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 @comment $Id: install.texi,v 1.24 2001/06/14 22:51:17 jsm28 Exp $
37 @c Copyright (C) 2001 Free Software Foundation, Inc.
38 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
40 @c Include everything if we're not making html
52 @c Part 2 Summary Description and Copyright
55 Copyright @copyright{} 2001 Free Software Foundation, Inc.
58 @c Part 3 Titlepage and Copyright
61 @comment The title is printed in a large font.
62 @center @titlefont{Installing GCC}
64 @c The following two commands start the copyright page.
66 @vskip 0pt plus 1filll
67 Copyright @copyright{} 2001 Free Software Foundation, Inc.
70 @c Part 4 Top node and Master Menu
73 @comment node-name, next, Previous, up
76 * Installing GCC:: This document describes the generic installation
77 procedure for GCC as well as detailing some target
78 specific installation instructions.
80 * Specific:: Host/target specific installation notes for GCC.
81 * Binaries:: Where to get pre-compiled binaries.
83 * Concept Index:: This index has two entries.
87 @c Part 5 The Body of the Document
88 @c ***Installing GCC**********************************************************
90 @comment node-name, next, previous, up
91 @node Installing GCC, Binaries, , Top
95 <h1 align="center">Installing GCC</h1>
98 @chapter Installing GCC
101 The latest version of this document is always available at
102 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
104 This document describes the generic installation procedure for GCC as well
105 as detailing some target specific installation instructions.
107 GCC includes several components that previously were separate distributions
108 with their own installation instructions. This document supersedes all
109 package specific installation instructions. We provide the component
110 specific installation information in the source distribution for historical
111 reference purposes only.
113 @emph{Before} starting the build/install procedure please check the
115 @xref{Specific, host/target specific installation notes}.
118 @uref{specific.html,,host/target specific installation notes}.
120 We recommend you browse the entire generic installation instructions before
123 The installation procedure itself is broken into five steps.
127 * Downloading the source::
130 * Testing:: (optional)
137 @uref{download.html,,Downloading the source}
139 @uref{configure.html,,Configuration}
141 @uref{build.html,,Building}
143 @uref{test.html,,Testing} (optional)
145 @uref{finalinstall.html,,Final install}
149 Please note that GCC does not support @samp{make uninstall} and probably
150 won't do so in the near future as this would open a can of worms. Instead,
151 we suggest that you install GCC into a directory of its own and simply
152 remove that directory when you do not need that specific version of GCC
160 @uref{./index.html,,Return to the GCC Installation page}
164 @c ***Downloading the source**************************************************
166 @comment node-name, next, previous, up
167 @node Downloading the source, Configuration, , Installing GCC
171 <h1 align="center">Downloading GCC</h1>
174 @chapter Downloading GCC
176 @cindex Downloading GCC
177 @cindex Downloading the Source
179 GCC is distributed via CVS and FTP tarballs compressed with @command{gzip} or
180 @command{bzip2}. It is possible to download a full distribution or specific
183 Please refer to our @uref{http://gcc.gnu.org/releases.html,,releases web page}
184 for information on how to obtain GCC.
186 The full distribution includes the C, C++, Objective-C, Fortran, Java,
187 and Chill compilers. The full distribution also includes runtime libraries
188 for C++, Objective-C, Fortran, Java and Chill. (GCC 3.0 does not
189 include Chill. Releases before 3.0 do not include the Java runtime
190 library.) In GCC 3.0 and later versions, GNU compiler testsuites
191 are also included in the full distribution.
193 If you choose to download specific components, you must download the core
194 gcc distribution plus any language specific distributions you wish to
195 use. The core distribution includes the C language front-end as well as the
196 shared components. Each language has a tarball which includes the language
197 front-end as well as the language runtime (when appropriate).
199 Unpack the core distribution as well as any language specific
200 distributions in the same directory.
202 If you also intend to build binutils (either to upgrade an existing
203 installation or for use in place of the corresponding tools of your
204 OS), unpack the binutils distribution either in the same directory or
205 a separate one. In the latter case, add symbolic links to any
206 components of the binutils you intend to build alongside the compiler
207 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
208 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
215 @uref{./index.html,,Return to the GCC Installation page}
219 @c ***Configuration***********************************************************
221 @comment node-name, next, previous, up
222 @node Configuration, Building, Downloading the source, Installing GCC
226 <h1 align="center">Installing GCC: Configuration</h1>
229 @chapter Installing GCC: Configuration
231 @cindex Configuration
232 @cindex Installing GCC: Configuration
234 Like most GNU software, GCC must be configured before it can be built.
235 This document describes the recommended configuration procedure
236 for both native and cross targets.
238 We use @var{srcdir} to refer to the toplevel source directory for
239 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
241 If you obtained the sources via CVS, @var{srcdir} must refer to the top
242 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
243 and not its @file{gcc} subdirectory, otherwise the build will fail.
245 First, we @strong{highly} recommend that GCC be built into a
246 separate directory than the sources which does @strong{not} reside
247 within the source tree. This is how we generally build GCC; building
248 where @var{srcdir} == @var{objdir} should still work, but doesn't
249 get extensive testing; building where @var{objdir} is a subdirectory
250 of @var{srcdir} is unsupported.
252 If you have built GNU CC previously in the same directory for a
253 different target machine, do @samp{make distclean} to delete all files
254 that might be invalid. One of the files this deletes is
255 @file{Makefile}; if @samp{make distclean} complains that @file{Makefile}
256 does not exist, it probably means that the directory is already suitably
257 clean. However, with the recommended method of building in a separate
258 @var{objdir}, you should simply use a different @var{objdir} for each
261 Second, when configuring a native system, either @command{cc} or
262 @command{gcc} must be in your path or you must set @env{CC} in
263 your environment before running configure. Otherwise the configuration
271 % @var{srcdir}/configure [@var{target}] [@var{options}]
275 @heading Target specification
278 GCC has code to correctly determine the correct value for @var{target}
279 for nearly all native systems. Therefore, we highly recommend you not
280 provide a configure target when configuring a native compiler.
283 @var{target} must be specified as @option{--target=@var{target}}
284 when configuring a cross compiler; examples of valid targets would be
285 i960-rtems, m68k-coff, sh-elf, etc.
288 Specifying just @var{target} instead of @option{--target=@var{target}}
289 implies that the host defaults to @var{target}.
293 @heading Options specification
295 Use @var{options} to override several configure time options for
296 GCC. A partial list of supported @var{options}:
299 @item --prefix=@var{dirname}
300 Specify the toplevel installation
301 directory. This is the recommended way to install the tools into a directory
302 other than the default. The toplevel installation directory defaults to
305 We @strong{highly} recommend against @var{dirname} being the same or a
306 subdirectory of @var{objdir} or vice versa.
308 These additional options control where certain parts of the distribution
309 are installed. Normally you should not need to use these options.
312 @item --exec-prefix=@var{dirname}
313 Specify the toplevel installation directory for architecture-dependent
314 files. The default is @file{@var{prefix}}.
316 @item --bindir=@var{dirname}
317 Specify the installation directory for the executables called by users
318 (such as @command{gcc} and @command{g++}). The default is
319 @file{@var{exec-prefix}/bin}.
321 @item --libdir=@var{dirname}
322 Specify the installation directory for object code libraries and
323 internal parts of GCC. The default is @file{@var{exec-prefix}/lib}.
325 @item --with-slibdir=@var{dirname}
326 Specify the installation directory for the shared libgcc library. The
327 default is @file{@var{libdir}}.
329 @item --infodir=@var{dirname}
330 Specify the installation directory for documentation in info format.
331 The default is @file{@var{prefix}/info}.
333 @item --mandir=@var{dirname}
334 Specify the installation directory for manual pages. The default is
335 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
336 the full GCC manuals, which are provided in Texinfo format. The
337 @command{g77} manpage is unmaintained and may be out of date; the others
338 are derived by an automatic conversion process from parts of the full
341 @item --with-gxx-include-dir=@var{dirname}
343 the installation directory for g++ header files. The default is
344 @file{@var{prefix}/include/g++-v3}.
348 @item --with-local-prefix=@var{dirname}
350 installation directory for local include files. The default is
351 @file{/usr/local}. Specify this option if you want the compiler to
352 search directory @file{@var{dirname}/include} for locally installed
353 header files @emph{instead} of @file{/usr/local/include}.
355 You should specify @option{--with-local-prefix} @strong{only} if your
356 site has a different convention (not @file{/usr/local}) for where to put
359 The default value for @option{--with-local-prefix} is @file{/usr/local}
360 regardless of the value of @option{--prefix}. Specifying
361 @option{--prefix} has no effect on which directory GCC searches for
362 local header files. This may seem counterintuitive, but actually it is
365 The purpose of @option{--prefix} is to specify where to @emph{install
366 GCC}. The local header files in @file{/usr/local/include}---if you put
367 any in that directory---are not part of GCC. They are part of other
368 programs---perhaps many others. (GCC installs its own header files in
369 another directory which is based on the @option{--prefix} value.)
371 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
372 The directory you use for @option{--with-local-prefix} @strong{must not}
373 contain any of the system's standard header files. If it did contain
374 them, certain programs would be miscompiled (including GNU Emacs, on
375 certain targets), because this would override and nullify the header
376 file corrections made by the @code{fixincludes} script.
378 Indications are that people who use this option use it based on mistaken
379 ideas of what it is for. People use it as if it specified where to
380 install part of GCC. Perhaps they make this assumption because
381 installing GCC creates the directory.
383 @item --enable-shared[=@var{package}[,@dots{}]]
384 Build shared versions of libraries, if shared libraries are supported on
385 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
386 are enabled by default on all platforms that support shared libraries.
388 If a list of packages is given as an argument, build shared libraries
389 only for the listed packages. For other packages, only static libraries
390 will be built. Package names currently recognized in the GCC tree are
391 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
392 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc} and
393 @samp{libjava}. Note that @samp{libobjc} does not recognize itself by
394 any name, so, if you list package names in @option{--enable-shared},
395 you'll only get static Objective C libraries. @samp{libf2c} and
396 @samp{libiberty} do not support shared libraries at all.
398 Use @option{--disable-shared} to build only static libraries. Note that
399 @option{--disable-shared} does not accept a list of package names as
400 argument, only @option{--enable-shared} does.
402 @item @anchor{with-gnu-as}--with-gnu-as
403 Specify that the compiler should assume that the
404 assembler it finds is the GNU assembler. However, this does not modify
405 the rules to find an assembler and will result in confusion if found
406 assembler is not actually the GNU assembler. If you have more than one
407 assembler installed on your system, you may want to use this option in
408 connection with @option{--with-as=@file{/path/to/gas}}.
410 @item --with-as=@file{/path/to/as}
412 compiler should use the assembler pointed to by @var{pathname}, rather
413 than the one found by the standard rules to find an assembler, which
418 @file{@var{exec_prefix}/lib/gcc-lib/@var{target}/@var{version}}
419 directory, where @var{exec_prefix} defaults to @var{prefix} which
420 defaults to @file{/usr/local} unless overridden by the
421 @option{--prefix=/pathname} switch described above. @var{target} is the
422 target system triple, such as @var{sparc-sun-solaris2.7}, and
423 @var{version} denotes the GCC version, such as 2.95.2.
425 Check operating system specific directories (e.g. @file{/usr/ccs/bin} on
428 Note that these rules do not check for the value of @env{PATH}. You may
429 want to use @option{--with-as} if no assembler is installed in the
430 directories listed above, or if you have multiple assemblers installed
431 and want to choose one that is not found by the above rules.
433 @item @anchor{with-gnu-ld}--with-gnu-ld
434 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
438 @item --with-ld=@file{/path/to/ld}
440 @option{--with-as}, but for the linker.
443 Specify that stabs debugging
444 information should be used instead of whatever format the host normally
445 uses. Normally GCC uses the same debug format as the host system.
447 @item --enable-multilib
448 Specify that multiple target
449 libraries should be built to support different target variants, calling
450 conventions, etc. This is the default.
452 @item --enable-threads
453 Specify that the target
454 supports threads. This affects the Objective-C compiler and runtime
455 library, and exception handling for other languages like C++ and Java.
456 On some systems, this is the default.
458 @item --enable-threads=@var{lib}
460 @var{lib} is the thread support library. This affects the Objective-C
461 compiler and runtime library, and exception handling for other languages
462 like C++ and Java. The possibilities for @var{lib} are:
470 DEC OSF/1 thread support.
472 SGI IRIX thread support.
474 Generic MACH thread support, known to work on NEXTSTEP.
476 IBM OS/2 thread support.
478 Generic POSIX thread support.
480 Same as @samp{posix}.
482 Disable thread support, should work for all platforms.
484 SUN Solaris thread support.
486 VxWorks thread support.
488 Microsoft Win32 API thread support.
491 @item --with-cpu=@var{cpu}
492 Specify which cpu variant the
493 compiler should generate code for by default. This is currently
494 only supported on the some ports, specifically arm, powerpc, and
495 SPARC. If configure does not recognize the model name (e.g. arm700,
496 603e, or ultrasparc) you provide, please check the configure script
497 for a complete list of supported models.
