1 \input texinfo.tex @c -*-texinfo-*-
4 @setfilename gccinstall.info
9 @include gcc-common.texi
11 @c Specify title for specific html page
13 @settitle Installing GCC
16 @settitle Host/Target specific installation notes for GCC
18 @ifset prerequisiteshtml
19 @settitle Prerequisites for GCC
22 @settitle Downloading GCC
25 @settitle Installing GCC: Configuration
28 @settitle Installing GCC: Building
31 @settitle Installing GCC: Testing
33 @ifset finalinstallhtml
34 @settitle Installing GCC: Final installation
37 @settitle Installing GCC: Binaries
40 @settitle Installing GCC: GNU Free Documentation License
43 @c Copyright (C) 1988-2021 Free Software Foundation, Inc.
44 @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com
46 @c IMPORTANT: whenever you modify this file, run `install.texi2html' to
47 @c test the generation of HTML documents for the gcc.gnu.org web pages.
49 @c Do not use @footnote{} in this file as it breaks install.texi2html!
51 @c Include everything if we're not making html
55 @set prerequisiteshtml
65 @c Part 2 Summary Description and Copyright
67 Copyright @copyright{} 1988-2021 Free Software Foundation, Inc.
69 Permission is granted to copy, distribute and/or modify this document
70 under the terms of the GNU Free Documentation License, Version 1.3 or
71 any later version published by the Free Software Foundation; with no
72 Invariant Sections, the Front-Cover texts being (a) (see below), and
73 with the Back-Cover Texts being (b) (see below). A copy of the
74 license is included in the section entitled ``@uref{./gfdl.html,,GNU
75 Free Documentation License}''.
77 (a) The FSF's Front-Cover Text is:
81 (b) The FSF's Back-Cover Text is:
83 You have freedom to copy and modify this GNU Manual, like GNU
84 software. Copies published by the Free Software Foundation raise
85 funds for GNU development.
90 @dircategory Software development
92 * gccinstall: (gccinstall). Installing the GNU Compiler Collection.
95 @c Part 3 Titlepage and Copyright
100 @c The following two commands start the copyright page.
102 @vskip 0pt plus 1filll
106 @c Part 4 Top node, Master Menu, and/or Table of Contents
109 @comment node-name, next, Previous, up
112 * Installing GCC:: This document describes the generic installation
113 procedure for GCC as well as detailing some target
114 specific installation instructions.
116 * Specific:: Host/target specific installation notes for GCC.
117 * Binaries:: Where to get pre-compiled binaries.
119 * GNU Free Documentation License:: How you can copy and share this manual.
120 * Concept Index:: This index has two entries.
128 @c Part 5 The Body of the Document
129 @c ***Installing GCC**********************************************************
131 @comment node-name, next, previous, up
132 @node Installing GCC, Binaries, , Top
136 @chapter Installing GCC
139 The latest version of this document is always available at
140 @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}.
141 It refers to the current development sources, instructions for
142 specific released versions are included with the sources.
144 This document describes the generic installation procedure for GCC as well
145 as detailing some target specific installation instructions.
147 GCC includes several components that previously were separate distributions
148 with their own installation instructions. This document supersedes all
149 package-specific installation instructions.
151 @emph{Before} starting the build/install procedure please check the
153 @ref{Specific, host/target specific installation notes}.
156 @uref{specific.html,,host/target specific installation notes}.
158 We recommend you browse the entire generic installation instructions before
161 Lists of successful builds for released versions of GCC are
162 available at @uref{http://gcc.gnu.org/buildstat.html}.
163 These lists are updated as new information becomes available.
165 The installation procedure itself is broken into five steps.
170 * Downloading the source::
173 * Testing:: (optional)
180 @uref{prerequisites.html,,Prerequisites}
182 @uref{download.html,,Downloading the source}
184 @uref{configure.html,,Configuration}
186 @uref{build.html,,Building}
188 @uref{test.html,,Testing} (optional)
190 @uref{finalinstall.html,,Final install}
194 Please note that GCC does not support @samp{make uninstall} and probably
195 won't do so in the near future as this would open a can of worms. Instead,
196 we suggest that you install GCC into a directory of its own and simply
197 remove that directory when you do not need that specific version of GCC
198 any longer, and, if shared libraries are installed there as well, no
199 more binaries exist that use them.
202 There are also some @uref{old.html,,old installation instructions},
203 which are mostly obsolete but still contain some information which has
204 not yet been merged into the main part of this manual.
212 @uref{./index.html,,Return to the GCC Installation page}
218 @c ***Prerequisites**************************************************
220 @comment node-name, next, previous, up
221 @node Prerequisites, Downloading the source, , Installing GCC
223 @ifset prerequisiteshtml
225 @chapter Prerequisites
227 @cindex Prerequisites
229 GCC requires that various tools and packages be available for use in the
230 build procedure. Modifying GCC sources requires additional tools
233 @heading Tools/packages necessary for building GCC
235 @item ISO C++11 compiler
236 Necessary to bootstrap GCC.
238 Versions of GCC prior to 11 also allow bootstrapping with an ISO C++98
239 compiler, versions of GCC prior to 4.8 also allow bootstrapping with a
240 ISO C89 compiler, and versions of GCC prior to 3.4 also allow
241 bootstrapping with a traditional (K&R) C compiler.
243 To build all languages in a cross-compiler or other configuration where
244 3-stage bootstrap is not performed, you need to start with an existing
245 GCC binary (version 4.8 or later) because source code for language
246 frontends other than C might use GCC extensions.
248 Note that to bootstrap GCC with versions of GCC earlier than 4.8, you
249 may need to use @option{--disable-stage1-checking}, though
250 bootstrapping the compiler with such earlier compilers is strongly
253 @item C standard library and headers
255 In order to build GCC, the C standard library and headers must be present
256 for all target variants for which target libraries will be built (and not
257 only the variant of the host C++ compiler).
259 This affects the popular @samp{x86_64-pc-linux-gnu} platform (among
260 other multilib targets), for which 64-bit (@samp{x86_64}) and 32-bit
261 (@samp{i386}) libc headers are usually packaged separately. If you do a
262 build of a native compiler on @samp{x86_64-pc-linux-gnu}, make sure you
263 either have the 32-bit libc developer package properly installed (the exact
264 name of the package depends on your distro) or you must build GCC as a
265 64-bit only compiler by configuring with the option
266 @option{--disable-multilib}. Otherwise, you may encounter an error such as
267 @samp{fatal error: gnu/stubs-32.h: No such file}
269 @item @anchor{GNAT-prerequisite}GNAT
271 In order to build GNAT, the Ada compiler, you need a working GNAT
272 compiler (GCC version 4.7 or later).
274 This includes GNAT tools such as @command{gnatmake} and
275 @command{gnatlink}, since the Ada front end is written in Ada and
276 uses some GNAT-specific extensions.
278 In order to build a cross compiler, it is strongly recommended to install
279 the new compiler as native first, and then use it to build the cross
280 compiler. Other native compiler versions may work but this is not guaranteed and
281 will typically fail with hard to understand compilation errors during the
284 Similarly, it is strongly recommended to use an older version of GNAT to build
285 GNAT. More recent versions of GNAT than the version built are not guaranteed
286 to work and will often fail during the build with compilation errors.
288 Note that @command{configure} does not test whether the GNAT installation works
289 and has a sufficiently recent version; if too old a GNAT version is
290 installed and @option{--enable-languages=ada} is used, the build will fail.
292 @env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables
293 must not be set when building the Ada compiler, the Ada tools, or the
294 Ada runtime libraries. You can check that your build environment is clean
295 by verifying that @samp{gnatls -v} lists only one explicit path in each
298 @item A ``working'' POSIX compatible shell, or GNU bash
300 Necessary when running @command{configure} because some
301 @command{/bin/sh} shells have bugs and may crash when configuring the
302 target libraries. In other cases, @command{/bin/sh} or @command{ksh}
303 have disastrous corner-case performance problems. This
304 can cause target @command{configure} runs to literally take days to
305 complete in some cases.
307 So on some platforms @command{/bin/ksh} is sufficient, on others it
308 isn't. See the host/target specific instructions for your platform, or
309 use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your
310 environment to your ``good'' shell prior to running
311 @command{configure}/@command{make}.
313 @command{zsh} is not a fully compliant POSIX shell and will not
314 work when configuring GCC@.
316 @item A POSIX or SVR4 awk
318 Necessary for creating some of the generated source files for GCC@.
319 If in doubt, use a recent GNU awk version, as some of the older ones
320 are broken. GNU awk version 3.1.5 is known to work.
324 Necessary in some circumstances, optional in others. See the
325 host/target specific instructions for your platform for the exact
328 @item gzip version 1.2.4 (or later) or
329 @itemx bzip2 version 1.0.2 (or later)
331 Necessary to uncompress GCC @command{tar} files when source code is
332 obtained via HTTPS mirror sites.
334 @item GNU make version 3.80 (or later)
336 You must have GNU make installed to build GCC@.
338 @item GNU tar version 1.14 (or later)
340 Necessary (only on some platforms) to untar the source code. Many
341 systems' @command{tar} programs will also work, only try GNU
342 @command{tar} if you have problems.
344 @item Perl version between 5.6.1 and 5.6.24
346 Necessary when targeting Darwin, building @samp{libstdc++},
347 and not using @option{--disable-symvers}.
348 Necessary when targeting Solaris 2 with Solaris @command{ld} and not using
349 @option{--disable-symvers}.
351 Necessary when regenerating @file{Makefile} dependencies in libiberty.
352 Necessary when regenerating @file{libiberty/functions.texi}.
353 Necessary when generating manpages from Texinfo manuals.
354 Used by various scripts to generate some files included in the source
355 repository (mainly Unicode-related and rarely changing) from source
358 Used by @command{automake}.
362 Several support libraries are necessary to build GCC, some are required,
363 others optional. While any sufficiently new version of required tools
364 usually work, library requirements are generally stricter. Newer
365 versions may work in some cases, but it's safer to use the exact
366 versions documented. We appreciate bug reports about problems with
367 newer versions, though. If your OS vendor provides packages for the
368 support libraries then using those packages may be the simplest way to
369 install the libraries.
372 @item GNU Multiple Precision Library (GMP) version 4.3.2 (or later)
374 Necessary to build GCC@. If a GMP source distribution is found in a
375 subdirectory of your GCC sources named @file{gmp}, it will be built
376 together with GCC. Alternatively, if GMP is already installed but it
377 is not in your library search path, you will have to configure with the
378 @option{--with-gmp} configure option. See also @option{--with-gmp-lib}
379 and @option{--with-gmp-include}.
380 The in-tree build is only supported with the GMP version that
381 download_prerequisites installs.
383 @item MPFR Library version 3.1.0 (or later)
385 Necessary to build GCC@. It can be downloaded from
386 @uref{https://www.mpfr.org}. If an MPFR source distribution is found
387 in a subdirectory of your GCC sources named @file{mpfr}, it will be
388 built together with GCC. Alternatively, if MPFR is already installed
389 but it is not in your default library search path, the
390 @option{--with-mpfr} configure option should be used. See also
391 @option{--with-mpfr-lib} and @option{--with-mpfr-include}.
392 The in-tree build is only supported with the MPFR version that
393 download_prerequisites installs.
395 @item MPC Library version 1.0.1 (or later)
397 Necessary to build GCC@. It can be downloaded from
398 @uref{http://www.multiprecision.org/mpc/}. If an MPC source distribution
399 is found in a subdirectory of your GCC sources named @file{mpc}, it
400 will be built together with GCC. Alternatively, if MPC is already
401 installed but it is not in your default library search path, the
402 @option{--with-mpc} configure option should be used. See also
403 @option{--with-mpc-lib} and @option{--with-mpc-include}.
404 The in-tree build is only supported with the MPC version that
405 download_prerequisites installs.
407 @item isl Library version 0.15 or later.
409 Necessary to build GCC with the Graphite loop optimizations.
410 It can be downloaded from @uref{https://gcc.gnu.org/pub/gcc/infrastructure/}.
411 If an isl source distribution is found
412 in a subdirectory of your GCC sources named @file{isl}, it will be
413 built together with GCC. Alternatively, the @option{--with-isl} configure
414 option should be used if isl is not installed in your default library
419 Necessary to build GCC with zstd compression used for LTO bytecode.
420 The library is searched in your default library patch search.
421 Alternatively, the @option{--with-zstd} configure option should be used.
425 @heading Tools/packages necessary for modifying GCC
427 @item autoconf version 2.69
428 @itemx GNU m4 version 1.4.6 (or later)
430 Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@:
431 to regenerate @file{configure} and @file{config.in} files.
433 @item automake version 1.15.1
435 Necessary when modifying a @file{Makefile.am} file to regenerate its
436 associated @file{Makefile.in}.
438 Much of GCC does not use automake, so directly edit the @file{Makefile.in}
439 file. Specifically this applies to the @file{gcc}, @file{intl},
440 @file{libcpp}, @file{libiberty}, @file{libobjc} directories as well
441 as any of their subdirectories.
443 For directories that use automake, GCC requires the latest release in
444 the 1.15 series, which is currently 1.15.1. When regenerating a directory
445 to a newer version, please update all the directories using an older 1.15
446 to the latest released version.
448 @item gettext version 0.14.5 (or later)
450 Needed to regenerate @file{gcc.pot}.
452 @item gperf version 2.7.2 (or later)
454 Necessary when modifying @command{gperf} input files, e.g.@:
455 @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@:
456 @file{gcc/cp/cfns.h}.
461 @c Once Tcl 8.5 or higher is required, remove any obsolete
462 @c compatibility workarounds:
463 @c git grep 'compatibility with earlier Tcl releases'
465 Necessary to run the GCC testsuite; see the section on testing for
468 @item autogen version 5.5.4 (or later) and
469 @itemx guile version 1.4.1 (or later)
471 Necessary to regenerate @file{fixinc/fixincl.x} from
472 @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}.
474 Necessary to run @samp{make check} for @file{fixinc}.
476 Necessary to regenerate the top level @file{Makefile.in} file from
477 @file{Makefile.tpl} and @file{Makefile.def}.
479 @item Flex version 2.5.4 (or later)
481 Necessary when modifying @file{*.l} files.
483 Necessary to build GCC during development because the generated output
484 files are not included in the version-controlled source repository.
485 They are included in releases.
487 @item Texinfo version 4.7 (or later)
489 Necessary for running @command{makeinfo} when modifying @file{*.texi}
490 files to test your changes.
492 Necessary for running @command{make dvi} or @command{make pdf} to
493 create printable documentation in DVI or PDF format. Texinfo version
494 4.8 or later is required for @command{make pdf}.
496 Necessary to build GCC documentation during development because the
497 generated output files are not included in the repository. They are
498 included in releases.
500 @item @TeX{} (any working version)
502 Necessary for running @command{texi2dvi} and @command{texi2pdf}, which
503 are used when running @command{make dvi} or @command{make pdf} to create
504 DVI or PDF files, respectively.
506 @item Sphinx version 1.0 (or later)
508 Necessary to regenerate @file{jit/docs/_build/texinfo} from the @file{.rst}
509 files in the directories below @file{jit/docs}.
511 @item git (any version)
512 @itemx SSH (any version)
514 Necessary to access the source repository. Public releases and weekly
515 snapshots of the development sources are also available via HTTPS@.
517 @item GNU diffutils version 2.7 (or later)
519 Useful when submitting patches for the GCC source code.
521 @item patch version 2.5.4 (or later)
523 Necessary when applying patches, created with @command{diff}, to one's
533 @uref{./index.html,,Return to the GCC Installation page}
537 @c ***Downloading the source**************************************************
539 @comment node-name, next, previous, up
540 @node Downloading the source, Configuration, Prerequisites, Installing GCC
544 @chapter Downloading GCC
546 @cindex Downloading GCC
547 @cindex Downloading the Source
549 GCC is distributed via @uref{http://gcc.gnu.org/git.html,,git} and via
550 HTTPS as tarballs compressed with @command{gzip} or @command{bzip2}.
552 Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page}
553 for information on how to obtain GCC@.
555 The source distribution includes the C, C++, Objective-C, Fortran,
556 and Ada (in the case of GCC 3.1 and later) compilers, as well as
557 runtime libraries for C++, Objective-C, and Fortran.
558 For previous versions these were downloadable as separate components such
559 as the core GCC distribution, which included the C language front end and
560 shared components, and language-specific distributions including the
561 language front end and the language runtime (where appropriate).
563 If you also intend to build binutils (either to upgrade an existing
564 installation or for use in place of the corresponding tools of your
565 OS), unpack the binutils distribution either in the same directory or
566 a separate one. In the latter case, add symbolic links to any
567 components of the binutils you intend to build alongside the compiler
568 (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld},
569 @file{opcodes}, @dots{}) to the directory containing the GCC sources.
571 Likewise the GMP, MPFR and MPC libraries can be automatically built
572 together with GCC. You may simply run the
573 @command{contrib/download_prerequisites} script in the GCC source directory
574 to set up everything.
575 Otherwise unpack the GMP, MPFR and/or MPC source
576 distributions in the directory containing the GCC sources and rename
577 their directories to @file{gmp}, @file{mpfr} and @file{mpc},
578 respectively (or use symbolic links with the same name).
585 @uref{./index.html,,Return to the GCC Installation page}
589 @c ***Configuration***********************************************************
591 @comment node-name, next, previous, up
592 @node Configuration, Building, Downloading the source, Installing GCC
596 @chapter Installing GCC: Configuration
598 @cindex Configuration
599 @cindex Installing GCC: Configuration
601 Like most GNU software, GCC must be configured before it can be built.
602 This document describes the recommended configuration procedure
603 for both native and cross targets.
605 We use @var{srcdir} to refer to the toplevel source directory for
606 GCC; we use @var{objdir} to refer to the toplevel build/object directory.
608 If you obtained the sources by cloning the repository, @var{srcdir}
609 must refer to the top @file{gcc} directory, the one where the
610 @file{MAINTAINERS} file can be found, and not its @file{gcc}
611 subdirectory, otherwise the build will fail.
613 If either @var{srcdir} or @var{objdir} is located on an automounted NFS
614 file system, the shell's built-in @command{pwd} command will return
615 temporary pathnames. Using these can lead to various sorts of build
616 problems. To avoid this issue, set the @env{PWDCMD} environment
617 variable to an automounter-aware @command{pwd} command, e.g.,
618 @command{pawd} or @samp{amq -w}, during the configuration and build
621 First, we @strong{highly} recommend that GCC be built into a
622 separate directory from the sources which does @strong{not} reside
623 within the source tree. This is how we generally build GCC; building
624 where @var{srcdir} == @var{objdir} should still work, but doesn't
625 get extensive testing; building where @var{objdir} is a subdirectory
626 of @var{srcdir} is unsupported.
628 If you have previously built GCC in the same directory for a
629 different target machine, do @samp{make distclean} to delete all files
630 that might be invalid. One of the files this deletes is @file{Makefile};
631 if @samp{make distclean} complains that @file{Makefile} does not exist
632 or issues a message like ``don't know how to make distclean'' it probably
633 means that the directory is already suitably clean. However, with the
634 recommended method of building in a separate @var{objdir}, you should
635 simply use a different @var{objdir} for each target.
637 Second, when configuring a native system, either @command{cc} or
638 @command{gcc} must be in your path or you must set @env{CC} in
639 your environment before running configure. Otherwise the configuration
643 Note that the bootstrap compiler and the resulting GCC must be link
644 compatible, else the bootstrap will fail with linker errors about
645 incompatible object file formats. Several multilibed targets are
646 affected by this requirement, see
648 @ref{Specific, host/target specific installation notes}.
651 @uref{specific.html,,host/target specific installation notes}.
660 % @var{srcdir}/configure [@var{options}] [@var{target}]
663 @heading Distributor options
665 If you will be distributing binary versions of GCC, with modifications
666 to the source code, you should use the options described in this
667 section to make clear that your version contains modifications.
670 @item --with-pkgversion=@var{version}
671 Specify a string that identifies your package. You may wish
672 to include a build number or build date. This version string will be
673 included in the output of @command{gcc --version}. This suffix does
674 not replace the default version string, only the @samp{GCC} part.
676 The default value is @samp{GCC}.
678 @item --with-bugurl=@var{url}
679 Specify the URL that users should visit if they wish to report a bug.
680 You are of course welcome to forward bugs reported to you to the FSF,
681 if you determine that they are not bugs in your modifications.
683 The default value refers to the FSF's GCC bug tracker.
685 @item --with-documentation-root-url=@var{url}
686 Specify the URL root that contains GCC option documentation. The @var{url}
687 should end with a @code{/} character.
689 The default value is @uref{https://gcc.gnu.org/onlinedocs/,,https://gcc.gnu.org/onlinedocs/}.
691 @item --with-changes-root-url=@var{url}
692 Specify the URL root that contains information about changes in GCC
693 releases like @code{gcc-@var{version}/changes.html}.
694 The @var{url} should end with a @code{/} character.
696 The default value is @uref{https://gcc.gnu.org/,,https://gcc.gnu.org/}.
700 @heading Host, Build and Target specification
702 Specify the host, build and target machine configurations. You do this
703 when you run the @file{configure} script.
705 The @dfn{build} machine is the system which you are using, the
706 @dfn{host} machine is the system where you want to run the resulting
707 compiler (normally the build machine), and the @dfn{target} machine is
708 the system for which you want the compiler to generate code.
710 If you are building a compiler to produce code for the machine it runs
711 on (a native compiler), you normally do not need to specify any operands
712 to @file{configure}; it will try to guess the type of machine you are on
713 and use that as the build, host and target machines. So you don't need
714 to specify a configuration when building a native compiler unless
715 @file{configure} cannot figure out what your configuration is or guesses
718 In those cases, specify the build machine's @dfn{configuration name}
719 with the @option{--host} option; the host and target will default to be
720 the same as the host machine.
725 ./configure --host=x86_64-pc-linux-gnu
728 A configuration name may be canonical or it may be more or less
729 abbreviated (@file{config.sub} script produces canonical versions).
731 A canonical configuration name has three parts, separated by dashes.
732 It looks like this: @samp{@var{cpu}-@var{company}-@var{system}}.
734 Here are the possible CPU types:
737 aarch64, aarch64_be, alpha, alpha64, amdgcn, arc, arceb, arm, armeb, avr, bfin,
738 bpf, cr16, cris, csky, epiphany, fido, fr30, frv, ft32, h8300, hppa, hppa2.0,
739 hppa64, i486, i686, ia64, iq2000, lm32, m32c, m32r, m32rle, m68k, mcore,
740 microblaze, microblazeel, mips, mips64, mips64el, mips64octeon, mips64orion,
741 mips64vr, mipsel, mipsisa32, mipsisa32r2, mipsisa64, mipsisa64r2,
742 mipsisa64r2el, mipsisa64sb1, mipsisa64sr71k, mipstx39, mmix, mn10300, moxie,
743 msp430, nds32be, nds32le, nios2, nvptx, or1k, pdp11, powerpc, powerpc64,
744 powerpc64le, powerpcle, pru, riscv32, riscv32be, riscv64, riscv64be, rl78, rx,
745 s390, s390x, sh, shle, sparc, sparc64, tic6x, tilegx, tilegxbe, tilepro, v850,
746 v850e, v850e1, vax, visium, x86_64, xstormy16, xtensa
749 Here is a list of system types:
752 aix@var{version}, amdhsa, aout, cygwin, darwin@var{version},
753 eabi, eabialtivec, eabisim, eabisimaltivec, elf, elf32,
754 elfbare, elfoabi, freebsd@var{version}, gnu, hpux, hpux@var{version},
755 kfreebsd-gnu, kopensolaris-gnu, linux-androideabi, linux-gnu,
756 linux-gnu_altivec, linux-musl, linux-uclibc, lynxos, mingw32, mingw32crt,
757 mmixware, msdosdjgpp, netbsd, netbsdelf@var{version}, nto-qnx, openbsd,
758 rtems, solaris@var{version}, symbianelf, tpf, uclinux, uclinux_eabi, vms,
759 vxworks, vxworksae, vxworksmils
762 @heading Options specification
764 Use @var{options} to override several configure time options for
765 GCC@. A list of supported @var{options} follows; @samp{configure
766 --help} may list other options, but those not listed below may not
767 work and should not normally be used.
