]>
Commit | Line | Data |
---|---|---|
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 | @c Specify title for specific html page | |
11 | @ifset indexhtml | |
12 | @settitle Installing GCC | |
13 | @end ifset | |
14 | @ifset specifichtml | |
15 | @settitle Host/Target specific installation notes for GCC | |
16 | @end ifset | |
17 | @ifset prerequisiteshtml | |
18 | @settitle Prerequisites for GCC | |
19 | @end ifset | |
20 | @ifset downloadhtml | |
21 | @settitle Downloading GCC | |
22 | @end ifset | |
23 | @ifset configurehtml | |
24 | @settitle Installing GCC: Configuration | |
25 | @end ifset | |
26 | @ifset buildhtml | |
27 | @settitle Installing GCC: Building | |
28 | @end ifset | |
29 | @ifset testhtml | |
30 | @settitle Installing GCC: Testing | |
31 | @end ifset | |
32 | @ifset finalinstallhtml | |
33 | @settitle Installing GCC: Final installation | |
34 | @end ifset | |
35 | @ifset binarieshtml | |
36 | @settitle Installing GCC: Binaries | |
37 | @end ifset | |
38 | @ifset oldhtml | |
39 | @settitle Installing GCC: Old documentation | |
40 | @end ifset | |
41 | @ifset gfdlhtml | |
42 | @settitle Installing GCC: GNU Free Documentation License | |
43 | @end ifset | |
44 | ||
45 | @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, | |
46 | @c 1999, 2000, 2001, 2002, 2003, 2004 Free Software Foundation, Inc. | |
47 | @c *** Converted to texinfo by Dean Wakerley, dean@wakerley.com | |
48 | ||
49 | @c Include everything if we're not making html | |
50 | @ifnothtml | |
51 | @set indexhtml | |
52 | @set specifichtml | |
53 | @set prerequisiteshtml | |
54 | @set downloadhtml | |
55 | @set configurehtml | |
56 | @set buildhtml | |
57 | @set testhtml | |
58 | @set finalinstallhtml | |
59 | @set binarieshtml | |
60 | @set oldhtml | |
61 | @set gfdlhtml | |
62 | @end ifnothtml | |
63 | ||
64 | @c Part 2 Summary Description and Copyright | |
65 | @copying | |
66 | Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, | |
67 | 1999, 2000, 2001, 2002, 2003, 2004 Free Software Foundation, Inc. | |
68 | @sp 1 | |
69 | Permission is granted to copy, distribute and/or modify this document | |
70 | under the terms of the GNU Free Documentation License, Version 1.2 or | |
71 | any later version published by the Free Software Foundation; with no | |
72 | Invariant Sections, the Front-Cover texts being (a) (see below), and | |
73 | with the Back-Cover Texts being (b) (see below). A copy of the | |
74 | license is included in the section entitled ``@uref{./gfdl.html,,GNU | |
75 | Free Documentation License}''. | |
76 | ||
77 | (a) The FSF's Front-Cover Text is: | |
78 | ||
79 | A GNU Manual | |
80 | ||
81 | (b) The FSF's Back-Cover Text is: | |
82 | ||
83 | You have freedom to copy and modify this GNU Manual, like GNU | |
84 | software. Copies published by the Free Software Foundation raise | |
85 | funds for GNU development. | |
86 | @end copying | |
87 | @ifinfo | |
88 | @insertcopying | |
89 | @end ifinfo | |
90 | ||
91 | @c Part 3 Titlepage and Copyright | |
92 | @titlepage | |
93 | @sp 10 | |
94 | @comment The title is printed in a large font. | |
95 | @center @titlefont{Installing GCC} | |
96 | ||
97 | @c The following two commands start the copyright page. | |
98 | @page | |
99 | @vskip 0pt plus 1filll | |
100 | @insertcopying | |
101 | @end titlepage | |
102 | ||
103 | @c Part 4 Top node and Master Menu | |
104 | @ifinfo | |
105 | @node Top, , , (dir) | |
106 | @comment node-name, next, Previous, up | |
107 | ||
108 | @menu | |
109 | * Installing GCC:: This document describes the generic installation | |
110 | procedure for GCC as well as detailing some target | |
111 | specific installation instructions. | |
112 | ||
113 | * Specific:: Host/target specific installation notes for GCC. | |
114 | * Binaries:: Where to get pre-compiled binaries. | |
115 | ||
116 | * Old:: Old installation documentation. | |
117 | ||
118 | * GNU Free Documentation License:: How you can copy and share this manual. | |
119 | * Concept Index:: This index has two entries. | |
120 | @end menu | |
121 | @end ifinfo | |
122 | ||
123 | @c Part 5 The Body of the Document | |
124 | @c ***Installing GCC********************************************************** | |
125 | @ifnothtml | |
126 | @comment node-name, next, previous, up | |
127 | @node Installing GCC, Binaries, , Top | |
128 | @end ifnothtml | |
129 | @ifset indexhtml | |
130 | @ifnothtml | |
131 | @chapter Installing GCC | |
132 | @end ifnothtml | |
133 | ||
134 | The latest version of this document is always available at | |
135 | @uref{http://gcc.gnu.org/install/,,http://gcc.gnu.org/install/}. | |
136 | ||
137 | This document describes the generic installation procedure for GCC as well | |
138 | as detailing some target specific installation instructions. | |
139 | ||
140 | GCC includes several components that previously were separate distributions | |
141 | with their own installation instructions. This document supersedes all | |
142 | package specific installation instructions. | |
143 | ||
144 | @emph{Before} starting the build/install procedure please check the | |
145 | @ifnothtml | |
146 | @ref{Specific, host/target specific installation notes}. | |
147 | @end ifnothtml | |
148 | @ifhtml | |
149 | @uref{specific.html,,host/target specific installation notes}. | |
150 | @end ifhtml | |
151 | We recommend you browse the entire generic installation instructions before | |
152 | you proceed. | |
153 | ||
154 | Lists of successful builds for released versions of GCC are | |
155 | available at @uref{http://gcc.gnu.org/buildstat.html}. | |
156 | These lists are updated as new information becomes available. | |
157 | ||
158 | The installation procedure itself is broken into five steps. | |
159 | ||
160 | @ifinfo | |
161 | @menu | |
162 | * Prerequisites:: | |
163 | * Downloading the source:: | |
164 | * Configuration:: | |
165 | * Building:: | |
166 | * Testing:: (optional) | |
167 | * Final install:: | |
168 | @end menu | |
169 | @end ifinfo | |
170 | @ifhtml | |
171 | @enumerate | |
172 | @item | |
173 | @uref{prerequisites.html,,Prerequisites} | |
174 | @item | |
175 | @uref{download.html,,Downloading the source} | |
176 | @item | |
177 | @uref{configure.html,,Configuration} | |
178 | @item | |
179 | @uref{build.html,,Building} | |
180 | @item | |
181 | @uref{test.html,,Testing} (optional) | |
182 | @item | |
183 | @uref{finalinstall.html,,Final install} | |
184 | @end enumerate | |
185 | @end ifhtml | |
186 | ||
187 | Please note that GCC does not support @samp{make uninstall} and probably | |
188 | won't do so in the near future as this would open a can of worms. Instead, | |
189 | we suggest that you install GCC into a directory of its own and simply | |
190 | remove that directory when you do not need that specific version of GCC | |
191 | any longer, and, if shared libraries are installed there as well, no | |
192 | more binaries exist that use them. | |
193 | ||
194 | @ifhtml | |
195 | There are also some @uref{old.html,,old installation instructions}, | |
196 | which are mostly obsolete but still contain some information which has | |
197 | not yet been merged into the main part of this manual. | |
198 | @end ifhtml | |
199 | ||
200 | @html | |
201 | <hr /> | |
202 | <p> | |
203 | @end html | |
204 | @ifhtml | |
205 | @uref{./index.html,,Return to the GCC Installation page} | |
206 | ||
207 | @insertcopying | |
208 | @end ifhtml | |
209 | @end ifset | |
210 | ||
211 | @c ***Prerequisites************************************************** | |
212 | @ifnothtml | |
213 | @comment node-name, next, previous, up | |
214 | @node Prerequisites, Downloading the source, , Installing GCC | |
215 | @end ifnothtml | |
216 | @ifset prerequisiteshtml | |
217 | @ifnothtml | |
218 | @chapter Prerequisites | |
219 | @end ifnothtml | |
220 | @cindex Prerequisites | |
221 | ||
222 | GCC requires that various tools and packages be available for use in the | |
223 | build procedure. Modifying GCC sources requires additional tools | |
224 | described below. | |
225 | ||
226 | @heading Tools/packages necessary for building GCC | |
227 | @table @asis | |
228 | @item ISO C90 compiler | |
229 | Necessary to bootstrap the GCC package, although versions of GCC prior | |
230 | to 3.4 also allow bootstrapping with a traditional (K&R) C compiler. | |
231 | ||
232 | To make all languages in a cross-compiler or other configuration where | |
233 | 3-stage bootstrap is not performed, you need to start with an existing | |
234 | GCC binary (version 2.95 or later) because source code for language | |
235 | frontends other than C might use GCC extensions. | |
236 | ||
237 | @item GNAT | |
238 | ||
239 | In order to build the Ada compiler (GNAT) you must already have GNAT | |
240 | installed because portions of the Ada frontend are written in Ada (with | |
241 | GNAT extensions.) Refer to the Ada installation instructions for more | |
242 | specific information. | |
243 | ||
244 | @item A ``working'' POSIX compatible shell, or GNU bash | |
245 | ||
246 | Necessary when running @command{configure} because some | |
247 | @command{/bin/sh} shells have bugs and may crash when configuring the | |
248 | target libraries. In other cases, @command{/bin/sh} or even some | |
249 | @command{ksh} have disastrous corner-case performance problems. This | |
250 | can cause target @command{configure} runs to literally take days to | |
251 | complete in some cases. | |
252 | ||
253 | So on some platforms @command{/bin/ksh} is sufficient, on others it | |
254 | isn't. See the host/target specific instructions for your platform, or | |
255 | use @command{bash} to be sure. Then set @env{CONFIG_SHELL} in your | |
256 | environment to your ``good'' shell prior to running | |
257 | @command{configure}/@command{make}. | |
258 | ||
259 | @command{zsh} is not a fully compliant POSIX shell and will not | |
260 | work when configuring GCC. | |
261 | ||
262 | @item GNU binutils | |
263 | ||
264 | Necessary in some circumstances, optional in others. See the | |
265 | host/target specific instructions for your platform for the exact | |
266 | requirements. | |
267 | ||
268 | @item gzip version 1.2.4 (or later) or | |
269 | @itemx bzip2 version 1.0.2 (or later) | |
270 | ||
271 | Necessary to uncompress GCC @command{tar} files when source code is | |
272 | obtained via FTP mirror sites. | |
273 | ||
274 | @item GNU make version 3.79.1 (or later) | |
275 | ||
276 | You must have GNU make installed to build GCC. | |
277 | ||
278 | @item GNU tar version 1.12 (or later) | |
279 | ||
280 | Necessary (only on some platforms) to untar the source code. Many | |
281 | systems' @command{tar} programs will also work, only try GNU | |
282 | @command{tar} if you have problems. | |
283 | ||
284 | @end table | |
285 | ||
286 | ||
287 | @heading Tools/packages necessary for modifying GCC | |
288 | @table @asis | |
289 | ||
290 | @item autoconf versions 2.13 and 2.59 | |
291 | @itemx GNU m4 version 1.4 (or later) | |
292 | ||
293 | Necessary when modifying @file{configure.in}, @file{aclocal.m4}, etc.@: | |
294 | to regenerate @file{configure} and @file{config.in} files. Most | |
295 | directories require autoconf 2.13 (exactly), but @file{libiberty}, | |
296 | @file{fastjar}, @file{libstdc++-v3}, @file{libjava/libltdl}, @file{boehm-gc}, | |
297 | @file{intl}, @file{libada}, @file{libffi} and @file{gcc} require autoconf | |
298 | 2.59 (exactly). | |
299 | ||
300 | @item automake versions 1.4-gcj and 1.7.9 | |
301 | ||
302 | Necessary when modifying a @file{Makefile.am} file to regenerate its | |
303 | associated @file{Makefile.in}. | |
304 | ||
305 | Much of GCC does not use automake, so directly edit the @file{Makefile.in} | |
306 | file. Specifically this applies to the @file{gcc}, @file{intl}, | |
307 | @file{libf2c}, @file{libiberty}, @file{libobjc} directories as well as any | |
308 | of their subdirectories. | |
309 | ||
310 | The @file{libstdc++-v3}, @file{libjava/libltdl}, @file{fastjar} and | |
311 | @file{libffi} directories require automake 1.7.9. However, the Java | |
312 | directories, which include @file{boehm-gc}, @file{libjava}, and @file{zlib}, | |
313 | require a modified version of automake 1.4 downloadable from | |
314 | @uref{ftp://gcc.gnu.org/pub/java/automake-gcj-1.4.tar.gz}. | |
315 | ||
316 | @item gettext version 0.12 (or later) | |
317 | ||
318 | Needed to regenerate @file{gcc.pot}. | |
319 | ||
320 | @item gperf version 2.7.2 (or later) | |
321 | ||
322 | Necessary when modifying @command{gperf} input files, e.g.@: | |
323 | @file{gcc/cp/cfns.gperf} to regenerate its associated header file, e.g.@: | |
324 | @file{gcc/cp/cfns.h}. | |
325 | ||
326 | @item expect version ??? | |
327 | @itemx tcl version ??? | |
328 | @itemx dejagnu version ??? | |
329 | ||
330 | Necessary to run the GCC testsuite. | |
331 | ||
332 | @item autogen version 5.5.4 (or later) and | |
333 | @itemx guile version 1.4.1 (or later) | |
334 | ||
335 | Necessary to regenerate @file{fixinc/fixincl.x} from | |
336 | @file{fixinc/inclhack.def} and @file{fixinc/*.tpl}. | |
337 | ||
338 | Necessary to run the @file{fixinc} @command{make check}. | |
339 | ||
340 | Necessary to regenerate the top level @file{Makefile.in} file from | |
341 | @file{Makefile.tpl} and @file{Makefile.def}. | |
342 | ||
343 | @item GNU Bison version 1.28 (or later) | |
344 | Berkeley @command{yacc} (@command{byacc}) is also reported to work other | |
345 | than for java. | |
346 | ||
347 | Necessary when modifying @file{*.y} files. | |
348 | ||
349 | Necessary to build GCC during development because the generated output | |
350 | files are not included in the CVS repository. They are included in | |
351 | releases. | |
352 | ||
353 | @item Flex version 2.5.4 (or later) | |
354 | ||
355 | Necessary when modifying @file{*.l} files. | |
356 | ||
357 | Necessary to build GCC during development because the generated output | |
358 | files are not included in the CVS repository. They are included in | |
359 | releases. | |
360 | ||
361 | @item Texinfo version 4.2 (or later) | |
362 | ||
363 | Necessary for running @command{makeinfo} when modifying @file{*.texi} | |
364 | files to test your changes. | |
365 | ||
366 | Necessary to build GCC documentation during development because the | |
367 | generated output files are not included in the CVS repository. They are | |
368 | included in releases. | |
369 | ||
370 | @item @TeX{} (any working version) | |
371 | ||
372 | Necessary for running @command{texi2dvi}, used when running | |
373 | @command{make dvi} to create DVI files. | |
374 | ||
375 | @item cvs version 1.10 (or later) | |
376 | @itemx ssh (any version) | |
377 | ||
378 | Necessary to access the CVS repository. Public releases and weekly | |
379 | snapshots of the development sources are also available via FTP. | |
380 | ||
381 | @item perl version 5.6.1 (or later) | |
382 | ||
383 | Necessary when regenerating @file{Makefile} dependencies in libiberty. | |
384 | Necessary when regenerating @file{libiberty/functions.texi}. | |
385 | Necessary when generating manpages from Texinfo manuals. | |
386 | Used by various scripts to generate some files included in CVS (mainly | |
387 | Unicode-related and rarely changing) from source tables. | |
388 | ||
389 | @item GNU diffutils version 2.7 (or later) | |
390 | ||
391 | Necessary when creating changes to GCC source code to submit for review. | |
392 | ||
393 | @item patch version 2.5.4 (or later) | |
394 | ||
395 | Necessary when applying patches, created with @command{diff}, to one's | |
396 | own sources. | |
397 | ||
398 | @end table | |
399 | ||
400 | @html | |
401 | <hr /> | |
402 | <p> | |
403 | @end html | |
404 | @ifhtml | |
405 | @uref{./index.html,,Return to the GCC Installation page} | |
406 | @end ifhtml | |
407 | @end ifset | |
408 | ||
409 | @c ***Downloading the source************************************************** | |
410 | @ifnothtml | |
411 | @comment node-name, next, previous, up | |
412 | @node Downloading the source, Configuration, Prerequisites, Installing GCC | |
413 | @end ifnothtml | |
414 | @ifset downloadhtml | |
415 | @ifnothtml | |
416 | @chapter Downloading GCC | |
417 | @end ifnothtml | |
418 | @cindex Downloading GCC | |
419 | @cindex Downloading the Source | |
420 | ||
421 | GCC is distributed via @uref{http://gcc.gnu.org/cvs.html,,CVS} and FTP | |
422 | tarballs compressed with @command{gzip} or | |
423 | @command{bzip2}. It is possible to download a full distribution or specific | |
424 | components. | |
425 | ||
426 | Please refer to our @uref{http://gcc.gnu.org/releases.html,,releases web page} | |
427 | for information on how to obtain GCC@. | |
428 | ||
429 | The full distribution includes the C, C++, Objective-C, Fortran, Java, | |
430 | and Ada (in case of GCC 3.1 and later) compilers. The full distribution | |
431 | also includes runtime libraries for C++, Objective-C, Fortran, and Java. | |
432 | In GCC 3.0 and later versions, GNU compiler testsuites are also included | |
433 | in the full distribution. | |
434 | ||
435 | If you choose to download specific components, you must download the core | |
436 | GCC distribution plus any language specific distributions you wish to | |
437 | use. The core distribution includes the C language front end as well as the | |
438 | shared components. Each language has a tarball which includes the language | |
439 | front end as well as the language runtime (when appropriate). | |
440 | ||
441 | Unpack the core distribution as well as any language specific | |
442 | distributions in the same directory. | |
443 | ||
444 | If you also intend to build binutils (either to upgrade an existing | |
445 | installation or for use in place of the corresponding tools of your | |
446 | OS), unpack the binutils distribution either in the same directory or | |
447 | a separate one. In the latter case, add symbolic links to any | |
448 | components of the binutils you intend to build alongside the compiler | |
449 | (@file{bfd}, @file{binutils}, @file{gas}, @file{gprof}, @file{ld}, | |
450 | @file{opcodes}, @dots{}) to the directory containing the GCC sources. | |
451 | ||
452 | @html | |
453 | <hr /> | |
454 | <p> | |
455 | @end html | |
456 | @ifhtml | |
457 | @uref{./index.html,,Return to the GCC Installation page} | |
458 | @end ifhtml | |
459 | @end ifset | |
460 | ||
461 | @c ***Configuration*********************************************************** | |
462 | @ifnothtml | |
463 | @comment node-name, next, previous, up | |
464 | @node Configuration, Building, Downloading the source, Installing GCC | |
465 | @end ifnothtml | |
466 | @ifset configurehtml | |
467 | @ifnothtml | |
468 | @chapter Installing GCC: Configuration | |
469 | @end ifnothtml | |
470 | @cindex Configuration | |
471 | @cindex Installing GCC: Configuration | |
472 | ||
473 | Like most GNU software, GCC must be configured before it can be built. | |
474 | This document describes the recommended configuration procedure | |
475 | for both native and cross targets. | |
476 | ||
477 | We use @var{srcdir} to refer to the toplevel source directory for | |
478 | GCC; we use @var{objdir} to refer to the toplevel build/object directory. | |
479 | ||
480 | If you obtained the sources via CVS, @var{srcdir} must refer to the top | |
481 | @file{gcc} directory, the one where the @file{MAINTAINERS} can be found, | |
482 | and not its @file{gcc} subdirectory, otherwise the build will fail. | |
483 | ||
484 | If either @var{srcdir} or @var{objdir} is located on an automounted NFS | |
485 | file system, the shell's built-in @command{pwd} command will return | |
486 | temporary pathnames. Using these can lead to various sorts of build | |
487 | problems. To avoid this issue, set the @env{PWDCMD} environment | |
488 | variable to an automounter-aware @command{pwd} command, e.g., | |
489 | @command{pawd} or @samp{amq -w}, during the configuration and build | |
490 | phases. | |
491 | ||
492 | First, we @strong{highly} recommend that GCC be built into a | |
493 | separate directory than the sources which does @strong{not} reside | |
494 | within the source tree. This is how we generally build GCC; building | |
495 | where @var{srcdir} == @var{objdir} should still work, but doesn't | |
496 | get extensive testing; building where @var{objdir} is a subdirectory | |
497 | of @var{srcdir} is unsupported. | |
498 | ||
499 | If you have previously built GCC in the same directory for a | |
500 | different target machine, do @samp{make distclean} to delete all files | |
501 | that might be invalid. One of the files this deletes is @file{Makefile}; | |
502 | if @samp{make distclean} complains that @file{Makefile} does not exist | |
503 | or issues a message like ``don't know how to make distclean'' it probably | |
504 | means that the directory is already suitably clean. However, with the | |
505 | recommended method of building in a separate @var{objdir}, you should | |
506 | simply use a different @var{objdir} for each target. | |
507 | ||
508 | Second, when configuring a native system, either @command{cc} or | |
509 | @command{gcc} must be in your path or you must set @env{CC} in | |
510 | your environment before running configure. Otherwise the configuration | |
511 | scripts may fail. | |
512 | ||
513 | Note that the bootstrap compiler and the resulting GCC must be link | |
514 | compatible, else the bootstrap will fail with linker errors about | |
515 | incompatible object file formats. Several multilibed targets are | |
516 | affected by this requirement, see | |
517 | @ifnothtml | |
518 | @ref{Specific, host/target specific installation notes}. | |
519 | @end ifnothtml | |
520 | @ifhtml | |
521 | @uref{specific.html,,host/target specific installation notes}. | |
522 | @end ifhtml | |
523 | ||
524 | To configure GCC: | |
525 | ||
526 | @smallexample | |
527 | % mkdir @var{objdir} | |
528 | % cd @var{objdir} | |
529 | % @var{srcdir}/configure [@var{options}] [@var{target}] | |
530 | @end smallexample | |
531 | ||
532 | ||
533 | @heading Target specification | |
534 | @itemize @bullet | |
535 | @item | |
536 | GCC has code to correctly determine the correct value for @var{target} | |
537 | for nearly all native systems. Therefore, we highly recommend you not | |
538 | provide a configure target when configuring a native compiler. | |
539 | ||
540 | @item | |
541 | @var{target} must be specified as @option{--target=@var{target}} | |
542 | when configuring a cross compiler; examples of valid targets would be | |
543 | m68k-coff, sh-elf, etc. | |
544 | ||
545 | @item | |
546 | Specifying just @var{target} instead of @option{--target=@var{target}} | |
547 | implies that the host defaults to @var{target}. | |
548 | @end itemize | |
549 | ||
550 | ||
551 | @heading Options specification | |
552 | ||
553 | Use @var{options} to override several configure time options for | |
