316 @heading @anchor{m68k-bull-sysv}m68k-bull-sysv
2317 Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GCC works
2318 either with native assembler or GNU assembler. You can use
2319 GNU assembler with native COFF generation by providing @option{--with-gnu-as} to
2320 the configure script or use GNU assembler with stabs-in-COFF encapsulation
2321 by providing @samp{--with-gnu-as --stabs}. For any problem with the native
2322 assembler or for availability of the DPX/2 port of GAS, contact
2323 @email{F.Pierresteguy@@frcl.bull.fr}.
2329 @heading @anchor{m68k-crds-unox}m68k-crds-unox
2330 Use @samp{configure unos} for building on Unos.
2332 The Unos assembler is named @code{casm} instead of @code{as}. For some
2333 strange reason linking @file{/bin/as} to @file{/bin/casm} changes the
2334 behavior, and does not work. So, when installing GCC, you should
2335 install the following script as @file{as} in the subdirectory where
2336 the passes of GCC are installed:
2343 The default Unos library is named @file{libunos.a} instead of
2344 @file{libc.a}. To allow GCC to function, either change all
2345 references to @option{-lc} in @file{gcc.c} to @option{-lunos} or link
2346 @file{/lib/libc.a} to @file{/lib/libunos.a}.
2348 @cindex @code{alloca}, for Unos
2349 When compiling GCC with the standard compiler, to overcome bugs in
2350 the support of @code{alloca}, do not use @option{-O} when making stage 2.
2351 Then use the stage 2 compiler with @option{-O} to make the stage 3
2352 compiler. This compiler will have the same characteristics as the usual
2353 stage 2 compiler on other systems. Use it to make a stage 4 compiler
2354 and compare that with stage 3 to verify proper compilation.
2356 (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in
2357 the comments there will make the above paragraph superfluous. Please
2358 inform us of whether this works.)
2360 Unos uses memory segmentation instead of demand paging, so you will need
2361 a lot of memory. 5 Mb is barely enough if no other tasks are running.
2362 If linking @file{cc1} fails, try putting the object files into a library
2363 and linking from that library.
2369 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
2370 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
2371 the assembler that prevents compilation of GCC@. This
2372 bug manifests itself during the first stage of compilation, while
2373 building @file{libgcc2.a}:
2377 cc1: warning: `-g' option not supported on this version of GCC
2378 cc1: warning: `-g1' option not supported on this version of GCC
2379 ./xgcc: Internal compiler error: program as got fatal signal 11
2382 A patched version of the assembler is available as the file
2383 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
2384 have HP software support, the patch can also be obtained directly from
2385 HP, as described in the following note:
2388 This is the patched assembler, to patch SR#1653-010439, where the
2389 assembler aborts on floating point constants.
2391 The bug is not really in the assembler, but in the shared library
2392 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
2393 SR#4701-078451. Anyway, the attached assembler uses the archive
2394 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
2397 This patch is also known as PHCO_4484.
2399 In addition, if you wish to use gas, you must use
2400 gas version 2.1 or later, and you must use the GNU linker version 2.1 or
2401 later. Earlier versions of gas relied upon a program which converted the
2402 gas output into the native HP-UX format, but that program has not been
2403 kept up to date. gdb does not understand that native HP-UX format, so
2404 you must use gas if you wish to use gdb.
2406 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
2407 @code{fixproto} shell script triggers a bug in the system shell. If you
2408 encounter this problem, upgrade your operating system or use BASH (the
2409 GNU shell) to run @code{fixproto}. This bug will cause the fixproto
2410 program to report an error of the form:
2413 ./fixproto: sh internal 1K buffer overflow
2416 To fix this, you can also change the first line of the fixproto script
2428 @heading @anchor{m68k-*-nextstep*}m68k-*-nextstep*
2430 Current GCC versions probably do not work on version 2 of the NeXT
2433 On NeXTStep 3.0, the Objective-C compiler does not work, due,
2434 apparently, to a kernel bug that it happens to trigger. This problem
2435 does not happen on 3.1.
2437 You absolutely @strong{must} use GNU sed and GNU make on this platform.
2440 On NeXTSTEP 3.x where x < 3 the build of GCC will abort during
2441 stage1 with an error message like this:
2445 /usr/tmp/ccbbsZ0U.s:987:Unknown pseudo-op: .section
2446 /usr/tmp/ccbbsZ0U.s:987:Rest of line ignored. 1st junk character
2450 The reason for this is the fact that NeXT's assembler for these
2451 versions of the operating system does not support the @samp{.section}
2452 pseudo op that's needed for full C++ exception functionality.