769 Note that each @option{--enable} option has a corresponding
770 @option{--disable} option and that each @option{--with} option has a
771 corresponding @option{--without} option.
774 @item --prefix=@var{dirname}
775 Specify the toplevel installation
776 directory. This is the recommended way to install the tools into a directory
777 other than the default. The toplevel installation directory defaults to
780 We @strong{highly} recommend against @var{dirname} being the same or a
781 subdirectory of @var{objdir} or vice versa. If specifying a directory
782 beneath a user's home directory tree, some shells will not expand
783 @var{dirname} correctly if it contains the @samp{~} metacharacter; use
786 The following standard @command{autoconf} options are supported. Normally you
787 should not need to use these options.
789 @item --exec-prefix=@var{dirname}
790 Specify the toplevel installation directory for architecture-dependent
791 files. The default is @file{@var{prefix}}.
793 @item --bindir=@var{dirname}
794 Specify the installation directory for the executables called by users
795 (such as @command{gcc} and @command{g++}). The default is
796 @file{@var{exec-prefix}/bin}.
798 @item --libdir=@var{dirname}
799 Specify the installation directory for object code libraries and
800 internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}.
802 @item --libexecdir=@var{dirname}
803 Specify the installation directory for internal executables of GCC@.
804 The default is @file{@var{exec-prefix}/libexec}.
806 @item --with-slibdir=@var{dirname}
807 Specify the installation directory for the shared libgcc library. The
808 default is @file{@var{libdir}}.
810 @item --datarootdir=@var{dirname}
811 Specify the root of the directory tree for read-only architecture-independent
812 data files referenced by GCC@. The default is @file{@var{prefix}/share}.
814 @item --infodir=@var{dirname}
815 Specify the installation directory for documentation in info format.
816 The default is @file{@var{datarootdir}/info}.
818 @item --datadir=@var{dirname}
819 Specify the installation directory for some architecture-independent
820 data files referenced by GCC@. The default is @file{@var{datarootdir}}.
822 @item --docdir=@var{dirname}
823 Specify the installation directory for documentation files (other
824 than Info) for GCC@. The default is @file{@var{datarootdir}/doc}.
826 @item --htmldir=@var{dirname}
827 Specify the installation directory for HTML documentation files.
828 The default is @file{@var{docdir}}.
830 @item --pdfdir=@var{dirname}
831 Specify the installation directory for PDF documentation files.
832 The default is @file{@var{docdir}}.
834 @item --mandir=@var{dirname}
835 Specify the installation directory for manual pages. The default is
836 @file{@var{datarootdir}/man}. (Note that the manual pages are only extracts
837 from the full GCC manuals, which are provided in Texinfo format. The manpages
838 are derived by an automatic conversion process from parts of the full
841 @item --with-gxx-include-dir=@var{dirname}
843 the installation directory for G++ header files. The default depends
844 on other configuration options, and differs between cross and native
847 @item --with-specs=@var{specs}
848 Specify additional command line driver SPECS.
849 This can be useful if you need to turn on a non-standard feature by
850 default without modifying the compiler's source code, for instance
851 @option{--with-specs=%@{!fcommon:%@{!fno-common:-fno-common@}@}}.
853 @xref{Spec Files,, Specifying subprocesses and the switches to pass to them,
854 gcc, Using the GNU Compiler Collection (GCC)},
857 See ``Spec Files'' in the main manual
862 @item --program-prefix=@var{prefix}
863 GCC supports some transformations of the names of its programs when
864 installing them. This option prepends @var{prefix} to the names of
865 programs to install in @var{bindir} (see above). For example, specifying
866 @option{--program-prefix=foo-} would result in @samp{gcc}
867 being installed as @file{/usr/local/bin/foo-gcc}.
869 @item --program-suffix=@var{suffix}
870 Appends @var{suffix} to the names of programs to install in @var{bindir}
871 (see above). For example, specifying @option{--program-suffix=-3.1}
872 would result in @samp{gcc} being installed as
873 @file{/usr/local/bin/gcc-3.1}.
875 @item --program-transform-name=@var{pattern}
876 Applies the @samp{sed} script @var{pattern} to be applied to the names
877 of programs to install in @var{bindir} (see above). @var{pattern} has to
878 consist of one or more basic @samp{sed} editing commands, separated by
879 semicolons. For example, if you want the @samp{gcc} program name to be
880 transformed to the installed program @file{/usr/local/bin/myowngcc} and
881 the @samp{g++} program name to be transformed to
882 @file{/usr/local/bin/gspecial++} without changing other program names,
883 you could use the pattern
884 @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'}
885 to achieve this effect.
887 All three options can be combined and used together, resulting in more
888 complex conversion patterns. As a basic rule, @var{prefix} (and
889 @var{suffix}) are prepended (appended) before further transformations
890 can happen with a special transformation script @var{pattern}.
892 As currently implemented, this option only takes effect for native
893 builds; cross compiler binaries' names are not transformed even when a
894 transformation is explicitly asked for by one of these options.
896 For native builds, some of the installed programs are also installed
897 with the target alias in front of their name, as in
898 @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen
899 before the target alias is prepended to the name---so, specifying
900 @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the
901 resulting binary would be installed as
902 @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}.
904 As a last shortcoming, none of the installed Ada programs are
905 transformed yet, which will be fixed in some time.
907 @item --with-local-prefix=@var{dirname}
909 installation directory for local include files. The default is
910 @file{/usr/local}. Specify this option if you want the compiler to
911 search directory @file{@var{dirname}/include} for locally installed
912 header files @emph{instead} of @file{/usr/local/include}.
914 You should specify @option{--with-local-prefix} @strong{only} if your
915 site has a different convention (not @file{/usr/local}) for where to put
918 The default value for @option{--with-local-prefix} is @file{/usr/local}
919 regardless of the value of @option{--prefix}. Specifying
920 @option{--prefix} has no effect on which directory GCC searches for
921 local header files. This may seem counterintuitive, but actually it is
924 The purpose of @option{--prefix} is to specify where to @emph{install
925 GCC}. The local header files in @file{/usr/local/include}---if you put
926 any in that directory---are not part of GCC@. They are part of other
927 programs---perhaps many others. (GCC installs its own header files in
928 another directory which is based on the @option{--prefix} value.)
930 Both the local-prefix include directory and the GCC-prefix include
931 directory are part of GCC's ``system include'' directories. Although these
932 two directories are not fixed, they need to be searched in the proper
933 order for the correct processing of the include_next directive. The
934 local-prefix include directory is searched before the GCC-prefix
935 include directory. Another characteristic of system include directories
936 is that pedantic warnings are turned off for headers in these directories.
938 Some autoconf macros add @option{-I @var{directory}} options to the
939 compiler command line, to ensure that directories containing installed
940 packages' headers are searched. When @var{directory} is one of GCC's
941 system include directories, GCC will ignore the option so that system
942 directories continue to be processed in the correct order. This
943 may result in a search order different from what was specified but the
944 directory will still be searched.
946 GCC automatically searches for ordinary libraries using
947 @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is
948 used for both GCC and packages, GCC will automatically search for
949 both headers and libraries. This provides a configuration that is
950 easy to use. GCC behaves in a manner similar to that when it is
951 installed as a system compiler in @file{/usr}.
953 Sites that need to install multiple versions of GCC may not want to
954 use the above simple configuration. It is possible to use the
955 @option{--program-prefix}, @option{--program-suffix} and
956 @option{--program-transform-name} options to install multiple versions
957 into a single directory, but it may be simpler to use different prefixes
958 and the @option{--with-local-prefix} option to specify the location of the
959 site-specific files for each version. It will then be necessary for
960 users to specify explicitly the location of local site libraries
961 (e.g., with @env{LIBRARY_PATH}).
963 The same value can be used for both @option{--with-local-prefix} and
964 @option{--prefix} provided it is not @file{/usr}. This can be used
965 to avoid the default search of @file{/usr/local/include}.
967 @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}!
968 The directory you use for @option{--with-local-prefix} @strong{must not}
969 contain any of the system's standard header files. If it did contain
970 them, certain programs would be miscompiled (including GNU Emacs, on
971 certain targets), because this would override and nullify the header
972 file corrections made by the @command{fixincludes} script.
974 Indications are that people who use this option use it based on mistaken
975 ideas of what it is for. People use it as if it specified where to
976 install part of GCC@. Perhaps they make this assumption because
977 installing GCC creates the directory.
979 @item --with-gcc-major-version-only
980 Specifies that GCC should use only the major number rather than
981 @var{major}.@var{minor}.@var{patchlevel} in filesystem paths.
983 @item --with-native-system-header-dir=@var{dirname}
984 Specifies that @var{dirname} is the directory that contains native system
985 header files, rather than @file{/usr/include}. This option is most useful
986 if you are creating a compiler that should be isolated from the system
987 as much as possible. It is most commonly used with the
988 @option{--with-sysroot} option and will cause GCC to search
989 @var{dirname} inside the system root specified by that option.
991 @item --enable-shared[=@var{package}[,@dots{}]]
992 Build shared versions of libraries, if shared libraries are supported on
993 the target platform. Unlike GCC 2.95.x and earlier, shared libraries
994 are enabled by default on all platforms that support shared libraries.
996 If a list of packages is given as an argument, build shared libraries
997 only for the listed packages. For other packages, only static libraries
998 will be built. Package names currently recognized in the GCC tree are
999 @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not
1000 @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc},
1001 @samp{ada}, @samp{libada}, @samp{libgo}, @samp{libobjc}, and @samp{libphobos}.
1002 Note @samp{libiberty} does not support shared libraries at all.
1004 Use @option{--disable-shared} to build only static libraries. Note that
1005 @option{--disable-shared} does not accept a list of package names as
1006 argument, only @option{--enable-shared} does.
1008 Contrast with @option{--enable-host-shared}, which affects @emph{host}
1011 @item --enable-host-shared
1012 Specify that the @emph{host} code should be built into position-independent
1013 machine code (with -fPIC), allowing it to be used within shared libraries,
1014 but yielding a slightly slower compiler.
1016 This option is required when building the libgccjit.so library.
1018 Contrast with @option{--enable-shared}, which affects @emph{target}
1021 @item @anchor{with-gnu-as}--with-gnu-as
1022 Specify that the compiler should assume that the
1023 assembler it finds is the GNU assembler. However, this does not modify
1024 the rules to find an assembler and will result in confusion if the
1025 assembler found is not actually the GNU assembler. (Confusion may also
1026 result if the compiler finds the GNU assembler but has not been
1027 configured with @option{--with-gnu-as}.) If you have more than one
1028 assembler installed on your system, you may want to use this option in
1029 connection with @option{--with-as=@var{pathname}} or
1030 @option{--with-build-time-tools=@var{pathname}}.
1032 The following systems are the only ones where it makes a difference
1033 whether you use the GNU assembler. On any other system,
1034 @option{--with-gnu-as} has no effect.
1037 @item @samp{hppa1.0-@var{any}-@var{any}}
1038 @item @samp{hppa1.1-@var{any}-@var{any}}
1039 @item @samp{sparc-sun-solaris2.@var{any}}
1040 @item @samp{sparc64-@var{any}-solaris2.@var{any}}
1043 @item @anchor{with-as}--with-as=@var{pathname}
1044 Specify that the compiler should use the assembler pointed to by
1045 @var{pathname}, rather than the one found by the standard rules to find
1046 an assembler, which are:
1049 Unless GCC is being built with a cross compiler, check the
1050 @file{@var{libexec}/gcc/@var{target}/@var{version}} directory.
1051 @var{libexec} defaults to @file{@var{exec-prefix}/libexec};
1052 @var{exec-prefix} defaults to @var{prefix}, which
1053 defaults to @file{/usr/local} unless overridden by the
1054 @option{--prefix=@var{pathname}} switch described above. @var{target}
1055 is the target system triple, such as @samp{sparc-sun-solaris2.7}, and
1056 @var{version} denotes the GCC version, such as 3.0.
1059 If the target system is the same that you are building on, check
1060 operating system specific directories (e.g.@: @file{/usr/ccs/bin} on
1064 Check in the @env{PATH} for a tool whose name is prefixed by the
1065 target system triple.
1068 Check in the @env{PATH} for a tool whose name is not prefixed by the
1069 target system triple, if the host and target system triple are
1070 the same (in other words, we use a host tool if it can be used for
1071 the target as well).
1074 You may want to use @option{--with-as} if no assembler
1075 is installed in the directories listed above, or if you have multiple
1076 assemblers installed and want to choose one that is not found by the
1079 @item @anchor{with-gnu-ld}--with-gnu-ld
1080 Same as @uref{#with-gnu-as,,@option{--with-gnu-as}}
1083 @item --with-ld=@var{pathname}
1084 Same as @uref{#with-as,,@option{--with-as}}
1088 Specify that stabs debugging
1089 information should be used instead of whatever format the host normally
1090 uses. Normally GCC uses the same debug format as the host system.
1092 @item --with-tls=@var{dialect}
1093 Specify the default TLS dialect, for systems were there is a choice.
1094 For ARM targets, possible values for @var{dialect} are @code{gnu} or
1095 @code{gnu2}, which select between the original GNU dialect and the GNU TLS
1096 descriptor-based dialect.
1098 @item --enable-multiarch
1099 Specify whether to enable or disable multiarch support. The default is
1100 to check for glibc start files in a multiarch location, and enable it
1101 if the files are found. The auto detection is enabled for native builds,
1102 and for cross builds configured with @option{--with-sysroot}, and without
1103 @option{--with-native-system-header-dir}.
1104 More documentation about multiarch can be found at
1105 @uref{https://wiki.debian.org/Multiarch}.
1107 @item --enable-sjlj-exceptions
1108 Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions.
1109 @samp{configure} ordinarily picks the correct value based on the platform.
1110 Only use this option if you are sure you need a different setting.
1112 @item --enable-vtable-verify
1113 Specify whether to enable or disable the vtable verification feature.
1114 Enabling this feature causes libstdc++ to be built with its virtual calls
1115 in verifiable mode. This means that, when linked with libvtv, every
1116 virtual call in libstdc++ will verify the vtable pointer through which the
1117 call will be made before actually making the call. If not linked with libvtv,
1118 the verifier will call stub functions (in libstdc++ itself) and do nothing.
1119 If vtable verification is disabled, then libstdc++ is not built with its
1120 virtual calls in verifiable mode at all. However the libvtv library will
1121 still be built (see @option{--disable-libvtv} to turn off building libvtv).
1122 @option{--disable-vtable-verify} is the default.
1124 @item --disable-gcov
1125 Specify that the run-time library used for coverage analysis
1126 and associated host tools should not be built.
1128 @item --disable-multilib
1129 Specify that multiple target
1130 libraries to support different target variants, calling
1131 conventions, etc.@: should not be built. The default is to build a
1132 predefined set of them.
1134 Some targets provide finer-grained control over which multilibs are built
1135 (e.g., @option{--disable-softfloat}):
1138 fpu, 26bit, underscore, interwork, biendian, nofmult.
1141 softfloat, m68881, m68000, m68020.
1144 single-float, biendian, softfloat.
1149 @item powerpc*-*-*, rs6000*-*-*
1150 aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
1155 @item --with-multilib-list=@var{list}
1156 @itemx --without-multilib-list
1157 Specify what multilibs to build. @var{list} is a comma separated list of
1158 values, possibly consisting of a single value. Currently only implemented
1159 for aarch64*-*-*, arm*-*-*, riscv*-*-*, sh*-*-* and x86-64-*-linux*. The
1160 accepted values and meaning for each target is given below.
1164 @var{list} is a comma separated list of @code{ilp32}, and @code{lp64}
1165 to enable ILP32 and LP64 run-time libraries, respectively. If
1166 @var{list} is empty, then there will be no multilibs and only the
1167 default run-time library will be built. If @var{list} is
1168 @code{default} or --with-multilib-list= is not specified, then the
1169 default set of libraries is selected based on the value of
1173 @var{list} is a comma separated list of @code{aprofile} and
1174 @code{rmprofile} to build multilibs for A or R and M architecture
1175 profiles respectively. Note that, due to some limitation of the current
1176 multilib framework, using the combined @code{aprofile,rmprofile}
1177 multilibs selects in some cases a less optimal multilib than when using
1178 the multilib profile for the architecture targetted. The special value
1179 @code{default} is also accepted and is equivalent to omitting the
1180 option, i.e., only the default run-time library will be enabled.
1182 @var{list} may instead contain @code{@@name}, to use the multilib
1183 configuration Makefile fragment @file{name} in @file{gcc/config/arm} in
1184 the source tree (it is part of the corresponding sources, after all).
1185 It is recommended, but not required, that files used for this purpose to
1186 be named starting with @file{t-ml-}, to make their intended purpose
1187 self-evident, in line with GCC conventions. Such files enable custom,
1188 user-chosen multilib lists to be configured. Whether multiple such
1189 files can be used together depends on the contents of the supplied
1190 files. See @file{gcc/config/arm/t-multilib} and its supplementary
1191 @file{gcc/config/arm/t-*profile} files for an example of what such
1192 Makefile fragments might look like for this version of GCC. The macros
1193 expected to be defined in these fragments are not stable across GCC
1194 releases, so make sure they define the @code{MULTILIB}-related macros
1195 expected by the version of GCC you are building.
1197 @xref{Target Fragment,, Target Makefile Fragments, gccint, GNU Compiler
1198 Collection (GCC) Internals}.
1201 See ``Target Makefile Fragments'' in the internals manual.
1204 The table below gives the combination of ISAs, architectures, FPUs and
1205 floating-point ABIs for which multilibs are built for each predefined
1206 profile. The union of these options is considered when specifying both
1207 @code{aprofile} and @code{rmprofile}.
1209 @multitable @columnfractions .15 .28 .30
1210 @item Option @tab aprofile @tab rmprofile
1212 @tab @code{-marm} and @code{-mthumb}
1214 @item Architectures@*@*@*@*@*@*
1215 @tab default architecture@*
1216 @code{-march=armv7-a}@*
1217 @code{-march=armv7ve}@*
1218 @code{-march=armv8-a}@*@*@*
1219 @tab default architecture@*
1220 @code{-march=armv6s-m}@*
1221 @code{-march=armv7-m}@*
1222 @code{-march=armv7e-m}@*
1223 @code{-march=armv8-m.base}@*
1224 @code{-march=armv8-m.main}@*
1226 @item FPUs@*@*@*@*@*
1228 @code{-mfpu=vfpv3-d16}@*
1230 @code{-mfpu=vfpv4-d16}@*
1231 @code{-mfpu=neon-vfpv4}@*
1232 @code{-mfpu=neon-fp-armv8}
1234 @code{-mfpu=vfpv3-d16}@*
1235 @code{-mfpu=fpv4-sp-d16}@*
1236 @code{-mfpu=fpv5-sp-d16}@*
1237 @code{-mfpu=fpv5-d16}@*
1238 @item floating-point@/ ABIs@*@*
1239 @tab @code{-mfloat-abi=soft}@*
1240 @code{-mfloat-abi=softfp}@*
1241 @code{-mfloat-abi=hard}
1242 @tab @code{-mfloat-abi=soft}@*
1243 @code{-mfloat-abi=softfp}@*
1244 @code{-mfloat-abi=hard}
1248 @var{list} is a single ABI name. The target architecture must be either
1249 @code{rv32gc} or @code{rv64gc}. This will build a single multilib for the
1250 specified architecture and ABI pair. If @code{--with-multilib-list} is not
1251 given, then a default set of multilibs is selected based on the value of
1252 @option{--target}. This is usually a large set of multilibs.
1255 @var{list} is a comma separated list of CPU names. These must be of the
1256 form @code{sh*} or @code{m*} (in which case they match the compiler option
1257 for that processor). The list should not contain any endian options -
1258 these are handled by @option{--with-endian}.
1260 If @var{list} is empty, then there will be no multilibs for extra
1261 processors. The multilib for the secondary endian remains enabled.
1263 As a special case, if an entry in the list starts with a @code{!}
1264 (exclamation point), then it is added to the list of excluded multilibs.
1265 Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES}
1266 (once the leading @code{!} has been stripped).
1268 If @option{--with-multilib-list} is not given, then a default set of
1269 multilibs is selected based on the value of @option{--target}. This is
1270 usually the complete set of libraries, but some targets imply a more
1273 Example 1: to configure a compiler for SH4A only, but supporting both
1274 endians, with little endian being the default:
1276 --with-cpu=sh4a --with-endian=little,big --with-multilib-list=
1279 Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with
1280 only little endian SH4AL:
1282 --with-cpu=sh4a --with-endian=little,big \
1283 --with-multilib-list=sh4al,!mb/m4al
1286 @item x86-64-*-linux*
1287 @var{list} is a comma separated list of @code{m32}, @code{m64} and
1288 @code{mx32} to enable 32-bit, 64-bit and x32 run-time libraries,
1289 respectively. If @var{list} is empty, then there will be no multilibs
1290 and only the default run-time library will be enabled.
1292 If @option{--with-multilib-list} is not given, then only 32-bit and
1293 64-bit run-time libraries will be enabled.
1296 @item --with-multilib-generator=@var{config}
1297 Specify what multilibs to build. @var{config} is a semicolon separated list of
1298 values, possibly consisting of a single value. Currently only implemented
1299 for riscv*-*-elf*. The accepted values and meanings are given below.
1302 Every config is constructed with four components: architecture string, ABI,
1303 reuse rule with architecture string and reuse rule with sub-extension.
1305 Example 1: Add multi-lib suppport for rv32i with ilp32.
1310 Example 2: Add multi-lib suppport for rv32i with ilp32 and rv32imafd with ilp32.
1312 rv32i-ilp32--;rv32imafd-ilp32--
1315 Example 3: Add multi-lib suppport for rv32i with ilp32; rv32im with ilp32 and
1316 rv32ic with ilp32 will reuse this multi-lib set.
1318 rv32i-ilp32-rv32im-c
1321 Example 4: Add multi-lib suppport for rv64ima with lp64; rv64imaf with lp64,
1322 rv64imac with lp64 and rv64imafc with lp64 will reuse this multi-lib set.
1324 rv64ima-lp64--f,c,fc
1327 @item --with-endian=@var{endians}
1328 Specify what endians to use.
1329 Currently only implemented for sh*-*-*.
1331 @var{endians} may be one of the following:
1334 Use big endian exclusively.
1336 Use little endian exclusively.
1338 Use big endian by default. Provide a multilib for little endian.
1340 Use little endian by default. Provide a multilib for big endian.
1343 @item --enable-threads
1344 Specify that the target
1345 supports threads. This affects the Objective-C compiler and runtime
1346 library, and exception handling for other languages like C++.
