1\input texinfo.tex @c -*-texinfo-*- 2@c @ifnothtml 3@c %**start of header 4@setfilename gccinstall.info 5@settitle Installing GCC 6@setchapternewpage odd 7@c %**end of header 8@c @end ifnothtml 9 10@include gcc-common.texi 11 12@c Specify title for specific html page 13@ifset indexhtml 14@settitle Installing GCC 15@end ifset 16@ifset specifichtml 17@settitle Host/Target specific installation notes for GCC 18@end ifset 19@ifset prerequisiteshtml 20@settitle Prerequisites for GCC 21@end ifset 22@ifset downloadhtml 23@settitle Downloading GCC 24@end ifset 25@ifset configurehtml 26@settitle Installing GCC: Configuration 27@end ifset 28@ifset buildhtml 29@settitle Installing GCC: Building 30@end ifset 31@ifset testhtml 32@settitle Installing GCC: Testing 33@end ifset 34@ifset finalinstallhtml 35@settitle Installing GCC: Final installation 36@end ifset 37@ifset binarieshtml 38@settitle Installing GCC: Binaries 39@end ifset 40@ifset oldhtml 41@settitle Installing GCC: Old documentation 42@end ifset 43@ifset gfdlhtml 44@settitle Installing GCC: GNU Free Documentation License 45@end ifset 46 47@c Copyright (C) 1988-2016 Free Software Foundation, Inc. 48@c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com 49 50@c IMPORTANT: whenever you modify this file, run `install.texi2html' to 51@c test the generation of HTML documents for the gcc.gnu.org web pages. 52@c 53@c Do not use @footnote{} in this file as it breaks install.texi2html! 54 55@c Include everything if we're not making html 56@ifnothtml 57@set indexhtml 58@set specifichtml 59@set prerequisiteshtml 60@set downloadhtml 61@set configurehtml 62@set buildhtml 63@set testhtml 64@set finalinstallhtml 65@set binarieshtml 66@set oldhtml 67@set gfdlhtml 68@end ifnothtml 69 70@c Part 2 Summary Description and Copyright 71@copying 72Copyright @copyright{} 1988-2017 Free Software Foundation, Inc. 73@sp 1 74Permission is granted to copy, distribute and/or modify this document 75under the terms of the GNU Free Documentation License, Version 1.3 or 76any later version published by the Free Software Foundation; with no 77Invariant Sections, the Front-Cover texts being (a) (see below), and 78with the Back-Cover Texts being (b) (see below). A copy of the 79license is included in the section entitled ``@uref{./gfdl.html,,GNU 80Free Documentation License}''. 81 82(a) The FSF's Front-Cover Text is: 83 84 A GNU Manual 85 86(b) The FSF's Back-Cover Text is: 87 88 You have freedom to copy and modify this GNU Manual, like GNU 89 software. Copies published by the Free Software Foundation raise 90 funds for GNU development. 91@end copying 92@ifinfo 93@insertcopying 94@end ifinfo 95@dircategory Software development 96@direntry 97* gccinstall: (gccinstall). Installing the GNU Compiler Collection. 98@end direntry 99 100@c Part 3 Titlepage and Copyright 101@titlepage 102@title Installing GCC 103@versionsubtitle 104 105@c The following two commands start the copyright page. 106@page 107@vskip 0pt plus 1filll 108@insertcopying 109@end titlepage 110 111@c Part 4 Top node, Master Menu, and/or Table of Contents 112@ifinfo 113@node Top, , , (dir) 114@comment node-name, next, Previous, up 115 116@menu 117* Installing GCC:: This document describes the generic installation 118 procedure for GCC as well as detailing some target 119 specific installation instructions. 120 121* Specific:: Host/target specific installation notes for GCC. 122* Binaries:: Where to get pre-compiled binaries. 123 124* Old:: Old installation documentation. 125 126* GNU Free Documentation License:: How you can copy and share this manual. 127* Concept Index:: This index has two entries. 128@end menu 129@end ifinfo 130 131@iftex 132@contents 133@end iftex 134 135@c Part 5 The Body of the Document 136@c ***Installing GCC********************************************************** 137@ifnothtml 138@comment node-name, next, previous, up 139@node Installing GCC, Binaries, , Top 140@end ifnothtml 141@ifset indexhtml 142@ifnothtml 143@chapter Installing GCC 144@end ifnothtml 145 146The latest version of this document is always available at 147@uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}. 148It refers to the current development sources, instructions for 149specific released versions are included with the sources. 150 151This document describes the generic installation procedure for GCC as well 152as detailing some target specific installation instructions. 153 154GCC includes several components that previously were separate distributions 155with their own installation instructions. This document supersedes all 156package-specific installation instructions. 157 158@emph{Before} starting the build/install procedure please check the 159@ifnothtml 160@ref{Specific, host/target specific installation notes}. 161@end ifnothtml 162@ifhtml 163@uref{specific.html,,host/target specific installation notes}. 164@end ifhtml 165We recommend you browse the entire generic installation instructions before 166you proceed. 167 168Lists of successful builds for released versions of GCC are 169available at @uref{http://gcc.gnu.org/buildstat.html}. 170These lists are updated as new information becomes available. 171 172The installation procedure itself is broken into five steps. 173 174@ifinfo 175@menu 176* Prerequisites:: 177* Downloading the source:: 178* Configuration:: 179* Building:: 180* Testing:: (optional) 181* Final install:: 182@end menu 183@end ifinfo 184@ifhtml 185@enumerate 186@item 187@uref{prerequisites.html,,Prerequisites} 188@item 189@uref{download.html,,Downloading the source} 190@item 191@uref{configure.html,,Configuration} 192@item 193@uref{build.html,,Building} 194@item 195@uref{test.html,,Testing} (optional) 196@item 197@uref{finalinstall.html,,Final install} 198@end enumerate 199@end ifhtml 200 201Please note that GCC does not support @samp{make uninstall} and probably 202won't do so in the near future as this would open a can of worms. Instead, 203we suggest that you install GCC into a directory of its own and simply 204remove that directory when you do not need that specific version of GCC 205any longer, and, if shared libraries are installed there as well, no 206more binaries exist that use them. 207 208@ifhtml 209There are also some @uref{old.html,,old installation instructions}, 210which are mostly obsolete but still contain some information which has 211not yet been merged into the main part of this manual. 212@end ifhtml 213 214@html 215<hr /> 216<p> 217@end html 218@ifhtml 219@uref{./index.html,,Return to the GCC Installation page} 220 221@insertcopying 222@end ifhtml 223@end ifset 224 225@c ***Prerequisites************************************************** 226@ifnothtml 227@comment node-name, next, previous, up 228@node Prerequisites, Downloading the source, , Installing GCC 229@end ifnothtml 230@ifset prerequisiteshtml 231@ifnothtml 232@chapter Prerequisites 233@end ifnothtml 234@cindex Prerequisites 235 236GCC requires that various tools and packages be available for use in the 237build procedure. Modifying GCC sources requires additional tools 238described below. 239 240@heading Tools/packages necessary for building GCC 241@table @asis 242@item ISO C++98 compiler 243Necessary to bootstrap GCC, although versions of GCC prior 244to 4.8 also allow bootstrapping with a ISO C89 compiler and versions 245of GCC prior to 3.4 also allow bootstrapping with a traditional 246(K&R) C compiler. 247 248To build all languages in a cross-compiler or other configuration where 2493-stage bootstrap is not performed, you need to start with an existing 250GCC binary (version 3.4 or later) because source code for language 251frontends other than C might use GCC extensions. 252 253Note that to bootstrap GCC with versions of GCC earlier than 3.4, you 254may need to use @option{--disable-stage1-checking}, though 255bootstrapping the compiler with such earlier compilers is strongly 256discouraged. 257 258@item C standard library and headers 259 260In order to build GCC, the C standard library and headers must be present 261for all target variants for which target libraries will be built (and not 262only the variant of the host C++ compiler). 263 264This affects the popular @samp{x86_64-unknown-linux-gnu} platform (among 265other multilib targets), for which 64-bit (@samp{x86_64}) and 32-bit 266(@samp{i386}) libc headers are usually packaged separately. If you do a 267build of a native compiler on @samp{x86_64-unknown-linux-gnu}, make sure you 268either have the 32-bit libc developer package properly installed (the exact 269name of the package depends on your distro) or you must build GCC as a 27064-bit only compiler by configuring with the option 271@option{--disable-multilib}. Otherwise, you may encounter an error such as 272@samp{fatal error: gnu/stubs-32.h: No such file} 273 274@item GNAT 275 276In order to build the Ada compiler (GNAT) you must already have GNAT 277installed because portions of the Ada frontend are written in Ada (with 278GNAT extensions.) Refer to the Ada installation instructions for more 279specific information. 280 281@item A ``working'' POSIX compatible shell, or GNU bash 282 283Necessary when running @command{configure} because some 284@command{/bin/sh} shells have bugs and may crash when configuring the 285target libraries. In other cases, @command{/bin/sh} or @command{ksh} 286have disastrous corner-case performance problems. This 287can cause target @command{configure} runs to literally take days to 288complete in some cases. 289 290So on some platforms @command{/bin/ksh} is sufficient, on others it 291isn't. See the host/target specific instructions for your platform, or 292use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your 293environment to your ``good'' shell prior to running 294@command{configure}/@command{make}. 295 296@command{zsh} is not a fully compliant POSIX shell and will not 297work when configuring GCC@. 298 299@item A POSIX or SVR4 awk 300 301Necessary for creating some of the generated source files for GCC@. 302If in doubt, use a recent GNU awk version, as some of the older ones 303are broken. GNU awk version 3.1.5 is known to work. 304 305@item GNU binutils 306 307Necessary in some circumstances, optional in others. See the 308host/target specific instructions for your platform for the exact 309requirements. 310 311@item gzip version 1.2.4 (or later) or 312@itemx bzip2 version 1.0.2 (or later) 313 314Necessary to uncompress GCC @command{tar} files when source code is 315obtained via FTP mirror sites. 316 317@item GNU make version 3.80 (or later) 318 319You must have GNU make installed to build GCC@. 320 321@item GNU tar version 1.14 (or later) 322 323Necessary (only on some platforms) to untar the source code. Many 324systems' @command{tar} programs will also work, only try GNU 325@command{tar} if you have problems. 326 327@item Perl version 5.6.1 (or later) 328 329Necessary when targeting Darwin, building @samp{libstdc++}, 330and not using @option{--disable-symvers}. 331Necessary when targeting Solaris 2 with Sun @command{ld} and not using 332@option{--disable-symvers}. The bundled @command{perl} in Solaris@tie{}8 333and up works. 334 335Necessary when regenerating @file{Makefile} dependencies in libiberty. 336Necessary when regenerating @file{libiberty/functions.texi}. 337Necessary when generating manpages from Texinfo manuals. 338Used by various scripts to generate some files included in SVN (mainly 339Unicode-related and rarely changing) from source tables. 340 341@item @command{jar}, or InfoZIP (@command{zip} and @command{unzip}) 342 343Necessary to build libgcj, the GCJ runtime. 344 345@end table 346 347Several support libraries are necessary to build GCC, some are required, 348others optional. While any sufficiently new version of required tools 349usually work, library requirements are generally stricter. Newer 350versions may work in some cases, but it's safer to use the exact 351versions documented. We appreciate bug reports about problems with 352newer versions, though. If your OS vendor provides packages for the 353support libraries then using those packages may be the simplest way to 354install the libraries. 355 356@table @asis 357@item GNU Multiple Precision Library (GMP) version 4.3.2 (or later) 358 359Necessary to build GCC@. If a GMP source distribution is found in a 360subdirectory of your GCC sources named @file{gmp}, it will be built 361together with GCC. Alternatively, if GMP is already installed but it 362is not in your library search path, you will have to configure with the 363@option{--with-gmp} configure option. See also @option{--with-gmp-lib} 364and @option{--with-gmp-include}. 365 366@item MPFR Library version 2.4.2 (or later) 367 368Necessary to build GCC@. It can be downloaded from 369@uref{http://www.mpfr.org/}. If an MPFR source distribution is found 370in a subdirectory of your GCC sources named @file{mpfr}, it will be 371built together with GCC. Alternatively, if MPFR is already installed 372but it is not in your default library search path, the 373@option{--with-mpfr} configure option should be used. See also 374@option{--with-mpfr-lib} and @option{--with-mpfr-include}. 375 376@item MPC Library version 0.8.1 (or later) 377 378Necessary to build GCC@. It can be downloaded from 379@uref{http://www.multiprecision.org/}. If an MPC source distribution 380is found in a subdirectory of your GCC sources named @file{mpc}, it 381will be built together with GCC. Alternatively, if MPC is already 382installed but it is not in your default library search path, the 383@option{--with-mpc} configure option should be used. See also 384@option{--with-mpc-lib} and @option{--with-mpc-include}. 385 386@item isl Library version 0.16, 0.15, or 0.14. 387 388Necessary to build GCC with the Graphite loop optimizations. 389It can be downloaded from @uref{ftp://gcc.gnu.org/pub/gcc/infrastructure/}. 390If an isl source distribution is found 391in a subdirectory of your GCC sources named @file{isl}, it will be 392built together with GCC. Alternatively, the @option{--with-isl} configure 393option should be used if isl is not installed in your default library 394search path. 395 396@end table 397 398@heading Tools/packages necessary for modifying GCC 399@table @asis 400@item autoconf version 2.64 401@itemx GNU m4 version 1.4.6 (or later) 402 403Necessary when modifying @file{configure.ac}, @file{aclocal.m4}, etc.@: 404to regenerate @file{configure} and @file{config.in} files. 405 406@item automake version 1.11.6 407 408Necessary when modifying a @file{Makefile.am} file to regenerate its 409associated @file{Makefile.in}. 410 411Much of GCC does not use automake, so directly edit the @file{Makefile.in} 412file. Specifically this applies to the @file{gcc}, @file{intl}, 413@file{libcpp}, @file{libiberty}, @file{libobjc} directories as well 414as any of their subdirectories. 415 416For directories that use automake, GCC requires the latest release in 417the 1.11 series, which is currently 1.11.6. When regenerating a directory 418to a newer version, please update all the directories using an older 1.11 419to the latest released version. 420 421@item gettext version 0.14.5 (or later) 422 423Needed to regenerate @file{gcc.pot}. 424 425@item gperf version 2.7.2 (or later) 426 427Necessary when modifying @command{gperf} input files, e.g.@: 428@file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@: 429@file{gcc/cp/cfns.h}. 430 431@item DejaGnu 1.4.4 432@itemx Expect 433@itemx Tcl 434 435Necessary to run the GCC testsuite; see the section on testing for 436details. Tcl 8.6 has a known regression in RE pattern handling that 437make parts of the testsuite fail. See 438@uref{http://core.tcl.tk/tcl/tktview/267b7e2334ee2e9de34c4b00d6e72e2f1997085f} 439for more information. This bug has been fixed in 8.6.1. 440 441@item autogen version 5.5.4 (or later) and 442@itemx guile version 1.4.1 (or later) 443 444Necessary to regenerate @file{fixinc/fixincl.x} from 445@file{fixinc/inclhack.def} and @file{fixinc/*.tpl}. 446 447Necessary to run @samp{make check} for @file{fixinc}. 448 449Necessary to regenerate the top level @file{Makefile.in} file from 450@file{Makefile.tpl} and @file{Makefile.def}. 451 452@item Flex version 2.5.4 (or later) 453 454Necessary when modifying @file{*.l} files. 455 456Necessary to build GCC during development because the generated output 457files are not included in the SVN repository. They are included in 458releases. 459 460@item Texinfo version 4.7 (or later) 461 462Necessary for running @command{makeinfo} when modifying @file{*.texi} 463files to test your changes. 464 465Necessary for running @command{make dvi} or @command{make pdf} to 466create printable documentation in DVI or PDF format. Texinfo version 4674.8 or later is required for @command{make pdf}. 468 469Necessary to build GCC documentation during development because the 470generated output files are not included in the SVN repository. They are 471included in releases. 472 473@item @TeX{} (any working version) 474 475Necessary for running @command{texi2dvi} and @command{texi2pdf}, which 476are used when running @command{make dvi} or @command{make pdf} to create 477DVI or PDF files, respectively. 478 479@item Sphinx version 1.0 (or later) 480 481Necessary to regenerate @file{jit/docs/_build/texinfo} from the @file{.rst} 482files in the directories below @file{jit/docs}. 483 484@item SVN (any version) 485@itemx SSH (any version) 486 487Necessary to access the SVN repository. Public releases and weekly 488snapshots of the development sources are also available via FTP@. 489 490@item GNU diffutils version 2.7 (or later) 491 492Useful when submitting patches for the GCC source code. 493 494@item patch version 2.5.4 (or later) 495 496Necessary when applying patches, created with @command{diff}, to one's 497own sources. 498 499@item ecj1 500@itemx gjavah 501 502If you wish to modify @file{.java} files in libjava, you will need to 503configure with @option{--enable-java-maintainer-mode}, and you will need 504to have executables named @command{ecj1} and @command{gjavah} in your path. 505The @command{ecj1} executable should run the Eclipse Java compiler via 506the GCC-specific entry point. You can download a suitable jar from 507@uref{ftp://sourceware.org/pub/java/}, or by running the script 508@command{contrib/download_ecj}. 509 510@item antlr.jar version 2.7.1 (or later) 511@itemx antlr binary 512 513If you wish to build the @command{gjdoc} binary in libjava, you will 514need to have an @file{antlr.jar} library available. The library is 515searched for in system locations but can be specified with 516@option{--with-antlr-jar=} instead. When configuring with 517@option{--enable-java-maintainer-mode}, you will need to have one of 518the executables named @command{cantlr}, @command{runantlr} or 519@command{antlr} in your path. 520 521@end table 522 523@html 524<hr /> 525<p> 526@end html 527@ifhtml 528@uref{./index.html,,Return to the GCC Installation page} 529@end ifhtml 530@end ifset 531 532@c ***Downloading the source************************************************** 533@ifnothtml 534@comment node-name, next, previous, up 535@node Downloading the source, Configuration, Prerequisites, Installing GCC 536@end ifnothtml 537@ifset downloadhtml 538@ifnothtml 539@chapter Downloading GCC 540@end ifnothtml 541@cindex Downloading GCC 542@cindex Downloading the Source 543 544GCC is distributed via @uref{http://gcc.gnu.org/svn.html,,SVN} and FTP 545tarballs compressed with @command{gzip} or 546@command{bzip2}. 547 548Please refer to the @uref{http://gcc.gnu.org/releases.html,,releases web page} 549for information on how to obtain GCC@. 550 551The source distribution includes the C, C++, Objective-C, Fortran, Java, 552and Ada (in the case of GCC 3.1 and later) compilers, as well as 553runtime libraries for C++, Objective-C, Fortran, and Java. 554For previous versions these were downloadable as separate components such 555as the core GCC distribution, which included the C language front end and 556shared components, and language-specific distributions including the 557language front end and the language runtime (where appropriate). 558 559If you also intend to build binutils (either to upgrade an existing 560installation or for use in place of the corresponding tools of your 561OS), unpack the binutils distribution either in the same directory or 562a separate one. In the latter case, add symbolic links to any 563components of the binutils you intend to build alongside the compiler 564(@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld}, 565@file{opcodes}, @dots{}) to the directory containing the GCC sources. 566 567Likewise the GMP, MPFR and MPC libraries can be automatically built 568together with GCC. You may simply run the 569@command{contrib/download_prerequisites} script in the GCC source directory 570to set up everything. 571Otherwise unpack the GMP, MPFR and/or MPC source 572distributions in the directory containing the GCC sources and rename 573their directories to @file{gmp}, @file{mpfr} and @file{mpc}, 574respectively (or use symbolic links with the same name). 575 576@html 577<hr /> 578<p> 579@end html 580@ifhtml 581@uref{./index.html,,Return to the GCC Installation page} 582@end ifhtml 583@end ifset 584 585@c ***Configuration*********************************************************** 586@ifnothtml 587@comment node-name, next, previous, up 588@node Configuration, Building, Downloading the source, Installing GCC 589@end ifnothtml 590@ifset configurehtml 591@ifnothtml 592@chapter Installing GCC: Configuration 593@end ifnothtml 594@cindex Configuration 595@cindex Installing GCC: Configuration 596 597Like most GNU software, GCC must be configured before it can be built. 598This document describes the recommended configuration procedure 599for both native and cross targets. 600 601We use @var{srcdir} to refer to the toplevel source directory for 602GCC; we use @var{objdir} to refer to the toplevel build/object directory. 603 604If you obtained the sources via SVN, @var{srcdir} must refer to the top 605@file{gcc} directory, the one where the @file{MAINTAINERS} file can be 606found, and not its @file{gcc} subdirectory, otherwise the build will fail. 607 608If either @var{srcdir} or @var{objdir} is located on an automounted NFS 609file system, the shell's built-in @command{pwd} command will return 610temporary pathnames. Using these can lead to various sorts of build 611problems. To avoid this issue, set the @env{PWDCMD} environment 612variable to an automounter-aware @command{pwd} command, e.g., 613@command{pawd} or @samp{amq -w}, during the configuration and build 614phases. 615 616First, we @strong{highly} recommend that GCC be built into a 617separate directory from the sources which does @strong{not} reside 618within the source tree. This is how we generally build GCC; building 619where @var{srcdir} == @var{objdir} should still work, but doesn't 620get extensive testing; building where @var{objdir} is a subdirectory 621of @var{srcdir} is unsupported. 622 623If you have previously built GCC in the same directory for a 624different target machine, do @samp{make distclean} to delete all files 625that might be invalid. One of the files this deletes is @file{Makefile}; 626if @samp{make distclean} complains that @file{Makefile} does not exist 627or issues a message like ``don't know how to make distclean'' it probably 628means that the directory is already suitably clean. However, with the 629recommended method of building in a separate @var{objdir}, you should 630simply use a different @var{objdir} for each target. 631 632Second, when configuring a native system, either @command{cc} or 633@command{gcc} must be in your path or you must set @env{CC} in 634your environment before running configure. Otherwise the configuration 635scripts may fail. 636 637@ignore 638Note that the bootstrap compiler and the resulting GCC must be link 639compatible, else the bootstrap will fail with linker errors about 640incompatible object file formats. Several multilibed targets are 641affected by this requirement, see 642@ifnothtml 643@ref{Specific, host/target specific installation notes}. 644@end ifnothtml 645@ifhtml 646@uref{specific.html,,host/target specific installation notes}. 647@end ifhtml 648@end ignore 649 650To configure GCC: 651 652@smallexample 653% mkdir @var{objdir} 654% cd @var{objdir} 655% @var{srcdir}/configure [@var{options}] [@var{target}] 656@end smallexample 657 658@heading Distributor options 659 660If you will be distributing binary versions of GCC, with modifications 661to the source code, you should use the options described in this 662section to make clear that your version contains modifications. 663 664@table @code 665@item --with-pkgversion=@var{version} 666Specify a string that identifies your package. You may wish 667to include a build number or build date. This version string will be 668included in the output of @command{gcc --version}. This suffix does 669not replace the default version string, only the @samp{GCC} part. 670 671The default value is @samp{GCC}. 672 673@item --with-bugurl=@var{url} 674Specify the URL that users should visit if they wish to report a bug. 675You are of course welcome to forward bugs reported to you to the FSF, 676if you determine that they are not bugs in your modifications. 677 678The default value refers to the FSF's GCC bug tracker. 679 680@end table 681 682@heading Target specification 683@itemize @bullet 684@item 685GCC has code to correctly determine the correct value for @var{target} 686for nearly all native systems. Therefore, we highly recommend you do 687not provide a configure target when configuring a native compiler. 688 689@item 690@var{target} must be specified as @option{--target=@var{target}} 691when configuring a cross compiler; examples of valid targets would be 692m68k-elf, sh-elf, etc. 693 694@item 695Specifying just @var{target} instead of @option{--target=@var{target}} 696implies that the host defaults to @var{target}. 697@end itemize 698 699 700@heading Options specification 701 702Use @var{options} to override several configure time options for 703GCC@. A list of supported @var{options} follows; @samp{configure 704--help} may list other options, but those not listed below may not 705work and should not normally be used. 706 707Note that each @option{--enable} option has a corresponding 708@option{--disable} option and that each @option{--with} option has a 709corresponding @option{--without} option. 