1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename install.info
5 @settitle Installing GCC
10 @c Specify title for specific html page
12 @settitle Installing GCC
15 @settitle Host/Target specific installation notes for GCC
18 @settitle Downloading GCC
21 @settitle Installing GCC: Configuration
24 @settitle Installing GCC: Building
27 @settitle Installing GCC: Testing
29 @ifset finalinstallhtml
30 @settitle Installing GCC: Final installation
33 @settitle Installing GCC: Binaries
36 @c Copyright (C) 2001 Free Software Foundation, Inc.
37 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
39 @c Include everything if we're not making html
51 @c Part 2 Summary Description and Copyright
54 Copyright @copyright{} 2001 Free Software Foundation, Inc.
57 @c Part 3 Titlepage and Copyright
60 @comment The title is printed in a large font.
61 @center @titlefont{Installing GCC}
63 @c The following two commands start the copyright page.
65 @vskip 0pt plus 1filll
66 Copyright @copyright{} 2001 Free Software Foundation, Inc.
69 @c Part 4 Top node and Master Menu
72 @comment node-name, next, Previous, up
75 * Installing GCC:: This document describes the generic installation
76 procedure for GCC as well as detailing some target
77 specific installation instructions.
79 * Specific:: Host/target specific installation notes for GCC.
80 * Binaries:: Where to get pre-compiled binaries.
82 * Concept Index:: This index has two entries.
86 @c Part 5 The Body of the Document
87 @c ***Installing GCC**********************************************************
89 @comment node-name, next, previous, up
90 @node Installing GCC, Binaries, , Top
94 <h1 align="center">Installing GCC</h1>
97 @chapter Installing GCC
100 The latest version of this document is always available at
101 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
103 This document describes the generic installation procedure for GCC as well
104 as detailing some target specific installation instructions.
106 GCC includes several components that previously were separate distributions
107 with their own installation instructions. This document supersedes all
108 package specific installation instructions.
110 @emph{Before} starting the build/install procedure please check the
112 @ref{Specific, host/target specific installation notes}.
115 @uref{specific.html,,host/target specific installation notes}.
117 We recommend you browse the entire generic installation instructions before
120 Lists of successful builds for released versions of GCC are
121 available at our web pages for
122 @uref{http://gcc.gnu.org/gcc-3.0/buildstat.html,,3.0}
124 @uref{http://gcc.gnu.org/gcc-2.95/buildstat.html,,2.95}.
125 These lists are updated as new information becomes available.
127 The installation procedure itself is broken into five steps.
131 * Downloading the source::
134 * Testing:: (optional)
141 @uref{download.html,,Downloading the source}
143 @uref{configure.html,,Configuration}
145 @uref{build.html,,Building}
147 @uref{test.html,,Testing} (optional)
149 @uref{finalinstall.html,,Final install}
153 Please note that GCC does not support @samp{make uninstall} and probably
154 won't do so in the near future as this would open a can of worms. Instead,
155 we suggest that you install GCC into a directory of its own and simply
156 remove that directory when you do not need that specific version of GCC
157 any longer, and, if shared libraries are installed there as well, no
158 more binaries exist that use them.
165 @uref{./index.html,,Return to the GCC Installation page}
169 @c ***Downloading the source**************************************************
171 @comment node-name, next, previous, up
172 @node Downloading the source, Configuration, , Installing GCC
176 <h1 align="center">Downloading GCC</h1>
179 @chapter Downloading GCC
181 @cindex Downloading GCC
182 @cindex Downloading the Source
184 GCC is distributed via @uref{http://gcc.gnu.org/cvs.html,,CVS} and FTP
185 tarballs compressed with @command{gzip} or
186 @command{bzip2}. It is possible to download a full distribution or specific
189 Please refer to our @uref{http://gcc.gnu.org/releases.html,,releases web page}
190 for information on how to obtain GCC@.
192 The full distribution includes the C, C++, Objective-C, Fortran, Java,
193 and CHILL compilers. The full distribution also includes runtime libraries
194 for C++, Objective-C, Fortran, Java and CHILL. (GCC 3.0 does not
195 include CHILL.) In GCC 3.0 and later versions, GNU compiler testsuites
196 are also included in the full distribution.
198 If you choose to download specific components, you must download the core
199 GCC distribution plus any language specific distributions you wish to
200 use. The core distribution includes the C language front end as well as the
201 shared components. Each language has a tarball which includes the language
202 front end as well as the language runtime (when appropriate).
204 Unpack the core distribution as well as any language specific
205 distributions in the same directory.
207 If you also intend to build binutils (either to upgrade an existing
208 installation or for use in place of the corresponding tools of your
209 OS), unpack the binutils distribution either in the same directory or
210 a separate one. In the latter case, add symbolic links to any
211 components of the binutils you intend to build alongside the compiler
212 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
213 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
220 @uref{./index.html,,Return to the GCC Installation page}
224 @c ***Configuration***********************************************************
226 @comment node-name, next, previous, up
227 @node Configuration, Building, Downloading the source, Installing GCC
231 <h1 align="center">Installing GCC: Configuration</h1>
234 @chapter Installing GCC: Configuration
236 @cindex Configuration
237 @cindex Installing GCC: Configuration
239 Like most GNU software, GCC must be configured before it can be built.
240 This document describes the recommended configuration procedure
241 for both native and cross targets.
243 We use @var{srcdir} to refer to the toplevel source directory for
244 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
246 If you obtained the sources via CVS, @var{srcdir} must refer to the top
247 @file{gcc} directory, the one where the @file{MAINTAINERS} can be found,
248 and not its @file{gcc} subdirectory, otherwise the build will fail.
250 First, we @strong{highly} recommend that GCC be built into a
251 separate directory than the sources which does @strong{not} reside
252 within the source tree. This is how we generally build GCC; building
253 where @var{srcdir} == @var{objdir} should still work, but doesn't
254 get extensive testing; building where @var{objdir} is a subdirectory
255 of @var{srcdir} is unsupported.
257 If you have previously built GCC in the same directory for a
258 different target machine, do @samp{make distclean} to delete all files
259 that might be invalid. One of the files this deletes is
260 @file{Makefile}; if @samp{make distclean} complains that @file{Makefile}
261 does not exist, it probably means that the directory is already suitably
262 clean. However, with the recommended method of building in a separate
263 @var{objdir}, you should simply use a different @var{objdir} for each
266 Second, when configuring a native system, either @command{cc} or
267 @command{gcc} must be in your path or you must set @env{CC} in
268 your environment before running configure. Otherwise the configuration
271 Note that the bootstrap compiler and the resulting GCC must be link
272 compatible, else the bootstrap will fail with linker errors about
273 incompatible object file formats. Several multilibed targets are
274 affected by this requirement, see
276 @ref{Specific, host/target specific installation notes}.
279 @uref{specific.html,,host/target specific installation notes}.
287 % @var{srcdir}/configure [@var{options}] [@var{target}]
291 @heading Target specification
294 GCC has code to correctly determine the correct value for @var{target}
295 for nearly all native systems. Therefore, we highly recommend you not
296 provide a configure target when configuring a native compiler.
299 @var{target} must be specified as @option{--target=@var{target}}
300 when configuring a cross compiler; examples of valid targets would be
301 i960-rtems, m68k-coff, sh-elf, etc.
304 Specifying just @var{target} instead of @option{--target=@var{target}}
305 implies that the host defaults to @var{target}.
309 @heading Options specification
311 Use @var{options} to override several configure time options for
312 GCC@. A partial list of supported @var{options}:
315 @item --prefix=@var{dirname}
316 Specify the toplevel installation
317 directory. This is the recommended way to install the tools into a directory
318 other than the default. The toplevel installation directory defaults to
321 We @strong{highly} recommend against @var{dirname} being the same or a
322 subdirectory of @var{objdir} or vice versa.
324 These additional options control where certain parts of the distribution
325 are installed. Normally you should not need to use these options.
328 @item --exec-prefix=@var{dirname}
329 Specify the toplevel installation directory for architecture-dependent
330 files. The default is @file{@var{prefix}}.
332 @item --bindir=@var{dirname}
333 Specify the installation directory for the executables called by users
334 (such as @command{gcc} and @command{g++}). The default is
335 @file{@var{exec-prefix}/bin}.
337 @item --libdir=@var{dirname}
338 Specify the installation directory for object code libraries and
339 internal parts of GCC@. The default is @file{@var{exec-prefix}/lib}.
341 @item --with-slibdir=@var{dirname}
342 Specify the installation directory for the shared libgcc library. The
343 default is @file{@var{libdir}}.
345 @item --infodir=@var{dirname}
346 Specify the installation directory for documentation in info format.
347 The default is @file{@var{prefix}/info}.
349 @item --mandir=@var{dirname}
350 Specify the installation directory for manual pages. The default is
351 @file{@var{prefix}/man}. (Note that the manual pages are only extracts from
352 the full GCC manuals, which are provided in Texinfo format. The
353 @command{g77} manpage is unmaintained and may be out of date; the others
354 are derived by an automatic conversion process from parts of the full
357 @item --with-gxx-include-dir=@var{dirname}
359 the installation directory for G++ header files. The default is
360 @file{@var{prefix}/include/g++-v3}.
364 @item --with-local-prefix=@var{dirname}
366 installation directory for local include files. The default is
367 @file{/usr/local}. Specify this option if you want the compiler to
368 search directory @file{@var{dirname}/include} for locally installed
369 header files @emph{instead} of @file{/usr/local/include}.
371 You should specify @option{--with-local-prefix} @strong{only} if your
372 site has a different convention (not @file{/usr/local}) for where to put
375 The default value for @option{--with-local-prefix} is @file{/usr/local}
376 regardless of the value of @option{--prefix}. Specifying
377 @option{--prefix} has no effect on which directory GCC searches for
378 local header files. This may seem counterintuitive, but actually it is
381 The purpose of @option{--prefix} is to specify where to @emph{install
382 GCC}. The local header files in @file{/usr/local/include}---if you put
383 any in that directory---are not part of GCC@. They are part of other
384 programs---perhaps many others. (GCC installs its own header files in
385 another directory which is based on the @option{--prefix} value.)
387 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
388 The directory you use for @option{--with-local-prefix} @strong{must not}
389 contain any of the system's standard header files. If it did contain
390 them, certain programs would be miscompiled (including GNU Emacs, on
391 certain targets), because this would override and nullify the header
392 file corrections made by the @code{fixincludes} script.
394 Indications are that people who use this option use it based on mistaken
395 ideas of what it is for. People use it as if it specified where to
396 install part of GCC@. Perhaps they make this assumption because
397 installing GCC creates the directory.
399 @item --enable-shared[=@var{package}[,@dots{}]]
400 Build shared versions of libraries, if shared libraries are supported on
401 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
402 are enabled by default on all platforms that support shared libraries,
403 except for @samp{libobjc} which is built as a static library only by
406 If a list of packages is given as an argument, build shared libraries
407 only for the listed packages. For other packages, only static libraries
408 will be built. Package names currently recognized in the GCC tree are
409 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
410 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc} and
411 @samp{libjava}. Note that @samp{libobjc} does not recognize itself by
412 any name, so, if you list package names in @option{--enable-shared},
413 you will only get static Objective-C libraries. @samp{libf2c} and
414 @samp{libiberty} do not support shared libraries at all.
416 Use @option{--disable-shared} to build only static libraries. Note that
417 @option{--disable-shared} does not accept a list of package names as
418 argument, only @option{--enable-shared} does.
420 @item @anchor{with-gnu-as}--with-gnu-as
421 Specify that the compiler should assume that the
422 assembler it finds is the GNU assembler. However, this does not modify
423 the rules to find an assembler and will result in confusion if found
424 assembler is not actually the GNU assembler. (Confusion will also
425 result if the compiler finds the GNU assembler but has not been
426 configured with @option{--with-gnu-as}.) If you have more than one
427 assembler installed on your system, you may want to use this option in
428 connection with @option{--with-as=@var{pathname}}.
430 The systems where it makes a difference whether you use the GNU assembler are
431 @samp{hppa1.0-@var{any}-@var{any}}, @samp{hppa1.1-@var{any}-@var{any}},
432 @samp{i386-@var{any}-sysv}, @samp{i386-@var{any}-isc},
433 @samp{i860-@var{any}-bsd}, @samp{m68k-bull-sysv},
434 @samp{m68k-hp-hpux}, @samp{m68k-sony-bsd},
435 @samp{m68k-altos-sysv}, @samp{m68000-hp-hpux},
436 @samp{m68000-att-sysv}, @samp{@var{any}-lynx-lynxos},
437 and @samp{mips-@var{any}}.
438 On any other system, @option{--with-gnu-as} has no effect.
440 On the systems listed above (except for the HP-PA, for ISC on the
441 386, and for @samp{mips-sgi-irix5.*}), if you use the GNU assembler,
442 you should also use the GNU linker (and specify @option{--with-gnu-ld}).
444 @item --with-as=@var{pathname}
446 compiler should use the assembler pointed to by @var{pathname}, rather
447 than the one found by the standard rules to find an assembler, which
452 @file{@var{exec_prefix}/lib/gcc-lib/@var{target}/@var{version}}
453 directory, where @var{exec_prefix} defaults to @var{prefix} which
454 defaults to @file{/usr/local} unless overridden by the
455 @option{--prefix=@var{pathname}} switch described above. @var{target} is the
456 target system triple, such as @samp{sparc-sun-solaris2.7}, and
457 @var{version} denotes the GCC version, such as 3.0.
459 Check operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
462 Note that these rules do not check for the value of @env{PATH}. You may
463 want to use @option{--with-as} if no assembler is installed in the
464 directories listed above, or if you have multiple assemblers installed
465 and want to choose one that is not found by the above rules.
467 @item @anchor{with-gnu-ld}--with-gnu-ld
468 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
472 @item --with-ld=@var{pathname}
474 @option{--with-as}, but for the linker.
477 Specify that stabs debugging
478 information should be used instead of whatever format the host normally
479 uses. Normally GCC uses the same debug format as the host system.