1347 On some systems, this is the default.
1349 In general, the best (and, in many cases, the only known) threading
1350 model available will be configured for use. Beware that on some
1351 systems, GCC has not been taught what threading models are generally
1352 available for the system. In this case, @option{--enable-threads} is an
1353 alias for @option{--enable-threads=single}.
1355 @item --disable-threads
1356 Specify that threading support should be disabled for the system.
1357 This is an alias for @option{--enable-threads=single}.
1359 @item --enable-threads=@var{lib}
1361 @var{lib} is the thread support library. This affects the Objective-C
1362 compiler and runtime library, and exception handling for other languages
1363 like C++. The possibilities for @var{lib} are:
1371 LynxOS thread support.
1373 MIPS SDE thread support.
1375 This is an alias for @samp{single}.
1377 Generic POSIX/Unix98 thread support.
1379 RTEMS thread support.
1381 Disable thread support, should work for all platforms.
1385 VxWorks thread support.
1387 Microsoft Win32 API thread support.
1391 Specify that the target supports TLS (Thread Local Storage). Usually
1392 configure can correctly determine if TLS is supported. In cases where
1393 it guesses incorrectly, TLS can be explicitly enabled or disabled with
1394 @option{--enable-tls} or @option{--disable-tls}. This can happen if
1395 the assembler supports TLS but the C library does not, or if the
1396 assumptions made by the configure test are incorrect.
1399 Specify that the target does not support TLS.
1400 This is an alias for @option{--enable-tls=no}.
1402 @item --disable-tm-clone-registry
1403 Disable TM clone registry in libgcc. It is enabled in libgcc by default.
1404 This option helps to reduce code size for embedded targets which do
1405 not use transactional memory.
1407 @item --with-cpu=@var{cpu}
1408 @itemx --with-cpu-32=@var{cpu}
1409 @itemx --with-cpu-64=@var{cpu}
1410 Specify which cpu variant the compiler should generate code for by default.
1411 @var{cpu} will be used as the default value of the @option{-mcpu=} switch.
1412 This option is only supported on some targets, including ARC, ARM, i386, M68k,
1413 PowerPC, and SPARC@. It is mandatory for ARC@. The @option{--with-cpu-32} and
1414 @option{--with-cpu-64} options specify separate default CPUs for
1415 32-bit and 64-bit modes; these options are only supported for aarch64, i386,
1416 x86-64, PowerPC, and SPARC@.
1418 @item --with-schedule=@var{cpu}
1419 @itemx --with-arch=@var{cpu}
1420 @itemx --with-arch-32=@var{cpu}
1421 @itemx --with-arch-64=@var{cpu}
1422 @itemx --with-tune=@var{cpu}
1423 @itemx --with-tune-32=@var{cpu}
1424 @itemx --with-tune-64=@var{cpu}
1425 @itemx --with-abi=@var{abi}
1426 @itemx --with-fpu=@var{type}
1427 @itemx --with-float=@var{type}
1428 These configure options provide default values for the @option{-mschedule=},
1429 @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=}
1430 options and for @option{-mhard-float} or @option{-msoft-float}. As with
1431 @option{--with-cpu}, which switches will be accepted and acceptable values
1432 of the arguments depend on the target.
1434 @item --with-mode=@var{mode}
1435 Specify if the compiler should default to @option{-marm} or @option{-mthumb}.
1436 This option is only supported on ARM targets.
1438 @item --with-stack-offset=@var{num}
1439 This option sets the default for the -mstack-offset=@var{num} option,
1440 and will thus generally also control the setting of this option for
1441 libraries. This option is only supported on Epiphany targets.
1443 @item --with-fpmath=@var{isa}
1444 This options sets @option{-mfpmath=sse} by default and specifies the default
1445 ISA for floating-point arithmetics. You can select either @samp{sse} which
1446 enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default.
1447 This option is only supported on i386 and x86-64 targets.
1449 @item --with-fp-32=@var{mode}
1450 On MIPS targets, set the default value for the @option{-mfp} option when using
1451 the o32 ABI. The possibilities for @var{mode} are:
1454 Use the o32 FP32 ABI extension, as with the @option{-mfp32} command-line
1457 Use the o32 FPXX ABI extension, as with the @option{-mfpxx} command-line
1460 Use the o32 FP64 ABI extension, as with the @option{-mfp64} command-line
1463 In the absence of this configuration option the default is to use the o32
1466 @item --with-odd-spreg-32
1467 On MIPS targets, set the @option{-modd-spreg} option by default when using
1470 @item --without-odd-spreg-32
1471 On MIPS targets, set the @option{-mno-odd-spreg} option by default when using
1472 the o32 ABI. This is normally used in conjunction with
1473 @option{--with-fp-32=64} in order to target the o32 FP64A ABI extension.
1475 @item --with-nan=@var{encoding}
1476 On MIPS targets, set the default encoding convention to use for the
1477 special not-a-number (NaN) IEEE 754 floating-point data. The
1478 possibilities for @var{encoding} are:
1481 Use the legacy encoding, as with the @option{-mnan=legacy} command-line
1484 Use the 754-2008 encoding, as with the @option{-mnan=2008} command-line
1487 To use this configuration option you must have an assembler version
1488 installed that supports the @option{-mnan=} command-line option too.
1489 In the absence of this configuration option the default convention is
1490 the legacy encoding, as when neither of the @option{-mnan=2008} and
1491 @option{-mnan=legacy} command-line options has been used.
1493 @item --with-divide=@var{type}
1494 Specify how the compiler should generate code for checking for
1495 division by zero. This option is only supported on the MIPS target.
1496 The possibilities for @var{type} are:
1499 Division by zero checks use conditional traps (this is the default on
1500 systems that support conditional traps).
1502 Division by zero checks use the break instruction.
1505 @c If you make --with-llsc the default for additional targets,
1506 @c update the --with-llsc description in the MIPS section below.
1509 On MIPS targets, make @option{-mllsc} the default when no
1510 @option{-mno-llsc} option is passed. This is the default for
1511 Linux-based targets, as the kernel will emulate them if the ISA does
1514 @item --without-llsc
1515 On MIPS targets, make @option{-mno-llsc} the default when no
1516 @option{-mllsc} option is passed.
1519 On MIPS targets, make @option{-msynci} the default when no
1520 @option{-mno-synci} option is passed.
1522 @item --without-synci
1523 On MIPS targets, make @option{-mno-synci} the default when no
1524 @option{-msynci} option is passed. This is the default.
1526 @item --with-lxc1-sxc1
1527 On MIPS targets, make @option{-mlxc1-sxc1} the default when no
1528 @option{-mno-lxc1-sxc1} option is passed. This is the default.
1530 @item --without-lxc1-sxc1
1531 On MIPS targets, make @option{-mno-lxc1-sxc1} the default when no
1532 @option{-mlxc1-sxc1} option is passed. The indexed load/store
1533 instructions are not directly a problem but can lead to unexpected
1534 behaviour when deployed in an application intended for a 32-bit address
1535 space but run on a 64-bit processor. The issue is seen because all
1536 known MIPS 64-bit Linux kernels execute o32 and n32 applications
1537 with 64-bit addressing enabled which affects the overflow behaviour
1538 of the indexed addressing mode. GCC will assume that ordinary
1539 32-bit arithmetic overflow behaviour is the same whether performed
1540 as an @code{addu} instruction or as part of the address calculation
1541 in @code{lwxc1} type instructions. This assumption holds true in a
1542 pure 32-bit environment and can hold true in a 64-bit environment if
1543 the address space is accurately set to be 32-bit for o32 and n32.
1546 On MIPS targets, make @option{-mmadd4} the default when no
1547 @option{-mno-madd4} option is passed. This is the default.
1549 @item --without-madd4
1550 On MIPS targets, make @option{-mno-madd4} the default when no
1551 @option{-mmadd4} option is passed. The @code{madd4} instruction
1552 family can be problematic when targeting a combination of cores that
1553 implement these instructions differently. There are two known cores
1554 that implement these as fused operations instead of unfused (where
1555 unfused is normally expected). Disabling these instructions is the
1556 only way to ensure compatible code is generated; this will incur
1557 a performance penalty.
1559 @item --with-mips-plt
1560 On MIPS targets, make use of copy relocations and PLTs.
1561 These features are extensions to the traditional
1562 SVR4-based MIPS ABIs and require support from GNU binutils
1563 and the runtime C library.
1565 @item --with-stack-clash-protection-guard-size=@var{size}
1566 On certain targets this option sets the default stack clash protection guard
1567 size as a power of two in bytes. On AArch64 @var{size} is required to be either
1568 12 (4KB) or 16 (64KB).
1570 @item --enable-__cxa_atexit
1571 Define if you want to use __cxa_atexit, rather than atexit, to
1572 register C++ destructors for local statics and global objects.
1573 This is essential for fully standards-compliant handling of
1574 destructors, but requires __cxa_atexit in libc. This option is currently
1575 only available on systems with GNU libc. When enabled, this will cause
1576 @option{-fuse-cxa-atexit} to be passed by default.
1578 @item --enable-gnu-indirect-function
1579 Define if you want to enable the @code{ifunc} attribute. This option is
1580 currently only available on systems with GNU libc on certain targets.
1582 @item --enable-target-optspace
1584 libraries should be optimized for code space instead of code speed.
1585 This is the default for the m32r platform.
1587 @item --with-cpp-install-dir=@var{dirname}
1588 Specify that the user visible @command{cpp} program should be installed
1589 in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}.
1591 @item --enable-comdat
1592 Enable COMDAT group support. This is primarily used to override the
1593 automatically detected value.
1595 @item --enable-initfini-array
1596 Force the use of sections @code{.init_array} and @code{.fini_array}
1597 (instead of @code{.init} and @code{.fini}) for constructors and
1598 destructors. Option @option{--disable-initfini-array} has the
1599 opposite effect. If neither option is specified, the configure script
1600 will try to guess whether the @code{.init_array} and
1601 @code{.fini_array} sections are supported and, if they are, use them.
1603 @item --enable-link-mutex
1604 When building GCC, use a mutex to avoid linking the compilers for
1605 multiple languages at the same time, to avoid thrashing on build
1606 systems with limited free memory. The default is not to use such a mutex.
1608 @item --enable-link-serialization
1609 When building GCC, use make dependencies to serialize linking the compilers for
1610 multiple languages, to avoid thrashing on build
1611 systems with limited free memory. The default is not to add such
1612 dependencies and thus with parallel make potentially link different
1613 compilers concurrently. If the argument is a positive integer, allow
1614 that number of concurrent link processes for the large binaries.
1616 @item --enable-maintainer-mode
1617 The build rules that regenerate the Autoconf and Automake output files as
1618 well as the GCC master message catalog @file{gcc.pot} are normally
1619 disabled. This is because it can only be rebuilt if the complete source
1620 tree is present. If you have changed the sources and want to rebuild the
1621 catalog, configuring with @option{--enable-maintainer-mode} will enable
1622 this. Note that you need a recent version of the @code{gettext} tools
1625 @item --disable-bootstrap
1626 For a native build, the default configuration is to perform
1627 a 3-stage bootstrap of the compiler when @samp{make} is invoked,
1628 testing that GCC can compile itself correctly. If you want to disable
1629 this process, you can configure with @option{--disable-bootstrap}.
1631 @item --enable-bootstrap
1632 In special cases, you may want to perform a 3-stage build
1633 even if the target and host triplets are different.
1634 This is possible when the host can run code compiled for
1635 the target (e.g.@: host is i686-linux, target is i486-linux).
1636 Starting from GCC 4.2, to do this you have to configure explicitly
1637 with @option{--enable-bootstrap}.
1639 @item --enable-generated-files-in-srcdir
1640 Neither the .c and .h files that are generated from Bison and flex nor the
1641 info manuals and man pages that are built from the .texi files are present
1642 in the repository development tree. When building GCC from that development tree,
1643 or from one of our snapshots, those generated files are placed in your
1644 build directory, which allows for the source to be in a readonly
1647 If you configure with @option{--enable-generated-files-in-srcdir} then those
1648 generated files will go into the source directory. This is mainly intended
1649 for generating release or prerelease tarballs of the GCC sources, since it
1650 is not a requirement that the users of source releases to have flex, Bison,
1653 @item --enable-version-specific-runtime-libs
1655 that runtime libraries should be installed in the compiler specific
1656 subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In
1657 addition, @samp{libstdc++}'s include files will be installed into
1658 @file{@var{libdir}} unless you overruled it by using
1659 @option{--with-gxx-include-dir=@var{dirname}}. Using this option is
1660 particularly useful if you intend to use several versions of GCC in
1661 parallel. The default is @samp{yes} for @samp{libada}, and @samp{no} for
1662 the remaining libraries.
1664 @item @anchor{WithAixSoname}--with-aix-soname=@samp{aix}, @samp{svr4} or @samp{both}
1665 Traditional AIX shared library versioning (versioned @code{Shared Object}
1666 files as members of unversioned @code{Archive Library} files named
1667 @samp{lib.a}) causes numerous headaches for package managers. However,
1668 @code{Import Files} as members of @code{Archive Library} files allow for
1669 @strong{filename-based versioning} of shared libraries as seen on Linux/SVR4,
1670 where this is called the "SONAME". But as they prevent static linking,
1671 @code{Import Files} may be used with @code{Runtime Linking} only, where the
1672 linker does search for @samp{libNAME.so} before @samp{libNAME.a} library
1673 filenames with the @samp{-lNAME} linker flag.
1675 @anchor{AixLdCommand}For detailed information please refer to the AIX
1676 @uref{https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22,,ld
1679 As long as shared library creation is enabled, upon:
1681 @item --with-aix-soname=aix
1682 @item --with-aix-soname=both
1683 A (traditional AIX) @code{Shared Archive Library} file is created:
1685 @item using the @samp{libNAME.a} filename scheme
1686 @item with the @code{Shared Object} file as archive member named
1687 @samp{libNAME.so.V} (except for @samp{libgcc_s}, where the @code{Shared
1688 Object} file is named @samp{shr.o} for backwards compatibility), which
1690 @item is used for runtime loading from inside the @samp{libNAME.a} file
1691 @item is used for dynamic loading via
1692 @code{dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)}
1693 @item is used for shared linking
1694 @item is used for static linking, so no separate @code{Static Archive
1695 Library} file is needed
1698 @item --with-aix-soname=both
1699 @item --with-aix-soname=svr4
1700 A (second) @code{Shared Archive Library} file is created:
1702 @item using the @samp{libNAME.so.V} filename scheme
1703 @item with the @code{Shared Object} file as archive member named
1706 @item is created with the @code{-G linker flag}
1707 @item has the @code{F_LOADONLY} flag set
1708 @item is used for runtime loading from inside the @samp{libNAME.so.V} file
1709 @item is used for dynamic loading via @code{dlopen("libNAME.so.V(shr.o)",
1712 @item with the @code{Import File} as archive member named @samp{shr.imp},
1715 @item refers to @samp{libNAME.so.V(shr.o)} as the "SONAME", to be recorded
1716 in the @code{Loader Section} of subsequent binaries
1717 @item indicates whether @samp{libNAME.so.V(shr.o)} is 32 or 64 bit
1718 @item lists all the public symbols exported by @samp{lib.so.V(shr.o)},
1719 eventually decorated with the @code{@samp{weak} Keyword}
1720 @item is necessary for shared linking against @samp{lib.so.V(shr.o)}
1723 A symbolic link using the @samp{libNAME.so} filename scheme is created:
1725 @item pointing to the @samp{libNAME.so.V} @code{Shared Archive Library} file
1726 @item to permit the @code{ld Command} to find @samp{lib.so.V(shr.imp)} via
1727 the @samp{-lNAME} argument (requires @code{Runtime Linking} to be enabled)
1728 @item to permit dynamic loading of @samp{lib.so.V(shr.o)} without the need
1729 to specify the version number via @code{dlopen("libNAME.so(shr.o)",
1734 As long as static library creation is enabled, upon:
1736 @item --with-aix-soname=svr4
1737 A @code{Static Archive Library} is created:
1739 @item using the @samp{libNAME.a} filename scheme
1740 @item with all the @code{Static Object} files as archive members, which
1742 @item are used for static linking
1747 While the aix-soname=@samp{svr4} option does not create @code{Shared Object}
1748 files as members of unversioned @code{Archive Library} files any more, package
1749 managers still are responsible to
1750 @uref{./specific.html#TransferAixShobj,,transfer} @code{Shared Object} files
1751 found as member of a previously installed unversioned @code{Archive Library}
1752 file into the newly installed @code{Archive Library} file with the same
1755 @emph{WARNING:} Creating @code{Shared Object} files with @code{Runtime Linking}
1756 enabled may bloat the TOC, eventually leading to @code{TOC overflow} errors,
1757 requiring the use of either the @option{-Wl,-bbigtoc} linker flag (seen to
1758 break with the @code{GDB} debugger) or some of the TOC-related compiler flags,
1760 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1761 Using the GNU Compiler Collection (GCC)}.
1764 see ``RS/6000 and PowerPC Options'' in the main manual.
1767 @option{--with-aix-soname} is currently supported by @samp{libgcc_s} only, so
1768 this option is still experimental and not for normal use yet.
1770 Default is the traditional behavior @option{--with-aix-soname=@samp{aix}}.
1772 @item --enable-languages=@var{lang1},@var{lang2},@dots{}
1773 Specify that only a particular subset of compilers and
1774 their runtime libraries should be built. For a list of valid values for
1775 @var{langN} you can issue the following command in the
1776 @file{gcc} directory of your GCC source tree:@*
1778 grep ^language= */config-lang.in
1780 Currently, you can use any of the following:
1781 @code{all}, @code{default}, @code{ada}, @code{c}, @code{c++}, @code{d},
1782 @code{fortran}, @code{go}, @code{jit}, @code{lto}, @code{objc}, @code{obj-c++}.
1783 Building the Ada compiler has special requirements, see below.
1784 If you do not pass this flag, or specify the option @code{default}, then the
1785 default languages available in the @file{gcc} sub-tree will be configured.
1786 Ada, D, Go, Jit, and Objective-C++ are not default languages. LTO is not a
1787 default language, but is built by default because @option{--enable-lto} is
1788 enabled by default. The other languages are default languages. If
1789 @code{all} is specified, then all available languages are built. An
1790 exception is @code{jit} language, which requires
1791 @option{--enable-host-shared} to be included with @code{all}.
1793 @item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{}
1794 Specify that a particular subset of compilers and their runtime
1795 libraries should be built with the system C compiler during stage 1 of
1796 the bootstrap process, rather than only in later stages with the
1797 bootstrapped C compiler. The list of valid values is the same as for
1798 @option{--enable-languages}, and the option @code{all} will select all
1799 of the languages enabled by @option{--enable-languages}. This option is
1800 primarily useful for GCC development; for instance, when a development
1801 version of the compiler cannot bootstrap due to compiler bugs, or when
1802 one is debugging front ends other than the C front end. When this
1803 option is used, one can then build the target libraries for the
1804 specified languages with the stage-1 compiler by using @command{make
1805 stage1-bubble all-target}, or run the testsuite on the stage-1 compiler
1806 for the specified languages using @command{make stage1-start check-gcc}.
1808 @item --disable-libada
1809 Specify that the run-time libraries and tools used by GNAT should not
1810 be built. This can be useful for debugging, or for compatibility with
1811 previous Ada build procedures, when it was required to explicitly
1812 do a @samp{make -C gcc gnatlib_and_tools}.
1814 @item --disable-libsanitizer
1815 Specify that the run-time libraries for the various sanitizers should
1818 @item --disable-libssp
1819 Specify that the run-time libraries for stack smashing protection
1820 should not be built or linked against. On many targets library support
1821 is provided by the C library instead.
1823 @item --disable-libquadmath
1824 Specify that the GCC quad-precision math library should not be built.
1825 On some systems, the library is required to be linkable when building
1826 the Fortran front end, unless @option{--disable-libquadmath-support}
1829 @item --disable-libquadmath-support
1830 Specify that the Fortran front end and @code{libgfortran} do not add
1831 support for @code{libquadmath} on systems supporting it.
1833 @item --disable-libgomp
1834 Specify that the GNU Offloading and Multi Processing Runtime Library
1835 should not be built.
1837 @item --disable-libvtv
1838 Specify that the run-time libraries used by vtable verification
1839 should not be built.
1842 Specify that the compiler should
1843 use DWARF 2 debugging information as the default.
1845 @item --with-advance-toolchain=@var{at}
1846 On 64-bit PowerPC Linux systems, configure the compiler to use the
1847 header files, library files, and the dynamic linker from the Advance
1848 Toolchain release @var{at} instead of the default versions that are
1849 provided by the Linux distribution. In general, this option is
1850 intended for the developers of GCC, and it is not intended for general
1853 @item --enable-targets=all
1854 @itemx --enable-targets=@var{target_list}
1855 Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers.
1856 These are compilers that are able to generate either 64-bit or 32-bit
1857 code. Typically, the corresponding 32-bit target, e.g.@:
1858 powerpc-linux for powerpc64-linux, only generates 32-bit code. This
1859 option enables the 32-bit target to be a bi-arch compiler, which is
1860 useful when you want a bi-arch compiler that defaults to 32-bit, and
1861 you are building a bi-arch or multi-arch binutils in a combined tree.
1862 On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64),
1864 Currently, this option only affects sparc-linux, powerpc-linux, x86-linux,
1865 mips-linux and s390-linux.
1867 @item --enable-default-pie
1868 Turn on @option{-fPIE} and @option{-pie} by default.
1870 @item --enable-secureplt
1871 This option enables @option{-msecure-plt} by default for powerpc-linux.
1873 @xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc,
1874 Using the GNU Compiler Collection (GCC)},
1877 See ``RS/6000 and PowerPC Options'' in the main manual
1880 @item --enable-default-ssp
1881 Turn on @option{-fstack-protector-strong} by default.
1884 This option enables @option{-mcld} by default for 32-bit x86 targets.
1886 @xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc,
1887 Using the GNU Compiler Collection (GCC)},
1890 See ``i386 and x86-64 Options'' in the main manual
1893 @item --enable-large-address-aware
1894 The @option{--enable-large-address-aware} option arranges for MinGW
1895 executables to be linked using the @option{--large-address-aware}
1896 option, that enables the use of more than 2GB of memory. If GCC is
1897 configured with this option, its effects can be reversed by passing the
1898 @option{-Wl,--disable-large-address-aware} option to the so-configured
1901 @item --enable-win32-registry
1902 @itemx --enable-win32-registry=@var{key}
1903 @itemx --disable-win32-registry
1904 The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC
1905 to look up installations paths in the registry using the following key:
1908 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}}
1911 @var{key} defaults to GCC version number, and can be overridden by the
1912 @option{--enable-win32-registry=@var{key}} option. Vendors and distributors
1913 who use custom installers are encouraged to provide a different key,
1914 perhaps one comprised of vendor name and GCC version number, to
1915 avoid conflict with existing installations. This feature is enabled
1916 by default, and can be disabled by @option{--disable-win32-registry}
1917 option. This option has no effect on the other hosts.