499 @item --enable-target-optspace
501 libraries should be optimized for code space instead of code speed.
502 This is the default for the m32r platform.
505 Specify that a user visible @command{cpp} program should not be installed.
507 @item --with-cpp-install-dir=@var{dirname}
508 Specify that the user visible @command{cpp} program should be installed
509 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
511 @item --enable-maintainer-mode
513 regenerate the GCC master message catalog @file{gcc.pot} are normally
514 disabled. This is because it can only be rebuilt if the complete source
515 tree is present. If you have changed the sources and want to rebuild the
516 catalog, configuring with @option{--enable-maintainer-mode} will enable
517 this. Note that you need a recent version of the @code{gettext} tools
520 @item --enable-version-specific-runtime-libs
522 that runtime libraries should be installed in the compiler specific
523 subdirectory (@file{@var{libsubdir}}) rather than the usual places. In
524 addition, libstdc++'s include files will be installed in
525 @file{@var{libsubdir}/include/g++} unless you overruled it by using
526 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
527 particularly useful if you intend to use several versions of GCC in
528 parallel. This is currently supported by @samp{libf2c} and
531 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
532 Specify that only a particular subset of compilers and
533 their runtime libraries should be built. For a list of valid values for
534 @var{langN} you can issue the following command in the
535 @file{gcc} directory of your GCC source tree:@* @samp{grep language=
536 */config-lang.in}@* Currently, you can use any of the following:
537 @code{c++}, @code{f77}, @code{java} and @code{objc}.
538 @code{CHILL} is not currently maintained, and will almost
539 certainly fail to compile. Note that this switch does not work with
540 EGCS 1.1.2 or older versions of egcs. It is supported in GCC 2.95
541 and newer versions.@*
542 If you do not pass this flag, all languages available in the @file{gcc}
543 sub-tree will be configured. Re-defining @code{LANGUAGES} when calling
544 @samp{make bootstrap} @strong{does not} work anymore, as those
545 language sub-directories might not have been configured!
547 @item --disable-libgcj
548 Specify that the run-time libraries
549 used by GCJ should not be built. This is useful in case you intend
550 to use GCJ with some other run-time, or you're going to install it
551 separately, or it just happens not to build on your particular
552 machine. In general, if the Java front-end is enabled, the GCJ
553 libraries will be enabled too, unless they're known to not work on
554 the target platform. If GCJ is enabled but libgcj isn't built, you
555 may need to port it; in this case, before modifying the top-level
556 @file{configure.in} so that libgcj is enabled by default on this platform,
557 you may use @option{--enable-libgcj} to override the default.
560 Specify that the compiler should
561 use DWARF2 debugging information as the default.
563 @item --enable-win32-registry
564 @itemx --enable-win32-registry=@var{KEY}
565 @itemx --disable-win32-registry
566 The @option{--enable-win32-registry} option enables Windows-hosted GCC
567 to look up installations paths in the registry using the following key:
570 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{KEY}}
573 @var{KEY} defaults to GCC version number, and can be overridden by the
574 @option{--enable-win32-registry=@var{KEY}} option. Vendors and distributors
575 who use custom installers are encouraged to provide a different key,
576 perhaps one comprised of vendor name and GCC version number, to
577 avoid conflict with existing installations. This feature is enabled
578 by default, and can be disabled by @option{--disable-win32-registry}
579 option. This option has no effect on the other hosts.
582 Specify that the machine does not have a floating point unit. This
583 option only applies to @samp{m68k-sun-sunos@var{n}} and
584 @samp{m68k-isi-bsd}. On any other system, @option{--nfp} has no effect.
586 @item --enable-checking
587 @itemx --enable-checking=@var{list}
588 When you specify this option, the compiler is built to perform checking
589 of tree node types when referencing fields of that node, and some other
590 internal consistency checks. This does not change the generated code,
591 but adds error checking within the compiler. This will slow down the
592 compiler and may only work properly if you are building the compiler
593 with GNU C. This is on by default when building from CVS or snapshots,
594 but off for releases. More control over the checks may be had by
595 specifying @var{list}; the categories of checks available are
596 @samp{misc}, @samp{tree}, @samp{gc}, @samp{rtl} and @samp{gcac}. The
597 default when @var{list} is not specified is @samp{misc,tree,gc}; the
598 checks @samp{rtl} and @samp{gcac} are very expensive.
602 The @option{--enable-nls} option enables Native Language Support (NLS),
603 which lets GCC output diagnostics in languages other than American
604 English. Native Language Support is enabled by default if not doing a
605 canadian cross build. The @option{--disable-nls} option disables NLS.
607 @item --with-included-gettext
608 If NLS is enabled, the @option{--with-included-gettext} option causes the build
609 procedure to prefer its copy of GNU @code{gettext}.
612 If NLS is enabled, and if the host lacks @code{gettext} but has the
613 inferior @code{catgets} interface, the GCC build procedure normally
614 ignores @code{catgets} and instead uses GCC's copy of the GNU
615 @code{gettext} library. The @option{--with-catgets} option causes the
616 build procedure to use the host's @code{catgets} in this situation.
619 Some options which only apply to building cross compilers:
621 @item --with-headers=@var{dir}
622 Specifies a directory
623 which has target include files.
624 @emph{This options is required} when building a cross
625 compiler, if @file{@var{prefix}/@var{target}/sys-include} doesn't pre-exist.
626 These include files will be copied into the @file{gcc} install directory.
627 Fixincludes will be run on these files to make them compatible with
629 @item --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
630 Specifies a list of directories which contain the target runtime
631 libraries. These libraries will be copied into the @file{gcc} install
634 Specifies that ``newlib'' is
635 being used as the target C library. This causes @code{__eprintf} to be
636 omitted from libgcc.a on the assumption that it will be provided by
640 Note that each @option{--enable} option has a corresponding
641 @option{--disable} option and that each @option{--with} option has a
642 corresponding @option{--without} option.
649 @uref{./index.html,,Return to the GCC Installation page}
653 @c ***Building****************************************************************
655 @comment node-name, next, previous, up
656 @node Building, Testing, Configuration, Installing GCC
660 <h1 align="center">Installing GCC: Building</h1>
665 @cindex Installing GCC: Building
667 Now that GCC is configured, you are ready to build the compiler and
670 We @strong{highly} recommend that GCC be built using GNU make;
671 other versions may work, then again they might not.
673 (For example, many broken versions of make will fail if you use the
674 recommended setup where @var{objdir} is different from @var{srcdir}.
675 Other broken versions may recompile parts of the compiler when
676 installing the compiler.)
678 Some commands executed when making the compiler may fail (return a
679 non-zero status) and be ignored by @code{make}. These failures, which
680 are often due to files that were not found, are expected, and can safely
683 It is normal to have compiler warnings when compiling certain files.
684 Unless you are a GCC developer, you can generally ignore these warnings
685 unless they cause compilation to fail.
687 On certain old systems, defining certain environment variables such as
688 @env{CC} can interfere with the functioning of @command{make}.
690 If you encounter seemingly strange errors when trying to build the
691 compiler in a directory other than the source directory, it could be
692 because you have previously configured the compiler in the source
693 directory. Make sure you have done all the necessary preparations.
695 If you build GCC on a BSD system using a directory stored in an old System
696 V file system, problems may occur in running @code{fixincludes} if the
697 System V file system doesn't support symbolic links. These problems
698 result in a failure to fix the declaration of @code{size_t} in
699 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
700 that type mismatches occur, this could be the cause.
702 The solution is not to use such a directory for building GCC.
704 When building from CVS or snapshots, or if you modify parser sources,
705 you need the Bison parser generator installed. Any version 1.25 or
706 later should work; older versions may also work. If you do not modify
707 parser sources, releases contain the Bison-generated files and you do
708 not need Bison installed to build them.
710 When building from CVS or snapshots, or if you modify Texinfo
711 documentation, you need version 4.0 or later of Texinfo installed if you
712 want Info documentation to be regenerated. Releases contain Info
713 documentation pre-built for the unmodified documentation in the release.
715 @section Building a native compiler
717 For a native build issue the command @samp{make bootstrap}. This
718 will build the entire GCC system, which includes the following steps:
722 Build host tools necessary to build the compiler such as texinfo, bison,
726 Build target tools for use by the compiler such as binutils (bfd,
727 binutils, gas, gprof, ld, and opcodes)@*
728 if they have been individually linked
729 or moved into the top level GCC source tree before configuring.
732 Perform a 3-stage bootstrap of the compiler.
735 Perform a comparison test of the stage2 and stage3 compilers.
738 Build runtime libraries using the stage3 compiler from the previous step.
742 If you are short on disk space you might consider @samp{make
743 bootstrap-lean} instead. This is identical to @samp{make
744 bootstrap} except that object files from the stage1 and
745 stage2 of the 3-stage bootstrap of the compiler are deleted as
746 soon as they are no longer needed.
749 If you want to save additional space during the bootstrap and in
750 the final installation as well, you can build the compiler binaries
751 without debugging information with @samp{make CFLAGS='-O' LIBCFLAGS='-g
752 -O2' LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap}. This will save
753 roughly 40% of disk space both for the bootstrap and the final installation.
754 (Libraries will still contain debugging information.)
756 If you wish to use non-default flags when compiling the stage2 and
757 stage3 compile, set @code{BOOT_CFLAGS} on the command line when doing
758 @samp{make bootstrap}. Non-default optimization flags are less well
759 tested here than the default of @samp{-g -O2}, but should still work.
760 In a few cases, you may find that you need to specify special flags such
761 as @option{-msoft-float} here to complete the bootstrap; or, if the
762 native compiler miscompiles the stage1 compiler, you may need to work
763 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
764 stage1 compiler that were miscompiled, or by using @samp{make
765 bootstrap4} to increase the number of stages of bootstrap.
767 If you used the flag @option{--enable-languages=@dots{}} to restrict
768 the compilers to be built, only those you've actually enabled will be
769 built. This will of course only build those runtime libraries, for
770 which the particular compiler has been built. Please note,
771 that re-defining LANGUAGES when calling @samp{make bootstrap}
772 @strong{does not} work anymore!
774 If the comparison of stage2 and stage3 fails, this normally indicates
775 that the stage 2 compiler has compiled GCC incorrectly, and is therefore
776 a potentially serious bug which you should investigate and report. (On
777 a few systems, meaningful comparison of object files is impossible; they
778 always appear ``different''. If you encounter this problem, you will
779 need to disable comparison in the @file{Makefile}.)
781 @section Building a cross compiler
783 We recommend reading the
784 @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ}
785 for information about building cross compilers.
787 When building a cross compiler, it is not generally possible to do a
788 3-stage bootstrap of the compiler. This makes for an interesting problem
789 as parts of GCC can only be built with GCC.
791 To build a cross compiler, we first recommend building and installing a
792 native compiler. You can then use the native GCC compiler to build the
795 Assuming you have already installed a native copy of GCC and configured
796 your cross compiler, issue the command @command{make}, which performs the
801 Build host tools necessary to build the compiler such as texinfo, bison,
805 Build target tools for use by the compiler such as binutils (bfd,
806 binutils, gas, gprof, ld, and opcodes)
807 if they have been individually linked or moved into the top level GCC source
808 tree before configuring.
811 Build the compiler (single stage only).
814 Build runtime libraries using the compiler from the previous step.
817 Note that if an error occurs in any step the make process will exit.
819 @section Building in parallel
821 If you have a multiprocessor system you can use @samp{make bootstrap
822 MAKE="make -j 2" -j 2} or just @samp{make -j 2 bootstrap}
823 for GNU Make 3.79 and above instead of just @samp{make bootstrap}
824 when building GCC. You can use a bigger number instead of two if
825 you like. In most cases, it won't help to use a number bigger than
826 the number of processors in your machine.
833 @uref{./index.html,,Return to the GCC Installation page}
837 @c ***Testing*****************************************************************
839 @comment node-name, next, previous, up
840 @node Testing, Final install, Building, Installing GCC
844 <h1 align="center">Installing GCC: Testing</h1>
847 @chapter Installing GCC: Testing
850 @cindex Installing GCC: Testing
853 @strong{Please note that this is only applicable
854 to current development versions of GCC and GCC 3.0 or later.
855 GCC 2.95.x does not come with a testsuite.}
857 Before you install GCC, you might wish to run the testsuite. This
858 step is optional and may require you to download additional software.
860 First, you must have @uref{download.html,,downloaded the testsuites}.
861 The full distribution contains testsuites; only if you downloaded the
862 ``core'' compiler plus any front ends, you do not have the testsuites.
864 Second, you must have a @uref{http://www.gnu.org/software/dejagnu/,,current version of DejaGnu} installed;
865 dejagnu 1.3 is not sufficient.