554 | GCC@. A list of supported @var{options} follows; @samp{configure | |
555 | --help} may list other options, but those not listed below may not | |
556 | work and should not normally be used. | |
557 | ||
558 | Note that each @option{--enable} option has a corresponding | |
559 | @option{--disable} option and that each @option{--with} option has a | |
560 | corresponding @option{--without} option. | |
561 | ||
562 | @table @code | |
563 | @item --prefix=@var{dirname} | |
564 | Specify the toplevel installation | |
565 | directory. This is the recommended way to install the tools into a directory | |
566 | other than the default. The toplevel installation directory defaults to | |
567 | @file{/usr/local}. | |
568 | ||
569 | We @strong{highly} recommend against @var{dirname} being the same or a | |
570 | subdirectory of @var{objdir} or vice versa. If specifying a directory | |
571 | beneath a user's home directory tree, some shells will not expand | |
572 | @var{dirname} correctly if it contains the @samp{~} metacharacter; use | |
573 | @env{$HOME} instead. | |
574 | ||
575 | The following standard @command{autoconf} options are supported. Normally you | |
576 | should not need to use these options. | |
577 | @table @code | |
578 | @item --exec-prefix=@var{dirname} | |
579 | Specify the toplevel installation directory for architecture-dependent | |
580 | files. The default is @file{@var{prefix}}. | |
581 | ||
582 | @item --bindir=@var{dirname} | |
583 | Specify the installation directory for the executables called by users | |
584 | (such as @command{gcc} and @command{g++}). The default is | |
585 | @file{@var{exec-prefix}/bin}. | |
586 | ||
587 | @item --libdir=@var{dirname} | |
588 | Specify the installation directory for object code libraries and | |
589 | internal data files of GCC@. The default is @file{@var{exec-prefix}/lib}. | |
590 | ||
591 | @item --libexecdir=@var{dirname} | |
592 | Specify the installation directory for internal executables of GCC@. | |
593 | The default is @file{@var{exec-prefix}/libexec}. | |
594 | ||
595 | @item --with-slibdir=@var{dirname} | |
596 | Specify the installation directory for the shared libgcc library. The | |
597 | default is @file{@var{libdir}}. | |
598 | ||
599 | @item --infodir=@var{dirname} | |
600 | Specify the installation directory for documentation in info format. | |
601 | The default is @file{@var{prefix}/info}. | |
602 | ||
603 | @item --datadir=@var{dirname} | |
604 | Specify the installation directory for some architecture-independent | |
605 | data files referenced by GCC@. The default is @file{@var{prefix}/share}. | |
606 | ||
607 | @item --mandir=@var{dirname} | |
608 | Specify the installation directory for manual pages. The default is | |
609 | @file{@var{prefix}/man}. (Note that the manual pages are only extracts from | |
610 | the full GCC manuals, which are provided in Texinfo format. The manpages | |
611 | are derived by an automatic conversion process from parts of the full | |
612 | manual.) | |
613 | ||
614 | @item --with-gxx-include-dir=@var{dirname} | |
615 | Specify | |
616 | the installation directory for G++ header files. The default is | |
617 | @file{@var{prefix}/include/c++/@var{version}}. | |
618 | ||
619 | @end table | |
620 | ||
621 | @item --program-prefix=@var{prefix} | |
622 | GCC supports some transformations of the names of its programs when | |
623 | installing them. This option prepends @var{prefix} to the names of | |
624 | programs to install in @var{bindir} (see above). For example, specifying | |
625 | @option{--program-prefix=foo-} would result in @samp{gcc} | |
626 | being installed as @file{/usr/local/bin/foo-gcc}. | |
627 | ||
628 | @item --program-suffix=@var{suffix} | |
629 | Appends @var{suffix} to the names of programs to install in @var{bindir} | |
630 | (see above). For example, specifying @option{--program-suffix=-3.1} | |
631 | would result in @samp{gcc} being installed as | |
632 | @file{/usr/local/bin/gcc-3.1}. | |
633 | ||
634 | @item --program-transform-name=@var{pattern} | |
635 | Applies the @samp{sed} script @var{pattern} to be applied to the names | |
636 | of programs to install in @var{bindir} (see above). @var{pattern} has to | |
637 | consist of one or more basic @samp{sed} editing commands, separated by | |
638 | semicolons. For example, if you want the @samp{gcc} program name to be | |
639 | transformed to the installed program @file{/usr/local/bin/myowngcc} and | |
640 | the @samp{g++} program name to be transformed to | |
641 | @file{/usr/local/bin/gspecial++} without changing other program names, | |
642 | you could use the pattern | |
643 | @option{--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'} | |
644 | to achieve this effect. | |
645 | ||
646 | All three options can be combined and used together, resulting in more | |
647 | complex conversion patterns. As a basic rule, @var{prefix} (and | |
648 | @var{suffix}) are prepended (appended) before further transformations | |
649 | can happen with a special transformation script @var{pattern}. | |
650 | ||
651 | As currently implemented, this option only takes effect for native | |
652 | builds; cross compiler binaries' names are not transformed even when a | |
653 | transformation is explicitly asked for by one of these options. | |
654 | ||
655 | For native builds, some of the installed programs are also installed | |
656 | with the target alias in front of their name, as in | |
657 | @samp{i686-pc-linux-gnu-gcc}. All of the above transformations happen | |
658 | before the target alias is prepended to the name - so, specifying | |
659 | @option{--program-prefix=foo-} and @option{program-suffix=-3.1}, the | |
660 | resulting binary would be installed as | |
661 | @file{/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1}. | |
662 | ||
663 | As a last shortcoming, none of the installed Ada programs are | |
664 | transformed yet, which will be fixed in some time. | |
665 | ||
666 | @item --with-local-prefix=@var{dirname} | |
667 | Specify the | |
668 | installation directory for local include files. The default is | |
669 | @file{/usr/local}. Specify this option if you want the compiler to | |
670 | search directory @file{@var{dirname}/include} for locally installed | |
671 | header files @emph{instead} of @file{/usr/local/include}. | |
672 | ||
673 | You should specify @option{--with-local-prefix} @strong{only} if your | |
674 | site has a different convention (not @file{/usr/local}) for where to put | |
675 | site-specific files. | |
676 | ||
677 | The default value for @option{--with-local-prefix} is @file{/usr/local} | |
678 | regardless of the value of @option{--prefix}. Specifying | |
679 | @option{--prefix} has no effect on which directory GCC searches for | |
680 | local header files. This may seem counterintuitive, but actually it is | |
681 | logical. | |
682 | ||
683 | The purpose of @option{--prefix} is to specify where to @emph{install | |
684 | GCC}. The local header files in @file{/usr/local/include}---if you put | |
685 | any in that directory---are not part of GCC@. They are part of other | |
686 | programs---perhaps many others. (GCC installs its own header files in | |
687 | another directory which is based on the @option{--prefix} value.) | |
688 | ||
689 | Both the local-prefix include directory and the GCC-prefix include | |
690 | directory are part of GCC's "system include" directories. Although these | |
691 | two directories are not fixed, they need to be searched in the proper | |
692 | order for the correct processing of the include_next directive. The | |
693 | local-prefix include directory is searched before the GCC-prefix | |
694 | include directory. Another characteristic of system include directories | |
695 | is that pedantic warnings are turned off for headers in these directories. | |
696 | ||
697 | Some autoconf macros add @option{-I @var{directory}} options to the | |
698 | compiler command line, to ensure that directories containing installed | |
699 | packages' headers are searched. When @var{directory} is one of GCC's | |
700 | system include directories, GCC will ignore the option so that system | |
701 | directories continue to be processed in the correct order. This | |
702 | may result in a search order different from what was specified but the | |
703 | directory will still be searched. | |
704 | ||
705 | GCC automatically searches for ordinary libraries using | |
706 | @env{GCC_EXEC_PREFIX}. Thus, when the same installation prefix is | |
707 | used for both GCC and packages, GCC will automatically search for | |
708 | both headers and libraries. This provides a configuration that is | |
709 | easy to use. GCC behaves in a manner similar to that when it is | |
710 | installed as a system compiler in @file{/usr}. | |
711 | ||
712 | Sites that need to install multiple versions of GCC may not want to | |
713 | use the above simple configuration. It is possible to use the | |
714 | @option{--program-prefix}, @option{--program-suffix} and | |
715 | @option{--program-transform-name} options to install multiple versions | |
716 | into a single directory, but it may be simpler to use different prefixes | |
717 | and the @option{--with-local-prefix} option to specify the location of the | |
718 | site-specific files for each version. It will then be necessary for | |
719 | users to specify explicitly the location of local site libraries | |
720 | (e.g., with @env{LIBRARY_PATH}). | |
721 | ||
722 | The same value can be used for both @option{--with-local-prefix} and | |
723 | @option{--prefix} provided it is not @file{/usr}. This can be used | |
724 | to avoid the default search of @file{/usr/local/include}. | |
725 | ||
726 | @strong{Do not} specify @file{/usr} as the @option{--with-local-prefix}! | |
727 | The directory you use for @option{--with-local-prefix} @strong{must not} | |
728 | contain any of the system's standard header files. If it did contain | |
729 | them, certain programs would be miscompiled (including GNU Emacs, on | |
730 | certain targets), because this would override and nullify the header | |
731 | file corrections made by the @command{fixincludes} script. | |
732 | ||
733 | Indications are that people who use this option use it based on mistaken | |
734 | ideas of what it is for. People use it as if it specified where to | |
735 | install part of GCC@. Perhaps they make this assumption because | |
736 | installing GCC creates the directory. | |
737 | ||
738 | @item --enable-shared[=@var{package}[,@dots{}]] | |
739 | Build shared versions of libraries, if shared libraries are supported on | |
740 | the target platform. Unlike GCC 2.95.x and earlier, shared libraries | |
741 | are enabled by default on all platforms that support shared libraries, | |
742 | except for @samp{libobjc} which is built as a static library only by | |
743 | default. | |
744 | ||
745 | If a list of packages is given as an argument, build shared libraries | |
746 | only for the listed packages. For other packages, only static libraries | |
747 | will be built. Package names currently recognized in the GCC tree are | |
748 | @samp{libgcc} (also known as @samp{gcc}), @samp{libstdc++} (not | |
749 | @samp{libstdc++-v3}), @samp{libffi}, @samp{zlib}, @samp{boehm-gc} and | |
750 | @samp{libjava}. Note that @samp{libobjc} does not recognize itself by | |
751 | any name, so, if you list package names in @option{--enable-shared}, | |
752 | you will only get static Objective-C libraries. @samp{libf2c} and | |
753 | @samp{libiberty} do not support shared libraries at all. | |
754 | ||
755 | Use @option{--disable-shared} to build only static libraries. Note that | |
756 | @option{--disable-shared} does not accept a list of package names as | |
757 | argument, only @option{--enable-shared} does. | |
758 | ||
759 | @item @anchor{with-gnu-as}--with-gnu-as | |
760 | Specify that the compiler should assume that the | |
761 | assembler it finds is the GNU assembler. However, this does not modify | |
762 | the rules to find an assembler and will result in confusion if the | |
763 | assembler found is not actually the GNU assembler. (Confusion may also | |
764 | result if the compiler finds the GNU assembler but has not been | |
765 | configured with @option{--with-gnu-as}.) If you have more than one | |
766 | assembler installed on your system, you may want to use this option in | |
767 | connection with @option{--with-as=@var{pathname}}. | |
768 | ||
769 | The following systems are the only ones where it makes a difference | |
770 | whether you use the GNU assembler. On any other system, | |
771 | @option{--with-gnu-as} has no effect. | |
772 | ||
773 | @itemize @bullet | |
774 | @item @samp{hppa1.0-@var{any}-@var{any}} | |
775 | @item @samp{hppa1.1-@var{any}-@var{any}} | |
776 | @item @samp{i386-@var{any}-sysv} | |
777 | @item @samp{m68k-bull-sysv} | |
778 | @item @samp{m68k-hp-hpux} | |
779 | @item @samp{m68000-hp-hpux} | |
780 | @item @samp{m68000-att-sysv} | |
781 | @item @samp{@var{any}-lynx-lynxos} | |
782 | @item @samp{mips-@var{any}} | |
783 | @item @samp{sparc-sun-solaris2.@var{any}} | |
784 | @item @samp{sparc64-@var{any}-solaris2.@var{any}} | |
785 | @end itemize | |
786 | ||
787 | On the systems listed above (except for the HP-PA, the SPARC, for ISC on | |
788 | the 386, and for @samp{mips-sgi-irix5.*}), if you use the GNU assembler, | |
789 | you should also use the GNU linker (and specify @option{--with-gnu-ld}). | |
790 | ||
791 | @item @anchor{with-as}--with-as=@var{pathname} | |
792 | Specify that the | |
793 | compiler should use the assembler pointed to by @var{pathname}, rather | |
794 | than the one found by the standard rules to find an assembler, which | |
795 | are: | |
796 | @itemize @bullet | |
797 | @item | |
798 | Check the @file{@var{libexec}/gcc/@var{target}/@var{version}} | |
799 | directory, where @var{libexec} defaults to | |
800 | @file{@var{exec-prefix}/libexec} and @var{exec-prefix} defaults to | |
801 | @var{prefix} which defaults to @file{/usr/local} unless overridden by | |
802 | the @option{--prefix=@var{pathname}} switch described | |
803 | above. @var{target} is the target system triple, such as | |
804 | @samp{sparc-sun-solaris2.7}, and @var{version} denotes the GCC | |
805 | version, such as 3.0. | |
806 | @item | |
807 | Check operating system specific directories (e.g.@: @file{/usr/ccs/bin} on | |
808 | Sun Solaris 2). | |
809 | @end itemize | |
810 | Note that these rules do not check for the value of @env{PATH}. You may | |
811 | want to use @option{--with-as} if no assembler is installed in the | |
812 | directories listed above, or if you have multiple assemblers installed | |
813 | and want to choose one that is not found by the above rules. | |
814 | ||
815 | @item @anchor{with-gnu-ld}--with-gnu-ld | |
816 | Same as @uref{#with-gnu-as,,@option{--with-gnu-as}} | |
817 | but for the linker. | |
818 | ||
819 | @item --with-ld=@var{pathname} | |
820 | Same as @uref{#with-as,,@option{--with-as}} | |
821 | but for the linker. | |
822 | ||
823 | @item --with-stabs | |
824 | Specify that stabs debugging | |
825 | information should be used instead of whatever format the host normally | |
826 | uses. Normally GCC uses the same debug format as the host system. | |
827 | ||
828 | On MIPS based systems and on Alphas, you must specify whether you want | |
829 | GCC to create the normal ECOFF debugging format, or to use BSD-style | |
830 | stabs passed through the ECOFF symbol table. The normal ECOFF debug | |
831 | format cannot fully handle languages other than C@. BSD stabs format can | |
832 | handle other languages, but it only works with the GNU debugger GDB@. | |
833 | ||
834 | Normally, GCC uses the ECOFF debugging format by default; if you | |
835 | prefer BSD stabs, specify @option{--with-stabs} when you configure GCC@. | |
836 | ||
837 | No matter which default you choose when you configure GCC, the user | |
838 | can use the @option{-gcoff} and @option{-gstabs+} options to specify explicitly | |
839 | the debug format for a particular compilation. | |
840 | ||
841 | @option{--with-stabs} is meaningful on the ISC system on the 386, also, if | |
842 | @option{--with-gas} is used. It selects use of stabs debugging | |
843 | information embedded in COFF output. This kind of debugging information | |
844 | supports C++ well; ordinary COFF debugging information does not. | |
845 | ||
846 | @option{--with-stabs} is also meaningful on 386 systems running SVR4. It | |
847 | selects use of stabs debugging information embedded in ELF output. The | |
848 | C++ compiler currently (2.6.0) does not support the DWARF debugging | |
849 | information normally used on 386 SVR4 platforms; stabs provide a | |
850 | workable alternative. This requires gas and gdb, as the normal SVR4 | |
851 | tools can not generate or interpret stabs. | |
852 | ||
853 | @item --disable-multilib | |
854 | Specify that multiple target | |
855 | libraries to support different target variants, calling | |
856 | conventions, etc should not be built. The default is to build a | |
857 | predefined set of them. | |
858 | ||
859 | Some targets provide finer-grained control over which multilibs are built | |
860 | (e.g., @option{--disable-softfloat}): | |
861 | @table @code | |
862 | @item arc-*-elf* | |
863 | biendian. | |
864 | ||
865 | @item arm-*-* | |
866 | fpu, 26bit, underscore, interwork, biendian, nofmult. | |
867 | ||
868 | @item m68*-*-* | |
869 | softfloat, m68881, m68000, m68020. | |
870 | ||
871 | @item mips*-*-* | |
872 | single-float, biendian, softfloat. | |
873 | ||
874 | @item powerpc*-*-*, rs6000*-*-* | |
875 | aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian, | |
876 | sysv, aix. | |
877 | ||
878 | @end table | |
879 | ||
880 | @item --enable-threads | |
881 | Specify that the target | |
882 | supports threads. This affects the Objective-C compiler and runtime | |
883 | library, and exception handling for other languages like C++ and Java. | |
884 | On some systems, this is the default. | |
885 | ||
886 | In general, the best (and, in many cases, the only known) threading | |
887 | model available will be configured for use. Beware that on some | |
888 | systems, gcc has not been taught what threading models are generally | |
889 | available for the system. In this case, @option{--enable-threads} is an | |
890 | alias for @option{--enable-threads=single}. | |
891 | ||
892 | @item --disable-threads | |
893 | Specify that threading support should be disabled for the system. | |
894 | This is an alias for @option{--enable-threads=single}. | |
895 | ||
896 | @item --enable-threads=@var{lib} | |
897 | Specify that | |
898 | @var{lib} is the thread support library. This affects the Objective-C | |
899 | compiler and runtime library, and exception handling for other languages | |
900 | like C++ and Java. The possibilities for @var{lib} are: | |
901 | ||
902 | @table @code | |
903 | @item aix | |
904 | AIX thread support. | |
905 | @item dce | |
906 | DCE thread support. | |
907 | @item gnat | |
908 | Ada tasking support. For non-Ada programs, this setting is equivalent | |
909 | to @samp{single}. When used in conjunction with the Ada run time, it | |
910 | causes GCC to use the same thread primitives as Ada uses. This option | |
911 | is necessary when using both Ada and the back end exception handling, | |
912 | which is the default for most Ada targets. | |
913 | @item mach | |
914 | Generic MACH thread support, known to work on NeXTSTEP@. (Please note | |
915 | that the file needed to support this configuration, @file{gthr-mach.h}, is | |
916 | missing and thus this setting will cause a known bootstrap failure.) | |
917 | @item no | |
918 | This is an alias for @samp{single}. | |
919 | @item posix | |
920 | Generic POSIX thread support. | |
921 | @item rtems | |
922 | RTEMS thread support. | |
923 | @item single | |
924 | Disable thread support, should work for all platforms. | |
925 | @item solaris | |
926 | Sun Solaris 2 thread support. | |
927 | @item vxworks | |
928 | VxWorks thread support. | |
929 | @item win32 | |
930 | Microsoft Win32 API thread support. | |
931 | @end table | |
932 | ||
933 | @item --with-cpu=@var{cpu} | |
934 | Specify which cpu variant the compiler should generate code for by default. | |
935 | @var{cpu} will be used as the default value of the @option{-mcpu=} switch. | |
936 | This option is only supported on some targets, including ARM, i386, PowerPC, | |
937 | and SPARC@. | |
938 | ||
939 | @item --with-schedule=@var{cpu} | |
940 | @itemx --with-arch=@var{cpu} | |
941 | @itemx --with-tune=@var{cpu} | |
942 | @itemx --with-abi=@var{abi} | |
943 | @itemx --with-fpu=@var{type} | |
944 | @itemx --with-float=@var{type} | |
945 | These configure options provide default values for the @option{-mschedule=}, | |
946 | @option{-march=}, @option{-mtune=}, @option{-mabi=}, and @option{-mfpu=} | |
947 | options and for @option{-mhard-float} or @option{-msoft-float}. As with | |
948 | @option{--with-cpu}, which switches will be accepted and acceptable values | |
949 | of the arguments depend on the target. | |
950 | ||
951 | @item --enable-altivec | |
952 | Specify that the target supports AltiVec vector enhancements. This | |
953 | option will adjust the ABI for AltiVec enhancements, as well as generate | |
954 | AltiVec code when appropriate. This option is only available for | |
955 | PowerPC systems. | |
956 | ||
957 | @item --enable-__cxa_atexit | |
958 | Define if you want to use __cxa_atexit, rather than atexit, to | |
959 | register C++ destructors for local statics and global objects. | |
960 | This is essential for fully standards-compliant handling of | |
961 | destructors, but requires __cxa_atexit in libc. This option is currently | |
962 | only available on systems with GNU libc. When enabled, this will cause | |
963 | @option{-fuse-cxa-exit} to be passed by default. | |
964 | ||
965 | @item --enable-target-optspace | |
966 | Specify that target | |
967 | libraries should be optimized for code space instead of code speed. | |
968 | This is the default for the m32r platform. | |
969 | ||
970 | @item --disable-cpp | |
971 | Specify that a user visible @command{cpp} program should not be installed. | |
972 | ||
973 | @item --with-cpp-install-dir=@var{dirname} | |
974 | Specify that the user visible @command{cpp} program should be installed | |
975 | in @file{@var{prefix}/@var{dirname}/cpp}, in addition to @var{bindir}. | |
976 | ||
977 | @item --enable-initfini-array | |
978 | Force the use of sections @code{.init_array} and @code{.fini_array} | |
979 | (instead of @code{.init} and @code{.fini}) for constructors and | |
980 | destructors. Option @option{--disable-initfini-array} has the | |
981 | opposite effect. If neither option is specified, the configure script | |