481 On MIPS based systems and on Alphas, you must specify whether you want
482 GCC to create the normal ECOFF debugging format, or to use BSD-style
483 stabs passed through the ECOFF symbol table. The normal ECOFF debug
484 format cannot fully handle languages other than C@. BSD stabs format can
485 handle other languages, but it only works with the GNU debugger GDB@.
487 Normally, GCC uses the ECOFF debugging format by default; if you
488 prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@.
490 No matter which default you choose when you configure GCC, the user
491 can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly
492 the debug format for a particular compilation.
494 @option{--with-stabs} is meaningful on the ISC system on the 386, also, if
495 @option{--with-gas} is used. It selects use of stabs debugging
496 information embedded in COFF output. This kind of debugging information
497 supports C++ well; ordinary COFF debugging information does not.
499 @option{--with-stabs} is also meaningful on 386 systems running SVR4. It
500 selects use of stabs debugging information embedded in ELF output. The
501 C++ compiler currently (2.6.0) does not support the DWARF debugging
502 information normally used on 386 SVR4 platforms; stabs provide a
503 workable alternative. This requires gas and gdb, as the normal SVR4
504 tools can not generate or interpret stabs.
506 @item --disable-multilib
507 Specify that multiple target
508 libraries to support different target variants, calling
509 conventions, etc should not be built. The default is to build a
510 predefined set of them.
512 Some targets provide finer-grained control over which multilibs are built
513 (e.g., @option{--disable-softfloat}):
520 fpu, 26bit, underscore, interwork, biendian, nofmult.
523 softfloat, m68881, m68000, m68020.
526 single-float, biendian, softfloat.
528 @item powerpc*-*-*, rs6000*-*-*
529 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
534 @item --enable-threads
535 Specify that the target
536 supports threads. This affects the Objective-C compiler and runtime
537 library, and exception handling for other languages like C++ and Java.
538 On some systems, this is the default.
540 In general, the best (and, in many cases, the only known) threading
541 model available will be configured for use. Beware that on some
542 systems, gcc has not been taught what threading models are generally
543 available for the system. In this case, @option{--enable-threads} is an
544 alias for @option{--enable-threads=single}.
546 @item --disable-threads
547 Specify that threading support should be disabled for the system.
548 This is an alias for @option{--enable-threads=single}.
550 @item --enable-threads=@var{lib}
552 @var{lib} is the thread support library. This affects the Objective-C
553 compiler and runtime library, and exception handling for other languages
554 like C++ and Java. The possibilities for @var{lib} are:
562 Generic MACH thread support, known to work on NeXTSTEP@. (Please note
563 that the file needed to support this configuration, @file{gthr-mach.h}, is
564 missing and thus this setting will cause a known bootstrap failure.)
566 This is an alias for @samp{single}.
568 Generic POSIX thread support.
570 Same as @samp{posix} on arm*-*-linux*, *-*-chorusos* and *-*-freebsd*
571 only. A future release of gcc might remove this alias or extend it
574 RTEMS thread support.
576 Disable thread support, should work for all platforms.
578 Sun Solaris 2 thread support.
580 VxWorks thread support.
582 Microsoft Win32 API thread support.
585 @item --with-cpu=@var{cpu}
586 Specify which cpu variant the
587 compiler should generate code for by default. This is currently
588 only supported on the some ports, specifically arm, powerpc, and
589 SPARC@. If configure does not recognize the model name (e.g.@: arm700,
590 603e, or ultrasparc) you provide, please check the configure script
591 for a complete list of supported models.
593 @item --enable-target-optspace
595 libraries should be optimized for code space instead of code speed.
596 This is the default for the m32r platform.
599 Specify that a user visible @command{cpp} program should not be installed.
601 @item --with-cpp-install-dir=@var{dirname}
602 Specify that the user visible @command{cpp} program should be installed
603 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
605 @item --enable-maintainer-mode
607 regenerate the GCC master message catalog @file{gcc.pot} are normally
608 disabled. This is because it can only be rebuilt if the complete source
609 tree is present. If you have changed the sources and want to rebuild the
610 catalog, configuring with @option{--enable-maintainer-mode} will enable
611 this. Note that you need a recent version of the @code{gettext} tools
614 @item --enable-version-specific-runtime-libs
616 that runtime libraries should be installed in the compiler specific
617 subdirectory (@file{@var{libsubdir}}) rather than the usual places. In
618 addition, @samp{libstdc++}'s include files will be installed in
619 @file{@var{libsubdir}/include/g++} unless you overruled it by using
620 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
621 particularly useful if you intend to use several versions of GCC in
622 parallel. This is currently supported by @samp{libf2c} and
623 @samp{libstdc++}, and is the default for @samp{libobjc} which cannot be
624 changed in this case.
626 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
627 Specify that only a particular subset of compilers and
628 their runtime libraries should be built. For a list of valid values for
629 @var{langN} you can issue the following command in the
630 @file{gcc} directory of your GCC source tree:@*
632 grep language= */config-lang.in
634 Currently, you can use any of the following:
635 @code{c}, @code{c++}, @code{f77}, @code{java} and @code{objc}.
636 @code{CHILL} is not currently maintained, and will almost
637 certainly fail to compile.@*
638 If you do not pass this flag, all languages available in the @file{gcc}
639 sub-tree will be configured. Re-defining @code{LANGUAGES} when calling
640 @samp{make bootstrap} @strong{does not} work anymore, as those
641 language sub-directories might not have been configured!
643 @item --disable-libgcj
644 Specify that the run-time libraries
645 used by GCJ should not be built. This is useful in case you intend
646 to use GCJ with some other run-time, or you're going to install it
647 separately, or it just happens not to build on your particular
648 machine. In general, if the Java front end is enabled, the GCJ
649 libraries will be enabled too, unless they're known to not work on
650 the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you
651 may need to port it; in this case, before modifying the top-level
652 @file{configure.in} so that @samp{libgcj} is enabled by default on this platform,
653 you may use @option{--enable-libgcj} to override the default.
656 Specify that the compiler should
657 use DWARF 2 debugging information as the default.
659 @item --enable-win32-registry
660 @itemx --enable-win32-registry=@var{key}
661 @itemx --disable-win32-registry
662 The @option{--enable-win32-registry} option enables Windows-hosted GCC
663 to look up installations paths in the registry using the following key:
666 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
669 @var{key} defaults to GCC version number, and can be overridden by the
670 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
671 who use custom installers are encouraged to provide a different key,
672 perhaps one comprised of vendor name and GCC version number, to
673 avoid conflict with existing installations. This feature is enabled
674 by default, and can be disabled by @option{--disable-win32-registry}
675 option. This option has no effect on the other hosts.
678 Specify that the machine does not have a floating point unit. This
679 option only applies to @samp{m68k-sun-sunos@var{n}} and
680 @samp{m68k-isi-bsd}. On any other system, @option{--nfp} has no effect.
682 @item --enable-checking
683 @itemx --enable-checking=@var{list}
684 When you specify this option, the compiler is built to perform checking
685 of tree node types when referencing fields of that node, and some other
686 internal consistency checks. This does not change the generated code,
687 but adds error checking within the compiler. This will slow down the
688 compiler and may only work properly if you are building the compiler
689 with GCC@. This is on by default when building from CVS or snapshots,
690 but off for releases. More control over the checks may be had by
691 specifying @var{list}; the categories of checks available are
692 @samp{misc}, @samp{tree}, @samp{gc}, @samp{rtl} and @samp{gcac}. The
693 default when @var{list} is not specified is @samp{misc,tree,gc}; the
694 checks @samp{rtl} and @samp{gcac} are very expensive.
698 The @option{--enable-nls} option enables Native Language Support (NLS),
699 which lets GCC output diagnostics in languages other than American
700 English. Native Language Support is enabled by default if not doing a
701 canadian cross build. The @option{--disable-nls} option disables NLS@.
703 @item --with-included-gettext
704 If NLS is enabled, the @option{--with-included-gettext} option causes the build
705 procedure to prefer its copy of GNU @command{gettext}.
708 If NLS is enabled, and if the host lacks @code{gettext} but has the
709 inferior @code{catgets} interface, the GCC build procedure normally
710 ignores @code{catgets} and instead uses GCC's copy of the GNU
711 @code{gettext} library. The @option{--with-catgets} option causes the
712 build procedure to use the host's @code{catgets} in this situation.
715 Some options which only apply to building cross compilers:
717 @item --with-headers=@var{dir}
718 Specifies a directory
719 which has target include files.
720 @emph{This options is required} when building a cross
721 compiler, if @file{@var{prefix}/@var{target}/sys-include} doesn't pre-exist.
722 These include files will be copied into the @file{gcc} install directory.
723 Fixincludes will be run on these files to make them compatible with
725 @item --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}''
726 Specifies a list of directories which contain the target runtime
727 libraries. These libraries will be copied into the @file{gcc} install
730 Specifies that @samp{newlib} is
731 being used as the target C library. This causes @code{__eprintf} to be
732 omitted from @file{libgcc.a} on the assumption that it will be provided by
736 Note that each @option{--enable} option has a corresponding
737 @option{--disable} option and that each @option{--with} option has a
738 corresponding @option{--without} option.
745 @uref{./index.html,,Return to the GCC Installation page}
749 @c ***Building****************************************************************
751 @comment node-name, next, previous, up
752 @node Building, Testing, Configuration, Installing GCC
756 <h1 align="center">Installing GCC: Building</h1>
761 @cindex Installing GCC: Building
763 Now that GCC is configured, you are ready to build the compiler and
766 We @strong{highly} recommend that GCC be built using GNU make;
767 other versions may work, then again they might not.
769 (For example, many broken versions of make will fail if you use the
770 recommended setup where @var{objdir} is different from @var{srcdir}.
771 Other broken versions may recompile parts of the compiler when
772 installing the compiler.)
774 Some commands executed when making the compiler may fail (return a
775 nonzero status) and be ignored by @code{make}. These failures, which
776 are often due to files that were not found, are expected, and can safely
779 It is normal to have compiler warnings when compiling certain files.
780 Unless you are a GCC developer, you can generally ignore these warnings
781 unless they cause compilation to fail.
783 On certain old systems, defining certain environment variables such as
784 @env{CC} can interfere with the functioning of @command{make}.
786 If you encounter seemingly strange errors when trying to build the
787 compiler in a directory other than the source directory, it could be
788 because you have previously configured the compiler in the source
789 directory. Make sure you have done all the necessary preparations.
791 If you build GCC on a BSD system using a directory stored in an old System
792 V file system, problems may occur in running @code{fixincludes} if the
793 System V file system doesn't support symbolic links. These problems
794 result in a failure to fix the declaration of @code{size_t} in
795 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
796 that type mismatches occur, this could be the cause.
798 The solution is not to use such a directory for building GCC@.
800 When building from CVS or snapshots, or if you modify parser sources,
801 you need the Bison parser generator installed. Any version 1.25 or
802 later should work; older versions may also work. If you do not modify
803 parser sources, releases contain the Bison-generated files and you do
804 not need Bison installed to build them.
806 When building from CVS or snapshots, or if you modify Texinfo
807 documentation, you need version 4.0 or later of Texinfo installed if you
808 want Info documentation to be regenerated. Releases contain Info
809 documentation pre-built for the unmodified documentation in the release.
811 @section Building a native compiler
813 For a native build issue the command @samp{make bootstrap}. This
814 will build the entire GCC system, which includes the following steps:
818 Build host tools necessary to build the compiler such as texinfo, bison,
822 Build target tools for use by the compiler such as binutils (bfd,
823 binutils, gas, gprof, ld, and opcodes)
824 if they have been individually linked
825 or moved into the top level GCC source tree before configuring.
828 Perform a 3-stage bootstrap of the compiler.
831 Perform a comparison test of the stage2 and stage3 compilers.
834 Build runtime libraries using the stage3 compiler from the previous step.
838 If you are short on disk space you might consider @samp{make
839 bootstrap-lean} instead. This is identical to @samp{make
840 bootstrap} except that object files from the stage1 and
841 stage2 of the 3-stage bootstrap of the compiler are deleted as
842 soon as they are no longer needed.
845 If you want to save additional space during the bootstrap and in
846 the final installation as well, you can build the compiler binaries
847 without debugging information with @samp{make CFLAGS='-O' LIBCFLAGS='-g
848 -O2' LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap}. This will save
849 roughly 40% of disk space both for the bootstrap and the final installation.
850 (Libraries will still contain debugging information.)
852 If you wish to use non-default GCC flags when compiling the stage2 and
853 stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing
854 @samp{make bootstrap}. Non-default optimization flags are less well
855 tested here than the default of @samp{-g -O2}, but should still work.
856 In a few cases, you may find that you need to specify special flags such
857 as @option{-msoft-float} here to complete the bootstrap; or, if the
858 native compiler miscompiles the stage1 compiler, you may need to work
859 around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the
860 stage1 compiler that were miscompiled, or by using @samp{make
861 bootstrap4} to increase the number of stages of bootstrap.
863 If you used the flag @option{--enable-languages=@dots{}} to restrict
864 the compilers to be built, only those you've actually enabled will be
865 built. This will of course only build those runtime libraries, for
866 which the particular compiler has been built. Please note,
867 that re-defining @env{LANGUAGES} when calling @samp{make bootstrap}
868 @strong{does not} work anymore!
870 If the comparison of stage2 and stage3 fails, this normally indicates
871 that the stage2 compiler has compiled GCC incorrectly, and is therefore
872 a potentially serious bug which you should investigate and report. (On
873 a few systems, meaningful comparison of object files is impossible; they
874 always appear ``different''. If you encounter this problem, you will
875 need to disable comparison in the @file{Makefile}.)
877 @section Building a cross compiler
879 We recommend reading the
880 @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ}
881 for information about building cross compilers.
883 When building a cross compiler, it is not generally possible to do a
884 3-stage bootstrap of the compiler. This makes for an interesting problem
885 as parts of GCC can only be built with GCC@.