710 711@table @code 712@item --prefix=@var{dirname} 713Specify the toplevel installation 714directory. This is the recommended way to install the tools into a directory 715other than the default. The toplevel installation directory defaults to 716@file{/usr/local}. 717 718We @strong{highly} recommend against @var{dirname} being the same or a 719subdirectory of @var{objdir} or vice versa. If specifying a directory 720beneath a user's home directory tree, some shells will not expand 721@var{dirname} correctly if it contains the @samp{~} metacharacter; use 722@env{$HOME} instead. 723 724The following standard @command{autoconf} options are supported. Normally you 725should not need to use these options. 726@table @code 727@item --exec-prefix=@var{dirname} 728Specify the toplevel installation directory for architecture-dependent 729files. The default is @file{@var{prefix}}. 730 731@item --bindir=@var{dirname} 732Specify the installation directory for the executables called by users 733(such as @command{gcc} and @command{g++}). The default is 734@file{@var{exec-prefix}/bin}. 735 736@item --libdir=@var{dirname} 737Specify the installation directory for object code libraries and 738internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}. 739 740@item --libexecdir=@var{dirname} 741Specify the installation directory for internal executables of GCC@. 742The default is @file{@var{exec-prefix}/libexec}. 743 744@item --with-slibdir=@var{dirname} 745Specify the installation directory for the shared libgcc library. The 746default is @file{@var{libdir}}. 747 748@item --datarootdir=@var{dirname} 749Specify the root of the directory tree for read-only architecture-independent 750data files referenced by GCC@. The default is @file{@var{prefix}/share}. 751 752@item --infodir=@var{dirname} 753Specify the installation directory for documentation in info format. 754The default is @file{@var{datarootdir}/info}. 755 756@item --datadir=@var{dirname} 757Specify the installation directory for some architecture-independent 758data files referenced by GCC@. The default is @file{@var{datarootdir}}. 759 760@item --docdir=@var{dirname} 761Specify the installation directory for documentation files (other 762than Info) for GCC@. The default is @file{@var{datarootdir}/doc}. 763 764@item --htmldir=@var{dirname} 765Specify the installation directory for HTML documentation files. 766The default is @file{@var{docdir}}. 767 768@item --pdfdir=@var{dirname} 769Specify the installation directory for PDF documentation files. 770The default is @file{@var{docdir}}. 771 772@item --mandir=@var{dirname} 773Specify the installation directory for manual pages. The default is 774@file{@var{datarootdir}/man}. (Note that the manual pages are only extracts 775from the full GCC manuals, which are provided in Texinfo format. The manpages 776are derived by an automatic conversion process from parts of the full 777manual.) 778 779@item --with-gxx-include-dir=@var{dirname} 780Specify 781the installation directory for G++ header files. The default depends 782on other configuration options, and differs between cross and native 783configurations. 784 785@item --with-specs=@var{specs} 786Specify additional command line driver SPECS. 787This can be useful if you need to turn on a non-standard feature by 788default without modifying the compiler's source code, for instance 789@option{--with-specs=%@{!fcommon:%@{!fno-common:-fno-common@}@}}. 790@ifnothtml 791@xref{Spec Files,, Specifying subprocesses and the switches to pass to them, 792gcc, Using the GNU Compiler Collection (GCC)}, 793@end ifnothtml 794@ifhtml 795See ``Spec Files'' in the main manual 796@end ifhtml 797 798@end table 799 800@item --program-prefix=@var{prefix} 801GCC supports some transformations of the names of its programs when 802installing them. This option prepends @var{prefix} to the names of 803programs to install in @var{bindir} (see above). For example, specifying 804@option{--program-prefix=foo-} would result in @samp{gcc} 805being installed as @file{/usr/local/bin/foo-gcc}. 806 807@item --program-suffix=@var{suffix} 808Appends @var{suffix} to the names of programs to install in @var{bindir} 809(see above). For example, specifying @option{--program-suffix=-3.1} 810would result in @samp{gcc} being installed as 811@file{/usr/local/bin/gcc-3.1}. 812 813@item --program-transform-name=@var{pattern} 814Applies the @samp{sed} script @var{pattern} to be applied to the names 815of programs to install in @var{bindir} (see above). @var{pattern} has to 816consist of one or more basic @samp{sed} editing commands, separated by 817semicolons. For example, if you want the @samp{gcc} program name to be 818transformed to the installed program @file{/usr/local/bin/myowngcc} and 819the @samp{g++} program name to be transformed to 820@file{/usr/local/bin/gspecial++} without changing other program names, 821you could use the pattern 822@option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'} 823to achieve this effect. 824 825All three options can be combined and used together, resulting in more 826complex conversion patterns. As a basic rule, @var{prefix} (and 827@var{suffix}) are prepended (appended) before further transformations 828can happen with a special transformation script @var{pattern}. 829 830As currently implemented, this option only takes effect for native 831builds; cross compiler binaries' names are not transformed even when a 832transformation is explicitly asked for by one of these options. 833 834For native builds, some of the installed programs are also installed 835with the target alias in front of their name, as in 836@samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen 837before the target alias is prepended to the name---so, specifying 838@option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the 839resulting binary would be installed as 840@file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}. 841 842As a last shortcoming, none of the installed Ada programs are 843transformed yet, which will be fixed in some time. 844 845@item --with-local-prefix=@var{dirname} 846Specify the 847installation directory for local include files. The default is 848@file{/usr/local}. Specify this option if you want the compiler to 849search directory @file{@var{dirname}/include} for locally installed 850header files @emph{instead} of @file{/usr/local/include}. 851 852You should specify @option{--with-local-prefix} @strong{only} if your 853site has a different convention (not @file{/usr/local}) for where to put 854site-specific files. 855 856The default value for @option{--with-local-prefix} is @file{/usr/local} 857regardless of the value of @option{--prefix}. Specifying 858@option{--prefix} has no effect on which directory GCC searches for 859local header files. This may seem counterintuitive, but actually it is 860logical. 861 862The purpose of @option{--prefix} is to specify where to @emph{install 863GCC}. The local header files in @file{/usr/local/include}---if you put 864any in that directory---are not part of GCC@. They are part of other 865programs---perhaps many others. (GCC installs its own header files in 866another directory which is based on the @option{--prefix} value.) 867 868Both the local-prefix include directory and the GCC-prefix include 869directory are part of GCC's ``system include'' directories. Although these 870two directories are not fixed, they need to be searched in the proper 871order for the correct processing of the include_next directive. The 872local-prefix include directory is searched before the GCC-prefix 873include directory. Another characteristic of system include directories 874is that pedantic warnings are turned off for headers in these directories. 875 876Some autoconf macros add @option{-I @var{directory}} options to the 877compiler command line, to ensure that directories containing installed 878packages' headers are searched. When @var{directory} is one of GCC's 879system include directories, GCC will ignore the option so that system 880directories continue to be processed in the correct order. This 881may result in a search order different from what was specified but the 882directory will still be searched. 883 884GCC automatically searches for ordinary libraries using 885@env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is 886used for both GCC and packages, GCC will automatically search for 887both headers and libraries. This provides a configuration that is 888easy to use. GCC behaves in a manner similar to that when it is 889installed as a system compiler in @file{/usr}. 890 891Sites that need to install multiple versions of GCC may not want to 892use the above simple configuration. It is possible to use the 893@option{--program-prefix}, @option{--program-suffix} and 894@option{--program-transform-name} options to install multiple versions 895into a single directory, but it may be simpler to use different prefixes 896and the @option{--with-local-prefix} option to specify the location of the 897site-specific files for each version. It will then be necessary for 898users to specify explicitly the location of local site libraries 899(e.g., with @env{LIBRARY_PATH}). 900 901The same value can be used for both @option{--with-local-prefix} and 902@option{--prefix} provided it is not @file{/usr}. This can be used 903to avoid the default search of @file{/usr/local/include}. 904 905@strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}! 906The directory you use for @option{--with-local-prefix} @strong{must not} 907contain any of the system's standard header files. If it did contain 908them, certain programs would be miscompiled (including GNU Emacs, on 909certain targets), because this would override and nullify the header 910file corrections made by the @command{fixincludes} script. 911 912Indications are that people who use this option use it based on mistaken 913ideas of what it is for. People use it as if it specified where to 914install part of GCC@. Perhaps they make this assumption because 915installing GCC creates the directory. 916 917@item --with-native-system-header-dir=@var{dirname} 918Specifies that @var{dirname} is the directory that contains native system 919header files, rather than @file{/usr/include}. This option is most useful 920if you are creating a compiler that should be isolated from the system 921as much as possible. It is most commonly used with the 922@option{--with-sysroot} option and will cause GCC to search 923@var{dirname} inside the system root specified by that option. 924 925@item --enable-shared[=@var{package}[,@dots{}]] 926Build shared versions of libraries, if shared libraries are supported on 927the target platform. Unlike GCC 2.95.x and earlier, shared libraries 928are enabled by default on all platforms that support shared libraries. 929 930If a list of packages is given as an argument, build shared libraries 931only for the listed packages. For other packages, only static libraries 932will be built. Package names currently recognized in the GCC tree are 933@samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not 934@samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc}, 935@samp{ada}, @samp{libada}, @samp{libjava}, @samp{libgo}, and @samp{libobjc}. 936Note @samp{libiberty} does not support shared libraries at all. 937 938Use @option{--disable-shared} to build only static libraries. Note that 939@option{--disable-shared} does not accept a list of package names as 940argument, only @option{--enable-shared} does. 941 942Contrast with @option{--enable-host-shared}, which affects @emph{host} 943code. 944 945@item --enable-host-shared 946Specify that the @emph{host} code should be built into position-independent 947machine code (with -fPIC), allowing it to be used within shared libraries, 948but yielding a slightly slower compiler. 949 950This option is required when building the libgccjit.so library. 951 952Contrast with @option{--enable-shared}, which affects @emph{target} 953libraries. 954 955@item @anchor{with-gnu-as}--with-gnu-as 956Specify that the compiler should assume that the 957assembler it finds is the GNU assembler. However, this does not modify 958the rules to find an assembler and will result in confusion if the 959assembler found is not actually the GNU assembler. (Confusion may also 960result if the compiler finds the GNU assembler but has not been 961configured with @option{--with-gnu-as}.) If you have more than one 962assembler installed on your system, you may want to use this option in 963connection with @option{--with-as=@var{pathname}} or 964@option{--with-build-time-tools=@var{pathname}}. 965 966The following systems are the only ones where it makes a difference 967whether you use the GNU assembler. On any other system, 968@option{--with-gnu-as} has no effect. 969 970@itemize @bullet 971@item @samp{hppa1.0-@var{any}-@var{any}} 972@item @samp{hppa1.1-@var{any}-@var{any}} 973@item @samp{sparc-sun-solaris2.@var{any}} 974@item @samp{sparc64-@var{any}-solaris2.@var{any}} 975@end itemize 976 977@item @anchor{with-as}--with-as=@var{pathname} 978Specify that the compiler should use the assembler pointed to by 979@var{pathname}, rather than the one found by the standard rules to find 980an assembler, which are: 981@itemize @bullet 982@item 983Unless GCC is being built with a cross compiler, check the 984@file{@var{libexec}/gcc/@var{target}/@var{version}} directory. 985@var{libexec} defaults to @file{@var{exec-prefix}/libexec}; 986@var{exec-prefix} defaults to @var{prefix}, which 987defaults to @file{/usr/local} unless overridden by the 988@option{--prefix=@var{pathname}} switch described above. @var{target} 989is the target system triple, such as @samp{sparc-sun-solaris2.7}, and 990@var{version} denotes the GCC version, such as 3.0. 991 992@item 993If the target system is the same that you are building on, check 994operating system specific directories (e.g.@: @file{/usr/ccs/bin} on 995Sun Solaris 2). 996 997@item 998Check in the @env{PATH} for a tool whose name is prefixed by the 999target system triple. 1000 1001@item 1002Check in the @env{PATH} for a tool whose name is not prefixed by the 1003target system triple, if the host and target system triple are 1004the same (in other words, we use a host tool if it can be used for 1005the target as well). 1006@end itemize 1007 1008You may want to use @option{--with-as} if no assembler 1009is installed in the directories listed above, or if you have multiple 1010assemblers installed and want to choose one that is not found by the 1011above rules. 1012 1013@item @anchor{with-gnu-ld}--with-gnu-ld 1014Same as @uref{#with-gnu-as,,@option{--with-gnu-as}} 1015but for the linker. 1016 1017@item --with-ld=@var{pathname} 1018Same as @uref{#with-as,,@option{--with-as}} 1019but for the linker. 1020 1021@item --with-stabs 1022Specify that stabs debugging 1023information should be used instead of whatever format the host normally 1024uses. Normally GCC uses the same debug format as the host system. 1025 1026On MIPS based systems and on Alphas, you must specify whether you want 1027GCC to create the normal ECOFF debugging format, or to use BSD-style 1028stabs passed through the ECOFF symbol table. The normal ECOFF debug 1029format cannot fully handle languages other than C@. BSD stabs format can 1030handle other languages, but it only works with the GNU debugger GDB@. 1031 1032Normally, GCC uses the ECOFF debugging format by default; if you 1033prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@. 1034 1035No matter which default you choose when you configure GCC, the user 1036can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly 1037the debug format for a particular compilation. 1038 1039@option{--with-stabs} is meaningful on the ISC system on the 386, also, if 1040@option{--with-gas} is used. It selects use of stabs debugging 1041information embedded in COFF output. This kind of debugging information 1042supports C++ well; ordinary COFF debugging information does not. 1043 1044@option{--with-stabs} is also meaningful on 386 systems running SVR4. It 1045selects use of stabs debugging information embedded in ELF output. The 1046C++ compiler currently (2.6.0) does not support the DWARF debugging 1047information normally used on 386 SVR4 platforms; stabs provide a 1048workable alternative. This requires gas and gdb, as the normal SVR4 1049tools can not generate or interpret stabs. 1050 1051@item --with-tls=@var{dialect} 1052Specify the default TLS dialect, for systems were there is a choice. 1053For ARM targets, possible values for @var{dialect} are @code{gnu} or 1054@code{gnu2}, which select between the original GNU dialect and the GNU TLS 1055descriptor-based dialect. 1056 1057@item --enable-multiarch 1058Specify whether to enable or disable multiarch support. The default is 1059to check for glibc start files in a multiarch location, and enable it 1060if the files are found. The auto detection is enabled for native builds, 1061and for cross builds configured with @option{--with-sysroot}, and without 1062@option{--with-native-system-header-dir}. 1063More documentation about multiarch can be found at 1064@uref{https://wiki.debian.org/Multiarch}. 1065 1066@item --enable-sjlj-exceptions 1067Force use of the @code{setjmp}/@code{longjmp}-based scheme for exceptions. 1068@samp{configure} ordinarily picks the correct value based on the platform. 1069Only use this option if you are sure you need a different setting. 1070 1071@item --enable-vtable-verify 1072Specify whether to enable or disable the vtable verification feature. 1073Enabling this feature causes libstdc++ to be built with its virtual calls 1074in verifiable mode. This means that, when linked with libvtv, every 1075virtual call in libstdc++ will verify the vtable pointer through which the 1076call will be made before actually making the call. If not linked with libvtv, 1077the verifier will call stub functions (in libstdc++ itself) and do nothing. 1078If vtable verification is disabled, then libstdc++ is not built with its 1079virtual calls in verifiable mode at all. However the libvtv library will 1080still be built (see @option{--disable-libvtv} to turn off building libvtv). 1081@option{--disable-vtable-verify} is the default. 1082 1083@item --disable-multilib 1084Specify that multiple target 1085libraries to support different target variants, calling 1086conventions, etc.@: should not be built. The default is to build a 1087predefined set of them. 1088 1089Some targets provide finer-grained control over which multilibs are built 1090(e.g., @option{--disable-softfloat}): 1091@table @code 1092@item arm-*-* 1093fpu, 26bit, underscore, interwork, biendian, nofmult. 1094 1095@item m68*-*-* 1096softfloat, m68881, m68000, m68020. 1097 1098@item mips*-*-* 1099single-float, biendian, softfloat. 1100 1101@item powerpc*-*-*, rs6000*-*-* 1102aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian, 1103sysv, aix. 1104 1105@end table 1106 1107@item --with-multilib-list=@var{list} 1108@itemx --without-multilib-list 1109Specify what multilibs to build. 1110Currently only implemented for arm*-*-*, sh*-*-* and x86-64-*-linux*. 1111 1112@table @code 1113@item arm*-*-* 1114@var{list} is either @code{default} or @code{aprofile}. Specifying 1115@code{default} is equivalent to omitting this option while specifying 1116@code{aprofile} builds multilibs for each combination of ISA (@code{-marm} or 1117@code{-mthumb}), architecture (@code{-march=armv7-a}, @code{-march=armv7ve}, 1118or @code{-march=armv8-a}), FPU available (none, @code{-mfpu=vfpv3-d16}, 1119@code{-mfpu=neon}, @code{-mfpu=vfpv4-d16}, @code{-mfpu=neon-vfpv4} or 1120@code{-mfpu=neon-fp-armv8} depending on architecture) and floating-point ABI 1121(@code{-mfloat-abi=softfp} or @code{-mfloat-abi=hard}). 1122 1123@item sh*-*-* 1124@var{list} is a comma separated list of CPU names. These must be of the 1125form @code{sh*} or @code{m*} (in which case they match the compiler option 1126for that processor). The list should not contain any endian options - 1127these are handled by @option{--with-endian}. 1128 1129If @var{list} is empty, then there will be no multilibs for extra 1130processors. The multilib for the secondary endian remains enabled. 1131 1132As a special case, if an entry in the list starts with a @code{!} 1133(exclamation point), then it is added to the list of excluded multilibs. 1134Entries of this sort should be compatible with @samp{MULTILIB_EXCLUDES} 1135(once the leading @code{!} has been stripped). 1136 1137If @option{--with-multilib-list} is not given, then a default set of 1138multilibs is selected based on the value of @option{--target}. This is 1139usually the complete set of libraries, but some targets imply a more 1140specialized subset. 1141 1142Example 1: to configure a compiler for SH4A only, but supporting both 1143endians, with little endian being the default: 1144@smallexample 1145--with-cpu=sh4a --with-endian=little,big --with-multilib-list= 1146@end smallexample 1147 1148Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with 1149only little endian SH4AL: 1150@smallexample 1151--with-cpu=sh4a --with-endian=little,big \ 1152--with-multilib-list=sh4al,!mb/m4al 1153@end smallexample 1154 1155@item x86-64-*-linux* 1156@var{list} is a comma separated list of @code{m32}, @code{m64} and 1157@code{mx32} to enable 32-bit, 64-bit and x32 run-time libraries, 1158respectively. If @var{list} is empty, then there will be no multilibs 1159and only the default run-time library will be enabled. 1160 1161If @option{--with-multilib-list} is not given, then only 32-bit and 116264-bit run-time libraries will be enabled. 1163@end table 1164 1165@item --with-endian=@var{endians} 1166Specify what endians to use. 1167Currently only implemented for sh*-*-*. 1168 1169@var{endians} may be one of the following: 1170@table @code 1171@item big 1172Use big endian exclusively. 1173@item little 1174Use little endian exclusively. 1175@item big,little 1176Use big endian by default. Provide a multilib for little endian. 1177@item little,big 1178Use little endian by default. Provide a multilib for big endian. 1179@end table 1180 1181@item --enable-threads 1182Specify that the target 1183supports threads. This affects the Objective-C compiler and runtime 1184library, and exception handling for other languages like C++ and Java. 1185On some systems, this is the default. 1186 1187In general, the best (and, in many cases, the only known) threading 1188model available will be configured for use. Beware that on some 1189systems, GCC has not been taught what threading models are generally 1190available for the system. In this case, @option{--enable-threads} is an 1191alias for @option{--enable-threads=single}. 1192 1193@item --disable-threads 1194Specify that threading support should be disabled for the system. 1195This is an alias for @option{--enable-threads=single}. 1196 1197@item --enable-threads=@var{lib} 1198Specify that 1199@var{lib} is the thread support library. This affects the Objective-C 1200compiler and runtime library, and exception handling for other languages 1201like C++ and Java. The possibilities for @var{lib} are: 1202 1203@table @code 1204@item aix 1205AIX thread support. 1206@item dce 1207DCE thread support. 1208@item lynx 1209LynxOS thread support. 1210@item mipssde 1211MIPS SDE thread support. 1212@item no 1213This is an alias for @samp{single}. 1214@item posix 1215Generic POSIX/Unix98 thread support. 1216@item rtems 1217RTEMS thread support. 1218@item single 1219Disable thread support, should work for all platforms. 1220@item tpf 1221TPF thread support. 1222@item vxworks 1223VxWorks thread support. 1224@item win32 1225Microsoft Win32 API thread support. 1226@end table 1227 1228@item --enable-tls 1229Specify that the target supports TLS (Thread Local Storage). Usually 1230configure can correctly determine if TLS is supported. In cases where 1231it guesses incorrectly, TLS can be explicitly enabled or disabled with 1232@option{--enable-tls} or @option{--disable-tls}. This can happen if 1233the assembler supports TLS but the C library does not, or if the 1234assumptions made by the configure test are incorrect. 1235 1236@item --disable-tls 1237Specify that the target does not support TLS. 1238This is an alias for @option{--enable-tls=no}. 1239 1240@item --with-cpu=@var{cpu} 1241@itemx --with-cpu-32=@var{cpu} 1242@itemx --with-cpu-64=@var{cpu} 1243Specify which cpu variant the compiler should generate code for by default. 1244@var{cpu} will be used as the default value of the @option{-mcpu=} switch. 1245This option is only supported on some targets, including ARC, ARM, i386, M68k, 1246PowerPC, and SPARC@. It is mandatory for ARC@. The @option{--with-cpu-32} and 1247@option{--with-cpu-64} options specify separate default CPUs for 124832-bit and 64-bit modes; these options are only supported for i386, 1249x86-64, PowerPC, and SPARC@. 1250 1251@item --with-schedule=@var{cpu} 1252@itemx --with-arch=@var{cpu} 1253@itemx --with-arch-32=@var{cpu} 1254@itemx --with-arch-64=@var{cpu} 1255@itemx --with-tune=@var{cpu} 1256@itemx --with-tune-32=@var{cpu} 1257@itemx --with-tune-64=@var{cpu} 1258@itemx --with-abi=@var{abi} 1259@itemx --with-fpu=@var{type} 1260@itemx --with-float=@var{type} 1261These configure options provide default values for the @option{-mschedule=}, 1262@option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=} 1263options and for @option{-mhard-float} or @option{-msoft-float}. As with 1264@option{--with-cpu}, which switches will be accepted and acceptable values 1265of the arguments depend on the target. 1266 1267@item --with-mode=@var{mode} 1268Specify if the compiler should default to @option{-marm} or @option{-mthumb}. 1269This option is only supported on ARM targets. 1270 1271@item --with-stack-offset=@var{num} 1272This option sets the default for the -mstack-offset=@var{num} option, 1273and will thus generally also control the setting of this option for 1274libraries. This option is only supported on Epiphany targets. 1275 1276@item --with-fpmath=@var{isa} 1277This options sets @option{-mfpmath=sse} by default and specifies the default 1278ISA for floating-point arithmetics. You can select either @samp{sse} which 1279enables @option{-msse2} or @samp{avx} which enables @option{-mavx} by default. 