982 | will try to guess whether the @code{.init_array} and | |
983 | @code{.fini_array} sections are supported and, if they are, use them. | |
984 | ||
985 | @item --enable-maintainer-mode | |
986 | The build rules that | |
987 | regenerate the GCC master message catalog @file{gcc.pot} are normally | |
988 | disabled. This is because it can only be rebuilt if the complete source | |
989 | tree is present. If you have changed the sources and want to rebuild the | |
990 | catalog, configuring with @option{--enable-maintainer-mode} will enable | |
991 | this. Note that you need a recent version of the @code{gettext} tools | |
992 | to do so. | |
993 | ||
994 | @item --enable-generated-files-in-srcdir | |
995 | Neither the .c and .h files that are generated from bison and flex nor the | |
996 | info manuals and man pages that are built from the .texi files are present | |
997 | in the CVS development tree. When building GCC from that development tree, | |
998 | or from a snapshot which are created from CVS, then those generated files | |
999 | are placed in your build directory, which allows for the source to be in a | |
1000 | readonly directory. | |
1001 | ||
1002 | If you configure with @option{--enable-generated-files-in-srcdir} then those | |
1003 | generated files will go into the source directory. This is mainly intended | |
1004 | for generating release or prerelease tarballs of the GCC sources, since it | |
1005 | is not a requirement that the users of source releases to have flex, bison, or | |
1006 | makeinfo. | |
1007 | ||
1008 | @item --enable-version-specific-runtime-libs | |
1009 | Specify | |
1010 | that runtime libraries should be installed in the compiler specific | |
1011 | subdirectory (@file{@var{libdir}/gcc}) rather than the usual places. In | |
1012 | addition, @samp{libstdc++}'s include files will be installed into | |
1013 | @file{@var{libdir}} unless you overruled it by using | |
1014 | @option{--with-gxx-include-dir=@var{dirname}}. Using this option is | |
1015 | particularly useful if you intend to use several versions of GCC in | |
1016 | parallel. This is currently supported by @samp{libf2c} and | |
1017 | @samp{libstdc++}, and is the default for @samp{libobjc} which cannot be | |
1018 | changed in this case. | |
1019 | ||
1020 | @item --enable-languages=@var{lang1},@var{lang2},@dots{} | |
1021 | Specify that only a particular subset of compilers and | |
1022 | their runtime libraries should be built. For a list of valid values for | |
1023 | @var{langN} you can issue the following command in the | |
1024 | @file{gcc} directory of your GCC source tree:@* | |
1025 | @smallexample | |
1026 | grep language= */config-lang.in | |
1027 | @end smallexample | |
1028 | Currently, you can use any of the following: | |
1029 | @code{ada}, @code{c}, @code{c++}, @code{f77}, @code{java}, @code{objc}. | |
1030 | Building the Ada compiler has special requirements, see below.@* | |
1031 | If you do not pass this flag, all languages available in the @file{gcc} | |
1032 | sub-tree will be configured. Re-defining @code{LANGUAGES} when calling | |
1033 | @samp{make bootstrap} @strong{does not} work anymore, as those | |
1034 | language sub-directories might not have been configured! | |
1035 | ||
1036 | @item --disable-libada | |
1037 | Specify that the run-time libraries and tools used by GNAT should not | |
1038 | be built. This can be useful for debugging, or for compatibility with | |
1039 | previous Ada build procedures, when it was required to explicitly | |
1040 | do a @samp{make -C gcc gnatlib_and_tools}. | |
1041 | ||
1042 | @item --with-dwarf2 | |
1043 | Specify that the compiler should | |
1044 | use DWARF 2 debugging information as the default. | |
1045 | ||
1046 | @item --enable-win32-registry | |
1047 | @itemx --enable-win32-registry=@var{key} | |
1048 | @itemx --disable-win32-registry | |
1049 | The @option{--enable-win32-registry} option enables Microsoft Windows-hosted GCC | |
1050 | to look up installations paths in the registry using the following key: | |
1051 | ||
1052 | @smallexample | |
1053 | @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\@var{key}} | |
1054 | @end smallexample | |
1055 | ||
1056 | @var{key} defaults to GCC version number, and can be overridden by the | |
1057 | @option{--enable-win32-registry=@var{key}} option. Vendors and distributors | |
1058 | who use custom installers are encouraged to provide a different key, | |
1059 | perhaps one comprised of vendor name and GCC version number, to | |
1060 | avoid conflict with existing installations. This feature is enabled | |
1061 | by default, and can be disabled by @option{--disable-win32-registry} | |
1062 | option. This option has no effect on the other hosts. | |
1063 | ||
1064 | @item --nfp | |
1065 | Specify that the machine does not have a floating point unit. This | |
1066 | option only applies to @samp{m68k-sun-sunos@var{n}}. On any other | |
1067 | system, @option{--nfp} has no effect. | |
1068 | ||
1069 | @item --enable-werror | |
1070 | @itemx --disable-werror | |
1071 | @itemx --enable-werror=yes | |
1072 | @itemx --enable-werror=no | |
1073 | When you specify this option, it controls whether certain files in the | |
1074 | compiler are built with @option{-Werror} in bootstrap stage2 and later. | |
1075 | If you don't specify it, @option{-Werror} is turned on for the main | |
1076 | development trunk. However it defaults to off for release branches and | |
1077 | final releases. The specific files which get @option{-Werror} are | |
1078 | controlled by the Makefiles. | |
1079 | ||
1080 | @item --enable-checking | |
1081 | @itemx --enable-checking=@var{list} | |
1082 | When you specify this option, the compiler is built to perform checking | |
1083 | of tree node types when referencing fields of that node, and some other | |
1084 | internal consistency checks. This does not change the generated code, | |
1085 | but adds error checking within the compiler. This will slow down the | |
1086 | compiler and may only work properly if you are building the compiler | |
1087 | with GCC@. This is on by default when building from CVS or snapshots, | |
1088 | but off for releases. More control over the checks may be had by | |
1089 | specifying @var{list}; the categories of checks available are | |
1090 | @samp{misc}, @samp{tree}, @samp{gc}, @samp{rtl}, @samp{rtlflag}, | |
1091 | @samp{fold}, @samp{gcac} and @samp{valgrind}. The check @samp{valgrind} | |
1092 | requires the external @command{valgrind} simulator, available from | |
1093 | @uref{http://valgrind.kde.org/}. The default when @var{list} is | |
1094 | not specified is @samp{misc,tree,gc,rtlflag}; the checks @samp{rtl}, | |
1095 | @samp{gcac} and @samp{valgrind} are very expensive. | |
1096 | ||
1097 | @item --enable-coverage | |
1098 | @itemx --enable-coverage=@var{level} | |
1099 | With this option, the compiler is built to collect self coverage | |
1100 | information, every time it is run. This is for internal development | |
1101 | purposes, and only works when the compiler is being built with gcc. The | |
1102 | @var{level} argument controls whether the compiler is built optimized or | |
1103 | not, values are @samp{opt} and @samp{noopt}. For coverage analysis you | |
1104 | want to disable optimization, for performance analysis you want to | |
1105 | enable optimization. When coverage is enabled, the default level is | |
1106 | without optimization. | |
1107 | ||
1108 | @item --enable-gather-detailed-mem-stats | |
1109 | When this option is specified more detailed information on memory | |
1110 | allocation is gathered. This information is printed when using | |
1111 | @option{-fmem-report}. | |
1112 | ||
1113 | @item --enable-nls | |
1114 | @itemx --disable-nls | |
1115 | The @option{--enable-nls} option enables Native Language Support (NLS), | |
1116 | which lets GCC output diagnostics in languages other than American | |
1117 | English. Native Language Support is enabled by default if not doing a | |
1118 | canadian cross build. The @option{--disable-nls} option disables NLS@. | |
1119 | ||
1120 | @item --with-included-gettext | |
1121 | If NLS is enabled, the @option{--with-included-gettext} option causes the build | |
1122 | procedure to prefer its copy of GNU @command{gettext}. | |
1123 | ||
1124 | @item --with-catgets | |
1125 | If NLS is enabled, and if the host lacks @code{gettext} but has the | |
1126 | inferior @code{catgets} interface, the GCC build procedure normally | |
1127 | ignores @code{catgets} and instead uses GCC's copy of the GNU | |
1128 | @code{gettext} library. The @option{--with-catgets} option causes the | |
1129 | build procedure to use the host's @code{catgets} in this situation. | |
1130 | ||
1131 | @item --with-libiconv-prefix=@var{dir} | |
1132 | Search for libiconv header files in @file{@var{dir}/include} and | |
1133 | libiconv library files in @file{@var{dir}/lib}. | |
1134 | ||
1135 | @item --enable-obsolete | |
1136 | Enable configuration for an obsoleted system. If you attempt to | |
1137 | configure GCC for a system (build, host, or target) which has been | |
1138 | obsoleted, and you do not specify this flag, configure will halt with an | |
1139 | error message. | |
1140 | ||
1141 | All support for systems which have been obsoleted in one release of GCC | |
1142 | is removed entirely in the next major release, unless someone steps | |
1143 | forward to maintain the port. | |
1144 | @end table | |
1145 | ||
1146 | @subheading Cross-Compiler-Specific Options | |
1147 | The following options only apply to building cross compilers. | |
1148 | @table @code | |
1149 | @item --with-sysroot | |
1150 | @itemx --with-sysroot=@var{dir} | |
1151 | Tells GCC to consider @var{dir} as the root of a tree that contains a | |
1152 | (subset of) the root filesystem of the target operating system. | |
1153 | Target system headers, libraries and run-time object files will be | |
1154 | searched in there. The specified directory is not copied into the | |
1155 | install tree, unlike the options @option{--with-headers} and | |
1156 | @option{--with-libs} that this option obsoletes. The default value, | |
1157 | in case @option{--with-sysroot} is not given an argument, is | |
1158 | @option{$@{gcc_tooldir@}/sys-root}. If the specified directory is a | |
1159 | subdirectory of @option{$@{exec_prefix@}}, then it will be found relative to | |
1160 | the GCC binaries if the installation tree is moved. | |
1161 | ||
1162 | @item --with-headers | |
1163 | @itemx --with-headers=@var{dir} | |
1164 | Deprecated in favor of @option{--with-sysroot}. | |
1165 | Specifies that target headers are available when building a cross compiler. | |
1166 | The @var{dir} argument specifies a directory which has the target include | |
1167 | files. These include files will be copied into the @file{gcc} install | |
1168 | directory. @emph{This option with the @var{dir} argument is required} when | |
1169 | building a cross compiler, if @file{@var{prefix}/@var{target}/sys-include} | |
1170 | doesn't pre-exist. If @file{@var{prefix}/@var{target}/sys-include} does | |
1171 | pre-exist, the @var{dir} argument may be omitted. @command{fixincludes} | |
1172 | will be run on these files to make them compatible with GCC. | |
1173 | ||
1174 | @item --without-headers | |
1175 | Tells GCC not use any target headers from a libc when building a cross | |
1176 | compiler. When crossing to GNU/Linux, you need the headers so gcc | |
1177 | can build the exception handling for libgcc. | |
1178 | See @uref{http://www.objsw.com/CrossGCC/,,CrossGCC} for more information | |
1179 | on this option. | |
1180 | ||
1181 | @item --with-libs | |
1182 | @itemx --with-libs=``@var{dir1} @var{dir2} @dots{} @var{dirN}'' | |
1183 | Deprecated in favor of @option{--with-sysroot}. | |
1184 | Specifies a list of directories which contain the target runtime | |
1185 | libraries. These libraries will be copied into the @file{gcc} install | |
1186 | directory. If the directory list is omitted, this option has no | |
1187 | effect. | |
1188 | @item --with-newlib | |
1189 | Specifies that @samp{newlib} is | |
1190 | being used as the target C library. This causes @code{__eprintf} to be | |
1191 | omitted from @file{libgcc.a} on the assumption that it will be provided by | |
1192 | @samp{newlib}. | |
1193 | @end table | |
1194 | ||
1195 | @subheading Java-Specific Options | |
1196 | ||
1197 | The following option applies to the build of the Java front end. | |
1198 | ||
1199 | @table @code | |
1200 | @item --disable-libgcj | |
1201 | Specify that the run-time libraries | |
1202 | used by GCJ should not be built. This is useful in case you intend | |
1203 | to use GCJ with some other run-time, or you're going to install it | |
1204 | separately, or it just happens not to build on your particular | |
1205 | machine. In general, if the Java front end is enabled, the GCJ | |
1206 | libraries will be enabled too, unless they're known to not work on | |
1207 | the target platform. If GCJ is enabled but @samp{libgcj} isn't built, you | |
1208 | may need to port it; in this case, before modifying the top-level | |
1209 | @file{configure.in} so that @samp{libgcj} is enabled by default on this platform, | |
1210 | you may use @option{--enable-libgcj} to override the default. | |
1211 | ||
1212 | @end table | |
1213 | ||
1214 | The following options apply to building @samp{libgcj}. | |
1215 | ||
1216 | @subsubheading General Options | |
1217 | ||
1218 | @table @code | |
1219 | @item --disable-getenv-properties | |
1220 | Don't set system properties from @env{GCJ_PROPERTIES}. | |
1221 | ||
1222 | @item --enable-hash-synchronization | |
1223 | Use a global hash table for monitor locks. Ordinarily, | |
1224 | @samp{libgcj}'s @samp{configure} script automatically makes | |
1225 | the correct choice for this option for your platform. Only use | |
1226 | this if you know you need the library to be configured differently. | |
1227 | ||
1228 | @item --enable-interpreter | |
1229 | Enable the Java interpreter. The interpreter is automatically | |
1230 | enabled by default on all platforms that support it. This option | |
1231 | is really only useful if you want to disable the interpreter | |
1232 | (using @option{--disable-interpreter}). | |
1233 | ||
1234 | @item --disable-java-net | |
1235 | Disable java.net. This disables the native part of java.net only, | |
1236 | using non-functional stubs for native method implementations. | |
1237 | ||
1238 | @item --disable-jvmpi | |
1239 | Disable JVMPI support. | |
1240 | ||
1241 | @item --with-ecos | |
1242 | Enable runtime eCos target support. | |
1243 | ||
1244 | @item --without-libffi | |
1245 | Don't use @samp{libffi}. This will disable the interpreter and JNI | |
1246 | support as well, as these require @samp{libffi} to work. | |
1247 | ||
1248 | @item --enable-libgcj-debug | |
1249 | Enable runtime debugging code. | |
1250 | ||
1251 | @item --enable-libgcj-multifile | |
1252 | If specified, causes all @file{.java} source files to be | |
1253 | compiled into @file{.class} files in one invocation of | |
1254 | @samp{gcj}. This can speed up build time, but is more | |
1255 | resource-intensive. If this option is unspecified or | |
1256 | disabled, @samp{gcj} is invoked once for each @file{.java} | |
1257 | file to compile into a @file{.class} file. | |
1258 | ||
1259 | @item --with-libiconv-prefix=DIR | |
1260 | Search for libiconv in @file{DIR/include} and @file{DIR/lib}. | |
1261 | ||
1262 | @item --enable-sjlj-exceptions | |
1263 | Force use of @code{builtin_setjmp} for exceptions. @samp{configure} | |
1264 | ordinarily picks the correct value based on the platform. Only use | |
1265 | this option if you are sure you need a different setting. | |
1266 | ||
1267 | @item --with-system-zlib | |
1268 | Use installed @samp{zlib} rather than that included with GCC@. | |
1269 | ||
1270 | @item --with-win32-nlsapi=ansi, unicows or unicode | |
1271 | Indicates how MinGW @samp{libgcj} translates between UNICODE | |
1272 | characters and the Win32 API. | |
1273 | @table @code | |
1274 | @item ansi | |
1275 | Use the single-byte @code{char} and the Win32 A functions natively, | |
1276 | translating to and from UNICODE when using these functions. If | |
1277 | unspecified, this is the default. | |
1278 | ||
1279 | @item unicows | |
1280 | Use the @code{WCHAR} and Win32 W functions natively. Adds | |
1281 | @code{-lunicows} to @file{libgcj.spec} to link with @samp{libunicows}. | |
1282 | @file{unicows.dll} needs to be deployed on Microsoft Windows 9X machines | |
1283 | running built executables. @file{libunicows.a}, an open-source | |
1284 | import library around Microsoft's @code{unicows.dll}, is obtained from | |
1285 | @uref{http://libunicows.sourceforge.net/}, which also gives details | |
1286 | on getting @file{unicows.dll} from Microsoft. | |
1287 | ||
1288 | @item unicode | |
1289 | Use the @code{WCHAR} and Win32 W functions natively. Does @emph{not} | |
1290 | add @code{-lunicows} to @file{libgcj.spec}. The built executables will | |
1291 | only run on Microsoft Windows NT and above. | |
1292 | @end table | |
1293 | @end table | |
1294 | ||
1295 | @subsubheading AWT-Specific Options | |
1296 | ||
1297 | @table @code | |
1298 | @item --with-x | |
1299 | Use the X Window System. | |
1300 | ||
1301 | @item --enable-java-awt=PEER(S) | |
1302 | Specifies the AWT peer library or libraries to build alongside | |
1303 | @samp{libgcj}. If this option is unspecified or disabled, AWT | |
1304 | will be non-functional. Current valid values are @option{gtk} and | |
1305 | @option{xlib}. Multiple libraries should be separated by a | |
1306 | comma (i.e. @option{--enable-java-awt=gtk,xlib}). | |
1307 | ||
1308 | @item --enable-gtk-cairo | |
1309 | Build the cairo Graphics2D implementation on GTK. | |
1310 | ||
1311 | @item --enable-java-gc=TYPE | |
1312 | Choose garbage collector. Defaults to @option{boehm} if unspecified. | |
1313 | ||
1314 | @item --disable-gtktest | |
1315 | Do not try to compile and run a test GTK+ program. | |
1316 | ||
1317 | @item --disable-glibtest | |
1318 | Do not try to compile and run a test GLIB program. | |
1319 | ||
1320 | @item --with-libart-prefix=PFX | |
1321 | Prefix where libart is installed (optional). | |
1322 | ||
1323 | @item --with-libart-exec-prefix=PFX | |
1324 | Exec prefix where libart is installed (optional). | |
1325 | ||
1326 | @item --disable-libarttest | |
1327 | Do not try to compile and run a test libart program. | |
1328 | ||
1329 | @end table | |
1330 | ||
1331 | @html | |
1332 | <hr /> | |
1333 | <p> | |
1334 | @end html | |
1335 | @ifhtml | |
1336 | @uref{./index.html,,Return to the GCC Installation page} | |
1337 | @end ifhtml | |
1338 | @end ifset | |
1339 | ||
1340 | @c ***Building**************************************************************** | |
1341 | @ifnothtml | |
1342 | @comment node-name, next, previous, up | |
1343 | @node Building, Testing, Configuration, Installing GCC | |
1344 | @end ifnothtml | |
1345 | @ifset buildhtml | |
1346 | @ifnothtml | |
1347 | @chapter Building | |
1348 | @end ifnothtml | |
1349 | @cindex Installing GCC: Building | |
1350 | ||
1351 | Now that GCC is configured, you are ready to build the compiler and | |
1352 | runtime libraries. | |
1353 | ||
1354 | We @strong{highly} recommend that GCC be built using GNU make; | |
1355 | other versions may work, then again they might not. | |
1356 | GNU make is required for compiling GNAT (the Ada compiler) and the Java | |
1357 | runtime library. | |
1358 | ||
1359 | (For example, many broken versions of make will fail if you use the | |
1360 | recommended setup where @var{objdir} is different from @var{srcdir}. | |
1361 | Other broken versions may recompile parts of the compiler when | |
1362 | installing the compiler.) | |
1363 | ||
1364 | Some commands executed when making the compiler may fail (return a | |
1365 | nonzero status) and be ignored by @command{make}. These failures, which | |
1366 | are often due to files that were not found, are expected, and can safely | |
1367 | be ignored. | |
1368 | ||
1369 | It is normal to have compiler warnings when compiling certain files. | |
1370 | Unless you are a GCC developer, you can generally ignore these warnings | |
1371 | unless they cause compilation to fail. Developers should attempt to fix | |
1372 | any warnings encountered, however they can temporarily continue past | |
1373 | warnings-as-errors by specifying the configure flag | |
1374 | @option{--disable-werror}. | |
1375 | ||
1376 | On certain old systems, defining certain environment variables such as | |
1377 | @env{CC} can interfere with the functioning of @command{make}. | |
1378 | ||
1379 | If you encounter seemingly strange errors when trying to build the | |
1380 | compiler in a directory other than the source directory, it could be | |
1381 | because you have previously configured the compiler in the source | |
1382 | directory. Make sure you have done all the necessary preparations. | |
1383 | ||
1384 | If you build GCC on a BSD system using a directory stored in an old System | |
1385 | V file system, problems may occur in running @command{fixincludes} if the | |
1386 | System V file system doesn't support symbolic links. These problems | |
1387 | result in a failure to fix the declaration of @code{size_t} in | |
1388 | @file{sys/types.h}. If you find that @code{size_t} is a signed type and | |
1389 | that type mismatches occur, this could be the cause. | |
1390 | ||
1391 | The solution is not to use such a directory for building GCC@. | |
1392 | ||
1393 | When building from CVS or snapshots, or if you modify parser sources, | |
1394 | you need the Bison parser generator installed. Any version 1.25 or | |
1395 | later should work; older versions may also work. If you do not modify | |
1396 | parser sources, releases contain the Bison-generated files and you do | |
1397 | not need Bison installed to build them. | |
1398 | ||
1399 | When building from CVS or snapshots, or if you modify Texinfo | |
1400 | documentation, you need version 4.2 or later of Texinfo installed if you | |
1401 | want Info documentation to be regenerated. Releases contain Info | |
1402 | documentation pre-built for the unmodified documentation in the release. | |
1403 | ||
1404 | @section Building a native compiler | |
1405 | ||
1406 | For a native build issue the command @samp{make bootstrap}. This | |
1407 | will build the entire GCC system, which includes the following steps: | |
1408 | ||
1409 | @itemize @bullet | |
1410 | @item | |
1411 | Build host tools necessary to build the compiler such as texinfo, bison, | |
1412 | gperf. | |
1413 | ||
1414 | @item | |
1415 | Build target tools for use by the compiler such as binutils (bfd, | |
1416 | binutils, gas, gprof, ld, and opcodes) | |
1417 | if they have been individually linked | |