887 To build a cross compiler, we first recommend building and installing a
888 native compiler. You can then use the native GCC compiler to build the
891 Assuming you have already installed a native copy of GCC and configured
892 your cross compiler, issue the command @command{make}, which performs the
897 Build host tools necessary to build the compiler such as texinfo, bison,
901 Build target tools for use by the compiler such as binutils (bfd,
902 binutils, gas, gprof, ld, and opcodes)
903 if they have been individually linked or moved into the top level GCC source
904 tree before configuring.
907 Build the compiler (single stage only).
910 Build runtime libraries using the compiler from the previous step.
913 Note that if an error occurs in any step the make process will exit.
915 @section Building in parallel
917 If you have a multiprocessor system you can use @samp{make bootstrap
918 MAKE="make -j 2" -j 2} or just @samp{make -j 2 bootstrap}
919 for GNU Make 3.79 and above instead of just @samp{make bootstrap}
920 when building GCC@. You can use a bigger number instead of two if
921 you like. In most cases, it won't help to use a number bigger than
922 the number of processors in your machine.
929 @uref{./index.html,,Return to the GCC Installation page}
933 @c ***Testing*****************************************************************
935 @comment node-name, next, previous, up
936 @node Testing, Final install, Building, Installing GCC
940 <h1 align="center">Installing GCC: Testing</h1>
943 @chapter Installing GCC: Testing
946 @cindex Installing GCC: Testing
949 Before you install GCC, you might wish to run the testsuite. This
950 step is optional and may require you to download additional software.
952 First, you must have @uref{download.html,,downloaded the testsuites}.
953 The full distribution contains testsuites; only if you downloaded the
954 ``core'' compiler plus any front ends, you do not have the testsuites.
956 Second, you must have a @uref{http://www.gnu.org/software/dejagnu/,,current version of DejaGnu} installed;
957 dejagnu 1.3 is not sufficient.
959 Now you may need specific preparations:
964 The following environment variables may need to be set appropriately, as in
965 the following example (which assumes that DejaGnu has been installed
966 under @file{/usr/local}):
969 TCL_LIBRARY = /usr/local/share/tcl8.0
970 DEJAGNULIBS = /usr/local/share/dejagnu
973 On systems such as Cygwin, these paths are required to be actual
974 paths, not mounts or links; presumably this is due to some lack of
975 portability in the DejaGnu code.
977 If the directories where @command{runtest} and @command{expect} were
978 installed are in the @env{PATH}, it should not be necessary to set these
979 environment variables.
983 Finally, you can run the testsuite (which may take a long time):
985 cd @var{objdir}; make -k check
988 The testing process will try to test as many components in the GCC
989 distribution as possible, including the C, C++, Objective-C and Fortran
990 compilers as well as the C++ and Java runtime libraries.
992 @section How can I run the test suite on selected tests?
994 As a first possibility to cut down the number of tests that are run it is
995 possible to use @samp{make check-gcc} or @samp{make check-g++}
996 in the @file{gcc} subdirectory of the object directory. To further cut down the
997 tests the following is possible:
1000 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
1003 This will run all @command{gcc} execute tests in the testsuite.
1006 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
1009 This will run the @command{g++} ``old-deja'' tests in the testsuite where the filename
1010 matches @samp{9805*}.
1012 The @file{*.exp} files are located in the testsuite directories of the GCC
1013 source, the most important ones being @file{compile.exp},
1014 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
1015 To get a list of the possible @file{*.exp} files, pipe the
1016 output of @samp{make check} into a file and look at the
1017 @samp{Running @dots{} .exp} lines.
1019 @section How to interpret test results
1021 After the testsuite has run you'll find various @file{*.sum} and @file{*.log}
1022 files in the testsuite subdirectories. The @file{*.log} files contain a
1023 detailed log of the compiler invocations and the corresponding
1024 results, the @file{*.sum} files summarize the results. These summaries list
1025 all the tests that have been run with a corresponding status code:
1029 PASS: the test passed as expected
1031 XPASS: the test unexpectedly passed
1033 FAIL: the test unexpectedly failed
1035 XFAIL: the test failed as expected
1037 UNSUPPORTED: the test is not supported on this platform
1039 ERROR: the testsuite detected an error
1041 WARNING: the testsuite detected a possible problem
1044 It is normal for some tests to report unexpected failures. At the
1045 current time our testing harness does not allow fine grained control
1046 over whether or not a test is expected to fail. We expect to fix this
1047 problem in future releases.
1050 @section Submitting test results
1052 If you want to report the results to the GCC project, use the
1053 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
1056 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
1057 -m gcc-testresults@@gcc.gnu.org |sh
1060 This script uses the @command{Mail} program to send the results, so
1061 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
1062 prepended to the testsuite summary and should contain any special
1063 remarks you have on your results or your build environment. Please
1064 do not edit the testsuite result block or the subject line, as these
1065 messages are automatically parsed and presented at the
1066 @uref{http://gcc.gnu.org/testresults/,,GCC testresults} web
1067 page. Here you can also gather information on how specific tests
1068 behave on different platforms and compare them with your results. A
1069 few failing testcases are possible even on released versions and you
1070 should look here first if you think your results are unreasonable.
1074 @c ***Final install***********************************************************
1076 @comment node-name, next, previous, up
1077 @node Final install, , Testing, Installing GCC
1079 @ifset finalinstallhtml
1081 <h1 align="center">Installing GCC: Final installation</h1>
1084 @chapter Installing GCC: Final installation
1087 Now that GCC has been built (and optionally tested), you can install it with
1089 cd @var{objdir}; make install
1092 That step completes the installation of GCC; user level binaries can
1093 be found in @file{@var{prefix}/bin} where @var{prefix} is the value you
1094 specified with the @option{--prefix} to configure (or @file{/usr/local}
1095 by default). (If you specified @option{--bindir}, that directory will
1096 be used instead; otherwise, if you specified @option{--exec-prefix},
1097 @file{@var{exec-prefix}/bin} will be used.) Headers for the C++ and
1098 Java libraries are installed in @file{@var{prefix}/include}; libraries
1099 in @file{@var{libdir}} (normally @file{@var{prefix}/lib}); internal
1100 parts of the compiler in @file{@var{libdir}/gcc-lib}; documentation in
1101 info format in @file{@var{infodir}} (normally @file{@var{prefix}/info}).
1103 If you built a released version of GCC then if you don't mind, please
1104 quickly review the build status page for
1105 @uref{http://gcc.gnu.org/gcc-3.0/buildstat.html,,3.0} or
1106 @uref{http://gcc.gnu.org/gcc-2.95/buildstat.html,,2.95}.
1107 If your system is not listed, send a note to
1108 @email{gcc@@gcc.gnu.org} indicating
1109 that you successfully built and installed GCC.
1111 Include the output from running @file{@var{srcdir}/config.guess}. (Do
1112 not send us the @file{config.guess} file itself, just the one-line output from
1113 running it!) Also specify which version you built.
1114 If the build was for GNU/Linux, also include the distribution name and version
1115 (e.g., Red Hat 7.1 or Debian 2.2.3, available from @file{/etc/issue})
1116 and the version of glibc you used; for RPM-based systems like Red Hat,
1117 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
1118 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
1120 We'd also like to know if the
1122 @ref{Specific, host/target specific installation notes}
1125 @uref{specific.html,,host/target specific installation notes}
1127 didn't include your host/target information or if that information is
1128 incomplete or out of date. Send a note to
1129 @email{gcc@@gcc.gnu.org} telling us how the information should be changed.
1131 If you find a bug, please report it following our
1132 @uref{../bugs.html,,bug reporting guidelines}.
1134 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
1135 dvi}. You will need to have @command{texi2dvi} (version at least 4.0)
1136 and @TeX{} installed. This creates a number of @file{.dvi} files in
1137 subdirectories of @file{@var{objdir}}; these may be converted for
1138 printing with programs such as @command{dvips}. You can also
1139 @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the
1140 Free Software Foundation}, though such manuals may not be for the most
1141 recent version of GCC@.
1148 @uref{./index.html,,Return to the GCC Installation page}
1152 @c ***Binaries****************************************************************
1154 @comment node-name, next, previous, up
1155 @node Binaries, Specific, Installing GCC, Top
1159 <h1 align="center">Installing GCC: Binaries</h1>
1162 @chapter Installing GCC: Binaries
1165 @cindex Installing GCC: Binaries
1167 We are often asked about pre-compiled versions of GCC@. While we cannot
1168 provide these for all platforms, below you'll find links to binaries for
1169 various platforms where creating them by yourself is not easy due to various
1172 Please note that we did not create these binaries, nor do we
1173 support them. If you have any problems installing them, please
1174 contact their makers.
1181 @uref{http://freeware.bull.net,,Bull's Freeware and Shareware Archive for AIX};
1184 @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX};
1188 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP};
1194 @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center};
1197 @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}.
1201 @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO
1202 OpenServer/Unixware};
1205 Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware};
1208 SGI---@uref{http://freeware.sgi.com/,,SGI Freeware};
1211 Windows 95, 98, and NT:
1214 The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project;
1216 @uref{http://www.xraylith.wisc.edu/~khan/software/gnu-win32/,,GNU Win32}
1217 related projects by Mumit Khan.
1221 @uref{ftp://ftp.thewrittenword.com/packages/free/by-name/gcc-2.95.2/,,The
1222 Written Word} offers binaries for Solaris 2.5.1, 2.6, 2.7/SPARC, 2.7/Intel,
1223 IRIX 6.2, 6.5, Digital UNIX 4.0D, HP-UX 10.20, and HP-UX 11.00.
1226 Hitachi H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU
1227 Development Tools for the Hitachi H8/300[HS] Series}
1231 In addition to those specific offerings, you can get a binary
1232 distribution CD-ROM from the
1233 @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}.
1234 It contains binaries for a number of platforms, and
1235 includes not only GCC, but other stuff as well. The current CD does
1236 not contain the latest version of GCC, but it should allow
1237 bootstrapping the compiler. An updated version of that disk is in the
1245 @uref{./index.html,,Return to the GCC Installation page}
1249 @c ***Specific****************************************************************
1251 @comment node-name, next, previous, up
1252 @node Specific, Concept Index, Binaries, Top
1256 <h1 align="center">Host/target specific installation notes for GCC</h1>
1259 @chapter Host/target specific installation notes for GCC
1262 @cindex Specific installation notes
1263 @cindex Target specific installation
1264 @cindex Host specific installation
1265 @cindex Target specific installation notes
1267 Please read this document carefully @emph{before} installing the
1268 GNU Compiler Collection on your machine.
1270 Lists of successful builds for released versions of GCC are
1271 available at our web pages for
1272 @uref{http://gcc.gnu.org/gcc-3.0/buildstat.html,,3.0}
1274 @uref{http://gcc.gnu.org/gcc-2.95/buildstat.html,,2.95}.