1920 Specify that the machine does not have a floating point unit. This
1921 option only applies to @samp{m68k-sun-sunos@var{n}}. On any other
1922 system, @option{--nfp} has no effect.
1924 @item --enable-werror
1925 @itemx --disable-werror
1926 @itemx --enable-werror=yes
1927 @itemx --enable-werror=no
1928 When you specify this option, it controls whether certain files in the
1929 compiler are built with @option{-Werror} in bootstrap stage2 and later.
1930 If you don't specify it, @option{-Werror} is turned on for the main
1931 development trunk. However it defaults to off for release branches and
1932 final releases. The specific files which get @option{-Werror} are
1933 controlled by the Makefiles.
1935 @item --enable-checking
1936 @itemx --disable-checking
1937 @itemx --enable-checking=@var{list}
1938 This option controls performing internal consistency checks in the compiler.
1939 It does not change the generated code, but adds error checking of the
1940 requested complexity. This slows down the compiler and may only work
1941 properly if you are building the compiler with GCC@.
1943 When the option is not specified, the active set of checks depends on context.
1944 Namely, bootstrap stage 1 defaults to @samp{--enable-checking=yes}, builds
1945 from release branches or release archives default to
1946 @samp{--enable-checking=release}, and otherwise
1947 @samp{--enable-checking=yes,extra} is used. When the option is
1948 specified without a @var{list}, the result is the same as
1949 @samp{--enable-checking=yes}. Likewise, @samp{--disable-checking} is
1950 equivalent to @samp{--enable-checking=no}.
1952 The categories of checks available in @var{list} are @samp{yes} (most common
1953 checks @samp{assert,misc,gc,gimple,rtlflag,runtime,tree,types}), @samp{no}
1954 (no checks at all), @samp{all} (all but @samp{valgrind}), @samp{release}
1955 (cheapest checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}).
1956 @samp{release} checks are always on and to disable them
1957 @samp{--disable-checking} or @samp{--enable-checking=no[,<other checks>]}
1958 must be explicitly requested. Disabling assertions makes the compiler and
1959 runtime slightly faster but increases the risk of undetected internal errors
1960 causing wrong code to be generated.
1962 Individual checks can be enabled with these flags: @samp{assert}, @samp{df},
1963 @samp{extra}, @samp{fold}, @samp{gc}, @samp{gcac}, @samp{gimple},
1964 @samp{misc}, @samp{rtl}, @samp{rtlflag}, @samp{runtime}, @samp{tree},
1965 @samp{types} and @samp{valgrind}. @samp{extra} extends @samp{misc}
1966 checking with extra checks that might affect code generation and should
1967 therefore not differ between stage1 and later stages in bootstrap.
1969 The @samp{valgrind} check requires the external @command{valgrind} simulator,
1970 available from @uref{https://valgrind.org}. The @samp{rtl} checks are
1971 expensive and the @samp{df}, @samp{gcac} and @samp{valgrind} checks are very
1974 @item --disable-stage1-checking
1975 @itemx --enable-stage1-checking
1976 @itemx --enable-stage1-checking=@var{list}
1977 This option affects only bootstrap build. If no @option{--enable-checking}
1978 option is specified the stage1 compiler is built with @samp{yes} checking
1979 enabled, otherwise the stage1 checking flags are the same as specified by
1980 @option{--enable-checking}. To build the stage1 compiler with
1981 different checking options use @option{--enable-stage1-checking}.
1982 The list of checking options is the same as for @option{--enable-checking}.
1983 If your system is too slow or too small to bootstrap a released compiler
1984 with checking for stage1 enabled, you can use @samp{--disable-stage1-checking}
1985 to disable checking for the stage1 compiler.
1987 @item --enable-coverage
1988 @itemx --enable-coverage=@var{level}
1989 With this option, the compiler is built to collect self coverage
1990 information, every time it is run. This is for internal development
1991 purposes, and only works when the compiler is being built with gcc. The
1992 @var{level} argument controls whether the compiler is built optimized or
1993 not, values are @samp{opt} and @samp{noopt}. For coverage analysis you
1994 want to disable optimization, for performance analysis you want to
1995 enable optimization. When coverage is enabled, the default level is
1996 without optimization.
1998 @item --enable-gather-detailed-mem-stats
1999 When this option is specified more detailed information on memory
2000 allocation is gathered. This information is printed when using
2001 @option{-fmem-report}.
2003 @item --enable-valgrind-annotations
2004 Mark selected memory related operations in the compiler when run under
2005 valgrind to suppress false positives.
2008 @itemx --disable-nls
2009 The @option{--enable-nls} option enables Native Language Support (NLS),
2010 which lets GCC output diagnostics in languages other than American
2011 English. Native Language Support is enabled by default if not doing a
2012 canadian cross build. The @option{--disable-nls} option disables NLS@.
2014 @item --with-included-gettext
2015 If NLS is enabled, the @option{--with-included-gettext} option causes the build
2016 procedure to prefer its copy of GNU @command{gettext}.
2018 @item --with-catgets
2019 If NLS is enabled, and if the host lacks @code{gettext} but has the
2020 inferior @code{catgets} interface, the GCC build procedure normally
2021 ignores @code{catgets} and instead uses GCC's copy of the GNU
2022 @code{gettext} library. The @option{--with-catgets} option causes the
2023 build procedure to use the host's @code{catgets} in this situation.
2025 @item --with-libiconv-prefix=@var{dir}
2026 Search for libiconv header files in @file{@var{dir}/include} and
2027 libiconv library files in @file{@var{dir}/lib}.
2029 @item --enable-obsolete
2030 Enable configuration for an obsoleted system. If you attempt to
2031 configure GCC for a system (build, host, or target) which has been
2032 obsoleted, and you do not specify this flag, configure will halt with an
2035 All support for systems which have been obsoleted in one release of GCC
2036 is removed entirely in the next major release, unless someone steps
2037 forward to maintain the port.
2039 @item --enable-decimal-float
2040 @itemx --enable-decimal-float=yes
2041 @itemx --enable-decimal-float=no
2042 @itemx --enable-decimal-float=bid
2043 @itemx --enable-decimal-float=dpd
2044 @itemx --disable-decimal-float
2045 Enable (or disable) support for the C decimal floating point extension
2046 that is in the IEEE 754-2008 standard. This is enabled by default only
2047 on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also
2048 support it, but require the user to specifically enable it. You can
2049 optionally control which decimal floating point format is used (either
2050 @samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal)
2051 format is default on i386 and x86_64 systems, and the @samp{dpd}
2052 (densely packed decimal) format is default on PowerPC systems.
2054 @item --enable-fixed-point
2055 @itemx --disable-fixed-point
2056 Enable (or disable) support for C fixed-point arithmetic.
2057 This option is enabled by default for some targets (such as MIPS) which
2058 have hardware-support for fixed-point operations. On other targets, you
2059 may enable this option manually.
2061 @item --with-long-double-128
2062 Specify if @code{long double} type should be 128-bit by default on selected
2063 GNU/Linux architectures. If using @code{--without-long-double-128},
2064 @code{long double} will be by default 64-bit, the same as @code{double} type.
2065 When neither of these configure options are used, the default will be
2066 128-bit @code{long double} when built against GNU C Library 2.4 and later,
2067 64-bit @code{long double} otherwise.
2069 @item --with-long-double-format=ibm
2070 @itemx --with-long-double-format=ieee
2071 Specify whether @code{long double} uses the IBM extended double format
2072 or the IEEE 128-bit floating point format on PowerPC Linux systems.
2073 This configuration switch will only work on little endian PowerPC
2074 Linux systems and on big endian 64-bit systems where the default cpu
2075 is at least power7 (i.e.@: @option{--with-cpu=power7},
2076 @option{--with-cpu=power8}, or @option{--with-cpu=power9} is used).
2078 If you use the @option{--with-long-double-64} configuration option,
2079 the @option{--with-long-double-format=ibm} and
2080 @option{--with-long-double-format=ieee} options are ignored.
2082 The default @code{long double} format is to use IBM extended double.
2083 Until all of the libraries are converted to use IEEE 128-bit floating
2084 point, it is not recommended to use
2085 @option{--with-long-double-format=ieee}.
2087 On little endian PowerPC Linux systems, if you explicitly set the
2088 @code{long double} type, it will build multilibs to allow you to
2089 select either @code{long double} format, unless you disable multilibs
2090 with the @code{--disable-multilib} option. At present,
2091 @code{long double} multilibs are not built on big endian PowerPC Linux
2092 systems. If you are building multilibs, you will need to configure
2093 the compiler using the @option{--with-system-zlib} option.
2095 If you do not set the @code{long double} type explicitly, no multilibs
2098 @item --enable-fdpic
2099 On SH Linux systems, generate ELF FDPIC code.
2101 @item --with-gmp=@var{pathname}
2102 @itemx --with-gmp-include=@var{pathname}
2103 @itemx --with-gmp-lib=@var{pathname}
2104 @itemx --with-mpfr=@var{pathname}
2105 @itemx --with-mpfr-include=@var{pathname}
2106 @itemx --with-mpfr-lib=@var{pathname}
2107 @itemx --with-mpc=@var{pathname}
2108 @itemx --with-mpc-include=@var{pathname}
2109 @itemx --with-mpc-lib=@var{pathname}
2110 If you want to build GCC but do not have the GMP library, the MPFR
2111 library and/or the MPC library installed in a standard location and
2112 do not have their sources present in the GCC source tree then you
2113 can explicitly specify the directory where they are installed
2114 (@samp{--with-gmp=@var{gmpinstalldir}},
2115 @samp{--with-mpfr=@/@var{mpfrinstalldir}},
2116 @samp{--with-mpc=@/@var{mpcinstalldir}}). The
2117 @option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for
2118 @option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and
2119 @option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the
2120 @option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for
2121 @option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and
2122 @option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the
2123 @option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for
2124 @option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and
2125 @option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these
2126 shorthand assumptions are not correct, you can use the explicit
2127 include and lib options directly. You might also need to ensure the
2128 shared libraries can be found by the dynamic linker when building and
2129 using GCC, for example by setting the runtime shared library path
2130 variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems).
2132 These flags are applicable to the host platform only. When building
2133 a cross compiler, they will not be used to configure target libraries.
2135 @item --with-isl=@var{pathname}
2136 @itemx --with-isl-include=@var{pathname}
2137 @itemx --with-isl-lib=@var{pathname}
2138 If you do not have the isl library installed in a standard location and you
2139 want to build GCC, you can explicitly specify the directory where it is
2140 installed (@samp{--with-isl=@/@var{islinstalldir}}). The
2141 @option{--with-isl=@/@var{islinstalldir}} option is shorthand for
2142 @option{--with-isl-lib=@/@var{islinstalldir}/lib} and
2143 @option{--with-isl-include=@/@var{islinstalldir}/include}. If this
2144 shorthand assumption is not correct, you can use the explicit
2145 include and lib options directly.
2147 These flags are applicable to the host platform only. When building
2148 a cross compiler, they will not be used to configure target libraries.
2150 @item --with-stage1-ldflags=@var{flags}
2151 This option may be used to set linker flags to be used when linking
2152 stage 1 of GCC. These are also used when linking GCC if configured with
2153 @option{--disable-bootstrap}. If @option{--with-stage1-libs} is not set to a
2154 value, then the default is @samp{-static-libstdc++ -static-libgcc}, if
2157 @item --with-stage1-libs=@var{libs}
2158 This option may be used to set libraries to be used when linking stage 1
2159 of GCC. These are also used when linking GCC if configured with
2160 @option{--disable-bootstrap}.
2162 @item --with-boot-ldflags=@var{flags}
2163 This option may be used to set linker flags to be used when linking
2164 stage 2 and later when bootstrapping GCC. If --with-boot-libs
2165 is not is set to a value, then the default is
2166 @samp{-static-libstdc++ -static-libgcc}.
2168 @item --with-boot-libs=@var{libs}
2169 This option may be used to set libraries to be used when linking stage 2
2170 and later when bootstrapping GCC.
2172 @item --with-debug-prefix-map=@var{map}
2173 Convert source directory names using @option{-fdebug-prefix-map} when
2174 building runtime libraries. @samp{@var{map}} is a space-separated
2175 list of maps of the form @samp{@var{old}=@var{new}}.
2177 @item --enable-linker-build-id
2178 Tells GCC to pass @option{--build-id} option to the linker for all final
2179 links (links performed without the @option{-r} or @option{--relocatable}
2180 option), if the linker supports it. If you specify
2181 @option{--enable-linker-build-id}, but your linker does not
2182 support @option{--build-id} option, a warning is issued and the
2183 @option{--enable-linker-build-id} option is ignored. The default is off.
2185 @item --with-linker-hash-style=@var{choice}
2186 Tells GCC to pass @option{--hash-style=@var{choice}} option to the
2187 linker for all final links. @var{choice} can be one of
2188 @samp{sysv}, @samp{gnu}, and @samp{both} where @samp{sysv} is the default.
2190 @item --enable-gnu-unique-object
2191 @itemx --disable-gnu-unique-object
2192 Tells GCC to use the gnu_unique_object relocation for C++ template
2193 static data members and inline function local statics. Enabled by
2194 default for a toolchain with an assembler that accepts it and
2195 GLIBC 2.11 or above, otherwise disabled.
2197 @item --with-diagnostics-color=@var{choice}
2198 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-color=}
2199 option (if not used explicitly on the command line). @var{choice}
2200 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2201 where @samp{auto} is the default. @samp{auto-if-env} makes
2202 @option{-fdiagnostics-color=auto} the default if @env{GCC_COLORS}
2203 is present and non-empty in the environment of the compiler, and
2204 @option{-fdiagnostics-color=never} otherwise.
2206 @item --with-diagnostics-urls=@var{choice}
2207 Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-urls=}
2208 option (if not used explicitly on the command line). @var{choice}
2209 can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env}
2210 where @samp{auto} is the default. @samp{auto-if-env} makes
2211 @option{-fdiagnostics-urls=auto} the default if @env{GCC_URLS}
2212 or @env{TERM_URLS} is present and non-empty in the environment of the
2213 compiler, and @option{-fdiagnostics-urls=never} otherwise.
2216 @itemx --disable-lto
2217 Enable support for link-time optimization (LTO). This is enabled by
2218 default, and may be disabled using @option{--disable-lto}.
2220 @item --enable-linker-plugin-configure-flags=FLAGS
2221 @itemx --enable-linker-plugin-flags=FLAGS
2222 By default, linker plugins (such as the LTO plugin) are built for the
2223 host system architecture. For the case that the linker has a
2224 different (but run-time compatible) architecture, these flags can be
2225 specified to build plugins that are compatible to the linker. For
2226 example, if you are building GCC for a 64-bit x86_64
2227 (@samp{x86_64-pc-linux-gnu}) host system, but have a 32-bit x86
2228 GNU/Linux (@samp{i686-pc-linux-gnu}) linker executable (which is
2229 executable on the former system), you can configure GCC as follows for
2230 getting compatible linker plugins:
2233 % @var{srcdir}/configure \
2234 --host=x86_64-pc-linux-gnu \
2235 --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \
2236 --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib'
2239 @item --with-plugin-ld=@var{pathname}
2240 Enable an alternate linker to be used at link-time optimization (LTO)
2241 link time when @option{-fuse-linker-plugin} is enabled.
2242 This linker should have plugin support such as gold starting with
2243 version 2.20 or GNU ld starting with version 2.21.
2244 See @option{-fuse-linker-plugin} for details.
2246 @item --enable-canonical-system-headers
2247 @itemx --disable-canonical-system-headers
2248 Enable system header path canonicalization for @file{libcpp}. This can
2249 produce shorter header file paths in diagnostics and dependency output
2250 files, but these changed header paths may conflict with some compilation
2251 environments. Enabled by default, and may be disabled using
2252 @option{--disable-canonical-system-headers}.
2254 @item --with-glibc-version=@var{major}.@var{minor}
2255 Tell GCC that when the GNU C Library (glibc) is used on the target it
2256 will be version @var{major}.@var{minor} or later. Normally this can
2257 be detected from the C library's header files, but this option may be
2258 needed when bootstrapping a cross toolchain without the header files
2259 available for building the initial bootstrap compiler.
2261 If GCC is configured with some multilibs that use glibc and some that
2262 do not, this option applies only to the multilibs that use glibc.
2263 However, such configurations may not work well as not all the relevant
2264 configuration in GCC is on a per-multilib basis.
2266 @item --enable-as-accelerator-for=@var{target}
2267 Build as offload target compiler. Specify offload host triple by @var{target}.
2269 @item --enable-offload-targets=@var{target1}[=@var{path1}],@dots{},@var{targetN}[=@var{pathN}]
2270 Enable offloading to targets @var{target1}, @dots{}, @var{targetN}.
2271 Offload compilers are expected to be already installed. Default search
2272 path for them is @file{@var{exec-prefix}}, but it can be changed by
2273 specifying paths @var{path1}, @dots{}, @var{pathN}.
2276 % @var{srcdir}/configure \
2277 --enable-offload-targets=x86_64-intelmicemul-linux-gnu=/path/to/x86_64/compiler,nvptx-none
2280 @item --enable-offload-defaulted
2282 Tell GCC that configured but not installed offload compilers and libgomp
2283 plugins are silently ignored. Useful for distribution compilers where
2284 those are in separate optional packages and where the presence or absence
2285 of those optional packages should determine the actual supported offloading
2286 target set rather than the GCC configure-time selection.
2288 @item --with-hsa-runtime=@var{pathname}
2289 @itemx --with-hsa-runtime-include=@var{pathname}
2290 @itemx --with-hsa-runtime-lib=@var{pathname}
2292 If you configure GCC with offloading which uses an HSA run-time such as
2293 AMDGCN but do not have the HSA run-time library installed in a standard
2294 location then you can explicitly specify the directory where they are
2295 installed. The @option{--with-hsa-runtime=@/@var{hsainstalldir}} option
2297 @option{--with-hsa-runtime-lib=@/@var{hsainstalldir}/lib} and
2298 @option{--with-hsa-runtime-include=@/@var{hsainstalldir}/include}.
2301 @itemx --disable-cet
2302 Enable building target run-time libraries with control-flow
2303 instrumentation, see @option{-fcf-protection} option. When
2304 @code{--enable-cet} is specified target libraries are configured
2305 to add @option{-fcf-protection} and, if needed, other target
2306 specific options to a set of building options.
2308 @code{--enable-cet=auto} is default. CET is enabled on Linux/x86 if
2309 target binutils supports @code{Intel CET} instructions and disabled
2310 otherwise. In this case, the target libraries are configured to get
2311 additional @option{-fcf-protection} option.
2313 @item --with-riscv-attribute=@samp{yes}, @samp{no} or @samp{default}
2314 Generate RISC-V attribute by default, in order to record extra build
2315 information in object.
2317 The option is disabled by default. It is enabled on RISC-V/ELF (bare-metal)
2318 target if target binutils supported.
2320 @item --enable-s390-excess-float-precision
2321 @itemx --disable-s390-excess-float-precision
2322 On s390(x) targets, enable treatment of float expressions with double precision
2323 when in standards-compliant mode (e.g., when @code{--std=c99} or
2324 @code{-fexcess-precision=standard} are given).
2326 For a native build and cross compiles that have target headers, the option's
2327 default is derived from glibc's behavior. When glibc clamps float_t to double,
2328 GCC follows and enables the option. For other cross compiles, the default is
2332 @subheading Cross-Compiler-Specific Options
2333 The following options only apply to building cross compilers.
2336 @item --with-toolexeclibdir=@var{dir}
2337 Specify the installation directory for libraries built with a cross compiler.
2338 The default is @option{$@{gcc_tooldir@}/lib}.
2340 @item --with-sysroot
2341 @itemx --with-sysroot=@var{dir}
2342 Tells GCC to consider @var{dir} as the root of a tree that contains
2343 (a subset of) the root filesystem of the target operating system.
2344 Target system headers, libraries and run-time object files will be
2345 searched for in there. More specifically, this acts as if
2346 @option{--sysroot=@var{dir}} was added to the default options of the built
2347 compiler. The specified directory is not copied into the
2348 install tree, unlike the options @option{--with-headers} and
2349 @option{--with-libs} that this option obsoletes. The default value,
2350 in case @option{--with-sysroot} is not given an argument, is
2351 @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a
2352 subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to
2353 the GCC binaries if the installation tree is moved.
2355 This option affects the system root for the compiler used to build
2356 target libraries (which runs on the build system) and the compiler newly
2357 installed with @code{make install}; it does not affect the compiler which is
2358 used to build GCC itself.
2360 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2361 option then the compiler will search that directory within @var{dirname} for
2362 native system headers rather than the default @file{/usr/include}.
2364 @item --with-build-sysroot
2365 @itemx --with-build-sysroot=@var{dir}
2366 Tells GCC to consider @var{dir} as the system root (see
2367 @option{--with-sysroot}) while building target libraries, instead of
2368 the directory specified with @option{--with-sysroot}. This option is
2369 only useful when you are already using @option{--with-sysroot}. You
2370 can use @option{--with-build-sysroot} when you are configuring with
2371 @option{--prefix} set to a directory that is different from the one in
2372 which you are installing GCC and your target libraries.
2374 This option affects the system root for the compiler used to build
2375 target libraries (which runs on the build system); it does not affect
2376 the compiler which is used to build GCC itself.
2378 If you specify the @option{--with-native-system-header-dir=@var{dirname}}
2379 option then the compiler will search that directory within @var{dirname} for
2380 native system headers rather than the default @file{/usr/include}.
2382 @item --with-headers
2383 @itemx --with-headers=@var{dir}
2384 Deprecated in favor of @option{--with-sysroot}.
2385 Specifies that target headers are available when building a cross compiler.
2386 The @var{dir} argument specifies a directory which has the target include
2387 files. These include files will be copied into the @file{gcc} install
2388 directory. @emph{This option with the @var{dir} argument is required} when
2389 building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include}
2390 doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does
2391 pre-exist, the @var{dir} argument may be omitted. @command{fixincludes}
2392 will be run on these files to make them compatible with GCC@.
2394 @item --without-headers
2395 Tells GCC not use any target headers from a libc when building a cross
2396 compiler. When crossing to GNU/Linux, you need the headers so GCC
2397 can build the exception handling for libgcc.
2400 @itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}"
2401 Deprecated in favor of @option{--with-sysroot}.
2402 Specifies a list of directories which contain the target runtime
2403 libraries. These libraries will be copied into the @file{gcc} install
2404 directory. If the directory list is omitted, this option has no
2408 Specifies that @samp{newlib} is
2409 being used as the target C library. This causes @code{__eprintf} to be
2410 omitted from @file{libgcc.a} on the assumption that it will be provided by
2416 @item --with-avrlibc
2417 Only supported for the AVR target. Specifies that @samp{AVR-Libc} is
2418 being used as the target C@tie{} library. This causes float support
2419 functions like @code{__addsf3} to be omitted from @file{libgcc.a} on
2420 the assumption that it will be provided by @file{libm.a}. For more
2421 technical details, cf. @uref{http://gcc.gnu.org/PR54461,,PR54461}.
2422 It is not supported for
2423 RTEMS configurations, which currently use newlib. The option is
2424 supported since version 4.7.2 and is the default in 4.8.0 and newer.