1280This option is only supported on i386 and x86-64 targets. 1281 1282@item --with-fp-32=@var{mode} 1283On MIPS targets, set the default value for the @option{-mfp} option when using 1284the o32 ABI. The possibilities for @var{mode} are: 1285@table @code 1286@item 32 1287Use the o32 FP32 ABI extension, as with the @option{-mfp32} command-line 1288option. 1289@item xx 1290Use the o32 FPXX ABI extension, as with the @option{-mfpxx} command-line 1291option. 1292@item 64 1293Use the o32 FP64 ABI extension, as with the @option{-mfp64} command-line 1294option. 1295@end table 1296In the absence of this configuration option the default is to use the o32 1297FP32 ABI extension. 1298 1299@item --with-odd-spreg-32 1300On MIPS targets, set the @option{-modd-spreg} option by default when using 1301the o32 ABI. 1302 1303@item --without-odd-spreg-32 1304On MIPS targets, set the @option{-mno-odd-spreg} option by default when using 1305the o32 ABI. This is normally used in conjunction with 1306@option{--with-fp-32=64} in order to target the o32 FP64A ABI extension. 1307 1308@item --with-nan=@var{encoding} 1309On MIPS targets, set the default encoding convention to use for the 1310special not-a-number (NaN) IEEE 754 floating-point data. The 1311possibilities for @var{encoding} are: 1312@table @code 1313@item legacy 1314Use the legacy encoding, as with the @option{-mnan=legacy} command-line 1315option. 1316@item 2008 1317Use the 754-2008 encoding, as with the @option{-mnan=2008} command-line 1318option. 1319@end table 1320To use this configuration option you must have an assembler version 1321installed that supports the @option{-mnan=} command-line option too. 1322In the absence of this configuration option the default convention is 1323the legacy encoding, as when neither of the @option{-mnan=2008} and 1324@option{-mnan=legacy} command-line options has been used. 1325 1326@item --with-divide=@var{type} 1327Specify how the compiler should generate code for checking for 1328division by zero. This option is only supported on the MIPS target. 1329The possibilities for @var{type} are: 1330@table @code 1331@item traps 1332Division by zero checks use conditional traps (this is the default on 1333systems that support conditional traps). 1334@item breaks 1335Division by zero checks use the break instruction. 1336@end table 1337 1338@c If you make --with-llsc the default for additional targets, 1339@c update the --with-llsc description in the MIPS section below. 1340 1341@item --with-llsc 1342On MIPS targets, make @option{-mllsc} the default when no 1343@option{-mno-llsc} option is passed. This is the default for 1344Linux-based targets, as the kernel will emulate them if the ISA does 1345not provide them. 1346 1347@item --without-llsc 1348On MIPS targets, make @option{-mno-llsc} the default when no 1349@option{-mllsc} option is passed. 1350 1351@item --with-synci 1352On MIPS targets, make @option{-msynci} the default when no 1353@option{-mno-synci} option is passed. 1354 1355@item --without-synci 1356On MIPS targets, make @option{-mno-synci} the default when no 1357@option{-msynci} option is passed. This is the default. 1358 1359@item --with-mips-plt 1360On MIPS targets, make use of copy relocations and PLTs. 1361These features are extensions to the traditional 1362SVR4-based MIPS ABIs and require support from GNU binutils 1363and the runtime C library. 1364 1365@item --enable-__cxa_atexit 1366Define if you want to use __cxa_atexit, rather than atexit, to 1367register C++ destructors for local statics and global objects. 1368This is essential for fully standards-compliant handling of 1369destructors, but requires __cxa_atexit in libc. This option is currently 1370only available on systems with GNU libc. When enabled, this will cause 1371@option{-fuse-cxa-atexit} to be passed by default. 1372 1373@item --enable-gnu-indirect-function 1374Define if you want to enable the @code{ifunc} attribute. This option is 1375currently only available on systems with GNU libc on certain targets. 1376 1377@item --enable-target-optspace 1378Specify that target 1379libraries should be optimized for code space instead of code speed. 1380This is the default for the m32r platform. 1381 1382@item --with-cpp-install-dir=@var{dirname} 1383Specify that the user visible @command{cpp} program should be installed 1384in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}. 1385 1386@item --enable-comdat 1387Enable COMDAT group support. This is primarily used to override the 1388automatically detected value. 1389 1390@item --enable-initfini-array 1391Force the use of sections @code{.init_array} and @code{.fini_array} 1392(instead of @code{.init} and @code{.fini}) for constructors and 1393destructors. Option @option{--disable-initfini-array} has the 1394opposite effect. If neither option is specified, the configure script 1395will try to guess whether the @code{.init_array} and 1396@code{.fini_array} sections are supported and, if they are, use them. 1397 1398@item --enable-link-mutex 1399When building GCC, use a mutex to avoid linking the compilers for 1400multiple languages at the same time, to avoid thrashing on build 1401systems with limited free memory. The default is not to use such a mutex. 1402 1403@item --enable-maintainer-mode 1404The build rules that regenerate the Autoconf and Automake output files as 1405well as the GCC master message catalog @file{gcc.pot} are normally 1406disabled. This is because it can only be rebuilt if the complete source 1407tree is present. If you have changed the sources and want to rebuild the 1408catalog, configuring with @option{--enable-maintainer-mode} will enable 1409this. Note that you need a recent version of the @code{gettext} tools 1410to do so. 1411 1412@item --disable-bootstrap 1413For a native build, the default configuration is to perform 1414a 3-stage bootstrap of the compiler when @samp{make} is invoked, 1415testing that GCC can compile itself correctly. If you want to disable 1416this process, you can configure with @option{--disable-bootstrap}. 1417 1418@item --enable-bootstrap 1419In special cases, you may want to perform a 3-stage build 1420even if the target and host triplets are different. 1421This is possible when the host can run code compiled for 1422the target (e.g.@: host is i686-linux, target is i486-linux). 1423Starting from GCC 4.2, to do this you have to configure explicitly 1424with @option{--enable-bootstrap}. 1425 1426@item --enable-generated-files-in-srcdir 1427Neither the .c and .h files that are generated from Bison and flex nor the 1428info manuals and man pages that are built from the .texi files are present 1429in the SVN development tree. When building GCC from that development tree, 1430or from one of our snapshots, those generated files are placed in your 1431build directory, which allows for the source to be in a readonly 1432directory. 1433 1434If you configure with @option{--enable-generated-files-in-srcdir} then those 1435generated files will go into the source directory. This is mainly intended 1436for generating release or prerelease tarballs of the GCC sources, since it 1437is not a requirement that the users of source releases to have flex, Bison, 1438or makeinfo. 1439 1440@item --enable-version-specific-runtime-libs 1441Specify 1442that runtime libraries should be installed in the compiler specific 1443subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In 1444addition, @samp{libstdc++}'s include files will be installed into 1445@file{@var{libdir}} unless you overruled it by using 1446@option{--with-gxx-include-dir=@var{dirname}}. Using this option is 1447particularly useful if you intend to use several versions of GCC in 1448parallel. This is currently supported by @samp{libgfortran}, 1449@samp{libjava}, @samp{libstdc++}, and @samp{libobjc}. 1450 1451@item @anchor{WithAixSoname}--with-aix-soname=@samp{aix}, @samp{svr4} or @samp{both} 1452Traditional AIX shared library versioning (versioned @code{Shared Object} 1453files as members of unversioned @code{Archive Library} files named 1454@samp{lib.a}) causes numerous headaches for package managers. However, 1455@code{Import Files} as members of @code{Archive Library} files allow for 1456@strong{filename-based versioning} of shared libraries as seen on Linux/SVR4, 1457where this is called the "SONAME". But as they prevent static linking, 1458@code{Import Files} may be used with @code{Runtime Linking} only, where the 1459linker does search for @samp{libNAME.so} before @samp{libNAME.a} library 1460filenames with the @samp{-lNAME} linker flag. 1461 1462@anchor{AixLdCommand}For detailed information please refer to the AIX 1463@uref{http://www-01.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22,,ld 1464Command} reference. 1465 1466As long as shared library creation is enabled, upon: 1467@table @code 1468@item --with-aix-soname=aix 1469@item --with-aix-soname=both 1470 A (traditional AIX) @code{Shared Archive Library} file is created: 1471 @itemize @bullet 1472 @item using the @samp{libNAME.a} filename scheme 1473 @item with the @code{Shared Object} file as archive member named 1474 @samp{libNAME.so.V} (except for @samp{libgcc_s}, where the @code{Shared 1475 Object} file is named @samp{shr.o} for backwards compatibility), which 1476 @itemize @minus 1477 @item is used for runtime loading from inside the @samp{libNAME.a} file 1478 @item is used for dynamic loading via 1479 @code{dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)} 1480 @item is used for shared linking 1481 @item is used for static linking, so no separate @code{Static Archive 1482 Library} file is needed 1483 @end itemize 1484 @end itemize 1485@item --with-aix-soname=both 1486@item --with-aix-soname=svr4 1487 A (second) @code{Shared Archive Library} file is created: 1488 @itemize @bullet 1489 @item using the @samp{libNAME.so.V} filename scheme 1490 @item with the @code{Shared Object} file as archive member named 1491 @samp{shr.o}, which 1492 @itemize @minus 1493 @item is created with the @code{-G linker flag} 1494 @item has the @code{F_LOADONLY} flag set 1495 @item is used for runtime loading from inside the @samp{libNAME.so.V} file 1496 @item is used for dynamic loading via @code{dlopen("libNAME.so.V(shr.o)", 1497 RTLD_MEMBER)} 1498 @end itemize 1499 @item with the @code{Import File} as archive member named @samp{shr.imp}, 1500 which 1501 @itemize @minus 1502 @item refers to @samp{libNAME.so.V(shr.o)} as the "SONAME", to be recorded 1503 in the @code{Loader Section} of subsequent binaries 1504 @item indicates whether @samp{libNAME.so.V(shr.o)} is 32 or 64 bit 1505 @item lists all the public symbols exported by @samp{lib.so.V(shr.o)}, 1506 eventually decorated with the @code{@samp{weak} Keyword} 1507 @item is necessary for shared linking against @samp{lib.so.V(shr.o)} 1508 @end itemize 1509 @end itemize 1510 A symbolic link using the @samp{libNAME.so} filename scheme is created: 1511 @itemize @bullet 1512 @item pointing to the @samp{libNAME.so.V} @code{Shared Archive Library} file 1513 @item to permit the @code{ld Command} to find @samp{lib.so.V(shr.imp)} via 1514 the @samp{-lNAME} argument (requires @code{Runtime Linking} to be enabled) 1515 @item to permit dynamic loading of @samp{lib.so.V(shr.o)} without the need 1516 to specify the version number via @code{dlopen("libNAME.so(shr.o)", 1517 RTLD_MEMBER)} 1518 @end itemize 1519@end table 1520 1521As long as static library creation is enabled, upon: 1522@table @code 1523@item --with-aix-soname=svr4 1524 A @code{Static Archive Library} is created: 1525 @itemize @bullet 1526 @item using the @samp{libNAME.a} filename scheme 1527 @item with all the @code{Static Object} files as archive members, which 1528 @itemize @minus 1529 @item are used for static linking 1530 @end itemize 1531 @end itemize 1532@end table 1533 1534While the aix-soname=@samp{svr4} option does not create @code{Shared Object} 1535files as members of unversioned @code{Archive Library} files any more, package 1536managers still are responsible to 1537@uref{./specific.html#TransferAixShobj,,transfer} @code{Shared Object} files 1538found as member of a previously installed unversioned @code{Archive Library} 1539file into the newly installed @code{Archive Library} file with the same 1540filename. 1541 1542@emph{WARNING:} Creating @code{Shared Object} files with @code{Runtime Linking} 1543enabled may bloat the TOC, eventually leading to @code{TOC overflow} errors, 1544requiring the use of either the @option{-Wl,-bbigtoc} linker flag (seen to 1545break with the @code{GDB} debugger) or some of the TOC-related compiler flags, 1546@ifnothtml 1547@xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc, 1548Using the GNU Compiler Collection (GCC)}. 1549@end ifnothtml 1550@ifhtml 1551see ``RS/6000 and PowerPC Options'' in the main manual. 1552@end ifhtml 1553 1554@option{--with-aix-soname} is currently supported by @samp{libgcc_s} only, so 1555this option is still experimental and not for normal use yet. 1556 1557Default is the traditional behavior @option{--with-aix-soname=@samp{aix}}. 1558 1559@item --enable-languages=@var{lang1},@var{lang2},@dots{} 1560Specify that only a particular subset of compilers and 1561their runtime libraries should be built. For a list of valid values for 1562@var{langN} you can issue the following command in the 1563@file{gcc} directory of your GCC source tree:@* 1564@smallexample 1565grep ^language= */config-lang.in 1566@end smallexample 1567Currently, you can use any of the following: 1568@code{all}, @code{ada}, @code{c}, @code{c++}, @code{fortran}, 1569@code{go}, @code{java}, @code{jit}, @code{lto}, @code{objc}, @code{obj-c++}. 1570Building the Ada compiler has special requirements, see below. 1571If you do not pass this flag, or specify the option @code{all}, then all 1572default languages available in the @file{gcc} sub-tree will be configured. 1573Ada, Go, Jit, and Objective-C++ are not default languages. LTO is not a 1574default language, but is built by default because @option{--enable-lto} is 1575enabled by default. The other languages are default languages. 1576 1577@item --enable-stage1-languages=@var{lang1},@var{lang2},@dots{} 1578Specify that a particular subset of compilers and their runtime 1579libraries should be built with the system C compiler during stage 1 of 1580the bootstrap process, rather than only in later stages with the 1581bootstrapped C compiler. The list of valid values is the same as for 1582@option{--enable-languages}, and the option @code{all} will select all 1583of the languages enabled by @option{--enable-languages}. This option is 1584primarily useful for GCC development; for instance, when a development 1585version of the compiler cannot bootstrap due to compiler bugs, or when 1586one is debugging front ends other than the C front end. When this 1587option is used, one can then build the target libraries for the 1588specified languages with the stage-1 compiler by using @command{make 1589stage1-bubble all-target}, or run the testsuite on the stage-1 compiler 1590for the specified languages using @command{make stage1-start check-gcc}. 1591 1592@item --disable-libada 1593Specify that the run-time libraries and tools used by GNAT should not 1594be built. This can be useful for debugging, or for compatibility with 1595previous Ada build procedures, when it was required to explicitly 1596do a @samp{make -C gcc gnatlib_and_tools}. 1597 1598@item --disable-libsanitizer 1599Specify that the run-time libraries for the various sanitizers should 1600not be built. 1601 1602@item --disable-libssp 1603Specify that the run-time libraries for stack smashing protection 1604should not be built. 1605 1606@item --disable-libquadmath 1607Specify that the GCC quad-precision math library should not be built. 1608On some systems, the library is required to be linkable when building 1609the Fortran front end, unless @option{--disable-libquadmath-support} 1610is used. 1611 1612@item --disable-libquadmath-support 1613Specify that the Fortran front end and @code{libgfortran} do not add 1614support for @code{libquadmath} on systems supporting it. 1615 1616@item --disable-libgomp 1617Specify that the GNU Offloading and Multi Processing Runtime Library 1618should not be built. 1619 1620@item --disable-libvtv 1621Specify that the run-time libraries used by vtable verification 1622should not be built. 1623 1624@item --with-dwarf2 1625Specify that the compiler should 1626use DWARF 2 debugging information as the default. 1627 1628@item --with-advance-toolchain=@var{at} 1629On 64-bit PowerPC Linux systems, configure the compiler to use the 1630header files, library files, and the dynamic linker from the Advance 1631Toolchain release @var{at} instead of the default versions that are 1632provided by the Linux distribution. In general, this option is 1633intended for the developers of GCC, and it is not intended for general 1634use. 1635 1636@item --enable-targets=all 1637@itemx --enable-targets=@var{target_list} 1638Some GCC targets, e.g.@: powerpc64-linux, build bi-arch compilers. 1639These are compilers that are able to generate either 64-bit or 32-bit 1640code. Typically, the corresponding 32-bit target, e.g.@: 1641powerpc-linux for powerpc64-linux, only generates 32-bit code. This 1642option enables the 32-bit target to be a bi-arch compiler, which is 1643useful when you want a bi-arch compiler that defaults to 32-bit, and 1644you are building a bi-arch or multi-arch binutils in a combined tree. 1645On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64), 1646defaulted to o32. 1647Currently, this option only affects sparc-linux, powerpc-linux, x86-linux, 1648mips-linux and s390-linux. 1649 1650@item --enable-default-pie 1651Turn on @option{-fPIE} and @option{-pie} by default. 1652 1653@item --enable-secureplt 1654This option enables @option{-msecure-plt} by default for powerpc-linux. 1655@ifnothtml 1656@xref{RS/6000 and PowerPC Options,, RS/6000 and PowerPC Options, gcc, 1657Using the GNU Compiler Collection (GCC)}, 1658@end ifnothtml 1659@ifhtml 1660See ``RS/6000 and PowerPC Options'' in the main manual 1661@end ifhtml 1662 1663@item --enable-default-ssp 1664Turn on @option{-fstack-protector-strong} by default. 1665 1666@item --enable-cld 1667This option enables @option{-mcld} by default for 32-bit x86 targets. 1668@ifnothtml 1669@xref{i386 and x86-64 Options,, i386 and x86-64 Options, gcc, 1670Using the GNU Compiler Collection (GCC)}, 1671@end ifnothtml 1672@ifhtml 1673See ``i386 and x86-64 Options'' in the main manual 1674@end ifhtml 1675 1676@item --enable-win32-registry 1677@itemx --enable-win32-registry=@var{key} 1678@itemx --disable-win32-registry 1679The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC 1680to look up installations paths in the registry using the following key: 1681 1682@smallexample 1683@code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}} 1684@end smallexample 1685 1686@var{key} defaults to GCC version number, and can be overridden by the 1687@option{--enable-win32-registry=@var{key}} option. Vendors and distributors 1688who use custom installers are encouraged to provide a different key, 1689perhaps one comprised of vendor name and GCC version number, to 1690avoid conflict with existing installations. This feature is enabled 1691by default, and can be disabled by @option{--disable-win32-registry} 1692option. This option has no effect on the other hosts. 1693 1694@item --nfp 1695Specify that the machine does not have a floating point unit. This 1696option only applies to @samp{m68k-sun-sunos@var{n}}. On any other 1697system, @option{--nfp} has no effect. 1698 1699@item --enable-werror 1700@itemx --disable-werror 1701@itemx --enable-werror=yes 1702@itemx --enable-werror=no 1703When you specify this option, it controls whether certain files in the 1704compiler are built with @option{-Werror} in bootstrap stage2 and later. 1705If you don't specify it, @option{-Werror} is turned on for the main 1706development trunk. However it defaults to off for release branches and 1707final releases. The specific files which get @option{-Werror} are 1708controlled by the Makefiles. 1709 1710@item --enable-checking 1711@itemx --enable-checking=@var{list} 1712When you specify this option, the compiler is built to perform internal 1713consistency checks of the requested complexity. This does not change the 1714generated code, but adds error checking within the compiler. This will 1715slow down the compiler and may only work properly if you are building 1716the compiler with GCC@. This is @samp{yes} by default when building 1717from SVN or snapshots, but @samp{release} for releases. The default 1718for building the stage1 compiler is @samp{yes}. More control 1719over the checks may be had by specifying @var{list}. The categories of 1720checks available are @samp{yes} (most common checks 1721@samp{assert,misc,tree,gc,rtlflag,runtime}), @samp{no} (no checks at 1722all), @samp{all} (all but @samp{valgrind}), @samp{release} (cheapest 1723checks @samp{assert,runtime}) or @samp{none} (same as @samp{no}). 1724Individual checks can be enabled with these flags @samp{assert}, 1725@samp{df}, @samp{fold}, @samp{gc}, @samp{gcac} @samp{misc}, @samp{rtl}, 1726@samp{rtlflag}, @samp{runtime}, @samp{tree}, and @samp{valgrind}. 1727 1728The @samp{valgrind} check requires the external @command{valgrind} 1729simulator, available from @uref{http://valgrind.org/}. The 1730@samp{df}, @samp{rtl}, @samp{gcac} and @samp{valgrind} checks are very expensive. 1731To disable all checking, @samp{--disable-checking} or 1732@samp{--enable-checking=none} must be explicitly requested. Disabling 1733assertions will make the compiler and runtime slightly faster but 1734increase the risk of undetected internal errors causing wrong code to be 1735generated. 1736 1737@item --disable-stage1-checking 1738@itemx --enable-stage1-checking 1739@itemx --enable-stage1-checking=@var{list} 1740If no @option{--enable-checking} option is specified the stage1 1741compiler will be built with @samp{yes} checking enabled, otherwise 1742the stage1 checking flags are the same as specified by 1743@option{--enable-checking}. To build the stage1 compiler with 1744different checking options use @option{--enable-stage1-checking}. 1745The list of checking options is the same as for @option{--enable-checking}. 1746If your system is too slow or too small to bootstrap a released compiler 1747with checking for stage1 enabled, you can use @samp{--disable-stage1-checking} 1748to disable checking for the stage1 compiler. 1749 1750@item --enable-coverage 1751@itemx --enable-coverage=@var{level} 1752With this option, the compiler is built to collect self coverage 1753information, every time it is run. This is for internal development 1754purposes, and only works when the compiler is being built with gcc. The 1755@var{level} argument controls whether the compiler is built optimized or 1756not, values are @samp{opt} and @samp{noopt}. For coverage analysis you 1757want to disable optimization, for performance analysis you want to 1758enable optimization. When coverage is enabled, the default level is 1759without optimization. 1760 1761@item --enable-gather-detailed-mem-stats 1762When this option is specified more detailed information on memory 1763allocation is gathered. This information is printed when using 1764@option{-fmem-report}. 1765 1766@item --enable-valgrind-annotations 1767Mark selected memory related operations in the compiler when run under 1768valgrind to suppress false positives. 1769 1770@item --enable-nls 1771@itemx --disable-nls 1772The @option{--enable-nls} option enables Native Language Support (NLS), 1773which lets GCC output diagnostics in languages other than American 1774English. Native Language Support is enabled by default if not doing a 1775canadian cross build. The @option{--disable-nls} option disables NLS@. 1776 1777@item --with-included-gettext 1778If NLS is enabled, the @option{--with-included-gettext} option causes the build 1779procedure to prefer its copy of GNU @command{gettext}. 1780 1781@item --with-catgets 1782If NLS is enabled, and if the host lacks @code{gettext} but has the 1783inferior @code{catgets} interface, the GCC build procedure normally 1784ignores @code{catgets} and instead uses GCC's copy of the GNU 1785@code{gettext} library. The @option{--with-catgets} option causes the 1786build procedure to use the host's @code{catgets} in this situation. 1787 1788@item --with-libiconv-prefix=@var{dir} 1789Search for libiconv header files in @file{@var{dir}/include} and 1790libiconv library files in @file{@var{dir}/lib}. 1791 1792@item --enable-obsolete 1793Enable configuration for an obsoleted system. If you attempt to 1794configure GCC for a system (build, host, or target) which has been 1795obsoleted, and you do not specify this flag, configure will halt with an 1796error message. 1797 1798All support for systems which have been obsoleted in one release of GCC 1799is removed entirely in the next major release, unless someone steps 1800forward to maintain the port. 1801 1802@item --enable-decimal-float 1803@itemx --enable-decimal-float=yes 1804@itemx --enable-decimal-float=no 1805@itemx --enable-decimal-float=bid 1806@itemx --enable-decimal-float=dpd 1807@itemx --disable-decimal-float 1808Enable (or disable) support for the C decimal floating point extension 1809that is in the IEEE 754-2008 standard. This is enabled by default only 1810on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also 1811support it, but require the user to specifically enable it. You can 1812optionally control which decimal floating point format is used (either 1813@samp{bid} or @samp{dpd}). The @samp{bid} (binary integer decimal) 1814format is default on i386 and x86_64 systems, and the @samp{dpd} 1815(densely packed decimal) format is default on PowerPC systems. 1816 1817@item --enable-fixed-point 1818@itemx --disable-fixed-point 1819Enable (or disable) support for C fixed-point arithmetic. 1820This option is enabled by default for some targets (such as MIPS) which 1821have hardware-support for fixed-point operations. On other targets, you 1822may enable this option manually. 1823 1824@item --with-long-double-128 1825Specify if @code{long double} type should be 128-bit by default on selected 1826GNU/Linux architectures. If using @code{--without-long-double-128}, 1827@code{long double} will be by default 64-bit, the same as @code{double} type. 1828When neither of these configure options are used, the default will be 1829128-bit @code{long double} when built against GNU C Library 2.4 and later, 183064-bit @code{long double} otherwise. 