1275 These lists are updated as new information becomes available.
1280 @uref{#1750a-*-*,,1750a-*-*}
1284 @uref{#a29k-*-bsd,,a29k-*-bsd}
1286 @uref{#alpha*-*-*,,alpha*-*-*}
1288 @uref{#alpha*-dec-osf*,,alpha*-dec-osf*}
1290 @uref{#alphaev5-cray-unicosmk*,,alphaev5-cray-unicosmk*}
1292 @uref{#arc-*-elf,,arc-*-elf}
1294 @uref{#arm-*-aout,,arm-*-aout}
1296 @uref{#arm-*-elf,,arm-*-elf}
1298 @uref{#arm*-*-linux-gnu,,arm*-*-linux-gnu}
1300 @uref{#arm-*-riscix,,arm-*-riscix}
1308 @uref{#dsp16xx,,dsp16xx}
1310 @uref{#elxsi-elxsi-bsd,,elxsi-elxsi-bsd}
1312 @uref{#*-*-freebsd*,,*-*-freebsd*}
1314 @uref{#h8300-hms,,h8300-hms}
1316 @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*}
1318 @uref{#hppa*-hp-hpux9,,hppa*-hp-hpux9}
1320 @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10}
1322 @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11}
1324 @uref{#i370-*-*,,i370-*-*}
1326 @uref{#*-*-linux-gnu,,*-*-linux-gnu}
1328 @uref{#ix86-*-linux*oldld,,i?86-*-linux*oldld}
1330 @uref{#ix86-*-linux*aout,,i?86-*-linux*aout}
1332 @uref{#ix86-*-linux*,,i?86-*-linux*}
1334 @uref{#ix86-*-sco,,i?86-*-sco}
1336 @uref{#ix86-*-sco3.2v4,,i?86-*-sco3.2v4}
1338 @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*}
1340 @uref{#ix86-*-udk,,i?86-*-udk}
1342 @uref{#ix86-*-isc,,i?86-*-isc}
1344 @uref{#ix86-*-esix,,i?86-*-esix}
1346 @uref{#ix86-ibm-aix,,i?86-ibm-aix}
1348 @uref{#ix86-sequent-bsd,,i?86-sequent-bsd}
1350 @uref{#ix86-sequent-ptx1*,,i?86-sequent-ptx1*, i?86-sequent-ptx2*}
1352 @uref{#ix86-*-sysv3*,,i?86-*-sysv3*}
1354 @uref{#i860-intel-osf*,,i860-intel-osf*}
1356 @uref{#ia64-*-linux,,ia64-*-linux}
1358 @uref{#*-lynx-lynxos,,*-lynx-lynxos}
1360 @uref{#*-ibm-aix*,,*-ibm-aix*}
1362 @uref{#m32r-*-elf,,m32r-*-elf}
1364 @uref{#m68000-hp-bsd,,m68000-hp-bsd}
1366 @uref{#m6811-elf,,m6811-elf}
1368 @uref{#m6812-elf,,m6812-elf}
1370 @uref{#m68k-altos,,m68k-altos}
1372 @uref{#m68k-apple-aux,,m68k-apple-aux}
1374 @uref{#m68k-att-sysv,,m68k-att-sysv}
1376 @uref{#m68k-bull-sysv,,m68k-bull-sysv}
1378 @uref{#m68k-crds-unox,,m68k-crds-unox}
1380 @uref{#m68k-hp-hpux,,m68k-hp-hpux}
1382 @uref{#m68k-*-nextstep*,,m68k-*-nextstep*}
1384 @uref{#m68k-ncr-*,,m68k-ncr-*}
1386 @uref{#m68k-sun,,m68k-sun}
1388 @uref{#m68k-sun-sunos4.1.1,,m68k-sun-sunos4.1.1}
1390 @uref{#m88k-*-svr3,,m88k-*-svr3}
1392 @uref{#m88k-*-dgux,,m88k-*-dgux}
1394 @uref{#m88k-tektronix-sysv3,,m88k-tektronix-sysv3}
1396 @uref{#mips-*-*,,mips-*-*}
1398 @uref{#mips-dec-*,,mips-dec-*}
1400 @uref{#mips-mips-bsd,,mips-mips-bsd}
1402 @uref{#mips-mips-riscos*,,mips-mips-riscos*}
1404 @uref{#mips-sgi-irix4,,mips-sgi-irix4}
1406 @uref{#mips-sgi-irix5,,mips-sgi-irix5}
1408 @uref{#mips-sgi-irix6,,mips-sgi-irix6}
1410 @uref{#mips-sony-sysv,,mips-sony-sysv}
1412 @uref{#ns32k-encore,,ns32k-encore}
1414 @uref{#ns32k-*-genix,,ns32k-*-genix}
1416 @uref{#ns32k-sequent,,ns32k-sequent}
1418 @uref{#ns32k-utek,,ns32k-utek}
1420 @uref{#powerpc*-*-*,,powerpc*-*-*, powerpc-*-sysv4}
1422 @uref{#powerpc-*-darwin*,,powerpc-*-darwin*}
1424 @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4}
1426 @uref{#powerpc-*-linux-gnu*,,powerpc-*-linux-gnu*}
1428 @uref{#powerpc-*-netbsd*,,powerpc-*-netbsd*}
1430 @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix}
1432 @uref{#powerpc-*-eabisim,,powerpc-*-eabisim}
1434 @uref{#powerpc-*-eabi,,powerpc-*-eabi}
1436 @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4}
1438 @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim}
1440 @uref{#powerpcle-*-eabi,,powerpcle-*-eabi}
1442 @uref{#powerpcle-*-winnt,,powerpcle-*-winnt, powerpcle-*-pe}
1444 @uref{#romp-*-aos,,romp-*-aos, romp-*-mach}
1446 @uref{#s390-*-linux*}
1448 @uref{#s390x-*-linux*}
1450 @uref{#*-*-solaris2*,,*-*-solaris2*}
1452 @uref{#sparc-sun-solaris2*,,sparc-sun-solaris2*}
1454 @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7}
1456 @uref{#*-*-solaris2.8,,*-*-solaris2.8}
1458 @uref{#sparc-sun-sunos4*,,sparc-sun-sunos4*}
1460 @uref{#sparc-unknown-linux-gnulibc1,,sparc-unknown-linux-gnulibc1}
1462 @uref{#sparc64-*-*,,sparc64-*-*}
1464 @uref{#*-*-sysv*,,*-*-sysv*}
1466 @uref{#vax-dec-ultrix,,vax-dec-ultrix}
1468 @uref{#we32k-*-*,,we32k-*-*}
1470 @uref{#windows,,Microsoft Windows}
1474 @uref{#older,,Older systems}
1479 @uref{#elf_targets,,all ELF targets} (SVR4, Solaris 2, etc.)
1485 <!-- -------- host/target specific issues start here ---------------- -->
1488 @heading @anchor{1750a-*-*}1750a-*-*
1489 MIL-STD-1750A processors.
1491 The MIL-STD-1750A cross configuration produces output for
1492 @code{as1750}, an assembler/linker available under the GNU General Public
1493 License for the 1750A@. @code{as1750} can be obtained at
1494 @uref{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}.
1495 A similarly licensed simulator for
1496 the 1750A is available from same address.
1498 You should ignore a fatal error during the building of @samp{libgcc}
1499 (@samp{libgcc} is not yet implemented for the 1750A@.)
1501 The @code{as1750} assembler requires the file @file{ms1750.inc}, which is
1502 found in the directory @file{gcc/config/1750a}.
1504 GCC produced the same sections as the Fairchild F9450 C Compiler,
1509 The program code section.
1512 The read/write (RAM) data section.
1515 The read-only (ROM) constants section.
1518 Initialization section (code to copy KREL to SREL)@.
1521 The smallest addressable unit is 16 bits (@code{BITS_PER_UNIT} is 16). This
1522 means that type @code{char} is represented with a 16-bit word per character.
1523 The 1750A's ``Load/Store Upper/Lower Byte'' instructions are not used by
1530 @heading @anchor{a29k}a29k
1531 AMD Am29k-family processors. These are normally used in embedded
1532 applications. There are no standard Unix configurations.
1534 corresponds to AMD's standard calling sequence and binary interface
1535 and is compatible with other 29k tools.
1537 You may need to make a variant of the file @file{a29k.h} for your
1538 particular configuration.
1544 @heading @anchor{a29k-*-bsd}a29k-*-bsd
1545 AMD Am29050 used in a system running a variant of BSD Unix.
1551 @heading @anchor{alpha*-*-*}alpha*-*-*
1553 This section contains general configuration information for all
1554 alpha-based platforms using ELF (in particular, ignore this section for
1555 DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this
1556 section, please read all other sections that match your target.
1558 We require binutils 2.11.2 or newer.
1559 Previous binutils releases had a number of problems with DWARF 2
1560 debugging information, not the least of which is incorrect linking of
1567 @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf*
1568 Systems using processors that implement the DEC Alpha architecture and
1569 are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq
1570 Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
1572 In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not
1573 currently (2001-06-13) work with @command{mips-tfile}. As a workaround,
1574 we need to use the old assembler, invoked via the barely documented
1575 @option{-oldas} option. To bootstrap GCC, you either need to use the
1579 % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}]
1582 or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0:
1585 % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}]
1588 As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld}
1589 are supported on Tru64 UNIX, so you must not configure GCC with
1590 @option{--with-gnu-as} or @option{--with-gnu-ld}.
1592 The @option{--enable-threads} options isn't supported yet. A patch is
1593 in preparation for a future release. The Java runtime library has been
1594 reported to work on Tru64 UNIX V4.0F, V5.0, and V5.1, so you may try
1595 @option{--enable-libgcj} and report your results.
1597 GCC writes a @samp{.verstamp} directive to the assembler output file
1598 unless it is built as a cross-compiler. It gets the version to use from
1599 the system header file @file{/usr/include/stamp.h}. If you install a
1600 new version of DEC Unix, you should rebuild GCC to pick up the new version
1603 Note that since the Alpha is a 64-bit architecture, cross-compilers from
1604 32-bit machines will not generate code as efficient as that generated
1605 when the compiler is running on a 64-bit machine because many
1606 optimizations that depend on being able to represent a word on the
1607 target in an integral value on the host cannot be performed. Building
1608 cross-compilers on the Alpha for 32-bit machines has only been tested in
1609 a few cases and may not work properly.
1611 @code{make compare} may fail on old versions of DEC Unix unless you add
1612 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
1613 assembler input file is stored in the object file, and that makes
1614 comparison fail if it differs between the @code{stage1} and
1615 @code{stage2} compilations. The option @option{-save-temps} forces a
1616 fixed name to be used for the assembler input file, instead of a
1617 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
1618 unless the comparisons fail without that option. If you add
1619 @option{-save-temps}, you will have to manually delete the @samp{.i} and
1620 @samp{.s} files after each series of compilations.
1622 GCC now supports both the native (ECOFF) debugging format used by DBX
1623 and GDB and an encapsulated STABS format for use only with GDB@. See the
1624 discussion of the @option{--with-stabs} option of @file{configure} above
1625 for more information on these formats and how to select them.
1627 There is a bug in DEC's assembler that produces incorrect line numbers
1628 for ECOFF format when the @samp{.align} directive is used. To work
1629 around this problem, GCC will not emit such alignment directives
1630 while writing ECOFF format debugging information even if optimization is
1631 being performed. Unfortunately, this has the very undesirable
1632 side-effect that code addresses when @option{-O} is specified are
1633 different depending on whether or not @option{-g} is also specified.
1635 To avoid this behavior, specify @option{-gstabs+} and use GDB instead of
1636 DBX@. DEC is now aware of this problem with the assembler and hopes to
1637 provide a fix shortly.
1643 @heading @anchor{alphaev5-cray-unicosmk*}alphaev5-cray-unicosmk*
1644 Cray T3E systems running Unicos/Mk.
1646 This port is incomplete and has many known bugs. We hope to improve the
1647 support for this target soon. Currently, only the C front end is supported,
1648 and it is not possible to build parallel applications. Cray modules are not
1649 supported; in particular, Craylibs are assumed to be in
1650 @file{/opt/ctl/craylibs/craylibs}.
1652 You absolutely @strong{must} use GNU make on this platform. Also, you
1653 need to tell GCC where to find the assembler and the linker. The
1654 simplest way to do so is by providing @option{--with-as} and
1655 @option{--with-ld} to @file{configure}, e.g.@:
1657 @samp{configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld
1658 --enable-languages=c}
1660 The comparison test during @samp{make bootstrap} fails on Unicos/Mk
1661 because the assembler inserts timestamps into object files. You should
1662 be able to work around this by doing @samp{make all} after getting this
1669 @heading @anchor{arc-*-elf}arc-*-elf
1670 Argonaut ARC processor.
1671 This configuration is intended for embedded systems.
1677 @heading @anchor{arm-*-aout}arm-*-aout
1678 Advanced RISC Machines ARM-family processors. These are often used in
1679 embedded applications. There are no standard Unix configurations.
1680 This configuration corresponds to the basic instruction sequences and will
1681 produce @file{a.out} format object modules.
1683 You may need to make a variant of the file @file{arm.h} for your particular
1690 @heading @anchor{arm-*-elf}arm-*-elf
1691 This configuration is intended for embedded systems.
1697 @heading @anchor{arm*-*-linux-gnu}arm*-*-linux-gnu
1699 We require GNU binutils 2.10 or newer.
1705 @heading @anchor{arm-*-riscix}arm-*-riscix
1706 The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix.
1707 If you are running a version of RISC iX prior to 1.2 then you must
1708 specify the version number during configuration. Note that the
1709 assembler shipped with RISC iX does not support stabs debugging
1710 information; a new version of the assembler, with stabs support
1711 included, is now available from Acorn and via ftp
1712 @uref{ftp://ftp.acorn.com/pub/riscix/as+xterm.tar.Z}. To enable stabs
1713 debugging, pass @option{--with-gnu-as} to configure.
1715 You will need to install GNU @command{sed} before you can run configure.
1721 @heading @anchor{avr}avr
1723 ATMEL AVR-family micro controllers. These are used in embedded
1724 applications. There are no standard Unix configurations.
1726 @xref{AVR Options,, AVR Options, gcc, Using and Porting the GNU Compiler
1730 See ``AVR Options'' in the main manual
1732 for the list of supported MCU types.
1734 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
1736 Further installation notes and other useful information about AVR tools
1737 can also be obtained from:
1741 @uref{http://home.overta.ru/users/denisc,,http://home.overta.ru/users/denisc}
1743 @uref{http://www.itnet.pl/amelektr/avr,,http://www.itnet.pl/amelektr/avr}
1746 We @emph{strongly} recommend using binutils 2.11 or newer.
1748 The following error:
1750 Error: register required
1753 indicates that you should upgrade to a newer version of the binutils.
1759 @heading @anchor{c4x}c4x
1761 Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal
1762 Processors. These are used in embedded applications. There are no
1763 standard Unix configurations.
1765 @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using and
1766 Porting the GNU Compiler Collection (GCC)},
1769 See ``TMS320C3x/C4x Options'' in the main manual
1771 for the list of supported MCU types.
1773 GCC can be configured as a cross compiler for both the C3x and C4x
1774 architectures on the same system. Use @samp{configure --target=c4x
1775 --enable-languages="c,c++"} to configure.
1778 Further installation notes and other useful information about C4x tools
1779 can also be obtained from:
1783 @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/}
1790 @heading @anchor{cris}CRIS
1792 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
1793 series. These are used in embedded applications.
1796 @xref{CRIS Options,, CRIS Options, gcc, Using and Porting the GNU Compiler
1800 See ``CRIS Options'' in the main manual
1802 for a list of CRIS-specific options.
1804 There are a few different CRIS targets:
1806 @item cris-axis-aout
1807 Old target. Includes a multilib for the @samp{elinux} a.out-based
1808 target. No multilibs for newer architecture variants.
1810 Mainly for monolithic embedded systems. Includes a multilib for the
1811 @samp{v10} core used in @samp{ETRAX 100 LX}.
1812 @item cris-axis-linux-gnu
1813 A GNU/Linux port for the CRIS architecture, currently targeting
1814 @samp{ETRAX 100 LX} by default.
1817 For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11
1818 or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer.
1820 Pre-packaged tools can be obtained from
1821 @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More
1822 information about this platform is available at
1823 @uref{http://developer.axis.com/}.
1829 @heading @anchor{dos}DOS
1831 Please have a look at our @uref{binaries.html,,binaries page}.
1833 You cannot install GCC by itself on MSDOS; it will not compile under
1834 any MSDOS compiler except itself. You need to get the complete
1835 compilation package DJGPP, which includes binaries as well as sources,
1836 and includes all the necessary compilation tools and libraries.
1842 @heading @anchor{dsp16xx}dsp16xx
1843 A port to the AT&T DSP1610 family of processors.
1849 @heading @anchor{*-*-freebsd*}*-*-freebsd*
1851 The version of binutils installed in @file{/usr/bin} is known to work unless
1852 otherwise specified in any per-architecture notes. However, binutils
1853 2.11 is known to improve overall testsuite results.