2426 @item --with-double=@{32|64|32,64|64,32@}
2427 @itemx --with-long-double=@{32|64|32,64|64,32|double@}
2428 Only supported for the AVR target since version@tie{}10.
2429 Specify the default layout available for the C/C++ @samp{double}
2430 and @samp{long double} type, respectively. The following rules apply:
2433 The first value after the @samp{=} specifies the default layout (in bits)
2434 of the type and also the default for the @option{-mdouble=} resp.
2435 @option{-mlong-double=} compiler option.
2437 If more than one value is specified, respective multilib variants are
2438 available, and @option{-mdouble=} resp. @option{-mlong-double=} acts
2439 as a multilib option.
2441 If @option{--with-long-double=double} is specified, @samp{double} and
2442 @samp{long double} will have the same layout.
2444 The defaults are @option{--with-long-double=64,32} and
2445 @option{--with-double=32,64}. The default @samp{double} layout imposed by
2446 the latter is compatible with older versions of the compiler that implement
2447 @samp{double} as a 32-bit type, which does not comply to the language standard.
2449 Not all combinations of @option{--with-double=} and
2450 @option{--with-long-double=} are valid. For example, the combination
2451 @option{--with-double=32,64} @option{--with-long-double=32} will be
2452 rejected because the first option specifies the availability of
2453 multilibs for @samp{double}, whereas the second option implies
2454 that @samp{long double} --- and hence also @samp{double} --- is always
2457 @item --with-double-comparison=@{tristate|bool|libf7@}
2458 Only supported for the AVR target since version@tie{}10.
2459 Specify what result format is returned by library functions that
2460 compare 64-bit floating point values (@code{DFmode}).
2461 The GCC default is @samp{tristate}. If the floating point
2462 implementation returns a boolean instead, set it to @samp{bool}.
2464 @item --with-libf7=@{libgcc|math|math-symbols|no@}
2465 Only supported for the AVR target since version@tie{}10.
2466 Specify to which degree code from LibF7 is included in libgcc.
2467 LibF7 is an ad-hoc, AVR-specific, 64-bit floating point emulation
2468 written in C and (inline) assembly. @samp{libgcc} adds support
2469 for functions that one would usually expect in libgcc like double addition,
2470 double comparisons and double conversions. @samp{math} also adds routines
2471 that one would expect in @file{libm.a}, but with @code{__} (two underscores)
2472 prepended to the symbol names as specified by @file{math.h}.
2473 @samp{math-symbols} also defines weak aliases for the functions
2474 declared in @file{math.h}. However, @code{--with-libf7} won't
2475 install no @file{math.h} header file whatsoever, this file must come
2476 from elsewhere. This option sets @option{--with-double-comparison}
2479 @item --with-nds32-lib=@var{library}
2480 Specifies that @var{library} setting is used for building @file{libgcc.a}.
2481 Currently, the valid @var{library} is @samp{newlib} or @samp{mculib}.
2482 This option is only supported for the NDS32 target.
2484 @item --with-build-time-tools=@var{dir}
2485 Specifies where to find the set of target tools (assembler, linker, etc.)
2486 that will be used while building GCC itself. This option can be useful
2487 if the directory layouts are different between the system you are building
2488 GCC on, and the system where you will deploy it.
2490 For example, on an @samp{ia64-hp-hpux} system, you may have the GNU
2491 assembler and linker in @file{/usr/bin}, and the native tools in a
2492 different path, and build a toolchain that expects to find the
2493 native tools in @file{/usr/bin}.
2495 When you use this option, you should ensure that @var{dir} includes
2496 @command{ar}, @command{as}, @command{ld}, @command{nm},
2497 @command{ranlib} and @command{strip} if necessary, and possibly
2498 @command{objdump}. Otherwise, GCC may use an inconsistent set of
2502 @subsubheading Overriding @command{configure} test results
2504 Sometimes, it might be necessary to override the result of some
2505 @command{configure} test, for example in order to ease porting to a new
2506 system or work around a bug in a test. The toplevel @command{configure}
2507 script provides three variables for this:
2511 @item build_configargs
2512 @cindex @code{build_configargs}
2513 The contents of this variable is passed to all build @command{configure}
2516 @item host_configargs
2517 @cindex @code{host_configargs}
2518 The contents of this variable is passed to all host @command{configure}
2521 @item target_configargs
2522 @cindex @code{target_configargs}
2523 The contents of this variable is passed to all target @command{configure}
2528 In order to avoid shell and @command{make} quoting issues for complex
2529 overrides, you can pass a setting for @env{CONFIG_SITE} and set
2530 variables in the site file.
2532 @subheading Objective-C-Specific Options
2534 The following options apply to the build of the Objective-C runtime library.
2537 @item --enable-objc-gc
2538 Specify that an additional variant of the GNU Objective-C runtime library
2539 is built, using an external build of the Boehm-Demers-Weiser garbage
2540 collector (@uref{https://www.hboehm.info/gc/}). This library needs to be
2541 available for each multilib variant, unless configured with
2542 @option{--enable-objc-gc=@samp{auto}} in which case the build of the
2543 additional runtime library is skipped when not available and the build
2546 @item --with-target-bdw-gc=@var{list}
2547 @itemx --with-target-bdw-gc-include=@var{list}
2548 @itemx --with-target-bdw-gc-lib=@var{list}
2549 Specify search directories for the garbage collector header files and
2550 libraries. @var{list} is a comma separated list of key value pairs of the
2551 form @samp{@var{multilibdir}=@var{path}}, where the default multilib key
2552 is named as @samp{.} (dot), or is omitted (e.g.@:
2553 @samp{--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32}).
2555 The options @option{--with-target-bdw-gc-include} and
2556 @option{--with-target-bdw-gc-lib} must always be specified together
2557 for each multilib variant and they take precedence over
2558 @option{--with-target-bdw-gc}. If @option{--with-target-bdw-gc-include}
2559 is missing values for a multilib, then the value for the default
2560 multilib is used (e.g.@: @samp{--with-target-bdw-gc-include=/opt/bdw-gc/include}
2561 @samp{--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32}).
2562 If none of these options are specified, the library is assumed in
2566 @subheading D-Specific Options
2568 The following options apply to the build of the D runtime library.
2571 @item --enable-libphobos-checking
2572 @itemx --disable-libphobos-checking
2573 @itemx --enable-libphobos-checking=@var{list}
2574 This option controls whether run-time checks and contracts are compiled into
2575 the D runtime library. When the option is not specified, the library is built
2576 with @samp{release} checking. When the option is specified without a
2577 @var{list}, the result is the same as @samp{--enable-libphobos-checking=yes}.
2578 Likewise, @samp{--disable-libphobos-checking} is equivalent to
2579 @samp{--enable-libphobos-checking=no}.
2581 The categories of checks available in @var{list} are @samp{yes} (compiles
2582 libphobos with @option{-fno-release}), @samp{no} (compiles libphobos with
2583 @option{-frelease}), @samp{all} (same as @samp{yes}), @samp{none} or
2584 @samp{release} (same as @samp{no}).
2586 Individual checks available in @var{list} are @samp{assert} (compiles libphobos
2587 with an extra option @option{-fassert}).
2589 @item --with-libphobos-druntime-only
2590 @itemx --with-libphobos-druntime-only=@var{choice}
2591 Specify whether to build only the core D runtime library (druntime), or both
2592 the core and standard library (phobos) into libphobos. This is useful for
2593 targets that have full support in druntime, but no or incomplete support
2594 in phobos. @var{choice} can be one of @samp{auto}, @samp{yes}, and @samp{no}
2595 where @samp{auto} is the default.
2597 When the option is not specified, the default choice @samp{auto} means that it
2598 is inferred whether the target has support for the phobos standard library.
2599 When the option is specified without a @var{choice}, the result is the same as
2600 @samp{--with-libphobos-druntime-only=yes}.
2602 @item --with-target-system-zlib
2603 Use installed @samp{zlib} rather than that included with GCC@. This needs
2604 to be available for each multilib variant, unless configured with
2605 @option{--with-target-system-zlib=@samp{auto}} in which case the GCC@ included
2606 @samp{zlib} is only used when the system installed library is not available.
2614 @uref{./index.html,,Return to the GCC Installation page}
2618 @c ***Building****************************************************************
2620 @comment node-name, next, previous, up
2621 @node Building, Testing, Configuration, Installing GCC
2627 @cindex Installing GCC: Building
2629 Now that GCC is configured, you are ready to build the compiler and
2632 Some commands executed when making the compiler may fail (return a
2633 nonzero status) and be ignored by @command{make}. These failures, which
2634 are often due to files that were not found, are expected, and can safely
2637 It is normal to have compiler warnings when compiling certain files.
2638 Unless you are a GCC developer, you can generally ignore these warnings
2639 unless they cause compilation to fail. Developers should attempt to fix
2640 any warnings encountered, however they can temporarily continue past
2641 warnings-as-errors by specifying the configure flag
2642 @option{--disable-werror}.
2644 On certain old systems, defining certain environment variables such as
2645 @env{CC} can interfere with the functioning of @command{make}.
2647 If you encounter seemingly strange errors when trying to build the
2648 compiler in a directory other than the source directory, it could be
2649 because you have previously configured the compiler in the source
2650 directory. Make sure you have done all the necessary preparations.
2652 If you build GCC on a BSD system using a directory stored in an old System
2653 V file system, problems may occur in running @command{fixincludes} if the
2654 System V file system doesn't support symbolic links. These problems
2655 result in a failure to fix the declaration of @code{size_t} in
2656 @file{sys/types.h}. If you find that @code{size_t} is a signed type and
2657 that type mismatches occur, this could be the cause.
2659 The solution is not to use such a directory for building GCC@.
2661 Similarly, when building from the source repository or snapshots, or if you modify
2662 @file{*.l} files, you need the Flex lexical analyzer generator
2663 installed. If you do not modify @file{*.l} files, releases contain
2664 the Flex-generated files and you do not need Flex installed to build
2665 them. There is still one Flex-based lexical analyzer (part of the
2666 build machinery, not of GCC itself) that is used even if you only
2667 build the C front end.
2669 When building from the source repository or snapshots, or if you modify Texinfo
2670 documentation, you need version 4.7 or later of Texinfo installed if you
2671 want Info documentation to be regenerated. Releases contain Info
2672 documentation pre-built for the unmodified documentation in the release.
2674 @section Building a native compiler
2676 For a native build, the default configuration is to perform
2677 a 3-stage bootstrap of the compiler when @samp{make} is invoked.
2678 This will build the entire GCC system and ensure that it compiles
2679 itself correctly. It can be disabled with the @option{--disable-bootstrap}
2680 parameter to @samp{configure}, but bootstrapping is suggested because
2681 the compiler will be tested more completely and could also have
2684 The bootstrapping process will complete the following steps:
2688 Build tools necessary to build the compiler.
2691 Perform a 3-stage bootstrap of the compiler. This includes building
2692 three times the target tools for use by the compiler such as binutils
2693 (bfd, binutils, gas, gprof, ld, and opcodes) if they have been
2694 individually linked or moved into the top level GCC source tree before
2698 Perform a comparison test of the stage2 and stage3 compilers.
2701 Build runtime libraries using the stage3 compiler from the previous step.
2705 If you are short on disk space you might consider @samp{make
2706 bootstrap-lean} instead. The sequence of compilation is the
2707 same described above, but object files from the stage1 and
2708 stage2 of the 3-stage bootstrap of the compiler are deleted as
2709 soon as they are no longer needed.
2711 If you wish to use non-default GCC flags when compiling the stage2
2712 and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when
2713 doing @samp{make}. For example, if you want to save additional space
2714 during the bootstrap and in the final installation as well, you can
2715 build the compiler binaries without debugging information as in the
2716 following example. This will save roughly 40% of disk space both for
2717 the bootstrap and the final installation. (Libraries will still contain
2718 debugging information.)
2721 make BOOT_CFLAGS='-O' bootstrap
2724 You can place non-default optimization flags into @code{BOOT_CFLAGS}; they
2725 are less well tested here than the default of @samp{-g -O2}, but should
2726 still work. In a few cases, you may find that you need to specify special
2727 flags such as @option{-msoft-float} here to complete the bootstrap; or,
2728 if the native compiler miscompiles the stage1 compiler, you may need
2729 to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts
2730 of the stage1 compiler that were miscompiled, or by using @samp{make
2731 bootstrap4} to increase the number of stages of bootstrap.
2733 @code{BOOT_CFLAGS} does not apply to bootstrapped target libraries.
2734 Since these are always compiled with the compiler currently being
2735 bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their
2736 compilation flags, as for non-bootstrapped target libraries.
2737 Again, if the native compiler miscompiles the stage1 compiler, you may
2738 need to work around this by avoiding non-working parts of the stage1
2739 compiler. Use @code{STAGE1_TFLAGS} to this end.
2741 If you used the flag @option{--enable-languages=@dots{}} to restrict
2742 the compilers to be built, only those you've actually enabled will be
2743 built. This will of course only build those runtime libraries, for
2744 which the particular compiler has been built. Please note,
2745 that re-defining @env{LANGUAGES} when calling @samp{make}
2746 @strong{does not} work anymore!
2748 If the comparison of stage2 and stage3 fails, this normally indicates
2749 that the stage2 compiler has compiled GCC incorrectly, and is therefore
2750 a potentially serious bug which you should investigate and report. (On
2751 a few systems, meaningful comparison of object files is impossible; they
2752 always appear ``different''. If you encounter this problem, you will
2753 need to disable comparison in the @file{Makefile}.)
2755 If you do not want to bootstrap your compiler, you can configure with
2756 @option{--disable-bootstrap}. In particular cases, you may want to
2757 bootstrap your compiler even if the target system is not the same as
2758 the one you are building on: for example, you could build a
2759 @code{powerpc-unknown-linux-gnu} toolchain on a
2760 @code{powerpc64-unknown-linux-gnu} host. In this case, pass
2761 @option{--enable-bootstrap} to the configure script.
2763 @code{BUILD_CONFIG} can be used to bring in additional customization
2764 to the build. It can be set to a whitespace-separated list of names.
2765 For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will
2766 be included by the top-level @file{Makefile}, bringing in any settings
2767 it contains. The default @code{BUILD_CONFIG} can be set using the
2768 configure option @option{--with-build-config=@code{NAME}...}. Some
2769 examples of supported build configurations are:
2772 @item @samp{bootstrap-O1}
2773 Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds
2774 @option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to
2775 @samp{BOOT_CFLAGS='-g -O1'}.
2777 @item @samp{bootstrap-O3}
2778 @itemx @samp{bootstrap-Og}
2779 Analogous to @code{bootstrap-O1}.
2781 @item @samp{bootstrap-lto}
2782 Enables Link-Time Optimization for host tools during bootstrapping.
2783 @samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding
2784 @option{-flto} to @samp{BOOT_CFLAGS}. This option assumes that the host
2785 supports the linker plugin (e.g.@: GNU ld version 2.21 or later or GNU gold
2786 version 2.21 or later).
2788 @item @samp{bootstrap-lto-noplugin}
2789 This option is similar to @code{bootstrap-lto}, but is intended for
2790 hosts that do not support the linker plugin. Without the linker plugin
2791 static libraries are not compiled with link-time optimizations. Since
2792 the GCC middle end and back end are in @file{libbackend.a} this means
2793 that only the front end is actually LTO optimized.
2795 @item @samp{bootstrap-lto-lean}
2796 This option is similar to @code{bootstrap-lto}, but is intended for
2797 faster build by only using LTO in the final bootstrap stage.
2798 With @samp{make profiledbootstrap} the LTO frontend
2799 is trained only on generator files.
2801 @item @samp{bootstrap-debug}
2802 Verifies that the compiler generates the same executable code, whether
2803 or not it is asked to emit debug information. To this end, this
2804 option builds stage2 host programs without debug information, and uses
2805 @file{contrib/compare-debug} to compare them with the stripped stage3
2806 object files. If @code{BOOT_CFLAGS} is overridden so as to not enable
2807 debug information, stage2 will have it, and stage3 won't. This option
2808 is enabled by default when GCC bootstrapping is enabled, if
2809 @code{strip} can turn object files compiled with and without debug
2810 info into identical object files. In addition to better test
2811 coverage, this option makes default bootstraps faster and leaner.
2813 @item @samp{bootstrap-debug-big}
2814 Rather than comparing stripped object files, as in
2815 @code{bootstrap-debug}, this option saves internal compiler dumps
2816 during stage2 and stage3 and compares them as well, which helps catch
2817 additional potential problems, but at a great cost in terms of disk
2818 space. It can be specified in addition to @samp{bootstrap-debug}.
2820 @item @samp{bootstrap-debug-lean}
2821 This option saves disk space compared with @code{bootstrap-debug-big},
2822 but at the expense of some recompilation. Instead of saving the dumps
2823 of stage2 and stage3 until the final compare, it uses
2824 @option{-fcompare-debug} to generate, compare and remove the dumps
2825 during stage3, repeating the compilation that already took place in
2826 stage2, whose dumps were not saved.
2828 @item @samp{bootstrap-debug-lib}
2829 This option tests executable code invariance over debug information
2830 generation on target libraries, just like @code{bootstrap-debug-lean}
2831 tests it on host programs. It builds stage3 libraries with
2832 @option{-fcompare-debug}, and it can be used along with any of the
2833 @code{bootstrap-debug} options above.
2835 There aren't @code{-lean} or @code{-big} counterparts to this option
2836 because most libraries are only build in stage3, so bootstrap compares
2837 would not get significant coverage. Moreover, the few libraries built
2838 in stage2 are used in stage3 host programs, so we wouldn't want to
2839 compile stage2 libraries with different options for comparison purposes.
2841 @item @samp{bootstrap-debug-ckovw}
2842 Arranges for error messages to be issued if the compiler built on any
2843 stage is run without the option @option{-fcompare-debug}. This is
2844 useful to verify the full @option{-fcompare-debug} testing coverage. It
2845 must be used along with @code{bootstrap-debug-lean} and
2846 @code{bootstrap-debug-lib}.
2848 @item @samp{bootstrap-cet}
2849 This option enables Intel CET for host tools during bootstrapping.
2850 @samp{BUILD_CONFIG=bootstrap-cet} is equivalent to adding
2851 @option{-fcf-protection} to @samp{BOOT_CFLAGS}. This option
2852 assumes that the host supports Intel CET (e.g.@: GNU assembler version
2855 @item @samp{bootstrap-time}
2856 Arranges for the run time of each program started by the GCC driver,
2857 built in any stage, to be logged to @file{time.log}, in the top level of
2860 @item @samp{bootstrap-asan}
2861 Compiles GCC itself using Address Sanitization in order to catch invalid memory
2862 accesses within the GCC code.
2864 @item @samp{bootstrap-hwasan}
2865 Compiles GCC itself using HWAddress Sanitization in order to catch invalid
2866 memory accesses within the GCC code. This option is only available on AArch64
2867 systems that are running Linux kernel version 5.4 or later.
2871 @section Building a cross compiler
2873 When building a cross compiler, it is not generally possible to do a
2874 3-stage bootstrap of the compiler. This makes for an interesting problem
2875 as parts of GCC can only be built with GCC@.
2877 To build a cross compiler, we recommend first building and installing a
2878 native compiler. You can then use the native GCC compiler to build the
2879 cross compiler. The installed native compiler needs to be GCC version
2882 Assuming you have already installed a native copy of GCC and configured
2883 your cross compiler, issue the command @command{make}, which performs the
2888 Build host tools necessary to build the compiler.
2891 Build target tools for use by the compiler such as binutils (bfd,
2892 binutils, gas, gprof, ld, and opcodes)
2893 if they have been individually linked or moved into the top level GCC source
2894 tree before configuring.
2897 Build the compiler (single stage only).
2900 Build runtime libraries using the compiler from the previous step.
2903 Note that if an error occurs in any step the make process will exit.
2905 If you are not building GNU binutils in the same source tree as GCC,
2906 you will need a cross-assembler and cross-linker installed before
2907 configuring GCC@. Put them in the directory
2908 @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools
2909 you should put in this directory:
2913 This should be the cross-assembler.
2916 This should be the cross-linker.
2919 This should be the cross-archiver: a program which can manipulate
2920 archive files (linker libraries) in the target machine's format.
2923 This should be a program to construct a symbol table in an archive file.
2926 The installation of GCC will find these programs in that directory,
2927 and copy or link them to the proper place to for the cross-compiler to
2928 find them when run later.
2930 The easiest way to provide these files is to build the Binutils package.
2931 Configure it with the same @option{--host} and @option{--target}
2932 options that you use for configuring GCC, then build and install
2933 them. They install their executables automatically into the proper
2934 directory. Alas, they do not support all the targets that GCC
2937 If you are not building a C library in the same source tree as GCC,
2938 you should also provide the target libraries and headers before
2939 configuring GCC, specifying the directories with
2940 @option{--with-sysroot} or @option{--with-headers} and
2941 @option{--with-libs}. Many targets also require ``start files'' such
2942 as @file{crt0.o} and
2943 @file{crtn.o} which are linked into each executable. There may be several
2944 alternatives for @file{crt0.o}, for use with profiling or other
2945 compilation options. Check your target's definition of
2946 @code{STARTFILE_SPEC} to find out what start files it uses.
2948 @section Building in parallel
2950 GNU Make 3.80 and above, which is necessary to build GCC, support
2951 building in parallel. To activate this, you can use @samp{make -j 2}
2952 instead of @samp{make}. You can also specify a bigger number, and
2953 in most cases using a value greater than the number of processors in
2954 your machine will result in fewer and shorter I/O latency hits, thus
2955 improving overall throughput; this is especially true for slow drives
2956 and network filesystems.
2958 @section Building the Ada compiler
2961 @ref{GNAT-prerequisite}.
2964 @uref{prerequisites.html#GNAT-prerequisite,,GNAT prerequisites}.
2967 @section Building with profile feedback
2969 It is possible to use profile feedback to optimize the compiler itself. This
2970 should result in a faster compiler binary. Experiments done on x86 using gcc
2971 3.3 showed approximately 7 percent speedup on compiling C programs. To
2972 bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}.
2974 When @samp{make profiledbootstrap} is run, it will first build a @code{stage1}
2975 compiler. This compiler is used to build a @code{stageprofile} compiler
2976 instrumented to collect execution counts of instruction and branch
2977 probabilities. Training run is done by building @code{stagetrain}
2978 compiler. Finally a @code{stagefeedback} compiler is built
2979 using the information collected.
2981 Unlike standard bootstrap, several additional restrictions apply. The
2982 compiler used to build @code{stage1} needs to support a 64-bit integral type.
2983 It is recommended to only use GCC for this.
2985 On Linux/x86_64 hosts with some restrictions (no virtualization) it is
2986 also possible to do autofdo build with @samp{make
2987 autoprofiledback}. This uses Linux perf to sample branches in the
2988 binary and then rebuild it with feedback derived from the profile.
2989 Linux perf and the @code{autofdo} toolkit needs to be installed for
2992 Only the profile from the current build is used, so when an error
2993 occurs it is recommended to clean before restarting. Otherwise
2994 the code quality may be much worse.
3001 @uref{./index.html,,Return to the GCC Installation page}
3005 @c ***Testing*****************************************************************
3007 @comment node-name, next, previous, up
3008 @node Testing, Final install, Building, Installing GCC
3012 @chapter Installing GCC: Testing
3015 @cindex Installing GCC: Testing
3018 Before you install GCC, we encourage you to run the testsuites and to
3019 compare your results with results from a similar configuration that have
3020 been submitted to the
3021 @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}.