1831 1832@item --enable-fdpic 1833On SH Linux systems, generate ELF FDPIC code. 1834 1835@item --with-gmp=@var{pathname} 1836@itemx --with-gmp-include=@var{pathname} 1837@itemx --with-gmp-lib=@var{pathname} 1838@itemx --with-mpfr=@var{pathname} 1839@itemx --with-mpfr-include=@var{pathname} 1840@itemx --with-mpfr-lib=@var{pathname} 1841@itemx --with-mpc=@var{pathname} 1842@itemx --with-mpc-include=@var{pathname} 1843@itemx --with-mpc-lib=@var{pathname} 1844If you want to build GCC but do not have the GMP library, the MPFR 1845library and/or the MPC library installed in a standard location and 1846do not have their sources present in the GCC source tree then you 1847can explicitly specify the directory where they are installed 1848(@samp{--with-gmp=@var{gmpinstalldir}}, 1849@samp{--with-mpfr=@/@var{mpfrinstalldir}}, 1850@samp{--with-mpc=@/@var{mpcinstalldir}}). The 1851@option{--with-gmp=@/@var{gmpinstalldir}} option is shorthand for 1852@option{--with-gmp-lib=@/@var{gmpinstalldir}/lib} and 1853@option{--with-gmp-include=@/@var{gmpinstalldir}/include}. Likewise the 1854@option{--with-mpfr=@/@var{mpfrinstalldir}} option is shorthand for 1855@option{--with-mpfr-lib=@/@var{mpfrinstalldir}/lib} and 1856@option{--with-mpfr-include=@/@var{mpfrinstalldir}/include}, also the 1857@option{--with-mpc=@/@var{mpcinstalldir}} option is shorthand for 1858@option{--with-mpc-lib=@/@var{mpcinstalldir}/lib} and 1859@option{--with-mpc-include=@/@var{mpcinstalldir}/include}. If these 1860shorthand assumptions are not correct, you can use the explicit 1861include and lib options directly. You might also need to ensure the 1862shared libraries can be found by the dynamic linker when building and 1863using GCC, for example by setting the runtime shared library path 1864variable (@env{LD_LIBRARY_PATH} on GNU/Linux and Solaris systems). 1865 1866These flags are applicable to the host platform only. When building 1867a cross compiler, they will not be used to configure target libraries. 1868 1869@item --with-isl=@var{pathname} 1870@itemx --with-isl-include=@var{pathname} 1871@itemx --with-isl-lib=@var{pathname} 1872If you do not have the isl library installed in a standard location and you 1873want to build GCC, you can explicitly specify the directory where it is 1874installed (@samp{--with-isl=@/@var{islinstalldir}}). The 1875@option{--with-isl=@/@var{islinstalldir}} option is shorthand for 1876@option{--with-isl-lib=@/@var{islinstalldir}/lib} and 1877@option{--with-isl-include=@/@var{islinstalldir}/include}. If this 1878shorthand assumption is not correct, you can use the explicit 1879include and lib options directly. 1880 1881These flags are applicable to the host platform only. When building 1882a cross compiler, they will not be used to configure target libraries. 1883 1884@item --with-stage1-ldflags=@var{flags} 1885This option may be used to set linker flags to be used when linking 1886stage 1 of GCC. These are also used when linking GCC if configured with 1887@option{--disable-bootstrap}. If @option{--with-stage1-libs} is not set to a 1888value, then the default is @samp{-static-libstdc++ -static-libgcc}, if 1889supported. 1890 1891@item --with-stage1-libs=@var{libs} 1892This option may be used to set libraries to be used when linking stage 1 1893of GCC. These are also used when linking GCC if configured with 1894@option{--disable-bootstrap}. 1895 1896@item --with-boot-ldflags=@var{flags} 1897This option may be used to set linker flags to be used when linking 1898stage 2 and later when bootstrapping GCC. If --with-boot-libs 1899is not is set to a value, then the default is 1900@samp{-static-libstdc++ -static-libgcc}. 1901 1902@item --with-boot-libs=@var{libs} 1903This option may be used to set libraries to be used when linking stage 2 1904and later when bootstrapping GCC. 1905 1906@item --with-debug-prefix-map=@var{map} 1907Convert source directory names using @option{-fdebug-prefix-map} when 1908building runtime libraries. @samp{@var{map}} is a space-separated 1909list of maps of the form @samp{@var{old}=@var{new}}. 1910 1911@item --enable-linker-build-id 1912Tells GCC to pass @option{--build-id} option to the linker for all final 1913links (links performed without the @option{-r} or @option{--relocatable} 1914option), if the linker supports it. If you specify 1915@option{--enable-linker-build-id}, but your linker does not 1916support @option{--build-id} option, a warning is issued and the 1917@option{--enable-linker-build-id} option is ignored. The default is off. 1918 1919@item --with-linker-hash-style=@var{choice} 1920Tells GCC to pass @option{--hash-style=@var{choice}} option to the 1921linker for all final links. @var{choice} can be one of 1922@samp{sysv}, @samp{gnu}, and @samp{both} where @samp{sysv} is the default. 1923 1924@item --enable-gnu-unique-object 1925@itemx --disable-gnu-unique-object 1926Tells GCC to use the gnu_unique_object relocation for C++ template 1927static data members and inline function local statics. Enabled by 1928default for a toolchain with an assembler that accepts it and 1929GLIBC 2.11 or above, otherwise disabled. 1930 1931@item --with-diagnostics-color=@var{choice} 1932Tells GCC to use @var{choice} as the default for @option{-fdiagnostics-color=} 1933option (if not used explicitly on the command line). @var{choice} 1934can be one of @samp{never}, @samp{auto}, @samp{always}, and @samp{auto-if-env} 1935where @samp{auto} is the default. @samp{auto-if-env} means that 1936@option{-fdiagnostics-color=auto} will be the default if @code{GCC_COLORS} 1937is present and non-empty in the environment, and 1938@option{-fdiagnostics-color=never} otherwise. 1939 1940@item --enable-lto 1941@itemx --disable-lto 1942Enable support for link-time optimization (LTO). This is enabled by 1943default, and may be disabled using @option{--disable-lto}. 1944 1945@item --enable-linker-plugin-configure-flags=FLAGS 1946@itemx --enable-linker-plugin-flags=FLAGS 1947By default, linker plugins (such as the LTO plugin) are built for the 1948host system architecture. For the case that the linker has a 1949different (but run-time compatible) architecture, these flags can be 1950specified to build plugins that are compatible to the linker. For 1951example, if you are building GCC for a 64-bit x86_64 1952(@samp{x86_64-unknown-linux-gnu}) host system, but have a 32-bit x86 1953GNU/Linux (@samp{i686-pc-linux-gnu}) linker executable (which is 1954executable on the former system), you can configure GCC as follows for 1955getting compatible linker plugins: 1956 1957@smallexample 1958% @var{srcdir}/configure \ 1959 --host=x86_64-unknown-linux-gnu \ 1960 --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \ 1961 --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib' 1962@end smallexample 1963 1964@item --with-plugin-ld=@var{pathname} 1965Enable an alternate linker to be used at link-time optimization (LTO) 1966link time when @option{-fuse-linker-plugin} is enabled. 1967This linker should have plugin support such as gold starting with 1968version 2.20 or GNU ld starting with version 2.21. 1969See @option{-fuse-linker-plugin} for details. 1970 1971@item --enable-canonical-system-headers 1972@itemx --disable-canonical-system-headers 1973Enable system header path canonicalization for @file{libcpp}. This can 1974produce shorter header file paths in diagnostics and dependency output 1975files, but these changed header paths may conflict with some compilation 1976environments. Enabled by default, and may be disabled using 1977@option{--disable-canonical-system-headers}. 1978 1979@item --with-glibc-version=@var{major}.@var{minor} 1980Tell GCC that when the GNU C Library (glibc) is used on the target it 1981will be version @var{major}.@var{minor} or later. Normally this can 1982be detected from the C library's header files, but this option may be 1983needed when bootstrapping a cross toolchain without the header files 1984available for building the initial bootstrap compiler. 1985 1986If GCC is configured with some multilibs that use glibc and some that 1987do not, this option applies only to the multilibs that use glibc. 1988However, such configurations may not work well as not all the relevant 1989configuration in GCC is on a per-multilib basis. 1990 1991@item --enable-as-accelerator-for=@var{target} 1992Build as offload target compiler. Specify offload host triple by @var{target}. 1993 1994@item --enable-offload-targets=@var{target1}[=@var{path1}],@dots{},@var{targetN}[=@var{pathN}] 1995Enable offloading to targets @var{target1}, @dots{}, @var{targetN}. 1996Offload compilers are expected to be already installed. Default search 1997path for them is @file{@var{exec-prefix}}, but it can be changed by 1998specifying paths @var{path1}, @dots{}, @var{pathN}. 1999 2000@smallexample 2001% @var{srcdir}/configure \ 2002 --enable-offload-target=i686-unknown-linux-gnu=/path/to/i686/compiler,x86_64-pc-linux-gnu 2003@end smallexample 2004 2005If @samp{hsa} is specified as one of the targets, the compiler will be 2006built with support for HSA GPU accelerators. Because the same 2007compiler will emit the accelerator code, no path should be specified. 2008 2009@item --with-hsa-runtime=@var{pathname} 2010@itemx --with-hsa-runtime-include=@var{pathname} 2011@itemx --with-hsa-runtime-lib=@var{pathname} 2012 2013If you configure GCC with HSA offloading but do not have the HSA 2014run-time library installed in a standard location then you can 2015explicitly specify the directory where they are installed. The 2016@option{--with-hsa-runtime=@/@var{hsainstalldir}} option is a 2017shorthand for 2018@option{--with-hsa-runtime-lib=@/@var{hsainstalldir}/lib} and 2019@option{--with-hsa-runtime-include=@/@var{hsainstalldir}/include}. 2020 2021@item --with-hsa-kmt-lib=@var{pathname} 2022 2023If you configure GCC with HSA offloading but do not have the HSA 2024KMT library installed in a standard location then you can 2025explicitly specify the directory where it resides. 2026@end table 2027 2028@subheading Cross-Compiler-Specific Options 2029The following options only apply to building cross compilers. 2030 2031@table @code 2032@item --with-sysroot 2033@itemx --with-sysroot=@var{dir} 2034Tells GCC to consider @var{dir} as the root of a tree that contains 2035(a subset of) the root filesystem of the target operating system. 2036Target system headers, libraries and run-time object files will be 2037searched for in there. More specifically, this acts as if 2038@option{--sysroot=@var{dir}} was added to the default options of the built 2039compiler. The specified directory is not copied into the 2040install tree, unlike the options @option{--with-headers} and 2041@option{--with-libs} that this option obsoletes. The default value, 2042in case @option{--with-sysroot} is not given an argument, is 2043@option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a 2044subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to 2045the GCC binaries if the installation tree is moved. 2046 2047This option affects the system root for the compiler used to build 2048target libraries (which runs on the build system) and the compiler newly 2049installed with @code{make install}; it does not affect the compiler which is 2050used to build GCC itself. 2051 2052If you specify the @option{--with-native-system-header-dir=@var{dirname}} 2053option then the compiler will search that directory within @var{dirname} for 2054native system headers rather than the default @file{/usr/include}. 2055 2056@item --with-build-sysroot 2057@itemx --with-build-sysroot=@var{dir} 2058Tells GCC to consider @var{dir} as the system root (see 2059@option{--with-sysroot}) while building target libraries, instead of 2060the directory specified with @option{--with-sysroot}. This option is 2061only useful when you are already using @option{--with-sysroot}. You 2062can use @option{--with-build-sysroot} when you are configuring with 2063@option{--prefix} set to a directory that is different from the one in 2064which you are installing GCC and your target libraries. 2065 2066This option affects the system root for the compiler used to build 2067target libraries (which runs on the build system); it does not affect 2068the compiler which is used to build GCC itself. 2069 2070If you specify the @option{--with-native-system-header-dir=@var{dirname}} 2071option then the compiler will search that directory within @var{dirname} for 2072native system headers rather than the default @file{/usr/include}. 2073 2074@item --with-headers 2075@itemx --with-headers=@var{dir} 2076Deprecated in favor of @option{--with-sysroot}. 2077Specifies that target headers are available when building a cross compiler. 2078The @var{dir} argument specifies a directory which has the target include 2079files. These include files will be copied into the @file{gcc} install 2080directory. @emph{This option with the @var{dir} argument is required} when 2081building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include} 2082doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does 2083pre-exist, the @var{dir} argument may be omitted. @command{fixincludes} 2084will be run on these files to make them compatible with GCC@. 2085 2086@item --without-headers 2087Tells GCC not use any target headers from a libc when building a cross 2088compiler. When crossing to GNU/Linux, you need the headers so GCC 2089can build the exception handling for libgcc. 2090 2091@item --with-libs 2092@itemx --with-libs="@var{dir1} @var{dir2} @dots{} @var{dirN}" 2093Deprecated in favor of @option{--with-sysroot}. 2094Specifies a list of directories which contain the target runtime 2095libraries. These libraries will be copied into the @file{gcc} install 2096directory. If the directory list is omitted, this option has no 2097effect. 2098 2099@item --with-newlib 2100Specifies that @samp{newlib} is 2101being used as the target C library. This causes @code{__eprintf} to be 2102omitted from @file{libgcc.a} on the assumption that it will be provided by 2103@samp{newlib}. 2104 2105@item --with-avrlibc 2106Specifies that @samp{AVR-Libc} is 2107being used as the target C library. This causes float support 2108functions like @code{__addsf3} to be omitted from @file{libgcc.a} on 2109the assumption that it will be provided by @file{libm.a}. For more 2110technical details, cf. @uref{http://gcc.gnu.org/PR54461,,PR54461}. 2111This option is only supported for the AVR target. It is not supported for 2112RTEMS configurations, which currently use newlib. The option is 2113supported since version 4.7.2 and is the default in 4.8.0 and newer. 2114 2115@item --with-nds32-lib=@var{library} 2116Specifies that @var{library} setting is used for building @file{libgcc.a}. 2117Currently, the valid @var{library} is @samp{newlib} or @samp{mculib}. 2118This option is only supported for the NDS32 target. 2119 2120@item --with-build-time-tools=@var{dir} 2121Specifies where to find the set of target tools (assembler, linker, etc.) 2122that will be used while building GCC itself. This option can be useful 2123if the directory layouts are different between the system you are building 2124GCC on, and the system where you will deploy it. 2125 2126For example, on an @samp{ia64-hp-hpux} system, you may have the GNU 2127assembler and linker in @file{/usr/bin}, and the native tools in a 2128different path, and build a toolchain that expects to find the 2129native tools in @file{/usr/bin}. 2130 2131When you use this option, you should ensure that @var{dir} includes 2132@command{ar}, @command{as}, @command{ld}, @command{nm}, 2133@command{ranlib} and @command{strip} if necessary, and possibly 2134@command{objdump}. Otherwise, GCC may use an inconsistent set of 2135tools. 2136@end table 2137 2138@subsubheading Overriding @command{configure} test results 2139 2140Sometimes, it might be necessary to override the result of some 2141@command{configure} test, for example in order to ease porting to a new 2142system or work around a bug in a test. The toplevel @command{configure} 2143script provides three variables for this: 2144 2145@table @code 2146 2147@item build_configargs 2148@cindex @code{build_configargs} 2149The contents of this variable is passed to all build @command{configure} 2150scripts. 2151 2152@item host_configargs 2153@cindex @code{host_configargs} 2154The contents of this variable is passed to all host @command{configure} 2155scripts. 2156 2157@item target_configargs 2158@cindex @code{target_configargs} 2159The contents of this variable is passed to all target @command{configure} 2160scripts. 2161 2162@end table 2163 2164In order to avoid shell and @command{make} quoting issues for complex 2165overrides, you can pass a setting for @env{CONFIG_SITE} and set 2166variables in the site file. 2167 2168@subheading Java-Specific Options 2169 2170The following option applies to the build of the Java front end. 2171 2172@table @code 2173@item --disable-libgcj 2174Specify that the run-time libraries 2175used by GCJ should not be built. This is useful in case you intend 2176to use GCJ with some other run-time, or you're going to install it 2177separately, or it just happens not to build on your particular 2178machine. In general, if the Java front end is enabled, the GCJ 2179libraries will be enabled too, unless they're known to not work on 2180the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you 2181may need to port it; in this case, before modifying the top-level 2182@file{configure.ac} so that @samp{libgcj} is enabled by default on this platform, 2183you may use @option{--enable-libgcj} to override the default. 2184 2185@end table 2186 2187The following options apply to building @samp{libgcj}. 2188 2189@subsubheading General Options 2190 2191@table @code 2192@item --enable-java-maintainer-mode 2193By default the @samp{libjava} build will not attempt to compile the 2194@file{.java} source files to @file{.class}. Instead, it will use the 2195@file{.class} files from the source tree. If you use this option you 2196must have executables named @command{ecj1} and @command{gjavah} in your path 2197for use by the build. You must use this option if you intend to 2198modify any @file{.java} files in @file{libjava}. 2199 2200@item --with-java-home=@var{dirname} 2201This @samp{libjava} option overrides the default value of the 2202@samp{java.home} system property. It is also used to set 2203@samp{sun.boot.class.path} to @file{@var{dirname}/lib/rt.jar}. By 2204default @samp{java.home} is set to @file{@var{prefix}} and 2205@samp{sun.boot.class.path} to 2206@file{@var{datadir}/java/libgcj-@var{version}.jar}. 2207 2208@item --with-ecj-jar=@var{filename} 2209This option can be used to specify the location of an external jar 2210file containing the Eclipse Java compiler. A specially modified 2211version of this compiler is used by @command{gcj} to parse 2212@file{.java} source files. If this option is given, the 2213@samp{libjava} build will create and install an @file{ecj1} executable 2214which uses this jar file at runtime. 2215 2216If this option is not given, but an @file{ecj.jar} file is found in 2217the topmost source tree at configure time, then the @samp{libgcj} 2218build will create and install @file{ecj1}, and will also install the 2219discovered @file{ecj.jar} into a suitable place in the install tree. 2220 2221If @file{ecj1} is not installed, then the user will have to supply one 2222on his path in order for @command{gcj} to properly parse @file{.java} 2223source files. A suitable jar is available from 2224@uref{ftp://sourceware.org/pub/java/}. 2225 2226@item --disable-getenv-properties 2227Don't set system properties from @env{GCJ_PROPERTIES}. 2228 2229@item --enable-hash-synchronization 2230Use a global hash table for monitor locks. Ordinarily, 2231@samp{libgcj}'s @samp{configure} script automatically makes 2232the correct choice for this option for your platform. Only use 2233this if you know you need the library to be configured differently. 2234 2235@item --enable-interpreter 2236Enable the Java interpreter. The interpreter is automatically 2237enabled by default on all platforms that support it. This option 2238is really only useful if you want to disable the interpreter 2239(using @option{--disable-interpreter}). 2240 2241@item --disable-java-net 2242Disable java.net. This disables the native part of java.net only, 2243using non-functional stubs for native method implementations. 2244 2245@item --disable-jvmpi 2246Disable JVMPI support. 2247 2248@item --disable-libgcj-bc 2249Disable BC ABI compilation of certain parts of libgcj. By default, 2250some portions of libgcj are compiled with @option{-findirect-dispatch} 2251and @option{-fno-indirect-classes}, allowing them to be overridden at 2252run-time. 2253 2254If @option{--disable-libgcj-bc} is specified, libgcj is built without 2255these options. This allows the compile-time linker to resolve 2256dependencies when statically linking to libgcj. However it makes it 2257impossible to override the affected portions of libgcj at run-time. 2258 2259@item --enable-reduced-reflection 2260Build most of libgcj with @option{-freduced-reflection}. This reduces 2261the size of libgcj at the expense of not being able to do accurate 2262reflection on the classes it contains. This option is safe if you 2263know that code using libgcj will never use reflection on the standard 2264runtime classes in libgcj (including using serialization, RMI or CORBA). 2265 2266@item --with-ecos 2267Enable runtime eCos target support. 2268 2269@item --without-libffi 2270Don't use @samp{libffi}. This will disable the interpreter and JNI 2271support as well, as these require @samp{libffi} to work. 2272 2273@item --enable-libgcj-debug 2274Enable runtime debugging code. 2275 2276@item --enable-libgcj-multifile 2277If specified, causes all @file{.java} source files to be 2278compiled into @file{.class} files in one invocation of 2279@samp{gcj}. This can speed up build time, but is more 2280resource-intensive. If this option is unspecified or 2281disabled, @samp{gcj} is invoked once for each @file{.java} 2282file to compile into a @file{.class} file. 2283 2284@item --with-libiconv-prefix=DIR 2285Search for libiconv in @file{DIR/include} and @file{DIR/lib}. 2286 2287@item --with-system-zlib 2288Use installed @samp{zlib} rather than that included with GCC@. 2289 2290@item --with-win32-nlsapi=ansi, unicows or unicode 2291Indicates how MinGW @samp{libgcj} translates between UNICODE 2292characters and the Win32 API@. 2293 2294@item --enable-java-home 2295If enabled, this creates a JPackage compatible SDK environment during install. 2296Note that if --enable-java-home is used, --with-arch-directory=ARCH must also 2297be specified. 2298 2299@item --with-arch-directory=ARCH 2300Specifies the name to use for the @file{jre/lib/ARCH} directory in the SDK 2301environment created when --enable-java-home is passed. Typical names for this 2302directory include i386, amd64, ia64, etc. 2303 2304@item --with-os-directory=DIR 2305Specifies the OS directory for the SDK include directory. This is set to auto 2306detect, and is typically 'linux'. 2307 2308@item --with-origin-name=NAME 2309Specifies the JPackage origin name. This defaults to the 'gcj' in 2310java-1.5.0-gcj. 2311 2312@item --with-arch-suffix=SUFFIX 2313Specifies the suffix for the sdk directory. Defaults to the empty string. 2314Examples include '.x86_64' in 'java-1.5.0-gcj-1.5.0.0.x86_64'. 2315 2316@item --with-jvm-root-dir=DIR 2317Specifies where to install the SDK. Default is $(prefix)/lib/jvm. 2318 2319@item --with-jvm-jar-dir=DIR 2320Specifies where to install jars. Default is $(prefix)/lib/jvm-exports. 2321 2322@item --with-python-dir=DIR 2323Specifies where to install the Python modules used for aot-compile. DIR should 2324not include the prefix used in installation. For example, if the Python modules 2325are to be installed in /usr/lib/python2.5/site-packages, then 2326--with-python-dir=/lib/python2.5/site-packages should be passed. If this is 2327not specified, then the Python modules are installed in $(prefix)/share/python. 2328 2329@item --enable-aot-compile-rpm 2330Adds aot-compile-rpm to the list of installed scripts. 2331 2332@item --enable-browser-plugin 2333Build the gcjwebplugin web browser plugin. 2334 2335@item --enable-static-libjava 2336Build static libraries in libjava. The default is to only build shared 2337libraries. 2338 2339@table @code 2340@item ansi 2341Use the single-byte @code{char} and the Win32 A functions natively, 2342translating to and from UNICODE when using these functions. If 2343unspecified, this is the default. 2344 2345@item unicows 2346Use the @code{WCHAR} and Win32 W functions natively. Adds 2347@code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}. 2348@file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines 2349running built executables. @file{libunicows.a}, an open-source 2350import library around Microsoft's @code{unicows.dll}, is obtained from 2351@uref{http://libunicows.sourceforge.net/}, which also gives details 2352on getting @file{unicows.dll} from Microsoft. 2353 2354@item unicode 2355Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not} 2356add @code{-lunicows} to @file{libgcj.spec}. The built executables will 2357only run on Microsoft Windows NT and above. 2358@end table 2359@end table 2360 2361@subsubheading AWT-Specific Options 2362 2363@table @code 2364@item --with-x 2365Use the X Window System. 2366 2367@item --enable-java-awt=PEER(S) 2368Specifies the AWT peer library or libraries to build alongside 2369@samp{libgcj}. If this option is unspecified or disabled, AWT 2370will be non-functional. Current valid values are @option{gtk} and 2371@option{xlib}. Multiple libraries should be separated by a 2372comma (i.e.@: @option{--enable-java-awt=gtk,xlib}). 2373 2374@item --enable-gtk-cairo 2375Build the cairo Graphics2D implementation on GTK@. 2376 2377@item --enable-java-gc=TYPE 2378Choose garbage collector. Defaults to @option{boehm} if unspecified. 2379 2380@item --disable-gtktest 2381Do not try to compile and run a test GTK+ program. 2382 2383@item --disable-glibtest 2384Do not try to compile and run a test GLIB program. 