1418 | or moved into the top level GCC source tree before configuring. | |
1419 | ||
1420 | @item | |
1421 | Perform a 3-stage bootstrap of the compiler. | |
1422 | ||
1423 | @item | |
1424 | Perform a comparison test of the stage2 and stage3 compilers. | |
1425 | ||
1426 | @item | |
1427 | Build runtime libraries using the stage3 compiler from the previous step. | |
1428 | ||
1429 | @end itemize | |
1430 | ||
1431 | If you are short on disk space you might consider @samp{make | |
1432 | bootstrap-lean} instead. This is identical to @samp{make | |
1433 | bootstrap} except that object files from the stage1 and | |
1434 | stage2 of the 3-stage bootstrap of the compiler are deleted as | |
1435 | soon as they are no longer needed. | |
1436 | ||
1437 | If you want to save additional space during the bootstrap and in | |
1438 | the final installation as well, you can build the compiler binaries | |
1439 | without debugging information as in the following example. This will save | |
1440 | roughly 40% of disk space both for the bootstrap and the final installation. | |
1441 | (Libraries will still contain debugging information.) | |
1442 | ||
1443 | @smallexample | |
1444 | make CFLAGS='-O' LIBCFLAGS='-g -O2' \ | |
1445 | LIBCXXFLAGS='-g -O2 -fno-implicit-templates' bootstrap | |
1446 | @end smallexample | |
1447 | ||
1448 | If you wish to use non-default GCC flags when compiling the stage2 and | |
1449 | stage3 compilers, set @code{BOOT_CFLAGS} on the command line when doing | |
1450 | @samp{make bootstrap}. Non-default optimization flags are less well | |
1451 | tested here than the default of @samp{-g -O2}, but should still work. | |
1452 | In a few cases, you may find that you need to specify special flags such | |
1453 | as @option{-msoft-float} here to complete the bootstrap; or, if the | |
1454 | native compiler miscompiles the stage1 compiler, you may need to work | |
1455 | around this, by choosing @code{BOOT_CFLAGS} to avoid the parts of the | |
1456 | stage1 compiler that were miscompiled, or by using @samp{make | |
1457 | bootstrap4} to increase the number of stages of bootstrap. | |
1458 | ||
1459 | If you used the flag @option{--enable-languages=@dots{}} to restrict | |
1460 | the compilers to be built, only those you've actually enabled will be | |
1461 | built. This will of course only build those runtime libraries, for | |
1462 | which the particular compiler has been built. Please note, | |
1463 | that re-defining @env{LANGUAGES} when calling @samp{make bootstrap} | |
1464 | @strong{does not} work anymore! | |
1465 | ||
1466 | If the comparison of stage2 and stage3 fails, this normally indicates | |
1467 | that the stage2 compiler has compiled GCC incorrectly, and is therefore | |
1468 | a potentially serious bug which you should investigate and report. (On | |
1469 | a few systems, meaningful comparison of object files is impossible; they | |
1470 | always appear ``different''. If you encounter this problem, you will | |
1471 | need to disable comparison in the @file{Makefile}.) | |
1472 | ||
1473 | @section Building a cross compiler | |
1474 | ||
1475 | We recommend reading the | |
1476 | @uref{http://www.objsw.com/CrossGCC/,,crossgcc FAQ} | |
1477 | for information about building cross compilers. | |
1478 | ||
1479 | When building a cross compiler, it is not generally possible to do a | |
1480 | 3-stage bootstrap of the compiler. This makes for an interesting problem | |
1481 | as parts of GCC can only be built with GCC@. | |
1482 | ||
1483 | To build a cross compiler, we first recommend building and installing a | |
1484 | native compiler. You can then use the native GCC compiler to build the | |
1485 | cross compiler. The installed native compiler needs to be GCC version | |
1486 | 2.95 or later. | |
1487 | ||
1488 | Assuming you have already installed a native copy of GCC and configured | |
1489 | your cross compiler, issue the command @command{make}, which performs the | |
1490 | following steps: | |
1491 | ||
1492 | @itemize @bullet | |
1493 | @item | |
1494 | Build host tools necessary to build the compiler such as texinfo, bison, | |
1495 | gperf. | |
1496 | ||
1497 | @item | |
1498 | Build target tools for use by the compiler such as binutils (bfd, | |
1499 | binutils, gas, gprof, ld, and opcodes) | |
1500 | if they have been individually linked or moved into the top level GCC source | |
1501 | tree before configuring. | |
1502 | ||
1503 | @item | |
1504 | Build the compiler (single stage only). | |
1505 | ||
1506 | @item | |
1507 | Build runtime libraries using the compiler from the previous step. | |
1508 | @end itemize | |
1509 | ||
1510 | Note that if an error occurs in any step the make process will exit. | |
1511 | ||
1512 | If you are not building GNU binutils in the same source tree as GCC, | |
1513 | you will need a cross-assembler and cross-linker installed before | |
1514 | configuring GCC@. Put them in the directory | |
1515 | @file{@var{prefix}/@var{target}/bin}. Here is a table of the tools | |
1516 | you should put in this directory: | |
1517 | ||
1518 | @table @file | |
1519 | @item as | |
1520 | This should be the cross-assembler. | |
1521 | ||
1522 | @item ld | |
1523 | This should be the cross-linker. | |
1524 | ||
1525 | @item ar | |
1526 | This should be the cross-archiver: a program which can manipulate | |
1527 | archive files (linker libraries) in the target machine's format. | |
1528 | ||
1529 | @item ranlib | |
1530 | This should be a program to construct a symbol table in an archive file. | |
1531 | @end table | |
1532 | ||
1533 | The installation of GCC will find these programs in that directory, | |
1534 | and copy or link them to the proper place to for the cross-compiler to | |
1535 | find them when run later. | |
1536 | ||
1537 | The easiest way to provide these files is to build the Binutils package. | |
1538 | Configure it with the same @option{--host} and @option{--target} | |
1539 | options that you use for configuring GCC, then build and install | |
1540 | them. They install their executables automatically into the proper | |
1541 | directory. Alas, they do not support all the targets that GCC | |
1542 | supports. | |
1543 | ||
1544 | If you are not building a C library in the same source tree as GCC, | |
1545 | you should also provide the target libraries and headers before | |
1546 | configuring GCC, specifying the directories with | |
1547 | @option{--with-sysroot} or @option{--with-headers} and | |
1548 | @option{--with-libs}. Many targets also require ``start files'' such | |
1549 | as @file{crt0.o} and | |
1550 | @file{crtn.o} which are linked into each executable. There may be several | |
1551 | alternatives for @file{crt0.o}, for use with profiling or other | |
1552 | compilation options. Check your target's definition of | |
1553 | @code{STARTFILE_SPEC} to find out what start files it uses. | |
1554 | ||
1555 | @section Building in parallel | |
1556 | ||
1557 | You can use @samp{make bootstrap MAKE="make -j 2" -j 2}, or just | |
1558 | @samp{make -j 2 bootstrap} for GNU Make 3.79 and above, instead of | |
1559 | @samp{make bootstrap} to build GCC in parallel. | |
1560 | You can also specify a bigger number, and in most cases using a value | |
1561 | greater than the number of processors in your machine will result in | |
1562 | fewer and shorter I/O latency hits, thus improving overall throughput; | |
1563 | this is especially true for slow drives and network filesystems. | |
1564 | ||
1565 | @section Building the Ada compiler | |
1566 | ||
1567 | In order to build GNAT, the Ada compiler, you need a working GNAT | |
1568 | compiler (GNAT version 3.14 or later, or GCC version 3.1 or later), | |
1569 | including GNAT tools such as @command{gnatmake} and @command{gnatlink}, | |
1570 | since the Ada front end is written in Ada (with some | |
1571 | GNAT-specific extensions), and GNU make. | |
1572 | ||
1573 | @command{configure} does not test whether the GNAT installation works | |
1574 | and has a sufficiently recent version; if too old a GNAT version is | |
1575 | installed, the build will fail unless @option{--enable-languages} is | |
1576 | used to disable building the Ada front end. | |
1577 | ||
1578 | @section Building with profile feedback | |
1579 | ||
1580 | It is possible to use profile feedback to optimize the compiler itself. This | |
1581 | should result in a faster compiler binary. Experiments done on x86 using gcc | |
1582 | 3.3 showed approximately 7 percent speedup on compiling C programs. To | |
1583 | bootstrap compiler with profile feedback, use @code{make profiledbootstrap}. | |
1584 | ||
1585 | When @samp{make profiledbootstrap} is run, it will first build a @code{stage1} | |
1586 | compiler. This compiler is used to build a @code{stageprofile} compiler | |
1587 | instrumented to collect execution counts of instruction and branch | |
1588 | probabilities. Then runtime libraries are compiled with profile collected. | |
1589 | Finally a @code{stagefeedback} compiler is built using the information collected. | |
1590 | ||
1591 | Unlike @samp{make bootstrap} several additional restrictions apply. The | |
1592 | compiler used to build @code{stage1} needs to support a 64-bit integral type. | |
1593 | It is recommended to only use GCC for this. Also parallel make is currently | |
1594 | not supported since collisions in profile collecting may occur. | |
1595 | ||
1596 | @html | |
1597 | <hr /> | |
1598 | <p> | |
1599 | @end html | |
1600 | @ifhtml | |
1601 | @uref{./index.html,,Return to the GCC Installation page} | |
1602 | @end ifhtml | |
1603 | @end ifset | |
1604 | ||
1605 | @c ***Testing***************************************************************** | |
1606 | @ifnothtml | |
1607 | @comment node-name, next, previous, up | |
1608 | @node Testing, Final install, Building, Installing GCC | |
1609 | @end ifnothtml | |
1610 | @ifset testhtml | |
1611 | @ifnothtml | |
1612 | @chapter Installing GCC: Testing | |
1613 | @end ifnothtml | |
1614 | @cindex Testing | |
1615 | @cindex Installing GCC: Testing | |
1616 | @cindex Testsuite | |
1617 | ||
1618 | Before you install GCC, we encourage you to run the testsuites and to | |
1619 | compare your results with results from a similar configuration that have | |
1620 | been submitted to the | |
1621 | @uref{http://gcc.gnu.org/ml/gcc-testresults/,,gcc-testresults mailing list}. | |
1622 | Some of these archived results are linked from the build status lists | |
1623 | at @uref{http://gcc.gnu.org/buildstat.html}, although not everyone who | |
1624 | reports a successful build runs the testsuites and submits the results. | |
1625 | This step is optional and may require you to download additional software, | |
1626 | but it can give you confidence in your new GCC installation or point out | |
1627 | problems before you install and start using your new GCC. | |
1628 | ||
1629 | First, you must have @uref{download.html,,downloaded the testsuites}. | |
1630 | These are part of the full distribution, but if you downloaded the | |
1631 | ``core'' compiler plus any front ends, you must download the testsuites | |
1632 | separately. | |
1633 | ||
1634 | Second, you must have the testing tools installed. This includes | |
1635 | @uref{http://www.gnu.org/software/dejagnu/,,DejaGnu} 1.4.1 or 1.4.3 | |
1636 | and later, Tcl, and Expect; the DejaGnu site has links to these. | |
1637 | ||
1638 | If the directories where @command{runtest} and @command{expect} were | |
1639 | installed are not in the @env{PATH}, you may need to set the following | |
1640 | environment variables appropriately, as in the following example (which | |
1641 | assumes that DejaGnu has been installed under @file{/usr/local}): | |
1642 | ||
1643 | @smallexample | |
1644 | TCL_LIBRARY = /usr/local/share/tcl8.0 | |
1645 | DEJAGNULIBS = /usr/local/share/dejagnu | |
1646 | @end smallexample | |
1647 | ||
1648 | (On systems such as Cygwin, these paths are required to be actual | |
1649 | paths, not mounts or links; presumably this is due to some lack of | |
1650 | portability in the DejaGnu code.) | |
1651 | ||
1652 | ||
1653 | Finally, you can run the testsuite (which may take a long time): | |
1654 | @smallexample | |
1655 | cd @var{objdir}; make -k check | |
1656 | @end smallexample | |
1657 | ||
1658 | This will test various components of GCC, such as compiler | |
1659 | front ends and runtime libraries. While running the testsuite, DejaGnu | |
1660 | might emit some harmless messages resembling | |
1661 | @samp{WARNING: Couldn't find the global config file.} or | |
1662 | @samp{WARNING: Couldn't find tool init file} that can be ignored. | |
1663 | ||
1664 | @section How can I run the test suite on selected tests? | |
1665 | ||
1666 | In order to run sets of tests selectively, there are targets | |
1667 | @samp{make check-gcc} and @samp{make check-g++} | |
1668 | in the @file{gcc} subdirectory of the object directory. You can also | |
1669 | just run @samp{make check} in a subdirectory of the object directory. | |
1670 | ||
1671 | ||
1672 | A more selective way to just run all @command{gcc} execute tests in the | |
1673 | testsuite is to use | |
1674 | ||
1675 | @smallexample | |
1676 | make check-gcc RUNTESTFLAGS="execute.exp @var{other-options}" | |
1677 | @end smallexample | |
1678 | ||
1679 | Likewise, in order to run only the @command{g++} ``old-deja'' tests in | |
1680 | the testsuite with filenames matching @samp{9805*}, you would use | |
1681 | ||
1682 | @smallexample | |
1683 | make check-g++ RUNTESTFLAGS="old-deja.exp=9805* @var{other-options}" | |
1684 | @end smallexample | |
1685 | ||
1686 | The @file{*.exp} files are located in the testsuite directories of the GCC | |
1687 | source, the most important ones being @file{compile.exp}, | |
1688 | @file{execute.exp}, @file{dg.exp} and @file{old-deja.exp}. | |
1689 | To get a list of the possible @file{*.exp} files, pipe the | |
1690 | output of @samp{make check} into a file and look at the | |
1691 | @samp{Running @dots{} .exp} lines. | |
1692 | ||
1693 | @section Passing options and running multiple testsuites | |
1694 | ||
1695 | You can pass multiple options to the testsuite using the | |
1696 | @samp{--target_board} option of DejaGNU, either passed as part of | |
1697 | @samp{RUNTESTFLAGS}, or directly to @command{runtest} if you prefer to | |
1698 | work outside the makefiles. For example, | |
1699 | ||
1700 | @smallexample | |
1701 | make check-g++ RUNTESTFLAGS="--target_board=unix/-O3/-fno-strength-reduce" | |
1702 | @end smallexample | |
1703 | ||
1704 | will run the standard @command{g++} testsuites (``unix'' is the target name | |
1705 | for a standard native testsuite situation), passing | |
1706 | @samp{-O3 -fno-strength-reduce} to the compiler on every test, i.e., | |
1707 | slashes separate options. | |
1708 | ||
1709 | You can run the testsuites multiple times using combinations of options | |
1710 | with a syntax similar to the brace expansion of popular shells: | |
1711 | ||
1712 | @smallexample | |
1713 | @dots{}"--target_board=arm-sim@{-mhard-float,-msoft-float@}@{-O1,-O2,-O3,@}" | |
1714 | @end smallexample | |
1715 | ||
1716 | (Note the empty option caused by the trailing comma in the final group.) | |
1717 | The following will run each testsuite eight times using the @samp{arm-sim} | |
1718 | target, as if you had specified all possible combinations yourself: | |
1719 | ||
1720 | @smallexample | |
1721 | --target_board=arm-sim/-mhard-float/-O1 | |
1722 | --target_board=arm-sim/-mhard-float/-O2 | |
1723 | --target_board=arm-sim/-mhard-float/-O3 | |
1724 | --target_board=arm-sim/-mhard-float | |
1725 | --target_board=arm-sim/-msoft-float/-O1 | |
1726 | --target_board=arm-sim/-msoft-float/-O2 | |
1727 | --target_board=arm-sim/-msoft-float/-O3 | |
1728 | --target_board=arm-sim/-msoft-float | |
1729 | @end smallexample | |
1730 | ||
1731 | They can be combined as many times as you wish, in arbitrary ways. This | |
1732 | list: | |
1733 | ||
1734 | @smallexample | |
1735 | @dots{}"--target_board=unix/-Wextra@{-O3,-fno-strength-reduce@}@{-fomit-frame-pointer,@}" | |
1736 | @end smallexample | |
1737 | ||
1738 | will generate four combinations, all involving @samp{-Wextra}. | |
1739 | ||
1740 | The disadvantage to this method is that the testsuites are run in serial, | |
1741 | which is a waste on multiprocessor systems. For users with GNU Make and | |
1742 | a shell which performs brace expansion, you can run the testsuites in | |
1743 | parallel by having the shell perform the combinations and @command{make} | |
1744 | do the parallel runs. Instead of using @samp{--target_board}, use a | |
1745 | special makefile target: | |
1746 | ||
1747 | @smallexample | |
1748 | make -j@var{N} check-@var{testsuite}//@var{test-target}/@var{option1}/@var{option2}/@dots{} | |
1749 | @end smallexample | |
1750 | ||
1751 | For example, | |
1752 | ||
1753 | @smallexample | |
1754 | make -j3 check-gcc//sh-hms-sim/@{-m1,-m2,-m3,-m3e,-m4@}/@{,-nofpu@} | |
1755 | @end smallexample | |
1756 | ||
1757 | will run three concurrent ``make-gcc'' testsuites, eventually testing all | |
1758 | ten combinations as described above. Note that this is currently only | |
1759 | supported in the @file{gcc} subdirectory. (To see how this works, try | |
1760 | typing @command{echo} before the example given here.) | |
1761 | ||
1762 | ||
1763 | @section Additional testing for Java Class Libraries | |
1764 | ||
1765 | The Java runtime tests can be executed via @samp{make check} | |
1766 | in the @file{@var{target}/libjava/testsuite} directory in | |
1767 | the build tree. | |
1768 | ||
1769 | The @uref{http://sources.redhat.com/mauve/,,Mauve Project} provides | |
1770 | a suite of tests for the Java Class Libraries. This suite can be run | |
1771 | as part of libgcj testing by placing the Mauve tree within the libjava | |
1772 | testsuite at @file{libjava/testsuite/libjava.mauve/mauve}, or by | |
1773 | specifying the location of that tree when invoking @samp{make}, as in | |
1774 | @samp{make MAUVEDIR=~/mauve check}. | |
1775 | ||
1776 | @uref{http://www-124.ibm.com/developerworks/oss/cvs/jikes/~checkout~/jacks/jacks.html,,Jacks} | |
1777 | is a free test suite that tests Java compiler front ends. This suite | |
1778 | can be run as part of libgcj testing by placing the Jacks tree within | |
1779 | the libjava testsuite at @file{libjava/testsuite/libjava.jacks/jacks}. | |
1780 | ||
1781 | @section How to interpret test results | |
1782 | ||
1783 | The result of running the testsuite are various @file{*.sum} and @file{*.log} | |
1784 | files in the testsuite subdirectories. The @file{*.log} files contain a | |
1785 | detailed log of the compiler invocations and the corresponding | |
1786 | results, the @file{*.sum} files summarize the results. These summaries | |
1787 | contain status codes for all tests: | |
1788 | ||
1789 | @itemize @bullet | |
1790 | @item | |
1791 | PASS: the test passed as expected | |
1792 | @item | |
1793 | XPASS: the test unexpectedly passed | |
1794 | @item | |
1795 | FAIL: the test unexpectedly failed | |
1796 | @item | |
1797 | XFAIL: the test failed as expected | |
1798 | @item | |
1799 | UNSUPPORTED: the test is not supported on this platform | |
1800 | @item | |
1801 | ERROR: the testsuite detected an error | |
1802 | @item | |
1803 | WARNING: the testsuite detected a possible problem | |
1804 | @end itemize | |
1805 | ||
1806 | It is normal for some tests to report unexpected failures. At the | |
1807 | current time our testing harness does not allow fine grained control | |
1808 | over whether or not a test is expected to fail. We expect to fix this | |
1809 | problem in future releases. | |
1810 | ||
1811 | ||
1812 | @section Submitting test results | |
1813 | ||
1814 | If you want to report the results to the GCC project, use the | |
1815 | @file{contrib/test_summary} shell script. Start it in the @var{objdir} with | |
1816 | ||
1817 | @smallexample | |
1818 | @var{srcdir}/contrib/test_summary -p your_commentary.txt \ | |
1819 | -m gcc-testresults@@gcc.gnu.org |sh | |
1820 | @end smallexample | |
1821 | ||
1822 | This script uses the @command{Mail} program to send the results, so | |
1823 | make sure it is in your @env{PATH}. The file @file{your_commentary.txt} is | |
1824 | prepended to the testsuite summary and should contain any special | |
1825 | remarks you have on your results or your build environment. Please | |
1826 | do not edit the testsuite result block or the subject line, as these | |
1827 | messages may be automatically processed. | |
1828 | ||
1829 | @html | |
1830 | <hr /> | |
1831 | <p> | |
1832 | @end html | |
1833 | @ifhtml | |
1834 | @uref{./index.html,,Return to the GCC Installation page} | |
1835 | @end ifhtml | |
1836 | @end ifset | |
1837 | ||
1838 | @c ***Final install*********************************************************** | |
1839 | @ifnothtml | |
1840 | @comment node-name, next, previous, up | |
1841 | @node Final install, , Testing, Installing GCC | |
1842 | @end ifnothtml | |
1843 | @ifset finalinstallhtml | |
1844 | @ifnothtml | |
1845 | @chapter Installing GCC: Final installation | |
1846 | @end ifnothtml | |
1847 | ||
1848 | Now that GCC has been built (and optionally tested), you can install it with | |
1849 | @smallexample | |
1850 | cd @var{objdir}; make install | |
1851 | @end smallexample | |
1852 | ||
1853 | We strongly recommend to install into a target directory where there is | |
1854 | no previous version of GCC present. | |
1855 | ||
1856 | That step completes the installation of GCC; user level binaries can | |
1857 | be found in @file{@var{prefix}/bin} where @var{prefix} is the value | |
1858 | you specified with the @option{--prefix} to configure (or | |
1859 | @file{/usr/local} by default). (If you specified @option{--bindir}, | |
1860 | that directory will be used instead; otherwise, if you specified | |
1861 | @option{--exec-prefix}, @file{@var{exec-prefix}/bin} will be used.) | |
1862 | Headers for the C++ and Java libraries are installed in | |
1863 | @file{@var{prefix}/include}; libraries in @file{@var{libdir}} | |
1864 | (normally @file{@var{prefix}/lib}); internal parts of the compiler in | |
1865 | @file{@var{libdir}/gcc} and @file{@var{libexecdir}/gcc}; documentation | |
1866 | in info format in @file{@var{infodir}} (normally | |
1867 | @file{@var{prefix}/info}). | |
1868 | ||
1869 | When installing cross-compilers, GCC's executables | |
1870 | are not only installed into @file{@var{bindir}}, that | |
1871 | is, @file{@var{exec-prefix}/bin}, but additionally into | |
1872 | @file{@var{exec-prefix}/@var{target-alias}/bin}, if that directory | |
1873 | exists. Typically, such @dfn{tooldirs} hold target-specific | |
1874 | binutils, including assembler and linker. | |
1875 | ||
1876 | Installation into a temporary staging area or into a @command{chroot} | |
