1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
2 @c This is part of the GCC manual.
3 @c For copying conditions, see the file gcc.texi.
5 @c The text of this file appears in the file INSTALL
6 @c in the GCC distribution, as well as in the GCC manual.
8 Note most of this information is out of date and superseded by the EGCS
9 install procedures. It is provided for historical reference only.
13 @chapter Installing GNU CC
15 @cindex installing GNU CC
18 * Configuration Files:: Files created by running @code{configure}.
19 * Configurations:: Configurations Supported by GNU CC.
20 * Other Dir:: Compiling in a separate directory (not where the source is).
21 * Cross-Compiler:: Building and installing a cross-compiler.
22 * Sun Install:: See below for installation on the Sun.
23 * VMS Install:: See below for installation on VMS.
24 * Collect2:: How @code{collect2} works; how it finds @code{ld}.
25 * Header Dirs:: Understanding the standard header file directories.
28 Here is the procedure for installing GNU CC on a GNU or Unix system.
29 See @ref{VMS Install}, for VMS systems. In this section we assume you
30 compile in the same directory that contains the source files; see
31 @ref{Other Dir}, to find out how to compile in a separate directory on
34 You cannot install GNU C by itself on MSDOS; it will not compile under
35 any MSDOS compiler except itself. You need to get the complete
36 compilation package DJGPP, which includes binaries as well as sources,
37 and includes all the necessary compilation tools and libraries.
41 If you have built GNU CC previously in the same directory for a
42 different target machine, do @samp{make distclean} to delete all files
43 that might be invalid. One of the files this deletes is
44 @file{Makefile}; if @samp{make distclean} complains that @file{Makefile}
45 does not exist, it probably means that the directory is already suitably
49 On a System V release 4 system, make sure @file{/usr/bin} precedes
50 @file{/usr/ucb} in @code{PATH}. The @code{cc} command in
51 @file{/usr/ucb} uses libraries which have bugs.
53 @cindex Bison parser generator
54 @cindex parser generator, Bison
56 Make sure the Bison parser generator is installed. (This is unnecessary
57 if the Bison output file @file{c-parse.c} is more recent than
58 @file{c-parse.y},and you do not plan to change the @samp{.y} file.)
60 Bison versions older than Sept 8, 1988 will produce incorrect output
64 If you have chosen a configuration for GNU CC which requires other GNU
65 tools (such as GAS or the GNU linker) instead of the standard system
66 tools, install the required tools in the build directory under the names
67 @file{as}, @file{ld} or whatever is appropriate. This will enable the
68 compiler to find the proper tools for compilation of the program
71 Alternatively, you can do subsequent compilation using a value of the
72 @code{PATH} environment variable such that the necessary GNU tools come
73 before the standard system tools.
76 Specify the host, build and target machine configurations. You do this
77 when you run the @file{configure} script.
79 The @dfn{build} machine is the system which you are using, the
80 @dfn{host} machine is the system where you want to run the resulting
81 compiler (normally the build machine), and the @dfn{target} machine is
82 the system for which you want the compiler to generate code.
84 If you are building a compiler to produce code for the machine it runs
85 on (a native compiler), you normally do not need to specify any operands
86 to @file{configure}; it will try to guess the type of machine you are on
87 and use that as the build, host and target machines. So you don't need
88 to specify a configuration when building a native compiler unless
89 @file{configure} cannot figure out what your configuration is or guesses
92 In those cases, specify the build machine's @dfn{configuration name}
93 with the @samp{--host} option; the host and target will default to be
94 the same as the host machine. (If you are building a cross-compiler,
95 see @ref{Cross-Compiler}.)
100 ./configure --host=sparc-sun-sunos4.1
103 A configuration name may be canonical or it may be more or less
106 A canonical configuration name has three parts, separated by dashes.
107 It looks like this: @samp{@var{cpu}-@var{company}-@var{system}}.
108 (The three parts may themselves contain dashes; @file{configure}
109 can figure out which dashes serve which purpose.) For example,
110 @samp{m68k-sun-sunos4.1} specifies a Sun 3.
112 You can also replace parts of the configuration by nicknames or aliases.
113 For example, @samp{sun3} stands for @samp{m68k-sun}, so
114 @samp{sun3-sunos4.1} is another way to specify a Sun 3. You can also
115 use simply @samp{sun3-sunos}, since the version of SunOS is assumed by
116 default to be version 4.
118 You can specify a version number after any of the system types, and some
119 of the CPU types. In most cases, the version is irrelevant, and will be
120 ignored. So you might as well specify the version if you know it.
122 See @ref{Configurations}, for a list of supported configuration names and
123 notes on many of the configurations. You should check the notes in that
124 section before proceeding any further with the installation of GNU CC.