2385 2386@item --with-libart-prefix=PFX 2387Prefix where libart is installed (optional). 2388 2389@item --with-libart-exec-prefix=PFX 2390Exec prefix where libart is installed (optional). 2391 2392@item --disable-libarttest 2393Do not try to compile and run a test libart program. 2394 2395@end table 2396 2397 2398@html 2399<hr /> 2400<p> 2401@end html 2402@ifhtml 2403@uref{./index.html,,Return to the GCC Installation page} 2404@end ifhtml 2405@end ifset 2406 2407@c ***Building**************************************************************** 2408@ifnothtml 2409@comment node-name, next, previous, up 2410@node Building, Testing, Configuration, Installing GCC 2411@end ifnothtml 2412@ifset buildhtml 2413@ifnothtml 2414@chapter Building 2415@end ifnothtml 2416@cindex Installing GCC: Building 2417 2418Now that GCC is configured, you are ready to build the compiler and 2419runtime libraries. 2420 2421Some commands executed when making the compiler may fail (return a 2422nonzero status) and be ignored by @command{make}. These failures, which 2423are often due to files that were not found, are expected, and can safely 2424be ignored. 2425 2426It is normal to have compiler warnings when compiling certain files. 2427Unless you are a GCC developer, you can generally ignore these warnings 2428unless they cause compilation to fail. Developers should attempt to fix 2429any warnings encountered, however they can temporarily continue past 2430warnings-as-errors by specifying the configure flag 2431@option{--disable-werror}. 2432 2433On certain old systems, defining certain environment variables such as 2434@env{CC} can interfere with the functioning of @command{make}. 2435 2436If you encounter seemingly strange errors when trying to build the 2437compiler in a directory other than the source directory, it could be 2438because you have previously configured the compiler in the source 2439directory. Make sure you have done all the necessary preparations. 2440 2441If you build GCC on a BSD system using a directory stored in an old System 2442V file system, problems may occur in running @command{fixincludes} if the 2443System V file system doesn't support symbolic links. These problems 2444result in a failure to fix the declaration of @code{size_t} in 2445@file{sys/types.h}. If you find that @code{size_t} is a signed type and 2446that type mismatches occur, this could be the cause. 2447 2448The solution is not to use such a directory for building GCC@. 2449 2450Similarly, when building from SVN or snapshots, or if you modify 2451@file{*.l} files, you need the Flex lexical analyzer generator 2452installed. If you do not modify @file{*.l} files, releases contain 2453the Flex-generated files and you do not need Flex installed to build 2454them. There is still one Flex-based lexical analyzer (part of the 2455build machinery, not of GCC itself) that is used even if you only 2456build the C front end. 2457 2458When building from SVN or snapshots, or if you modify Texinfo 2459documentation, you need version 4.7 or later of Texinfo installed if you 2460want Info documentation to be regenerated. Releases contain Info 2461documentation pre-built for the unmodified documentation in the release. 2462 2463@section Building a native compiler 2464 2465For a native build, the default configuration is to perform 2466a 3-stage bootstrap of the compiler when @samp{make} is invoked. 2467This will build the entire GCC system and ensure that it compiles 2468itself correctly. It can be disabled with the @option{--disable-bootstrap} 2469parameter to @samp{configure}, but bootstrapping is suggested because 2470the compiler will be tested more completely and could also have 2471better performance. 2472 2473The bootstrapping process will complete the following steps: 2474 2475@itemize @bullet 2476@item 2477Build tools necessary to build the compiler. 2478 2479@item 2480Perform a 3-stage bootstrap of the compiler. This includes building 2481three times the target tools for use by the compiler such as binutils 2482(bfd, binutils, gas, gprof, ld, and opcodes) if they have been 2483individually linked or moved into the top level GCC source tree before 2484configuring. 2485 2486@item 2487Perform a comparison test of the stage2 and stage3 compilers. 2488 2489@item 2490Build runtime libraries using the stage3 compiler from the previous step. 2491 2492@end itemize 2493 2494If you are short on disk space you might consider @samp{make 2495bootstrap-lean} instead. The sequence of compilation is the 2496same described above, but object files from the stage1 and 2497stage2 of the 3-stage bootstrap of the compiler are deleted as 2498soon as they are no longer needed. 2499 2500If you wish to use non-default GCC flags when compiling the stage2 2501and stage3 compilers, set @code{BOOT_CFLAGS} on the command line when 2502doing @samp{make}. For example, if you want to save additional space 2503during the bootstrap and in the final installation as well, you can 2504build the compiler binaries without debugging information as in the 2505following example. This will save roughly 40% of disk space both for 2506the bootstrap and the final installation. (Libraries will still contain 2507debugging information.) 2508 2509@smallexample 2510make BOOT_CFLAGS='-O' bootstrap 2511@end smallexample 2512 2513You can place non-default optimization flags into @code{BOOT_CFLAGS}; they 2514are less well tested here than the default of @samp{-g -O2}, but should 2515still work. In a few cases, you may find that you need to specify special 2516flags such as @option{-msoft-float} here to complete the bootstrap; or, 2517if the native compiler miscompiles the stage1 compiler, you may need 2518to work around this, by choosing @code{BOOT_CFLAGS} to avoid the parts 2519of the stage1 compiler that were miscompiled, or by using @samp{make 2520bootstrap4} to increase the number of stages of bootstrap. 2521 2522@code{BOOT_CFLAGS} does not apply to bootstrapped target libraries. 2523Since these are always compiled with the compiler currently being 2524bootstrapped, you can use @code{CFLAGS_FOR_TARGET} to modify their 2525compilation flags, as for non-bootstrapped target libraries. 2526Again, if the native compiler miscompiles the stage1 compiler, you may 2527need to work around this by avoiding non-working parts of the stage1 2528compiler. Use @code{STAGE1_TFLAGS} to this end. 2529 2530If you used the flag @option{--enable-languages=@dots{}} to restrict 2531the compilers to be built, only those you've actually enabled will be 2532built. This will of course only build those runtime libraries, for 2533which the particular compiler has been built. Please note, 2534that re-defining @env{LANGUAGES} when calling @samp{make} 2535@strong{does not} work anymore! 2536 2537If the comparison of stage2 and stage3 fails, this normally indicates 2538that the stage2 compiler has compiled GCC incorrectly, and is therefore 2539a potentially serious bug which you should investigate and report. (On 2540a few systems, meaningful comparison of object files is impossible; they 2541always appear ``different''. If you encounter this problem, you will 2542need to disable comparison in the @file{Makefile}.) 2543 2544If you do not want to bootstrap your compiler, you can configure with 2545@option{--disable-bootstrap}. In particular cases, you may want to 2546bootstrap your compiler even if the target system is not the same as 2547the one you are building on: for example, you could build a 2548@code{powerpc-unknown-linux-gnu} toolchain on a 2549@code{powerpc64-unknown-linux-gnu} host. In this case, pass 2550@option{--enable-bootstrap} to the configure script. 2551 2552@code{BUILD_CONFIG} can be used to bring in additional customization 2553to the build. It can be set to a whitespace-separated list of names. 2554For each such @code{NAME}, top-level @file{config/@code{NAME}.mk} will 2555be included by the top-level @file{Makefile}, bringing in any settings 2556it contains. The default @code{BUILD_CONFIG} can be set using the 2557configure option @option{--with-build-config=@code{NAME}...}. Some 2558examples of supported build configurations are: 2559 2560@table @asis 2561@item @samp{bootstrap-O1} 2562Removes any @option{-O}-started option from @code{BOOT_CFLAGS}, and adds 2563@option{-O1} to it. @samp{BUILD_CONFIG=bootstrap-O1} is equivalent to 2564@samp{BOOT_CFLAGS='-g -O1'}. 2565 2566@item @samp{bootstrap-O3} 2567Analogous to @code{bootstrap-O1}. 2568 2569@item @samp{bootstrap-lto} 2570Enables Link-Time Optimization for host tools during bootstrapping. 2571@samp{BUILD_CONFIG=bootstrap-lto} is equivalent to adding 2572@option{-flto} to @samp{BOOT_CFLAGS}. This option assumes that the host 2573supports the linker plugin (e.g. GNU ld version 2.21 or later or GNU gold 2574version 2.21 or later). 2575 2576@item @samp{bootstrap-lto-noplugin} 2577This option is similar to @code{bootstrap-lto}, but is intended for 2578hosts that do not support the linker plugin. Without the linker plugin 2579static libraries are not compiled with link-time optimizations. Since 2580the GCC middle end and back end are in @file{libbackend.a} this means 2581that only the front end is actually LTO optimized. 2582 2583@item @samp{bootstrap-debug} 2584Verifies that the compiler generates the same executable code, whether 2585or not it is asked to emit debug information. To this end, this 2586option builds stage2 host programs without debug information, and uses 2587@file{contrib/compare-debug} to compare them with the stripped stage3 2588object files. If @code{BOOT_CFLAGS} is overridden so as to not enable 2589debug information, stage2 will have it, and stage3 won't. This option 2590is enabled by default when GCC bootstrapping is enabled, if 2591@code{strip} can turn object files compiled with and without debug 2592info into identical object files. In addition to better test 2593coverage, this option makes default bootstraps faster and leaner. 2594 2595@item @samp{bootstrap-debug-big} 2596Rather than comparing stripped object files, as in 2597@code{bootstrap-debug}, this option saves internal compiler dumps 2598during stage2 and stage3 and compares them as well, which helps catch 2599additional potential problems, but at a great cost in terms of disk 2600space. It can be specified in addition to @samp{bootstrap-debug}. 2601 2602@item @samp{bootstrap-debug-lean} 2603This option saves disk space compared with @code{bootstrap-debug-big}, 2604but at the expense of some recompilation. Instead of saving the dumps 2605of stage2 and stage3 until the final compare, it uses 2606@option{-fcompare-debug} to generate, compare and remove the dumps 2607during stage3, repeating the compilation that already took place in 2608stage2, whose dumps were not saved. 2609 2610@item @samp{bootstrap-debug-lib} 2611This option tests executable code invariance over debug information 2612generation on target libraries, just like @code{bootstrap-debug-lean} 2613tests it on host programs. It builds stage3 libraries with 2614@option{-fcompare-debug}, and it can be used along with any of the 2615@code{bootstrap-debug} options above. 2616 2617There aren't @code{-lean} or @code{-big} counterparts to this option 2618because most libraries are only build in stage3, so bootstrap compares 2619would not get significant coverage. Moreover, the few libraries built 2620in stage2 are used in stage3 host programs, so we wouldn't want to 2621compile stage2 libraries with different options for comparison purposes. 2622 2623@item @samp{bootstrap-debug-ckovw} 2624Arranges for error messages to be issued if the compiler built on any 2625stage is run without the option @option{-fcompare-debug}. This is 2626useful to verify the full @option{-fcompare-debug} testing coverage. It 2627must be used along with @code{bootstrap-debug-lean} and 2628@code{bootstrap-debug-lib}. 2629 2630@item @samp{bootstrap-time} 2631Arranges for the run time of each program started by the GCC driver, 2632built in any stage, to be logged to @file{time.log}, in the top level of 2633the build tree. 2634 2635@end table 2636 2637@section Building a cross compiler 2638 2639When building a cross compiler, it is not generally possible to do a 26403-stage bootstrap of the compiler. This makes for an interesting problem 2641as parts of GCC can only be built with GCC@. 2642 2643To build a cross compiler, we recommend first building and installing a 2644native compiler. You can then use the native GCC compiler to build the 2645cross compiler. The installed native compiler needs to be GCC version 26462.95 or later. 2647 2648If the cross compiler is to be built with support for the Java 2649programming language and the ability to compile .java source files is 2650desired, the installed native compiler used to build the cross 2651compiler needs to be the same GCC version as the cross compiler. In 2652addition the cross compiler needs to be configured with 2653@option{--with-ecj-jar=@dots{}}. 2654 2655Assuming you have already installed a native copy of GCC and configured 2656your cross compiler, issue the command @command{make}, which performs the 2657following steps: 2658 2659@itemize @bullet 2660@item 2661Build host tools necessary to build the compiler. 2662 2663@item 2664Build target tools for use by the compiler such as binutils (bfd, 2665binutils, gas, gprof, ld, and opcodes) 2666if they have been individually linked or moved into the top level GCC source 2667tree before configuring. 2668 2669@item 2670Build the compiler (single stage only). 2671 2672@item 2673Build runtime libraries using the compiler from the previous step. 2674@end itemize 2675 2676Note that if an error occurs in any step the make process will exit. 2677 2678If you are not building GNU binutils in the same source tree as GCC, 2679you will need a cross-assembler and cross-linker installed before 2680configuring GCC@. Put them in the directory 2681@file{@var{prefix}/@var{target}/bin}. Here is a table of the tools 2682you should put in this directory: 2683 2684@table @file 2685@item as 2686This should be the cross-assembler. 2687 2688@item ld 2689This should be the cross-linker. 2690 2691@item ar 2692This should be the cross-archiver: a program which can manipulate 2693archive files (linker libraries) in the target machine's format. 2694 2695@item ranlib 2696This should be a program to construct a symbol table in an archive file. 2697@end table 2698 2699The installation of GCC will find these programs in that directory, 2700and copy or link them to the proper place to for the cross-compiler to 2701find them when run later. 2702 2703The easiest way to provide these files is to build the Binutils package. 2704Configure it with the same @option{--host} and @option{--target} 2705options that you use for configuring GCC, then build and install 2706them. They install their executables automatically into the proper 2707directory. Alas, they do not support all the targets that GCC 2708supports. 2709 2710If you are not building a C library in the same source tree as GCC, 2711you should also provide the target libraries and headers before 2712configuring GCC, specifying the directories with 2713@option{--with-sysroot} or @option{--with-headers} and 2714@option{--with-libs}. Many targets also require ``start files'' such 2715as @file{crt0.o} and 2716@file{crtn.o} which are linked into each executable. There may be several 2717alternatives for @file{crt0.o}, for use with profiling or other 2718compilation options. Check your target's definition of 2719@code{STARTFILE_SPEC} to find out what start files it uses. 2720 2721@section Building in parallel 2722 2723GNU Make 3.80 and above, which is necessary to build GCC, support 2724building in parallel. To activate this, you can use @samp{make -j 2} 2725instead of @samp{make}. You can also specify a bigger number, and 2726in most cases using a value greater than the number of processors in 2727your machine will result in fewer and shorter I/O latency hits, thus 2728improving overall throughput; this is especially true for slow drives 2729and network filesystems. 2730 2731@section Building the Ada compiler 2732 2733In order to build GNAT, the Ada compiler, you need a working GNAT 2734compiler (GCC version 4.0 or later). 2735This includes GNAT tools such as @command{gnatmake} and 2736@command{gnatlink}, since the Ada front end is written in Ada and 2737uses some GNAT-specific extensions. 2738 2739In order to build a cross compiler, it is suggested to install 2740the new compiler as native first, and then use it to build the cross 2741compiler. 2742 2743@command{configure} does not test whether the GNAT installation works 2744and has a sufficiently recent version; if too old a GNAT version is 2745installed, the build will fail unless @option{--enable-languages} is 2746used to disable building the Ada front end. 2747 2748@env{ADA_INCLUDE_PATH} and @env{ADA_OBJECT_PATH} environment variables 2749must not be set when building the Ada compiler, the Ada tools, or the 2750Ada runtime libraries. You can check that your build environment is clean 2751by verifying that @samp{gnatls -v} lists only one explicit path in each 2752section. 2753 2754@section Building with profile feedback 2755 2756It is possible to use profile feedback to optimize the compiler itself. This 2757should result in a faster compiler binary. Experiments done on x86 using gcc 27583.3 showed approximately 7 percent speedup on compiling C programs. To 2759bootstrap the compiler with profile feedback, use @code{make profiledbootstrap}. 2760 2761When @samp{make profiledbootstrap} is run, it will first build a @code{stage1} 2762compiler. This compiler is used to build a @code{stageprofile} compiler 2763instrumented to collect execution counts of instruction and branch 2764probabilities. Then runtime libraries are compiled with profile collected. 2765Finally a @code{stagefeedback} compiler is built using the information collected. 2766 2767Unlike standard bootstrap, several additional restrictions apply. The 2768compiler used to build @code{stage1} needs to support a 64-bit integral type. 2769It is recommended to only use GCC for this. 2770 2771@html 2772<hr /> 2773<p> 2774@end html 2775@ifhtml 2776@uref{./index.html,,Return to the GCC Installation page} 2777@end ifhtml 2778@end ifset 2779 2780@c ***Testing***************************************************************** 2781@ifnothtml 2782@comment node-name, next, previous, up 2783@node Testing, Final install, Building, Installing GCC 2784@end ifnothtml 2785@ifset testhtml 2786@ifnothtml 2787@chapter Installing GCC: Testing 2788@end ifnothtml 2789@cindex Testing 2790@cindex Installing GCC: Testing 2791@cindex Testsuite 2792 2793Before you install GCC, we encourage you to run the testsuites and to 2794compare your results with results from a similar configuration that have 2795been submitted to the 2796@uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}. 2797Some of these archived results are linked from the build status lists 2798at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who 2799reports a successful build runs the testsuites and submits the results. 2800This step is optional and may require you to download additional software, 2801but it can give you confidence in your new GCC installation or point out 2802problems before you install and start using your new GCC@. 2803 2804First, you must have @uref{download.html,,downloaded the testsuites}. 2805These are part of the full distribution, but if you downloaded the 2806``core'' compiler plus any front ends, you must download the testsuites 2807separately. 2808 2809Second, you must have the testing tools installed. This includes 2810@uref{http://www.gnu.org/software/dejagnu/,,DejaGnu}, Tcl, and Expect; 2811the DejaGnu site has links to these. 2812 2813If the directories where @command{runtest} and @command{expect} were 2814installed are not in the @env{PATH}, you may need to set the following 2815environment variables appropriately, as in the following example (which 2816assumes that DejaGnu has been installed under @file{/usr/local}): 2817 2818@smallexample 2819TCL_LIBRARY = /usr/local/share/tcl8.0 2820DEJAGNULIBS = /usr/local/share/dejagnu 2821@end smallexample 2822 2823(On systems such as Cygwin, these paths are required to be actual 2824paths, not mounts or links; presumably this is due to some lack of 2825portability in the DejaGnu code.) 2826 2827 2828Finally, you can run the testsuite (which may take a long time): 2829@smallexample 2830cd @var{objdir}; make -k check 2831@end smallexample 2832 2833This will test various components of GCC, such as compiler 2834front ends and runtime libraries. While running the testsuite, DejaGnu 2835might emit some harmless messages resembling 2836@samp{WARNING: Couldn't find the global config file.} or 2837@samp{WARNING: Couldn't find tool init file} that can be ignored. 2838 2839If you are testing a cross-compiler, you may want to run the testsuite 2840on a simulator as described at @uref{http://gcc.gnu.org/simtest-howto.html}. 2841 2842@section How can you run the testsuite on selected tests? 2843 2844In order to run sets of tests selectively, there are targets 2845@samp{make check-gcc} and language specific @samp{make check-c}, 2846@samp{make check-c++}, @samp{make check-fortran}, @samp{make check-java}, 2847@samp{make check-ada}, @samp{make check-objc}, @samp{make check-obj-c++}, 2848@samp{make check-lto} 2849in the @file{gcc} subdirectory of the object directory. You can also 2850just run @samp{make check} in a subdirectory of the object directory. 2851 2852 2853A more selective way to just run all @command{gcc} execute tests in the 2854testsuite is to use 2855 2856@smallexample 2857make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}" 2858@end smallexample 2859 2860Likewise, in order to run only the @command{g++} ``old-deja'' tests in 2861the testsuite with filenames matching @samp{9805*}, you would use 2862 2863@smallexample 2864make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}" 2865@end smallexample 2866 2867The @file{*.exp} files are located in the testsuite directories of the GCC 2868source, the most important ones being @file{compile.exp}, 2869@file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}. 2870To get a list of the possible @file{*.exp} files, pipe the 2871output of @samp{make check} into a file and look at the 2872@samp{Running @dots{} .exp} lines. 2873 2874@section Passing options and running multiple testsuites 2875 2876You can pass multiple options to the testsuite using the 2877@samp{--target_board} option of DejaGNU, either passed as part of 2878@samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to 2879work outside the makefiles. For example, 2880 2881@smallexample 2882make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fmerge-constants" 2883@end smallexample 2884 2885will run the standard @command{g++} testsuites (``unix'' is the target name 2886for a standard native testsuite situation), passing 2887@samp{-O3 -fmerge-constants} to the compiler on every test, i.e., 2888slashes separate options. 2889 2890You can run the testsuites multiple times using combinations of options 2891with a syntax similar to the brace expansion of popular shells: 2892 2893@smallexample 2894@dots{}"--target_board=arm-sim\@{-mhard-float,-msoft-float\@}\@{-O1,-O2,-O3,\@}" 2895@end smallexample 2896 2897(Note the empty option caused by the trailing comma in the final group.) 2898The following will run each testsuite eight times using the @samp{arm-sim} 2899target, as if you had specified all possible combinations yourself: 2900 2901@smallexample 2902--target_board='arm-sim/-mhard-float/-O1 \ 2903 arm-sim/-mhard-float/-O2 \ 2904 arm-sim/-mhard-float/-O3 \ 2905 arm-sim/-mhard-float \ 2906 arm-sim/-msoft-float/-O1 \ 2907 arm-sim/-msoft-float/-O2 \ 2908 arm-sim/-msoft-float/-O3 \ 2909 arm-sim/-msoft-float' 2910@end smallexample 2911 2912They can be combined as many times as you wish, in arbitrary ways. This 2913list: 2914 2915@smallexample 2916@dots{}"--target_board=unix/-Wextra\@{-O3,-fno-strength\@}\@{-fomit-frame,\@}" 2917@end smallexample 2918 2919will generate four combinations, all involving @samp{-Wextra}. 2920 2921The disadvantage to this method is that the testsuites are run in serial, 2922which is a waste on multiprocessor systems. For users with GNU Make and 2923a shell which performs brace expansion, you can run the testsuites in 2924parallel by having the shell perform the combinations and @command{make} 2925do the parallel runs. Instead of using @samp{--target_board}, use a 2926special makefile target: 2927 2928@smallexample 2929make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{} 2930@end smallexample 2931 2932For example, 2933 2934@smallexample 2935make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@} 2936@end smallexample 2937 2938will run three concurrent ``make-gcc'' testsuites, eventually testing all 2939ten combinations as described above. Note that this is currently only 2940supported in the @file{gcc} subdirectory. (To see how this works, try 2941typing @command{echo} before the example given here.) 2942 2943 2944@section Additional testing for Java Class Libraries 2945 2946The Java runtime tests can be executed via @samp{make check} 2947in the @file{@var{target}/libjava/testsuite} directory in 2948the build tree. 2949 2950The @uref{http://sourceware.org/mauve/,,Mauve Project} provides 2951a suite of tests for the Java Class Libraries. This suite can be run 2952as part of libgcj testing by placing the Mauve tree within the libjava 2953testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by 2954specifying the location of that tree when invoking @samp{make}, as in 2955@samp{make MAUVEDIR=~/mauve check}. 2956 2957@section How to interpret test results 2958 2959The result of running the testsuite are various @file{*.sum} and @file{*.log} 2960files in the testsuite subdirectories. The @file{*.log} files contain a 2961detailed log of the compiler invocations and the corresponding 2962results, the @file{*.sum} files summarize the results. These summaries 2963contain status codes for all tests: 2964 2965@itemize @bullet 2966@item 2967PASS: the test passed as expected 2968@item 2969XPASS: the test unexpectedly passed 2970@item 2971FAIL: the test unexpectedly failed 2972@item 2973XFAIL: the test failed as expected 2974@item 2975UNSUPPORTED: the test is not supported on this platform 2976@item 2977ERROR: the testsuite detected an error 2978@item 2979WARNING: the testsuite detected a possible problem 2980@end itemize 2981 2982It is normal for some tests to report unexpected failures. At the 2983current time the testing harness does not allow fine grained control 2984over whether or not a test is expected to fail. This problem should 2985be fixed in future releases. 