2454 As NeXT's assembler is a derived work from GNU as, a free
2455 replacement that does can be obtained at
2456 @uref{ftp://ftp.next.peak.org:/next-ftp/next/apps/devtools/as.3.3.NIHS.s.tar.gz,,ftp://ftp.next.peak.org:/next-ftp/next/apps/devtools/as.3.3.NIHS.s.tar.gz}.
2458 If you try to build the integrated C++ & C++ runtime libraries on this system
2459 you will run into trouble with include files. The way to get around this is
2460 to use the following sequence. Note you must have write permission to
2461 the directory @var{prefix} you specified in the configuration process of GCC
2462 for this sequence to work.
2466 make all-texinfo all-bison all-byacc all-binutils all-gas all-ld
2469 make install-headers-tar
2478 @heading @anchor{m68k-ncr-*}m68k-ncr-*
2479 On the Tower models 4@var{n}0 and 6@var{n}0, by default a process is not
2480 allowed to have more than one megabyte of memory. GCC cannot compile
2481 itself (or many other programs) with @option{-O} in that much memory.
2483 To solve this problem, reconfigure the kernel adding the following line
2484 to the configuration file:
2495 @heading @anchor{m68k-sun}m68k-sun
2496 Sun 3. We do not provide a configuration file to use the Sun FPA by
2497 default, because programs that establish signal handlers for floating
2498 point traps inherently cannot work with the FPA@.
2504 @heading @anchor{m68k-sun-sunos4.1.1}m68k-sun-sunos4.1.1
2506 It is reported that you may need the GNU assembler on this platform.
2513 @heading @anchor{m88k-*-svr3}m88k-*-svr3
2514 Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port.
2515 These systems tend to use the Green Hills C, revision 1.8.5, as the
2516 standard C compiler. There are apparently bugs in this compiler that
2517 result in object files differences between stage 2 and stage 3. If this
2518 happens, make the stage 4 compiler and compare it to the stage 3
2519 compiler. If the stage 3 and stage 4 object files are identical, this
2520 suggests you encountered a problem with the standard C compiler; the
2521 stage 3 and 4 compilers may be usable.
2523 It is best, however, to use an older version of GCC for bootstrapping
2530 @heading @anchor{m88k-*-dgux}m88k-*-dgux
2531 Motorola m88k running DG/UX@. To build 88open BCS native or cross
2532 compilers on DG/UX, specify the configuration name as
2533 @samp{m88k-*-dguxbcs} and build in the 88open BCS software development
2534 environment. To build ELF native or cross compilers on DG/UX, specify
2535 @samp{m88k-*-dgux} and build in the DG/UX ELF development environment.
2536 You set the software development environment by issuing
2537 @samp{sde-target} command and specifying either @samp{m88kbcs} or
2538 @samp{m88kdguxelf} as the operand.
2540 If you do not specify a configuration name, @file{configure} guesses the
2541 configuration based on the current software development environment.
2547 @heading @anchor{m88k-tektronix-sysv3}m88k-tektronix-sysv3
2548 Tektronix XD88 running UTekV 3.2e. Do not turn on
2549 optimization while building stage1 if you bootstrap with
2550 the buggy Green Hills compiler. Also, the bundled LAI
2551 System V NFS is buggy so if you build in an NFS mounted
2552 directory, start from a fresh reboot, or avoid NFS all together.
2553 Otherwise you may have trouble getting clean comparisons
2560 @heading @anchor{mips-*-*}mips-*-*
2561 If you use the 1.31 version of the MIPS assembler (such as was shipped
2562 with Ultrix 3.1), you will need to use the @option{-fno-delayed-branch} switch
2563 when optimizing floating point code. Otherwise, the assembler will
2564 complain when the GCC compiler fills a branch delay slot with a
2565 floating point instruction, such as @code{add.d}.
2567 If on a MIPS system you get an error message saying ``does not have gp
2568 sections for all it's [sic] sectons [sic]'', don't worry about it. This
2569 happens whenever you use GAS with the MIPS linker, but there is not
2570 really anything wrong, and it is okay to use the output file. You can
2571 stop such warnings by installing the GNU linker.
2573 It would be nice to extend GAS to produce the gp tables, but they are
2574 optional, and there should not be a warning about their absence.
2576 Users have reported some problems with version 2.0 of the MIPS
2577 compiler tools that were shipped with Ultrix 4.1. Version 2.10
2578 which came with Ultrix 4.2 seems to work fine.