867 Now you may need specific preparations:
871 In order to run the libio tests in GCC 2.95 and earlier versions of GCC
872 on targets which do not fully
873 support Unix/POSIX commands (e.g. Cygwin), the references to the @file{dbz}
874 directory have to be deleted from @file{libio/configure.in}.
877 The following environment variables may need to be set appropriately, as in
878 the following example (which assumes that DejaGnu has been installed
879 under @file{/usr/local}):
882 TCL_LIBRARY = /usr/local/share/tcl8.0
883 DEJAGNULIBS = /usr/local/share/dejagnu
886 On systems such as Cygwin, these paths are required to be actual
887 paths, not mounts or links; presumably this is due to some lack of
888 portability in the DejaGnu code.
890 If the directories where @command{runtest} and @command{expect} were
891 installed are in the @env{PATH}, it should not be necessary to set these
892 environment variables.
896 Finally, you can run the testsuite (which may take a long time):
898 cd @var{objdir}; make -k check
901 The testing process will try to test as many components in the GCC
902 distribution as possible, including the C, C++, Objective C and Fortran
903 compilers as well as the C++ and Java runtime libraries.
905 @section How can I run the test suite on selected tests?
907 As a first possibility to cut down the number of tests that are run it is
908 possible to use @samp{make check-gcc} or @samp{make check-g++}
909 in the gcc subdirectory of the object directory. To further cut down the
910 tests the following is possible:
913 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
916 This will run all gcc execute tests in the testsuite.
919 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
922 This will run the g++ ``old-deja'' tests in the testsuite where the filename
923 matches @samp{9805*}.
925 The @file{*.exp} files are located in the testsuite directories of the GCC
926 source, the most important ones being @file{compile.exp},
927 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
928 To get a list of the possible @file{*.exp} files, pipe the
929 output of @samp{make check} into a file and look at the
930 @samp{Running @dots{} .exp} lines.
932 @section How to interpret test results
934 After the testsuite has run you'll find various @file{*.sum} and @file{*.log}
935 files in the testsuite subdirectories. The @file{*.log} files contain a
936 detailed log of the compiler invocations and the corresponding
937 results, the @file{*.sum} files summarize the results. These summaries list
938 all the tests that have been run with a corresponding status code:
942 PASS: the test passed as expected
944 XPASS: the test unexpectedly passed
946 FAIL: the test unexpectedly failed
948 XFAIL: the test failed as expected
950 UNSUPPORTED: the test is not supported on this platform
952 ERROR: the testsuite detected an error
954 WARNING: the testsuite detected a possible problem
957 It is normal for some tests to report unexpected failures. At the
958 current time our testing harness does not allow fine grained control
959 over whether or not a test is expected to fail. We expect to fix this
960 problem in future releases.
963 @section Submitting test results
965 If you want to report the results to the GCC project, use the
966 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
969 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
970 -m gcc-testresults@@gcc.gnu.org |sh
973 This script uses the @command{Mail} program to send the results, so
974 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
975 prepended to the testsuite summary and should contain any special
976 remarks you have on your results or your build environment. Please
977 do not edit the testsuite result block or the subject line, as these
978 messages are automatically parsed and presented at the
979 @uref{http://gcc.gnu.org/testresults/,,GCC testresults} web
980 page. Here you can also gather information on how specific tests
981 behave on different platforms and compare them with your results. A
982 few failing testcases are possible even on released versions and you
983 should look here first if you think your results are unreasonable.
987 @c ***Final install***********************************************************
989 @comment node-name, next, previous, up
990 @node Final install, , Testing, Installing GCC
992 @ifset finalinstallhtml
994 <h1 align="center">Installing GCC: Final installation</h1>
997 @chapter Installing GCC: Final installation
1000 Now that GCC has been built and tested, you can install it with
1001 @samp{cd @var{objdir}; make install}.
1003 That step completes the installation of GCC; user level binaries can
1004 be found in @file{@var{prefix}/bin} where @var{prefix} is the value you
1005 specified with the @option{--prefix} to configure (or @file{/usr/local}
1006 by default). (If you specified @option{--bindir}, that directory will
1007 be used instead; otherwise, if you specified @option{--exec-prefix},
1008 @file{@var{exec-prefix}/bin} will be used.) Headers for the C++ and
1009 Java libraries are installed in @file{@var{prefix}/include}; libraries
1010 in @file{@var{libdir}} (normally @file{@var{prefix}/lib}); internal
1011 parts of the compiler in @file{@var{libdir}/gcc-lib}; documentation in
1012 info format in @file{@var{infodir}} (normally @file{@var{prefix}/info}).
1014 If you don't mind, please quickly review the
1015 @uref{http://gcc.gnu.org/gcc-3.0/buildstat.html,,build status page}.
1016 If your system is not listed, send a note to
1017 @uref{mailto:gcc@@gcc.gnu.org,,gcc@@gcc.gnu.org} indicating
1018 that you successfully built and installed GCC.
1020 Include the output from running @file{@var{srcdir}/config.guess}. (Do
1021 not send us the config.guess file itself, just the one-line output from
1024 If you find a bug, please report it following our
1025 @uref{../bugs.html,,bug reporting guidelines}.
1027 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
1028 dvi}. You will need to have @command{texi2dvi} (version at least 4.0)
1029 and @TeX{} installed. This creates a number of @file{.dvi} files in
1030 subdirectories of @file{@var{objdir}}; these may be converted for
1031 printing with programs such as @command{dvips}. You can also
1032 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
1033 Free Software Foundation}, though such manuals may not be for the most
1034 recent version of GCC.
1041 @uref{./index.html,,Return to the GCC Installation page}
1045 @c ***Binaries****************************************************************
1047 @comment node-name, next, previous, up
1048 @node Binaries, Specific, Installing GCC, Top
1052 <h1 align="center">Installing GCC: Binaries</h1>
1055 @chapter Installing GCC: Binaries
1058 @cindex Installing GCC: Binaries
1060 We are often asked about pre-compiled versions of GCC. While we cannot
1061 provide these for all platforms, below you'll find links to binaries for
1062 various platforms where creating them by yourself is not easy due to various
1065 Please note that we did not create these binaries, nor do we
1066 support them. If you have any problems installing them, please
1067 contact their makers.
1074 @uref{http://www-frec.bull.com/docs/download.htm,,Bull's Freeware and
1075 Shareware Archive for AIX};
1078 @uref{http://aixpdlib.seas.ucla.edu,,UCLA Software Library for AIX};
1082 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP};
1085 @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center};
1088 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
1089 OpenServer/Unixware};
1092 Solaris (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware};
1095 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware};
1098 Windows 95, 98, and NT:
1101 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
1103 @uref{http://www.xraylith.wisc.edu/~khan/software/gnu-win32/,,GNU Win32}
1104 related projects by Mumit Khan.
1108 @uref{ftp://ftp.thewrittenword.com/packages/free/by-name/gcc-2.95.2/,,The
1109 Written Word} offers binaries for Solaris 2.5.1, 2.6, 2.7/SPARC, 2.7/Intel,
1110 IRIX 6.2, 6.5, Digital UNIX 4.0D, HP-UX 10.20, and HP-UX 11.00.
1113 Hitachi H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
1114 Development Tools for the Hitachi H8/300[HS] Series}
1118 In addition to those specific offerings, you can get a binary
1119 distribution CD-ROM from the
1120 @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}.
1121 It contains binaries for a number of platforms, and
1122 includes not only GCC, but other stuff as well. The current CD does
1123 not contain the latest version of GCC, but it should allow
1124 bootstrapping the compiler. An updated version of that disk is in the
1132 @uref{./index.html,,Return to the GCC Installation page}
1136 @c ***Specific****************************************************************
1138 @comment node-name, next, previous, up
1139 @node Specific, Concept Index, Binaries, Top
1143 <h1 align="center">Host/target specific installation notes for GCC</h1>
1146 @chapter Host/target specific installation notes for GCC
1149 @cindex Specific installation notes
1150 @cindex Target specific installation
1151 @cindex Host specific installation
1152 @cindex Target specific installation notes
1154 Please read this document carefully @emph{before} installing the
1155 GNU Compiler Collection on your machine.
1160 @uref{#1750a-*-*,,1750a-*-*}
1164 @uref{#a29k-*-bsd,,a29k-*-bsd}
1166 @uref{#alpha*-*-*,,alpha*-*-*}
1168 @uref{#alpha-*-osf1,,alpha-*-osf1}
1170 @uref{#alpha*-dec-osf*,,alpha*-dec-osf*}
1172 @uref{#arc-*-elf,,arc-*-elf}
1174 @uref{#arm-*-aout,,arm-*-aout}
1176 @uref{#arm-*-elf,,arm-*-elf}
1178 @uref{#arm*-*-linux-gnu,,arm*-*-linux-gnu}
1180 @uref{#arm-*-riscix,,arm-*-riscix}
1186 @uref{#decstation-*,,decstation-*}
1190 @uref{#dsp16xx,,dsp16xx}
1192 @uref{#elxsi-elxsi-bsd,,elxsi-elxsi-bsd}
1194 @uref{#h8300-hms,,h8300-hms}
1196 @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*}
1198 @uref{#hppa*-hp-hpux9,,hppa*-hp-hpux9}
1200 @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10}
1202 @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11}
1204 @uref{#*-*-freebsd*,,*-*-freebsd*}
1206 @uref{#i370-*-*,,i370-*-*}
1208 @uref{#*-*-linux-gnu,,*-*-linux-gnu}
1210 @uref{#ix86-*-linux*oldld,,i?86-*-linux*oldld}
1212 @uref{#ix86-*-linux*aout,,i?86-*-linux*aout}
1214 @uref{#ix86-*-linux*,,i?86-*-linux*}
1216 @uref{#ix86-*-sco,,i?86-*-sco}
1218 @uref{#ix86-*-sco3.2v4,,i?86-*-sco3.2v4}
1220 @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*}
1222 @uref{#ix86-*-udk,,i?86-*-udk}
1224 @uref{#ix86-*-isc,,i?86-*-isc}
1226 @uref{#ix86-*-esix,,i?86-*-esix}
1228 @uref{#ix86-ibm-aix,,i?86-ibm-aix}
1230 @uref{#ix86-sequent-bsd,,i?86-sequent-bsd}
1232 @uref{#ix86-sequent-ptx1*,,i?86-sequent-ptx1*, i?86-sequent-ptx2*}
1234 @uref{#ix86-*-sysv3*,,i?86-*-sysv3*}
1236 @uref{#i860-intel-osf*,,i860-intel-osf*}
1238 @uref{#*-lynx-lynxos,,*-lynx-lynxos}
1240 @uref{#*-ibm-aix*,,*-ibm-aix*}
1242 @uref{#m32r-*-elf,,m32r-*-elf}
1244 @uref{#m68000-hp-bsd,,m68000-hp-bsd}
1246 @uref{#m6811-elf,,m6811-elf}
1248 @uref{#m6812-elf,,m6812-elf}
1250 @uref{#m68k-altos,,m68k-altos}
1252 @uref{#m68k-apple-aux,,m68k-apple-aux}
1254 @uref{#m68k-att-sysv,,m68k-att-sysv}
1256 @uref{#m68k-bull-sysv,,m68k-bull-sysv}
1258 @uref{#m68k-crds-unox,,m68k-crds-unox}
1260 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
1262 @uref{#m68k-*-nextstep*,,m68k-*-nextstep*}
1264 @uref{#m68k-ncr-*,,m68k-ncr-*}
1266 @uref{#m68k-sun,,m68k-sun}
1268 @uref{#m68k-sun-sunos4.1.1,,m68k-sun-sunos4.1.1}
1270 @uref{#m88k-*-svr3,,m88k-*-svr3}
1272 @uref{#m88k-*-dgux,,m88k-*-dgux}
1274 @uref{#m88k-tektronix-sysv3,,m88k-tektronix-sysv3}
1276 @uref{#mips-*-*,,mips-*-*}
1278 @uref{#mips-mips-bsd,,mips-mips-bsd}
1280 @uref{#mips-mips-riscos*,,mips-mips-riscos*}
1282 @uref{#mips*-sgi-irix[45],,mips*-sgi-irix[45]}
1284 @uref{#mips*-sgi-irix6,,mips*-sgi-irix6}
1286 @uref{#mips-sony-sysv,,mips-sony-sysv}
1288 @uref{#ns32k-encore,,ns32k-encore}
1290 @uref{#ns32k-*-genix,,ns32k-*-genix}
1292 @uref{#ns32k-sequent,,ns32k-sequent}
1294 @uref{#ns32k-utek,,ns32k-utek}
1296 @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4}
1298 @uref{#powerpc-*-linux-gnu*,,powerpc-*-linux-gnu*}
1300 @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix}
1302 @uref{#powerpc-*-eabisim,,powerpc-*-eabisim}
1304 @uref{#powerpc-*-eabi,,powerpc-*-eabi}
1306 @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
1308 @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim}
1310 @uref{#powerpcle-*-eabi,,powerpcle-*-eabi}
1312 @uref{#powerpcle-*-winnt,,powerpcle-*-winnt, powerpcle-*-pe}
1314 @uref{#romp-*-aos,,romp-*-aos, romp-*-mach}
1316 @uref{#*-*-solaris*,,*-*-solaris*}
1318 @uref{#sparc-sun-*,,sparc-sun-*}
1320 @uref{#sparc-sun-solaris*,,sparc-sun-solaris*}
1322 @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7}
1324 @uref{#*-sun-solaris2.8,,*-sun-solaris2.8}
1326 @uref{#sunv5,,Sun V5.0 Compiler Bugs}
1328 @uref{#sparc-sun-sunos*,,sparc-sun-sunos*}
1330 @uref{#sparc-unknown-linux-gnulibc1,,sparc-unknown-linux-gnulibc1}
1332 @uref{#sparc64-*-*,,sparc64-*-*}
1334 @uref{#*-*-sysv*,,*-*-sysv*}
1336 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
1338 @uref{#we32k-*-*,,we32k-*-*}
1340 @uref{#windows,,Microsoft Windows}
1344 @uref{#older,,Older systems}
1349 @uref{#elf_targets,,all ELF targets} (SVR4, Solaris, etc.)