2986 2987 2988@section Submitting test results 2989 2990If you want to report the results to the GCC project, use the 2991@file{contrib/test_summary} shell script. Start it in the @var{objdir} with 2992 2993@smallexample 2994@var{srcdir}/contrib/test_summary -p your_commentary.txt \ 2995 -m gcc-testresults@@gcc.gnu.org |sh 2996@end smallexample 2997 2998This script uses the @command{Mail} program to send the results, so 2999make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is 3000prepended to the testsuite summary and should contain any special 3001remarks you have on your results or your build environment. Please 3002do not edit the testsuite result block or the subject line, as these 3003messages may be automatically processed. 3004 3005@html 3006<hr /> 3007<p> 3008@end html 3009@ifhtml 3010@uref{./index.html,,Return to the GCC Installation page} 3011@end ifhtml 3012@end ifset 3013 3014@c ***Final install*********************************************************** 3015@ifnothtml 3016@comment node-name, next, previous, up 3017@node Final install, , Testing, Installing GCC 3018@end ifnothtml 3019@ifset finalinstallhtml 3020@ifnothtml 3021@chapter Installing GCC: Final installation 3022@end ifnothtml 3023 3024Now that GCC has been built (and optionally tested), you can install it with 3025@smallexample 3026cd @var{objdir} && make install 3027@end smallexample 3028 3029We strongly recommend to install into a target directory where there is 3030no previous version of GCC present. Also, the GNAT runtime should not 3031be stripped, as this would break certain features of the debugger that 3032depend on this debugging information (catching Ada exceptions for 3033instance). 3034 3035That step completes the installation of GCC; user level binaries can 3036be found in @file{@var{prefix}/bin} where @var{prefix} is the value 3037you specified with the @option{--prefix} to configure (or 3038@file{/usr/local} by default). (If you specified @option{--bindir}, 3039that directory will be used instead; otherwise, if you specified 3040@option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.) 3041Headers for the C++ and Java libraries are installed in 3042@file{@var{prefix}/include}; libraries in @file{@var{libdir}} 3043(normally @file{@var{prefix}/lib}); internal parts of the compiler in 3044@file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation 3045in info format in @file{@var{infodir}} (normally 3046@file{@var{prefix}/info}). 3047 3048When installing cross-compilers, GCC's executables 3049are not only installed into @file{@var{bindir}}, that 3050is, @file{@var{exec-prefix}/bin}, but additionally into 3051@file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory 3052exists. Typically, such @dfn{tooldirs} hold target-specific 3053binutils, including assembler and linker. 3054 3055Installation into a temporary staging area or into a @command{chroot} 3056jail can be achieved with the command 3057 3058@smallexample 3059make DESTDIR=@var{path-to-rootdir} install 3060@end smallexample 3061 3062@noindent 3063where @var{path-to-rootdir} is the absolute path of 3064a directory relative to which all installation paths will be 3065interpreted. Note that the directory specified by @code{DESTDIR} 3066need not exist yet; it will be created if necessary. 3067 3068There is a subtle point with tooldirs and @code{DESTDIR}: 3069If you relocate a cross-compiler installation with 3070e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory 3071@file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will 3072be filled with duplicated GCC executables only if it already exists, 3073it will not be created otherwise. This is regarded as a feature, 3074not as a bug, because it gives slightly more control to the packagers 3075using the @code{DESTDIR} feature. 3076 3077You can install stripped programs and libraries with 3078 3079@smallexample 3080make install-strip 3081@end smallexample 3082 3083If you are bootstrapping a released version of GCC then please 3084quickly review the build status page for your release, available from 3085@uref{http://gcc.gnu.org/buildstat.html}. 3086If your system is not listed for the version of GCC that you built, 3087send a note to 3088@email{gcc@@gcc.gnu.org} indicating 3089that you successfully built and installed GCC@. 3090Include the following information: 3091 3092@itemize @bullet 3093@item 3094Output from running @file{@var{srcdir}/config.guess}. Do not send 3095that file itself, just the one-line output from running it. 3096 3097@item 3098The output of @samp{gcc -v} for your newly installed @command{gcc}. 3099This tells us which version of GCC you built and the options you passed to 3100configure. 3101 3102@item 3103Whether you enabled all languages or a subset of them. If you used a 3104full distribution then this information is part of the configure 3105options in the output of @samp{gcc -v}, but if you downloaded the 3106``core'' compiler plus additional front ends then it isn't apparent 3107which ones you built unless you tell us about it. 3108 3109@item 3110If the build was for GNU/Linux, also include: 3111@itemize @bullet 3112@item 3113The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3); 3114this information should be available from @file{/etc/issue}. 3115 3116@item 3117The version of the Linux kernel, available from @samp{uname --version} 3118or @samp{uname -a}. 3119 3120@item 3121The version of glibc you used; for RPM-based systems like Red Hat, 3122Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version, 3123and on systems like Debian and Progeny use @samp{dpkg -l libc6}. 3124@end itemize 3125For other systems, you can include similar information if you think it is 3126relevant. 3127 3128@item 3129Any other information that you think would be useful to people building 3130GCC on the same configuration. The new entry in the build status list 3131will include a link to the archived copy of your message. 3132@end itemize 3133 3134We'd also like to know if the 3135@ifnothtml 3136@ref{Specific, host/target specific installation notes} 3137@end ifnothtml 3138@ifhtml 3139@uref{specific.html,,host/target specific installation notes} 3140@end ifhtml 3141didn't include your host/target information or if that information is 3142incomplete or out of date. Send a note to 3143@email{gcc@@gcc.gnu.org} detailing how the information should be changed. 3144 3145If you find a bug, please report it following the 3146@uref{../bugs/,,bug reporting guidelines}. 3147 3148If you want to print the GCC manuals, do @samp{cd @var{objdir}; make 3149dvi}. You will need to have @command{texi2dvi} (version at least 4.7) 3150and @TeX{} installed. This creates a number of @file{.dvi} files in 3151subdirectories of @file{@var{objdir}}; these may be converted for 3152printing with programs such as @command{dvips}. Alternately, by using 3153@samp{make pdf} in place of @samp{make dvi}, you can create documentation 3154in the form of @file{.pdf} files; this requires @command{texi2pdf}, which 3155is included with Texinfo version 4.8 and later. You can also 3156@uref{http://shop.fsf.org/,,buy printed manuals from the 3157Free Software Foundation}, though such manuals may not be for the most 3158recent version of GCC@. 3159 3160If you would like to generate online HTML documentation, do @samp{cd 3161@var{objdir}; make html} and HTML will be generated for the gcc manuals in 3162@file{@var{objdir}/gcc/HTML}. 3163 3164@html 3165<hr /> 3166<p> 3167@end html 3168@ifhtml 3169@uref{./index.html,,Return to the GCC Installation page} 3170@end ifhtml 3171@end ifset 3172 3173@c ***Binaries**************************************************************** 3174@ifnothtml 3175@comment node-name, next, previous, up 3176@node Binaries, Specific, Installing GCC, Top 3177@end ifnothtml 3178@ifset binarieshtml 3179@ifnothtml 3180@chapter Installing GCC: Binaries 3181@end ifnothtml 3182@cindex Binaries 3183@cindex Installing GCC: Binaries 3184 3185We are often asked about pre-compiled versions of GCC@. While we cannot 3186provide these for all platforms, below you'll find links to binaries for 3187various platforms where creating them by yourself is not easy due to various 3188reasons. 3189 3190Please note that we did not create these binaries, nor do we 3191support them. If you have any problems installing them, please 3192contact their makers. 3193 3194@itemize 3195@item 3196AIX: 3197@itemize 3198@item 3199@uref{http://www.bullfreeware.com,,Bull's Open Source Software Archive for 3200for AIX 5L and AIX 6}; 3201 3202@item 3203@uref{http://www.perzl.org/aix/,,AIX Open Source Packages (AIX5L AIX 6.1 3204AIX 7.1)}. 3205@end itemize 3206 3207@item 3208DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}. 3209 3210@item 3211HP-UX: 3212@itemize 3213@item 3214@uref{http://hpux.connect.org.uk/,,HP-UX Porting Center}; 3215@end itemize 3216 3217@item 3218Solaris 2 (SPARC, Intel): 3219@itemize 3220@item 3221@uref{https://www.opencsw.org/,,OpenCSW} 3222 3223@item 3224@uref{http://jupiterrise.com/tgcware/,,TGCware} 3225@end itemize 3226 3227@item 3228Microsoft Windows: 3229@itemize 3230@item 3231The @uref{http://sourceware.org/cygwin/,,Cygwin} project; 3232@item 3233The @uref{http://www.mingw.org/,,MinGW} and 3234@uref{http://mingw-w64.org/,,mingw-w64} projects. 3235@end itemize 3236 3237@item 3238@uref{http://www.openpkg.org/,,OpenPKG} offers binaries for quite a 3239number of platforms. 3240 3241@item 3242The @uref{http://gcc.gnu.org/wiki/GFortranBinaries,,GFortran Wiki} has 3243links to GNU Fortran binaries for several platforms. 3244@end itemize 3245 3246@html 3247<hr /> 3248<p> 3249@end html 3250@ifhtml 3251@uref{./index.html,,Return to the GCC Installation page} 3252@end ifhtml 3253@end ifset 3254 3255@c ***Specific**************************************************************** 3256@ifnothtml 3257@comment node-name, next, previous, up 3258@node Specific, Old, Binaries, Top 3259@end ifnothtml 3260@ifset specifichtml 3261@ifnothtml 3262@chapter Host/target specific installation notes for GCC 3263@end ifnothtml 3264@cindex Specific 3265@cindex Specific installation notes 3266@cindex Target specific installation 3267@cindex Host specific installation 3268@cindex Target specific installation notes 3269 3270Please read this document carefully @emph{before} installing the 3271GNU Compiler Collection on your machine. 3272 3273Note that this list of install notes is @emph{not} a list of supported 3274hosts or targets. Not all supported hosts and targets are listed 3275here, only the ones that require host-specific or target-specific 3276information have to. 3277 3278@ifhtml 3279@itemize 3280@item 3281@uref{#aarch64-x-x,,aarch64*-*-*} 3282@item 3283@uref{#alpha-x-x,,alpha*-*-*} 3284@item 3285@uref{#alpha-dec-osf51,,alpha*-dec-osf5.1} 3286@item 3287@uref{#amd64-x-solaris210,,amd64-*-solaris2.10} 3288@item 3289@uref{#arm-x-eabi,,arm-*-eabi} 3290@item 3291@uref{#avr,,avr} 3292@item 3293@uref{#bfin,,Blackfin} 3294@item 3295@uref{#dos,,DOS} 3296@item 3297@uref{#x-x-freebsd,,*-*-freebsd*} 3298@item 3299@uref{#h8300-hms,,h8300-hms} 3300@item 3301@uref{#hppa-hp-hpux,,hppa*-hp-hpux*} 3302@item 3303@uref{#hppa-hp-hpux10,,hppa*-hp-hpux10} 3304@item 3305@uref{#hppa-hp-hpux11,,hppa*-hp-hpux11} 3306@item 3307@uref{#x-x-linux-gnu,,*-*-linux-gnu} 3308@item 3309@uref{#ix86-x-linux,,i?86-*-linux*} 3310@item 3311@uref{#ix86-x-solaris210,,i?86-*-solaris2.10} 3312@item 3313@uref{#ia64-x-linux,,ia64-*-linux} 3314@item 3315@uref{#ia64-x-hpux,,ia64-*-hpux*} 3316@item 3317@uref{#x-ibm-aix,,*-ibm-aix*} 3318@item 3319@uref{#iq2000-x-elf,,iq2000-*-elf} 3320@item 3321@uref{#lm32-x-elf,,lm32-*-elf} 3322@item 3323@uref{#lm32-x-uclinux,,lm32-*-uclinux} 3324@item 3325@uref{#m32c-x-elf,,m32c-*-elf} 3326@item 3327@uref{#m32r-x-elf,,m32r-*-elf} 3328@item 3329@uref{#m68k-x-x,,m68k-*-*} 3330@item 3331@uref{#m68k-uclinux,,m68k-uclinux} 3332@item 3333@uref{#mep-x-elf,,mep-*-elf} 3334@item 3335@uref{#microblaze-x-elf,,microblaze-*-elf} 3336@item 3337@uref{#mips-x-x,,mips-*-*} 3338@item 3339@uref{#mips-sgi-irix5,,mips-sgi-irix5} 3340@item 3341@uref{#mips-sgi-irix6,,mips-sgi-irix6} 3342@item 3343@uref{#nds32le-x-elf,,nds32le-*-elf} 3344@item 3345@uref{#nds32be-x-elf,,nds32be-*-elf} 3346@item 3347@uref{#nvptx-x-none,,nvptx-*-none} 3348@item 3349@uref{#powerpc-x-x,,powerpc*-*-*} 3350@item 3351@uref{#powerpc-x-darwin,,powerpc-*-darwin*} 3352@item 3353@uref{#powerpc-x-elf,,powerpc-*-elf} 3354@item 3355@uref{#powerpc-x-linux-gnu,,powerpc*-*-linux-gnu*} 3356@item 3357@uref{#powerpc-x-netbsd,,powerpc-*-netbsd*} 3358@item 3359@uref{#powerpc-x-eabisim,,powerpc-*-eabisim} 3360@item 3361@uref{#powerpc-x-eabi,,powerpc-*-eabi} 3362@item 3363@uref{#powerpcle-x-elf,,powerpcle-*-elf} 3364@item 3365@uref{#powerpcle-x-eabisim,,powerpcle-*-eabisim} 3366@item 3367@uref{#powerpcle-x-eabi,,powerpcle-*-eabi} 3368@item 3369@uref{#s390-x-linux,,s390-*-linux*} 3370@item 3371@uref{#s390x-x-linux,,s390x-*-linux*} 3372@item 3373@uref{#s390x-ibm-tpf,,s390x-ibm-tpf*} 3374@item 3375@uref{#x-x-solaris2,,*-*-solaris2*} 3376@item 3377@uref{#sparc-x-x,,sparc*-*-*} 3378@item 3379@uref{#sparc-sun-solaris2,,sparc-sun-solaris2*} 3380@item 3381@uref{#sparc-sun-solaris210,,sparc-sun-solaris2.10} 3382@item 3383@uref{#sparc-x-linux,,sparc-*-linux*} 3384@item 3385@uref{#sparc64-x-solaris2,,sparc64-*-solaris2*} 3386@item 3387@uref{#sparcv9-x-solaris2,,sparcv9-*-solaris2*} 3388@item 3389@uref{#c6x-x-x,,c6x-*-*} 3390@item 3391@uref{#tilegx-x-linux,,tilegx-*-linux*} 3392@item 3393@uref{#tilegxbe-x-linux,,tilegxbe-*-linux*} 3394@item 3395@uref{#tilepro-x-linux,,tilepro-*-linux*} 3396@item 3397@uref{#visium-x-elf, visium-*-elf} 3398@item 3399@uref{#x-x-vxworks,,*-*-vxworks*} 3400@item 3401@uref{#x86-64-x-x,,x86_64-*-*, amd64-*-*} 3402@item 3403@uref{#x86-64-x-solaris210,,x86_64-*-solaris2.1[0-9]*} 3404@item 3405@uref{#xtensa-x-elf,,xtensa*-*-elf} 3406@item 3407@uref{#xtensa-x-linux,,xtensa*-*-linux*} 3408@item 3409@uref{#windows,,Microsoft Windows} 3410@item 3411@uref{#x-x-cygwin,,*-*-cygwin} 3412@item 3413@uref{#x-x-interix,,*-*-interix} 3414@item 3415@uref{#x-x-mingw32,,*-*-mingw32} 3416@item 3417@uref{#os2,,OS/2} 3418@item 3419@uref{#older,,Older systems} 3420@end itemize 3421 3422@itemize 3423@item 3424@uref{#elf,,all ELF targets} (SVR4, Solaris 2, etc.) 3425@end itemize 3426@end ifhtml 3427 3428 3429@html 3430<!-- -------- host/target specific issues start here ---------------- --> 3431<hr /> 3432@end html 3433@anchor{aarch64-x-x} 3434@heading aarch64*-*-* 3435Binutils pre 2.24 does not have support for selecting @option{-mabi} and 3436does not support ILP32. If it is used to build GCC 4.9 or later, GCC will 3437not support option @option{-mabi=ilp32}. 3438 3439To enable a workaround for the Cortex-A53 erratum number 835769 by default 3440(for all CPUs regardless of -mcpu option given) at configure time use the 3441@option{--enable-fix-cortex-a53-835769} option. This will enable the fix by 3442default and can be explicitly disabled during compilation by passing the 3443@option{-mno-fix-cortex-a53-835769} option. Conversely, 3444@option{--disable-fix-cortex-a53-835769} will disable the workaround by 3445default. The workaround is disabled by default if neither of 3446@option{--enable-fix-cortex-a53-835769} or 3447@option{--disable-fix-cortex-a53-835769} is given at configure time. 3448 3449To enable a workaround for the Cortex-A53 erratum number 843419 by default 3450(for all CPUs regardless of -mcpu option given) at configure time use the 3451@option{--enable-fix-cortex-a53-843419} option. This workaround is applied at 3452link time. Enabling the workaround will cause GCC to pass the relevant option 3453to the linker. It can be explicitly disabled during compilation by passing the 3454@option{-mno-fix-cortex-a53-843419} option. Conversely, 3455@option{--disable-fix-cortex-a53-843419} will disable the workaround by default. 3456The workaround is disabled by default if neither of 3457@option{--enable-fix-cortex-a53-843419} or 3458@option{--disable-fix-cortex-a53-843419} is given at configure time. 3459 3460@html 3461<hr /> 3462@end html 3463@anchor{alpha-x-x} 3464@heading alpha*-*-* 3465This section contains general configuration information for all 3466alpha-based platforms using ELF (in particular, ignore this section for 3467DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this 3468section, please read all other sections that match your target. 3469 3470We require binutils 2.11.2 or newer. 3471Previous binutils releases had a number of problems with DWARF 2 3472debugging information, not the least of which is incorrect linking of 3473shared libraries. 3474 3475@html 3476<hr /> 3477@end html 3478@anchor{alpha-dec-osf51} 3479@heading alpha*-dec-osf5.1 3480Systems using processors that implement the DEC Alpha architecture and 3481are running the DEC/Compaq/HP Unix (DEC OSF/1, Digital UNIX, or Compaq/HP 3482Tru64 UNIX) operating system, for example the DEC Alpha AXP systems. 3483 3484Support for Tru64 UNIX V5.1 has been removed in GCC 4.8. As of GCC 4.6, 3485support for Tru64 UNIX V4.0 and V5.0 has been removed. As of GCC 3.2, 3486versions before @code{alpha*-dec-osf4} are no longer supported. (These 3487are the versions which identify themselves as DEC OSF/1.) 3488 3489@html 3490<hr /> 3491@end html 3492@anchor{amd64-x-solaris210} 3493@heading amd64-*-solaris2.1[0-9]* 3494This is a synonym for @samp{x86_64-*-solaris2.1[0-9]*}. 3495 3496@html 3497<hr /> 3498@end html 3499@anchor{arc-x-elf32} 3500@heading arc-*-elf32 3501 3502Use @samp{configure --target=arc-elf32 --with-cpu=@var{cpu} --enable-languages="c,c++"} 3503to configure GCC, with @var{cpu} being one of @samp{arc600}, @samp{arc601}, 3504or @samp{arc700}@. 3505 3506@html 3507<hr /> 3508@end html 3509@anchor{arc-linux-uclibc} 3510@heading arc-linux-uclibc 3511 3512Use @samp{configure --target=arc-linux-uclibc --with-cpu=arc700 --enable-languages="c,c++"} to configure GCC@. 3513 3514@html 3515<hr /> 3516@end html 3517@anchor{arm-x-eabi} 3518@heading arm-*-eabi 3519ARM-family processors. Subtargets that use the ELF object format 3520require GNU binutils 2.13 or newer. Such subtargets include: 3521@code{arm-*-netbsdelf}, @code{arm-*-*linux-*} 3522and @code{arm-*-rtemseabi}. 3523 3524Building the Ada frontend commonly fails (an infinite loop executing 3525@code{xsinfo}) if the host compiler is GNAT 4.8. Host compilers built from the 3526GNAT 4.6, 4.9 or 5 release branches are known to succeed. 3527 3528@html 3529<hr /> 3530@end html 3531@anchor{avr} 3532@heading avr 3533ATMEL AVR-family micro controllers. These are used in embedded 3534applications. There are no standard Unix configurations. 3535@ifnothtml 3536@xref{AVR Options,, AVR Options, gcc, Using the GNU Compiler 3537Collection (GCC)}, 3538@end ifnothtml 3539@ifhtml 3540See ``AVR Options'' in the main manual 3541@end ifhtml 3542for the list of supported MCU types. 3543 3544Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@. 3545 3546Further installation notes and other useful information about AVR tools 3547can also be obtained from: 3548 3549@itemize @bullet 3550@item 3551@uref{http://www.nongnu.org/avr/,,http://www.nongnu.org/avr/} 3552@item 3553@uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/} 3554@end itemize 3555 3556We @emph{strongly} recommend using binutils 2.13 or newer. 3557 3558The following error: 3559@smallexample 3560Error: register required 3561@end smallexample 3562 3563indicates that you should upgrade to a newer version of the binutils. 3564 3565@html 3566<hr /> 3567@end html 3568@anchor{bfin} 3569@heading Blackfin 3570The Blackfin processor, an Analog Devices DSP. 3571@ifnothtml 3572@xref{Blackfin Options,, Blackfin Options, gcc, Using the GNU Compiler 3573Collection (GCC)}, 3574@end ifnothtml 3575@ifhtml 3576See ``Blackfin Options'' in the main manual 3577@end ifhtml 3578 3579More information, and a version of binutils with support for this processor, 3580is available at @uref{http://blackfin.uclinux.org} 3581 3582@html 3583<hr /> 3584@end html 3585@anchor{cr16} 3586@heading CR16 3587The CR16 CompactRISC architecture is a 16-bit architecture. This 3588architecture is used in embedded applications. 3589 3590@ifnothtml 3591@xref{CR16 Options,, CR16 Options, gcc, Using and Porting the GNU Compiler 3592Collection (GCC)}, 3593@end ifnothtml 3594 3595@ifhtml 3596See ``CR16 Options'' in the main manual for a list of CR16-specific options. 3597@end ifhtml 3598 3599Use @samp{configure --target=cr16-elf --enable-languages=c,c++} to configure 3600GCC@ for building a CR16 elf cross-compiler. 3601 3602Use @samp{configure --target=cr16-uclinux --enable-languages=c,c++} to 3603configure GCC@ for building a CR16 uclinux cross-compiler. 3604 3605@html 3606<hr /> 3607@end html 3608@anchor{cris} 3609@heading CRIS 3610CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip 3611series. These are used in embedded applications. 3612 3613@ifnothtml 3614@xref{CRIS Options,, CRIS Options, gcc, Using the GNU Compiler 3615Collection (GCC)}, 3616@end ifnothtml 3617@ifhtml 3618See ``CRIS Options'' in the main manual 3619@end ifhtml 3620for a list of CRIS-specific options. 3621 3622There are a few different CRIS targets: 3623@table @code 3624@item cris-axis-elf 3625Mainly for monolithic embedded systems. Includes a multilib for the 3626@samp{v10} core used in @samp{ETRAX 100 LX}. 3627@item cris-axis-linux-gnu 3628A GNU/Linux port for the CRIS architecture, currently targeting 3629@samp{ETRAX 100 LX} by default. 3630@end table 3631 3632For @code{cris-axis-elf} you need binutils 2.11 3633or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer. 3634 3635Pre-packaged tools can be obtained from 3636@uref{ftp://ftp.axis.com/@/pub/@/axis/@/tools/@/cris/@/compiler-kit/}. More 3637information about this platform is available at 3638@uref{http://developer.axis.com/}. 3639 3640@html 3641<hr /> 3642@end html 3643@anchor{dos} 3644@heading DOS 3645Please have a look at the @uref{binaries.html,,binaries page}. 3646 3647You cannot install GCC by itself on MSDOS; it will not compile under 3648any MSDOS compiler except itself. You need to get the complete 3649compilation package DJGPP, which includes binaries as well as sources, 3650and includes all the necessary compilation tools and libraries. 3651 3652@html 3653<hr /> 3654@end html 3655@anchor{epiphany-x-elf} 3656@heading epiphany-*-elf 3657Adapteva Epiphany. 3658This configuration is intended for embedded systems. 3659 3660@html 3661<hr /> 3662@end html 3663@anchor{x-x-freebsd} 3664@heading *-*-freebsd* 3665Support for FreeBSD 1 was discontinued in GCC 3.2. Support for 3666FreeBSD 2 (and any mutant a.out variants of FreeBSD 3) was 3667discontinued in GCC 4.0. 3668 3669In order to better utilize FreeBSD base system functionality and match 3670the configuration of the system compiler, GCC 4.5 and above as well as 3671GCC 4.4 past 2010-06-20 leverage SSP support in libc (which is present 3672on FreeBSD 7 or later) and the use of @code{__cxa_atexit} by default 3673(on FreeBSD 6 or later). The use of @code{dl_iterate_phdr} inside 3674@file{libgcc_s.so.1} and boehm-gc (on FreeBSD 7 or later) is enabled 3675by GCC 4.5 and above. 3676 3677We support FreeBSD using the ELF file format with DWARF 2 debugging 3678for all CPU architectures. You may use @option{-gstabs} instead of 3679@option{-g}, if you really want the old debugging format. There are 3680no known issues with mixing object files and libraries with different 3681debugging formats. Otherwise, this release of GCC should now match 3682more of the configuration used in the stock FreeBSD configuration of 3683GCC@. In particular, @option{--enable-threads} is now configured by 3684default. However, as a general user, do not attempt to replace the 3685system compiler with this release. Known to bootstrap and check with 3686good results on FreeBSD 7.2-STABLE@. In the past, known to bootstrap 3687and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4, 36884.5, 4.8, 4.9 and 5-CURRENT@. 3689 3690The version of binutils installed in @file{/usr/bin} probably works 3691with this release of GCC@. Bootstrapping against the latest GNU 3692binutils and/or the version found in @file{/usr/ports/devel/binutils} has 3693been known to enable additional features and improve overall testsuite 3694results. However, it is currently known that boehm-gc (which itself 3695is required for java) may not configure properly on FreeBSD prior to 3696the FreeBSD 7.0 release with GNU binutils after 2.16.1. 3697 3698@html 3699<hr /> 3700@end html 3701@anchor{ft32-x-elf} 3702@heading ft32-*-elf 3703The FT32 processor. 3704This configuration is intended for embedded systems. 3705 3706@html 3707<hr /> 3708@end html 3709@anchor{h8300-hms} 3710@heading h8300-hms 3711Renesas H8/300 series of processors. 3712 3713Please have a look at the @uref{binaries.html,,binaries page}. 3714 3715The calling convention and structure layout has changed in release 2.6. 3716All code must be recompiled. The calling convention now passes the 3717first three arguments in function calls in registers. Structures are no 3718longer a multiple of 2 bytes. 3719 3720@html 3721<hr /> 3722@end html 3723@anchor{hppa-hp-hpux} 3724@heading hppa*-hp-hpux* 3725Support for HP-UX version 9 and older was discontinued in GCC 3.4. 3726 3727We require using gas/binutils on all hppa platforms. Version 2.19 or 3728later is recommended. 3729 3730It may be helpful to configure GCC with the 3731@uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and 3732@option{--with-as=@dots{}} options to ensure that GCC can find GAS@. 3733 3734The HP assembler should not be used with GCC. It is rarely tested and may 3735not work. It shouldn't be used with any languages other than C due to its 3736many limitations. 3737 3738Specifically, @option{-g} does not work (HP-UX uses a peculiar debugging 3739format which GCC does not know about). It also inserts timestamps 3740into each object file it creates, causing the 3-stage comparison test to 3741fail during a bootstrap. You should be able to continue by saying 3742@samp{make all-host all-target} after getting the failure from @samp{make}. 