1877 | jail can be achieved with the command | |
1878 | ||
1879 | @smallexample | |
1880 | make DESTDIR=@var{path-to-rootdir} install | |
1881 | @end smallexample | |
1882 | ||
1883 | @noindent where @var{path-to-rootdir} is the absolute path of | |
1884 | a directory relative to which all installation paths will be | |
1885 | interpreted. Note that the directory specified by @code{DESTDIR} | |
1886 | need not exist yet; it will be created if necessary. | |
1887 | ||
1888 | There is a subtle point with tooldirs and @code{DESTDIR}: | |
1889 | If you relocate a cross-compiler installation with | |
1890 | e.g.@: @samp{DESTDIR=@var{rootdir}}, then the directory | |
1891 | @file{@var{rootdir}/@var{exec-prefix}/@var{target-alias}/bin} will | |
1892 | be filled with duplicated GCC executables only if it already exists, | |
1893 | it will not be created otherwise. This is regarded as a feature, | |
1894 | not as a bug, because it gives slightly more control to the packagers | |
1895 | using the @code{DESTDIR} feature. | |
1896 | ||
1897 | If you built a released version of GCC using @samp{make bootstrap} then please | |
1898 | quickly review the build status page for your release, available from | |
1899 | @uref{http://gcc.gnu.org/buildstat.html}. | |
1900 | If your system is not listed for the version of GCC that you built, | |
1901 | send a note to | |
1902 | @email{gcc@@gcc.gnu.org} indicating | |
1903 | that you successfully built and installed GCC. | |
1904 | Include the following information: | |
1905 | ||
1906 | @itemize @bullet | |
1907 | @item | |
1908 | Output from running @file{@var{srcdir}/config.guess}. Do not send us | |
1909 | that file itself, just the one-line output from running it. | |
1910 | ||
1911 | @item | |
1912 | The output of @samp{gcc -v} for your newly installed gcc. | |
1913 | This tells us which version of GCC you built and the options you passed to | |
1914 | configure. | |
1915 | ||
1916 | @item | |
1917 | Whether you enabled all languages or a subset of them. If you used a | |
1918 | full distribution then this information is part of the configure | |
1919 | options in the output of @samp{gcc -v}, but if you downloaded the | |
1920 | ``core'' compiler plus additional front ends then it isn't apparent | |
1921 | which ones you built unless you tell us about it. | |
1922 | ||
1923 | @item | |
1924 | If the build was for GNU/Linux, also include: | |
1925 | @itemize @bullet | |
1926 | @item | |
1927 | The distribution name and version (e.g., Red Hat 7.1 or Debian 2.2.3); | |
1928 | this information should be available from @file{/etc/issue}. | |
1929 | ||
1930 | @item | |
1931 | The version of the Linux kernel, available from @samp{uname --version} | |
1932 | or @samp{uname -a}. | |
1933 | ||
1934 | @item | |
1935 | The version of glibc you used; for RPM-based systems like Red Hat, | |
1936 | Mandrake, and SuSE type @samp{rpm -q glibc} to get the glibc version, | |
1937 | and on systems like Debian and Progeny use @samp{dpkg -l libc6}. | |
1938 | @end itemize | |
1939 | For other systems, you can include similar information if you think it is | |
1940 | relevant. | |
1941 | ||
1942 | @item | |
1943 | Any other information that you think would be useful to people building | |
1944 | GCC on the same configuration. The new entry in the build status list | |
1945 | will include a link to the archived copy of your message. | |
1946 | @end itemize | |
1947 | ||
1948 | We'd also like to know if the | |
1949 | @ifnothtml | |
1950 | @ref{Specific, host/target specific installation notes} | |
1951 | @end ifnothtml | |
1952 | @ifhtml | |
1953 | @uref{specific.html,,host/target specific installation notes} | |
1954 | @end ifhtml | |
1955 | didn't include your host/target information or if that information is | |
1956 | incomplete or out of date. Send a note to | |
1957 | @email{gcc@@gcc.gnu.org} telling us how the information should be changed. | |
1958 | ||
1959 | If you find a bug, please report it following our | |
1960 | @uref{../bugs.html,,bug reporting guidelines}. | |
1961 | ||
1962 | If you want to print the GCC manuals, do @samp{cd @var{objdir}; make | |
1963 | dvi}. You will need to have @command{texi2dvi} (version at least 4.2) | |
1964 | and @TeX{} installed. This creates a number of @file{.dvi} files in | |
1965 | subdirectories of @file{@var{objdir}}; these may be converted for | |
1966 | printing with programs such as @command{dvips}. You can also | |
1967 | @uref{http://www.gnu.org/order/order.html,,buy printed manuals from the | |
1968 | Free Software Foundation}, though such manuals may not be for the most | |
1969 | recent version of GCC@. | |
1970 | ||
1971 | @html | |
1972 | <hr /> | |
1973 | <p> | |
1974 | @end html | |
1975 | @ifhtml | |
1976 | @uref{./index.html,,Return to the GCC Installation page} | |
1977 | @end ifhtml | |
1978 | @end ifset | |
1979 | ||
1980 | @c ***Binaries**************************************************************** | |
1981 | @ifnothtml | |
1982 | @comment node-name, next, previous, up | |
1983 | @node Binaries, Specific, Installing GCC, Top | |
1984 | @end ifnothtml | |
1985 | @ifset binarieshtml | |
1986 | @ifnothtml | |
1987 | @chapter Installing GCC: Binaries | |
1988 | @end ifnothtml | |
1989 | @cindex Binaries | |
1990 | @cindex Installing GCC: Binaries | |
1991 | ||
1992 | We are often asked about pre-compiled versions of GCC@. While we cannot | |
1993 | provide these for all platforms, below you'll find links to binaries for | |
1994 | various platforms where creating them by yourself is not easy due to various | |
1995 | reasons. | |
1996 | ||
1997 | Please note that we did not create these binaries, nor do we | |
1998 | support them. If you have any problems installing them, please | |
1999 | contact their makers. | |
2000 | ||
2001 | @itemize | |
2002 | @item | |
2003 | AIX: | |
2004 | @itemize | |
2005 | @item | |
2006 | @uref{http://www.bullfreeware.com,,Bull's Freeware and Shareware Archive for AIX}; | |
2007 | ||
2008 | @item | |
2009 | @uref{http://aixpdslib.seas.ucla.edu,,UCLA Software Library for AIX}. | |
2010 | @end itemize | |
2011 | ||
2012 | @item | |
2013 | DOS---@uref{http://www.delorie.com/djgpp/,,DJGPP}. | |
2014 | ||
2015 | @item | |
2016 | Renesas H8/300[HS]---@uref{http://h8300-hms.sourceforge.net/,,GNU | |
2017 | Development Tools for the Renesas H8/300[HS] Series}. | |
2018 | ||
2019 | @item | |
2020 | HP-UX: | |
2021 | @itemize | |
2022 | @item | |
2023 | @uref{http://hpux.cae.wisc.edu/,,HP-UX Porting Center}; | |
2024 | ||
2025 | @item | |
2026 | @uref{ftp://sunsite.informatik.rwth-aachen.de/pub/packages/gcc_hpux/,,Binaries for HP-UX 11.00 at Aachen University of Technology}. | |
2027 | @end itemize | |
2028 | ||
2029 | @item | |
2030 | Motorola 68HC11/68HC12---@uref{http://www.gnu-m68hc11.org,,GNU | |
2031 | Development Tools for the Motorola 68HC11/68HC12}. | |
2032 | ||
2033 | @item | |
2034 | @uref{http://www.sco.com/skunkware/devtools/index.html#gcc,,SCO | |
2035 | OpenServer/Unixware}. | |
2036 | ||
2037 | @item | |
2038 | Sinix/Reliant Unix---@uref{ftp://ftp.fujitsu-siemens.com/pub/pd/gnu/gcc/,,Siemens}. | |
2039 | ||
2040 | @item | |
2041 | Solaris 2 (SPARC, Intel)---@uref{http://www.sunfreeware.com/,,Sunfreeware}. | |
2042 | ||
2043 | @item | |
2044 | SGI---@uref{http://freeware.sgi.com/,,SGI Freeware}. | |
2045 | ||
2046 | @item | |
2047 | Microsoft Windows: | |
2048 | @itemize | |
2049 | @item | |
2050 | The @uref{http://sources.redhat.com/cygwin/,,Cygwin} project; | |
2051 | @item | |
2052 | The @uref{http://www.mingw.org/,,MinGW} project. | |
2053 | @end itemize | |
2054 | ||
2055 | @item | |
2056 | @uref{ftp://ftp.thewrittenword.com/packages/by-name/,,The | |
2057 | Written Word} offers binaries for | |
2058 | AIX 4.3.2. | |
2059 | IRIX 6.5, | |
2060 | Digital UNIX 4.0D and 5.1, | |
2061 | GNU/Linux (i386), | |
2062 | HP-UX 10.20, 11.00, and 11.11, and | |
2063 | Solaris/SPARC 2.5.1, 2.6, 2.7, 8, and 9, | |
2064 | @end itemize | |
2065 | ||
2066 | In addition to those specific offerings, you can get a binary | |
2067 | distribution CD-ROM from the | |
2068 | @uref{http://www.fsf.org/order/order.html,,Free Software Foundation}. | |
2069 | It contains binaries for a number of platforms, and | |
2070 | includes not only GCC, but other stuff as well. The current CD does | |
2071 | not contain the latest version of GCC, but it should allow | |
2072 | bootstrapping the compiler. An updated version of that disk is in the | |
2073 | works. | |
2074 | ||
2075 | @html | |
2076 | <hr /> | |
2077 | <p> | |
2078 | @end html | |
2079 | @ifhtml | |
2080 | @uref{./index.html,,Return to the GCC Installation page} | |
2081 | @end ifhtml | |
2082 | @end ifset | |
2083 | ||
2084 | @c ***Specific**************************************************************** | |
2085 | @ifnothtml | |
2086 | @comment node-name, next, previous, up | |
2087 | @node Specific, Old, Binaries, Top | |
2088 | @end ifnothtml | |
2089 | @ifset specifichtml | |
2090 | @ifnothtml | |
2091 | @chapter Host/target specific installation notes for GCC | |
2092 | @end ifnothtml | |
2093 | @cindex Specific | |
2094 | @cindex Specific installation notes | |
2095 | @cindex Target specific installation | |
2096 | @cindex Host specific installation | |
2097 | @cindex Target specific installation notes | |
2098 | ||
2099 | Please read this document carefully @emph{before} installing the | |
2100 | GNU Compiler Collection on your machine. | |
2101 | ||
2102 | @ifhtml | |
2103 | @itemize | |
2104 | @item | |
2105 | @uref{#alpha*-*-*,,alpha*-*-*} | |
2106 | @item | |
2107 | @uref{#alpha*-dec-osf*,,alpha*-dec-osf*} | |
2108 | @item | |
2109 | @uref{#alphaev5-cray-unicosmk*,,alphaev5-cray-unicosmk*} | |
2110 | @item | |
2111 | @uref{#arc-*-elf,,arc-*-elf} | |
2112 | @item | |
2113 | @uref{#arm-*-elf,,arm-*-elf} | |
2114 | @uref{#arm-*-coff,,arm-*-coff} | |
2115 | @uref{#arm-*-aout,,arm-*-aout} | |
2116 | @item | |
2117 | @uref{#xscale-*-*,,xscale-*-*} | |
2118 | @item | |
2119 | @uref{#avr,,avr} | |
2120 | @item | |
2121 | @uref{#c4x,,c4x} | |
2122 | @item | |
2123 | @uref{#dos,,DOS} | |
2124 | @item | |
2125 | @uref{#*-*-freebsd*,,*-*-freebsd*} | |
2126 | @item | |
2127 | @uref{#h8300-hms,,h8300-hms} | |
2128 | @item | |
2129 | @uref{#hppa*-hp-hpux*,,hppa*-hp-hpux*} | |
2130 | @item | |
2131 | @uref{#hppa*-hp-hpux10,,hppa*-hp-hpux10} | |
2132 | @item | |
2133 | @uref{#hppa*-hp-hpux11,,hppa*-hp-hpux11} | |
2134 | @item | |
2135 | @uref{#*-*-linux-gnu,,*-*-linux-gnu} | |
2136 | @item | |
2137 | @uref{#ix86-*-linux*aout,,i?86-*-linux*aout} | |
2138 | @item | |
2139 | @uref{#ix86-*-linux*,,i?86-*-linux*} | |
2140 | @item | |
2141 | @uref{#ix86-*-sco3.2v5*,,i?86-*-sco3.2v5*} | |
2142 | @item | |
2143 | @uref{#ix86-*-udk,,i?86-*-udk} | |
2144 | @item | |
2145 | @uref{#ix86-*-esix,,i?86-*-esix} | |
2146 | @item | |
2147 | @uref{#ia64-*-linux,,ia64-*-linux} | |
2148 | @item | |
2149 | @uref{#ia64-*-hpux*,,ia64-*-hpux*} | |
2150 | @item | |
2151 | @uref{#*-ibm-aix*,,*-ibm-aix*} | |
2152 | @item | |
2153 | @uref{#ip2k-*-elf,,ip2k-*-elf} | |
2154 | @item | |
2155 | @uref{#iq2000-*-elf,,iq2000-*-elf} | |
2156 | @item | |
2157 | @uref{#m32r-*-elf,,m32r-*-elf} | |
2158 | @item | |
2159 | @uref{#m6811-elf,,m6811-elf} | |
2160 | @item | |
2161 | @uref{#m6812-elf,,m6812-elf} | |
2162 | @item | |
2163 | @uref{#m68k-hp-hpux,,m68k-hp-hpux} | |
2164 | @item | |
2165 | @uref{#mips-*-*,,mips-*-*} | |
2166 | @item | |
2167 | @uref{#mips-sgi-irix5,,mips-sgi-irix5} | |
2168 | @item | |
2169 | @uref{#mips-sgi-irix6,,mips-sgi-irix6} | |
2170 | @item | |
2171 | @uref{#powerpc*-*-*,,powerpc*-*-*, powerpc-*-sysv4} | |
2172 | @item | |
2173 | @uref{#powerpc-*-darwin*,,powerpc-*-darwin*} | |
2174 | @item | |
2175 | @uref{#powerpc-*-elf,,powerpc-*-elf, powerpc-*-sysv4} | |
2176 | @item | |
2177 | @uref{#powerpc-*-linux-gnu*,,powerpc-*-linux-gnu*} | |
2178 | @item | |
2179 | @uref{#powerpc-*-netbsd*,,powerpc-*-netbsd*} | |
2180 | @item | |
2181 | @uref{#powerpc-*-eabiaix,,powerpc-*-eabiaix} | |
2182 | @item | |
2183 | @uref{#powerpc-*-eabisim,,powerpc-*-eabisim} | |
2184 | @item | |
2185 | @uref{#powerpc-*-eabi,,powerpc-*-eabi} | |
2186 | @item | |
2187 | @uref{#powerpcle-*-elf,,powerpcle-*-elf, powerpcle-*-sysv4} | |
2188 | @item | |
2189 | @uref{#powerpcle-*-eabisim,,powerpcle-*-eabisim} | |
2190 | @item | |
2191 | @uref{#powerpcle-*-eabi,,powerpcle-*-eabi} | |
2192 | @item | |
2193 | @uref{#s390-*-linux*,,s390-*-linux*} | |
2194 | @item | |
2195 | @uref{#s390x-*-linux*,,s390x-*-linux*} | |
2196 | @item | |
2197 | @uref{#s390x-ibm-tpf*,,s390x-ibm-tpf*} | |
2198 | @item | |
2199 | @uref{#*-*-solaris2*,,*-*-solaris2*} | |
2200 | @item | |
2201 | @uref{#sparc-sun-solaris2*,,sparc-sun-solaris2*} | |
2202 | @item | |
2203 | @uref{#sparc-sun-solaris2.7,,sparc-sun-solaris2.7} | |
2204 | @item | |
2205 | @uref{#sparc-*-linux*,,sparc-*-linux*} | |
2206 | @item | |
2207 | @uref{#sparc64-*-solaris2*,,sparc64-*-solaris2*} | |
2208 | @item | |
2209 | @uref{#sparcv9-*-solaris2*,,sparcv9-*-solaris2*} | |
2210 | @item | |
2211 | @uref{#*-*-sysv*,,*-*-sysv*} | |
2212 | @item | |
2213 | @uref{#vax-dec-ultrix,,vax-dec-ultrix} | |
2214 | @item | |
2215 | @uref{#*-*-vxworks*,,*-*-vxworks*} | |
2216 | @item | |
2217 | @uref{#x86_64-*-*,,x86_64-*-*, amd64-*-*} | |
2218 | @item | |
2219 | @uref{#xtensa-*-elf,,xtensa-*-elf} | |
2220 | @item | |
2221 | @uref{#xtensa-*-linux*,,xtensa-*-linux*} | |
2222 | @item | |
2223 | @uref{#windows,,Microsoft Windows} | |
2224 | @item | |
2225 | @uref{#os2,,OS/2} | |
2226 | @item | |
2227 | @uref{#older,,Older systems} | |
2228 | @end itemize | |
2229 | ||
2230 | @itemize | |
2231 | @item | |
2232 | @uref{#elf_targets,,all ELF targets} (SVR4, Solaris 2, etc.) | |
2233 | @end itemize | |
2234 | @end ifhtml | |
2235 | ||
2236 | ||
2237 | @html | |
2238 | <!-- -------- host/target specific issues start here ---------------- --> | |
2239 | <hr /> | |
2240 | @end html | |
2241 | @heading @anchor{alpha*-*-*}alpha*-*-* | |
2242 | ||
2243 | This section contains general configuration information for all | |
2244 | alpha-based platforms using ELF (in particular, ignore this section for | |
2245 | DEC OSF/1, Digital UNIX and Tru64 UNIX)@. In addition to reading this | |
2246 | section, please read all other sections that match your target. | |
2247 | ||
2248 | We require binutils 2.11.2 or newer. | |
2249 | Previous binutils releases had a number of problems with DWARF 2 | |
2250 | debugging information, not the least of which is incorrect linking of | |
2251 | shared libraries. | |
2252 | ||
2253 | @html | |
2254 | <hr /> | |
2255 | @end html | |
2256 | @heading @anchor{alpha*-dec-osf*}alpha*-dec-osf* | |
2257 | Systems using processors that implement the DEC Alpha architecture and | |
2258 | are running the DEC/Compaq Unix (DEC OSF/1, Digital UNIX, or Compaq | |
2259 | Tru64 UNIX) operating system, for example the DEC Alpha AXP systems. | |
2260 | ||
2261 | As of GCC 3.2, versions before @code{alpha*-dec-osf4} are no longer | |
2262 | supported. (These are the versions which identify themselves as DEC | |
2263 | OSF/1.) | |
2264 | ||
2265 | In Digital Unix V4.0, virtual memory exhausted bootstrap failures | |
2266 | may be fixed by configuring with @option{--with-gc=simple}, | |
2267 | reconfiguring Kernel Virtual Memory and Swap parameters | |
2268 | per the @command{/usr/sbin/sys_check} Tuning Suggestions, | |
2269 | or applying the patch in | |
2270 | @uref{http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html}. | |
2271 | ||
2272 | In Tru64 UNIX V5.1, Compaq introduced a new assembler that does not | |
2273 | currently (2001-06-13) work with @command{mips-tfile}. As a workaround, | |
2274 | we need to use the old assembler, invoked via the barely documented | |
2275 | @option{-oldas} option. To bootstrap GCC, you either need to use the | |
2276 | Compaq C Compiler: | |
2277 | ||
2278 | @smallexample | |
2279 | % CC=cc @var{srcdir}/configure [@var{options}] [@var{target}] | |
2280 | @end smallexample | |
2281 | ||
2282 | or you can use a copy of GCC 2.95.3 or higher built on Tru64 UNIX V4.0: | |
2283 | ||
2284 | @smallexample | |
2285 | % CC=gcc -Wa,-oldas @var{srcdir}/configure [@var{options}] [@var{target}] | |
2286 | @end smallexample | |
2287 | ||
2288 | As of GNU binutils 2.11.2, neither GNU @command{as} nor GNU @command{ld} | |
2289 | are supported on Tru64 UNIX, so you must not configure GCC with | |
2290 | @option{--with-gnu-as} or @option{--with-gnu-ld}. | |
2291 | ||
2292 | GCC writes a @samp{.verstamp} directive to the assembler output file | |
2293 | unless it is built as a cross-compiler. It gets the version to use from | |
2294 | the system header file @file{/usr/include/stamp.h}. If you install a | |
2295 | new version of DEC Unix, you should rebuild GCC to pick up the new version | |
2296 | stamp. | |
2297 | ||
2298 | Note that since the Alpha is a 64-bit architecture, cross-compilers from | |
2299 | 32-bit machines will not generate code as efficient as that generated | |
2300 | when the compiler is running on a 64-bit machine because many | |
2301 | optimizations that depend on being able to represent a word on the | |
2302 | target in an integral value on the host cannot be performed. Building | |
2303 | cross-compilers on the Alpha for 32-bit machines has only been tested in | |
2304 | a few cases and may not work properly. | |
2305 | ||
2306 | @samp{make compare} may fail on old versions of DEC Unix unless you add | |
2307 | @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the | |
2308 | assembler input file is stored in the object file, and that makes | |
2309 | comparison fail if it differs between the @code{stage1} and | |
2310 | @code{stage2} compilations. The option @option{-save-temps} forces a | |
2311 | fixed name to be used for the assembler input file, instead of a | |
2312 | randomly chosen name in @file{/tmp}. Do not add @option{-save-temps} | |
2313 | unless the comparisons fail without that option. If you add | |
2314 | @option{-save-temps}, you will have to manually delete the @samp{.i} and | |
2315 | @samp{.s} files after each series of compilations. | |
2316 | ||
2317 | GCC now supports both the native (ECOFF) debugging format used by DBX | |
2318 | and GDB and an encapsulated STABS format for use only with GDB@. See the | |
2319 | discussion of the @option{--with-stabs} option of @file{configure} above | |
2320 | for more information on these formats and how to select them. | |
2321 | ||
2322 | There is a bug in DEC's assembler that produces incorrect line numbers | |
2323 | for ECOFF format when the @samp{.align} directive is used. To work | |
2324 | around this problem, GCC will not emit such alignment directives | |
2325 | while writing ECOFF format debugging information even if optimization is | |
2326 | being performed. Unfortunately, this has the very undesirable | |
2327 | side-effect that code addresses when @option{-O} is specified are | |
2328 | different depending on whether or not @option{-g} is also specified. | |
2329 | ||
2330 | To avoid this behavior, specify @option{-gstabs+} and use GDB instead of | |
2331 | DBX@. DEC is now aware of this problem with the assembler and hopes to | |
2332 | provide a fix shortly. | |
2333 | ||
2334 | @html | |
2335 | <hr /> | |
2336 | @end html | |
2337 | @heading @anchor{alphaev5-cray-unicosmk*}alphaev5-cray-unicosmk* | |
2338 | Cray T3E systems running Unicos/Mk. | |
2339 | ||
2340 | This port is incomplete and has many known bugs. We hope to improve the | |
2341 | support for this target soon. Currently, only the C front end is supported, | |
2342 | and it is not possible to build parallel applications. Cray modules are not | |
2343 | supported; in particular, Craylibs are assumed to be in | |
2344 | @file{/opt/ctl/craylibs/craylibs}. | |
2345 | ||
2346 | You absolutely @strong{must} use GNU make on this platform. Also, you | |
2347 | need to tell GCC where to find the assembler and the linker. The | |
2348 | simplest way to do so is by providing @option{--with-as} and | |
2349 | @option{--with-ld} to @file{configure}, e.g.@: | |
2350 | ||
2351 | @smallexample | |
2352 | configure --with-as=/opt/ctl/bin/cam --with-ld=/opt/ctl/bin/cld \ | |
2353 | --enable-languages=c | |
2354 | @end smallexample | |
2355 | ||
2356 | The comparison test during @samp{make bootstrap} fails on Unicos/Mk | |
2357 | because the assembler inserts timestamps into object files. You should | |
2358 | be able to work around this by doing @samp{make all} after getting this | |
2359 | failure. | |
2360 | ||
2361 | @html | |
2362 | <hr /> | |
2363 | @end html | |
2364 | @heading @anchor{arc-*-elf}arc-*-elf | |
2365 | Argonaut ARC processor. | |
2366 | This configuration is intended for embedded systems. | |
2367 | ||
2368 | @html | |
2369 | <hr /> | |
2370 | @end html | |
2371 | @heading @anchor{arm-*-elf}arm-*-elf | |
2372 | @heading @anchor{xscale-*-*}xscale-*-* | |
2373 | ARM-family processors. Subtargets that use the ELF object format | |
2374 | require GNU binutils 2.13 or newer. Such subtargets include: | |
2375 | @code{arm-*-freebsd}, @code{arm-*-netbsdelf}, @code{arm-*-*linux}, | |
2376 | @code{arm-*-rtems} and @code{arm-*-kaos}. | |
2377 | ||
2378 | @html | |
2379 | <hr /> | |
2380 | @end html | |
2381 | @heading @anchor{arm-*-coff}arm-*-coff | |
2382 | ARM-family processors. Note that there are two different varieties | |
2383 | of PE format subtarget supported: @code{arm-wince-pe} and | |
2384 | @code{arm-pe} as well as a standard COFF target @code{arm-*-coff}. | |
2385 | ||
2386 | @html | |
2387 | <hr /> | |
2388 | @end html | |
2389 | @heading @anchor{arm-*-aout}arm-*-aout | |
2390 | ARM-family processors. These targets support the AOUT file format: | |
2391 | @code{arm-*-aout}, @code{arm-*-netbsd}. | |
2392 | ||
2393 | @html | |
2394 | <hr /> | |
2395 | @end html | |
2396 | @heading @anchor{avr}avr | |
2397 | ||
2398 | ATMEL AVR-family micro controllers. These are used in embedded | |
2399 | applications. There are no standard Unix configurations. | |
2400 | @ifnothtml | |
2401 | @xref{AVR Options,, AVR Options, gcc, Using and Porting the GNU Compiler | |
2402 | Collection (GCC)}, | |
2403 | @end ifnothtml | |
2404 | @ifhtml | |
2405 | See ``AVR Options'' in the main manual | |
2406 | @end ifhtml | |
2407 | for the list of supported MCU types. | |
2408 | ||
2409 | Use @samp{configure --target=avr --enable-languages="c"} to configure GCC@. | |
2410 | ||
2411 | Further installation notes and other useful information about AVR tools | |
2412 | can also be obtained from: | |
2413 | ||
2414 | @itemize @bullet | |
2415 | @item | |
2416 | @uref{http://www.openavr.org,,http://www.openavr.org} | |
2417 | @item | |
2418 | @uref{http://home.overta.ru/users/denisc/,,http://home.overta.ru/users/denisc/} | |