2580 Users have also reported some problems with version 2.20 of the
2581 MIPS compiler tools that were shipped with RISC/os 4.x. The earlier
2582 version 2.11 seems to work fine.
2584 Some versions of the MIPS linker will issue an assertion failure
2585 when linking code that uses @code{alloca} against shared
2586 libraries on RISC-OS 5.0, and DEC's OSF/1 systems. This is a bug
2587 in the linker, that is supposed to be fixed in future revisions.
2588 To protect against this, GCC passes @option{-non_shared} to the
2589 linker unless you pass an explicit @option{-shared} or
2590 @option{-call_shared} switch.
2592 @heading @anchor{mips-mips-bsd}mips-mips-bsd
2593 MIPS machines running the MIPS operating system in BSD mode. It's
2594 possible that some old versions of the system lack the functions
2595 @code{memcpy}, @code{memmove}, @code{memcmp}, and @code{memset}. If your
2596 system lacks these, you must remove or undo the definition of
2597 @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}.
2599 If you use the MIPS C compiler to bootstrap, it may be necessary
2600 to increase its table size for switch statements with the
2601 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2602 optimization option, you also need to use @option{-Olimit 3000}.
2603 Both of these options are automatically generated in the
2604 @file{Makefile} that the shell script @file{configure} builds.
2605 If you override the @code{CC} make variable and use the MIPS
2606 compilers, you may need to add @option{-Wf,-XNg1500 -Olimit 3000}.
2612 @heading @anchor{mips-dec-*}mips-dec-*
2613 MIPS-based DECstations can support three different personalities:
2614 Ultrix, DEC OSF/1, and OSF/rose. (Alpha-based DECstation products have
2615 a configuration name beginning with @samp{alpha*-dec}.) To configure GCC
2616 for these platforms use the following configurations:
2619 @item mips-dec-ultrix
2620 Ultrix configuration.
2623 DEC's version of OSF/1.
2625 @item mips-dec-osfrose
2626 Open Software Foundation reference port of OSF/1 which uses the
2627 OSF/rose object file format instead of ECOFF@. Normally, you
2628 would not select this configuration.
2631 If you use the MIPS C compiler to bootstrap, it may be necessary
2632 to increase its table size for switch statements with the
2633 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2634 optimization option, you also need to use @option{-Olimit 3000}.
2635 Both of these options are automatically generated in the
2636 @file{Makefile} that the shell script @file{configure} builds.
2637 If you override the @code{CC} make variable and use the MIPS
2638 compilers, you may need to add @option{-Wf,-XNg1500 -Olimit 3000}.
2644 @heading @anchor{mips-mips-riscos*}mips-mips-riscos*
2645 If you use the MIPS C compiler to bootstrap, it may be necessary
2646 to increase its table size for switch statements with the
2647 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2648 optimization option, you also need to use @option{-Olimit 3000}.
2649 Both of these options are automatically generated in the
2650 @file{Makefile} that the shell script @file{configure} builds.
2651 If you override the @code{CC} make variable and use the MIPS
2652 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
2654 MIPS computers running RISC-OS can support four different
2655 personalities: default, BSD 4.3, System V.3, and System V.4
2656 (older versions of RISC-OS don't support V.4). To configure GCC
2657 for these platforms use the following configurations:
2660 @item mips-mips-riscos@var{rev}
2661 Default configuration for RISC-OS, revision @var{rev}.
2663 @item mips-mips-riscos@var{rev}bsd
2664 BSD 4.3 configuration for RISC-OS, revision @var{rev}.
2666 @item mips-mips-riscos@var{rev}sysv4
2667 System V.4 configuration for RISC-OS, revision @var{rev}.
2673 @item mips-mips-riscos@var{rev}sysv
2674 System V.3 configuration for RISC-OS, revision @var{rev}.
2677 The revision @code{rev} mentioned above is the revision of
2678 RISC-OS to use. You must reconfigure GCC when going from a
2679 RISC-OS revision 4 to RISC-OS revision 5. This has the effect of
2680 avoiding a linker bug.
2686 @heading @anchor{mips-sgi-irix4}mips-sgi-irix4
2688 In order to compile GCC on an SGI running IRIX 4, the ``c.hdr.lib''
2689 option must be installed from the CD-ROM supplied from Silicon Graphics.