1855 For FreeBSD 1, FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All
1856 configuration support and files as shipped with GCC 2.95 are still in
1857 place. FreeBSD 2.2.7 has been known to bootstrap completely; however,
1858 it is unknown which version of binutils was used (it is assumed that it
1859 was the system copy in @file{/usr/bin}) and C++ EH failures were noted.
1861 For FreeBSD using the ELF file format: DWARF 2 debugging is now the
1862 default for all CPU architectures. It had been the default on
1863 FreeBSD/alpha since its inception. You may use @option{-gstabs} instead
1864 of @option{-g}, if you really want the old debugging format. There are
1865 no known issues with mixing object files and libraries with different
1866 debugging formats. Otherwise, this release of GCC should now match more
1867 of the configuration used in the stock FreeBSD configuration of GCC. In
1868 particular, @option{--enable-threads} is now configured by default.
1869 However, as a general user, do not attempt to replace the system
1870 compiler with this release. Known to bootstrap and check with good
1871 results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3 and 5-CURRENT@.
1873 At this time, @option{--enable-threads} is not compatible with
1874 @option{--enable-libgcj} on FreeBSD@.
1880 @heading @anchor{elxsi-elxsi-bsd}elxsi-elxsi-bsd
1881 The Elxsi's C compiler has known limitations that prevent it from
1882 compiling GCC@. Please contact @email{mrs@@wrs.com} for more details.
1888 @heading @anchor{h8300-hms}h8300-hms
1889 Hitachi H8/300 series of processors.
1891 Please have a look at our @uref{binaries.html,,binaries page}.
1893 The calling convention and structure layout has changed in release 2.6.
1894 All code must be recompiled. The calling convention now passes the
1895 first three arguments in function calls in registers. Structures are no
1896 longer a multiple of 2 bytes.
1902 @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux*
1904 We @emph{highly} recommend using gas/binutils 2.8 or newer on all hppa
1905 platforms; you may encounter a variety of problems when using the HP
1908 Specifically, @option{-g} does not work on HP-UX (since that system
1909 uses a peculiar debugging format which GCC does not know about), unless you
1910 use GAS and GDB and configure GCC with the
1911 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
1912 @option{--with-as=@dots{}} options.
1914 If you wish to use pa-risc 2.0 architecture support, you must use either
1915 the HP assembler, gas/binutils 2.11 or a recent
1916 @uref{ftp://sources.redhat.com/pub/binutils/snapshots,,snapshot of gas}.
1918 More specific information to @samp{hppa*-hp-hpux*} targets follows.
1924 @heading @anchor{hppa*-hp-hpux9}hppa*-hp-hpux9
1926 The HP assembler has major problems on this platform. We've tried to work
1927 around the worst of the problems. However, those workarounds may be causing
1928 linker crashes in some circumstances; the workarounds also probably prevent
1929 shared libraries from working. Use the GNU assembler to avoid these problems.
1932 The configuration scripts for GCC will also trigger a bug in the hpux9
1933 shell. To avoid this problem set @env{CONFIG_SHELL} to @file{/bin/ksh}
1934 and @env{SHELL} to @file{/bin/ksh} in your environment.
1941 @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10
1943 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
1944 @code{PHCO_19798} from HP@. HP has two sites which provide patches free of
1950 <a href="http://us-support.external.hp.com">US, Canada, Asia-Pacific, and
1954 @uref{http://us-support.external.hp.com,,}US, Canada, Asia-Pacific, and
1958 @uref{http://europe-support.external.hp.com,,Europe}
1961 The HP assembler on these systems is much better than the hpux9 assembler,
1962 but still has some problems. Most notably the assembler inserts timestamps
1963 into each object file it creates, causing the 3-stage comparison test to fail
1964 during a @samp{make bootstrap}. You should be able to continue by
1965 saying @samp{make all} after getting the failure from @samp{make
1973 @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11
1975 GCC 3.0 supports HP-UX 11. You must use GNU binutils 2.11 or above on
1982 @heading @anchor{i370-*-*}i370-*-*
1983 This port is very preliminary and has many known bugs. We hope to
1984 have a higher-quality port for this machine soon.
1990 @heading @anchor{*-*-linux-gnu}*-*-linux-gnu
1992 If you use glibc 2.2 (or 2.1.9x), GCC 2.95.2 won't install
1993 out-of-the-box. You'll get compile errors while building @samp{libstdc++}.
1994 The patch @uref{glibc-2.2.patch,,glibc-2.2.patch}, that is to be
1995 applied in the GCC source tree, fixes the compatibility problems.
2005 Currently Glibc 2.2.3 (and older releases) and GCC 3.0 are out of sync
2006 since the latest exception handling changes for GCC@. Compiling glibc
2007 with GCC 3.0 will give a binary incompatible glibc and therefore cause
2008 lots of problems and might make your system completly unusable. This
2009 will definitly need fixes in glibc but might also need fixes in GCC@. We
2010 strongly advise to wait for glibc 2.2.4 and to read the release notes of
2011 glibc 2.2.4 whether patches for GCC 3.0 are needed. You can use glibc
2012 2.2.3 with GCC 3.0, just do not try to recompile it.
2018 @heading @anchor{ix86-*-linux*oldld}i?86-*-linux*oldld
2019 Use this configuration to generate @file{a.out} binaries on Linux-based
2020 GNU systems if you do not have gas/binutils version 2.5.2 or later
2021 installed. This is an obsolete configuration.
2027 @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout
2028 Use this configuration to generate @file{a.out} binaries on Linux-based
2029 GNU systems. This configuration is being superseded. You must use
2030 gas/binutils version 2.5.2 or later.
2036 @heading @anchor{ix86-*-linux*}i?86-*-linux*
2038 You will need binutils 2.9.1.0.15 or newer for exception handling to work.
2040 If you receive Signal 11 errors when building on GNU/Linux, then it is
2041 possible you have a hardware problem. Further information on this can be
2042 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
2048 @heading @anchor{ix86-*-sco}i?86-*-sco
2049 Compilation with RCC is recommended. Also, it may be a good idea to
2050 link with GNU malloc instead of the malloc that comes with the system.
2056 @heading @anchor{ix86-*-sco3.2v4}i?86-*-sco3.2v4
2057 Use this configuration for SCO release 3.2 version 4.
2063 @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5*
2064 Use this for the SCO OpenServer Release 5 family of operating systems.
2066 Unlike earlier versions of GCC, the ability to generate COFF with this
2067 target is no longer provided.
2069 Earlier versions of GCC emitted DWARF 1 when generating ELF to allow
2070 the system debugger to be used. That support was too burdensome to
2071 maintain. GCC now emits only DWARF 2 for this target. This means you
2072 may use either the UDK debugger or GDB to debug programs built by this
2075 Use of the @option{-march=pentiumpro} flag can result in
2076 unrecognized opcodes when using the native assembler on OS versions before
2077 5.0.6. (Support for P6 opcodes was added to the native ELF assembler in
2078 that version.) While it's rather rare to see these emitted by GCC yet,
2079 errors of the basic form:
2082 /usr/tmp/ccaNlqBc.s:22:unknown instruction: fcomip
2083 /usr/tmp/ccaNlqBc.s:50:unknown instruction: fucomip
2086 are symptoms of this problem. You may work around this by not
2087 building affected files with that flag, by using the GNU assembler, or
2088 by using the assembler provided with the current version of the OS@.
2089 Users of GNU assembler should see the note below for hazards on doing
2092 The native SCO assembler that is provided with the OS at no
2093 charge is normally required. If, however, you must be able to use
2094 the GNU assembler (perhaps you're compiling code with asms that
2095 require GAS syntax) you may configure this package using the flags
2096 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You must
2097 use a recent version of GNU binutils; versions past 2.9.1 seem to work
2100 In general, the @option{--with-gnu-as} option isn't as well tested
2101 as the native assembler.
2103 Look in @file{gcc/config/i386/sco5.h} (search for ``messy'') for
2104 additional OpenServer-specific flags.
2106 Systems based on OpenServer before 5.0.4 (@samp{uname -X}
2107 will tell you what you're running) require TLS597 from
2108 @uref{ftp://ftp.sco.com/TLS/,,ftp://ftp.sco.com/TLS/}
2109 for C++ constructors and destructors to work right.
2111 The system linker in (at least) 5.0.4 and 5.0.5 will sometimes
2112 do the wrong thing for a construct that GCC will emit for PIC
2113 code. This can be seen as execution testsuite failures when using
2114 @option{-fPIC} on @file{921215-1.c}, @file{931002-1.c}, @file{nestfunc-1.c}, and @file{gcov-1.c}.
2115 For 5.0.5, an updated linker that will cure this problem is
2116 available. You must install both
2117 @uref{ftp://ftp.sco.com/Supplements/rs505a/,,ftp://ftp.sco.com/Supplements/rs505a/}
2118 and @uref{ftp://ftp.sco.com/SLS/,,OSS499A}.
2120 The dynamic linker in OpenServer 5.0.5 (earlier versions may show
2121 the same problem) aborts on certain G77-compiled programs. It's particularly
2122 likely to be triggered by building Fortran code with the @option{-fPIC} flag.
2123 Although it's conceivable that the error could be triggered by other
2124 code, only G77-compiled code has been observed to cause this abort.
2125 If you are getting core dumps immediately upon execution of your
2126 G77 program---and especially if it's compiled with @option{-fPIC}---try applying
2127 @uref{sco_osr5_g77.patch,,@file{sco_osr5_g77.patch}} to your @samp{libf2c} and
2129 Affected faults, when analyzed in a debugger, will show a stack
2130 backtrace with a fault occurring in @code{rtld()} and the program
2131 running as @file{/usr/lib/ld.so.1}. This problem has been reported to SCO
2132 engineering and will hopefully be addressed in later releases.
2139 @heading @anchor{ix86-*-udk}i?86-*-udk
2141 This target emulates the SCO Universal Development Kit and requires that
2142 package be installed. (If it is installed, you will have a
2143 @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the
2144 @samp{i?86-*-unixware7*} target
2145 but is meant to be used when hosting on a system where UDK isn't the
2146 default compiler such as OpenServer 5 or Unixware 2. This target will
2147 generate binaries that will run on OpenServer, Unixware 2, or Unixware 7,
2148 with the same warnings and caveats as the SCO UDK@.
2150 This target is a little tricky to build because we have to distinguish
2151 it from the native tools (so it gets headers, startups, and libraries
2152 from the right place) while making the tools not think we're actually
2153 building a cross compiler. The easiest way to do this is with a configure
2156 @samp{CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure
2157 --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk-}
2159 @emph{You should substitute @samp{i686} in the above command with the appropriate
2160 processor for your host.}
2162 After the usual @samp{make bootstrap} and
2163 @samp{make install}, you can then access the UDK-targeted GCC
2164 tools by adding @command{udk-} before the commonly known name. For
2165 example, to invoke the C compiler, you would use @command{udk-gcc}.
2166 They will coexist peacefully with any native-target GCC tools you may
2174 @heading @anchor{ix86-*-isc}i?86-*-isc
2175 It may be a good idea to link with GNU malloc instead of the malloc that
2176 comes with the system.
2178 In ISC version 4.1, @command{sed} core dumps when building
2179 @file{deduced.h}. Use the version of @command{sed} from version 4.0.
2185 @heading @anchor{ix86-*-esix}i?86-*-esix
2186 It may be good idea to link with GNU malloc instead of the malloc that
2187 comes with the system.
2193 @heading @anchor{ix86-ibm-aix}i?86-ibm-aix
2194 You need to use GAS version 2.1 or later, and LD from
2195 GNU binutils version 2.2 or later.
2201 @heading @anchor{ix86-sequent-bsd}i?86-sequent-bsd
2202 Go to the Berkeley universe before compiling.
2208 @heading @anchor{ix86-sequent-ptx1*}i?86-sequent-ptx1*, i?86-sequent-ptx2*
2209 You must install GNU @command{sed} before running @command{configure}.
2215 @heading @anchor{#ix86-*-sysv3*}i?86-*-sysv3*
2216 The @code{fixproto} shell script may trigger a bug in the system shell.
2217 If you encounter this problem, upgrade your operating system or
2218 use @command{bash} (the GNU shell) to run @code{fixproto}.
2225 @heading @anchor{i860-intel-osf*}i860-intel-osf*
2226 On the Intel Paragon (an i860 machine), if you are using operating
2227 system version 1.0, you will get warnings or errors about redefinition
2228 of @code{va_arg} when you build GCC@.
2230 If this happens, then you need to link most programs with the library
2231 @file{iclib.a}. You must also modify @file{stdio.h} as follows: before
2235 #if defined(__i860__) && !defined(_VA_LIST)
2236 #include <va_list.h>
2250 extern int vprintf(const char *, va_list );
2251 extern int vsprintf(char *, const char *, va_list );
2259 #endif /* __PGC__ */
2262 These problems don't exist in operating system version 1.1.
2268 @heading @anchor{ia64-*-linux}ia64-*-linux
2269 IA-64 processor (also known as IPF, or Itanium Processor Family)
2272 The toolchain is not completely finished, so requirements will continue
2274 GCC 3.0.1 and later require glibc 2.2.4.
2275 GCC 3.0.2 requires binutils from 2001-09-05 or later.
2276 GCC 3.0.1 requires binutils 2.11.1 or later.
2278 None of the following versions of GCC has an ABI that is compatible
2279 with any of the other versions in this list, with the exception that
2280 Red Hat 2.96 and Trillian 000171 are compatible with each other:
2281 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
2282 This primarily affects C++ programs and programs that create shared libraries.
2283 Because of these ABI incompatibilities, GCC 3.0.2 is not recommended for
2284 user programs on GNU/Linux systems built using earlier compiler releases.
2285 GCC 3.0.2 is recommended for compiling linux, the kernel.
2286 GCC 3.0.2 is believed to be fully ABI compliant, and hence no more major
2287 ABI changes are expected.
2293 @heading @anchor{*-lynx-lynxos}*-lynx-lynxos
2294 LynxOS 2.2 and earlier comes with GCC 1.x already installed as
2295 @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}.