1355 <!-- -------- host/target specific issues start here ---------------- -->
1358 @heading @anchor{1750a-*-*}1750a-*-*
1359 MIL-STD-1750A processors.
1361 The MIL-STD-1750A cross configuration produces output for
1362 @code{as1750}, an assembler/linker available under the GNU Public
1363 License for the 1750A. @code{as1750} can be obtained at
1364 @uref{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}.
1365 A similarly licensed simulator for
1366 the 1750A is available from same address.
1368 You should ignore a fatal error during the building of libgcc (libgcc is
1369 not yet implemented for the 1750A.)
1371 The @code{as1750} assembler requires the file @file{ms1750.inc}, which is
1372 found in the directory @file{config/1750a}.
1374 GNU CC produced the same sections as the Fairchild F9450 C Compiler,
1379 The program code section.
1382 The read/write (RAM) data section.
1385 The read-only (ROM) constants section.
1388 Initialization section (code to copy KREL to SREL).
1391 The smallest addressable unit is 16 bits (BITS_PER_UNIT is 16). This
1392 means that type @code{char} is represented with a 16-bit word per character.
1393 The 1750A's ``Load/Store Upper/Lower Byte'' instructions are not used by
1400 @heading @anchor{a29k}a29k
1401 AMD Am29k-family processors. These are normally used in embedded
1402 applications. There are no standard Unix configurations.
1404 corresponds to AMD's standard calling sequence and binary interface
1405 and is compatible with other 29k tools.
1407 You may need to make a variant of the file @file{a29k.h} for your
1408 particular configuration.
1414 @heading @anchor{a29k-*-bsd}a29k-*-bsd
1415 AMD Am29050 used in a system running a variant of BSD Unix.
1421 @heading @anchor{alpha*-*-*}alpha*-*-*
1423 This section contains general configuration information for all
1424 alpha-based platforms using ELF (in particular, ignore this section for
1425 OSF and tru64). In addition to reading this section, please read all
1426 other sections that match your target.
1428 We require binutils 2.11.1 (as of yet unreleased), binutils with
1429 @samp{binutils-2_11-branch} tag after May 31, 2001 (as taken below), or newer.
1430 Previous binutils releases had a number of problems with DWARF2
1431 debugging information, not the least of which is incorrect linking of
1434 Until binutils 2.11.1 is released, these sample commands may be useful:
1437 mkdir binutils-2.11.X; cd binutils-2.11.X
1438 cvs -d :pserver:anoncvs@@anoncvs.cygnus.com:/cvs/src \
1439 co -rbinutils-2_11-branch -P binutils
1441 ../src/configure --prefix=@emph{an-absolute-path}
1442 make all check install
1445 When configuring gcc, provide explicit @option{--with-gnu-as}
1446 @option{--with-as=@emph{an-absolute-path/bin/as}} and
1447 @option{--with-gnu-ld} @option{--with-ld=@emph{an-absolute-path/bin/ld}}
1448 options to point into the prefix used above.
1454 @heading @anchor{alpha-*-osf1}alpha-*-osf1
1455 Systems using processors that implement the DEC Alpha architecture and
1456 are running the DEC Unix (OSF/1) operating system, for example the DEC
1457 Alpha AXP systems.CC.)
1459 GNU CC writes a @samp{.verstamp} directive to the assembler output file
1460 unless it is built as a cross-compiler. It gets the version to use from
1461 the system header file @file{/usr/include/stamp.h}. If you install a
1462 new version of DEC Unix, you should rebuild GCC to pick up the new version
1465 Note that since the Alpha is a 64-bit architecture, cross-compilers from
1466 32-bit machines will not generate code as efficient as that generated
1467 when the compiler is running on a 64-bit machine because many
1468 optimizations that depend on being able to represent a word on the
1469 target in an integral value on the host cannot be performed. Building
1470 cross-compilers on the Alpha for 32-bit machines has only been tested in
1471 a few cases and may not work properly.
1473 @code{make compare} may fail on old versions of DEC Unix unless you add
1474 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
1475 assembler input file is stored in the object file, and that makes
1476 comparison fail if it differs between the @code{stage1} and
1477 @code{stage2} compilations. The option @option{-save-temps} forces a
1478 fixed name to be used for the assembler input file, instead of a
1479 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
1480 unless the comparisons fail without that option. If you add
1481 @option{-save-temps}, you will have to manually delete the @samp{.i} and
1482 @samp{.s} files after each series of compilations.
1484 GNU CC now supports both the native (ECOFF) debugging format used by DBX
1485 and GDB and an encapsulated STABS format for use only with GDB. See the
1486 discussion of the @option{--with-stabs} option of @file{configure} above
1487 for more information on these formats and how to select them.
1489 There is a bug in DEC's assembler that produces incorrect line numbers
1490 for ECOFF format when the @samp{.align} directive is used. To work
1491 around this problem, GNU CC will not emit such alignment directives
1492 while writing ECOFF format debugging information even if optimization is
1493 being performed. Unfortunately, this has the very undesirable
1494 side-effect that code addresses when @option{-O} is specified are
1495 different depending on whether or not @option{-g} is also specified.
1497 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
1498 DBX. DEC is now aware of this problem with the assembler and hopes to
1499 provide a fix shortly.
1505 @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf*
1507 If you install a shared libstdc++ and, when you link a non-trivial C++
1508 program (for example, @file{gcc/testsuite/g++.other/delete3.C}),
1509 the linker reports a couple of errors about multiply-defined symbols
1510 (for example, @code{nothrow}, @code{__throw} and
1511 @code{terminate(void)}), you've probably got a linker bug, for
1512 which there's no known fix. The officially recommended work-around is
1513 to remove the shared libstdc++.
1515 An alternative solution is to arrange that all symbols from
1516 @code{libgcc} get copied to the shared @code{libstdc++};
1517 see detailed solution below. (Surprising as it may seem, this does
1518 indeed fix the problem!) @emph{Beware} that this may bring you
1519 binary-compatibility problems in the future, if you don't use the same
1520 work-around next time you build @code{libstdc++}: if programs
1521 start to depend on @code{libstdc++} to provide symbols that used
1522 to be only in @code{libgcc}, you must arrange that
1523 @code{libstdc++} keeps providing them, otherwise the programs
1524 will have to be relinked.
1526 The magic spell is to add @option{-Wl,-all,-lgcc,-none} to the
1527 definition of macro @code{SHDEPS} in
1528 @file{libstdc++/config/dec-osf.ml} @emph{before}
1529 @file{alpha*-dec-osf*/libstdc++/Makefile} is created (a
1530 @uref{dec-osf-shlibstdc++.patch,,patch}
1531 that does just that is available). If the Makefile already exists, run
1532 @file{./config.status} within directory
1533 @file{alpha*-dec-osf*/libstdc++} (and
1534 @file{alpha*-dec-osf*/ieee/libstdc++}, if it also exists).
1535 Remove any existing @file{libstdc++.so*} from such directories,
1536 and run @samp{make all-target-libstdc++} in the top-level
1537 directory, then @samp{make install-target-libstdc++}.
1539 If you have already removed the build tree, you may just remove
1540 @file{libstdc++.so.2.10.0} from the install tree and re-create
1542 @samp{gcc -shared -o libstdc++.so.2.10.0 -Wl,-all,-lstdc++,-lgcc,-none -lm}.
1544 sub-directory exists, repeat this command in it, with the additional
1545 flag @option{-mieee}.
1551 @heading @anchor{arc-*-elf}arc-*-elf
1552 Argonaut ARC processor.
1553 This configuration is intended for embedded systems.
1559 @heading @anchor{arm-*-aout}arm-*-aout
1560 Advanced RISC Machines ARM-family processors. These are often used in
1561 embedded applications. There are no standard Unix configurations.
1562 This configuration corresponds to the basic instruction sequences and will
1563 produce @file{a.out} format object modules.
1565 You may need to make a variant of the file @file{arm.h} for your particular
1572 @heading @anchor{arm-*-elf}arm-*-elf
1573 This configuration is intended for embedded systems.
1579 @heading @anchor{arm*-*-linux-gnu}arm*-*-linux-gnu
1581 We require GNU binutils 2.10 or newer.
1587 @heading @anchor{arm-*-riscix}arm-*-riscix
1588 The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix.
1589 If you are running a version of RISC iX prior to 1.2 then you must
1590 specify the version number during configuration. Note that the
1591 assembler shipped with RISC iX does not support stabs debugging
1592 information; a new version of the assembler, with stabs support
1593 included, is now available from Acorn and via ftp
1594 @uref{ftp://ftp.acorn.com/pub/riscix/as+xterm.tar.Z}. To enable stabs
1595 debugging, pass @option{--with-gnu-as} to configure.
1597 You will need to install GNU @command{sed} before you can run configure.
1603 @heading @anchor{avr}avr
1605 ATMEL AVR-family micro controllers. These are used in embedded
1606 applications. There are no standard Unix configurations. @xref{AVR
1607 Options,, AVR Options, gcc, Using and Porting the GNU Compiler
1608 Collection (GCC)}, for the list of supported MCU types.
1610 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC.
1612 Further installation notes and other useful information about AVR tools
1613 can also be obtained from:
1617 @uref{http://home.overta.ru/users/denisc,,http://home.overta.ru/users/denisc}
1619 @uref{http://www.itnet.pl/amelektr/avr,,http://www.itnet.pl/amelektr/avr}
1622 We @emph{strongly} recommend using binutils 2.11 or newer.
1624 The following error:
1626 Error: register required
1629 indicates that you should upgrade to a newer version of the binutils.
1635 @heading @anchor{c4x}c4x
1637 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
1638 Processors. These are used in embedded applications. There are no
1639 standard Unix configurations. @xref{C4x Options,, C4x Options, gcc,
1640 Using and Porting the GNU Compiler Collection (GCC)}, for the list of
1641 supported MCU types.
1643 GCC can be configured as a cross compiler for both the C3x and C4x
1644 architectures on the same system. Use @samp{configure --target=c4x
1645 --enable-languages="c,c++"} to configure.
1648 Further installation notes and other useful information about C4x tools
1649 can also be obtained from:
1653 @uref{http://www.elec.canterbury.ac.nz/c4x,,http://www.elec.canterbury.ac.nz/c4x}
1660 @heading @anchor{decstation-*}decstation-*
1661 MIPS-based DECstations can support three different personalities:
1662 Ultrix, DEC OSF/1, and OSF/rose. (Alpha-based DECstation products have
1663 a configuration name beginning with @samp{alpha-dec}.) To configure GCC
1664 for these platforms use the following configurations:
1667 @item decstation-ultrix
1668 Ultrix configuration.
1670 @item decstation-osf1
1671 Dec's version of OSF/1.
1673 @item decstation-osfrose
1674 Open Software Foundation reference port of OSF/1 which uses the
1675 OSF/rose object file format instead of ECOFF. Normally, you
1676 would not select this configuration.
1679 The MIPS C compiler needs to be told to increase its table size
1680 for switch statements with the @option{-Wf,-XNg1500} option in
1681 order to compile @file{cp/parse.c}. If you use the @option{-O2}
1682 optimization option, you also need to use @option{-Olimit 3000}.
1683 Both of these options are automatically generated in the
1684 @file{Makefile} that the shell script @file{configure} builds.
1685 If you override the @code{CC} make variable and use the MIPS
1686 compilers, you may need to add @option{-Wf,-XNg1500 -Olimit 3000}.