2690 This is found on the 2nd CD in release 4.0.1.
2692 On IRIX version 4.0.5F, and perhaps on some other versions as well,
2693 there is an assembler bug that reorders instructions incorrectly. To
2694 work around it, specify the target configuration
2695 @samp{mips-sgi-irix4loser}. This configuration inhibits assembler
2698 In a compiler configured with target @samp{mips-sgi-irix4}, you can turn
2699 off assembler optimization by using the @option{-noasmopt} option. This
2700 compiler option passes the option @option{-O0} to the assembler, to
2703 The @option{-noasmopt} option can be useful for testing whether a problem
2704 is due to erroneous assembler reordering. Even if a problem does not go
2705 away with @option{-noasmopt}, it may still be due to assembler
2706 reordering---perhaps GCC itself was miscompiled as a result.
2708 You may get the following warning on IRIX 4 platforms, it can be safely
2711 warning: foo.o does not have gp tables for all its sections.
2718 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
2720 This configuration has considerable problems, which will be fixed in a
2723 In order to compile GCC on an SGI running IRIX 5, the ``compiler_dev.hdr''
2724 subsystem must be installed from the IDO CD-ROM supplied by Silicon
2725 Graphics. It is also available for download from
2726 @uref{http://www.sgi.com/developers/devtools/apis/ido.html,,http://www.sgi.com/developers/devtools/apis/ido.html}.
2728 @code{make compare} may fail on version 5 of IRIX unless you add
2729 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2730 assembler input file is stored in the object file, and that makes
2731 comparison fail if it differs between the @code{stage1} and
2732 @code{stage2} compilations. The option @option{-save-temps} forces a
2733 fixed name to be used for the assembler input file, instead of a
2734 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2735 unless the comparisons fail without that option. If you do you
2736 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2737 @samp{.s} files after each series of compilations.
2739 If you use the MIPS C compiler to bootstrap, it may be necessary
2740 to increase its table size for switch statements with the
2741 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2742 optimization option, you also need to use @option{-Olimit 3000}.
2744 To enable debugging under IRIX 5, you must use GNU @command{as} 2.11.2
2746 and use the @option{--with-gnu-as} configure option when configuring GCC.
2747 GNU @command{as} is distributed as part of the binutils package.
2748 When using release 2.11.2, you need to apply a patch
2749 @uref{http://sources.redhat.com/ml/binutils/2001-07/msg00352.html,,http://sources.redhat.com/ml/binutils/2001-07/msg00352.html}
2750 which will be included in the next release of binutils.
2752 When building GCC, the build process loops rebuilding @command{cc1} over
2753 and over again. This happens on @samp{mips-sgi-irix5.2}, and possibly
2754 other platforms. It has been reported that this is a known bug in the
2755 @command{make} shipped with IRIX 5.2. We recommend you use GNU
2756 @command{make} instead of the vendor supplied @command{make} program;
2757 however, you may have success with @command{smake} on IRIX 5.2 if you do
2758 not have GNU @command{make} available.
2764 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
2766 If you are using IRIX @command{cc} as your bootstrap compiler, you must
2767 ensure that the N32 ABI is in use. To test this, compile a simple C
2768 file with @command{cc} and then run @command{file} on the
2769 resulting object file. The output should look like:
2772 test.o: ELF N32 MSB @dots{}
2778 test.o: ELF 32-bit MSB @dots{}
2784 test.o: ELF 64-bit MSB @dots{}
2787 then your version of @command{cc} uses the O32 or N64 ABI by default. You
2788 should set the environment variable @env{CC} to @samp{cc -n32}
2789 before configuring GCC@.
2791 GCC on IRIX 6 is usually built to support both the N32 and N64 ABIs. If
2792 you build GCC on a system that doesn't have the N64 libraries installed,
2793 you need to configure with @option{--disable-multilib} so GCC doesn't
2794 try to use them. Look for @file{/usr/lib64/libc.so.1} to see if you
2795 have the 64-bit libraries installed.
2797 You must @emph{not} use GNU @command{as} (which isn't built anyway as of
2798 binutils 2.11.2) on IRIX 6 platforms; doing so will only cause problems.
2800 GCC does not currently support generating O32 ABI binaries in the
2801 @samp{mips-sgi-irix6} configurations. It is possible to create a GCC
2802 with O32 ABI only support by configuring it for the @samp{mips-sgi-irix5}
2803 target and using a patched GNU @command{as} 2.11.2 as documented in the
2804 @uref{#mips-sgi-irix5,,@samp{mips-sgi-irix5}} section above. Using the
2805 native assembler requires patches to GCC which will be included in a
2806 future release. It is
2807 expected that O32 ABI support will be available again in a future release.