127 When running @code{configure}, you may also need to specify certain
128 additional options that describe variant hardware and software
129 configurations. These are @samp{--with-gnu-as}, @samp{--with-gnu-ld},
130 @samp{--with-stabs} and @samp{--nfp}.
134 If you will use GNU CC with the GNU assembler (GAS), you should declare
135 this by using the @samp{--with-gnu-as} option when you run
138 Using this option does not install GAS. It only modifies the output of
139 GNU CC to work with GAS. Building and installing GAS is up to you.
141 Conversely, if you @emph{do not} wish to use GAS and do not specify
142 @samp{--with-gnu-as} when building GNU CC, it is up to you to make sure
143 that GAS is not installed. GNU CC searches for a program named
144 @code{as} in various directories; if the program it finds is GAS, then
145 it runs GAS. If you are not sure where GNU CC finds the assembler it is
146 using, try specifying @samp{-v} when you run it.
148 The systems where it makes a difference whether you use GAS are@*
149 @samp{hppa1.0-@var{any}-@var{any}}, @samp{hppa1.1-@var{any}-@var{any}},
150 @samp{i386-@var{any}-sysv}, @samp{i386-@var{any}-isc},@*
151 @samp{i860-@var{any}-bsd}, @samp{m68k-bull-sysv},@*
152 @samp{m68k-hp-hpux}, @samp{m68k-sony-bsd},@*
153 @samp{m68k-altos-sysv}, @samp{m68000-hp-hpux},@*
154 @samp{m68000-att-sysv}, @samp{@var{any}-lynx-lynxos},
155 and @samp{mips-@var{any}}).
156 On any other system, @samp{--with-gnu-as} has no effect.
158 On the systems listed above (except for the HP-PA, for ISC on the
159 386, and for @samp{mips-sgi-irix5.*}), if you use GAS, you should also
160 use the GNU linker (and specify @samp{--with-gnu-ld}).
163 Specify the option @samp{--with-gnu-ld} if you plan to use the GNU
166 This option does not cause the GNU linker to be installed; it just
167 modifies the behavior of GNU CC to work with the GNU linker.
168 @c Specifically, it inhibits the installation of @code{collect2}, a program
169 @c which otherwise serves as a front-end for the system's linker on most
173 On MIPS based systems and on Alphas, you must specify whether you want
174 GNU CC to create the normal ECOFF debugging format, or to use BSD-style
175 stabs passed through the ECOFF symbol table. The normal ECOFF debug
176 format cannot fully handle languages other than C. BSD stabs format can
177 handle other languages, but it only works with the GNU debugger GDB.
179 Normally, GNU CC uses the ECOFF debugging format by default; if you
180 prefer BSD stabs, specify @samp{--with-stabs} when you configure GNU
183 No matter which default you choose when you configure GNU CC, the user
184 can use the @samp{-gcoff} and @samp{-gstabs+} options to specify explicitly
185 the debug format for a particular compilation.
187 @samp{--with-stabs} is meaningful on the ISC system on the 386, also, if
188 @samp{--with-gas} is used. It selects use of stabs debugging
189 information embedded in COFF output. This kind of debugging information
190 supports C++ well; ordinary COFF debugging information does not.
192 @samp{--with-stabs} is also meaningful on 386 systems running SVR4. It
193 selects use of stabs debugging information embedded in ELF output. The
194 C++ compiler currently (2.6.0) does not support the DWARF debugging
195 information normally used on 386 SVR4 platforms; stabs provide a
196 workable alternative. This requires gas and gdb, as the normal SVR4
197 tools can not generate or interpret stabs.
200 On certain systems, you must specify whether the machine has a floating
201 point unit. These systems include @samp{m68k-sun-sunos@var{n}} and
202 @samp{m68k-isi-bsd}. On any other system, @samp{--nfp} currently has no
203 effect, though perhaps there are other systems where it could usefully
206 @cindex Haifa scheduler
207 @cindex scheduler, experimental
209 @itemx --disable-haifa
210 Use @samp{--enable-haifa} to enable use of an experimental instruction
211 scheduler (from IBM Haifa). This may or may not produce better code.
212 Some targets on which it is known to be a win enable it by default; use
213 @samp{--disable-haifa} to disable it in these cases. @code{configure}
214 will print out whether the Haifa scheduler is enabled when it is run.
216 @cindex Objective C threads
217 @cindex threads, Objective C
218 @item --enable-threads=@var{type}
219 Certain systems, notably Linux-based GNU systems, can't be relied on to
220 supply a threads facility for the Objective C runtime and so will
221 default to single-threaded runtime. They may, however, have a library
222 threads implementation available, in which case threads can be enabled
223 with this option by supplying a suitable @var{type}, probably
224 @samp{posix}. The possibilities for @var{type} are @samp{single},
225 @samp{posix}, @samp{win32}, @samp{solaris}, @samp{irix} and @samp{mach}.