1692 @heading @anchor{dos}DOS
1694 Please have a look at our @uref{binaries.html,,binaries page}.
1696 You cannot install GNU C by itself on MSDOS; it will not compile under
1697 any MSDOS compiler except itself. You need to get the complete
1698 compilation package DJGPP, which includes binaries as well as sources,
1699 and includes all the necessary compilation tools and libraries.
1705 @heading @anchor{dsp16xx}dsp16xx
1706 A port to the AT&T DSP1610 family of processors.
1712 @heading @anchor{elxsi-elxsi-bsd}elxsi-elxsi-bsd
1713 The Elxsi's C compiler has known limitations that prevent it from
1714 compiling GNU C. Please contact @email{mrs@@cygnus.com} for more details.
1720 @heading @anchor{h8300-hms}h8300-hms
1721 Hitachi H8/300 series of processors.
1723 Please have a look at our @uref{binaries.html,,binaries page}.
1725 The calling convention and structure layout has changed in release 2.6.
1726 All code must be recompiled. The calling convention now passes the
1727 first three arguments in function calls in registers. Structures are no
1728 longer a multiple of 2 bytes.
1734 @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux*
1736 We @emph{highly} recommend using gas/binutils-2.8 or newer on all hppa
1737 platforms; you may encounter a variety of problems when using the HP
1740 Specifically, @option{-g} does not work on HP-UX (since that system
1741 uses a peculiar debugging format which GCC does not know about), unless you
1742 use GAS and GDB and configure GCC with the
1743 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
1744 @option{--with-as=@dots{}} options.
1746 If you wish to use pa-risc 2.0 architecture support, you must use either
1747 the HP assembler, gas/binutils-2.11 or a recent
1748 @uref{ftp://sources.redhat.com/pub/binutils/snapshots,,snapshot of gas}.
1750 More specific information to hppa*-hp-hpux* targets follows.
1756 @heading @anchor{hppa*-hp-hpux9}hppa*-hp-hpux9
1758 The HP assembler has major problems on this platform. We've tried to work
1759 around the worst of the problems. However, those workarounds may be causing
1760 linker crashes in some circumstances; the workarounds also probably prevent
1761 shared libraries from working. Use the GNU assembler to avoid these problems.
1764 The configuration scripts for GCC will also trigger a bug in the hpux9
1765 shell. To avoid this problem set @env{CONFIG_SHELL} to @file{/bin/ksh}
1766 and @env{SHELL} to @file{/bin/ksh} in your environment.
1773 @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10
1775 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
1776 @code{PHCO_19798} from HP. HP has two sites which provide patches free of
1782 <a href="http://us-support.external.hp.com">US, Canada, Asia-Pacific, and
1786 @uref{http://us-support.external.hp.com,,}US, Canada, Asia-Pacific, and
1790 @uref{http://europe-support.external.hp.com,,Europe}
1793 The HP assembler on these systems is much better than the hpux9 assembler,
1794 but still has some problems. Most notably the assembler inserts timestamps
1795 into each object file it creates, causing the 3-stage comparison test to fail
1796 during a @samp{make bootstrap}. You should be able to continue by
1797 saying @samp{make all} after getting the failure from @samp{make
1805 @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11
1807 GCC 3.0 supports HP-UX 11. You must use GNU binutils 2.11 or above on
1814 @heading @anchor{*-*-freebsd*}*-*-freebsd*
1816 The version of binutils installed in @file{/usr/bin} is known to work unless
1817 otherwise specified in any per-architecture notes. However, binutils
1818 2.11 is known to improve overall testsuite results.
1820 For FreeBSD 1, FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
1821 configuration support and files as shipped with gcc 2.95 are still in
1822 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
1823 it is unknown which version of binutils was used (it is assumed that it
1824 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
1826 For FreeBSD using the ELF file format: DWARF2 debugging is now the
1827 default for all CPU architectures. It had been the default on
1828 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
1829 of @option{-g}, if you really want the old debugging format. There are
1830 no known issues with mixing object files and libraries with different
1831 debugging formats. Otherwise, this release of gcc should now match more
1832 of the configuration used in the stock FreeBSD configuration of gcc. In
1833 particular, @option{--enable-threads} is now configured by default.
1834 However, as a general user, do not attempt to replace the system
1835 compiler with this release. Known to bootstrap and check with good
1836 results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3 and 5-CURRENT.
1838 At this time, @option{--enable-threads} is not compatible with
1839 @option{--enable-libgcj} on FreeBSD.
1845 @heading @anchor{i370-*-*}i370-*-*
1846 This port is very preliminary and has many known bugs. We hope to
1847 have a higher-quality port for this machine soon.
1853 @heading @anchor{*-*-linux-gnu}*-*-linux-gnu
1855 If you use glibc 2.2 (or 2.1.9x), GCC 2.95.2 won't install
1856 out-of-the-box. You'll get compile errors while building libstdc++.
1857 The patch @uref{glibc-2.2.patch,,glibc-2.2.patch}, that is to be
1858 applied in the GCC source tree, fixes the compatibility problems.
1864 @heading @anchor{ix86-*-linux*oldld}i?86-*-linux*oldld
1865 Use this configuration to generate @file{a.out} binaries on Linux-based
1866 GNU systems if you do not have gas/binutils version 2.5.2 or later
1867 installed. This is an obsolete configuration.
1873 @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout
1874 Use this configuration to generate @file{a.out} binaries on Linux-based
1875 GNU systems. This configuration is being superseded. You must use
1876 gas/binutils version 2.5.2 or later.
1882 @heading @anchor{ix86-*-linux*}i?86-*-linux*
1884 You will need binutils-2.9.1.0.15 or newer for exception handling to work.
1886 If you receive Signal 11 errors when building on GNU/Linux, then it is
1887 possible you have a hardware problem. Further information on this can be
1888 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
1894 @heading @anchor{ix86-*-sco}i?86-*-sco
1895 Compilation with RCC is recommended. Also, it may be a good idea to
1896 link with GNU malloc instead of the malloc that comes with the system.
1902 @heading @anchor{ix86-*-sco3.2v4}i?86-*-sco3.2v4
1903 Use this configuration for SCO release 3.2 version 4.
1909 @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5*
1910 Use this for the SCO OpenServer Release 5 family of operating systems.
1912 Unlike earlier versions of GCC, the ability to generate COFF with this
1913 target is no longer provided.
1915 Earlier versions of GCC emitted Dwarf-1 when generating ELF to allow
1916 the system debugger to be used. That support was too burdensome to
1917 maintain. GCC now emits only dwarf-2 for this target. This means you
1918 may use either the UDK debugger or GDB to debug programs built by this
1921 If you are building languages other than C, you must follow the instructions
1922 about invoking @samp{make bootstrap} because the native OpenServer
1923 compiler will build a @command{cc1plus} that will not correctly parse many
1924 valid C++ programs including those in @file{libgcc.a}.
1925 @strong{You must do a @samp{make bootstrap} if you are building with the
1928 Use of the @option{-march-pentiumpro} flag can result in
1929 unrecognized opcodes when using the native assembler on OS versions before
1930 5.0.6. (Support for P6 opcodes was added to the native ELF assembler in
1931 that version.) While it's rather rare to see these emitted by GCC yet,
1932 errors of the basic form:
1935 /usr/tmp/ccaNlqBc.s:22:unknown instruction: fcomip
1936 /usr/tmp/ccaNlqBc.s:50:unknown instruction: fucomip
1939 are symptoms of this problem. You may work around this by not
1940 building affected files with that flag, by using the GNU assembler, or
1941 by using the assembler provided with the current version of the OS.
1942 Users of GNU assembler should see the note below for hazards on doing
1945 The native SCO assembler that is provided with the OS at no
1946 charge is normally required. If, however, you must be able to use
1947 the GNU assembler (perhaps you're compiling code with asms that
1948 require GAS syntax) you may configure this package using the flags
1949 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You must
1950 use a recent version of GNU binutils; versions past 2.9.1 seem to work
1953 In general, the @option{--with-gnu-as} option isn't as well tested
1954 as the native assembler.
1956 Look in @file{gcc/config/i386/sco5.h} (search for ``messy'') for
1957 additional OpenServer-specific flags.
1959 Systems based on OpenServer before 5.0.4 (@samp{uname -X}
1960 will tell you what you're running) require TLS597 from ftp.sco.com/TLS
1961 for C++ constructors and destructors to work right.
1963 The system linker in (at least) 5.0.4 and 5.0.5 will sometimes
1964 do the wrong thing for a construct that GCC will emit for PIC
1965 code. This can be seen as execution testsuite failures when using
1966 @option{-fPIC} on @file{921215-1.c}, @file{931002-1.c}, @file{nestfunc-1.c}, and @file{gcov-1.c}.
1967 For 5.0.5, an updated linker that will cure this problem is
1968 available. You must install both
1969 @uref{ftp://ftp.sco.com/Supplements/rs505a/,,ftp://ftp.sco.com/Supplements/rs505a/}
1970 and @uref{ftp://ftp.sco.com/SLS/,,OSS499A}.
1972 The dynamic linker in OpenServer 5.0.5 (earlier versions may show
1973 the same problem) aborts on certain g77-compiled programs. It's particularly
1974 likely to be triggered by building Fortran code with the @option{-fPIC} flag.
1975 Although it's conceivable that the error could be triggered by other
1976 code, only G77-compiled code has been observed to cause this abort.
1977 If you are getting core dumps immediately upon execution of your
1978 g77 program---and especially if it's compiled with @option{-fPIC}---try applying
1979 @uref{sco_osr5_g77.patch,,@file{sco_osr5_g77.patch}} to your libf2c and
1981 Affected faults, when analyzed in a debugger, will show a stack
1982 backtrace with a fault occurring in @code{rtld()} and the program
1983 running as @file{/usr/lib/ld.so.1}. This problem has been reported to SCO
1984 engineering and will hopefully be addressed in later releases.
1991 @heading @anchor{ix86-*-udk}i?86-*-udk
1993 This target emulates the SCO Universal Development Kit and requires that
1994 package be installed. (If it is installed, you will have a
1995 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
1996 @code{i?86-*-unixware7*} target
1997 but is meant to be used when hosting on a system where UDK isn't the
1998 default compiler such as OpenServer 5 or Unixware 2. This target will
1999 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
2000 with the same warnings and caveats as the SCO UDK.
2002 You can stage1 with either your native compiler or with UDK. If you
2003 don't do a full bootstrap when initially building with your native compiler
2004 you will have an utterly unusable pile of bits as your reward.
2006 This target is a little tricky to build because we have to distinguish
2007 it from the native tools (so it gets headers, startups, and libraries
2008 from the right place) while making the tools not think we're actually
2009 building a cross compiler. The easiest way to do this is with a configure
2012 @samp{CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure
2013 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-}
2015 @emph{You should substitute @samp{i686} in the above command with the appropriate
2016 processor for your host.}
2018 You should follow this with a @samp{make bootstrap} then
2019 @samp{make install}. You can then access the UDK-targeted GCC
2020 tools by adding @command{udk-} before the commonly known name. For
2021 example, to invoke the C compiler, you would use @command{udk-gcc}.
2022 They will coexist peacefully with any native-target GCC tools you may
2030 @heading @anchor{ix86-*-isc}i?86-*-isc
2031 It may be a good idea to link with GNU malloc instead of the malloc that
2032 comes with the system.
2034 In ISC version 4.1, @command{sed} core dumps when building
2035 @file{deduced.h}. Use the version of @command{sed} from version 4.0.
2041 @heading @anchor{ix86-*-esix}i?86-*-esix
2042 It may be good idea to link with GNU malloc instead of the malloc that
2043 comes with the system.
2049 @heading @anchor{ix86-ibm-aix}i?86-ibm-aix
2050 You need to use GAS version 2.1 or later, and LD from
2051 GNU binutils version 2.2 or later.
2057 @heading @anchor{ix86-sequent-bsd}i?86-sequent-bsd
2058 Go to the Berkeley universe before compiling.
2064 @heading @anchor{ix86-sequent-ptx1*}i?86-sequent-ptx1*, i?86-sequent-ptx2*
2065 You must install GNU @file{sed} before running @file{configure}.
2071 @heading @anchor{#ix86-*-sysv3*}i?86-*-sysv3*
2072 The @code{fixproto} shell script may trigger a bug in the system shell.
2073 If you encounter this problem, upgrade your operating system or
2074 use BASH (the GNU shell) to run @code{fixproto}.
2081 @heading @anchor{i860-intel-osf*}i860-intel-osf*
2082 On the Intel Paragon (an i860 machine), if you are using operating
2083 system version 1.0, you will get warnings or errors about redefinition
2084 of @code{va_arg} when you build GCC.
2086 If this happens, then you need to link most programs with the library
2087 @file{iclib.a}. You must also modify @file{stdio.h} as follows: before
2091 #if defined(__i860__) && !defined(_VA_LIST)
2092 #include <va_list.h>
2106 extern int vprintf(const char *, va_list );
2107 extern int vsprintf(char *, const char *, va_list );
2115 #endif /* __PGC__ */
2118 These problems don't exist in operating system version 1.1.
2124 @heading @anchor{*-lynx-lynxos}*-lynx-lynxos
2125 LynxOS 2.2 and earlier comes with GNU CC 1.x already installed as
2126 @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}.