2809 The @option{--enable-threads} option doesn't currently work, a patch is
2810 in preparation for a future release. The @option{--enable-libgcj}
2811 option is disabled by default: IRIX 6 uses a very low default limit
2812 (20480) for the command line length. Although libtool contains a
2813 workaround for this problem, at least the N64 @samp{libgcj} is known not
2814 to build despite this, running into an internal error of the native
2815 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
2816 its maximum of 262144 bytes. If you have root access, you can use the
2817 @command{systune} command to do this.
2819 GCC does not correctly pass/return structures which are
2820 smaller than 16 bytes and which are not 8 bytes. The problem is very
2821 involved and difficult to fix. It affects a number of other targets also,
2822 but IRIX 6 is affected the most, because it is a 64 bit target, and 4 byte
2823 structures are common. The exact problem is that structures are being padded
2824 at the wrong end, e.g.@: a 4 byte structure is loaded into the lower 4 bytes
2825 of the register when it should be loaded into the upper 4 bytes of the
2828 GCC is consistent with itself, but not consistent with the SGI C compiler
2829 (and the SGI supplied runtime libraries), so the only failures that can
2830 happen are when there are library functions that take/return such
2831 structures. There are very few such library functions. Currently this
2832 is known to affect @code{inet_ntoa}, @code{inet_lnaof},
2833 @code{inet_netof}, @code{inet_makeaddr}, and @code{semctl}. Until the
2834 bug is fixed, GCC contains workarounds for the known affected functions.
2836 See @uref{http://freeware.sgi.com/,,http://freeware.sgi.com/} for more
2837 information about using GCC on IRIX platforms.
2843 @heading @anchor{mips-sony-sysv}mips-sony-sysv
2844 Sony MIPS NEWS@. This works in NEWSOS 5.0.1, but not in 5.0.2 (which
2845 uses ELF instead of COFF)@. Support for 5.0.2 will probably be provided
2846 soon by volunteers. In particular, the linker does not like the
2847 code generated by GCC when shared libraries are linked in.
2854 @heading @anchor{ns32k-encore}ns32k-encore
2855 Encore ns32000 system. Encore systems are supported only under BSD@.
2861 @heading @anchor{ns32k-*-genix}ns32k-*-genix
2862 National Semiconductor ns32000 system. Genix has bugs in @code{alloca}
2863 and @code{malloc}; you must get the compiled versions of these from GNU
2870 @heading @anchor{ns32k-sequent}ns32k-sequent
2871 Go to the Berkeley universe before compiling.
2877 @heading @anchor{ns32k-utek}ns32k-utek
2878 UTEK ns32000 system (``merlin''). The C compiler that comes with this
2879 system cannot compile GCC; contact @samp{tektronix!reed!mason} to get
2880 binaries of GCC for bootstrapping.
2887 @heading @anchor{powerpc*-*-*}powerpc-*-*
2889 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2890 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2896 @heading @anchor{powerpc-*-darwin*}powerpc-*-darwin*
2897 PowerPC running Darwin (Mac OS X kernel).
2899 GCC 3.0 does not support Darwin, but 3.1 and later releases will work.
2901 Pre-installed versions of Mac OS X may not include any developer tools,
2902 meaning that you will not be able to build GCC from source. Tool
2903 binaries are available at
2904 @uref{http://www.opensource.apple.com/projects/darwin} (free
2905 registration required).
2907 Versions of the assembler prior to ``cctools-364'' cannot handle the
2908 4-argument form of rlwinm and related mask-using instructions. Darwin
2909 1.3 (Mac OS X 10.0) uses cctools-353 for instance. To get cctools-364,
2910 check out @file{cctools} with tag @samp{Apple-364}, build it, and
2911 install the assembler as @file{usr/bin/as}. See
2912 @uref{http://www.opensource.apple.com/tools/cvs/docs.html} for details.
2914 Also, the default stack limit of 512K is too small, and a bootstrap will
2915 typically fail when self-compiling @file{expr.c}. Set the stack to 800K
2916 or more, for instance by doing @samp{limit stack 800}. It's also
2917 convenient to use the GNU preprocessor instead of Apple's during the
2918 first stage of bootstrapping; this is automatic when doing @samp{make
2919 bootstrap}, but to do it from the toplevel objdir you will need to say
2920 @samp{make CC='cc -no-cpp-precomp' bootstrap}.
2922 Note that the version of GCC shipped by Apple typically includes a
2923 number of extensions not available in a standard GCC release. These
2924 extensions are generally specific to Mac programming.