3022 Some of these archived results are linked from the build status lists
3023 at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who
3024 reports a successful build runs the testsuites and submits the results.
3025 This step is optional and may require you to download additional software,
3026 but it can give you confidence in your new GCC installation or point out
3027 problems before you install and start using your new GCC@.
3029 First, you must have @uref{download.html,,downloaded the testsuites}.
3030 These are part of the full distribution, but if you downloaded the
3031 ``core'' compiler plus any front ends, you must download the testsuites
3034 Second, you must have the testing tools installed. This includes
3035 @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect;
3036 the DejaGnu site has links to these.
3037 Some optional tests also require Python3 and pytest module.
3039 If the directories where @command{runtest} and @command{expect} were
3040 installed are not in the @env{PATH}, you may need to set the following
3041 environment variables appropriately, as in the following example (which
3042 assumes that DejaGnu has been installed under @file{/usr/local}):
3045 TCL_LIBRARY = /usr/local/share/tcl8.0
3046 DEJAGNULIBS = /usr/local/share/dejagnu
3049 (On systems such as Cygwin, these paths are required to be actual
3050 paths, not mounts or links; presumably this is due to some lack of
3051 portability in the DejaGnu code.)
3054 Finally, you can run the testsuite (which may take a long time):
3056 cd @var{objdir}; make -k check
3059 This will test various components of GCC, such as compiler
3060 front ends and runtime libraries. While running the testsuite, DejaGnu
3061 might emit some harmless messages resembling
3062 @samp{WARNING: Couldn't find the global config file.} or
3063 @samp{WARNING: Couldn't find tool init file} that can be ignored.
3065 If you are testing a cross-compiler, you may want to run the testsuite
3066 on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}.
3068 @section How can you run the testsuite on selected tests?
3070 In order to run sets of tests selectively, there are targets
3071 @samp{make check-gcc} and language specific @samp{make check-c},
3072 @samp{make check-c++}, @samp{make check-d} @samp{make check-fortran},
3073 @samp{make check-ada}, @samp{make check-objc}, @samp{make check-obj-c++},
3074 @samp{make check-lto}
3075 in the @file{gcc} subdirectory of the object directory. You can also
3076 just run @samp{make check} in a subdirectory of the object directory.
3079 A more selective way to just run all @command{gcc} execute tests in the
3083 make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}"
3086 Likewise, in order to run only the @command{g++} ``old-deja'' tests in
3087 the testsuite with filenames matching @samp{9805*}, you would use
3090 make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}"
3093 The file-matching expression following @var{filename}@command{.exp=} is treated
3094 as a series of whitespace-delimited glob expressions so that multiple patterns
3095 may be passed, although any whitespace must either be escaped or surrounded by
3096 single quotes if multiple expressions are desired. For example,
3099 make check-g++ RUNTESTFLAGS="old-deja.exp=9805*\ virtual2.c @var{other-options}"
3100 make check-g++ RUNTESTFLAGS="'old-deja.exp=9805* virtual2.c' @var{other-options}"
3103 The @file{*.exp} files are located in the testsuite directories of the GCC
3104 source, the most important ones being @file{compile.exp},
3105 @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}.
3106 To get a list of the possible @file{*.exp} files, pipe the
3107 output of @samp{make check} into a file and look at the
3108 @samp{Running @dots{} .exp} lines.
3110 @section Passing options and running multiple testsuites
3112 You can pass multiple options to the testsuite using the
3113 @samp{--target_board} option of DejaGNU, either passed as part of
3114 @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to
3115 work outside the makefiles. For example,
3118 make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants"
3121 will run the standard @command{g++} testsuites (``unix'' is the target name
3122 for a standard native testsuite situation), passing
3123 @samp{-O3 -fmerge-constants} to the compiler on every test, i.e.,
3124 slashes separate options.
3126 You can run the testsuites multiple times using combinations of options
3127 with a syntax similar to the brace expansion of popular shells:
3130 @dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}"
3133 (Note the empty option caused by the trailing comma in the final group.)
3134 The following will run each testsuite eight times using the @samp{arm-sim}
3135 target, as if you had specified all possible combinations yourself:
3138 --target_board='arm-sim/-mhard-float/-O1 \
3139 arm-sim/-mhard-float/-O2 \
3140 arm-sim/-mhard-float/-O3 \
3141 arm-sim/-mhard-float \
3142 arm-sim/-msoft-float/-O1 \
3143 arm-sim/-msoft-float/-O2 \
3144 arm-sim/-msoft-float/-O3 \
3145 arm-sim/-msoft-float'
3148 They can be combined as many times as you wish, in arbitrary ways. This
3152 @dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}"
3155 will generate four combinations, all involving @samp{-Wextra}.
3157 The disadvantage to this method is that the testsuites are run in serial,
3158 which is a waste on multiprocessor systems. For users with GNU Make and
3159 a shell which performs brace expansion, you can run the testsuites in
3160 parallel by having the shell perform the combinations and @command{make}
3161 do the parallel runs. Instead of using @samp{--target_board}, use a
3162 special makefile target:
3165 make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{}
3171 make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@}
3174 will run three concurrent ``make-gcc'' testsuites, eventually testing all
3175 ten combinations as described above. Note that this is currently only
3176 supported in the @file{gcc} subdirectory. (To see how this works, try
3177 typing @command{echo} before the example given here.)
3180 @section How to interpret test results
3182 The result of running the testsuite are various @file{*.sum} and @file{*.log}
3183 files in the testsuite subdirectories. The @file{*.log} files contain a
3184 detailed log of the compiler invocations and the corresponding
3185 results, the @file{*.sum} files summarize the results. These summaries
3186 contain status codes for all tests:
3190 PASS: the test passed as expected
3192 XPASS: the test unexpectedly passed
3194 FAIL: the test unexpectedly failed
3196 XFAIL: the test failed as expected
3198 UNSUPPORTED: the test is not supported on this platform
3200 ERROR: the testsuite detected an error
3202 WARNING: the testsuite detected a possible problem
3205 It is normal for some tests to report unexpected failures. At the
3206 current time the testing harness does not allow fine grained control
3207 over whether or not a test is expected to fail. This problem should
3208 be fixed in future releases.
3211 @section Submitting test results
3213 If you want to report the results to the GCC project, use the
3214 @file{contrib/test_summary} shell script. Start it in the @var{objdir} with
3217 @var{srcdir}/contrib/test_summary -p your_commentary.txt \
3218 -m gcc-testresults@@gcc.gnu.org |sh
3221 This script uses the @command{Mail} program to send the results, so
3222 make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is
3223 prepended to the testsuite summary and should contain any special
3224 remarks you have on your results or your build environment. Please
3225 do not edit the testsuite result block or the subject line, as these
3226 messages may be automatically processed.
3233 @uref{./index.html,,Return to the GCC Installation page}
3237 @c ***Final install***********************************************************
3239 @comment node-name, next, previous, up
3240 @node Final install, , Testing, Installing GCC
3242 @ifset finalinstallhtml
3244 @chapter Installing GCC: Final installation
3247 Now that GCC has been built (and optionally tested), you can install it with
3249 cd @var{objdir} && make install
3252 We strongly recommend to install into a target directory where there is
3253 no previous version of GCC present. Also, the GNAT runtime should not
3254 be stripped, as this would break certain features of the debugger that
3255 depend on this debugging information (catching Ada exceptions for
3258 That step completes the installation of GCC; user level binaries can
3259 be found in @file{@var{prefix}/bin} where @var{prefix} is the value
3260 you specified with the @option{--prefix} to configure (or
3261 @file{/usr/local} by default). (If you specified @option{--bindir},
3262 that directory will be used instead; otherwise, if you specified
3263 @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.)
3264 Headers for the C++ library are installed in
3265 @file{@var{prefix}/include}; libraries in @file{@var{libdir}}
3266 (normally @file{@var{prefix}/lib}); internal parts of the compiler in
3267 @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation
3268 in info format in @file{@var{infodir}} (normally
3269 @file{@var{prefix}/info}).
3271 When installing cross-compilers, GCC's executables
3272 are not only installed into @file{@var{bindir}}, that
3273 is, @file{@var{exec-prefix}/bin}, but additionally into
3274 @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory
3275 exists. Typically, such @dfn{tooldirs} hold target-specific
3276 binutils, including assembler and linker.
3278 Installation into a temporary staging area or into a @command{chroot}
3279 jail can be achieved with the command
3282 make DESTDIR=@var{path-to-rootdir} install
3286 where @var{path-to-rootdir} is the absolute path of
3287 a directory relative to which all installation paths will be
3288 interpreted. Note that the directory specified by @code{DESTDIR}
3289 need not exist yet; it will be created if necessary.
3291 There is a subtle point with tooldirs and @code{DESTDIR}:
3292 If you relocate a cross-compiler installation with
3293 e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory
3294 @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will
3295 be filled with duplicated GCC executables only if it already exists,
3296 it will not be created otherwise. This is regarded as a feature,
3297 not as a bug, because it gives slightly more control to the packagers
3298 using the @code{DESTDIR} feature.
3300 You can install stripped programs and libraries with
3306 If you are bootstrapping a released version of GCC then please
3307 quickly review the build status page for your release, available from
3308 @uref{http://gcc.gnu.org/buildstat.html}.
3309 If your system is not listed for the version of GCC that you built,
3311 @email{gcc@@gcc.gnu.org} indicating
3312 that you successfully built and installed GCC@.
3313 Include the following information:
3317 Output from running @file{@var{srcdir}/config.guess}. Do not send
3318 that file itself, just the one-line output from running it.
3321 The output of @samp{gcc -v} for your newly installed @command{gcc}.
3322 This tells us which version of GCC you built and the options you passed to
3326 Whether you enabled all languages or a subset of them. If you used a
3327 full distribution then this information is part of the configure
3328 options in the output of @samp{gcc -v}, but if you downloaded the
3329 ``core'' compiler plus additional front ends then it isn't apparent
3330 which ones you built unless you tell us about it.
3333 If the build was for GNU/Linux, also include:
3336 The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3);
3337 this information should be available from @file{/etc/issue}.
3340 The version of the Linux kernel, available from @samp{uname --version}
3344 The version of glibc you used; for RPM-based systems like Red Hat,
3345 Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version,
3346 and on systems like Debian and Progeny use @samp{dpkg -l libc6}.
3348 For other systems, you can include similar information if you think it is
3352 Any other information that you think would be useful to people building
3353 GCC on the same configuration. The new entry in the build status list
3354 will include a link to the archived copy of your message.
3357 We'd also like to know if the
3359 @ref{Specific, host/target specific installation notes}
3362 @uref{specific.html,,host/target specific installation notes}
3364 didn't include your host/target information or if that information is
3365 incomplete or out of date. Send a note to
3366 @email{gcc@@gcc.gnu.org} detailing how the information should be changed.
3368 If you find a bug, please report it following the
3369 @uref{../bugs/,,bug reporting guidelines}.
3371 If you want to print the GCC manuals, do @samp{cd @var{objdir}; make
3372 dvi}. You will need to have @command{texi2dvi} (version at least 4.7)
3373 and @TeX{} installed. This creates a number of @file{.dvi} files in
3374 subdirectories of @file{@var{objdir}}; these may be converted for
3375 printing with programs such as @command{dvips}. Alternately, by using
3376 @samp{make pdf} in place of @samp{make dvi}, you can create documentation
3377 in the form of @file{.pdf} files; this requires @command{texi2pdf}, which
3378 is included with Texinfo version 4.8 and later. You can also
3379 @uref{https://shop.fsf.org/,,buy printed manuals from the
3380 Free Software Foundation}, though such manuals may not be for the most
3381 recent version of GCC@.
3383 If you would like to generate online HTML documentation, do @samp{cd
3384 @var{objdir}; make html} and HTML will be generated for the gcc manuals in
3385 @file{@var{objdir}/gcc/HTML}.
3392 @uref{./index.html,,Return to the GCC Installation page}
3396 @c ***Binaries****************************************************************
3398 @comment node-name, next, previous, up
3399 @node Binaries, Specific, Installing GCC, Top
3403 @chapter Installing GCC: Binaries
3406 @cindex Installing GCC: Binaries
3408 We are often asked about pre-compiled versions of GCC@. While we cannot
3409 provide these for all platforms, below you'll find links to binaries for
3410 various platforms where creating them by yourself is not easy due to various
3413 Please note that we did not create these binaries, nor do we
3414 support them. If you have any problems installing them, please
3415 contact their makers.
3422 @uref{http://www.bullfreeware.com,,Bull's Open Source Software Archive for
3423 for AIX 6 and AIX 7};
3426 @uref{http://www.perzl.org/aix/,,AIX Open Source Packages (AIX5L AIX 6.1
3431 DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}.
3437 @uref{http://hpux.connect.org.uk/,,HP-UX Porting Center};
3441 Solaris 2 (SPARC, Intel):
3444 @uref{https://www.opencsw.org/,,OpenCSW}
3451 The @uref{https://brew.sh,,Homebrew} package manager;
3453 @uref{https://www.macports.org,,MacPorts}.
3460 The @uref{https://sourceware.org/cygwin/,,Cygwin} project;
3462 The @uref{http://www.mingw.org/,,MinGW} and
3463 @uref{http://mingw-w64.org/doku.php,,mingw-w64} projects.
3467 @uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a
3468 number of platforms.
3471 The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has
3472 links to GNU Fortran binaries for several platforms.
3480 @uref{./index.html,,Return to the GCC Installation page}
3484 @c ***Specific****************************************************************
3486 @comment node-name, next, previous, up
3487 @node Specific, GNU Free Documentation License, Binaries, Top
3491 @chapter Host/target specific installation notes for GCC
3494 @cindex Specific installation notes
3495 @cindex Target specific installation
3496 @cindex Host specific installation
3497 @cindex Target specific installation notes
3499 Please read this document carefully @emph{before} installing the
3500 GNU Compiler Collection on your machine.
3502 Note that this list of install notes is @emph{not} a list of supported
3503 hosts or targets. Not all supported hosts and targets are listed
3504 here, only the ones that require host-specific or target-specific
3505 information have to.
3510 @uref{#aarch64-x-x,,aarch64*-*-*}
3512 @uref{#alpha-x-x,,alpha*-*-*}
3514 @uref{#amd64-x-solaris2,,amd64-*-solaris2*}
3516 @uref{#arm-x-eabi,,arm-*-eabi}
3520 @uref{#bfin,,Blackfin}
3524 @uref{#x-x-freebsd,,*-*-freebsd*}
3526 @uref{#h8300-hms,,h8300-hms}
3528 @uref{#hppa-hp-hpux,,hppa*-hp-hpux*}
3530 @uref{#hppa-hp-hpux10,,hppa*-hp-hpux10}
3532 @uref{#hppa-hp-hpux11,,hppa*-hp-hpux11}
3534 @uref{#x-x-linux-gnu,,*-*-linux-gnu}
3536 @uref{#ix86-x-linux,,i?86-*-linux*}
3538 @uref{#ix86-x-solaris2,,i?86-*-solaris2*}
3540 @uref{#ia64-x-linux,,ia64-*-linux}
3542 @uref{#ia64-x-hpux,,ia64-*-hpux*}
3544 @uref{#x-ibm-aix,,*-ibm-aix*}
3546 @uref{#iq2000-x-elf,,iq2000-*-elf}
3548 @uref{#lm32-x-elf,,lm32-*-elf}
3550 @uref{#lm32-x-uclinux,,lm32-*-uclinux}
3552 @uref{#m32c-x-elf,,m32c-*-elf}
3554 @uref{#m32r-x-elf,,m32r-*-elf}
3556 @uref{#m68k-x-x,,m68k-*-*}
3558 @uref{#m68k-uclinux,,m68k-uclinux}
3560 @uref{#microblaze-x-elf,,microblaze-*-elf}
3562 @uref{#mips-x-x,,mips-*-*}
3564 @uref{#nds32le-x-elf,,nds32le-*-elf}
3566 @uref{#nds32be-x-elf,,nds32be-*-elf}
3568 @uref{#nvptx-x-none,,nvptx-*-none}
3570 @uref{#or1k-x-elf,,or1k-*-elf}
3572 @uref{#or1k-x-linux,,or1k-*-linux}
3574 @uref{#powerpc-x-x,,powerpc*-*-*}
3576 @uref{#powerpc-x-darwin,,powerpc-*-darwin*}
3578 @uref{#powerpc-x-elf,,powerpc-*-elf}
3580 @uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*}
3582 @uref{#powerpc-x-netbsd,,powerpc-*-netbsd*}
3584 @uref{#powerpc-x-eabisim,,powerpc-*-eabisim}
3586 @uref{#powerpc-x-eabi,,powerpc-*-eabi}
3588 @uref{#powerpcle-x-elf,,powerpcle-*-elf}
3590 @uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim}
3592 @uref{#powerpcle-x-eabi,,powerpcle-*-eabi}
3594 @uref{#riscv32-x-elf,,riscv32-*-elf}
3596 @uref{#riscv32-x-linux,,riscv32-*-linux}
3598 @uref{#riscv64-x-elf,,riscv64-*-elf}
3600 @uref{#riscv64-x-linux,,riscv64-*-linux}
3602 @uref{#s390-x-linux,,s390-*-linux*}
3604 @uref{#s390x-x-linux,,s390x-*-linux*}
3606 @uref{#s390x-ibm-tpf,,s390x-ibm-tpf*}
3608 @uref{#x-x-solaris2,,*-*-solaris2*}
3610 @uref{#sparc-x-x,,sparc*-*-*}
3612 @uref{#sparc-sun-solaris2,,sparc-sun-solaris2*}
3614 @uref{#sparc-x-linux,,sparc-*-linux*}
3616 @uref{#sparc64-x-solaris2,,sparc64-*-solaris2*}
3618 @uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*}
3620 @uref{#c6x-x-x,,c6x-*-*}
3622 @uref{#tilegx-x-linux,,tilegx-*-linux*}
3624 @uref{#tilegxbe-x-linux,,tilegxbe-*-linux*}
3626 @uref{#tilepro-x-linux,,tilepro-*-linux*}
3628 @uref{#visium-x-elf, visium-*-elf}
3630 @uref{#x-x-vxworks,,*-*-vxworks*}
3632 @uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*}
3634 @uref{#x86-64-x-solaris2,,x86_64-*-solaris2*}
3636 @uref{#xtensa-x-elf,,xtensa*-*-elf}
3638 @uref{#xtensa-x-linux,,xtensa*-*-linux*}
3640 @uref{#windows,,Microsoft Windows}
3642 @uref{#x-x-cygwin,,*-*-cygwin}
3644 @uref{#x-x-mingw32,,*-*-mingw32}
3648 @uref{#older,,Older systems}
3653 @uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.)
3659 <!-- -------- host/target specific issues start here ---------------- -->
3662 @anchor{aarch64-x-x}
3663 @heading aarch64*-*-*
3664 Binutils pre 2.24 does not have support for selecting @option{-mabi} and
3665 does not support ILP32. If it is used to build GCC 4.9 or later, GCC will
3666 not support option @option{-mabi=ilp32}.
3668 To enable a workaround for the Cortex-A53 erratum number 835769 by default
3669 (for all CPUs regardless of -mcpu option given) at configure time use the
3670 @option{--enable-fix-cortex-a53-835769} option. This will enable the fix by
3671 default and can be explicitly disabled during compilation by passing the
3672 @option{-mno-fix-cortex-a53-835769} option. Conversely,
3673 @option{--disable-fix-cortex-a53-835769} will disable the workaround by
3674 default. The workaround is disabled by default if neither of
3675 @option{--enable-fix-cortex-a53-835769} or
3676 @option{--disable-fix-cortex-a53-835769} is given at configure time.
3678 To enable a workaround for the Cortex-A53 erratum number 843419 by default
3679 (for all CPUs regardless of -mcpu option given) at configure time use the
3680 @option{--enable-fix-cortex-a53-843419} option. This workaround is applied at
3681 link time. Enabling the workaround will cause GCC to pass the relevant option
3682 to the linker. It can be explicitly disabled during compilation by passing the
3683 @option{-mno-fix-cortex-a53-843419} option. Conversely,
3684 @option{--disable-fix-cortex-a53-843419} will disable the workaround by default.
3685 The workaround is disabled by default if neither of
3686 @option{--enable-fix-cortex-a53-843419} or
3687 @option{--disable-fix-cortex-a53-843419} is given at configure time.
3689 To enable Branch Target Identification Mechanism and Return Address Signing by
3690 default at configure time use the @option{--enable-standard-branch-protection}
3691 option. This is equivalent to having @option{-mbranch-protection=standard}
3692 during compilation. This can be explicitly disabled during compilation by
3693 passing the @option{-mbranch-protection=none} option which turns off all
3694 types of branch protections. Conversely,
3695 @option{--disable-standard-branch-protection} will disable both the
3696 protections by default. This mechanism is turned off by default if neither
3697 of the options are given at configure time.
3704 This section contains general configuration information for all
3705 Alpha-based platforms using ELF@. In addition to reading this
3706 section, please read all other sections that match your target.
3711 @anchor{amd64-x-solaris2}
3712 @heading amd64-*-solaris2*
3713 This is a synonym for @samp{x86_64-*-solaris2*}.
3718 @anchor{amdgcn-x-amdhsa}
3719 @heading amdgcn-*-amdhsa
3722 Instead of GNU Binutils, you will need to install LLVM 6, or later, and copy
3723 @file{bin/llvm-mc} to @file{amdgcn-amdhsa/bin/as},
3724 @file{bin/lld} to @file{amdgcn-amdhsa/bin/ld},
3725 @file{bin/llvm-nm} to @file{amdgcn-amdhsa/bin/nm}, and
3726 @file{bin/llvm-ar} to both @file{bin/amdgcn-amdhsa-ar} and
3727 @file{bin/amdgcn-amdhsa-ranlib}.
3729 Use Newlib (2019-01-16, or newer).
3731 To run the binaries, install the HSA Runtime from the
3732 @uref{https://rocm.github.io,,ROCm Platform}, and use
3733 @file{libexec/gcc/amdhsa-amdhsa/@var{version}/gcn-run} to launch them
3739 @anchor{arc-x-elf32}
3740 @heading arc-*-elf32
3742 Use @samp{configure --target=arc-elf32 --with-cpu=@var{cpu} --enable-languages="c,c++"}
3743 to configure GCC, with @var{cpu} being one of @samp{arc600}, @samp{arc601},
3749 @anchor{arc-linux-uclibc}
3750 @heading arc-linux-uclibc
3752 Use @samp{configure --target=arc-linux-uclibc --with-cpu=arc700 --enable-languages="c,c++"} to configure GCC@.
3759 ARM-family processors.
3761 Building the Ada frontend commonly fails (an infinite loop executing
3762 @code{xsinfo}) if the host compiler is GNAT 4.8. Host compilers built from the
3763 GNAT 4.6, 4.9 or 5 release branches are known to succeed.
3770 ATMEL AVR-family micro controllers. These are used in embedded
3771 applications. There are no standard Unix configurations.
3773 @xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler
3777 See ``AVR Options'' in the main manual
3779 for the list of supported MCU types.