2127 You can tell GNU CC to use the GNU assembler and linker, by specifying
2128 @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce
2129 COFF format object files and executables; otherwise GNU CC will use the
2130 installed tools, which produce @file{a.out} format executables.
2135 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
2137 @heading @anchor{*-ibm-aix*}*-ibm-aix*
2139 AIX Make frequently has problems with GCC makefiles. GNU Make 3.76 or
2140 newer is recommended to build on this platform.
2142 Errors involving @code{alloca} when building GCC generally are due
2143 to an incorrect definition of @var{CC} in the Makefile or mixing files
2144 compiled with the native C compiler and GCC. During the stage1 phase of
2145 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
2146 (not @command{xlc}). Once @command{configure} has been informed of
2147 @command{xlc}, one needs to use @samp{make distclean} to remove the
2148 configure cache files and ensure that @env{CC} environment variable
2149 does not provide a definition that will confuse @command{configure}.
2150 If this error occurs during stage2 or later, then the problem most likely
2151 is the version of Make (see above).
2153 Binutils 2.10 does not support AIX 4.3. Binutils available from the
2154 @uref{http://www-1.ibm.com/servers/aix/products/aixos/linux/,,AIX
2155 Toolbox for Linux: GNU and Open Source tools for AIX};
2156 website does work. Binutils 2.11 is expected to include AIX 4.3
2157 support. The GNU Assembler is necessary for libstdc++ to build. The
2158 AIX native ld still is recommended. The native AIX tools do
2159 interoperate with GCC.
2161 Linking executables and shared libraries may produce warnings of
2162 duplicate symbols. The assembly files generated by GCC for AIX always
2163 have included multiple symbol definitions for certain global variable
2164 and function declarations in the original program. The warnings should
2165 not prevent the linker from producing a correct library or runnable
2168 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
2169 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
2170 to parse archive libraries did not handle the new format correctly.
2171 These routines are used by GCC and result in error messages during
2172 linking such as ``not a COFF file''. The version of the routines shipped
2173 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
2174 option of the archive command may be used to create archives of 32-bit
2175 objects using the original ``small format''. A correct version of the
2176 routines is shipped with AIX 4.3.2.
2178 Some versions of the AIX binder (linker) can fail with a relocation
2179 overflow severe error when the @option{-bbigtoc} option is used to link
2180 GCC-produced object files into an executable that overflows the TOC. A fix
2181 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
2182 available from IBM Customer Support and from its
2183 @uref{http://service.boulder.ibm.com/,,service.boulder.ibm.com}
2184 website as PTF U455193.
2186 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
2187 with a segmentation fault when invoked by any version of GCC. A fix for
2188 APAR IX87327 is available from IBM Customer Support and from its
2189 @uref{http://service.boulder.ibm.com/,,service.boulder.ibm.com}
2190 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
2192 The initial assembler shipped with AIX 4.3.0 generates incorrect object
2193 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
2194 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
2195 @uref{http://service.boulder.ibm.com/,,service.boulder.ibm.com}
2196 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
2198 AIX provides National Language Support (NLS). Compilers and assemblers
2199 use NLS to support locale-specific representations of various data
2200 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
2201 separating decimal fractions). There have been problems reported where
2202 GCC does not produce the same floating-point formats that the assembler
2203 expects. If one encounters this problem, set the @env{LANG}
2204 environment variable to @samp{C} or @samp{En_US}.
2206 By default, GCC for AIX 4.1 and above produces code that can be used on
2207 both Power or PowerPC processors.
2209 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2210 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2216 @heading @anchor{m32r-*-elf}m32r-*-elf
2217 Mitsubishi M32R processor.
2218 This configuration is intended for embedded systems.
2224 @heading @anchor{m68000-hp-bsd}m68000-hp-bsd
2225 HP 9000 series 200 running BSD. Note that the C compiler that comes
2226 with this system cannot compile GNU CC; contact @email{law@@cygnus.com}
2227 to get binaries of GNU CC for bootstrapping.
2233 @heading @anchor{m6811-elf}m6811-elf
2234 Motorola 68HC11 family micro controllers. These are used in embedded
2235 applications. There are no standard Unix configurations.
2241 @heading @anchor{m6812-elf}m6812-elf
2242 Motorola 68HC12 family micro controllers. These are used in embedded
2243 applications. There are no standard Unix configurations.
2249 @heading @anchor{m68k-altos}m68k-altos
2250 Altos 3068. You must use the GNU assembler, linker and debugger.
2251 Also, you must fix a kernel bug. Details in the file @file{README.ALTOS}.
2257 @heading @anchor{m68k-apple-aux}m68k-apple-aux
2258 Apple Macintosh running A/UX.
2259 You may configure GCC to use either the system assembler and
2260 linker or the GNU assembler and linker. You should use the GNU configuration
2261 if you can, especially if you also want to use GNU C++. You enabled
2262 that configuration with + the @option{--with-gnu-as} and @option{--with-gnu-ld}
2263 options to @code{configure}.
2265 Note the C compiler that comes
2266 with this system cannot compile GNU CC. You can find binaries of GNU CC
2267 for bootstrapping on @code{jagubox.gsfc.nasa.gov}.
2268 You will also a patched version of @file{/bin/ld} there that
2269 raises some of the arbitrary limits found in the original.
2275 @heading @anchor{m68k-att-sysv}m68k-att-sysv
2276 AT&T 3b1, a.k.a. 7300 PC. This version of GNU CC cannot
2277 be compiled with the system C compiler, which is too buggy.
2278 You will need to get a previous version of GCC and use it to
2279 bootstrap. Binaries are available from the OSU-CIS archive, at
2280 @uref{ftp://archive.cis.ohio-state.edu/pub/att7300/}.
2286 @heading @anchor{m68k-bull-sysv}m68k-bull-sysv
2287 Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GNU CC works
2288 either with native assembler or GNU assembler. You can use
2289 GNU assembler with native coff generation by providing @option{--with-gnu-as} to
2290 the configure script or use GNU assembler with dbx-in-coff encapsulation
2291 by providing @samp{--with-gnu-as --stabs}. For any problem with native
2292 assembler or for availability of the DPX/2 port of GAS, contact
2293 @email{F.Pierresteguy@@frcl.bull.fr}.
2299 @heading @anchor{m68k-crds-unox}m68k-crds-unox
2300 Use @samp{configure unos} for building on Unos.
2302 The Unos assembler is named @code{casm} instead of @code{as}. For some
2303 strange reason linking @file{/bin/as} to @file{/bin/casm} changes the
2304 behavior, and does not work. So, when installing GNU CC, you should
2305 install the following script as @file{as} in the subdirectory where
2306 the passes of GCC are installed:
2313 The default Unos library is named @file{libunos.a} instead of
2314 @file{libc.a}. To allow GNU CC to function, either change all
2315 references to @option{-lc} in @file{gcc.c} to @option{-lunos} or link
2316 @file{/lib/libc.a} to @file{/lib/libunos.a}.
2318 @cindex @code{alloca}, for Unos
2319 When compiling GNU CC with the standard compiler, to overcome bugs in
2320 the support of @code{alloca}, do not use @option{-O} when making stage 2.
2321 Then use the stage 2 compiler with @option{-O} to make the stage 3
2322 compiler. This compiler will have the same characteristics as the usual
2323 stage 2 compiler on other systems. Use it to make a stage 4 compiler
2324 and compare that with stage 3 to verify proper compilation.
2326 (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in
2327 the comments there will make the above paragraph superfluous. Please
2328 inform us of whether this works.)
2330 Unos uses memory segmentation instead of demand paging, so you will need
2331 a lot of memory. 5 Mb is barely enough if no other tasks are running.
2332 If linking @file{cc1} fails, try putting the object files into a library
2333 and linking from that library.
2339 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
2340 HP 9000 series 300 or 400 running HP-UX. HP-UX version 8.0 has a bug in
2341 the assembler that prevents compilation of GNU CC. This
2342 bug manifests itself during the first stage of compilation, while
2343 building @file{libgcc2.a}:
2347 cc1: warning: `-g' option not supported on this version of GCC
2348 cc1: warning: `-g1' option not supported on this version of GCC
2349 ./xgcc: Internal compiler error: program as got fatal signal 11
2352 A patched version of the assembler is available as the file
2353 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
2354 have HP software support, the patch can also be obtained directly from
2355 HP, as described in the following note:
2358 This is the patched assembler, to patch SR#1653-010439, where the
2359 assembler aborts on floating point constants.
2361 The bug is not really in the assembler, but in the shared library
2362 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
2363 SR#4701-078451. Anyway, the attached assembler uses the archive
2364 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
2367 This patch is also known as PHCO_4484.
2369 In addition, if you wish to use gas @option{--with-gnu-as} you must use
2370 gas version 2.1 or later, and you must use the GNU linker version 2.1 or
2371 later. Earlier versions of gas relied upon a program which converted the
2372 gas output into the native HP-UX format, but that program has not been
2373 kept up to date. gdb does not understand that native HP-UX format, so
2374 you must use gas if you wish to use gdb.
2376 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
2377 @code{fixproto} shell script triggers a bug in the system shell. If you
2378 encounter this problem, upgrade your operating system or use BASH (the
2379 GNU shell) to run @code{fixproto}. This bug will cause the fixproto
2380 program to report an error of the form:
2383 ./fixproto: sh internal 1K buffer overflow
2386 To fix this, you can also change the first line of the fixproto script
2398 @heading @anchor{m68k-*-nextstep*}m68k-*-nextstep*
2400 Current GCC versions probably do not work on version 2 of the NeXT
2403 On NeXTStep 3.0, the Objective C compiler does not work, due,
2404 apparently, to a kernel bug that it happens to trigger. This problem
2405 does not happen on 3.1.
2407 You absolutely @strong{must} use GNU sed and GNU make on this platform.
2410 On NEXTSTEP 3.x where x < 3 the build of GCC will abort during
2411 stage1 with an error message like this:
2415 /usr/tmp/ccbbsZ0U.s:987:Unknown pseudo-op: .section
2416 /usr/tmp/ccbbsZ0U.s:987:Rest of line ignored. 1st junk character
2420 The reason for this is the fact that NeXT's assembler for these
2421 versions of the operating system does not support the @samp{.section}
2422 pseudo op that's needed for full C++ exception functionality.
2424 As NeXT's assembler is a derived work from GNU as, a free
2425 replacement that does can be obtained at
2426 @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}.
2428 If you try to build the integrated C++ & C++ runtime libraries on this system
2429 you will run into trouble with include files. The way to get around this is
2430 to use the following sequence. Note you must have write permission to
2431 the directory @var{prefix} you specified in the configuration process of GCC
2432 for this sequence to work.
2436 make all-texinfo all-bison all-byacc all-binutils all-gas all-ld
2439 make install-headers-tar
2448 @heading @anchor{m68k-ncr-*}m68k-ncr-*
2449 On the Tower models 4@var{n}0 and 6@var{n}0, by default a process is not
2450 allowed to have more than one megabyte of memory. GCC cannot compile
2451 itself (or many other programs) with @option{-O} in that much memory.
2453 To solve this problem, reconfigure the kernel adding the following line
2454 to the configuration file:
2465 @heading @anchor{m68k-sun}m68k-sun
2466 Sun 3. We do not provide a configuration file to use the Sun FPA by
2467 default, because programs that establish signal handlers for floating
2468 point traps inherently cannot work with the FPA.
2474 @heading @anchor{m68k-sun-sunos4.1.1}m68k-sun-sunos4.1.1
2476 It is reported that you may need the GNU assembler on this platform.
2483 @heading @anchor{m88k-*-svr3}m88k-*-svr3
2484 Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port.
2485 These systems tend to use the Green Hills C, revision 1.8.5, as the
2486 standard C compiler. There are apparently bugs in this compiler that
2487 result in object files differences between stage 2 and stage 3. If this
2488 happens, make the stage 4 compiler and compare it to the stage 3
2489 compiler. If the stage 3 and stage 4 object files are identical, this
2490 suggests you encountered a problem with the standard C compiler; the
2491 stage 3 and 4 compilers may be usable.
2493 It is best, however, to use an older version of GNU CC for bootstrapping
2500 @heading @anchor{m88k-*-dgux}m88k-*-dgux
2501 Motorola m88k running DG/UX. To build 88open BCS native or cross
2502 compilers on DG/UX, specify the configuration name as
2503 @samp{m88k-*-dguxbcs} and build in the 88open BCS software development
2504 environment. To build ELF native or cross compilers on DG/UX, specify
2505 @samp{m88k-*-dgux} and build in the DG/UX ELF development environment.