2930 @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4
2931 PowerPC system in big endian mode, running System V.4.
2937 @heading @anchor{powerpc-*-linux-gnu*}powerpc-*-linux-gnu*
2940 @uref{ftp://ftp.varesearch.com/pub/support/hjl/binutils,,binutils 2.9.4.0.8}
2941 or newer for a working GCC@. It is strongly recommended to recompile binutils
2942 if you initially built it with gcc-2.7.2.x.
2948 @heading @anchor{powerpc-*-netbsd*}powerpc-*-netbsd*
2949 PowerPC system in big endian mode running NetBSD@. To build the
2950 documentation you will need Texinfo version 4.0 (NetBSD 1.5.1 included
2951 Texinfo version 3.12).
2957 @heading @anchor{powerpc-*-eabiaix}powerpc-*-eabiaix
2958 Embedded PowerPC system in big endian mode with @option{-mcall-aix} selected as
2965 @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim
2966 Embedded PowerPC system in big endian mode for use in running under the
2973 @heading @anchor{powerpc-*-eabi}powerpc-*-eabi
2974 Embedded PowerPC system in big endian mode.
2980 @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4
2981 PowerPC system in little endian mode, running System V.4.
2987 @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim
2988 Embedded PowerPC system in little endian mode for use in running under
2995 @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi
2996 Embedded PowerPC system in little endian mode.
3002 @heading @anchor{powerpcle-*-winnt}powerpcle-*-winnt, powerpcle-*-pe
3003 PowerPC system in little endian mode running Windows NT@.
3009 @heading @anchor{romp-*-aos}romp-*-aos, romp-*-mach
3010 The only operating systems supported for the IBM RT PC are AOS and
3011 MACH@. GCC does not support AIX running on the RT@. We recommend you
3012 compile GCC with an earlier version of itself; if you compile GCC
3013 with @command{hc}, the Metaware compiler, it will work, but you will get
3014 mismatches between the stage 2 and stage 3 compilers in various files.
3015 These errors are minor differences in some floating-point constants and
3016 can be safely ignored; the stage 3 compiler is correct.
3022 @heading @anchor{s390-*-linux*}s390-*-linux*
3023 S/390 system running Linux for S/390@.
3029 @heading @anchor{s390x-*-linux*}s390x-*-linux*
3030 zSeries system (64 Bit) running Linux for zSeries@.
3036 @heading @anchor{*-*-solaris*}*-*-solaris*
3038 Sun does not ship a C compiler with Solaris. To bootstrap and install
3039 GCC you first have to install a pre-built compiler, see our
3040 @uref{binaries.html,,binaries page} for details.
3042 Solaris' @file{/bin/sh} will often fail to configure @file{libstdc++-v3}, @file{boehm-gc} or
3043 @file{libjava}. If you encounter this problem, set @env{CONFIG_SHELL} to
3044 @file{/bin/ksh} in your environment and run @command{make bootstrap} again.
3045 Another possibility that sometimes helps is to remove
3046 @file{*-*-solaris*/config.cache}.
3048 Solaris 2 comes with a number of optional OS packages. Some of these
3049 packages are needed to use GCC fully, namely @code{SUNWarc},
3050 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3051 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3052 optional packages when installing Solaris, you will need to verify that
3053 the packages that GCC needs are installed.
3055 To check whether an optional package is installed, use
3056 the @command{pkginfo} command. To add an optional package, use the
3057 @command{pkgadd} command. For further details, see the Solaris
3060 On some versions of Solaris, trying to use the linker and other tools in
3061 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3062 For example, the linker may hang indefinitely. The fix is to remove
3063 @file{/usr/ucb} from your @code{PATH}.
3065 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3066 platform. We recommend the use of GNU binutils 2.11.2 or the vendor
3067 tools (Sun @command{as}, Sun @command{ld}).
3069 Unfortunately, C++ shared libraries, including @samp{libstdc++}, won't work
3070 properly if assembled with Sun @command{as}: the linker will complain about
3071 relocations in read-only sections, in the definition of virtual
3072 tables. Also, Sun @command{as} fails to process long symbols resulting from
3073 mangling template-heavy C++ function names.
3079 @heading @anchor{sparc-sun-solaris*}sparc-sun-solaris*
3081 Sun @command{as} 4.X is broken in that it cannot cope with long symbol names.
3082 A typical error message might look similar to the following:
3084 @samp{/usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041:
3085 error: can't compute value of an expression involving an external symbol.}
3087 This is Sun bug 4237974. This is fixed with patch 108908-02 and has
3088 been fixed in later (5.x) versions of the assembler.