3743 3744Various GCC features are not supported. For example, it does not support weak 3745symbols or alias definitions. As a result, explicit template instantiations 3746are required when using C++. This makes it difficult if not impossible to 3747build many C++ applications. 3748 3749There are two default scheduling models for instructions. These are 3750PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc 3751architecture specified for the target machine when configuring. 3752PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when 3753the target is a @samp{hppa1*} machine. 3754 3755The PROCESSOR_8000 model is not well suited to older processors. Thus, 3756it is important to completely specify the machine architecture when 3757configuring if you want a model other than PROCESSOR_8000. The macro 3758TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different 3759default scheduling model is desired. 3760 3761As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10 3762through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later. 3763This namespace change might cause problems when bootstrapping with 3764an earlier version of GCC or the HP compiler as essentially the same 3765namespace is required for an entire build. This problem can be avoided 3766in a number of ways. With HP cc, @env{UNIX_STD} can be set to @samp{95} 3767or @samp{98}. Another way is to add an appropriate set of predefines 3768to @env{CC}. The description for the @option{munix=} option contains 3769a list of the predefines used with each standard. 3770 3771More specific information to @samp{hppa*-hp-hpux*} targets follows. 3772 3773@html 3774<hr /> 3775@end html 3776@anchor{hppa-hp-hpux10} 3777@heading hppa*-hp-hpux10 3778For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch 3779@code{PHCO_19798} from HP@. 3780 3781The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are 3782used for one-only code and data. This resolves many of the previous 3783problems in using C++ on this target. However, the ABI is not compatible 3784with the one implemented under HP-UX 11 using secondary definitions. 3785 3786@html 3787<hr /> 3788@end html 3789@anchor{hppa-hp-hpux11} 3790@heading hppa*-hp-hpux11 3791GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot 3792be used to compile GCC 3.0 and up. 3793 3794The libffi and libjava libraries haven't been ported to 64-bit HP-UX@ 3795and don't build. 3796 3797Refer to @uref{binaries.html,,binaries} for information about obtaining 3798precompiled GCC binaries for HP-UX@. Precompiled binaries must be obtained 3799to build the Ada language as it can't be bootstrapped using C@. Ada is 3800only available for the 32-bit PA-RISC runtime. 3801 3802Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The 3803bundled compiler supports only traditional C; you will need either HP's 3804unbundled compiler, or a binary distribution of GCC@. 3805 3806It is possible to build GCC 3.3 starting with the bundled HP compiler, 3807but the process requires several steps. GCC 3.3 can then be used to 3808build later versions. The fastjar program contains ISO C code and 3809can't be built with the HP bundled compiler. This problem can be 3810avoided by not building the Java language. For example, use the 3811@option{--enable-languages="c,c++,f77,objc"} option in your configure 3812command. 3813 3814There are several possible approaches to building the distribution. 3815Binutils can be built first using the HP tools. Then, the GCC 3816distribution can be built. The second approach is to build GCC 3817first using the HP tools, then build binutils, then rebuild GCC@. 3818There have been problems with various binary distributions, so it 3819is best not to start from a binary distribution. 3820 3821On 64-bit capable systems, there are two distinct targets. Different 3822installation prefixes must be used if both are to be installed on 3823the same system. The @samp{hppa[1-2]*-hp-hpux11*} target generates code 3824for the 32-bit PA-RISC runtime architecture and uses the HP linker. 3825The @samp{hppa64-hp-hpux11*} target generates 64-bit code for the 3826PA-RISC 2.0 architecture. 3827 3828The script config.guess now selects the target type based on the compiler 3829detected during configuration. You must define @env{PATH} or @env{CC} so 3830that configure finds an appropriate compiler for the initial bootstrap. 3831When @env{CC} is used, the definition should contain the options that are 3832needed whenever @env{CC} is used. 3833 3834Specifically, options that determine the runtime architecture must be 3835in @env{CC} to correctly select the target for the build. It is also 3836convenient to place many other compiler options in @env{CC}. For example, 3837@env{CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"} 3838can be used to bootstrap the GCC 3.3 branch with the HP compiler in 383964-bit K&R/bundled mode. The @option{+DA2.0W} option will result in 3840the automatic selection of the @samp{hppa64-hp-hpux11*} target. The 3841macro definition table of cpp needs to be increased for a successful 3842build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to 3843be defined when building with the bundled compiler, or when using the 3844@option{-Ac} option. These defines aren't necessary with @option{-Ae}. 3845 3846It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target 3847with the @option{--with-ld=@dots{}} option. This overrides the standard 3848search for ld. The two linkers supported on this target require different 3849commands. The default linker is determined during configuration. As a 3850result, it's not possible to switch linkers in the middle of a GCC build. 3851This has been reported to sometimes occur in unified builds of binutils 3852and GCC@. 3853 3854A recent linker patch must be installed for the correct operation of 3855GCC 3.3 and later. @code{PHSS_26559} and @code{PHSS_24304} are the 3856oldest linker patches that are known to work. They are for HP-UX 385711.00 and 11.11, respectively. @code{PHSS_24303}, the companion to 3858@code{PHSS_24304}, might be usable but it hasn't been tested. These 3859patches have been superseded. Consult the HP patch database to obtain 3860the currently recommended linker patch for your system. 3861 3862The patches are necessary for the support of weak symbols on the 386332-bit port, and for the running of initializers and finalizers. Weak 3864symbols are implemented using SOM secondary definition symbols. Prior 3865to HP-UX 11, there are bugs in the linker support for secondary symbols. 3866The patches correct a problem of linker core dumps creating shared 3867libraries containing secondary symbols, as well as various other 3868linking issues involving secondary symbols. 3869 3870GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to 3871run initializers and finalizers on the 64-bit port. The 32-bit port 3872uses the linker @option{+init} and @option{+fini} options for the same 3873purpose. The patches correct various problems with the +init/+fini 3874options, including program core dumps. Binutils 2.14 corrects a 3875problem on the 64-bit port resulting from HP's non-standard use of 3876the .init and .fini sections for array initializers and finalizers. 3877 3878Although the HP and GNU linkers are both supported for the 3879@samp{hppa64-hp-hpux11*} target, it is strongly recommended that the 3880HP linker be used for link editing on this target. 3881 3882At this time, the GNU linker does not support the creation of long 3883branch stubs. As a result, it can't successfully link binaries 3884containing branch offsets larger than 8 megabytes. In addition, 3885there are problems linking shared libraries, linking executables 3886with @option{-static}, and with dwarf2 unwind and exception support. 3887It also doesn't provide stubs for internal calls to global functions 3888in shared libraries, so these calls can't be overloaded. 3889 3890The HP dynamic loader does not support GNU symbol versioning, so symbol 3891versioning is not supported. It may be necessary to disable symbol 3892versioning with @option{--disable-symvers} when using GNU ld. 3893 3894POSIX threads are the default. The optional DCE thread library is not 3895supported, so @option{--enable-threads=dce} does not work. 3896 3897@html 3898<hr /> 3899@end html 3900@anchor{x-x-linux-gnu} 3901@heading *-*-linux-gnu 3902Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present 3903in glibc 2.2.5 and later. More information is available in the 3904libstdc++-v3 documentation. 3905 3906@html 3907<hr /> 3908@end html 3909@anchor{ix86-x-linux} 3910@heading i?86-*-linux* 3911As of GCC 3.3, binutils 2.13.1 or later is required for this platform. 3912See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information. 3913 3914If you receive Signal 11 errors when building on GNU/Linux, then it is 3915possible you have a hardware problem. Further information on this can be 3916found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}. 3917 3918@html 3919<hr /> 3920@end html 3921@anchor{ix86-x-solaris210} 3922@heading i?86-*-solaris2.10 3923Use this for Solaris 10 or later on x86 and x86-64 systems. Starting 3924with GCC 4.7, there is also a 64-bit @samp{amd64-*-solaris2.1[0-9]*} or 3925@samp{x86_64-*-solaris2.1[0-9]*} configuration that corresponds to 3926@samp{sparcv9-sun-solaris2*}. 3927 3928It is recommended that you configure GCC to use the GNU assembler. The 3929versions included in Solaris 10, from GNU binutils 2.15 (in 3930@file{/usr/sfw/bin/gas}), and Solaris 11, from GNU binutils 2.19 or 3931newer (also available as @file{/usr/bin/gas} and 3932@file{/usr/gnu/bin/as}), work fine. Please note that the current 3933version, from GNU binutils 2.26, only works on Solaris 12 when using the 3934Solaris linker. On Solaris 10 and 11, you either have to wait for GNU 3935binutils 2.26.1 or newer, or stay with GNU binutils 2.25.1. Recent 3936versions of the Solaris assembler in @file{/usr/ccs/bin/as} work almost 3937as well, though. 3938@c FIXME: as patch requirements? 3939 3940For linking, the Solaris linker, is preferred. If you want to use the GNU 3941linker instead, note that due to a packaging bug the version in Solaris 394210, from GNU binutils 2.15 (in @file{/usr/sfw/bin/gld}), cannot be used, 3943while the version in Solaris 11, from GNU binutils 2.19 or newer (also 3944in @file{/usr/gnu/bin/ld} and @file{/usr/bin/gld}), works, as does the 3945latest version, from GNU binutils 2.26. 3946 3947To use GNU @command{as}, configure with the options 3948@option{--with-gnu-as --with-as=@//usr/@/sfw/@/bin/@/gas}. It may be necessary 3949to configure with @option{--without-gnu-ld --with-ld=@//usr/@/ccs/@/bin/@/ld} to 3950guarantee use of Sun @command{ld}. 3951@c FIXME: why --without-gnu-ld --with-ld? 3952 3953@html 3954<hr /> 3955@end html 3956@anchor{ia64-x-linux} 3957@heading ia64-*-linux 3958IA-64 processor (also known as IPF, or Itanium Processor Family) 3959running GNU/Linux. 3960 3961If you are using the installed system libunwind library with 3962@option{--with-system-libunwind}, then you must use libunwind 0.98 or 3963later. 3964 3965None of the following versions of GCC has an ABI that is compatible 3966with any of the other versions in this list, with the exception that 3967Red Hat 2.96 and Trillian 000171 are compatible with each other: 39683.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717. 3969This primarily affects C++ programs and programs that create shared libraries. 3970GCC 3.1 or later is recommended for compiling linux, the kernel. 3971As of version 3.1 GCC is believed to be fully ABI compliant, and hence no 3972more major ABI changes are expected. 3973 3974@html 3975<hr /> 3976@end html 3977@anchor{ia64-x-hpux} 3978@heading ia64-*-hpux* 3979Building GCC on this target requires the GNU Assembler. The bundled HP 3980assembler will not work. To prevent GCC from using the wrong assembler, 3981the option @option{--with-gnu-as} may be necessary. 3982 3983The GCC libunwind library has not been ported to HPUX@. This means that for 3984GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions} 3985is required to build GCC@. For GCC 3.3 and later, this is the default. 3986For gcc 3.4.3 and later, @option{--enable-libunwind-exceptions} is 3987removed and the system libunwind library will always be used. 3988 3989@html 3990<hr /> 3991<!-- rs6000-ibm-aix*, powerpc-ibm-aix* --> 3992@end html 3993@anchor{x-ibm-aix} 3994@heading *-ibm-aix* 3995Support for AIX version 3 and older was discontinued in GCC 3.4. 3996Support for AIX version 4.2 and older was discontinued in GCC 4.5. 3997 3998``out of memory'' bootstrap failures may indicate a problem with 3999process resource limits (ulimit). Hard limits are configured in the 4000@file{/etc/security/limits} system configuration file. 4001 4002GCC 4.9 and above require a C++ compiler for bootstrap. IBM VAC++ / xlC 4003cannot bootstrap GCC. xlc can bootstrap an older version of GCC and 4004G++ can bootstrap recent releases of GCC. 4005 4006GCC can bootstrap with recent versions of IBM XLC, but bootstrapping 4007with an earlier release of GCC is recommended. Bootstrapping with XLC 4008requires a larger data segment, which can be enabled through the 4009@var{LDR_CNTRL} environment variable, e.g., 4010 4011@smallexample 4012% LDR_CNTRL=MAXDATA=0x50000000 4013% export LDR_CNTRL 4014@end smallexample 4015 4016One can start with a pre-compiled version of GCC to build from 4017sources. One may delete GCC's ``fixed'' header files when starting 4018with a version of GCC built for an earlier release of AIX. 4019 4020To speed up the configuration phases of bootstrapping and installing GCC, 4021one may use GNU Bash instead of AIX @command{/bin/sh}, e.g., 4022 4023@smallexample 4024% CONFIG_SHELL=/opt/freeware/bin/bash 4025% export CONFIG_SHELL 4026@end smallexample 4027 4028and then proceed as described in @uref{build.html,,the build 4029instructions}, where we strongly recommend specifying an absolute path 4030to invoke @var{srcdir}/configure. 4031 4032Because GCC on AIX is built as a 32-bit executable by default, 4033(although it can generate 64-bit programs) the GMP and MPFR libraries 4034required by gfortran must be 32-bit libraries. Building GMP and MPFR 4035as static archive libraries works better than shared libraries. 4036 4037Errors involving @code{alloca} when building GCC generally are due 4038to an incorrect definition of @code{CC} in the Makefile or mixing files 4039compiled with the native C compiler and GCC@. During the stage1 phase of 4040the build, the native AIX compiler @strong{must} be invoked as @command{cc} 4041(not @command{xlc}). Once @command{configure} has been informed of 4042@command{xlc}, one needs to use @samp{make distclean} to remove the 4043configure cache files and ensure that @env{CC} environment variable 4044does not provide a definition that will confuse @command{configure}. 4045If this error occurs during stage2 or later, then the problem most likely 4046is the version of Make (see above). 4047 4048The native @command{as} and @command{ld} are recommended for 4049bootstrapping on AIX@. The GNU Assembler, GNU Linker, and GNU 4050Binutils version 2.20 is the minimum level that supports bootstrap on 4051AIX 5@. The GNU Assembler has not been updated to support AIX 6@ or 4052AIX 7. The native AIX tools do interoperate with GCC@. 4053 4054AIX 7.1 added partial support for DWARF debugging, but full support 4055requires AIX 7.1 TL03 SP7 that supports additional DWARF sections and 4056fixes a bug in the assembler. AIX 7.1 TL03 SP5 distributed a version 4057of libm.a missing important symbols; a fix for IV77796 will be 4058included in SP6. 4059 4060AIX 5.3 TL10, AIX 6.1 TL05 and AIX 7.1 TL00 introduced an AIX 4061assembler change that sometimes produces corrupt assembly files 4062causing AIX linker errors. The bug breaks GCC bootstrap on AIX and 4063can cause compilation failures with existing GCC installations. An 4064AIX iFix for AIX 5.3 is available (APAR IZ98385 for AIX 5.3 TL10, APAR 4065IZ98477 for AIX 5.3 TL11 and IZ98134 for AIX 5.3 TL12). AIX 5.3 TL11 SP8, 4066AIX 5.3 TL12 SP5, AIX 6.1 TL04 SP11, AIX 6.1 TL05 SP7, AIX 6.1 TL06 SP6, 4067AIX 6.1 TL07 and AIX 7.1 TL01 should include the fix. 4068 4069Building @file{libstdc++.a} requires a fix for an AIX Assembler bug 4070APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a 4071fix for another AIX Assembler bug and a co-dependent AIX Archiver fix 4072referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1) 4073 4074@anchor{TransferAixShobj} 4075@samp{libstdc++} in GCC 3.4 increments the major version number of the 4076shared object and GCC installation places the @file{libstdc++.a} 4077shared library in a common location which will overwrite the and GCC 40783.3 version of the shared library. Applications either need to be 4079re-linked against the new shared library or the GCC 3.1 and GCC 3.3 4080versions of the @samp{libstdc++} shared object needs to be available 4081to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if 4082present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be 4083installed for runtime dynamic loading using the following steps to set 4084the @samp{F_LOADONLY} flag in the shared object for @emph{each} 4085multilib @file{libstdc++.a} installed: 4086 4087Extract the shared objects from the currently installed 4088@file{libstdc++.a} archive: 4089@smallexample 4090% ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5 4091@end smallexample 4092 4093Enable the @samp{F_LOADONLY} flag so that the shared object will be 4094available for runtime dynamic loading, but not linking: 4095@smallexample 4096% strip -e libstdc++.so.4 libstdc++.so.5 4097@end smallexample 4098 4099Archive the runtime-only shared object in the GCC 3.4 4100@file{libstdc++.a} archive: 4101@smallexample 4102% ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5 4103@end smallexample 4104 4105Eventually, the 4106@uref{./configure.html#WithAixSoname,,@option{--with-aix-soname=svr4}} 4107configure option may drop the need for this procedure for libraries that 4108support it. 4109 4110Linking executables and shared libraries may produce warnings of 4111duplicate symbols. The assembly files generated by GCC for AIX always 4112have included multiple symbol definitions for certain global variable 4113and function declarations in the original program. The warnings should 4114not prevent the linker from producing a correct library or runnable 4115executable. 4116 4117AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and 411864-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1 4119to parse archive libraries did not handle the new format correctly. 4120These routines are used by GCC and result in error messages during 4121linking such as ``not a COFF file''. The version of the routines shipped 4122with AIX 4.3.1 should work for a 32-bit environment. The @option{-g} 4123option of the archive command may be used to create archives of 32-bit 4124objects using the original ``small format''. A correct version of the 4125routines is shipped with AIX 4.3.2 and above. 4126 4127Some versions of the AIX binder (linker) can fail with a relocation 4128overflow severe error when the @option{-bbigtoc} option is used to link 4129GCC-produced object files into an executable that overflows the TOC@. A fix 4130for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is 4131available from IBM Customer Support and from its 4132@uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com} 4133website as PTF U455193. 4134 4135The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core 4136with a segmentation fault when invoked by any version of GCC@. A fix for 4137APAR IX87327 is available from IBM Customer Support and from its 4138@uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com} 4139website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above. 4140 4141The initial assembler shipped with AIX 4.3.0 generates incorrect object 4142files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS 4143TO ASSEMBLE/BIND) is available from IBM Customer Support and from its 4144@uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com} 4145website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above. 4146 4147AIX provides National Language Support (NLS)@. Compilers and assemblers 4148use NLS to support locale-specific representations of various data 4149formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for 4150separating decimal fractions). There have been problems reported where 4151GCC does not produce the same floating-point formats that the assembler 4152expects. If one encounters this problem, set the @env{LANG} 4153environment variable to @samp{C} or @samp{En_US}. 4154 4155A default can be specified with the @option{-mcpu=@var{cpu_type}} 4156switch and using the configure option @option{--with-cpu-@var{cpu_type}}. 4157 4158@html 4159<hr /> 4160@end html 4161@anchor{iq2000-x-elf} 4162@heading iq2000-*-elf 4163Vitesse IQ2000 processors. These are used in embedded 4164applications. There are no standard Unix configurations. 4165 4166@html 4167<hr /> 4168@end html 4169@anchor{lm32-x-elf} 4170@heading lm32-*-elf 4171Lattice Mico32 processor. 4172This configuration is intended for embedded systems. 4173 4174@html 4175<hr /> 4176@end html 4177@anchor{lm32-x-uclinux} 4178@heading lm32-*-uclinux 4179Lattice Mico32 processor. 4180This configuration is intended for embedded systems running uClinux. 4181 4182@html 4183<hr /> 4184@end html 4185@anchor{m32c-x-elf} 4186@heading m32c-*-elf 4187Renesas M32C processor. 4188This configuration is intended for embedded systems. 4189 4190@html 4191<hr /> 4192@end html 4193@anchor{m32r-x-elf} 4194@heading m32r-*-elf 4195Renesas M32R processor. 4196This configuration is intended for embedded systems. 4197 4198@html 4199<hr /> 4200@end html 4201@anchor{m68k-x-x} 4202@heading m68k-*-* 4203By default, 4204@samp{m68k-*-elf*}, @samp{m68k-*-rtems}, @samp{m68k-*-uclinux} and 4205@samp{m68k-*-linux} 4206build libraries for both M680x0 and ColdFire processors. If you only 4207need the M680x0 libraries, you can omit the ColdFire ones by passing 4208@option{--with-arch=m68k} to @command{configure}. Alternatively, you 4209can omit the M680x0 libraries by passing @option{--with-arch=cf} to 4210@command{configure}. These targets default to 5206 or 5475 code as 4211appropriate for the target system when 4212configured with @option{--with-arch=cf} and 68020 code otherwise. 4213 4214The @samp{m68k-*-netbsd} and 4215@samp{m68k-*-openbsd} targets also support the @option{--with-arch} 4216option. They will generate ColdFire CFV4e code when configured with 4217@option{--with-arch=cf} and 68020 code otherwise. 4218 4219You can override the default processors listed above by configuring 4220with @option{--with-cpu=@var{target}}. This @var{target} can either 4221be a @option{-mcpu} argument or one of the following values: 4222@samp{m68000}, @samp{m68010}, @samp{m68020}, @samp{m68030}, 4223@samp{m68040}, @samp{m68060}, @samp{m68020-40} and @samp{m68020-60}. 4224 4225GCC requires at least binutils version 2.17 on these targets. 4226 4227@html 4228<hr /> 4229@end html 4230@anchor{m68k-x-uclinux} 4231@heading m68k-*-uclinux 4232GCC 4.3 changed the uClinux configuration so that it uses the 4233@samp{m68k-linux-gnu} ABI rather than the @samp{m68k-elf} ABI. 4234It also added improved support for C++ and flat shared libraries, 4235both of which were ABI changes. 4236 4237@html 4238<hr /> 4239@end html 4240@anchor{mep-x-elf} 4241@heading mep-*-elf 4242Toshiba Media embedded Processor. 4243This configuration is intended for embedded systems. 4244 4245@html 4246<hr /> 4247@end html 4248@anchor{microblaze-x-elf} 4249@heading microblaze-*-elf 4250Xilinx MicroBlaze processor. 4251This configuration is intended for embedded systems. 4252 4253@html 4254<hr /> 4255@end html 4256@anchor{mips-x-x} 4257@heading mips-*-* 4258If on a MIPS system you get an error message saying ``does not have gp 4259sections for all it's [sic] sectons [sic]'', don't worry about it. This 4260happens whenever you use GAS with the MIPS linker, but there is not 4261really anything wrong, and it is okay to use the output file. You can 4262stop such warnings by installing the GNU linker. 4263 4264It would be nice to extend GAS to produce the gp tables, but they are 4265optional, and there should not be a warning about their absence. 4266 4267The libstdc++ atomic locking routines for MIPS targets requires MIPS II 4268and later. A patch went in just after the GCC 3.3 release to 4269make @samp{mips*-*-*} use the generic implementation instead. You can also 4270configure for @samp{mipsel-elf} as a workaround. The 4271@samp{mips*-*-linux*} target continues to use the MIPS II routines. More 4272work on this is expected in future releases. 4273 4274@c If you make --with-llsc the default for another target, please also 4275@c update the description of the --with-llsc option. 4276 4277The built-in @code{__sync_*} functions are available on MIPS II and 4278later systems and others that support the @samp{ll}, @samp{sc} and 4279@samp{sync} instructions. This can be overridden by passing 4280@option{--with-llsc} or @option{--without-llsc} when configuring GCC. 4281Since the Linux kernel emulates these instructions if they are 4282missing, the default for @samp{mips*-*-linux*} targets is 4283@option{--with-llsc}. The @option{--with-llsc} and 4284@option{--without-llsc} configure options may be overridden at compile 4285time by passing the @option{-mllsc} or @option{-mno-llsc} options to 4286the compiler. 4287 4288MIPS systems check for division by zero (unless 4289@option{-mno-check-zero-division} is passed to the compiler) by 4290generating either a conditional trap or a break instruction. Using 4291trap results in smaller code, but is only supported on MIPS II and 4292later. Also, some versions of the Linux kernel have a bug that 4293prevents trap from generating the proper signal (@code{SIGFPE}). To enable 4294the use of break, use the @option{--with-divide=breaks} 4295@command{configure} option when configuring GCC@. The default is to 4296use traps on systems that support them. 4297 4298The assembler from GNU binutils 2.17 and earlier has a bug in the way 4299it sorts relocations for REL targets (o32, o64, EABI). This can cause 4300bad code to be generated for simple C++ programs. Also the linker 4301from GNU binutils versions prior to 2.17 has a bug which causes the 4302runtime linker stubs in very large programs, like @file{libgcj.so}, to 4303be incorrectly generated. GNU Binutils 2.18 and later (and snapshots 4304made after Nov. 9, 2006) should be free from both of these problems. 