3781 Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@.
3783 Further installation notes and other useful information about AVR tools
3784 can also be obtained from:
3788 @uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/}
3790 @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/}
3793 The following error:
3795 Error: register required
3798 indicates that you should upgrade to a newer version of the binutils.
3805 The Blackfin processor, an Analog Devices DSP.
3807 @xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler
3811 See ``Blackfin Options'' in the main manual
3814 More information, and a version of binutils with support for this processor,
3815 are available at @uref{https://sourceforge.net/projects/adi-toolchain/}.
3822 The CR16 CompactRISC architecture is a 16-bit architecture. This
3823 architecture is used in embedded applications.
3826 @xref{CR16 Options,, CR16 Options, gcc, Using and Porting the GNU Compiler
3831 See ``CR16 Options'' in the main manual for a list of CR16-specific options.
3834 Use @samp{configure --target=cr16-elf --enable-languages=c,c++} to configure
3835 GCC@ for building a CR16 elf cross-compiler.
3837 Use @samp{configure --target=cr16-uclinux --enable-languages=c,c++} to
3838 configure GCC@ for building a CR16 uclinux cross-compiler.
3845 CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
3846 series. These are used in embedded applications.
3849 @xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler
3853 See ``CRIS Options'' in the main manual
3855 for a list of CRIS-specific options.
3857 There are a few different CRIS targets:
3860 Mainly for monolithic embedded systems. Includes a multilib for the
3861 @samp{v10} core used in @samp{ETRAX 100 LX}.
3862 @item cris-axis-linux-gnu
3863 A GNU/Linux port for the CRIS architecture, currently targeting
3864 @samp{ETRAX 100 LX} by default.
3867 Pre-packaged tools can be obtained from
3868 @uref{ftp://ftp.axis.com/@/pub/@/axis/@/tools/@/cris/@/compiler-kit/}. More
3869 information about this platform is available at
3870 @uref{http://developer.axis.com/}.
3877 Please have a look at the @uref{binaries.html,,binaries page}.
3879 You cannot install GCC by itself on MSDOS; it will not compile under
3880 any MSDOS compiler except itself. You need to get the complete
3881 compilation package DJGPP, which includes binaries as well as sources,
3882 and includes all the necessary compilation tools and libraries.
3887 @anchor{epiphany-x-elf}
3888 @heading epiphany-*-elf
3890 This configuration is intended for embedded systems.
3895 @anchor{x-x-freebsd}
3896 @heading *-*-freebsd*
3897 Support for FreeBSD 1 was discontinued in GCC 3.2. Support for
3898 FreeBSD 2 (and any mutant a.out variants of FreeBSD 3) was
3899 discontinued in GCC 4.0.
3901 In order to better utilize FreeBSD base system functionality and match
3902 the configuration of the system compiler, GCC 4.5 and above as well as
3903 GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present
3904 on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default
3905 (on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside
3906 @file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled
3907 by GCC 4.5 and above.
3909 We support FreeBSD using the ELF file format with DWARF 2 debugging
3910 for all CPU architectures. You may use @option{-gstabs} instead of
3911 @option{-g}, if you really want the old debugging format. There are
3912 no known issues with mixing object files and libraries with different
3913 debugging formats. Otherwise, this release of GCC should now match
3914 more of the configuration used in the stock FreeBSD configuration of
3915 GCC@. In particular, @option{--enable-threads} is now configured by
3916 default. However, as a general user, do not attempt to replace the
3917 system compiler with this release. Known to bootstrap and check with
3918 good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap
3919 and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4,
3920 4.5, 4.8, 4.9 and 5-CURRENT@.
3922 The version of binutils installed in @file{/usr/bin} probably works
3923 with this release of GCC@. Bootstrapping against the latest GNU
3924 binutils and/or the version found in @file{/usr/ports/devel/binutils} has
3925 been known to enable additional features and improve overall testsuite
3926 results. However, it is currently known that boehm-gc may not configure
3927 properly on FreeBSD prior to the FreeBSD 7.0 release with GNU binutils
3936 This configuration is intended for embedded systems.
3943 Renesas H8/300 series of processors.
3945 Please have a look at the @uref{binaries.html,,binaries page}.
3947 The calling convention and structure layout has changed in release 2.6.
3948 All code must be recompiled. The calling convention now passes the
3949 first three arguments in function calls in registers. Structures are no
3950 longer a multiple of 2 bytes.
3955 @anchor{hppa-hp-hpux}
3956 @heading hppa*-hp-hpux*
3957 Support for HP-UX version 9 and older was discontinued in GCC 3.4.
3959 We require using gas/binutils on all hppa platforms. Version 2.19 or
3960 later is recommended.
3962 It may be helpful to configure GCC with the
3963 @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and
3964 @option{--with-as=@dots{}} options to ensure that GCC can find GAS@.
3966 The HP assembler should not be used with GCC. It is rarely tested and may
3967 not work. It shouldn't be used with any languages other than C due to its
3970 Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging
3971 format which GCC does not know about). It also inserts timestamps
3972 into each object file it creates, causing the 3-stage comparison test to
3973 fail during a bootstrap. You should be able to continue by saying
3974 @samp{make all-host all-target} after getting the failure from @samp{make}.
3976 Various GCC features are not supported. For example, it does not support weak
3977 symbols or alias definitions. As a result, explicit template instantiations
3978 are required when using C++. This makes it difficult if not impossible to
3979 build many C++ applications.
3981 There are two default scheduling models for instructions. These are
3982 PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
3983 architecture specified for the target machine when configuring.
3984 PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
3985 the target is a @samp{hppa1*} machine.
3987 The PROCESSOR_8000 model is not well suited to older processors. Thus,
3988 it is important to completely specify the machine architecture when
3989 configuring if you want a model other than PROCESSOR_8000. The macro
3990 TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
3991 default scheduling model is desired.
3993 As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
3994 through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
3995 This namespace change might cause problems when bootstrapping with
3996 an earlier version of GCC or the HP compiler as essentially the same
3997 namespace is required for an entire build. This problem can be avoided
3998 in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95}
3999 or @samp{98}. Another way is to add an appropriate set of predefines
4000 to @env{CC}. The description for the @option{munix=} option contains
4001 a list of the predefines used with each standard.
4003 More specific information to @samp{hppa*-hp-hpux*} targets follows.
4008 @anchor{hppa-hp-hpux10}
4009 @heading hppa*-hp-hpux10
4010 For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch
4011 @code{PHCO_19798} from HP@.
4013 The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
4014 used for one-only code and data. This resolves many of the previous
4015 problems in using C++ on this target. However, the ABI is not compatible
4016 with the one implemented under HP-UX 11 using secondary definitions.
4021 @anchor{hppa-hp-hpux11}
4022 @heading hppa*-hp-hpux11
4023 GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
4024 be used to compile GCC 3.0 and up.
4026 The libffi library haven't been ported to 64-bit HP-UX@ and doesn't build.
4028 Refer to @uref{binaries.html,,binaries} for information about obtaining
4029 precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained
4030 to build the Ada language as it cannot be bootstrapped using C@. Ada is
4031 only available for the 32-bit PA-RISC runtime.
4033 Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
4034 bundled compiler supports only traditional C; you will need either HP's
4035 unbundled compiler, or a binary distribution of GCC@.
4037 It is possible to build GCC 3.3 starting with the bundled HP compiler,
4038 but the process requires several steps. GCC 3.3 can then be used to
4039 build later versions.
4041 There are several possible approaches to building the distribution.
4042 Binutils can be built first using the HP tools. Then, the GCC
4043 distribution can be built. The second approach is to build GCC
4044 first using the HP tools, then build binutils, then rebuild GCC@.
4045 There have been problems with various binary distributions, so it
4046 is best not to start from a binary distribution.
4048 On 64-bit capable systems, there are two distinct targets. Different
4049 installation prefixes must be used if both are to be installed on
4050 the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code
4051 for the 32-bit PA-RISC runtime architecture and uses the HP linker.
4052 The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the
4053 PA-RISC 2.0 architecture.
4055 The script config.guess now selects the target type based on the compiler
4056 detected during configuration. You must define @env{PATH} or @env{CC} so
4057 that configure finds an appropriate compiler for the initial bootstrap.
4058 When @env{CC} is used, the definition should contain the options that are
4059 needed whenever @env{CC} is used.
4061 Specifically, options that determine the runtime architecture must be
4062 in @env{CC} to correctly select the target for the build. It is also
4063 convenient to place many other compiler options in @env{CC}. For example,
4064 @env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"}
4065 can be used to bootstrap the GCC 3.3 branch with the HP compiler in
4066 64-bit K&R/bundled mode. The @option{+DA2.0W} option will result in
4067 the automatic selection of the @samp{hppa64-hp-hpux11*} target. The
4068 macro definition table of cpp needs to be increased for a successful
4069 build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
4070 be defined when building with the bundled compiler, or when using the
4071 @option{-Ac} option. These defines aren't necessary with @option{-Ae}.
4073 It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target
4074 with the @option{--with-ld=@dots{}} option. This overrides the standard
4075 search for ld. The two linkers supported on this target require different
4076 commands. The default linker is determined during configuration. As a
4077 result, it's not possible to switch linkers in the middle of a GCC build.
4078 This has been reported to sometimes occur in unified builds of binutils
4081 A recent linker patch must be installed for the correct operation of
4082 GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the
4083 oldest linker patches that are known to work. They are for HP-UX
4084 11.00 and 11.11, respectively. @code{PHSS_24303}, the companion to
4085 @code{PHSS_24304}, might be usable but it hasn't been tested. These
4086 patches have been superseded. Consult the HP patch database to obtain
4087 the currently recommended linker patch for your system.
4089 The patches are necessary for the support of weak symbols on the
4090 32-bit port, and for the running of initializers and finalizers. Weak
4091 symbols are implemented using SOM secondary definition symbols. Prior
4092 to HP-UX 11, there are bugs in the linker support for secondary symbols.
4093 The patches correct a problem of linker core dumps creating shared
4094 libraries containing secondary symbols, as well as various other
4095 linking issues involving secondary symbols.
4097 GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
4098 run initializers and finalizers on the 64-bit port. The 32-bit port
4099 uses the linker @option{+init} and @option{+fini} options for the same
4100 purpose. The patches correct various problems with the +init/+fini
4101 options, including program core dumps. Binutils 2.14 corrects a
4102 problem on the 64-bit port resulting from HP's non-standard use of
4103 the .init and .fini sections for array initializers and finalizers.
4105 Although the HP and GNU linkers are both supported for the
4106 @samp{hppa64-hp-hpux11*} target, it is strongly recommended that the
4107 HP linker be used for link editing on this target.
4109 At this time, the GNU linker does not support the creation of long
4110 branch stubs. As a result, it cannot successfully link binaries
4111 containing branch offsets larger than 8 megabytes. In addition,
4112 there are problems linking shared libraries, linking executables
4113 with @option{-static}, and with dwarf2 unwind and exception support.
4114 It also doesn't provide stubs for internal calls to global functions
4115 in shared libraries, so these calls cannot be overloaded.
4117 The HP dynamic loader does not support GNU symbol versioning, so symbol
4118 versioning is not supported. It may be necessary to disable symbol
4119 versioning with @option{--disable-symvers} when using GNU ld.
4121 POSIX threads are the default. The optional DCE thread library is not
4122 supported, so @option{--enable-threads=dce} does not work.
4127 @anchor{x-x-linux-gnu}
4128 @heading *-*-linux-gnu
4129 The @code{.init_array} and @code{.fini_array} sections are enabled
4130 unconditionally which requires at least glibc 2.1 and binutils 2.12.
4132 Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
4133 in glibc 2.2.5 and later. More information is available in the
4134 libstdc++-v3 documentation.
4139 @anchor{ix86-x-linux}
4140 @heading i?86-*-linux*
4141 As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
4142 See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information.
4144 If you receive Signal 11 errors when building on GNU/Linux, then it is
4145 possible you have a hardware problem. Further information on this can be
4146 found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}.
4151 @anchor{ix86-x-solaris2}
4152 @heading i?86-*-solaris2*
4153 Use this for Solaris 11.3 or later on x86 and x86-64 systems. Starting
4154 with GCC 4.7, there is also a 64-bit @samp{amd64-*-solaris2*} or
4155 @samp{x86_64-*-solaris2*} configuration that corresponds to
4156 @samp{sparcv9-sun-solaris2*}.
4158 It is recommended that you configure GCC to use the GNU assembler. The
4159 versions included in Solaris 11.3, from GNU binutils 2.23.1 or
4160 newer (available as @file{/usr/bin/gas} and
4161 @file{/usr/gnu/bin/as}), work fine. The current version, from GNU
4162 binutils 2.34, is known to work. Recent versions of the Solaris assembler in
4163 @file{/usr/bin/as} work almost as well, though.
4165 For linking, the Solaris linker is preferred. If you want to use the GNU
4166 linker instead, the version in Solaris 11.3, from GNU binutils 2.23.1 or
4167 newer (in @file{/usr/gnu/bin/ld} and @file{/usr/bin/gld}), works,
4168 as does the latest version, from GNU binutils 2.34.
4170 To use GNU @command{as}, configure with the options
4171 @option{--with-gnu-as --with-as=@//usr/@/gnu/@/bin/@/as}. It may be necessary
4172 to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to
4173 guarantee use of Solaris @command{ld}.
4174 @c FIXME: why --without-gnu-ld --with-ld?
4179 @anchor{ia64-x-linux}
4180 @heading ia64-*-linux
4181 IA-64 processor (also known as IPF, or Itanium Processor Family)
4184 If you are using the installed system libunwind library with
4185 @option{--with-system-libunwind}, then you must use libunwind 0.98 or
4188 None of the following versions of GCC has an ABI that is compatible
4189 with any of the other versions in this list, with the exception that
4190 Red Hat 2.96 and Trillian 000171 are compatible with each other:
4191 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
4192 This primarily affects C++ programs and programs that create shared libraries.
4193 GCC 3.1 or later is recommended for compiling linux, the kernel.
4194 As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
4195 more major ABI changes are expected.
4200 @anchor{ia64-x-hpux}
4201 @heading ia64-*-hpux*
4202 Building GCC on this target requires the GNU Assembler. The bundled HP
4203 assembler will not work. To prevent GCC from using the wrong assembler,
4204 the option @option{--with-gnu-as} may be necessary.
4206 The GCC libunwind library has not been ported to HPUX@. This means that for
4207 GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions}
4208 is required to build GCC@. For GCC 3.3 and later, this is the default.
4209 For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is
4210 removed and the system libunwind library will always be used.
4214 <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
4218 Support for AIX version 3 and older was discontinued in GCC 3.4.
4219 Support for AIX version 4.2 and older was discontinued in GCC 4.5.
4221 ``out of memory'' bootstrap failures may indicate a problem with
4222 process resource limits (ulimit). Hard limits are configured in the
4223 @file{/etc/security/limits} system configuration file.
4225 GCC 4.9 and above require a C++ compiler for bootstrap. IBM VAC++ / xlC
4226 cannot bootstrap GCC. xlc can bootstrap an older version of GCC and
4227 G++ can bootstrap recent releases of GCC.
4229 GCC can bootstrap with recent versions of IBM XLC, but bootstrapping
4230 with an earlier release of GCC is recommended. Bootstrapping with XLC
4231 requires a larger data segment, which can be enabled through the
4232 @var{LDR_CNTRL} environment variable, e.g.,
4235 % LDR_CNTRL=MAXDATA=0x50000000
4239 One can start with a pre-compiled version of GCC to build from
4240 sources. One may delete GCC's ``fixed'' header files when starting
4241 with a version of GCC built for an earlier release of AIX.
4243 To speed up the configuration phases of bootstrapping and installing GCC,
4244 one may use GNU Bash instead of AIX @command{/bin/sh}, e.g.,
4247 % CONFIG_SHELL=/opt/freeware/bin/bash
4248 % export CONFIG_SHELL
4251 and then proceed as described in @uref{build.html,,the build
4252 instructions}, where we strongly recommend specifying an absolute path
4253 to invoke @var{srcdir}/configure.
4255 Because GCC on AIX is built as a 32-bit executable by default,
4256 (although it can generate 64-bit programs) the GMP and MPFR libraries
4257 required by gfortran must be 32-bit libraries. Building GMP and MPFR
4258 as static archive libraries works better than shared libraries.
4260 Errors involving @code{alloca} when building GCC generally are due
4261 to an incorrect definition of @code{CC} in the Makefile or mixing files
4262 compiled with the native C compiler and GCC@. During the stage1 phase of
4263 the build, the native AIX compiler @strong{must} be invoked as @command{cc}
4264 (not @command{xlc}). Once @command{configure} has been informed of
4265 @command{xlc}, one needs to use @samp{make distclean} to remove the
4266 configure cache files and ensure that @env{CC} environment variable
4267 does not provide a definition that will confuse @command{configure}.
4268 If this error occurs during stage2 or later, then the problem most likely
4269 is the version of Make (see above).
4271 The native @command{as} and @command{ld} are recommended for
4272 bootstrapping on AIX@. The GNU Assembler, GNU Linker, and GNU
4273 Binutils version 2.20 is the minimum level that supports bootstrap on
4274 AIX 5@. The GNU Assembler has not been updated to support AIX 6@ or
4275 AIX 7. The native AIX tools do interoperate with GCC@.
4277 AIX 7.1 added partial support for DWARF debugging, but full support
4278 requires AIX 7.1 TL03 SP7 that supports additional DWARF sections and
4279 fixes a bug in the assembler. AIX 7.1 TL03 SP5 distributed a version
4280 of libm.a missing important symbols; a fix for IV77796 will be
4283 AIX 5.3 TL10, AIX 6.1 TL05 and AIX 7.1 TL00 introduced an AIX
4284 assembler change that sometimes produces corrupt assembly files
4285 causing AIX linker errors. The bug breaks GCC bootstrap on AIX and
4286 can cause compilation failures with existing GCC installations. An
4287 AIX iFix for AIX 5.3 is available (APAR IZ98385 for AIX 5.3 TL10, APAR
4288 IZ98477 for AIX 5.3 TL11 and IZ98134 for AIX 5.3 TL12). AIX 5.3 TL11 SP8,
4289 AIX 5.3 TL12 SP5, AIX 6.1 TL04 SP11, AIX 6.1 TL05 SP7, AIX 6.1 TL06 SP6,
4290 AIX 6.1 TL07 and AIX 7.1 TL01 should include the fix.
4292 Building @file{libstdc++.a} requires a fix for an AIX Assembler bug
4293 APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
4294 fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
4295 referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1)
4297 @anchor{TransferAixShobj}
4298 @samp{libstdc++} in GCC 3.4 increments the major version number of the
4299 shared object and GCC installation places the @file{libstdc++.a}
4300 shared library in a common location which will overwrite the and GCC
4301 3.3 version of the shared library. Applications either need to be
4302 re-linked against the new shared library or the GCC 3.1 and GCC 3.3
4303 versions of the @samp{libstdc++} shared object needs to be available
4304 to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if
4305 present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be
4306 installed for runtime dynamic loading using the following steps to set
4307 the @samp{F_LOADONLY} flag in the shared object for @emph{each}
4308 multilib @file{libstdc++.a} installed:
4310 Extract the shared objects from the currently installed
4311 @file{libstdc++.a} archive:
4313 % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
4316 Enable the @samp{F_LOADONLY} flag so that the shared object will be
4317 available for runtime dynamic loading, but not linking:
4319 % strip -e libstdc++.so.4 libstdc++.so.5
4322 Archive the runtime-only shared object in the GCC 3.4
4323 @file{libstdc++.a} archive:
4325 % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
4329 @uref{./configure.html#WithAixSoname,,@option{--with-aix-soname=svr4}}
4330 configure option may drop the need for this procedure for libraries that
4333 Linking executables and shared libraries may produce warnings of
4334 duplicate symbols. The assembly files generated by GCC for AIX always
4335 have included multiple symbol definitions for certain global variable
4336 and function declarations in the original program. The warnings should
4337 not prevent the linker from producing a correct library or runnable
4340 AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and
4341 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
4342 to parse archive libraries did not handle the new format correctly.
4343 These routines are used by GCC and result in error messages during
4344 linking such as ``not a COFF file''. The version of the routines shipped
4345 with AIX 4.3.1 should work for a 32-bit environment. The @option{-g}
4346 option of the archive command may be used to create archives of 32-bit
4347 objects using the original ``small format''. A correct version of the
4348 routines is shipped with AIX 4.3.2 and above.
4350 Some versions of the AIX binder (linker) can fail with a relocation
4351 overflow severe error when the @option{-bbigtoc} option is used to link
4352 GCC-produced object files into an executable that overflows the TOC@. A fix
4353 for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
4354 available from IBM Customer Support and from its
4355 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4356 website as PTF U455193.
4358 The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
4359 with a segmentation fault when invoked by any version of GCC@. A fix for
4360 APAR IX87327 is available from IBM Customer Support and from its
4361 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4362 website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
4364 The initial assembler shipped with AIX 4.3.0 generates incorrect object
4365 files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
4366 TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
4367 @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com}
4368 website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
4370 AIX provides National Language Support (NLS)@. Compilers and assemblers
4371 use NLS to support locale-specific representations of various data
4372 formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for
4373 separating decimal fractions). There have been problems reported where
4374 GCC does not produce the same floating-point formats that the assembler
4375 expects. If one encounters this problem, set the @env{LANG}
4376 environment variable to @samp{C} or @samp{En_US}.
4378 A default can be specified with the @option{-mcpu=@var{cpu_type}}
4379 switch and using the configure option @option{--with-cpu-@var{cpu_type}}.
4384 @anchor{iq2000-x-elf}
4385 @heading iq2000-*-elf
4386 Vitesse IQ2000 processors. These are used in embedded
4387 applications. There are no standard Unix configurations.
4394 Lattice Mico32 processor.
4395 This configuration is intended for embedded systems.
4400 @anchor{lm32-x-uclinux}
4401 @heading lm32-*-uclinux
4402 Lattice Mico32 processor.
4403 This configuration is intended for embedded systems running uClinux.
4410 Renesas M32C processor.
4411 This configuration is intended for embedded systems.
4418 Renesas M32R processor.
4419 This configuration is intended for embedded systems.
4427 @samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and
4429 build libraries for both M680x0 and ColdFire processors. If you only
4430 need the M680x0 libraries, you can omit the ColdFire ones by passing
4431 @option{--with-arch=m68k} to @command{configure}. Alternatively, you
4432 can omit the M680x0 libraries by passing @option{--with-arch=cf} to
4433 @command{configure}. These targets default to 5206 or 5475 code as
4434 appropriate for the target system when
4435 configured with @option{--with-arch=cf} and 68020 code otherwise.
4437 The @samp{m68k-*-netbsd} and
4438 @samp{m68k-*-openbsd} targets also support the @option{--with-arch}
4439 option. They will generate ColdFire CFV4e code when configured with
4440 @option{--with-arch=cf} and 68020 code otherwise.
4442 You can override the default processors listed above by configuring
4443 with @option{--with-cpu=@var{target}}. This @var{target} can either
4444 be a @option{-mcpu} argument or one of the following values:
4445 @samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030},
4446 @samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}.
4448 GCC requires at least binutils version 2.17 on these targets.
4453 @anchor{m68k-x-uclinux}
4454 @heading m68k-*-uclinux
4455 GCC 4.3 changed the uClinux configuration so that it uses the
4456 @samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI.