3094 @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7
3096 Sun patch 107058-01 (1999-01-13) for SPARC Solaris 7 triggers a bug in
3097 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3098 and later, including all EGCS releases. Sun formerly recommended
3099 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3100 recommend it only for people who use Sun's compilers.
3102 Here are some workarounds to this problem:
3105 Do not install Sun patch 107058-01 until after Sun releases a
3106 complete patch for bug 4210064. This is the simplest course to take,
3107 unless you must also use Sun's C compiler. Unfortunately 107058-01
3108 is preinstalled on some new Solaris-based hosts, so you may have to
3112 Copy the original, unpatched Solaris 7
3113 @command{/usr/ccs/bin/as} into
3114 @command{/usr/local/lib/gcc-lib/sparc-sun-solaris2.7/3.0/as},
3115 adjusting the latter name to fit your local conventions and software
3119 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3120 both 107058-01 and 106950-03 installed has reported the bug with GCC
3121 and Sun's dynamic linker. This last course of action is riskiest,
3122 for two reasons. First, you must install 106950 on all hosts that
3123 run code generated by GCC; it doesn't suffice to install it only on
3124 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3125 only a partial fix for bug 4210064, but Sun doesn't know whether the
3126 partial fix is adequate for GCC@. Revision -08 or later should fix
3127 the bug, but (as of 1999-10-06) it is still being tested.
3134 <!-- ripped from the same FAQ that I answered -->
3136 @heading @anchor{*-sun-solaris2.8}*-sun-solaris2.8
3138 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3139 newer: @command{g++} will complain that types are missing. These headers assume
3140 that omitting the type means @code{int}; this assumption worked for C89 but
3141 is wrong for C++, and is now wrong for C99 also.
3143 @command{g++} accepts such (illegal) constructs with the option @option{-fpermissive}; it
3144 will assume that any missing type is @code{int} (as defined by C89).
3146 For Solaris 8, this is fixed by revision 24 or later of patch 108652
3147 (for SPARCs) or 108653 (for Intels).
3149 Solaris 8's linker fails to link some @samp{libjava} programs if
3150 previously-installed GCC java libraries already exist in the configured
3151 prefix. For this reason, @samp{libgcj} is disabled by default on Solaris 8.
3152 If you use GNU ld, or if you don't have a previously-installed @samp{libgcj} in
3153 the same prefix, use @option{--enable-libgcj} to build and install the
3160 @heading @anchor{sparc-sun-sunos*}sparc-sun-sunos*
3162 A bug in the SunOS 4 linker will cause it to crash when linking
3163 @option{-fPIC} compiled objects (and will therefore not allow you to build
3166 To fix this problem you can either use the most recent version of
3167 binutils or get the latest SunOS 4 linker patch (patch ID 100170-10)
3168 from Sun's patch site.
3170 Sometimes on a Sun 4 you may observe a crash in the program
3171 @command{genflags} or @command{genoutput} while building GCC. This is said to
3172 be due to a bug in @command{sh}. You can probably get around it by running
3173 @command{genflags} or @command{genoutput} manually and then retrying the
3180 @heading @anchor{sparc-unknown-linux-gnulibc1}sparc-unknown-linux-gnulibc1
3182 It has been reported that you might need
3183 @uref{ftp://ftp.yggdrasil.com/private/hjl,,binutils 2.8.1.0.23}
3184 for this platform, too.
3191 @heading @anchor{sparc64-*-*}sparc64-*-*
3193 GCC version 2.95 is not able to compile code correctly for
3194 @code{sparc64} targets. Users of the Linux kernel, at least,
3195 can use the @code{sparc32} program to start up a new shell
3196 invocation with an environment that causes @command{configure} to
3197 recognize (via @samp{uname -a}) the system as @samp{sparc-*-*} instead.
3204 @heading @anchor{#*-*-sysv*}*-*-sysv*
3205 On System V release 3, you may get this error message
3209 ld fatal: failed to write symbol name @var{something}
3210 in strings table for file @var{whatever}
3213 This probably indicates that the disk is full or your ulimit won't allow
3214 the file to be as large as it needs to be.
3216 This problem can also result because the kernel parameter @code{MAXUMEM}
3217 is too small. If so, you must regenerate the kernel and make the value
3218 much larger. The default value is reported to be 1024; a value of 32768
3219 is said to work. Smaller values may also work.