2296 You can tell GCC to use the GNU assembler and linker, by specifying
2297 @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce
2298 COFF format object files and executables; otherwise GCC will use the
2299 installed tools, which produce @file{a.out} format executables.
2304 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
2306 @heading @anchor{*-ibm-aix*}*-ibm-aix*
2308 AIX Make frequently has problems with GCC makefiles. GNU Make 3.76 or
2309 newer is recommended to build on this platform.
2311 Errors involving @code{alloca} when building GCC generally are due
2312 to an incorrect definition of @code{CC} in the Makefile or mixing files
2313 compiled with the native C compiler and GCC@. During the stage1 phase of
2314 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
2315 (not @command{xlc}). Once @command{configure} has been informed of
2316 @command{xlc}, one needs to use @samp{make distclean} to remove the
2317 configure cache files and ensure that @env{CC} environment variable
2318 does not provide a definition that will confuse @command{configure}.
2319 If this error occurs during stage2 or later, then the problem most likely
2320 is the version of Make (see above).
2322 Binutils 2.10 does not support AIX 4.3. Binutils available from the
2323 @uref{http://www-1.ibm.com/servers/aix/products/aixos/linux/,,AIX
2324 Toolbox for Linux: GNU and Open Source tools for AIX};
2325 website does work. Binutils 2.11 is expected to include AIX 4.3
2326 support. The GNU Assembler is necessary for @samp{libstdc++} to build. The
2327 AIX native @command{ld} still is recommended. The native AIX tools do
2328 interoperate with GCC@.
2330 Linking executables and shared libraries may produce warnings of
2331 duplicate symbols. The assembly files generated by GCC for AIX always
2332 have included multiple symbol definitions for certain global variable
2333 and function declarations in the original program. The warnings should
2334 not prevent the linker from producing a correct library or runnable
2337 GCC's exception handling implementation stores process-specific data in
2338 the shared library which prevents exception handling from working
2339 correctly on AIX in a default installation. To work around this, the
2340 shared objects need to be loaded in the process private segment to
2341 prevent them from being shared and marked read-only. This is
2342 accomplished on AIX by installing the shared libraries
2343 (@file{libgcc_s.a} and @file{libstdc++.a}) with file permissions
2344 disallowing read-other (@samp{chmod a+x,o-r}). If the shared libraries
2345 have been used, the shared library segment can be cleaned using the
2346 @samp{/usr/sbin/slibclean} command.
2348 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
2349 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
2350 to parse archive libraries did not handle the new format correctly.
2351 These routines are used by GCC and result in error messages during
2352 linking such as ``not a COFF file''. The version of the routines shipped
2353 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
2354 option of the archive command may be used to create archives of 32-bit
2355 objects using the original ``small format''. A correct version of the
2356 routines is shipped with AIX 4.3.2 and above.
2358 Some versions of the AIX binder (linker) can fail with a relocation
2359 overflow severe error when the @option{-bbigtoc} option is used to link
2360 GCC-produced object files into an executable that overflows the TOC@. A fix
2361 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
2362 available from IBM Customer Support and from its
2363 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2364 website as PTF U455193.
2366 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
2367 with a segmentation fault when invoked by any version of GCC@. A fix for
2368 APAR IX87327 is available from IBM Customer Support and from its
2369 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2370 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
2372 The initial assembler shipped with AIX 4.3.0 generates incorrect object
2373 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
2374 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
2375 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
2376 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
2378 AIX provides National Language Support (NLS)@. Compilers and assemblers
2379 use NLS to support locale-specific representations of various data
2380 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
2381 separating decimal fractions). There have been problems reported where
2382 GCC does not produce the same floating-point formats that the assembler
2383 expects. If one encounters this problem, set the @env{LANG}
2384 environment variable to @samp{C} or @samp{En_US}.
2386 By default, GCC for AIX 4.1 and above produces code that can be used on
2387 both Power or PowerPC processors.
2389 A default can be specified with the @option{-mcpu=@var{cpu_type}}
2390 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
2396 @heading @anchor{m32r-*-elf}m32r-*-elf
2397 Mitsubishi M32R processor.
2398 This configuration is intended for embedded systems.
2404 @heading @anchor{m68000-hp-bsd}m68000-hp-bsd
2405 HP 9000 series 200 running BSD@. Note that the C compiler that comes
2406 with this system cannot compile GCC; contact @email{law@@cygnus.com}
2407 to get binaries of GCC for bootstrapping.
2413 @heading @anchor{m6811-elf}m6811-elf
2414 Motorola 68HC11 family micro controllers. These are used in embedded
2415 applications. There are no standard Unix configurations.
2421 @heading @anchor{m6812-elf}m6812-elf
2422 Motorola 68HC12 family micro controllers. These are used in embedded
2423 applications. There are no standard Unix configurations.
2429 @heading @anchor{m68k-altos}m68k-altos
2430 Altos 3068. You must use the GNU assembler, linker and debugger.
2431 Also, you must fix a kernel bug.
2437 @heading @anchor{m68k-apple-aux}m68k-apple-aux
2438 Apple Macintosh running A/UX@.
2439 You may configure GCC to use either the system assembler and
2440 linker or the GNU assembler and linker. You should use the GNU configuration
2441 if you can, especially if you also want to use G++. You enable
2442 that configuration with the @option{--with-gnu-as} and @option{--with-gnu-ld}
2443 options to @code{configure}.
2445 Note the C compiler that comes
2446 with this system cannot compile GCC@. You can find binaries of GCC
2447 for bootstrapping on @code{jagubox.gsfc.nasa.gov}.
2448 You will also a patched version of @file{/bin/ld} there that
2449 raises some of the arbitrary limits found in the original.
2455 @heading @anchor{m68k-att-sysv}m68k-att-sysv
2456 AT&T 3b1, a.k.a.@: 7300 PC@. This version of GCC cannot
2457 be compiled with the system C compiler, which is too buggy.
2458 You will need to get a previous version of GCC and use it to
2459 bootstrap. Binaries are available from the OSU-CIS archive, at
2460 @uref{ftp://archive.cis.ohio-state.edu/pub/att7300/}.
2466 @heading @anchor{m68k-bull-sysv}m68k-bull-sysv
2467 Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GCC works
2468 either with native assembler or GNU assembler. You can use
2469 GNU assembler with native COFF generation by providing @option{--with-gnu-as} to
2470 the configure script or use GNU assembler with stabs-in-COFF encapsulation
2471 by providing @samp{--with-gnu-as --stabs}. For any problem with the native
2472 assembler or for availability of the DPX/2 port of GAS, contact
2473 @email{F.Pierresteguy@@frcl.bull.fr}.
2479 @heading @anchor{m68k-crds-unox}m68k-crds-unox
2480 Use @samp{configure unos} for building on Unos.
2482 The Unos assembler is named @code{casm} instead of @code{as}. For some
2483 strange reason linking @file{/bin/as} to @file{/bin/casm} changes the
2484 behavior, and does not work. So, when installing GCC, you should
2485 install the following script as @file{as} in the subdirectory where
2486 the passes of GCC are installed:
2493 The default Unos library is named @file{libunos.a} instead of
2494 @file{libc.a}. To allow GCC to function, either change all
2495 references to @option{-lc} in @file{gcc.c} to @option{-lunos} or link
2496 @file{/lib/libc.a} to @file{/lib/libunos.a}.
2498 @cindex @code{alloca}, for Unos
2499 When compiling GCC with the standard compiler, to overcome bugs in
2500 the support of @code{alloca}, do not use @option{-O} when making stage 2.
2501 Then use the stage 2 compiler with @option{-O} to make the stage 3
2502 compiler. This compiler will have the same characteristics as the usual
2503 stage 2 compiler on other systems. Use it to make a stage 4 compiler
2504 and compare that with stage 3 to verify proper compilation.
2506 (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in
2507 the comments there will make the above paragraph superfluous. Please
2508 inform us of whether this works.)
2510 Unos uses memory segmentation instead of demand paging, so you will need
2511 a lot of memory. 5 Mb is barely enough if no other tasks are running.
2512 If linking @file{cc1} fails, try putting the object files into a library
2513 and linking from that library.
2519 @heading @anchor{m68k-hp-hpux}m68k-hp-hpux
2520 HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in
2521 the assembler that prevents compilation of GCC@. This
2522 bug manifests itself during the first stage of compilation, while
2523 building @file{libgcc2.a}:
2527 cc1: warning: `-g' option not supported on this version of GCC
2528 cc1: warning: `-g1' option not supported on this version of GCC
2529 ./xgcc: Internal compiler error: program as got fatal signal 11
2532 A patched version of the assembler is available as the file
2533 @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you
2534 have HP software support, the patch can also be obtained directly from
2535 HP, as described in the following note:
2538 This is the patched assembler, to patch SR#1653-010439, where the
2539 assembler aborts on floating point constants.
2541 The bug is not really in the assembler, but in the shared library
2542 version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is
2543 SR#4701-078451. Anyway, the attached assembler uses the archive
2544 library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
2547 This patch is also known as PHCO_4484.
2549 In addition, if you wish to use gas, you must use
2550 gas version 2.1 or later, and you must use the GNU linker version 2.1 or
2551 later. Earlier versions of gas relied upon a program which converted the
2552 gas output into the native HP-UX format, but that program has not been
2553 kept up to date. gdb does not understand that native HP-UX format, so
2554 you must use gas if you wish to use gdb.
2556 On HP-UX version 8.05, but not on 8.07 or more recent versions, the
2557 @code{fixproto} shell script triggers a bug in the system shell. If you
2558 encounter this problem, upgrade your operating system or use BASH (the
2559 GNU shell) to run @code{fixproto}. This bug will cause the fixproto
2560 program to report an error of the form:
2563 ./fixproto: sh internal 1K buffer overflow
2566 To fix this, you can also change the first line of the fixproto script
2578 @heading @anchor{m68k-*-nextstep*}m68k-*-nextstep*
2580 Current GCC versions probably do not work on version 2 of the NeXT
2583 On NeXTStep 3.0, the Objective-C compiler does not work, due,
2584 apparently, to a kernel bug that it happens to trigger. This problem
2585 does not happen on 3.1.
2587 You absolutely @strong{must} use GNU sed and GNU make on this platform.
2590 On NeXTSTEP 3.x where x < 3 the build of GCC will abort during
2591 stage1 with an error message like this:
2595 /usr/tmp/ccbbsZ0U.s:987:Unknown pseudo-op: .section
2596 /usr/tmp/ccbbsZ0U.s:987:Rest of line ignored. 1st junk character
2600 The reason for this is the fact that NeXT's assembler for these
2601 versions of the operating system does not support the @samp{.section}
2602 pseudo op that's needed for full C++ exception functionality.
2604 As NeXT's assembler is a derived work from GNU as, a free
2605 replacement that does can be obtained at
2606 @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}.
2608 If you try to build the integrated C++ & C++ runtime libraries on this system
2609 you will run into trouble with include files. The way to get around this is
2610 to use the following sequence. Note you must have write permission to
2611 the directory @var{prefix} you specified in the configuration process of GCC
2612 for this sequence to work.
2616 make all-texinfo all-bison all-byacc all-binutils all-gas all-ld
2619 make install-headers-tar
2628 @heading @anchor{m68k-ncr-*}m68k-ncr-*
2629 On the Tower models 4@var{n}0 and 6@var{n}0, by default a process is not
2630 allowed to have more than one megabyte of memory. GCC cannot compile
2631 itself (or many other programs) with @option{-O} in that much memory.
2633 To solve this problem, reconfigure the kernel adding the following line
2634 to the configuration file:
2645 @heading @anchor{m68k-sun}m68k-sun
2646 Sun 3. We do not provide a configuration file to use the Sun FPA by
2647 default, because programs that establish signal handlers for floating
2648 point traps inherently cannot work with the FPA@.
2654 @heading @anchor{m68k-sun-sunos4.1.1}m68k-sun-sunos4.1.1
2656 It is reported that you may need the GNU assembler on this platform.
2663 @heading @anchor{m88k-*-svr3}m88k-*-svr3
2664 Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port.
2665 These systems tend to use the Green Hills C, revision 1.8.5, as the
2666 standard C compiler. There are apparently bugs in this compiler that
2667 result in object files differences between stage 2 and stage 3. If this
2668 happens, make the stage 4 compiler and compare it to the stage 3
2669 compiler. If the stage 3 and stage 4 object files are identical, this
2670 suggests you encountered a problem with the standard C compiler; the
2671 stage 3 and 4 compilers may be usable.
2673 It is best, however, to use an older version of GCC for bootstrapping
2680 @heading @anchor{m88k-*-dgux}m88k-*-dgux
2681 Motorola m88k running DG/UX@. To build 88open BCS native or cross
2682 compilers on DG/UX, specify the configuration name as
2683 @samp{m88k-*-dguxbcs} and build in the 88open BCS software development
2684 environment. To build ELF native or cross compilers on DG/UX, specify
2685 @samp{m88k-*-dgux} and build in the DG/UX ELF development environment.
2686 You set the software development environment by issuing
2687 @samp{sde-target} command and specifying either @samp{m88kbcs} or
2688 @samp{m88kdguxelf} as the operand.
2690 If you do not specify a configuration name, @file{configure} guesses the
2691 configuration based on the current software development environment.
2697 @heading @anchor{m88k-tektronix-sysv3}m88k-tektronix-sysv3
2698 Tektronix XD88 running UTekV 3.2e. Do not turn on
2699 optimization while building stage1 if you bootstrap with
2700 the buggy Green Hills compiler. Also, the bundled LAI
2701 System V NFS is buggy so if you build in an NFS mounted
2702 directory, start from a fresh reboot, or avoid NFS all together.