2419 | @item | |
2420 | @uref{http://www.amelek.gda.pl/avr/,,http://www.amelek.gda.pl/avr/} | |
2421 | @end itemize | |
2422 | ||
2423 | We @emph{strongly} recommend using binutils 2.13 or newer. | |
2424 | ||
2425 | The following error: | |
2426 | @smallexample | |
2427 | Error: register required | |
2428 | @end smallexample | |
2429 | ||
2430 | indicates that you should upgrade to a newer version of the binutils. | |
2431 | ||
2432 | @html | |
2433 | <hr /> | |
2434 | @end html | |
2435 | @heading @anchor{c4x}c4x | |
2436 | ||
2437 | Texas Instruments TMS320C3x and TMS320C4x Floating Point Digital Signal | |
2438 | Processors. These are used in embedded applications. There are no | |
2439 | standard Unix configurations. | |
2440 | @ifnothtml | |
2441 | @xref{TMS320C3x/C4x Options,, TMS320C3x/C4x Options, gcc, Using and | |
2442 | Porting the GNU Compiler Collection (GCC)}, | |
2443 | @end ifnothtml | |
2444 | @ifhtml | |
2445 | See ``TMS320C3x/C4x Options'' in the main manual | |
2446 | @end ifhtml | |
2447 | for the list of supported MCU types. | |
2448 | ||
2449 | GCC can be configured as a cross compiler for both the C3x and C4x | |
2450 | architectures on the same system. Use @samp{configure --target=c4x | |
2451 | --enable-languages="c,c++"} to configure. | |
2452 | ||
2453 | ||
2454 | Further installation notes and other useful information about C4x tools | |
2455 | can also be obtained from: | |
2456 | ||
2457 | @itemize @bullet | |
2458 | @item | |
2459 | @uref{http://www.elec.canterbury.ac.nz/c4x/,,http://www.elec.canterbury.ac.nz/c4x/} | |
2460 | @end itemize | |
2461 | ||
2462 | @html | |
2463 | <hr /> | |
2464 | @end html | |
2465 | @heading @anchor{cris}CRIS | |
2466 | ||
2467 | CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip | |
2468 | series. These are used in embedded applications. | |
2469 | ||
2470 | @ifnothtml | |
2471 | @xref{CRIS Options,, CRIS Options, gcc, Using and Porting the GNU Compiler | |
2472 | Collection (GCC)}, | |
2473 | @end ifnothtml | |
2474 | @ifhtml | |
2475 | See ``CRIS Options'' in the main manual | |
2476 | @end ifhtml | |
2477 | for a list of CRIS-specific options. | |
2478 | ||
2479 | There are a few different CRIS targets: | |
2480 | @table @code | |
2481 | @item cris-axis-aout | |
2482 | Old target. Includes a multilib for the @samp{elinux} a.out-based | |
2483 | target. No multilibs for newer architecture variants. | |
2484 | @item cris-axis-elf | |
2485 | Mainly for monolithic embedded systems. Includes a multilib for the | |
2486 | @samp{v10} core used in @samp{ETRAX 100 LX}. | |
2487 | @item cris-axis-linux-gnu | |
2488 | A GNU/Linux port for the CRIS architecture, currently targeting | |
2489 | @samp{ETRAX 100 LX} by default. | |
2490 | @end table | |
2491 | ||
2492 | For @code{cris-axis-aout} and @code{cris-axis-elf} you need binutils 2.11 | |
2493 | or newer. For @code{cris-axis-linux-gnu} you need binutils 2.12 or newer. | |
2494 | ||
2495 | Pre-packaged tools can be obtained from | |
2496 | @uref{ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/}. More | |
2497 | information about this platform is available at | |
2498 | @uref{http://developer.axis.com/}. | |
2499 | ||
2500 | @html | |
2501 | <hr /> | |
2502 | @end html | |
2503 | @heading @anchor{dos}DOS | |
2504 | ||
2505 | Please have a look at our @uref{binaries.html,,binaries page}. | |
2506 | ||
2507 | You cannot install GCC by itself on MSDOS; it will not compile under | |
2508 | any MSDOS compiler except itself. You need to get the complete | |
2509 | compilation package DJGPP, which includes binaries as well as sources, | |
2510 | and includes all the necessary compilation tools and libraries. | |
2511 | ||
2512 | @html | |
2513 | <hr /> | |
2514 | @end html | |
2515 | @heading @anchor{*-*-freebsd*}*-*-freebsd* | |
2516 | ||
2517 | The version of binutils installed in @file{/usr/bin} probably works with | |
2518 | this release of GCC. However, on FreeBSD 4, bootstrapping against the | |
2519 | latest FSF binutils is known to improve overall testsuite results; and, | |
2520 | on FreeBSD/alpha, using binutils 2.14 or later is required to build libjava. | |
2521 | ||
2522 | Support for FreeBSD 1 was discontinued in GCC 3.2. | |
2523 | ||
2524 | Support for FreeBSD 2 will be discontinued after GCC 3.4. The | |
2525 | following was true for GCC 3.1 but the current status is unknown. | |
2526 | For FreeBSD 2 or any mutant a.out versions of FreeBSD 3: All | |
2527 | configuration support and files as shipped with GCC 2.95 are still in | |
2528 | place. FreeBSD 2.2.7 has been known to bootstrap completely; however, | |
2529 | it is unknown which version of binutils was used (it is assumed that it | |
2530 | was the system copy in @file{/usr/bin}) and C++ EH failures were noted. | |
2531 | ||
2532 | For FreeBSD using the ELF file format: DWARF 2 debugging is now the | |
2533 | default for all CPU architectures. It had been the default on | |
2534 | FreeBSD/alpha since its inception. You may use @option{-gstabs} instead | |
2535 | of @option{-g}, if you really want the old debugging format. There are | |
2536 | no known issues with mixing object files and libraries with different | |
2537 | debugging formats. Otherwise, this release of GCC should now match more | |
2538 | of the configuration used in the stock FreeBSD configuration of GCC. In | |
2539 | particular, @option{--enable-threads} is now configured by default. | |
2540 | However, as a general user, do not attempt to replace the system | |
2541 | compiler with this release. Known to bootstrap and check with good | |
2542 | results on FreeBSD 4.9-STABLE and 5-CURRENT@. In the past, known to | |
2543 | bootstrap and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, | |
2544 | 4.3, 4.4, 4.5, 4.8-STABLE@. | |
2545 | ||
2546 | In principle, @option{--enable-threads} is now compatible with | |
2547 | @option{--enable-libgcj} on FreeBSD@. However, it has only been built | |
2548 | and tested on @samp{i386-*-freebsd[45]} and @samp{alpha-*-freebsd[45]}. | |
2549 | The static | |
2550 | library may be incorrectly built (symbols are missing at link time). | |
2551 | There is a rare timing-based startup hang (probably involves an | |
2552 | assumption about the thread library). Multi-threaded boehm-gc (required for | |
2553 | libjava) exposes severe threaded signal-handling bugs on FreeBSD before | |
2554 | 4.5-RELEASE@. Other CPU architectures | |
2555 | supported by FreeBSD will require additional configuration tuning in, at | |
2556 | the very least, both boehm-gc and libffi. | |
2557 | ||
2558 | Shared @file{libgcc_s.so} is now built and installed by default. | |
2559 | ||
2560 | @html | |
2561 | <hr /> | |
2562 | @end html | |
2563 | @heading @anchor{h8300-hms}h8300-hms | |
2564 | Renesas H8/300 series of processors. | |
2565 | ||
2566 | Please have a look at our @uref{binaries.html,,binaries page}. | |
2567 | ||
2568 | The calling convention and structure layout has changed in release 2.6. | |
2569 | All code must be recompiled. The calling convention now passes the | |
2570 | first three arguments in function calls in registers. Structures are no | |
2571 | longer a multiple of 2 bytes. | |
2572 | ||
2573 | @html | |
2574 | <hr /> | |
2575 | @end html | |
2576 | @heading @anchor{hppa*-hp-hpux*}hppa*-hp-hpux* | |
2577 | Support for HP-UX version 9 and older was discontinued in GCC 3.4. | |
2578 | ||
2579 | We @emph{highly} recommend using gas/binutils 2.8 or newer on all hppa | |
2580 | platforms; you may encounter a variety of problems when using the HP | |
2581 | assembler. | |
2582 | ||
2583 | Specifically, @option{-g} does not work on HP-UX (since that system | |
2584 | uses a peculiar debugging format which GCC does not know about), unless you | |
2585 | use GAS and GDB and configure GCC with the | |
2586 | @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}} and | |
2587 | @option{--with-as=@dots{}} options. | |
2588 | ||
2589 | If you wish to use the pa-risc 2.0 architecture support with a 32-bit | |
2590 | runtime, you must use either the HP assembler, gas/binutils 2.11 or newer, | |
2591 | or a recent | |
2592 | @uref{ftp://sources.redhat.com/pub/binutils/snapshots,,snapshot of gas}. | |
2593 | ||
2594 | There are two default scheduling models for instructions. These are | |
2595 | PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc | |
2596 | architecture specified for the target machine when configuring. | |
2597 | PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when | |
2598 | the target is a @samp{hppa1*} machine. | |
2599 | ||
2600 | The PROCESSOR_8000 model is not well suited to older processors. Thus, | |
2601 | it is important to completely specify the machine architecture when | |
2602 | configuring if you want a model other than PROCESSOR_8000. The macro | |
2603 | TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different | |
2604 | default scheduling model is desired. | |
2605 | ||
2606 | More specific information to @samp{hppa*-hp-hpux*} targets follows. | |
2607 | ||
2608 | @html | |
2609 | <hr /> | |
2610 | @end html | |
2611 | @heading @anchor{hppa*-hp-hpux10}hppa*-hp-hpux10 | |
2612 | ||
2613 | For hpux10.20, we @emph{highly} recommend you pick up the latest sed patch | |
2614 | @code{PHCO_19798} from HP@. HP has two sites which provide patches free of | |
2615 | charge: | |
2616 | ||
2617 | @itemize @bullet | |
2618 | @item | |
2619 | @html | |
2620 | <a href="http://us.itrc.hp.com/service/home/home.do">US, Canada, Asia-Pacific, and | |
2621 | Latin-America</a> | |
2622 | @end html | |
2623 | @ifnothtml | |
2624 | @uref{http://us.itrc.hp.com/service/home/home.do,,} US, Canada, Asia-Pacific, | |
2625 | and Latin-America. | |
2626 | @end ifnothtml | |
2627 | @item | |
2628 | @uref{http://europe.itrc.hp.com/service/home/home.do,,} Europe. | |
2629 | @end itemize | |
2630 | ||
2631 | The HP assembler on these systems has some problems. Most notably the | |
2632 | assembler inserts timestamps into each object file it creates, causing | |
2633 | the 3-stage comparison test to fail during a @samp{make bootstrap}. | |
2634 | You should be able to continue by saying @samp{make all} after getting | |
2635 | the failure from @samp{make bootstrap}. | |
2636 | ||
2637 | ||
2638 | @html | |
2639 | <hr /> | |
2640 | @end html | |
2641 | @heading @anchor{hppa*-hp-hpux11}hppa*-hp-hpux11 | |
2642 | ||
2643 | GCC 3.0 and up support HP-UX 11. On 64-bit capable systems, there | |
2644 | are two distinct ports. The @samp{hppa2.0w-hp-hpux11*} port generates | |
2645 | code for the 32-bit pa-risc runtime architecture. It uses the HP | |
2646 | linker. The @samp{hppa64-hp-hpux11*} port generates 64-bit code for the | |
2647 | pa-risc 2.0 architecture. The script config.guess now selects the port | |
2648 | type based on the type compiler detected during configuration. You must | |
2649 | set your @env{PATH} or define @env{CC} so that configure finds an appropriate | |
2650 | compiler for the initial bootstrap. Different prefixes must be used if | |
2651 | both ports are to be installed on the same system. | |
2652 | ||
2653 | It is best to explicitly configure the @samp{hppa64-hp-hpux11*} target | |
2654 | with the @option{--with-ld=@dots{}} option. We support both the HP | |
2655 | and GNU linkers for this target. The two linkers require different | |
2656 | link commands. Thus, it's not possible to switch linkers during a | |
2657 | GCC build. This has been been reported to occur in a unified build | |
2658 | of binutils and GCC. | |
2659 | ||
2660 | GCC 2.95.x is not supported under HP-UX 11 and cannot be used to | |
2661 | compile GCC 3.0 and up. Refer to @uref{binaries.html,,binaries} for | |
2662 | information about obtaining precompiled GCC binaries for HP-UX. | |
2663 | ||
2664 | You must use GNU binutils 2.11 or above with the 32-bit port. Thread | |
2665 | support is not currently implemented, so @option{--enable-threads} does | |
2666 | not work. See: | |
2667 | ||
2668 | @itemize | |
2669 | @item @uref{http://gcc.gnu.org/ml/gcc-prs/2002-01/msg00551.html} | |
2670 | @item @uref{http://gcc.gnu.org/ml/gcc-bugs/2002-01/msg00663.html} | |
2671 | @end itemize | |
2672 | ||
2673 | GCC 3.3 and later support weak symbols on the 32-bit port using SOM | |
2674 | secondary definition symbols. This feature is not enabled for earlier | |
2675 | versions of HP-UX since there have been bugs in the linker support for | |
2676 | secondary symbols. The HP linker patches @code{PHSS_26559} and | |
2677 | @code{PHSS_24304} for HP-UX 11.00 and 11.11, respectively, correct the | |
2678 | problem of linker core dumps creating C++ libraries. Earlier patches | |
2679 | may work but they have not been tested. | |
2680 | ||
2681 | GCC 3.3 nows uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capability | |
2682 | to run initializers and finalizers on the 64-bit port. The feature | |
2683 | requires CVS binutils as of January 2, 2003, or a subsequent release | |
2684 | to correct a problem arising from HP's non-standard use of the .init | |
2685 | and .fini sections. The 32-bit port uses the linker @option{+init} | |
2686 | and @option{+fini} options. As with the support for secondary symbols, | |
2687 | there have been bugs in the order in which these options are executed | |
2688 | by the HP linker. So, again a recent linker patch is recommended. | |
2689 | ||
2690 | The HP assembler has many limitations and is not recommended for either | |
2691 | the 32 or 64-bit ports. For example, it does not support weak symbols | |
2692 | or alias definitions. As a result, explicit template instantiations | |
2693 | are required when using C++. This will make it difficult if not | |
2694 | impossible to build many C++ applications. You also can't generate | |
2695 | debugging information when using the HP assembler with GCC. | |
2696 | ||
2697 | There are a number of issues to consider in selecting which linker to | |
2698 | use with the 64-bit port. The GNU 64-bit linker can only create dynamic | |
2699 | binaries. The @option{-static} option causes linking with archive | |
2700 | libraries but doesn't produce a truly static binary. Dynamic binaries | |
2701 | still require final binding by the dynamic loader to resolve a set of | |
2702 | dynamic-loader-defined symbols. The default behavior of the HP linker | |
2703 | is the same as the GNU linker. However, it can generate true 64-bit | |
2704 | static binaries using the @option{+compat} option. | |
2705 | ||
2706 | The HP 64-bit linker doesn't support linkonce semantics. As a | |
2707 | result, C++ programs have many more sections than they should. | |
2708 | ||
2709 | The GNU 64-bit linker has some issues with shared library support | |
2710 | and exceptions. As a result, we only support libgcc in archive | |
2711 | format. For similar reasons, dwarf2 unwind and exception support | |
2712 | are disabled. The GNU linker also has problems creating binaries | |
2713 | with @option{-static}. It doesn't provide stubs for internal | |
2714 | calls to global functions in shared libraries, so these calls | |
2715 | can't be overloaded. | |
2716 | ||
2717 | There are several possible approaches to building the distribution. | |
2718 | Binutils can be built first using the HP tools. Then, the GCC | |
2719 | distribution can be built. The second approach is to build GCC | |
2720 | first using the HP tools, then build binutils, then rebuild GCC. | |
2721 | There have been problems with various binary distributions, so | |
2722 | it is best not to start from a binary distribution. | |
2723 | ||
2724 | Starting with GCC 3.4 an ISO C compiler is required to bootstrap. | |
2725 | The bundled compiler supports only traditional C; you will need | |
2726 | either HP's unbundled compiler, or a binary distribution of GCC@. | |
2727 | ||
2728 | This port still is undergoing significant development. | |
2729 | ||
2730 | @html | |
2731 | <hr /> | |
2732 | @end html | |
2733 | @heading @anchor{*-*-linux-gnu}*-*-linux-gnu | |
2734 | ||
2735 | Versions of libstdc++-v3 starting with 3.2.1 require bugfixes present | |
2736 | in glibc 2.2.5 and later. More information is available in the | |
2737 | libstdc++-v3 documentation. | |
2738 | ||
2739 | @html | |
2740 | <hr /> | |
2741 | @end html | |
2742 | @heading @anchor{ix86-*-linux*aout}i?86-*-linux*aout | |
2743 | Use this configuration to generate @file{a.out} binaries on Linux-based | |
2744 | GNU systems. This configuration is being superseded. | |
2745 | ||
2746 | @html | |
2747 | <hr /> | |
2748 | @end html | |
2749 | @heading @anchor{ix86-*-linux*}i?86-*-linux* | |
2750 | ||
2751 | As of GCC 3.3, binutils 2.13.1 or later is required for this platform. | |
2752 | See @uref{http://gcc.gnu.org/PR10877,,bug 10877} for more information. | |
2753 | ||
2754 | If you receive Signal 11 errors when building on GNU/Linux, then it is | |
2755 | possible you have a hardware problem. Further information on this can be | |
2756 | found on @uref{http://www.bitwizard.nl/sig11/,,www.bitwizard.nl}. | |
2757 | ||
2758 | @html | |
2759 | <hr /> | |
2760 | @end html | |
2761 | @heading @anchor{ix86-*-sco3.2v5*}i?86-*-sco3.2v5* | |
2762 | Use this for the SCO OpenServer Release 5 family of operating systems. | |
2763 | ||
2764 | Unlike earlier versions of GCC, the ability to generate COFF with this | |
2765 | target is no longer provided. | |
2766 | ||
2767 | Earlier versions of GCC emitted DWARF 1 when generating ELF to allow | |
2768 | the system debugger to be used. That support was too burdensome to | |
2769 | maintain. GCC now emits only DWARF 2 for this target. This means you | |
2770 | may use either the UDK debugger or GDB to debug programs built by this | |
2771 | version of GCC@. | |
2772 | ||
2773 | GCC is now only supported on releases 5.0.4 and later, and requires that | |
2774 | you install Support Level Supplement OSS646B or later, and Support Level | |
2775 | Supplement OSS631C or later. If you are using release 5.0.7 of | |
2776 | OpenServer, you must have at least the first maintenance pack installed | |
2777 | (this includes the relevant portions of OSS646). OSS646, also known as | |
2778 | the "Execution Environment Update", provides updated link editors and | |
2779 | assemblers, as well as updated standard C and math libraries. The C | |
2780 | startup modules are also updated to support the System V gABI draft, and | |
2781 | GCC relies on that behavior. OSS631 provides a collection of commonly | |
2782 | used open source libraries, some of which GCC depends on (such as GNU | |
2783 | gettext and zlib). SCO OpenServer Release 5.0.7 has all of this built | |
2784 | in by default, but OSS631C and later also apply to that release. Please | |
2785 | visit | |
2786 | @uref{ftp://ftp.sco.com/pub/openserver5,,ftp://ftp.sco.com/pub/openserver5} | |
2787 | for the latest versions of these (and other potentially useful) | |
2788 | supplements. | |
2789 | ||
2790 | Although there is support for using the native assembler, it is | |
2791 | recommended that you configure GCC to use the GNU assembler. You do | |
2792 | this by using the flags | |
2793 | @uref{./configure.html#with-gnu-as,,@option{--with-gnu-as}}. You should | |
2794 | use a modern version of GNU binutils. Version 2.13.2.1 was used for all | |
2795 | testing. In general, only the @option{--with-gnu-as} option is tested. | |
2796 | A modern bintuils (as well as a plethora of other development related | |
2797 | GNU utilities) can be found in Support Level Supplement OSS658A, the | |
2798 | "GNU Development Tools" package. See the SCO web and ftp sites for details. | |
2799 | That package also contains the currently "officially supported" version of | |
2800 | GCC, version 2.95.3. It is useful for bootstrapping this version. | |
2801 | ||
2802 | @html | |
2803 | <hr /> | |
2804 | @end html | |
2805 | @heading @anchor{ix86-*-udk}i?86-*-udk | |
2806 | ||
2807 | This target emulates the SCO Universal Development Kit and requires that | |
2808 | package be installed. (If it is installed, you will have a | |
2809 | @file{/udk/usr/ccs/bin/cc} file present.) It's very much like the | |
2810 | @samp{i?86-*-unixware7*} target | |
2811 | but is meant to be used when hosting on a system where UDK isn't the | |
2812 | default compiler such as OpenServer 5 or Unixware 2. This target will | |
2813 | generate binaries that will run on OpenServer, Unixware 2, or Unixware 7, | |
2814 | with the same warnings and caveats as the SCO UDK@. | |
2815 | ||
2816 | This target is a little tricky to build because we have to distinguish | |
2817 | it from the native tools (so it gets headers, startups, and libraries | |
2818 | from the right place) while making the tools not think we're actually | |
2819 | building a cross compiler. The easiest way to do this is with a configure | |
2820 | command like this: | |
2821 | ||
2822 | @smallexample | |
2823 | CC=/udk/usr/ccs/bin/cc @var{/your/path/to}/gcc/configure \ | |
2824 | --host=i686-pc-udk --target=i686-pc-udk --program-prefix=udk- | |
2825 | @end smallexample | |
2826 | ||
2827 | @emph{You should substitute @samp{i686} in the above command with the appropriate | |
2828 | processor for your host.} | |
2829 | ||
2830 | After the usual @samp{make bootstrap} and | |
2831 | @samp{make install}, you can then access the UDK-targeted GCC | |
2832 | tools by adding @command{udk-} before the commonly known name. For | |
2833 | example, to invoke the C compiler, you would use @command{udk-gcc}. | |
2834 | They will coexist peacefully with any native-target GCC tools you may | |
2835 | have installed. | |
2836 | ||
2837 | ||
2838 | @html | |
2839 | <hr /> | |
2840 | @end html | |
2841 | @heading @anchor{ia64-*-linux}ia64-*-linux | |
2842 | IA-64 processor (also known as IPF, or Itanium Processor Family) | |
2843 | running GNU/Linux. | |
2844 | ||
2845 | If you are using the optional libunwind library, then you must use | |
2846 | libunwind 0.96 or later. | |
2847 | ||
2848 | None of the following versions of GCC has an ABI that is compatible | |
2849 | with any of the other versions in this list, with the exception that | |
2850 | Red Hat 2.96 and Trillian 000171 are compatible with each other: | |
2851 | 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717. | |
2852 | This primarily affects C++ programs and programs that create shared libraries. | |
2853 | GCC 3.1 or later is recommended for compiling linux, the kernel. | |
2854 | As of version 3.1 GCC is believed to be fully ABI compliant, and hence no | |
2855 | more major ABI changes are expected. | |
2856 | ||
2857 | @html | |
2858 | <hr /> | |
2859 | @end html | |
2860 | @heading @anchor{ia64-*-hpux*}ia64-*-hpux* | |
2861 | Building GCC on this target requires the GNU Assembler. The bundled HP | |
2862 | assembler will not work. To prevent GCC from using the wrong assembler, | |