2506 You set the software development environment by issuing
2507 @samp{sde-target} command and specifying either @samp{m88kbcs} or
2508 @samp{m88kdguxelf} as the operand.
2510 If you do not specify a configuration name, @file{configure} guesses the
2511 configuration based on the current software development environment.
2517 @heading @anchor{m88k-tektronix-sysv3}m88k-tektronix-sysv3
2518 Tektronix XD88 running UTekV 3.2e. Do not turn on
2519 optimization while building stage1 if you bootstrap with
2520 the buggy Green Hills compiler. Also, The bundled LAI
2521 System V NFS is buggy so if you build in an NFS mounted
2522 directory, start from a fresh reboot, or avoid NFS all together.
2523 Otherwise you may have trouble getting clean comparisons
2530 @heading @anchor{mips-*-*}mips-*-*
2531 If you use the 1.31 version of the MIPS assembler (such as was shipped
2532 with Ultrix 3.1), you will need to use the @option{-fno-delayed-branch} switch
2533 when optimizing floating point code. Otherwise, the assembler will
2534 complain when the GCC compiler fills a branch delay slot with a
2535 floating point instruction, such as @code{add.d}.
2537 If on a MIPS system you get an error message saying ``does not have gp
2538 sections for all it's [sic] sectons [sic]'', don't worry about it. This
2539 happens whenever you use GAS with the MIPS linker, but there is not
2540 really anything wrong, and it is okay to use the output file. You can
2541 stop such warnings by installing the GNU linker.
2543 It would be nice to extend GAS to produce the gp tables, but they are
2544 optional, and there should not be a warning about their absence.
2546 Users have reported some problems with version 2.0 of the MIPS
2547 compiler tools that were shipped with Ultrix 4.1. Version 2.10
2548 which came with Ultrix 4.2 seems to work fine.
2550 Users have also reported some problems with version 2.20 of the
2551 MIPS compiler tools that were shipped with RISC/os 4.x. The earlier
2552 version 2.11 seems to work fine.
2554 Some versions of the MIPS linker will issue an assertion failure
2555 when linking code that uses @code{alloca} against shared
2556 libraries on RISC-OS 5.0, and DEC's OSF/1 systems. This is a bug
2557 in the linker, that is supposed to be fixed in future revisions.
2558 To protect against this, GCC passes @option{-non_shared} to the
2559 linker unless you pass an explicit @option{-shared} or
2560 @option{-call_shared} switch.
2562 @heading @anchor{mips-mips-bsd}mips-mips-bsd
2563 MIPS machines running the MIPS operating system in BSD mode. It's
2564 possible that some old versions of the system lack the functions
2565 @code{memcpy}, @code{memmove}, @code{memcmp}, and @code{memset}. If your
2566 system lacks these, you must remove or undo the definition of
2567 @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}.
2569 The MIPS C compiler needs to be told to increase its table size
2570 for switch statements with the @option{-Wf,-XNg1500} option in
2571 order to compile @file{cp/parse.c}. If you use the @option{-O2}
2572 optimization option, you also need to use @option{-Olimit 3000}.
2573 Both of these options are automatically generated in the
2574 @file{Makefile} that the shell script @file{configure} builds.
2575 If you override the @code{CC} make variable and use the MIPS
2576 compilers, you may need to add @option{-Wf,-XNg1500 -Olimit 3000}.
2582 @heading @anchor{mips-mips-riscos*}mips-mips-riscos*
2583 The MIPS C compiler needs to be told to increase its table size
2584 for switch statements with the @option{-Wf,-XNg1500} option in
2585 order to compile @file{cp/parse.c}. If you use the @option{-O2}
2586 optimization option, you also need to use @option{-Olimit 3000}.
2587 Both of these options are automatically generated in the
2588 @file{Makefile} that the shell script @file{configure} builds.
2589 If you override the @code{CC} make variable and use the MIPS
2590 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
2592 MIPS computers running RISC-OS can support four different
2593 personalities: default, BSD 4.3, System V.3, and System V.4
2594 (older versions of RISC-OS don't support V.4). To configure GCC
2595 for these platforms use the following configurations:
2598 @item mips-mips-riscos@code{rev}
2599 Default configuration for RISC-OS, revision @code{rev}.
2601 @item mips-mips-riscos@code{rev}bsd
2602 BSD 4.3 configuration for RISC-OS, revision @code{rev}.
2604 @item mips-mips-riscos@code{rev}sysv4
2605 System V.4 configuration for RISC-OS, revision @code{rev}.
2611 @item mips-mips-riscos@code{rev}sysv
2612 System V.3 configuration for RISC-OS, revision @code{rev}.
2615 The revision @code{rev} mentioned above is the revision of
2616 RISC-OS to use. You must reconfigure GCC when going from a
2617 RISC-OS revision 4 to RISC-OS revision 5. This has the effect of
2618 avoiding a linker bug.
2624 @heading @anchor{mips*-sgi-irix[45]}mips*-sgi-irix[45]
2626 In order to compile GCC on an SGI running IRIX 4, the ``c.hdr.lib''
2627 option must be installed from the CD-ROM supplied from Silicon Graphics.
2628 This is found on the 2nd CD in release 4.0.1.
2630 In order to compile GCC on an SGI running IRIX 5, the ``compiler_dev.hdr''
2631 subsystem must be installed from the IDO CD-ROM supplied by Silicon
2634 @code{make compare} may fail on version 5 of IRIX unless you add
2635 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2636 assembler input file is stored in the object file, and that makes
2637 comparison fail if it differs between the @code{stage1} and
2638 @code{stage2} compilations. The option @option{-save-temps} forces a
2639 fixed name to be used for the assembler input file, instead of a
2640 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2641 unless the comparisons fail without that option. If you do you
2642 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2643 @samp{.s} files after each series of compilations.
2645 The MIPS C compiler needs to be told to increase its table size
2646 for switch statements with the @option{-Wf,-XNg1500} option in
2647 order to compile @file{cp/parse.c}. 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 On Irix version 4.0.5F, and perhaps on some other versions as well,
2655 there is an assembler bug that reorders instructions incorrectly. To
2656 work around it, specify the target configuration
2657 @samp{mips-sgi-irix4loser}. This configuration inhibits assembler
2660 In a compiler configured with target @samp{mips-sgi-irix4}, you can turn
2661 off assembler optimization by using the @option{-noasmopt} option. This
2662 compiler option passes the option @option{-O0} to the assembler, to
2665 The @option{-noasmopt} option can be useful for testing whether a problem
2666 is due to erroneous assembler reordering. Even if a problem does not go
2667 away with @option{-noasmopt}, it may still be due to assembler
2668 reordering---perhaps GNU CC itself was miscompiled as a result.
2670 To enable debugging under Irix 5, you must use GNU as 2.5 or later,
2671 and use the @option{--with-gnu-as} configure option when configuring gcc.
2672 GNU as is distributed as part of the binutils package.
2674 You must use GAS on these platforms, as the native assembler can not handle
2675 the code for exception handling support. Either of these messages indicates
2676 that you are using the MIPS assembler when instead you should be using GAS:
2678 @samp{ as0: Error: ./libgcc2.c, line 1:Badly delimited numeric literal
2679 .4byte $LECIE1-$LSCIE1
2680 as0: Error: ./libgcc2.c, line 1:malformed statement}
2684 @samp{ as0: Error: /src/bld-gcc/gcc/libgcc2.c, line 1:undefined symbol in expression
2685 .word $LECIE1-$LSCIE1}
2687 These systems don't have ranlib, which various components in GCC need; you
2688 should be able to avoid this problem by installing GNU binutils, which includes
2689 a functional ranlib for this system.
2691 You may get the following warning on irix4 platforms, it can be safely
2694 warning: foo.o does not have gp tables for all its sections.
2697 When building GCC, the build process loops rebuilding @command{cc1} over and
2698 over again. This happens on mips-sgi-irix5.2, and possibly other platforms.@*
2699 It has been reported that this is a known bug in the @command{make} shipped with
2700 IRIX 5.2. We recommend you use GNU @command{make} instead of the vendor supplied
2701 @command{make} program; however, you may have success with @command{smake} on IRIX 5.2 if
2702 you do not have GNU @command{make} available.
2704 See @uref{http://freeware.sgi.com/,,http://freeware.sgi.com/} for more
2705 information about using GCC on IRIX platforms.
2712 @heading @anchor{mips*-sgi-irix6}mips*-sgi-irix6
2714 You must @emph{not} use GAS on irix6 platforms; doing so will only
2717 These systems don't have @command{ranlib}, which various components in GCC need; you
2718 should be able to avoid this problem by making a dummy script called @command{ranlib}
2719 which just exits with zero status and placing it in your path.
2721 If you are using Irix @command{cc} as your bootstrap compiler, you must
2722 ensure that the N32 ABI is in use. To test this, compile a simple C
2723 file with @command{cc} and then run @command{file} on the
2724 resulting object file. The output should look like:
2727 @code{ test.o: ELF N32 MSB @dots{}}
2732 @code{ test.o: ELF 32-bit MSB}
2735 then your version of @command{cc} uses the O32 ABI default. You
2736 should set the environment variable @env{CC} to @samp{cc -n32}
2737 before configuring GCC.
2739 GCC does not currently support generating O32 ABI binaries in the
2740 mips-sgi-irix6 configurations. It used to be possible to create a GCC
2741 with O32 ABI only support by configuring it for the mips-sgi-irix5
2742 target. See the link below for details.
2744 GCC does not correctly pass/return structures which are
2745 smaller than 16 bytes and which are not 8 bytes. The problem is very
2746 involved and difficult to fix. It affects a number of other targets also,
2747 but IRIX 6 is affected the most, because it is a 64 bit target, and 4 byte
2748 structures are common. The exact problem is that structures are being padded
2749 at the wrong end, e.g. a 4 byte structure is loaded into the lower 4 bytes
2750 of the register when it should be loaded into the upper 4 bytes of the
2753 GCC is consistent with itself, but not consistent with the SGI C compiler
2754 (and the SGI supplied runtime libraries), so the only failures that can
2755 happen are when there are library functions that take/return such
2756 structures. There are very few such library functions. I can only recall
2757 seeing two of them: inet_ntoa, and semctl.
2759 See @uref{http://freeware.sgi.com/,,http://freeware.sgi.com/} for more
2760 information about using GCC on IRIX platforms.
2767 @heading @anchor{mips-sony-sysv}mips-sony-sysv
2768 Sony MIPS NEWS. This works in NEWSOS 5.0.1, but not in 5.0.2 (which
2769 uses ELF instead of COFF). Support for 5.0.2 will probably be provided
2770 soon by volunteers. In particular, the linker does not like the
2771 code generated by GCC when shared libraries are linked in.
2778 @heading @anchor{ns32k-encore}ns32k-encore
2779 Encore ns32000 system. Encore systems are supported only under BSD.
2785 @heading @anchor{ns32k-*-genix}ns32k-*-genix
2786 National Semiconductor ns32000 system. Genix has bugs in @code{alloca}
2787 and @code{malloc}; you must get the compiled versions of these from GNU
2794 @heading @anchor{ns32k-sequent}ns32k-sequent
2795 Go to the Berkeley universe before compiling.
2801 @heading @anchor{ns32k-utek}ns32k-utek
2802 UTEK ns32000 system (``merlin''). The C compiler that comes with this
2803 system cannot compile GNU CC; contact @samp{tektronix!reed!mason} to get
2804 binaries of GNU CC for bootstrapping.
2811 @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4
2812 PowerPC system in big endian mode, running System V.4.
2814 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2815 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2821 @heading @anchor{powerpc-*-linux-gnu*}powerpc-*-linux-gnu*
2824 @uref{ftp://ftp.varesearch.com/pub/support/hjl/binutils,,binutils-2.9.4.0.8}
2825 or newer for a working GCC. It is strongly recommended to recompile binutils
2826 if you initially built it with gcc-2.7.2.x.
2828 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2829 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2835 @heading @anchor{powerpc-*-eabiaix}powerpc-*-eabiaix
2836 Embedded PowerPC system in big endian mode with @option{-mcall-aix} selected as
2839 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2840 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2846 @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim
2847 Embedded PowerPC system in big endian mode for use in running under the
2850 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2851 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2857 @heading @anchor{powerpc-*-eabi}powerpc-*-eabi
2858 Embedded PowerPC system in big endian mode.
2860 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2861 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2867 @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4
2868 PowerPC system in little endian mode, running System V.4.
2870 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2871 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2877 @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim
2878 Embedded PowerPC system in little endian mode for use in running under
2885 @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi
2886 Embedded PowerPC system in little endian mode.
2888 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2889 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2895 @heading @anchor{powerpcle-*-winnt}powerpcle-*-winnt, powerpcle-*-pe
2896 PowerPC system in little endian mode running Windows NT.
2898 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
2899 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
2905 @heading @anchor{romp-*-aos}romp-*-aos, romp-*-mach
2906 The only operating systems supported for the IBM RT PC are AOS and
2907 MACH. GNU CC does not support AIX running on the RT. We recommend you
2908 compile GNU CC with an earlier version of itself; if you compile GNU CC
2909 with @code{hc}, the Metaware compiler, it will work, but you will get
2910 mismatches between the stage 2 and stage 3 compilers in various files.