2703 Otherwise you may have trouble getting clean comparisons
2710 @heading @anchor{mips-*-*}mips-*-*
2711 If you use the 1.31 version of the MIPS assembler (such as was shipped
2712 with Ultrix 3.1), you will need to use the @option{-fno-delayed-branch} switch
2713 when optimizing floating point code. Otherwise, the assembler will
2714 complain when the GCC compiler fills a branch delay slot with a
2715 floating point instruction, such as @code{add.d}.
2717 If on a MIPS system you get an error message saying ``does not have gp
2718 sections for all it's [sic] sectons [sic]'', don't worry about it. This
2719 happens whenever you use GAS with the MIPS linker, but there is not
2720 really anything wrong, and it is okay to use the output file. You can
2721 stop such warnings by installing the GNU linker.
2723 It would be nice to extend GAS to produce the gp tables, but they are
2724 optional, and there should not be a warning about their absence.
2726 Users have reported some problems with version 2.0 of the MIPS
2727 compiler tools that were shipped with Ultrix 4.1. Version 2.10
2728 which came with Ultrix 4.2 seems to work fine.
2730 Users have also reported some problems with version 2.20 of the
2731 MIPS compiler tools that were shipped with RISC/os 4.x. The earlier
2732 version 2.11 seems to work fine.
2734 Some versions of the MIPS linker will issue an assertion failure
2735 when linking code that uses @code{alloca} against shared
2736 libraries on RISC-OS 5.0, and DEC's OSF/1 systems. This is a bug
2737 in the linker, that is supposed to be fixed in future revisions.
2738 To protect against this, GCC passes @option{-non_shared} to the
2739 linker unless you pass an explicit @option{-shared} or
2740 @option{-call_shared} switch.
2742 @heading @anchor{mips-mips-bsd}mips-mips-bsd
2743 MIPS machines running the MIPS operating system in BSD mode. It's
2744 possible that some old versions of the system lack the functions
2745 @code{memcpy}, @code{memmove}, @code{memcmp}, and @code{memset}. If your
2746 system lacks these, you must remove or undo the definition of
2747 @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}.
2749 If you use the MIPS C compiler to bootstrap, it may be necessary
2750 to increase its table size for switch statements with the
2751 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2752 optimization option, you also need to use @option{-Olimit 3000}.
2753 Both of these options are automatically generated in the
2754 @file{Makefile} that the shell script @file{configure} builds.
2755 If you override the @code{CC} make variable and use the MIPS
2756 compilers, you may need to add @option{-Wf,-XNg1500 -Olimit 3000}.
2762 @heading @anchor{mips-dec-*}mips-dec-*
2763 MIPS-based DECstations can support three different personalities:
2764 Ultrix, DEC OSF/1, and OSF/rose. (Alpha-based DECstation products have
2765 a configuration name beginning with @samp{alpha*-dec}.) To configure GCC
2766 for these platforms use the following configurations:
2769 @item mips-dec-ultrix
2770 Ultrix configuration.
2773 DEC's version of OSF/1.
2775 @item mips-dec-osfrose
2776 Open Software Foundation reference port of OSF/1 which uses the
2777 OSF/rose object file format instead of ECOFF@. Normally, you
2778 would not select this configuration.
2781 If you use the MIPS C compiler to bootstrap, it may be necessary
2782 to increase its table size for switch statements with the
2783 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2784 optimization option, you also need to use @option{-Olimit 3000}.
2785 Both of these options are automatically generated in the
2786 @file{Makefile} that the shell script @file{configure} builds.
2787 If you override the @code{CC} make variable and use the MIPS
2788 compilers, you may need to add @option{-Wf,-XNg1500 -Olimit 3000}.
2794 @heading @anchor{mips-mips-riscos*}mips-mips-riscos*
2795 If you use the MIPS C compiler to bootstrap, it may be necessary
2796 to increase its table size for switch statements with the
2797 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2798 optimization option, you also need to use @option{-Olimit 3000}.
2799 Both of these options are automatically generated in the
2800 @file{Makefile} that the shell script @file{configure} builds.
2801 If you override the @code{CC} make variable and use the MIPS
2802 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
2804 MIPS computers running RISC-OS can support four different
2805 personalities: default, BSD 4.3, System V.3, and System V.4
2806 (older versions of RISC-OS don't support V.4). To configure GCC
2807 for these platforms use the following configurations:
2810 @item mips-mips-riscos@var{rev}
2811 Default configuration for RISC-OS, revision @var{rev}.
2813 @item mips-mips-riscos@var{rev}bsd
2814 BSD 4.3 configuration for RISC-OS, revision @var{rev}.
2816 @item mips-mips-riscos@var{rev}sysv4
2817 System V.4 configuration for RISC-OS, revision @var{rev}.
2823 @item mips-mips-riscos@var{rev}sysv
2824 System V.3 configuration for RISC-OS, revision @var{rev}.
2827 The revision @code{rev} mentioned above is the revision of
2828 RISC-OS to use. You must reconfigure GCC when going from a
2829 RISC-OS revision 4 to RISC-OS revision 5. This has the effect of
2830 avoiding a linker bug.
2836 @heading @anchor{mips-sgi-irix4}mips-sgi-irix4
2838 In order to compile GCC on an SGI running IRIX 4, the ``c.hdr.lib''
2839 option must be installed from the CD-ROM supplied from Silicon Graphics.
2840 This is found on the 2nd CD in release 4.0.1.
2842 On IRIX version 4.0.5F, and perhaps on some other versions as well,
2843 there is an assembler bug that reorders instructions incorrectly. To
2844 work around it, specify the target configuration
2845 @samp{mips-sgi-irix4loser}. This configuration inhibits assembler
2848 In a compiler configured with target @samp{mips-sgi-irix4}, you can turn
2849 off assembler optimization by using the @option{-noasmopt} option. This
2850 compiler option passes the option @option{-O0} to the assembler, to
2853 The @option{-noasmopt} option can be useful for testing whether a problem
2854 is due to erroneous assembler reordering. Even if a problem does not go
2855 away with @option{-noasmopt}, it may still be due to assembler
2856 reordering---perhaps GCC itself was miscompiled as a result.
2858 You may get the following warning on IRIX 4 platforms, it can be safely
2861 warning: foo.o does not have gp tables for all its sections.
2868 @heading @anchor{mips-sgi-irix5}mips-sgi-irix5
2870 This configuration has considerable problems, which will be fixed in a
2873 In order to compile GCC on an SGI running IRIX 5, the ``compiler_dev.hdr''
2874 subsystem must be installed from the IDO CD-ROM supplied by Silicon
2875 Graphics. It is also available for download from
2876 @uref{http://www.sgi.com/developers/devtools/apis/ido.html,,http://www.sgi.com/developers/devtools/apis/ido.html}.
2878 @code{make compare} may fail on version 5 of IRIX unless you add
2879 @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the
2880 assembler input file is stored in the object file, and that makes
2881 comparison fail if it differs between the @code{stage1} and
2882 @code{stage2} compilations. The option @option{-save-temps} forces a
2883 fixed name to be used for the assembler input file, instead of a
2884 randomly chosen name in @file{/tmp}. Do not add @option{-save-temps}
2885 unless the comparisons fail without that option. If you do you
2886 @option{-save-temps}, you will have to manually delete the @samp{.i} and
2887 @samp{.s} files after each series of compilations.
2889 If you use the MIPS C compiler to bootstrap, it may be necessary
2890 to increase its table size for switch statements with the
2891 @option{-Wf,-XNg1500} option. If you use the @option{-O2}
2892 optimization option, you also need to use @option{-Olimit 3000}.
2894 To enable debugging under IRIX 5, you must use GNU @command{as} 2.11.2
2896 and use the @option{--with-gnu-as} configure option when configuring GCC.
2897 GNU @command{as} is distributed as part of the binutils package.
2898 When using release 2.11.2, you need to apply a patch
2899 @uref{http://sources.redhat.com/ml/binutils/2001-07/msg00352.html,,http://sources.redhat.com/ml/binutils/2001-07/msg00352.html}
2900 which will be included in the next release of binutils.
2902 When building GCC, the build process loops rebuilding @command{cc1} over
2903 and over again. This happens on @samp{mips-sgi-irix5.2}, and possibly
2904 other platforms. It has been reported that this is a known bug in the
2905 @command{make} shipped with IRIX 5.2. We recommend you use GNU
2906 @command{make} instead of the vendor supplied @command{make} program;
2907 however, you may have success with @command{smake} on IRIX 5.2 if you do
2908 not have GNU @command{make} available.
2914 @heading @anchor{mips-sgi-irix6}mips-sgi-irix6
2916 If you are using IRIX @command{cc} as your bootstrap compiler, you must
2917 ensure that the N32 ABI is in use. To test this, compile a simple C
2918 file with @command{cc} and then run @command{file} on the
2919 resulting object file. The output should look like:
2922 test.o: ELF N32 MSB @dots{}
2928 test.o: ELF 32-bit MSB @dots{}
2934 test.o: ELF 64-bit MSB @dots{}
2937 then your version of @command{cc} uses the O32 or N64 ABI by default. You
2938 should set the environment variable @env{CC} to @samp{cc -n32}
2939 before configuring GCC@.
2941 GCC on IRIX 6 is usually built to support both the N32 and N64 ABIs. If
2942 you build GCC on a system that doesn't have the N64 libraries installed,
2943 you need to configure with @option{--disable-multilib} so GCC doesn't
2944 try to use them. Look for @file{/usr/lib64/libc.so.1} to see if you
2945 have the 64-bit libraries installed.
2947 You must @emph{not} use GNU @command{as} (which isn't built anyway as of
2948 binutils 2.11.2) on IRIX 6 platforms; doing so will only cause problems.
2950 GCC does not currently support generating O32 ABI binaries in the
2951 @samp{mips-sgi-irix6} configurations. It is possible to create a GCC
2952 with O32 ABI only support by configuring it for the @samp{mips-sgi-irix5}
2953 target and using a patched GNU @command{as} 2.11.2 as documented in the
2954 @uref{#mips-sgi-irix5,,@samp{mips-sgi-irix5}} section above. Using the
2955 native assembler requires patches to GCC which will be included in a
2956 future release. It is
2957 expected that O32 ABI support will be available again in a future release.
2959 The @option{--enable-threads} option doesn't currently work, a patch is
2960 in preparation for a future release. The @option{--enable-libgcj}
2961 option is disabled by default: IRIX 6 uses a very low default limit
2962 (20480) for the command line length. Although libtool contains a
2963 workaround for this problem, at least the N64 @samp{libgcj} is known not
2964 to build despite this, running into an internal error of the native
2965 @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to
2966 its maximum of 262144 bytes. If you have root access, you can use the
2967 @command{systune} command to do this.
2969 GCC does not correctly pass/return structures which are
2970 smaller than 16 bytes and which are not 8 bytes. The problem is very
2971 involved and difficult to fix. It affects a number of other targets also,
2972 but IRIX 6 is affected the most, because it is a 64 bit target, and 4 byte
2973 structures are common. The exact problem is that structures are being padded
2974 at the wrong end, e.g.@: a 4 byte structure is loaded into the lower 4 bytes
2975 of the register when it should be loaded into the upper 4 bytes of the
2978 GCC is consistent with itself, but not consistent with the SGI C compiler
2979 (and the SGI supplied runtime libraries), so the only failures that can
2980 happen are when there are library functions that take/return such
2981 structures. There are very few such library functions. Currently this
2982 is known to affect @code{inet_ntoa}, @code{inet_lnaof},
2983 @code{inet_netof}, @code{inet_makeaddr}, and @code{semctl}. Until the
2984 bug is fixed, GCC contains workarounds for the known affected functions.
2986 See @uref{http://freeware.sgi.com/,,http://freeware.sgi.com/} for more
2987 information about using GCC on IRIX platforms.
2993 @heading @anchor{mips-sony-sysv}mips-sony-sysv
2994 Sony MIPS NEWS@. This works in NEWSOS 5.0.1, but not in 5.0.2 (which
2995 uses ELF instead of COFF)@. Support for 5.0.2 will probably be provided
2996 soon by volunteers. In particular, the linker does not like the
2997 code generated by GCC when shared libraries are linked in.
3004 @heading @anchor{ns32k-encore}ns32k-encore
3005 Encore ns32000 system. Encore systems are supported only under BSD@.
3011 @heading @anchor{ns32k-*-genix}ns32k-*-genix
3012 National Semiconductor ns32000 system. Genix has bugs in @code{alloca}
3013 and @code{malloc}; you must get the compiled versions of these from GNU
3020 @heading @anchor{ns32k-sequent}ns32k-sequent
3021 Go to the Berkeley universe before compiling.
3027 @heading @anchor{ns32k-utek}ns32k-utek
3028 UTEK ns32000 system (``merlin''). The C compiler that comes with this
3029 system cannot compile GCC; contact @samp{tektronix!reed!mason} to get
3030 binaries of GCC for bootstrapping.
3037 @heading @anchor{powerpc*-*-*}powerpc-*-*
3039 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
3040 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
3046 @heading @anchor{powerpc-*-darwin*}powerpc-*-darwin*
3047 PowerPC running Darwin (Mac OS X kernel).
3049 GCC 3.0 does not support Darwin, but 3.1 and later releases will work.
3051 Pre-installed versions of Mac OS X may not include any developer tools,
3052 meaning that you will not be able to build GCC from source. Tool
3053 binaries are available at
3054 @uref{http://www.opensource.apple.com/projects/darwin} (free
3055 registration required).
3057 Versions of the assembler prior to ``cctools-364'' cannot handle the
3058 4-argument form of rlwinm and related mask-using instructions. Darwin
3059 1.3 (Mac OS X 10.0) uses cctools-353 for instance. To get cctools-364,
3060 check out @file{cctools} with tag @samp{Apple-364}, build it, and
3061 install the assembler as @file{usr/bin/as}. See
3062 @uref{http://www.opensource.apple.com/tools/cvs/docs.html} for details.