4305 4306@html 4307<hr /> 4308@end html 4309@anchor{mips-sgi-irix5} 4310@heading mips-sgi-irix5 4311Support for IRIX 5 has been removed in GCC 4.6. 4312 4313@html 4314<hr /> 4315@end html 4316@anchor{mips-sgi-irix6} 4317@heading mips-sgi-irix6 4318Support for IRIX 6.5 has been removed in GCC 4.8. Support for IRIX 6 4319releases before 6.5 has been removed in GCC 4.6, as well as support for 4320the O32 ABI. 4321 4322@html 4323<hr /> 4324@end html 4325@anchor{moxie-x-elf} 4326@heading moxie-*-elf 4327The moxie processor. 4328 4329@html 4330<hr /> 4331@end html 4332@anchor{msp430-x-elf} 4333@heading msp430-*-elf 4334TI MSP430 processor. 4335This configuration is intended for embedded systems. 4336 4337@html 4338<hr /> 4339@end html 4340@anchor{nds32le-x-elf} 4341@heading nds32le-*-elf 4342Andes NDS32 target in little endian mode. 4343 4344@html 4345<hr /> 4346@end html 4347@anchor{nds32be-x-elf} 4348@heading nds32be-*-elf 4349Andes NDS32 target in big endian mode. 4350 4351@html 4352<hr /> 4353@end html 4354@anchor{nvptx-x-none} 4355@heading nvptx-*-none 4356Nvidia PTX target. 4357 4358Instead of GNU binutils, you will need to install 4359@uref{https://github.com/MentorEmbedded/nvptx-tools/,,nvptx-tools}. 4360Tell GCC where to find it: 4361@option{--with-build-time-tools=[install-nvptx-tools]/nvptx-none/bin}. 4362 4363A nvptx port of newlib is available at 4364@uref{https://github.com/MentorEmbedded/nvptx-newlib/,,nvptx-newlib}. 4365It can be automatically built together with GCC@. For this, add a 4366symbolic link to nvptx-newlib's @file{newlib} directory to the 4367directory containing the GCC sources. 4368 4369Use the @option{--disable-sjlj-exceptions} and 4370@option{--enable-newlib-io-long-long} options when configuring. 4371 4372@html 4373<hr /> 4374@end html 4375@anchor{powerpc-x-x} 4376@heading powerpc-*-* 4377You can specify a default version for the @option{-mcpu=@var{cpu_type}} 4378switch by using the configure option @option{--with-cpu-@var{cpu_type}}. 4379 4380You will need 4381@uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.15} 4382or newer for a working GCC@. 4383 4384@html 4385<hr /> 4386@end html 4387@anchor{powerpc-x-darwin} 4388@heading powerpc-*-darwin* 4389PowerPC running Darwin (Mac OS X kernel). 4390 4391Pre-installed versions of Mac OS X may not include any developer tools, 4392meaning that you will not be able to build GCC from source. Tool 4393binaries are available at 4394@uref{http://opensource.apple.com/}. 4395 4396This version of GCC requires at least cctools-590.36. The 4397cctools-590.36 package referenced from 4398@uref{http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html} will not work 4399on systems older than 10.3.9 (aka darwin7.9.0). 4400 4401@html 4402<hr /> 4403@end html 4404@anchor{powerpc-x-elf} 4405@heading powerpc-*-elf 4406PowerPC system in big endian mode, running System V.4. 4407 4408@html 4409<hr /> 4410@end html 4411@anchor{powerpc-x-linux-gnu} 4412@heading powerpc*-*-linux-gnu* 4413PowerPC system in big endian mode running Linux. 4414 4415@html 4416<hr /> 4417@end html 4418@anchor{powerpc-x-netbsd} 4419@heading powerpc-*-netbsd* 4420PowerPC system in big endian mode running NetBSD@. 4421 4422@html 4423<hr /> 4424@end html 4425@anchor{powerpc-x-eabisim} 4426@heading powerpc-*-eabisim 4427Embedded PowerPC system in big endian mode for use in running under the 4428PSIM simulator. 4429 4430@html 4431<hr /> 4432@end html 4433@anchor{powerpc-x-eabi} 4434@heading powerpc-*-eabi 4435Embedded PowerPC system in big endian mode. 4436 4437@html 4438<hr /> 4439@end html 4440@anchor{powerpcle-x-elf} 4441@heading powerpcle-*-elf 4442PowerPC system in little endian mode, running System V.4. 4443 4444@html 4445<hr /> 4446@end html 4447@anchor{powerpcle-x-eabisim} 4448@heading powerpcle-*-eabisim 4449Embedded PowerPC system in little endian mode for use in running under 4450the PSIM simulator. 4451 4452@html 4453<hr /> 4454@end html 4455@anchor{powerpcle-x-eabi} 4456@heading powerpcle-*-eabi 4457Embedded PowerPC system in little endian mode. 4458 4459@html 4460<hr /> 4461@end html 4462@anchor{rl78-x-elf} 4463@heading rl78-*-elf 4464The Renesas RL78 processor. 4465This configuration is intended for embedded systems. 4466 4467@html 4468<hr /> 4469@end html 4470@anchor{rx-x-elf} 4471@heading rx-*-elf 4472The Renesas RX processor. See 4473@uref{http://eu.renesas.com/fmwk.jsp?cnt=rx600_series_landing.jsp&fp=/products/mpumcu/rx_family/rx600_series} 4474for more information about this processor. 4475 4476@html 4477<hr /> 4478@end html 4479@anchor{s390-x-linux} 4480@heading s390-*-linux* 4481S/390 system running GNU/Linux for S/390@. 4482 4483@html 4484<hr /> 4485@end html 4486@anchor{s390x-x-linux} 4487@heading s390x-*-linux* 4488zSeries system (64-bit) running GNU/Linux for zSeries@. 4489 4490@html 4491<hr /> 4492@end html 4493@anchor{s390x-ibm-tpf} 4494@heading s390x-ibm-tpf* 4495zSeries system (64-bit) running TPF@. This platform is 4496supported as cross-compilation target only. 4497 4498@html 4499<hr /> 4500@end html 4501@c Please use Solaris 2 to refer to all release of Solaris, starting 4502@c with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for 4503@c SunOS 4 releases which we don't use to avoid confusion. Solaris 4504@c alone is too unspecific and must be avoided. 4505@anchor{x-x-solaris2} 4506@heading *-*-solaris2* 4507Support for Solaris 9 has been removed in GCC 5. Support for Solaris 45088 has been removed in GCC 4.8. Support for Solaris 7 has been removed 4509in GCC 4.6. 4510 4511Sun does not ship a C compiler with Solaris 2 before Solaris 10, though 4512you can download the Sun Studio compilers for free. In Solaris 10 and 451311, GCC 3.4.3 is available as @command{/usr/sfw/bin/gcc}. Solaris 11 4514also provides GCC 4.5.2, 4.7.3, and 4.8.2 as 4515@command{/usr/gcc/4.5/bin/gcc} or similar. Alternatively, 4516you can install a pre-built GCC to bootstrap and install GCC. See the 4517@uref{binaries.html,,binaries page} for details. 4518 4519The Solaris 2 @command{/bin/sh} will often fail to configure 4520@samp{libstdc++-v3}, @samp{boehm-gc} or @samp{libjava}. We therefore 4521recommend using the following initial sequence of commands 4522 4523@smallexample 4524% CONFIG_SHELL=/bin/ksh 4525% export CONFIG_SHELL 4526@end smallexample 4527 4528@noindent 4529and proceed as described in @uref{configure.html,,the configure instructions}. 4530In addition we strongly recommend specifying an absolute path to invoke 4531@command{@var{srcdir}/configure}. 4532 4533Solaris 10 comes with a number of optional OS packages. Some of these 4534are needed to use GCC fully, namely @code{SUNWarc}, 4535@code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm}, 4536@code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all 4537optional packages when installing Solaris 10, you will need to verify that 4538the packages that GCC needs are installed. 4539To check whether an optional package is installed, use 4540the @command{pkginfo} command. To add an optional package, use the 4541@command{pkgadd} command. For further details, see the Solaris 10 4542documentation. 4543 4544Starting with Solaris 11, the package management has changed, so you 4545need to check for @code{system/header}, @code{system/linker}, and 4546@code{developer/assembler} packages. Checking for and installing 4547packages is done with the @command{pkg} command now. 4548 4549Trying to use the linker and other tools in 4550@file{/usr/ucb} to install GCC has been observed to cause trouble. 4551For example, the linker may hang indefinitely. The fix is to remove 4552@file{/usr/ucb} from your @env{PATH}. 4553 4554The build process works more smoothly with the legacy Sun tools so, if you 4555have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place 4556@file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build. 4557 4558We recommend the use of the Solaris assembler or the GNU assembler, in 4559conjunction with the Solaris linker. The GNU @command{as} 4560versions included in Solaris 10, from GNU binutils 2.15 (in 4561@file{/usr/sfw/bin/gas}), and Solaris 11, 4562from GNU binutils 2.19 or newer (also in @file{/usr/bin/gas} and 4563@file{/usr/gnu/bin/as}), are known to work. 4564Current versions of GNU binutils (2.26) 4565are known to work as well, with the caveat mentioned in 4566@uref{#ix86-x-solaris210,,i?86-*-solaris2.10} . Note that your mileage may vary 4567if you use a combination of the GNU tools and the Solaris tools: while the 4568combination GNU @command{as} + Sun @command{ld} should reasonably work, 4569the reverse combination Sun @command{as} + GNU @command{ld} may fail to 4570build or cause memory corruption at runtime in some cases for C++ programs. 4571@c FIXME: still? 4572GNU @command{ld} usually works as well, although the version included in 4573Solaris 10 cannot be used due to several bugs. Again, the current 4574version (2.26) is known to work, but generally lacks platform specific 4575features, so better stay with Solaris @command{ld}. To use the LTO linker 4576plugin (@option{-fuse-linker-plugin}) with GNU @command{ld}, GNU 4577binutils @emph{must} be configured with @option{--enable-largefile}. 4578 4579To enable symbol versioning in @samp{libstdc++} with the Solaris linker, 4580you need to have any version of GNU @command{c++filt}, which is part of 4581GNU binutils. @samp{libstdc++} symbol versioning will be disabled if no 4582appropriate version is found. Solaris @command{c++filt} from the Solaris 4583Studio compilers does @emph{not} work. 4584 4585GNU @command{make} version 3.81 or later is required to build libjava 4586with the Solaris linker. 4587 4588Sun bug 4927647 sometimes causes random spurious testsuite failures 4589related to missing diagnostic output. This bug doesn't affect GCC 4590itself, rather it is a kernel bug triggered by the @command{expect} 4591program which is used only by the GCC testsuite driver. When the bug 4592causes the @command{expect} program to miss anticipated output, extra 4593testsuite failures appear. 4594 4595@html 4596<hr /> 4597@end html 4598@anchor{sparc-x-x} 4599@heading sparc*-*-* 4600This section contains general configuration information for all 4601SPARC-based platforms. In addition to reading this section, please 4602read all other sections that match your target. 4603 4604Newer versions of the GNU Multiple Precision Library (GMP), the MPFR 4605library and the MPC library are known to be miscompiled by earlier 4606versions of GCC on these platforms. We therefore recommend the use 4607of the exact versions of these libraries listed as minimal versions 4608in @uref{prerequisites.html,,the prerequisites}. 4609 4610@html 4611<hr /> 4612@end html 4613@anchor{sparc-sun-solaris2} 4614@heading sparc-sun-solaris2* 4615When GCC is configured to use GNU binutils 2.14 or later, the binaries 4616produced are smaller than the ones produced using Sun's native tools; 4617this difference is quite significant for binaries containing debugging 4618information. 4619 4620Starting with Solaris 7, the operating system is capable of executing 462164-bit SPARC V9 binaries. GCC 3.1 and later properly supports 4622this; the @option{-m64} option enables 64-bit code generation. 4623However, if all you want is code tuned for the UltraSPARC CPU, you 4624should try the @option{-mtune=ultrasparc} option instead, which produces 4625code that, unlike full 64-bit code, can still run on non-UltraSPARC 4626machines. 4627 4628When configuring the GNU Multiple Precision Library (GMP), the MPFR 4629library or the MPC library on a Solaris 7 or later system, the canonical 4630target triplet must be specified as the @command{build} parameter on the 4631configure line. This target triplet can be obtained by invoking @command{./config.guess} in the toplevel source directory of GCC (and 4632not that of GMP or MPFR or MPC). For example on a Solaris 9 system: 4633 4634@smallexample 4635% ./configure --build=sparc-sun-solaris2.9 --prefix=xxx 4636@end smallexample 4637 4638@html 4639<hr /> 4640@end html 4641@anchor{sparc-sun-solaris210} 4642@heading sparc-sun-solaris2.10 4643There is a bug in older versions of the Sun assembler which breaks 4644thread-local storage (TLS). A typical error message is 4645 4646@smallexample 4647ld: fatal: relocation error: R_SPARC_TLS_LE_HIX22: file /var/tmp//ccamPA1v.o: 4648 symbol <unknown>: bad symbol type SECT: symbol type must be TLS 4649@end smallexample 4650 4651@noindent 4652This bug is fixed in Sun patch 118683-03 or later. 4653 4654@html 4655<hr /> 4656@end html 4657@anchor{sparc-x-linux} 4658@heading sparc-*-linux* 4659 4660GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4 4661or newer on this platform. All earlier binutils and glibc 4662releases mishandled unaligned relocations on @code{sparc-*-*} targets. 4663 4664 4665@html 4666<hr /> 4667@end html 4668@anchor{sparc64-x-solaris2} 4669@heading sparc64-*-solaris2* 4670When configuring the GNU Multiple Precision Library (GMP), the MPFR 4671library or the MPC library, the canonical target triplet must be specified 4672as the @command{build} parameter on the configure line. For example 4673on a Solaris 9 system: 4674 4675@smallexample 4676% ./configure --build=sparc64-sun-solaris2.9 --prefix=xxx 4677@end smallexample 4678 4679@html 4680<hr /> 4681@end html 4682@anchor{sparcv9-x-solaris2} 4683@heading sparcv9-*-solaris2* 4684This is a synonym for @samp{sparc64-*-solaris2*}. 4685 4686@html 4687<hr /> 4688@end html 4689@anchor{c6x-x-x} 4690@heading c6x-*-* 4691The C6X family of processors. This port requires binutils-2.22 or newer. 4692 4693@html 4694<hr /> 4695@end html 4696@anchor{tilegx-*-linux} 4697@heading tilegx-*-linux* 4698The TILE-Gx processor in little endian mode, running GNU/Linux. This 4699port requires binutils-2.22 or newer. 4700 4701@html 4702<hr /> 4703@end html 4704@anchor{tilegxbe-*-linux} 4705@heading tilegxbe-*-linux* 4706The TILE-Gx processor in big endian mode, running GNU/Linux. This 4707port requires binutils-2.23 or newer. 4708 4709@html 4710<hr /> 4711@end html 4712@anchor{tilepro-*-linux} 4713@heading tilepro-*-linux* 4714The TILEPro processor running GNU/Linux. This port requires 4715binutils-2.22 or newer. 4716 4717@html 4718<hr /> 4719@end html 4720@anchor{visium-x-elf} 4721@heading visium-*-elf 4722CDS VISIUMcore processor. 4723This configuration is intended for embedded systems. 4724 4725@html 4726<hr /> 4727@end html 4728@anchor{x-x-vxworks} 4729@heading *-*-vxworks* 4730Support for VxWorks is in flux. At present GCC supports @emph{only} the 4731very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC@. 4732We welcome patches for other architectures supported by VxWorks 5.5. 4733Support for VxWorks AE would also be welcome; we believe this is merely 4734a matter of writing an appropriate ``configlette'' (see below). We are 4735not interested in supporting older, a.out or COFF-based, versions of 4736VxWorks in GCC 3. 4737 4738VxWorks comes with an older version of GCC installed in 4739@file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it. 4740Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}. 4741Before running @command{configure}, create the directories @file{@var{prefix}} 4742and @file{@var{prefix}/bin}. Link or copy the appropriate assembler, 4743linker, etc.@: into @file{@var{prefix}/bin}, and set your @var{PATH} to 4744include that directory while running both @command{configure} and 4745@command{make}. 4746 4747You must give @command{configure} the 4748@option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can 4749find the VxWorks system headers. Since VxWorks is a cross compilation 4750target only, you must also specify @option{--target=@var{target}}. 4751@command{configure} will attempt to create the directory 4752@file{@var{prefix}/@var{target}/sys-include} and copy files into it; 4753make sure the user running @command{configure} has sufficient privilege 4754to do so. 4755 4756GCC's exception handling runtime requires a special ``configlette'' 4757module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in 4758that file to add the module to your kernel build. (Future versions of 4759VxWorks will incorporate this module.) 4760 4761@html 4762<hr /> 4763@end html 4764@anchor{x86-64-x-x} 4765@heading x86_64-*-*, amd64-*-* 4766GCC supports the x86-64 architecture implemented by the AMD64 processor 4767(amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD@. 4768On GNU/Linux the default is a bi-arch compiler which is able to generate 4769both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch). 4770 4771@html 4772<hr /> 4773@end html 4774@anchor{x86-64-x-solaris210} 4775@heading x86_64-*-solaris2.1[0-9]* 4776GCC also supports the x86-64 architecture implemented by the AMD64 4777processor (@samp{amd64-*-*} is an alias for @samp{x86_64-*-*}) on 4778Solaris 10 or later. Unlike other systems, without special options a 4779bi-arch compiler is built which generates 32-bit code by default, but 4780can generate 64-bit x86-64 code with the @option{-m64} switch. Since 4781GCC 4.7, there is also a configuration that defaults to 64-bit code, but 4782can generate 32-bit code with @option{-m32}. To configure and build 4783this way, you have to provide all support libraries like @file{libgmp} 4784as 64-bit code, configure with @option{--target=x86_64-pc-solaris2.1x} 4785and @samp{CC=gcc -m64}. 4786 4787@html 4788<hr /> 4789@end html 4790@anchor{xtensa-x-elf} 4791@heading xtensa*-*-elf 4792This target is intended for embedded Xtensa systems using the 4793@samp{newlib} C library. It uses ELF but does not support shared 4794objects. Designed-defined instructions specified via the 4795Tensilica Instruction Extension (TIE) language are only supported 4796through inline assembly. 4797 4798The Xtensa configuration information must be specified prior to 4799building GCC@. The @file{include/xtensa-config.h} header 4800file contains the configuration information. If you created your 4801own Xtensa configuration with the Xtensa Processor Generator, the 4802downloaded files include a customized copy of this header file, 4803which you can use to replace the default header file. 4804 4805@html 4806<hr /> 4807@end html 4808@anchor{xtensa-x-linux} 4809@heading xtensa*-*-linux* 4810This target is for Xtensa systems running GNU/Linux. It supports ELF 4811shared objects and the GNU C library (glibc). It also generates 4812position-independent code (PIC) regardless of whether the 4813@option{-fpic} or @option{-fPIC} options are used. In other 4814respects, this target is the same as the 4815@uref{#xtensa*-*-elf,,@samp{xtensa*-*-elf}} target. 4816 4817@html 4818<hr /> 4819@end html 4820@anchor{windows} 4821@heading Microsoft Windows 4822 4823@subheading Intel 16-bit versions 4824The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not 4825supported. 4826 4827However, the 32-bit port has limited support for Microsoft 4828Windows 3.11 in the Win32s environment, as a target only. See below. 4829 4830@subheading Intel 32-bit versions 4831The 32-bit versions of Windows, including Windows 95, Windows NT, Windows 4832XP, and Windows Vista, are supported by several different target 4833platforms. These targets differ in which Windows subsystem they target 4834and which C libraries are used. 4835 4836@itemize 4837@item Cygwin @uref{#x-x-cygwin,,*-*-cygwin}: Cygwin provides a user-space 4838Linux API emulation layer in the Win32 subsystem. 4839@item Interix @uref{#x-x-interix,,*-*-interix}: The Interix subsystem 4840provides native support for POSIX. 4841@item MinGW @uref{#x-x-mingw32,,*-*-mingw32}: MinGW is a native GCC port for 4842the Win32 subsystem that provides a subset of POSIX. 4843@item MKS i386-pc-mks: NuTCracker from MKS. See 4844@uref{http://www.mkssoftware.com/} for more information. 4845@end itemize 4846 4847@subheading Intel 64-bit versions 4848GCC contains support for x86-64 using the mingw-w64 4849runtime library, available from @uref{http://mingw-w64.sourceforge.net/}. 4850This library should be used with the target triple x86_64-pc-mingw32. 4851 4852Presently Windows for Itanium is not supported. 4853 4854@subheading Windows CE 4855Windows CE is supported as a target only on Hitachi 4856SuperH (sh-wince-pe), and MIPS (mips-wince-pe). 4857 4858@subheading Other Windows Platforms 4859GCC no longer supports Windows NT on the Alpha or PowerPC. 4860 4861GCC no longer supports the Windows POSIX subsystem. However, it does 4862support the Interix subsystem. See above. 4863 4864Old target names including *-*-winnt and *-*-windowsnt are no longer used. 4865 4866PW32 (i386-pc-pw32) support was never completed, and the project seems to 4867be inactive. See @uref{http://pw32.sourceforge.net/} for more information. 4868 4869UWIN support has been removed due to a lack of maintenance. 4870 4871@html 4872<hr /> 4873@end html 4874@anchor{x-x-cygwin} 4875@heading *-*-cygwin 4876Ports of GCC are included with the 4877@uref{http://www.cygwin.com/,,Cygwin environment}. 4878 4879GCC will build under Cygwin without modification; it does not build 4880with Microsoft's C++ compiler and there are no plans to make it do so. 4881 4882The Cygwin native compiler can be configured to target any 32-bit x86 4883cpu architecture desired; the default is i686-pc-cygwin. It should be 4884used with as up-to-date a version of binutils as possible; use either 4885the latest official GNU binutils release in the Cygwin distribution, 4886or version 2.20 or above if building your own. 4887 4888@html 4889<hr /> 4890@end html 4891@anchor{x-x-interix} 4892@heading *-*-interix 4893The Interix target is used by OpenNT, Interix, Services For UNIX (SFU), 4894and Subsystem for UNIX-based Applications (SUA). Applications compiled 4895with this target run in the Interix subsystem, which is separate from 4896the Win32 subsystem. This target was last known to work in GCC 3.3. 4897 4898@html 4899<hr /> 4900@end html 4901@anchor{x-x-mingw32} 4902@heading *-*-mingw32 4903GCC will build with and support only MinGW runtime 3.12 and later. 4904Earlier versions of headers are incompatible with the new default semantics 4905of @code{extern inline} in @code{-std=c99} and @code{-std=gnu99} modes. 4906 4907@html 4908<hr /> 4909@end html 4910@anchor{older} 4911@heading Older systems 4912GCC contains support files for many older (1980s and early 49131990s) Unix variants. For the most part, support for these systems 4914has not been deliberately removed, but it has not been maintained for 4915several years and may suffer from bitrot. 4916 4917Starting with GCC 3.1, each release has a list of ``obsoleted'' systems. 4918Support for these systems is still present in that release, but 4919@command{configure} will fail unless the @option{--enable-obsolete} 4920option is given. Unless a maintainer steps forward, support for these 4921systems will be removed from the next release of GCC@. 4922 4923Support for old systems as hosts for GCC can cause problems if the 4924workarounds for compiler, library and operating system bugs affect the 4925cleanliness or maintainability of the rest of GCC@. In some cases, to 4926bring GCC up on such a system, if still possible with current GCC, may 4927require first installing an old version of GCC which did work on that 4928system, and using it to compile a more recent GCC, to avoid bugs in the 4929vendor compiler. Old releases of GCC 1 and GCC 2 are available in the 4930@file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror 4931sites}. Header bugs may generally be avoided using 4932@command{fixincludes}, but bugs or deficiencies in libraries and the 4933operating system may still cause problems. 4934 4935Support for older systems as targets for cross-compilation is less 4936problematic than support for them as hosts for GCC; if an enthusiast 4937wishes to make such a target work again (including resurrecting any of 4938the targets that never worked with GCC 2, starting from the last 4939version before they were removed), patches 4940@uref{../contribute.html,,following the usual requirements} would be 4941likely to be accepted, since they should not affect the support for more 4942modern targets. 4943 4944For some systems, old versions of GNU binutils may also be useful, 4945and are available from @file{pub/binutils/old-releases} on 4946@uref{http://sourceware.org/mirrors.html,,sourceware.org mirror sites}. 4947 4948Some of the information on specific systems above relates to 4949such older systems, but much of the information 4950about GCC on such systems (which may no longer be applicable to 4951current GCC) is to be found in the GCC texinfo manual. 4952 4953@html 4954<hr /> 4955@end html 4956@anchor{elf} 4957@heading all ELF targets (SVR4, Solaris 2, etc.) 4958C++ support is significantly better on ELF targets if you use the 4959@uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of 4960inlines, vtables and template instantiations will be discarded 4961automatically. 4962 4963 4964@html 4965<hr /> 4966<p> 4967@end html 4968@ifhtml 4969@uref{./index.html,,Return to the GCC Installation page} 4970@end ifhtml 4971@end ifset 4972 4973@c ***Old documentation****************************************************** 4974@ifset oldhtml 4975@include install-old.texi 4976@html 4977<hr /> 4978<p> 4979@end html 4980@ifhtml 4981@uref{./index.html,,Return to the GCC Installation page} 4982@end ifhtml 4983@end ifset 4984 4985@c ***GFDL******************************************************************** 4986@ifset gfdlhtml 4987@include fdl.texi 4988@html 4989<hr /> 4990<p> 4991@end html 4992@ifhtml 4993@uref{./index.html,,Return to the GCC Installation page} 4994@end ifhtml 4995@end ifset 4996 4997@c *************************************************************************** 4998@c Part 6 The End of the Document 4999@ifinfo 5000@comment node-name, next, previous, up 5001@node Concept Index, , GNU Free Documentation License, Top 5002@end ifinfo 5003 5004@ifinfo 5005@unnumbered Concept Index 5006 5007@printindex cp 5008 5009@contents 5010@end ifinfo 5011@bye 5012