2911 These errors are minor differences in some floating-point constants and
2912 can be safely ignored; the stage 3 compiler is correct.
2918 @heading @anchor{*-*-solaris*}*-*-solaris*
2920 Starting with Solaris, Sun does not ship a C compiler any more. To
2921 bootstrap and install GCC you first have to install a pre-built
2922 compiler, see our @uref{binaries.html,,binaries page} for
2925 You must use GNU Make to build GCC on Solaris 2. If you don't have GNU
2926 Make installed, you can use the prebuilt compiler mentioned above to
2929 Sun as 4.X is broken in that it cannot cope with long symbol names.
2930 A typical error message might look similar to the following:
2932 @samp{/usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041:
2933 error: can't compute value of an expression involving an external symbol.}
2935 This is Sun bug 4237974. This is fixed with patch 108908-02 and has
2936 been fixed in later (5.x) versions of the assembler.
2938 Solaris' @file{/bin/sh} will often fail to configure libstdc++-v3, boehm-gc or
2939 libjava. If you encounter this problem, set @var{CONFIG_SHELL} to
2940 @file{/bin/ksh} in your environment and run @samp{make bootstrap} again.
2941 Another possibility that sometimes helps is to remove
2942 @file{*-*-solaris*/config.cache}.
2948 @heading @anchor{sparc-sun-*}sparc-sun-*
2949 Sometimes on a Sun 4 you may observe a crash in the program
2950 @code{genflags} or @code{genoutput} while building GCC. This is said to
2951 be due to a bug in @code{sh}. You can probably get around it by running
2952 @code{genflags} or @code{genoutput} manually and then retrying the
2959 @heading @anchor{sparc-sun-solaris*}sparc-sun-solaris*
2960 On Solaris 2, executables of GCC version 2.0.2 are commonly
2961 available, but they have a bug that shows up when compiling current
2962 versions of GCC: undefined symbol errors occur during assembly if you
2965 The solution is to compile the current version of GCC without
2966 @option{-g}. That makes a working compiler which you can use to recompile
2969 Solaris 2 comes with a number of optional OS packages. Some of these
2970 packages are needed to use GCC fully. If you did not install all
2971 optional packages when installing Solaris, you will need to verify that
2972 the packages that GCC needs are installed.
2974 To check whether an optional package is installed, use
2975 the @code{pkginfo} command. To add an optional package, use the
2976 @code{pkgadd} command. For further details, see the Solaris
2979 For Solaris 2.0 and 2.1, GCC needs six packages: @samp{SUNWarc},
2980 @samp{SUNWbtool}, @samp{SUNWesu}, @samp{SUNWhea}, @samp{SUNWlibm}, and
2983 For Solaris 2.2, GCC needs an additional seventh package: @samp{SUNWsprot}.
2985 On Solaris 2, trying to use the linker and other tools in
2986 @file{/usr/ucb} to install GCC has been observed to cause trouble.
2987 For example, the linker may hang indefinitely. The fix is to remove
2988 @file{/usr/ucb} from your @code{PATH}.
2990 All releases of GNU binutils prior to 2.11.1 have known bugs on this
2991 platform. We recommend the use of GNU binutils 2.11.1 or the vendor
2992 tools (Sun as, Sun ld).
2994 Unfortunately, C++ shared libraries, including libstdc++, won't work
2995 properly if assembled with Sun as: the linker will complain about
2996 relocations in read-only sections, in the definition of virtual
2997 tables. Also, Sun as fails to process long symbols resulting from
2998 mangling template-heavy C++ function names.
3004 @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7
3006 Sun patch 107058-01 (1999-01-13) for SPARC Solaris 7 triggers a bug in
3007 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3008 and later, including all EGCS releases. Sun formerly recommended
3009 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3010 recommend it only for people who use Sun's compilers.
3012 Here are some workarounds to this problem:
3015 Do not install Sun patch 107058-01 until after Sun releases a
3016 complete patch for bug 4210064. This is the simplest course to take,
3017 unless you must also use Sun's C compiler. Unfortunately 107058-01
3018 is preinstalled on some new Solaris-based hosts, so you may have to
3022 Copy the original, unpatched Solaris 7
3023 @command{/usr/ccs/bin/as} into
3024 @command{/usr/local/lib/gcc-lib/sparc-sun-solaris2.7/2.95.1/as},
3025 adjusting the latter name to fit your local conventions and software
3029 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3030 both 107058-01 and 106950-03 installed has reported the bug with GCC
3031 and Sun's dynamic linker. This last course of action is riskiest,
3032 for two reasons. First, you must install 106950 on all hosts that
3033 run code generated by GCC; it doesn't suffice to install it only on
3034 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3035 only a partial fix for bug 4210064, but Sun doesn't know whether the
3036 partial fix is adequate for GCC. Revision -08 or later should fix
3037 the bug, but (as of 1999-10-06) it is still being tested.
3044 <!-- ripped from the same FAQ that I answered -->
3046 @heading @anchor{*-sun-solaris2.8}*-sun-solaris2.8
3048 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3049 newer: @command{g++} will complain that types are missing. These headers assume
3050 that omitting the type means @code{int}; this assumption worked for C89 but
3051 is wrong for C++, and is now wrong for C99 also.
3053 @command{g++} accepts such (illegal) constructs with the option @option{-fpermissive}; it
3054 will assume that any missing type is @code{int} (as defined by C89).
3056 For Solaris 8, this is fixed by revision 24 or later of patch 108652
3057 (for SPARCs) or 108653 (for Intels).
3059 Solaris 8's linker fails to link some libjava programs if
3060 previously-installed GCC java libraries already exist in the configured
3061 prefix. For this reason, libgcj is disabled by default on Solaris 8.
3062 If you use GNU ld, or if you don't have a previously-installed libgcj in
3063 the same prefix, use @option{--enable-libgcj} to build and install the
3070 @heading @anchor{sunv5}Sun V5.0 Compiler Bugs
3072 The Sun V5.0 compilers are known to mis-compile GCC 2.95 and GCC 2.95.1,
3073 which in turn causes GCC to fail its bootstrap comparison test.
3074 GCC 2.95.2 has a workaround.
3081 @heading @anchor{sparc-sun-sunos*}sparc-sun-sunos*
3083 A bug in the SunOS4 linker will cause it to crash when linking
3084 @option{-fPIC} compiled objects (and will therefore not allow you to build
3087 To fix this problem you can either use the most recent version of
3088 binutils or get the latest SunOS4 linker patch (patch ID 100170-10)
3089 from Sun's patch site.
3096 @heading @anchor{sparc-unknown-linux-gnulibc1}sparc-unknown-linux-gnulibc1
3098 It has been reported that you might need
3099 @uref{ftp://ftp.yggdrasil.com/private/hjl,,binutils-2.8.1.0.23}
3100 for this platform, too.
3107 @heading @anchor{sparc64-*-*}sparc64-*-*
3109 GCC version 2.95 is not able to compile code correctly for
3110 @code{sparc64} targets. Users of the Linux kernel, at least,
3111 12~can use the @code{sparc32} program to start up a new shell
3112 invocation with an environment that causes @command{configure} to
3113 recognize (via @samp{uname -a}) the system as @var{sparc-*-*} instead.
3120 @heading @anchor{#*-*-sysv*}*-*-sysv*
3121 On System V release 3, you may get this error message
3125 ld fatal: failed to write symbol name @var{something}
3126 in strings table for file @var{whatever}
3129 This probably indicates that the disk is full or your ULIMIT won't allow
3130 the file to be as large as it needs to be.
3132 This problem can also result because the kernel parameter @code{MAXUMEM}
3133 is too small. If so, you must regenerate the kernel and make the value
3134 much larger. The default value is reported to be 1024; a value of 32768
3135 is said to work. Smaller values may also work.
3137 On System V, if you get an error like this,
3140 /usr/local/lib/bison.simple: In function `yyparse':
3141 /usr/local/lib/bison.simple:625: virtual memory exhausted
3145 that too indicates a problem with disk space, ULIMIT, or @code{MAXUMEM}.
3147 On a System V release 4 system, make sure @file{/usr/bin} precedes
3148 @file{/usr/ucb} in @code{PATH}. The @code{cc} command in
3149 @file{/usr/ucb} uses libraries which have bugs.
3155 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3156 Don't try compiling with Vax C (@code{vcc}). It produces incorrect code
3157 in some cases (for example, when @code{alloca} is used).
3159 Meanwhile, compiling @file{cp/parse.c} with pcc does not work because of
3160 an internal table size limitation in that compiler. To avoid this
3161 problem, compile just the GNU C compiler first, and use it to recompile
3162 building all the languages that you want to run.
3168 @heading @anchor{we32k-*-*}we32k-*-*
3169 These computers are also known as the 3b2, 3b5, 3b20 and other similar
3170 names. (However, the 3b1 is actually a 68000.)
3172 Don't use @option{-g} when compiling with the system's compiler. The
3173 system's linker seems to be unable to handle such a large program with
3174 debugging information.
3176 The system's compiler runs out of capacity when compiling @file{stmt.c}
3177 in GNU CC. You can work around this by building @file{cpp} in GNU CC
3178 first, then use that instead of the system's preprocessor with the
3179 system's C compiler to compile @file{stmt.c}. Here is how:
3182 mv /lib/cpp /lib/cpp.att
3184 echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp
3188 The system's compiler produces bad code for some of the GNU CC
3189 optimization files. So you must build the stage 2 compiler without
3190 optimization. Then build a stage 3 compiler with optimization.
3191 That executable should work. Here are the necessary commands:
3194 make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g"
3196 make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O"
3199 You may need to raise the ULIMIT setting to build a C++ compiler,
3200 as the file @file{cc1plus} is larger than one megabyte.
3206 @heading @anchor{windows}Microsoft Windows (32 bit)
3208 A port of GCC 2.95.x is included with the
3209 @uref{http://www.cygwin.com/,,Cygwin environment}.
3211 Current (as of early 2001) snapshots of GCC will build under Cygwin
3212 without modification.
3218 @heading @anchor{os2}OS/2
3220 GCC does not currently support OS/2. However, Andrew Zabolotny has been
3221 working on a generic OS/2 port with pgcc. The current code code can be found
3222 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
3224 An older copy of GCC 2.8.1 is included with the EMX tools available at
3225 @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,,
3226 ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}.
3232 @heading @anchor{older}Older systems
3234 GCC contains support files for many older (1980s and early
3235 1990s) Unix variants. For the most part, support for these systems
3236 has not been deliberately removed, but it has not been maintained for
3237 several years and may suffer from bitrot. Support from some systems
3238 has been removed from GCC 3: fx80, ns32-ns-genix, pyramid, tahoe,
3239 gmicro, spur; most of these targets had not been updated since GCC
3242 Support for older systems as targets for cross-compilation is less
3243 problematic than support for them as hosts for GCC; if an enthusiast
3244 wishes to make such a target work again (including resurrecting any
3245 of the targets that never worked with GCC 2, starting from the last
3246 CVS version before they were removed), patches
3247 @uref{../contribute.html,,following the usual requirements}
3248 would be likely to be accepted, since they should not affect the
3249 support for more modern targets.
3251 Support for old systems as hosts for GCC can cause problems if the
3252 workarounds for compiler, library and operating system bugs affect the
3253 cleanliness or maintainability of the rest of GCC. In some cases, to
3254 bring GCC up on such a system, if still possible with current GCC, may
3255 require first installing an old version of GCC which did work on that
3256 system, and using it to compile a more recent GCC, to avoid bugs in
3257 the vendor compiler. Old releases of GCC 1 and GCC 2 are available in
3258 the old-releases directory on the
3259 @uref{../mirrors.html,,GCC mirror sites}. Header bugs may generally
3260 be avoided using @command{fixincludes}, but bugs or deficiencies in
3261 libraries and the operating system may still cause problems.
3263 For some systems, old versions of GNU binutils may also be useful,
3264 and are available from pub/binutils/old-releases on
3265 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
3267 Some of the information on specific systems above relates to
3268 such older systems, but much of the information
3269 about GCC on such systems (which may no longer be applicable to
3270 current GCC) is to be found in the GCC texinfo manual.
3276 @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris, etc.)
3278 C++ support is significantly better on ELF targets if you use the
3279 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
3280 inlines, vtables and template instantiations will be discarded
3290 @uref{./index.html,,Return to the GCC Installation page}
3294 @c ***************************************************************************
3295 @c Part 6 The End of the Document
3297 @comment node-name, next, previous, up
3298 @node Concept Index, , Specific, Top
3302 @unnumbered Concept Index