3064 Also, the default stack limit of 512K is too small, and a bootstrap will
3065 typically fail when self-compiling @file{expr.c}. Set the stack to 800K
3066 or more, for instance by doing @samp{limit stack 800}. It's also
3067 convenient to use the GNU preprocessor instead of Apple's during the
3068 first stage of bootstrapping; this is automatic when doing @samp{make
3069 bootstrap}, but to do it from the toplevel objdir you will need to say
3070 @samp{make CC='cc -no-cpp-precomp' bootstrap}.
3072 Note that the version of GCC shipped by Apple typically includes a
3073 number of extensions not available in a standard GCC release. These
3074 extensions are generally specific to Mac programming.
3080 @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4
3081 PowerPC system in big endian mode, running System V.4.
3087 @heading @anchor{powerpc-*-linux-gnu*}powerpc-*-linux-gnu*
3090 @uref{ftp://ftp.varesearch.com/pub/support/hjl/binutils,,binutils 2.9.4.0.8}
3091 or newer for a working GCC@. It is strongly recommended to recompile binutils
3092 if you initially built it with gcc-2.7.2.x.
3098 @heading @anchor{powerpc-*-netbsd*}powerpc-*-netbsd*
3099 PowerPC system in big endian mode running NetBSD@. To build the
3100 documentation you will need Texinfo version 4.0 (NetBSD 1.5.1 included
3101 Texinfo version 3.12).
3107 @heading @anchor{powerpc-*-eabiaix}powerpc-*-eabiaix
3108 Embedded PowerPC system in big endian mode with @option{-mcall-aix} selected as
3115 @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim
3116 Embedded PowerPC system in big endian mode for use in running under the
3123 @heading @anchor{powerpc-*-eabi}powerpc-*-eabi
3124 Embedded PowerPC system in big endian mode.
3130 @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4
3131 PowerPC system in little endian mode, running System V.4.
3137 @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim
3138 Embedded PowerPC system in little endian mode for use in running under
3145 @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi
3146 Embedded PowerPC system in little endian mode.
3152 @heading @anchor{powerpcle-*-winnt}powerpcle-*-winnt, powerpcle-*-pe
3153 PowerPC system in little endian mode running Windows NT@.
3159 @heading @anchor{romp-*-aos}romp-*-aos, romp-*-mach
3160 The only operating systems supported for the IBM RT PC are AOS and
3161 MACH@. GCC does not support AIX running on the RT@. We recommend you
3162 compile GCC with an earlier version of itself; if you compile GCC
3163 with @command{hc}, the Metaware compiler, it will work, but you will get
3164 mismatches between the stage 2 and stage 3 compilers in various files.
3165 These errors are minor differences in some floating-point constants and
3166 can be safely ignored; the stage 3 compiler is correct.
3172 @heading @anchor{s390-*-linux*}s390-*-linux*
3173 S/390 system running Linux for S/390@.
3179 @heading @anchor{s390x-*-linux*}s390x-*-linux*
3180 zSeries system (64 Bit) running Linux for zSeries@.
3186 @c Please use Solaris 2 to refer to all release of Solaris, starting
3187 @c with 2.0 until 2.6, 7, and 8. Solaris 1 was a marketing name for
3188 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
3189 @c alone is too unspecific and must be avoided.
3190 @heading @anchor{*-*-solaris2*}*-*-solaris2*
3192 Sun does not ship a C compiler with Solaris 2. To bootstrap and install
3193 GCC you first have to install a pre-built compiler, see our
3194 @uref{binaries.html,,binaries page} for details.
3196 The Solaris 2 @command{/bin/sh} will often fail to configure
3197 @file{libstdc++-v3}, @file{boehm-gc} or
3198 @file{libjava}. If you encounter this problem, set @env{CONFIG_SHELL} to
3199 @command{/bin/ksh} in your environment and run @command{make bootstrap} again.
3200 Another possibility that sometimes helps is to remove
3201 @file{*-*-solaris2*/config.cache}.
3203 Solaris 2 comes with a number of optional OS packages. Some of these
3204 packages are needed to use GCC fully, namely @code{SUNWarc},
3205 @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm},
3206 @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all
3207 optional packages when installing Solaris 2, you will need to verify that
3208 the packages that GCC needs are installed.
3210 To check whether an optional package is installed, use
3211 the @command{pkginfo} command. To add an optional package, use the
3212 @command{pkgadd} command. For further details, see the Solaris 2
3215 Trying to use the linker and other tools in
3216 @file{/usr/ucb} to install GCC has been observed to cause trouble.
3217 For example, the linker may hang indefinitely. The fix is to remove
3218 @file{/usr/ucb} from your @env{PATH}.
3220 All releases of GNU binutils prior to 2.11.2 have known bugs on this
3221 platform. We recommend the use of GNU binutils 2.11.2 or the vendor
3222 tools (Sun @command{as}, Sun @command{ld}).
3224 Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
3225 newer: @command{g++} will complain that types are missing. These headers assume
3226 that omitting the type means @code{int}; this assumption worked for C89 but
3227 is wrong for C++, and is now wrong for C99 also.
3229 @command{g++} accepts such (illegal) constructs with the option
3230 @option{-fpermissive}; it
3231 will assume that any missing type is @code{int} (as defined by C89).
3233 There are patches for Solaris 2.6 (105633-56 or newer for SPARC,
3234 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC,
3235 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
3236 108653-22 for Intel) that fix this bug.
3242 @heading @anchor{sparc-sun-solaris2*}sparc-sun-solaris2*
3244 Sun @command{as} 4.x is broken in that it cannot cope with long symbol names.
3245 A typical error message might look similar to the following:
3247 @samp{/usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041:
3248 error: can't compute value of an expression involving an external symbol.}
3250 This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris
3251 2.6 and has been fixed in later (5.x) versions of the assembler,
3252 starting with Solaris 7.
3258 @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7
3260 Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
3261 the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
3262 and later, including all EGCS releases. Sun formerly recommended
3263 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
3264 recommend it only for people who use Sun's compilers.
3266 Here are some workarounds to this problem:
3269 Do not install Sun patch 107058-01 until after Sun releases a
3270 complete patch for bug 4210064. This is the simplest course to take,
3271 unless you must also use Sun's C compiler. Unfortunately 107058-01
3272 is preinstalled on some new Solaris 7-based hosts, so you may have to
3276 Copy the original, unpatched Solaris 7
3277 @command{/usr/ccs/bin/as} into
3278 @command{/usr/local/lib/gcc-lib/sparc-sun-solaris2.7/3.0/as},
3279 adjusting the latter name to fit your local conventions and software
3283 Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
3284 both 107058-01 and 106950-03 installed has reported the bug with GCC
3285 and Sun's dynamic linker. This last course of action is riskiest,
3286 for two reasons. First, you must install 106950 on all hosts that
3287 run code generated by GCC; it doesn't suffice to install it only on
3288 the hosts that run GCC itself. Second, Sun says that 106950-03 is
3289 only a partial fix for bug 4210064, but Sun doesn't know whether the
3290 partial fix is adequate for GCC@. Revision -08 or later should fix
3291 the bug. The current (as of 2001-09-24) revision is -14, and is included in
3292 the Solaris 7 Recommended Patch Cluster.
3300 @heading @anchor{*-*-solaris2.8}*-*-solaris2.8
3302 The Solaris 8 linker fails to link some @samp{libjava} programs if
3303 previously-installed GCC java libraries already exist in the configured
3304 prefix. For this reason, @samp{libgcj} is disabled by default on Solaris 8.
3305 If you use GNU @command{ld}, or if you don't have a previously-installed @samp{libgcj} in
3306 the same prefix, use @option{--enable-libgcj} to build and install the
3313 @heading @anchor{sparc-sun-sunos4*}sparc-sun-sunos4*
3315 A bug in the SunOS 4 linker will cause it to crash when linking
3316 @option{-fPIC} compiled objects (and will therefore not allow you to build
3319 To fix this problem you can either use the most recent version of
3320 binutils or get the latest SunOS 4 linker patch (patch ID 100170-10)
3321 from Sun's patch site.
3323 Sometimes on a Sun 4 you may observe a crash in the program
3324 @command{genflags} or @command{genoutput} while building GCC. This is said to
3325 be due to a bug in @command{sh}. You can probably get around it by running
3326 @command{genflags} or @command{genoutput} manually and then retrying the
3333 @heading @anchor{sparc-unknown-linux-gnulibc1}sparc-unknown-linux-gnulibc1
3335 It has been reported that you might need
3336 @uref{ftp://ftp.yggdrasil.com/private/hjl,,binutils 2.8.1.0.23}
3337 for this platform, too.
3344 @heading @anchor{sparc64-*-*}sparc64-*-*
3346 GCC version 2.95 is not able to compile code correctly for
3347 @code{sparc64} targets. Users of the Linux kernel, at least,
3348 can use the @code{sparc32} program to start up a new shell
3349 invocation with an environment that causes @command{configure} to
3350 recognize (via @samp{uname -a}) the system as @samp{sparc-*-*} instead.
3357 @heading @anchor{#*-*-sysv*}*-*-sysv*
3358 On System V release 3, you may get this error message
3362 ld fatal: failed to write symbol name @var{something}
3363 in strings table for file @var{whatever}
3366 This probably indicates that the disk is full or your ulimit won't allow
3367 the file to be as large as it needs to be.
3369 This problem can also result because the kernel parameter @code{MAXUMEM}
3370 is too small. If so, you must regenerate the kernel and make the value
3371 much larger. The default value is reported to be 1024; a value of 32768
3372 is said to work. Smaller values may also work.
3374 On System V, if you get an error like this,
3377 /usr/local/lib/bison.simple: In function `yyparse':
3378 /usr/local/lib/bison.simple:625: virtual memory exhausted
3382 that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}.
3384 On a System V release 4 system, make sure @file{/usr/bin} precedes
3385 @file{/usr/ucb} in @code{PATH}. The @code{cc} command in
3386 @file{/usr/ucb} uses libraries which have bugs.
3392 @heading @anchor{vax-dec-ultrix}vax-dec-ultrix
3393 Don't try compiling with VAX C (@code{vcc}). It produces incorrect code
3394 in some cases (for example, when @code{alloca} is used).
3400 @heading @anchor{we32k-*-*}we32k-*-*
3401 These computers are also known as the 3b2, 3b5, 3b20 and other similar
3402 names. (However, the 3b1 is actually a 68000.)
3404 Don't use @option{-g} when compiling with the system's compiler. The
3405 system's linker seems to be unable to handle such a large program with
3406 debugging information.
3408 The system's compiler runs out of capacity when compiling @file{stmt.c}
3409 in GCC@. You can work around this by building @file{cpp} in GCC
3410 first, then use that instead of the system's preprocessor with the
3411 system's C compiler to compile @file{stmt.c}. Here is how:
3414 mv /lib/cpp /lib/cpp.att
3416 echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp
3420 The system's compiler produces bad code for some of the GCC
3421 optimization files. So you must build the stage 2 compiler without
3422 optimization. Then build a stage 3 compiler with optimization.
3423 That executable should work. Here are the necessary commands:
3426 make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g"
3428 make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O"
3431 You may need to raise the ULIMIT setting to build a C++ compiler,
3432 as the file @file{cc1plus} is larger than one megabyte.
3438 @heading @anchor{windows}Microsoft Windows (32 bit)
3440 A port of GCC 2.95.x is included with the
3441 @uref{http://www.cygwin.com/,,Cygwin environment}.
3443 Current (as of early 2001) snapshots of GCC will build under Cygwin
3444 without modification.
3450 @heading @anchor{os2}OS/2
3452 GCC does not currently support OS/2. However, Andrew Zabolotny has been
3453 working on a generic OS/2 port with pgcc. The current code code can be found
3454 at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}.
3456 An older copy of GCC 2.8.1 is included with the EMX tools available at
3457 @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,,
3458 ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}.
3464 @heading @anchor{older}Older systems
3466 GCC contains support files for many older (1980s and early
3467 1990s) Unix variants. For the most part, support for these systems
3468 has not been deliberately removed, but it has not been maintained for
3469 several years and may suffer from bitrot. Support from some systems
3470 has been removed from GCC 3: fx80, ns32-ns-genix, pyramid, tahoe,
3471 gmicro, spur; most of these targets had not been updated since GCC
3474 Support for older systems as targets for cross-compilation is less
3475 problematic than support for them as hosts for GCC; if an enthusiast
3476 wishes to make such a target work again (including resurrecting any
3477 of the targets that never worked with GCC 2, starting from the last
3478 CVS version before they were removed), patches
3479 @uref{../contribute.html,,following the usual requirements}
3480 would be likely to be accepted, since they should not affect the
3481 support for more modern targets.
3483 Support for old systems as hosts for GCC can cause problems if the
3484 workarounds for compiler, library and operating system bugs affect the
3485 cleanliness or maintainability of the rest of GCC@. In some cases, to
3486 bring GCC up on such a system, if still possible with current GCC, may
3487 require first installing an old version of GCC which did work on that
3488 system, and using it to compile a more recent GCC, to avoid bugs in
3489 the vendor compiler. Old releases of GCC 1 and GCC 2 are available in
3490 the @file{old-releases} directory on the
3491 @uref{../mirrors.html,,GCC mirror sites}. Header bugs may generally
3492 be avoided using @command{fixincludes}, but bugs or deficiencies in
3493 libraries and the operating system may still cause problems.
3495 For some systems, old versions of GNU binutils may also be useful,
3496 and are available from @file{pub/binutils/old-releases} on
3497 @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}.
3499 Some of the information on specific systems above relates to
3500 such older systems, but much of the information
3501 about GCC on such systems (which may no longer be applicable to
3502 current GCC) is to be found in the GCC texinfo manual.
3508 @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris 2, etc.)
3510 C++ support is significantly better on ELF targets if you use the
3511 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
3512 inlines, vtables and template instantiations will be discarded
3522 @uref{./index.html,,Return to the GCC Installation page}
3526 @c ***************************************************************************
3527 @c Part 6 The End of the Document
3529 @comment node-name, next, previous, up
3530 @node Concept Index, , Specific, Top
3534 @unnumbered Concept Index