2863 | the option @option{--with-gnu-as} may be necessary. | |
2864 | ||
2865 | The GCC libunwind library has not been ported to HPUX. This means that for | |
2866 | GCC versions 3.2.3 and earlier, @option{--enable-libunwind-exceptions} | |
2867 | is required to build GCC. For GCC 3.3 and later, this is the default. | |
2868 | ||
2869 | @html | |
2870 | <hr /> | |
2871 | <!-- rs6000-ibm-aix*, powerpc-ibm-aix* --> | |
2872 | @end html | |
2873 | @heading @anchor{*-ibm-aix*}*-ibm-aix* | |
2874 | Support for AIX version 3 and older was discontinued in GCC 3.4. | |
2875 | ||
2876 | AIX Make frequently has problems with GCC makefiles. GNU Make 3.79.1 or | |
2877 | newer is recommended to build on this platform. | |
2878 | ||
2879 | To speed up the configuration phases of bootstrapping and installing GCC, | |
2880 | one may use GNU Bash instead of AIX @command{/bin/sh}, e.g., | |
2881 | ||
2882 | @smallexample | |
2883 | % CONFIG_SHELL=/opt/freeware/bin/bash | |
2884 | % export CONFIG_SHELL | |
2885 | @end smallexample | |
2886 | ||
2887 | and then proceed as described in @uref{build.html,,the build instructions}, | |
2888 | where we strongly recommend using GNU make and specifying an absolute path | |
2889 | to invoke @var{srcdir}/configure. | |
2890 | ||
2891 | Errors involving @code{alloca} when building GCC generally are due | |
2892 | to an incorrect definition of @code{CC} in the Makefile or mixing files | |
2893 | compiled with the native C compiler and GCC@. During the stage1 phase of | |
2894 | the build, the native AIX compiler @strong{must} be invoked as @command{cc} | |
2895 | (not @command{xlc}). Once @command{configure} has been informed of | |
2896 | @command{xlc}, one needs to use @samp{make distclean} to remove the | |
2897 | configure cache files and ensure that @env{CC} environment variable | |
2898 | does not provide a definition that will confuse @command{configure}. | |
2899 | If this error occurs during stage2 or later, then the problem most likely | |
2900 | is the version of Make (see above). | |
2901 | ||
2902 | The native @command{as} and @command{ld} are recommended for bootstrapping | |
2903 | on AIX 4 and required for bootstrapping on AIX 5L. The GNU Assembler | |
2904 | reports that it supports WEAK symbols on AIX 4, which causes GCC to try to | |
2905 | utilize weak symbol functionality although it is not supported. The GNU | |
2906 | Assembler and Linker do not support AIX 5L sufficiently to bootstrap GCC. | |
2907 | The native AIX tools do interoperate with GCC@. | |
2908 | ||
2909 | Building @file{libstdc++.a} requires a fix for an AIX Assembler bug | |
2910 | APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a | |
2911 | fix for another AIX Assembler bug and a co-dependent AIX Archiver fix | |
2912 | referenced as APAR IY53606 (AIX 5.2) or a APAR TBD (AIX 5.1) | |
2913 | ||
2914 | @samp{libstdc++} in GCC 3.4 increments the major version number of the | |
2915 | shared object and GCC installation places the @file{libstdc++.a} | |
2916 | shared library in a common location which will overwrite the and GCC | |
2917 | 3.3 version of the shared library. Applications either need to be | |
2918 | re-linked against the new shared library or the GCC 3.1 and GCC 3.3 | |
2919 | versions of the @samp{libstdc++} shared object needs to be available | |
2920 | to the AIX runtime loader. The GCC 3.1 @samp{libstdc++.so.4}, if | |
2921 | present, and GCC 3.3 @samp{libstdc++.so.5} shared objects can be | |
2922 | installed for runtime dynamic loading using the following steps to set | |
2923 | the @samp{F_LOADONLY} flag in the shared object for @emph{each} | |
2924 | multilib @file{libstdc++.a} installed: | |
2925 | ||
2926 | Extract the shared objects from the currently installed | |
2927 | @file{libstdc++.a} archive: | |
2928 | @smallexample | |
2929 | % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5 | |
2930 | @end smallexample | |
2931 | ||
2932 | Enable the @samp{F_LOADONLY} flag so that the shared object will be | |
2933 | available for runtime dynamic loading, but not linking: | |
2934 | @smallexample | |
2935 | % strip -e libstdc++.so.4 libstdc++.so.5 | |
2936 | @end smallexample | |
2937 | ||
2938 | Archive the runtime-only shared object in the GCC 3.4 | |
2939 | @file{libstdc++.a} archive: | |
2940 | @smallexample | |
2941 | % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5 | |
2942 | @end smallexample | |
2943 | ||
2944 | Linking executables and shared libraries may produce warnings of | |
2945 | duplicate symbols. The assembly files generated by GCC for AIX always | |
2946 | have included multiple symbol definitions for certain global variable | |
2947 | and function declarations in the original program. The warnings should | |
2948 | not prevent the linker from producing a correct library or runnable | |
2949 | executable. | |
2950 | ||
2951 | AIX 4.3 utilizes a ``large format'' archive to support both 32-bit and | |
2952 | 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1 | |
2953 | to parse archive libraries did not handle the new format correctly. | |
2954 | These routines are used by GCC and result in error messages during | |
2955 | linking such as ``not a COFF file''. The version of the routines shipped | |
2956 | with AIX 4.3.1 should work for a 32-bit environment. The @option{-g} | |
2957 | option of the archive command may be used to create archives of 32-bit | |
2958 | objects using the original ``small format''. A correct version of the | |
2959 | routines is shipped with AIX 4.3.2 and above. | |
2960 | ||
2961 | Some versions of the AIX binder (linker) can fail with a relocation | |
2962 | overflow severe error when the @option{-bbigtoc} option is used to link | |
2963 | GCC-produced object files into an executable that overflows the TOC@. A fix | |
2964 | for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is | |
2965 | available from IBM Customer Support and from its | |
2966 | @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com} | |
2967 | website as PTF U455193. | |
2968 | ||
2969 | The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core | |
2970 | with a segmentation fault when invoked by any version of GCC@. A fix for | |
2971 | APAR IX87327 is available from IBM Customer Support and from its | |
2972 | @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com} | |
2973 | website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above. | |
2974 | ||
2975 | The initial assembler shipped with AIX 4.3.0 generates incorrect object | |
2976 | files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS | |
2977 | TO ASSEMBLE/BIND) is available from IBM Customer Support and from its | |
2978 | @uref{http://techsupport.services.ibm.com/,,techsupport.services.ibm.com} | |
2979 | website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above. | |
2980 | ||
2981 | AIX provides National Language Support (NLS)@. Compilers and assemblers | |
2982 | use NLS to support locale-specific representations of various data | |
2983 | formats including floating-point numbers (e.g., @samp{.} vs @samp{,} for | |
2984 | separating decimal fractions). There have been problems reported where | |
2985 | GCC does not produce the same floating-point formats that the assembler | |
2986 | expects. If one encounters this problem, set the @env{LANG} | |
2987 | environment variable to @samp{C} or @samp{En_US}. | |
2988 | ||
2989 | By default, GCC for AIX 4.1 and above produces code that can be used on | |
2990 | both Power or PowerPC processors. | |
2991 | ||
2992 | A default can be specified with the @option{-mcpu=@var{cpu_type}} | |
2993 | switch and using the configure option @option{--with-cpu-@var{cpu_type}}. | |
2994 | ||
2995 | @html | |
2996 | <hr /> | |
2997 | @end html | |
2998 | @heading @anchor{ip2k-*-elf}ip2k-*-elf | |
2999 | Ubicom IP2022 micro controller. | |
3000 | This configuration is intended for embedded systems. | |
3001 | There are no standard Unix configurations. | |
3002 | ||
3003 | Use @samp{configure --target=ip2k-elf --enable-languages=c} to configure GCC@. | |
3004 | ||
3005 | @html | |
3006 | <hr /> | |
3007 | @end html | |
3008 | @heading @anchor{iq2000-*-elf}iq2000-*-elf | |
3009 | Vitesse IQ2000 processors. These are used in embedded | |
3010 | applications. There are no standard Unix configurations. | |
3011 | ||
3012 | @html | |
3013 | <hr /> | |
3014 | @end html | |
3015 | @heading @anchor{m32r-*-elf}m32r-*-elf | |
3016 | Renesas M32R processor. | |
3017 | This configuration is intended for embedded systems. | |
3018 | ||
3019 | @html | |
3020 | <hr /> | |
3021 | @end html | |
3022 | @heading @anchor{m6811-elf}m6811-elf | |
3023 | Motorola 68HC11 family micro controllers. These are used in embedded | |
3024 | applications. There are no standard Unix configurations. | |
3025 | ||
3026 | @html | |
3027 | <hr /> | |
3028 | @end html | |
3029 | @heading @anchor{m6812-elf}m6812-elf | |
3030 | Motorola 68HC12 family micro controllers. These are used in embedded | |
3031 | applications. There are no standard Unix configurations. | |
3032 | ||
3033 | @html | |
3034 | <hr /> | |
3035 | @end html | |
3036 | @heading @anchor{m68k-hp-hpux}m68k-hp-hpux | |
3037 | HP 9000 series 300 or 400 running HP-UX@. HP-UX version 8.0 has a bug in | |
3038 | the assembler that prevents compilation of GCC@. This | |
3039 | bug manifests itself during the first stage of compilation, while | |
3040 | building @file{libgcc2.a}: | |
3041 | ||
3042 | @smallexample | |
3043 | _floatdisf | |
3044 | cc1: warning: `-g' option not supported on this version of GCC | |
3045 | cc1: warning: `-g1' option not supported on this version of GCC | |
3046 | ./xgcc: Internal compiler error: program as got fatal signal 11 | |
3047 | @end smallexample | |
3048 | ||
3049 | A patched version of the assembler is available as the file | |
3050 | @uref{ftp://altdorf.ai.mit.edu/archive/cph/hpux-8.0-assembler}. If you | |
3051 | have HP software support, the patch can also be obtained directly from | |
3052 | HP, as described in the following note: | |
3053 | ||
3054 | @quotation | |
3055 | This is the patched assembler, to patch SR#1653-010439, where the | |
3056 | assembler aborts on floating point constants. | |
3057 | ||
3058 | The bug is not really in the assembler, but in the shared library | |
3059 | version of the function ``cvtnum(3c)''. The bug on ``cvtnum(3c)'' is | |
3060 | SR#4701-078451. Anyway, the attached assembler uses the archive | |
3061 | library version of ``cvtnum(3c)'' and thus does not exhibit the bug. | |
3062 | @end quotation | |
3063 | ||
3064 | This patch is also known as PHCO_4484. | |
3065 | ||
3066 | In addition gdb does not understand that native HP-UX format, so | |
3067 | you must use gas if you wish to use gdb. | |
3068 | ||
3069 | On HP-UX version 8.05, but not on 8.07 or more recent versions, the | |
3070 | @command{fixproto} shell script triggers a bug in the system shell. If you | |
3071 | encounter this problem, upgrade your operating system or use BASH (the | |
3072 | GNU shell) to run @command{fixproto}. This bug will cause the fixproto | |
3073 | program to report an error of the form: | |
3074 | ||
3075 | @smallexample | |
3076 | ./fixproto: sh internal 1K buffer overflow | |
3077 | @end smallexample | |
3078 | ||
3079 | To fix this, you can also change the first line of the fixproto script | |
3080 | to look like: | |
3081 | ||
3082 | @smallexample | |
3083 | #!/bin/ksh | |
3084 | @end smallexample | |
3085 | ||
3086 | @html | |
3087 | <hr /> | |
3088 | @end html | |
3089 | @heading @anchor{mips-*-*}mips-*-* | |
3090 | If on a MIPS system you get an error message saying ``does not have gp | |
3091 | sections for all it's [sic] sectons [sic]'', don't worry about it. This | |
3092 | happens whenever you use GAS with the MIPS linker, but there is not | |
3093 | really anything wrong, and it is okay to use the output file. You can | |
3094 | stop such warnings by installing the GNU linker. | |
3095 | ||
3096 | It would be nice to extend GAS to produce the gp tables, but they are | |
3097 | optional, and there should not be a warning about their absence. | |
3098 | ||
3099 | The libstdc++ atomic locking routines for MIPS targets requires MIPS II | |
3100 | and later. A patch went in just after the GCC 3.3 release to | |
3101 | make @samp{mips*-*-*} use the generic implementation instead. You can also | |
3102 | configure for @samp{mipsel-elf} as a workaround. The | |
3103 | @samp{mips*-*-linux*} target continues to use the MIPS II routines. More | |
3104 | work on this is expected in future releases. | |
3105 | ||
3106 | Cross-compilers for the Mips as target using the Mips assembler | |
3107 | currently do not work, because the auxiliary programs | |
3108 | @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on | |
3109 | anything but a Mips. It does work to cross compile for a Mips | |
3110 | if you use the GNU assembler and linker. | |
3111 | ||
3112 | @html | |
3113 | <hr /> | |
3114 | @end html | |
3115 | @heading @anchor{mips-sgi-irix5}mips-sgi-irix5 | |
3116 | ||
3117 | This configuration has considerable problems, which will be fixed in a | |
3118 | future release. | |
3119 | ||
3120 | In order to compile GCC on an SGI running IRIX 5, the ``compiler_dev.hdr'' | |
3121 | subsystem must be installed from the IDO CD-ROM supplied by Silicon | |
3122 | Graphics. It is also available for download from | |
3123 | @uref{http://www.sgi.com/developers/devtools/apis/ido.html,,http://www.sgi.com/developers/devtools/apis/ido.html}. | |
3124 | ||
3125 | @samp{make compare} may fail on version 5 of IRIX unless you add | |
3126 | @option{-save-temps} to @code{CFLAGS}. On these systems, the name of the | |
3127 | assembler input file is stored in the object file, and that makes | |
3128 | comparison fail if it differs between the @code{stage1} and | |
3129 | @code{stage2} compilations. The option @option{-save-temps} forces a | |
3130 | fixed name to be used for the assembler input file, instead of a | |
3131 | randomly chosen name in @file{/tmp}. Do not add @option{-save-temps} | |
3132 | unless the comparisons fail without that option. If you do you | |
3133 | @option{-save-temps}, you will have to manually delete the @samp{.i} and | |
3134 | @samp{.s} files after each series of compilations. | |
3135 | ||
3136 | If you use the MIPS C compiler to bootstrap, it may be necessary | |
3137 | to increase its table size for switch statements with the | |
3138 | @option{-Wf,-XNg1500} option. If you use the @option{-O2} | |
3139 | optimization option, you also need to use @option{-Olimit 3000}. | |
3140 | ||
3141 | To enable debugging under IRIX 5, you must use GNU @command{as} 2.11.2 | |
3142 | or later, | |
3143 | and use the @option{--with-gnu-as} configure option when configuring GCC. | |
3144 | GNU @command{as} is distributed as part of the binutils package. | |
3145 | When using release 2.11.2, you need to apply a patch | |
3146 | @uref{http://sources.redhat.com/ml/binutils/2001-07/msg00352.html,,http://sources.redhat.com/ml/binutils/2001-07/msg00352.html} | |
3147 | which will be included in the next release of binutils. | |
3148 | ||
3149 | When building GCC, the build process loops rebuilding @command{cc1} over | |
3150 | and over again. This happens on @samp{mips-sgi-irix5.2}, and possibly | |
3151 | other platforms. It has been reported that this is a known bug in the | |
3152 | @command{make} shipped with IRIX 5.2. We recommend you use GNU | |
3153 | @command{make} instead of the vendor supplied @command{make} program; | |
3154 | however, you may have success with @command{smake} on IRIX 5.2 if you do | |
3155 | not have GNU @command{make} available. | |
3156 | ||
3157 | @html | |
3158 | <hr /> | |
3159 | @end html | |
3160 | @heading @anchor{mips-sgi-irix6}mips-sgi-irix6 | |
3161 | ||
3162 | If you are using IRIX @command{cc} as your bootstrap compiler, you must | |
3163 | ensure that the N32 ABI is in use. To test this, compile a simple C | |
3164 | file with @command{cc} and then run @command{file} on the | |
3165 | resulting object file. The output should look like: | |
3166 | ||
3167 | @smallexample | |
3168 | test.o: ELF N32 MSB @dots{} | |
3169 | @end smallexample | |
3170 | ||
3171 | If you see: | |
3172 | ||
3173 | @smallexample | |
3174 | test.o: ELF 32-bit MSB @dots{} | |
3175 | @end smallexample | |
3176 | ||
3177 | or | |
3178 | ||
3179 | @smallexample | |
3180 | test.o: ELF 64-bit MSB @dots{} | |
3181 | @end smallexample | |
3182 | ||
3183 | then your version of @command{cc} uses the O32 or N64 ABI by default. You | |
3184 | should set the environment variable @env{CC} to @samp{cc -n32} | |
3185 | before configuring GCC@. | |
3186 | ||
3187 | If you want the resulting @command{gcc} to run on old 32-bit systems | |
3188 | with the MIPS R4400 CPU, you need to ensure that only code for the mips3 | |
3189 | instruction set architecture (ISA) is generated. While GCC 3.x does | |
3190 | this correctly, both GCC 2.95 and SGI's MIPSpro @command{cc} may change | |
3191 | the ISA depending on the machine where GCC is built. Using one of them | |
3192 | as the bootstrap compiler may result in mips4 code, which won't run at | |
3193 | all on mips3-only systems. For the test program above, you should see: | |
3194 | ||
3195 | @smallexample | |
3196 | test.o: ELF N32 MSB mips-3 @dots{} | |
3197 | @end smallexample | |
3198 | ||
3199 | If you get: | |
3200 | ||
3201 | @smallexample | |
3202 | test.o: ELF N32 MSB mips-4 @dots{} | |
3203 | @end smallexample | |
3204 | ||
3205 | instead, you should set the environment variable @env{CC} to @samp{cc | |
3206 | -n32 -mips3} or @samp{gcc -mips3} respectively before configuring GCC@. | |
3207 | ||
3208 | GCC on IRIX 6 is usually built to support both the N32 and N64 ABIs. If | |
3209 | you build GCC on a system that doesn't have the N64 libraries installed, | |
3210 | you need to configure with @option{--disable-multilib} so GCC doesn't | |
3211 | try to use them. Look for @file{/usr/lib64/libc.so.1} to see if you | |
3212 | have the 64-bit libraries installed. | |
3213 | ||
3214 | You must @emph{not} use GNU @command{as} (which isn't built anyway as of | |
3215 | binutils 2.11.2) on IRIX 6 platforms; doing so will only cause problems. | |
3216 | ||
3217 | GCC does not currently support generating O32 ABI binaries in the | |
3218 | @samp{mips-sgi-irix6} configurations. It is possible to create a GCC | |
3219 | with O32 ABI only support by configuring it for the @samp{mips-sgi-irix5} | |
3220 | target and using a patched GNU @command{as} 2.11.2 as documented in the | |
3221 | @uref{#mips-sgi-irix5,,@samp{mips-sgi-irix5}} section above. Using the | |
3222 | native assembler requires patches to GCC which will be included in a | |
3223 | future release. It is | |
3224 | expected that O32 ABI support will be available again in a future release. | |
3225 | ||
3226 | The @option{--enable-threads} option doesn't currently work, a patch is | |
3227 | in preparation for a future release. The @option{--enable-libgcj} | |
3228 | option is disabled by default: IRIX 6 uses a very low default limit | |
3229 | (20480) for the command line length. Although libtool contains a | |
3230 | workaround for this problem, at least the N64 @samp{libgcj} is known not | |
3231 | to build despite this, running into an internal error of the native | |
3232 | @command{ld}. A sure fix is to increase this limit (@samp{ncargs}) to | |
3233 | its maximum of 262144 bytes. If you have root access, you can use the | |
3234 | @command{systune} command to do this. | |
3235 | ||
3236 | GCC does not correctly pass/return structures which are | |
3237 | smaller than 16 bytes and which are not 8 bytes. The problem is very | |
3238 | involved and difficult to fix. It affects a number of other targets also, | |
3239 | but IRIX 6 is affected the most, because it is a 64-bit target, and 4 byte | |
3240 | structures are common. The exact problem is that structures are being padded | |
3241 | at the wrong end, e.g.@: a 4 byte structure is loaded into the lower 4 bytes | |
3242 | of the register when it should be loaded into the upper 4 bytes of the | |
3243 | register. | |
3244 | ||
3245 | GCC is consistent with itself, but not consistent with the SGI C compiler | |
3246 | (and the SGI supplied runtime libraries), so the only failures that can | |
3247 | happen are when there are library functions that take/return such | |
3248 | structures. There are very few such library functions. Currently this | |
3249 | is known to affect @code{inet_ntoa}, @code{inet_lnaof}, | |
3250 | @code{inet_netof}, @code{inet_makeaddr}, and @code{semctl}. Until the | |
3251 | bug is fixed, GCC contains workarounds for the known affected functions. | |
3252 | ||
3253 | See @uref{http://freeware.sgi.com/,,http://freeware.sgi.com/} for more | |
3254 | information about using GCC on IRIX platforms. | |
3255 | ||
3256 | @html | |
3257 | <hr /> | |
3258 | @end html | |
3259 | @heading @anchor{powerpc*-*-*}powerpc-*-* | |
3260 | ||
3261 | You can specify a default version for the @option{-mcpu=@var{cpu_type}} | |
3262 | switch by using the configure option @option{--with-cpu-@var{cpu_type}}. | |
3263 | ||
3264 | @html | |
3265 | <hr /> | |
3266 | @end html | |
3267 | @heading @anchor{powerpc-*-darwin*}powerpc-*-darwin* | |
3268 | PowerPC running Darwin (Mac OS X kernel). | |
3269 | ||
3270 | Pre-installed versions of Mac OS X may not include any developer tools, | |
3271 | meaning that you will not be able to build GCC from source. Tool | |
3272 | binaries are available at | |
3273 | @uref{http://developer.apple.com/tools/compilers.html} (free | |
3274 | registration required). | |
3275 | ||
3276 | The default stack limit of 512K is too small, which may cause compiles | |
3277 | to fail with 'Bus error'. Set the stack larger, for instance | |
3278 | by doing @samp{limit stack 800}. It's a good idea to use the GNU | |
3279 | preprocessor instead of Apple's @file{cpp-precomp} during the first stage of | |
3280 | bootstrapping; this is automatic when doing @samp{make bootstrap}, but | |
3281 | to do it from the toplevel objdir you will need to say @samp{make | |
3282 | CC='cc -no-cpp-precomp' bootstrap}. | |
3283 | ||
3284 | The version of GCC shipped by Apple typically includes a number of | |
3285 | extensions not available in a standard GCC release. These extensions | |
3286 | are generally specific to Mac programming. | |
3287 | ||
3288 | @html | |
3289 | <hr /> | |
3290 | @end html | |
3291 | @heading @anchor{powerpc-*-elf}powerpc-*-elf, powerpc-*-sysv4 | |
3292 | PowerPC system in big endian mode, running System V.4. | |
3293 | ||
3294 | @html | |
3295 | <hr /> | |
3296 | @end html | |
3297 | @heading @anchor{powerpc-*-linux-gnu*}powerpc-*-linux-gnu* | |
3298 | ||
3299 | You will need | |