4457 It also added improved support for C++ and flat shared libraries,
4458 both of which were ABI changes.
4463 @anchor{microblaze-x-elf}
4464 @heading microblaze-*-elf
4465 Xilinx MicroBlaze processor.
4466 This configuration is intended for embedded systems.
4473 If on a MIPS system you get an error message saying ``does not have gp
4474 sections for all it's [sic] sectons [sic]'', don't worry about it. This
4475 happens whenever you use GAS with the MIPS linker, but there is not
4476 really anything wrong, and it is okay to use the output file. You can
4477 stop such warnings by installing the GNU linker.
4479 It would be nice to extend GAS to produce the gp tables, but they are
4480 optional, and there should not be a warning about their absence.
4482 The libstdc++ atomic locking routines for MIPS targets requires MIPS II
4483 and later. A patch went in just after the GCC 3.3 release to
4484 make @samp{mips*-*-*} use the generic implementation instead. You can also
4485 configure for @samp{mipsel-elf} as a workaround. The
4486 @samp{mips*-*-linux*} target continues to use the MIPS II routines. More
4487 work on this is expected in future releases.
4489 @c If you make --with-llsc the default for another target, please also
4490 @c update the description of the --with-llsc option.
4492 The built-in @code{__sync_*} functions are available on MIPS II and
4493 later systems and others that support the @samp{ll}, @samp{sc} and
4494 @samp{sync} instructions. This can be overridden by passing
4495 @option{--with-llsc} or @option{--without-llsc} when configuring GCC.
4496 Since the Linux kernel emulates these instructions if they are
4497 missing, the default for @samp{mips*-*-linux*} targets is
4498 @option{--with-llsc}. The @option{--with-llsc} and
4499 @option{--without-llsc} configure options may be overridden at compile
4500 time by passing the @option{-mllsc} or @option{-mno-llsc} options to
4503 MIPS systems check for division by zero (unless
4504 @option{-mno-check-zero-division} is passed to the compiler) by
4505 generating either a conditional trap or a break instruction. Using
4506 trap results in smaller code, but is only supported on MIPS II and
4507 later. Also, some versions of the Linux kernel have a bug that
4508 prevents trap from generating the proper signal (@code{SIGFPE}). To enable
4509 the use of break, use the @option{--with-divide=breaks}
4510 @command{configure} option when configuring GCC@. The default is to
4511 use traps on systems that support them.
4516 @anchor{moxie-x-elf}
4517 @heading moxie-*-elf
4518 The moxie processor.
4523 @anchor{msp430-x-elf}
4524 @heading msp430-*-elf*
4525 TI MSP430 processor.
4526 This configuration is intended for embedded systems.
4528 @samp{msp430-*-elf} is the standard configuration with most GCC
4529 features enabled by default.
4531 @samp{msp430-*-elfbare} is tuned for a bare-metal environment, and disables
4532 features related to shared libraries and other functionality not used for
4533 this device. This reduces code and data usage of the GCC libraries, resulting
4534 in a minimal run-time environment by default.
4536 Features disabled by default include:
4538 @item transactional memory
4545 @anchor{nds32le-x-elf}
4546 @heading nds32le-*-elf
4547 Andes NDS32 target in little endian mode.
4552 @anchor{nds32be-x-elf}
4553 @heading nds32be-*-elf
4554 Andes NDS32 target in big endian mode.
4559 @anchor{nvptx-x-none}
4560 @heading nvptx-*-none
4563 Instead of GNU binutils, you will need to install
4564 @uref{https://github.com/MentorEmbedded/nvptx-tools/,,nvptx-tools}.
4565 Tell GCC where to find it:
4566 @option{--with-build-time-tools=[install-nvptx-tools]/nvptx-none/bin}.
4568 You will need newlib 3.0 git revision
4569 cd31fbb2aea25f94d7ecedc9db16dfc87ab0c316 or later. It can be
4570 automatically built together with GCC@. For this, add a symbolic link
4571 to nvptx-newlib's @file{newlib} directory to the directory containing
4574 Use the @option{--disable-sjlj-exceptions} and
4575 @option{--enable-newlib-io-long-long} options when configuring.
4582 The OpenRISC 1000 32-bit processor with delay slots.
4583 This configuration is intended for embedded systems.
4588 @anchor{or1k-x-linux}
4589 @heading or1k-*-linux
4590 The OpenRISC 1000 32-bit processor with delay slots.
4595 @anchor{powerpc-x-x}
4596 @heading powerpc-*-*
4597 You can specify a default version for the @option{-mcpu=@var{cpu_type}}
4598 switch by using the configure option @option{--with-cpu-@var{cpu_type}}.
4600 You will need GNU binutils 2.20 or newer.
4605 @anchor{powerpc-x-darwin}
4606 @heading powerpc-*-darwin*
4607 PowerPC running Darwin (Mac OS X kernel).
4609 Pre-installed versions of Mac OS X may not include any developer tools,
4610 meaning that you will not be able to build GCC from source. Tool
4611 binaries are available at
4612 @uref{https://opensource.apple.com}.
4614 This version of GCC requires at least cctools-590.36. The
4615 cctools-590.36 package referenced from
4616 @uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work
4617 on systems older than 10.3.9 (aka darwin7.9.0).
4622 @anchor{powerpc-x-elf}
4623 @heading powerpc-*-elf
4624 PowerPC system in big endian mode, running System V.4.
4629 @anchor{powerpc-x-linux-gnu}
4630 @heading powerpc*-*-linux-gnu*
4631 PowerPC system in big endian mode running Linux.
4636 @anchor{powerpc-x-netbsd}
4637 @heading powerpc-*-netbsd*
4638 PowerPC system in big endian mode running NetBSD@.
4643 @anchor{powerpc-x-eabisim}
4644 @heading powerpc-*-eabisim
4645 Embedded PowerPC system in big endian mode for use in running under the
4651 @anchor{powerpc-x-eabi}
4652 @heading powerpc-*-eabi
4653 Embedded PowerPC system in big endian mode.
4658 @anchor{powerpcle-x-elf}
4659 @heading powerpcle-*-elf
4660 PowerPC system in little endian mode, running System V.4.
4665 @anchor{powerpcle-x-eabisim}
4666 @heading powerpcle-*-eabisim
4667 Embedded PowerPC system in little endian mode for use in running under
4673 @anchor{powerpcle-x-eabi}
4674 @heading powerpcle-*-eabi
4675 Embedded PowerPC system in little endian mode.
4682 The Renesas RL78 processor.
4683 This configuration is intended for embedded systems.
4688 @anchor{riscv32-x-elf}
4689 @heading riscv32-*-elf
4690 The RISC-V RV32 instruction set.
4691 This configuration is intended for embedded systems.
4692 This (and all other RISC-V) targets require the binutils 2.30 release.
4697 @anchor{riscv32-x-linux}
4698 @heading riscv32-*-linux
4699 The RISC-V RV32 instruction set running GNU/Linux.
4700 This (and all other RISC-V) targets require the binutils 2.30 release.
4705 @anchor{riscv64-x-elf}
4706 @heading riscv64-*-elf
4707 The RISC-V RV64 instruction set.
4708 This configuration is intended for embedded systems.
4709 This (and all other RISC-V) targets require the binutils 2.30 release.
4714 @anchor{riscv64-x-linux}
4715 @heading riscv64-*-linux
4716 The RISC-V RV64 instruction set running GNU/Linux.
4717 This (and all other RISC-V) targets require the binutils 2.30 release.
4724 The Renesas RX processor.
4729 @anchor{s390-x-linux}
4730 @heading s390-*-linux*
4731 S/390 system running GNU/Linux for S/390@.
4736 @anchor{s390x-x-linux}
4737 @heading s390x-*-linux*
4738 zSeries system (64-bit) running GNU/Linux for zSeries@.
4743 @anchor{s390x-ibm-tpf}
4744 @heading s390x-ibm-tpf*
4745 zSeries system (64-bit) running TPF@. This platform is
4746 supported as cross-compilation target only.
4751 @c Please use Solaris 2 to refer to all release of Solaris, starting
4752 @c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for
4753 @c SunOS 4 releases which we don't use to avoid confusion. Solaris
4754 @c alone is too unspecific and must be avoided.
4755 @anchor{x-x-solaris2}
4756 @heading *-*-solaris2*
4757 Support for Solaris 10 has been removed in GCC 10. Support for Solaris
4758 9 has been removed in GCC 5. Support for Solaris 8 has been removed in
4759 GCC 4.8. Support for Solaris 7 has been removed in GCC 4.6.
4761 Solaris 11.3 provides GCC 4.5.2, 4.7.3, and 4.8.2 as
4762 @command{/usr/gcc/4.5/bin/gcc} or similar. Newer Solaris versions
4763 provide one or more of GCC 5, 7, and 9. Alternatively,
4764 you can install a pre-built GCC to bootstrap and install GCC. See the
4765 @uref{binaries.html,,binaries page} for details.
4767 The Solaris 2 @command{/bin/sh} will often fail to configure
4768 @samp{libstdc++-v3}. We therefore recommend using the
4769 following initial sequence of commands
4772 % CONFIG_SHELL=/bin/ksh
4773 % export CONFIG_SHELL
4777 and proceed as described in @uref{configure.html,,the configure instructions}.
4778 In addition we strongly recommend specifying an absolute path to invoke
4779 @command{@var{srcdir}/configure}.
4781 In Solaris 11, you need to check for @code{system/header},
4782 @code{system/linker}, and @code{developer/assembler} packages.
4784 Trying to use the linker and other tools in
4785 @file{/usr/ucb} to install GCC has been observed to cause trouble.
4786 For example, the linker may hang indefinitely. The fix is to remove
4787 @file{/usr/ucb} from your @env{PATH}.
4789 The build process works more smoothly with the legacy Solaris tools so, if you
4790 have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place
4791 @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build.
4793 We recommend the use of the Solaris assembler or the GNU assembler, in
4794 conjunction with the Solaris linker. The GNU @command{as}
4795 versions included in Solaris 11.3,
4796 from GNU binutils 2.23.1 or newer (in @file{/usr/bin/gas} and
4797 @file{/usr/gnu/bin/as}), are known to work.
4798 The current version, from GNU binutils 2.34,
4799 is known to work as well. Note that your mileage may vary
4800 if you use a combination of the GNU tools and the Solaris tools: while the
4801 combination GNU @command{as} + Solaris @command{ld} should reasonably work,
4802 the reverse combination Solaris @command{as} + GNU @command{ld} may fail to
4803 build or cause memory corruption at runtime in some cases for C++ programs.
4805 GNU @command{ld} usually works as well. Again, the current
4806 version (2.34) is known to work, but generally lacks platform specific
4807 features, so better stay with Solaris @command{ld}. To use the LTO linker
4808 plugin (@option{-fuse-linker-plugin}) with GNU @command{ld}, GNU
4809 binutils @emph{must} be configured with @option{--enable-largefile}.
4811 To enable symbol versioning in @samp{libstdc++} with the Solaris linker,
4812 you need to have any version of GNU @command{c++filt}, which is part of
4813 GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no
4814 appropriate version is found. Solaris @command{c++filt} from the Solaris
4815 Studio compilers does @emph{not} work.
4817 The versions of the GNU Multiple Precision Library (GMP), the MPFR
4818 library and the MPC library bundled with Solaris 11.3 and later are
4819 usually recent enough to match GCC's requirements. There are two
4824 While the version of the GMP library in Solaris 11.3 works with GCC, you
4825 need to configure with @option{--with-gmp-include=/usr/include/gmp}.
4828 The version of the MPFR libary included in Solaris 11.3 is too old; you
4829 need to provide a more recent one.
4837 This section contains general configuration information for all
4838 SPARC-based platforms. In addition to reading this section, please
4839 read all other sections that match your target.
4841 Newer versions of the GNU Multiple Precision Library (GMP), the MPFR
4842 library and the MPC library are known to be miscompiled by earlier
4843 versions of GCC on these platforms. We therefore recommend the use
4844 of the exact versions of these libraries listed as minimal versions
4845 in @uref{prerequisites.html,,the prerequisites}.
4850 @anchor{sparc-sun-solaris2}
4851 @heading sparc-sun-solaris2*
4852 When GCC is configured to use GNU binutils 2.14 or later, the binaries
4853 produced are smaller than the ones produced using Solaris native tools;
4854 this difference is quite significant for binaries containing debugging
4857 Starting with Solaris 7, the operating system is capable of executing
4858 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
4859 this; the @option{-m64} option enables 64-bit code generation.
4860 However, if all you want is code tuned for the UltraSPARC CPU, you
4861 should try the @option{-mtune=ultrasparc} option instead, which produces
4862 code that, unlike full 64-bit code, can still run on non-UltraSPARC
4865 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4866 library or the MPC library on a Solaris 7 or later system, the canonical
4867 target triplet must be specified as the @command{build} parameter on the
4868 configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and
4869 not that of GMP or MPFR or MPC). For example on a Solaris 11 system:
4872 % ./configure --build=sparc-sun-solaris2.11 --prefix=xxx
4878 @anchor{sparc-x-linux}
4879 @heading sparc-*-linux*
4884 @anchor{sparc64-x-solaris2}
4885 @heading sparc64-*-solaris2*
4886 When configuring a 64-bit-default GCC on Solaris/SPARC, you must use a
4887 build compiler that generates 64-bit code, either by default or by
4888 specifying @samp{CC='gcc -m64' CXX='gcc-m64'} to @command{configure}.
4889 Additionally, you @emph{must} pass @option{--build=sparc64-sun-solaris2.11}
4890 or @option{--build=sparcv9-sun-solaris2.11} because @file{config.guess}
4891 misdetects this situation, which can cause build failures.
4893 When configuring the GNU Multiple Precision Library (GMP), the MPFR
4894 library or the MPC library, the canonical target triplet must be specified
4895 as the @command{build} parameter on the configure line. For example
4896 on a Solaris 11 system:
4899 % ./configure --build=sparc64-sun-solaris2.11 --prefix=xxx
4905 @anchor{sparcv9-x-solaris2}
4906 @heading sparcv9-*-solaris2*
4907 This is a synonym for @samp{sparc64-*-solaris2*}.
4914 The C6X family of processors. This port requires binutils-2.22 or newer.
4919 @anchor{tilegx-*-linux}
4920 @heading tilegx-*-linux*
4921 The TILE-Gx processor in little endian mode, running GNU/Linux. This
4922 port requires binutils-2.22 or newer.
4927 @anchor{tilegxbe-*-linux}
4928 @heading tilegxbe-*-linux*
4929 The TILE-Gx processor in big endian mode, running GNU/Linux. This
4930 port requires binutils-2.23 or newer.
4935 @anchor{tilepro-*-linux}
4936 @heading tilepro-*-linux*
4937 The TILEPro processor running GNU/Linux. This port requires
4938 binutils-2.22 or newer.
4943 @anchor{visium-x-elf}
4944 @heading visium-*-elf
4945 CDS VISIUMcore processor.
4946 This configuration is intended for embedded systems.
4951 @anchor{x-x-vxworks}
4952 @heading *-*-vxworks*
4953 Support for VxWorks is in flux. At present GCC supports @emph{only} the
4954 very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@.
4955 We welcome patches for other architectures supported by VxWorks 5.5.
4956 Support for VxWorks AE would also be welcome; we believe this is merely
4957 a matter of writing an appropriate ``configlette'' (see below). We are
4958 not interested in supporting older, a.out or COFF-based, versions of
4961 VxWorks comes with an older version of GCC installed in
4962 @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it.
4963 Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}.
4964 Before running @command{configure}, create the directories @file{@var{prefix}}
4965 and @file{@var{prefix}/bin}. Link or copy the appropriate assembler,
4966 linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to
4967 include that directory while running both @command{configure} and
4970 You must give @command{configure} the
4971 @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can
4972 find the VxWorks system headers. Since VxWorks is a cross compilation
4973 target only, you must also specify @option{--target=@var{target}}.
4974 @command{configure} will attempt to create the directory
4975 @file{@var{prefix}/@var{target}/sys-include} and copy files into it;
4976 make sure the user running @command{configure} has sufficient privilege
4979 GCC's exception handling runtime requires a special ``configlette''
4980 module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in
4981 that file to add the module to your kernel build. (Future versions of
4982 VxWorks will incorporate this module.)
4988 @heading x86_64-*-*, amd64-*-*
4989 GCC supports the x86-64 architecture implemented by the AMD64 processor
4990 (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@.
4991 On GNU/Linux the default is a bi-arch compiler which is able to generate
4992 both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch).
4997 @anchor{x86-64-x-solaris2}
4998 @heading x86_64-*-solaris2*
4999 GCC also supports the x86-64 architecture implemented by the AMD64
5000 processor (@samp{amd64-*-*} is an alias for @samp{x86_64-*-*}) on
5001 Solaris 10 or later. Unlike other systems, without special options a
5002 bi-arch compiler is built which generates 32-bit code by default, but
5003 can generate 64-bit x86-64 code with the @option{-m64} switch. Since
5004 GCC 4.7, there is also a configuration that defaults to 64-bit code, but
5005 can generate 32-bit code with @option{-m32}. To configure and build
5006 this way, you have to provide all support libraries like @file{libgmp}
5007 as 64-bit code, configure with @option{--target=x86_64-pc-solaris2.11}
5008 and @samp{CC=gcc -m64}.
5013 @anchor{xtensa-x-elf}
5014 @heading xtensa*-*-elf
5015 This target is intended for embedded Xtensa systems using the
5016 @samp{newlib} C library. It uses ELF but does not support shared
5017 objects. Designed-defined instructions specified via the
5018 Tensilica Instruction Extension (TIE) language are only supported
5019 through inline assembly.
5021 The Xtensa configuration information must be specified prior to
5022 building GCC@. The @file{include/xtensa-config.h} header
5023 file contains the configuration information. If you created your
5024 own Xtensa configuration with the Xtensa Processor Generator, the
5025 downloaded files include a customized copy of this header file,
5026 which you can use to replace the default header file.
5031 @anchor{xtensa-x-linux}
5032 @heading xtensa*-*-linux*
5033 This target is for Xtensa systems running GNU/Linux. It supports ELF
5034 shared objects and the GNU C library (glibc). It also generates
5035 position-independent code (PIC) regardless of whether the
5036 @option{-fpic} or @option{-fPIC} options are used. In other
5037 respects, this target is the same as the
5038 @uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target.
5044 @heading Microsoft Windows
5046 @subheading Intel 16-bit versions
5047 The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
5050 However, the 32-bit port has limited support for Microsoft
5051 Windows 3.11 in the Win32s environment, as a target only. See below.
5053 @subheading Intel 32-bit versions
5054 The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
5055 XP, and Windows Vista, are supported by several different target
5056 platforms. These targets differ in which Windows subsystem they target
5057 and which C libraries are used.
5060 @item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space
5061 Linux API emulation layer in the Win32 subsystem.
5062 @item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for
5063 the Win32 subsystem that provides a subset of POSIX.
5064 @item MKS i386-pc-mks: NuTCracker from MKS. See
5065 @uref{https://www.mkssoftware.com} for more information.
5068 @subheading Intel 64-bit versions
5069 GCC contains support for x86-64 using the mingw-w64
5070 runtime library, available from @uref{http://mingw-w64.org/doku.php}.
5071 This library should be used with the target triple x86_64-pc-mingw32.
5073 Presently Windows for Itanium is not supported.
5075 @subheading Windows CE
5076 Windows CE is supported as a target only on Hitachi
5077 SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
5079 @subheading Other Windows Platforms
5080 GCC no longer supports Windows NT on the Alpha or PowerPC.
5082 GCC no longer supports the Windows POSIX subsystem. However, it does
5083 support the Interix subsystem. See above.
5085 Old target names including *-*-winnt and *-*-windowsnt are no longer used.
5087 PW32 (i386-pc-pw32) support was never completed, and the project seems to
5088 be inactive. See @uref{http://pw32.sourceforge.net/} for more information.
5090 UWIN support has been removed due to a lack of maintenance.
5097 Ports of GCC are included with the
5098 @uref{http://www.cygwin.com/,,Cygwin environment}.
5100 GCC will build under Cygwin without modification; it does not build
5101 with Microsoft's C++ compiler and there are no plans to make it do so.
5103 The Cygwin native compiler can be configured to target any 32-bit x86
5104 cpu architecture desired; the default is i686-pc-cygwin. It should be
5105 used with as up-to-date a version of binutils as possible; use either
5106 the latest official GNU binutils release in the Cygwin distribution,
5107 or version 2.20 or above if building your own.
5112 @anchor{x-x-mingw32}
5113 @heading *-*-mingw32
5114 GCC will build with and support only MinGW runtime 3.12 and later.
5115 Earlier versions of headers are incompatible with the new default semantics
5116 of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes.
5122 @heading Older systems
5123 GCC contains support files for many older (1980s and early
5124 1990s) Unix variants. For the most part, support for these systems
5125 has not been deliberately removed, but it has not been maintained for
5126 several years and may suffer from bitrot.
5128 Starting with GCC 3.1, each release has a list of ``obsoleted'' systems.
5129 Support for these systems is still present in that release, but
5130 @command{configure} will fail unless the @option{--enable-obsolete}
5131 option is given. Unless a maintainer steps forward, support for these
5132 systems will be removed from the next release of GCC@.
5134 Support for old systems as hosts for GCC can cause problems if the
5135 workarounds for compiler, library and operating system bugs affect the
5136 cleanliness or maintainability of the rest of GCC@. In some cases, to
5137 bring GCC up on such a system, if still possible with current GCC, may
5138 require first installing an old version of GCC which did work on that
5139 system, and using it to compile a more recent GCC, to avoid bugs in the
5140 vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
5141 @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror
5142 sites}. Header bugs may generally be avoided using
5143 @command{fixincludes}, but bugs or deficiencies in libraries and the
5144 operating system may still cause problems.
5146 Support for older systems as targets for cross-compilation is less
5147 problematic than support for them as hosts for GCC; if an enthusiast
5148 wishes to make such a target work again (including resurrecting any of
5149 the targets that never worked with GCC 2, starting from the last
5150 version before they were removed), patches
5151 @uref{../contribute.html,,following the usual requirements} would be
5152 likely to be accepted, since they should not affect the support for more
5155 For some systems, old versions of GNU binutils may also be useful,
5156 and are available from @file{pub/binutils/old-releases} on
5157 @uref{https://sourceware.org/mirrors.html,,sourceware.org mirror sites}.
5159 Some of the information on specific systems above relates to
5160 such older systems, but much of the information
5161 about GCC on such systems (which may no longer be applicable to
5162 current GCC) is to be found in the GCC texinfo manual.
5168 @heading all ELF targets (SVR4, Solaris 2, etc.)
5169 C++ support is significantly better on ELF targets if you use the
5170 @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of
5171 inlines, vtables and template instantiations will be discarded
5180 @uref{./index.html,,Return to the GCC Installation page}
5184 @c ***GFDL********************************************************************
5192 @uref{./index.html,,Return to the GCC Installation page}
5196 @c ***************************************************************************
5197 @c Part 6 The End of the Document
5199 @comment node-name, next, previous, up
5200 @node Concept Index, , GNU Free Documentation License, Top
5204 @unnumbered Concept Index