3221 On System V, if you get an error like this,
3224 /usr/local/lib/bison.simple: In function `yyparse':
3225 /usr/local/lib/bison.simple:625: virtual memory exhausted
3229 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
3231 On a System V release 4 system, make sure @file{/usr/bin} precedes
3232 @file{/usr/ucb} in @code{PATH}. The @code{cc} command in
3233 @file{/usr/ucb} uses libraries which have bugs.
3239 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3240 Don't try compiling with VAX C (@code{vcc}). It produces incorrect code
3241 in some cases (for example, when @code{alloca} is used).
3247 @heading @anchor{we32k-*-*}we32k-*-*
3248 These computers are also known as the 3b2, 3b5, 3b20 and other similar
3249 names. (However, the 3b1 is actually a 68000.)
3251 Don't use @option{-g} when compiling with the system's compiler. The
3252 system's linker seems to be unable to handle such a large program with
3253 debugging information.
3255 The system's compiler runs out of capacity when compiling @file{stmt.c}
3256 in GCC@. You can work around this by building @file{cpp} in GCC
3257 first, then use that instead of the system's preprocessor with the
3258 system's C compiler to compile @file{stmt.c}. Here is how:
3261 mv /lib/cpp /lib/cpp.att
3263 echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp
3267 The system's compiler produces bad code for some of the GCC
3268 optimization files. So you must build the stage 2 compiler without
3269 optimization. Then build a stage 3 compiler with optimization.
3270 That executable should work. Here are the necessary commands:
3273 make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g"
3275 make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O"
3278 You may need to raise the ULIMIT setting to build a C++ compiler,
3279 as the file @file{cc1plus} is larger than one megabyte.
3285 @heading @anchor{windows}Microsoft Windows (32 bit)
3287 A port of GCC 2.95.x is included with the
3288 @uref{http://www.cygwin.com/,,Cygwin environment}.
3290 Current (as of early 2001) snapshots of GCC will build under Cygwin
3291 without modification.
3297 @heading @anchor{os2}OS/2
3299 GCC does not currently support OS/2. However, Andrew Zabolotny has been
3300 working on a generic OS/2 port with pgcc. The current code code can be found
3301 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
3303 An older copy of GCC 2.8.1 is included with the EMX tools available at
3304 @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,,
3305 ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}.
3311 @heading @anchor{older}Older systems
3313 GCC contains support files for many older (1980s and early
3314 1990s) Unix variants. For the most part, support for these systems
3315 has not been deliberately removed, but it has not been maintained for
3316 several years and may suffer from bitrot. Support from some systems
3317 has been removed from GCC 3: fx80, ns32-ns-genix, pyramid, tahoe,
3318 gmicro, spur; most of these targets had not been updated since GCC
3321 Support for older systems as targets for cross-compilation is less
3322 problematic than support for them as hosts for GCC; if an enthusiast
3323 wishes to make such a target work again (including resurrecting any
3324 of the targets that never worked with GCC 2, starting from the last
3325 CVS version before they were removed), patches
3326 @uref{../contribute.html,,following the usual requirements}
3327 would be likely to be accepted, since they should not affect the
3328 support for more modern targets.
3330 Support for old systems as hosts for GCC can cause problems if the
3331 workarounds for compiler, library and operating system bugs affect the
3332 cleanliness or maintainability of the rest of GCC@. In some cases, to
3333 bring GCC up on such a system, if still possible with current GCC, may
3334 require first installing an old version of GCC which did work on that
3335 system, and using it to compile a more recent GCC, to avoid bugs in
3336 the vendor compiler. Old releases of GCC 1 and GCC 2 are available in
3337 the @file{old-releases} directory on the
3338 @uref{../mirrors.html,,GCC mirror sites}. Header bugs may generally
3339 be avoided using @command{fixincludes}, but bugs or deficiencies in
3340 libraries and the operating system may still cause problems.
3342 For some systems, old versions of GNU binutils may also be useful,
3343 and are available from @file{pub/binutils/old-releases} on
3344 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
3346 Some of the information on specific systems above relates to
3347 such older systems, but much of the information
3348 about GCC on such systems (which may no longer be applicable to
3349 current GCC) is to be found in the GCC texinfo manual.
3355 @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris, etc.)
3357 C++ support is significantly better on ELF targets if you use the
3358 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
3359 inlines, vtables and template instantiations will be discarded
3369 @uref{./index.html,,Return to the GCC Installation page}
3373 @c ***************************************************************************
3374 @c Part 6 The End of the Document
3376 @comment node-name, next, previous, up
3377 @node Concept Index, , Specific, Top
3381 @unnumbered Concept Index