3300 | @uref{ftp://ftp.kernel.org/pub/linux/devel/binutils,,binutils 2.13.90.0.10} | |
3301 | or newer for a working GCC@. | |
3302 | ||
3303 | @html | |
3304 | <hr /> | |
3305 | @end html | |
3306 | @heading @anchor{powerpc-*-netbsd*}powerpc-*-netbsd* | |
3307 | PowerPC system in big endian mode running NetBSD@. To build the | |
3308 | documentation you will need Texinfo version 4.2 (NetBSD 1.5.1 included | |
3309 | Texinfo version 3.12). | |
3310 | ||
3311 | @html | |
3312 | <hr /> | |
3313 | @end html | |
3314 | @heading @anchor{powerpc-*-eabisim}powerpc-*-eabisim | |
3315 | Embedded PowerPC system in big endian mode for use in running under the | |
3316 | PSIM simulator. | |
3317 | ||
3318 | @html | |
3319 | <hr /> | |
3320 | @end html | |
3321 | @heading @anchor{powerpc-*-eabi}powerpc-*-eabi | |
3322 | Embedded PowerPC system in big endian mode. | |
3323 | ||
3324 | @html | |
3325 | <hr /> | |
3326 | @end html | |
3327 | @heading @anchor{powerpcle-*-elf}powerpcle-*-elf, powerpcle-*-sysv4 | |
3328 | PowerPC system in little endian mode, running System V.4. | |
3329 | ||
3330 | @html | |
3331 | <hr /> | |
3332 | @end html | |
3333 | @heading @anchor{powerpcle-*-eabisim}powerpcle-*-eabisim | |
3334 | Embedded PowerPC system in little endian mode for use in running under | |
3335 | the PSIM simulator. | |
3336 | ||
3337 | @html | |
3338 | <hr /> | |
3339 | @end html | |
3340 | @heading @anchor{powerpcle-*-eabi}powerpcle-*-eabi | |
3341 | Embedded PowerPC system in little endian mode. | |
3342 | ||
3343 | @html | |
3344 | <hr /> | |
3345 | @end html | |
3346 | @heading @anchor{s390-*-linux*}s390-*-linux* | |
3347 | S/390 system running GNU/Linux for S/390@. | |
3348 | ||
3349 | @html | |
3350 | <hr /> | |
3351 | @end html | |
3352 | @heading @anchor{s390x-*-linux*}s390x-*-linux* | |
3353 | zSeries system (64-bit) running GNU/Linux for zSeries@. | |
3354 | ||
3355 | @html | |
3356 | <hr /> | |
3357 | @end html | |
3358 | @heading @anchor{s390x-ibm-tpf*}s390x-ibm-tpf* | |
3359 | zSeries system (64-bit) running TPF. This platform is | |
3360 | supported as cross-compilation target only. | |
3361 | ||
3362 | @html | |
3363 | <hr /> | |
3364 | @end html | |
3365 | @c Please use Solaris 2 to refer to all release of Solaris, starting | |
3366 | @c with 2.0 until 2.6, 7, and 8. Solaris 1 was a marketing name for | |
3367 | @c SunOS 4 releases which we don't use to avoid confusion. Solaris | |
3368 | @c alone is too unspecific and must be avoided. | |
3369 | @heading @anchor{*-*-solaris2*}*-*-solaris2* | |
3370 | ||
3371 | Sun does not ship a C compiler with Solaris 2. To bootstrap and install | |
3372 | GCC you first have to install a pre-built compiler, see our | |
3373 | @uref{binaries.html,,binaries page} for details. | |
3374 | ||
3375 | The Solaris 2 @command{/bin/sh} will often fail to configure | |
3376 | @file{libstdc++-v3}, @file{boehm-gc} or @file{libjava}. We therefore | |
3377 | recommend to use the following sequence of commands to bootstrap and | |
3378 | install GCC: | |
3379 | ||
3380 | @smallexample | |
3381 | % CONFIG_SHELL=/bin/ksh | |
3382 | % export CONFIG_SHELL | |
3383 | @end smallexample | |
3384 | ||
3385 | and then proceed as described in @uref{build.html,,the build instructions}, | |
3386 | where we strongly recommend using GNU make and specifying an absolute path | |
3387 | to invoke @var{srcdir}/configure. | |
3388 | ||
3389 | Solaris 2 comes with a number of optional OS packages. Some of these | |
3390 | are needed to use GCC fully, namely @code{SUNWarc}, | |
3391 | @code{SUNWbtool}, @code{SUNWesu}, @code{SUNWhea}, @code{SUNWlibm}, | |
3392 | @code{SUNWsprot}, and @code{SUNWtoo}. If you did not install all | |
3393 | optional packages when installing Solaris 2, you will need to verify that | |
3394 | the packages that GCC needs are installed. | |
3395 | ||
3396 | To check whether an optional package is installed, use | |
3397 | the @command{pkginfo} command. To add an optional package, use the | |
3398 | @command{pkgadd} command. For further details, see the Solaris 2 | |
3399 | documentation. | |
3400 | ||
3401 | Trying to use the linker and other tools in | |
3402 | @file{/usr/ucb} to install GCC has been observed to cause trouble. | |
3403 | For example, the linker may hang indefinitely. The fix is to remove | |
3404 | @file{/usr/ucb} from your @env{PATH}. | |
3405 | ||
3406 | The build process works more smoothly with the legacy Sun tools so, if you | |
3407 | have @file{/usr/xpg4/bin} in your @env{PATH}, we recommend that you place | |
3408 | @file{/usr/bin} before @file{/usr/xpg4/bin} for the duration of the build. | |
3409 | ||
3410 | All releases of GNU binutils prior to 2.11.2 have known bugs on this | |
3411 | platform. We recommend the use of GNU binutils 2.11.2 or the vendor | |
3412 | tools (Sun @command{as}, Sun @command{ld}). | |
3413 | ||
3414 | Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or | |
3415 | newer: @command{g++} will complain that types are missing. These headers assume | |
3416 | that omitting the type means @code{int}; this assumption worked for C89 but | |
3417 | is wrong for C++, and is now wrong for C99 also. | |
3418 | ||
3419 | @command{g++} accepts such (invalid) constructs with the option | |
3420 | @option{-fpermissive}; it | |
3421 | will assume that any missing type is @code{int} (as defined by C89). | |
3422 | ||
3423 | There are patches for Solaris 2.6 (105633-56 or newer for SPARC, | |
3424 | 106248-42 or newer for Intel), Solaris 7 (108376-21 or newer for SPARC, | |
3425 | 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC, | |
3426 | 108653-22 for Intel) that fix this bug. | |
3427 | ||
3428 | @html | |
3429 | <hr /> | |
3430 | @end html | |
3431 | @heading @anchor{sparc-sun-solaris2*}sparc-sun-solaris2* | |
3432 | ||
3433 | When GCC is configured to use binutils 2.11.2 or later the binaries | |
3434 | produced are smaller than the ones produced using Sun's native tools; | |
3435 | this difference is quite significant for binaries containing debugging | |
3436 | information. | |
3437 | ||
3438 | Sun @command{as} 4.x is broken in that it cannot cope with long symbol names. | |
3439 | A typical error message might look similar to the following: | |
3440 | ||
3441 | @smallexample | |
3442 | /usr/ccs/bin/as: "/var/tmp/ccMsw135.s", line 11041: error: | |
3443 | can't compute value of an expression involving an external symbol. | |
3444 | @end smallexample | |
3445 | ||
3446 | This is Sun bug 4237974. This is fixed with patch 108908-02 for Solaris | |
3447 | 2.6 and has been fixed in later (5.x) versions of the assembler, | |
3448 | starting with Solaris 7. | |
3449 | ||
3450 | Starting with Solaris 7, the operating system is capable of executing | |
3451 | 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports | |
3452 | this; the @option{-m64} option enables 64-bit code generation. | |
3453 | However, if all you want is code tuned for the UltraSPARC CPU, you | |
3454 | should try the @option{-mtune=ultrasparc} option instead, which produces | |
3455 | code that, unlike full 64-bit code, can still run on non-UltraSPARC | |
3456 | machines. | |
3457 | ||
3458 | When configuring on a Solaris 7 or later system that is running a kernel | |
3459 | that supports only 32-bit binaries, one must configure with | |
3460 | @option{--disable-multilib}, since we will not be able to build the | |
3461 | 64-bit target libraries. | |
3462 | ||
3463 | GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7) | |
3464 | and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap | |
3465 | failure in form of a miscompilation of the stage1 compiler by the Sun | |
3466 | compiler. This is Sun bug 4974440. This is fixed with patch 112760-07. | |
3467 | ||
3468 | @html | |
3469 | <hr /> | |
3470 | @end html | |
3471 | @heading @anchor{sparc-sun-solaris2.7}sparc-sun-solaris2.7 | |
3472 | ||
3473 | Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in | |
3474 | the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8 | |
3475 | and later, including all EGCS releases. Sun formerly recommended | |
3476 | 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to | |
3477 | recommend it only for people who use Sun's compilers. | |
3478 | ||
3479 | Here are some workarounds to this problem: | |
3480 | @itemize @bullet | |
3481 | @item | |
3482 | Do not install Sun patch 107058-01 until after Sun releases a | |
3483 | complete patch for bug 4210064. This is the simplest course to take, | |
3484 | unless you must also use Sun's C compiler. Unfortunately 107058-01 | |
3485 | is preinstalled on some new Solaris 7-based hosts, so you may have to | |
3486 | back it out. | |
3487 | ||
3488 | @item | |
3489 | Copy the original, unpatched Solaris 7 | |
3490 | @command{/usr/ccs/bin/as} into | |
3491 | @command{/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as}, | |
3492 | adjusting the latter name to fit your local conventions and software | |
3493 | version numbers. | |
3494 | ||
3495 | @item | |
3496 | Install Sun patch 106950-03 (1999-05-25) or later. Nobody with | |
3497 | both 107058-01 and 106950-03 installed has reported the bug with GCC | |
3498 | and Sun's dynamic linker. This last course of action is riskiest, | |
3499 | for two reasons. First, you must install 106950 on all hosts that | |
3500 | run code generated by GCC; it doesn't suffice to install it only on | |
3501 | the hosts that run GCC itself. Second, Sun says that 106950-03 is | |
3502 | only a partial fix for bug 4210064, but Sun doesn't know whether the | |
3503 | partial fix is adequate for GCC@. Revision -08 or later should fix | |
3504 | the bug. The current (as of 2001-09-24) revision is -14, and is included in | |
3505 | the Solaris 7 Recommended Patch Cluster. | |
3506 | @end itemize | |
3507 | ||
3508 | GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler, | |
3509 | which causes a bootstrap failure when linking the 64-bit shared version of | |
3510 | libgcc. A typical error message is: | |
3511 | ||
3512 | @smallexample | |
3513 | ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o: | |
3514 | symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned. | |
3515 | @end smallexample | |
3516 | ||
3517 | This bug has been fixed in the final 5.0 version of the assembler. | |
3518 | ||
3519 | @html | |
3520 | <hr /> | |
3521 | @end html | |
3522 | @heading @anchor{sparc-*-linux*}sparc-*-linux* | |
3523 | ||
3524 | GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4 | |
3525 | or newer on this platform. All earlier binutils and glibc | |
3526 | releases mishandled unaligned relocations on @code{sparc-*-*} targets. | |
3527 | ||
3528 | ||
3529 | @html | |
3530 | <hr /> | |
3531 | @end html | |
3532 | @heading @anchor{sparc64-*-solaris2*}sparc64-*-solaris2* | |
3533 | ||
3534 | The following compiler flags must be specified in the configure | |
3535 | step in order to bootstrap this target with the Sun compiler: | |
3536 | ||
3537 | @smallexample | |
3538 | % CC="cc -xildoff -xarch=v9" @var{srcdir}/configure [@var{options}] [@var{target}] | |
3539 | @end smallexample | |
3540 | ||
3541 | @option{-xildoff} turns off the incremental linker, and @option{-xarch=v9} | |
3542 | specifies the SPARC-V9 architecture to the Sun linker and assembler. | |
3543 | ||
3544 | @html | |
3545 | <hr /> | |
3546 | @end html | |
3547 | @heading @anchor{sparcv9-*-solaris2*}sparcv9-*-solaris2* | |
3548 | ||
3549 | This is a synonym for sparc64-*-solaris2*. | |
3550 | ||
3551 | @html | |
3552 | <hr /> | |
3553 | @end html | |
3554 | @heading @anchor{#*-*-sysv*}*-*-sysv* | |
3555 | On System V release 3, you may get this error message | |
3556 | while linking: | |
3557 | ||
3558 | @smallexample | |
3559 | ld fatal: failed to write symbol name @var{something} | |
3560 | in strings table for file @var{whatever} | |
3561 | @end smallexample | |
3562 | ||
3563 | This probably indicates that the disk is full or your ulimit won't allow | |
3564 | the file to be as large as it needs to be. | |
3565 | ||
3566 | This problem can also result because the kernel parameter @code{MAXUMEM} | |
3567 | is too small. If so, you must regenerate the kernel and make the value | |
3568 | much larger. The default value is reported to be 1024; a value of 32768 | |
3569 | is said to work. Smaller values may also work. | |
3570 | ||
3571 | On System V, if you get an error like this, | |
3572 | ||
3573 | @smallexample | |
3574 | /usr/local/lib/bison.simple: In function `yyparse': | |
3575 | /usr/local/lib/bison.simple:625: virtual memory exhausted | |
3576 | @end smallexample | |
3577 | ||
3578 | @noindent | |
3579 | that too indicates a problem with disk space, ulimit, or @code{MAXUMEM}. | |
3580 | ||
3581 | On a System V release 4 system, make sure @file{/usr/bin} precedes | |
3582 | @file{/usr/ucb} in @code{PATH}. The @command{cc} command in | |
3583 | @file{/usr/ucb} uses libraries which have bugs. | |
3584 | ||
3585 | @html | |
3586 | <hr /> | |
3587 | @end html | |
3588 | @heading @anchor{vax-dec-ultrix}vax-dec-ultrix | |
3589 | Don't try compiling with VAX C (@command{vcc}). It produces incorrect code | |
3590 | in some cases (for example, when @code{alloca} is used). | |
3591 | ||
3592 | @html | |
3593 | <hr /> | |
3594 | @end html | |
3595 | @heading @anchor{*-*-vxworks*}*-*-vxworks* | |
3596 | Support for VxWorks is in flux. At present GCC supports @emph{only} the | |
3597 | very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC. | |
3598 | We welcome patches for other architectures supported by VxWorks 5.5. | |
3599 | Support for VxWorks AE would also be welcome; we believe this is merely | |
3600 | a matter of writing an appropriate ``configlette'' (see below). We are | |
3601 | not interested in supporting older, a.out or COFF-based, versions of | |
3602 | VxWorks in GCC 3. | |
3603 | ||
3604 | VxWorks comes with an older version of GCC installed in | |
3605 | @file{@var{$WIND_BASE}/host}; we recommend you do not overwrite it. | |
3606 | Choose an installation @var{prefix} entirely outside @var{$WIND_BASE}. | |
3607 | Before running @command{configure}, create the directories @file{@var{prefix}} | |
3608 | and @file{@var{prefix}/bin}. Link or copy the appropriate assembler, | |
3609 | linker, etc. into @file{@var{prefix}/bin}, and set your @var{PATH} to | |
3610 | include that directory while running both @command{configure} and | |
3611 | @command{make}. | |
3612 | ||
3613 | You must give @command{configure} the | |
3614 | @option{--with-headers=@var{$WIND_BASE}/target/h} switch so that it can | |
3615 | find the VxWorks system headers. Since VxWorks is a cross compilation | |
3616 | target only, you must also specify @option{--target=@var{target}}. | |
3617 | @command{configure} will attempt to create the directory | |
3618 | @file{@var{prefix}/@var{target}/sys-include} and copy files into it; | |
3619 | make sure the user running @command{configure} has sufficient privilege | |
3620 | to do so. | |
3621 | ||
3622 | GCC's exception handling runtime requires a special ``configlette'' | |
3623 | module, @file{contrib/gthr_supp_vxw_5x.c}. Follow the instructions in | |
3624 | that file to add the module to your kernel build. (Future versions of | |
3625 | VxWorks will incorporate this module.) | |
3626 | ||
3627 | @html | |
3628 | <hr /> | |
3629 | @end html | |
3630 | @heading @anchor{x86_64-*-*}x86_64-*-*, amd64-*-* | |
3631 | ||
3632 | GCC supports the x86-64 architecture implemented by the AMD64 processor | |
3633 | (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD. | |
3634 | On GNU/Linux the default is a bi-arch compiler which is able to generate | |
3635 | both 64-bit x86-64 and 32-bit x86 code (via the @option{-m32} switch). | |
3636 | ||
3637 | @html | |
3638 | <hr /> | |
3639 | @end html | |
3640 | @heading @anchor{xtensa-*-elf}xtensa-*-elf | |
3641 | ||
3642 | This target is intended for embedded Xtensa systems using the | |
3643 | @samp{newlib} C library. It uses ELF but does not support shared | |
3644 | objects. Designed-defined instructions specified via the | |
3645 | Tensilica Instruction Extension (TIE) language are only supported | |
3646 | through inline assembly. | |
3647 | ||
3648 | The Xtensa configuration information must be specified prior to | |
3649 | building GCC@. The @file{include/xtensa-config.h} header | |
3650 | file contains the configuration information. If you created your | |
3651 | own Xtensa configuration with the Xtensa Processor Generator, the | |
3652 | downloaded files include a customized copy of this header file, | |
3653 | which you can use to replace the default header file. | |
3654 | ||
3655 | @html | |
3656 | <hr /> | |
3657 | @end html | |
3658 | @heading @anchor{xtensa-*-linux*}xtensa-*-linux* | |
3659 | ||
3660 | This target is for Xtensa systems running GNU/Linux. It supports ELF | |
3661 | shared objects and the GNU C library (glibc). It also generates | |
3662 | position-independent code (PIC) regardless of whether the | |
3663 | @option{-fpic} or @option{-fPIC} options are used. In other | |
3664 | respects, this target is the same as the | |
3665 | @uref{#xtensa-*-elf,,@samp{xtensa-*-elf}} target. | |
3666 | ||
3667 | @html | |
3668 | <hr /> | |
3669 | @end html | |
3670 | @heading @anchor{windows}Microsoft Windows (32-bit) | |
3671 | ||
3672 | A port of GCC 2.95.2 and 3.x is included with the | |
3673 | @uref{http://www.cygwin.com/,,Cygwin environment}. | |
3674 | ||
3675 | Current (as of early 2001) snapshots of GCC will build under Cygwin | |
3676 | without modification. | |
3677 | ||
3678 | GCC does not currently build with Microsoft's C++ compiler and there | |
3679 | are no plans to make it do so. | |
3680 | ||
3681 | @html | |
3682 | <hr /> | |
3683 | @end html | |
3684 | @heading @anchor{os2}OS/2 | |
3685 | ||
3686 | GCC does not currently support OS/2. However, Andrew Zabolotny has been | |
3687 | working on a generic OS/2 port with pgcc. The current code can be found | |
3688 | at @uref{http://www.goof.com/pcg/os2/,,http://www.goof.com/pcg/os2/}. | |
3689 | ||
3690 | An older copy of GCC 2.8.1 is included with the EMX tools available at | |
3691 | @uref{ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/,, | |
3692 | ftp://ftp.leo.org/pub/comp/os/os2/leo/devtools/emx+gcc/}. | |
3693 | ||
3694 | @html | |
3695 | <hr /> | |
3696 | @end html | |
3697 | @heading @anchor{older}Older systems | |
3698 | ||
3699 | GCC contains support files for many older (1980s and early | |
3700 | 1990s) Unix variants. For the most part, support for these systems | |
3701 | has not been deliberately removed, but it has not been maintained for | |
3702 | several years and may suffer from bitrot. | |
3703 | ||
3704 | Starting with GCC 3.1, each release has a list of ``obsoleted'' systems. | |
3705 | Support for these systems is still present in that release, but | |
3706 | @command{configure} will fail unless the @option{--enable-obsolete} | |
3707 | option is given. Unless a maintainer steps forward, support for these | |
3708 | systems will be removed from the next release of GCC@. | |
3709 | ||
3710 | Support for old systems as hosts for GCC can cause problems if the | |
3711 | workarounds for compiler, library and operating system bugs affect the | |
3712 | cleanliness or maintainability of the rest of GCC@. In some cases, to | |
3713 | bring GCC up on such a system, if still possible with current GCC, may | |
3714 | require first installing an old version of GCC which did work on that | |
3715 | system, and using it to compile a more recent GCC, to avoid bugs in the | |
3716 | vendor compiler. Old releases of GCC 1 and GCC 2 are available in the | |
3717 | @file{old-releases} directory on the @uref{../mirrors.html,,GCC mirror | |
3718 | sites}. Header bugs may generally be avoided using | |
3719 | @command{fixincludes}, but bugs or deficiencies in libraries and the | |
3720 | operating system may still cause problems. | |
3721 | ||
3722 | Support for older systems as targets for cross-compilation is less | |
3723 | problematic than support for them as hosts for GCC; if an enthusiast | |
3724 | wishes to make such a target work again (including resurrecting any of | |
3725 | the targets that never worked with GCC 2, starting from the last CVS | |
3726 | version before they were removed), patches | |
3727 | @uref{../contribute.html,,following the usual requirements} would be | |
3728 | likely to be accepted, since they should not affect the support for more | |
3729 | modern targets. | |
3730 | ||
3731 | For some systems, old versions of GNU binutils may also be useful, | |
3732 | and are available from @file{pub/binutils/old-releases} on | |
3733 | @uref{http://sources.redhat.com/mirrors.html,,sources.redhat.com mirror sites}. | |
3734 | ||
3735 | Some of the information on specific systems above relates to | |
3736 | such older systems, but much of the information | |
3737 | about GCC on such systems (which may no longer be applicable to | |
3738 | current GCC) is to be found in the GCC texinfo manual. | |
3739 | ||
3740 | @html | |
3741 | <hr /> | |
3742 | @end html | |
3743 | @heading @anchor{elf_targets}all ELF targets (SVR4, Solaris 2, etc.) | |
3744 | ||
3745 | C++ support is significantly better on ELF targets if you use the | |
3746 | @uref{./configure.html#with-gnu-ld,,GNU linker}; duplicate copies of | |
3747 | inlines, vtables and template instantiations will be discarded | |
3748 | automatically. | |
3749 | ||
3750 | ||
3751 | @html | |
3752 | <hr /> | |
3753 | <p> | |
3754 | @end html | |
3755 | @ifhtml | |
3756 | @uref{./index.html,,Return to the GCC Installation page} | |
3757 | @end ifhtml | |
3758 | @end ifset | |
3759 | ||
3760 | @c ***Old documentation****************************************************** | |
3761 | @ifset oldhtml | |
3762 | @include install-old.texi | |
3763 | @html | |
3764 | <hr /> | |
3765 | <p> | |
3766 | @end html | |
3767 | @ifhtml | |
3768 | @uref{./index.html,,Return to the GCC Installation page} | |
3769 | @end ifhtml | |
3770 | @end ifset | |
3771 | ||
3772 | @c ***GFDL******************************************************************** | |
3773 | @ifset gfdlhtml | |
3774 | @include fdl.texi | |
3775 | @html | |
3776 | <hr /> | |
3777 | <p> | |
3778 | @end html | |
3779 | @ifhtml | |
3780 | @uref{./index.html,,Return to the GCC Installation page} | |
3781 | @end ifhtml | |
3782 | @end ifset | |
3783 | ||
3784 | @c *************************************************************************** | |
3785 | @c Part 6 The End of the Document | |
3786 | @ifinfo | |
3787 | @comment node-name, next, previous, up | |
3788 | @node Concept Index, , GNU Free Documentation License, Top | |
3789 | @end ifinfo | |
3790 | ||
3791 | @ifinfo | |
3792 | @unnumbered Concept Index | |
3793 | ||
3794 | @printindex cp | |
3795 | ||
3796 | @contents | |
3797 | @end ifinfo | |
3798 | @bye |