227 @cindex Internal Compiler Checking
228 @item --enable-checking
229 When you specify this option, the compiler is built to perform checking
230 of tree node types when referencing fields of that node. This does not
231 change the generated code, but adds error checking within the compiler.
232 This will slow down the compiler and may only work properly if you
233 are building the compiler with GNU C.
235 The @file{configure} script searches subdirectories of the source
236 directory for other compilers that are to be integrated into GNU CC.
237 The GNU compiler for C++, called G++ is in a subdirectory named
238 @file{cp}. @file{configure} inserts rules into @file{Makefile} to build
239 all of those compilers.
241 Here we spell out what files will be set up by @code{configure}. Normally
242 you need not be concerned with these files.
247 A file named @file{config.h} is created that contains a @samp{#include}
248 of the top-level config file for the machine you will run the compiler
249 on (@pxref{Config}). This file is responsible for defining information
250 about the host machine. It includes @file{tm.h}.
253 A file named @file{config.h} is created that contains a @samp{#include}
254 of the top-level config file for the machine you will run the compiler
255 on (@pxref{Config,,The Configuration File, gcc.info, Using and Porting
256 GCC}). This file is responsible for defining information about the host
257 machine. It includes @file{tm.h}.
260 The top-level config file is located in the subdirectory @file{config}.
261 Its name is always @file{xm-@var{something}.h}; usually
262 @file{xm-@var{machine}.h}, but there are some exceptions.
264 If your system does not support symbolic links, you might want to
265 set up @file{config.h} to contain a @samp{#include} command which
266 refers to the appropriate file.
269 A file named @file{tconfig.h} is created which includes the top-level config
270 file for your target machine. This is used for compiling certain
271 programs to run on that machine.
274 A file named @file{tm.h} is created which includes the
275 machine-description macro file for your target machine. It should be in
276 the subdirectory @file{config} and its name is often
277 @file{@var{machine}.h}.
280 @cindex Native Language Support
284 The @samp{--enable-nls} option enables Native Language Support (NLS),
285 which lets GCC output diagnostics in languages other than American
286 English. Native Language Support is enabled by default if not doing a
287 canadian cross build. The @samp{--disable-nls} option disables NLS.
289 @cindex @code{gettext}
290 @item --with-included-gettext
291 If NLS is enbled, the @samp{--with-included-gettext} option causes the build
292 procedure to prefer its copy of GNU @code{gettext}. This is the default. If
293 you want the GCC build procedure to prefer the host's @code{gettext}
294 libraries, use @samp{--without-included-gettext}.
296 @cindex @code{catgets}
298 If NLS is enabled, and if the host lacks @code{gettext} but has the
299 inferior @code{catgets} interface, the GCC build procedure normally
300 ignores @code{catgets} and instead uses GCC's copy of the GNU
301 @code{gettext} library. The @samp{--with-catgets} option causes the
302 build procedure to use the host's @code{catgets} in this situation.
304 @cindex @code{maintainer-mode}
305 @item --enable-maintainer-mode
306 The build rules that regenerate the GCC master message catalog
307 @code{gcc.pot} are normally disabled. This is because it can only be rebuilt
308 if the complete source tree is present. If you have changed the sources and
309 want to rebuild the catalog, configuring with
310 @samp{--enable-maintainer-mode} will enable this. Note that you need a
311 special version of the @code{gettext} tools to do so.
313 @cindex Windows32 Registry support
314 @item --enable-win32-registry
315 @itemx --enable-win32-registry=@var{KEY}
316 @itemx --disable-win32-registry
317 The @samp{--enable-win32-registry} option enables Windows-hosted GCC
318 to look up installations paths in the registry using the following key:
321 @code{HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\<KEY>}
324 <KEY> defaults to GCC version number, and can be overridden by the
325 @code{--enable-win32-registry=KEY} option. Vendors and distributors
326 who use custom installers are encouraged to provide a different key,
327 perhaps one comprised of vendor name and GCC version number, to
328 avoid conflict with existing installations. This feature is enabled
329 by default, and can be disabled by @code{--disable-win32-registry}
330 option. This option has no effect on the other hosts.
334 In certain cases, you should specify certain other options when you run
339 The standard directory for installing GNU CC is @file{/usr/local/lib}.
340 If you want to install its files somewhere else, specify
341 @samp{--prefix=@var{dir}} when you run @file{configure}. Here @var{dir}
342 is a directory name to use instead of @file{/usr/local} for all purposes
343 with one exception: the directory @file{/usr/local/include} is searched
344 for header files no matter where you install the compiler. To override
345 this name, use the @code{--with-local-prefix} option below. The directory
346 you specify need not exist, but its parent directory must exist.
349 Specify @samp{--with-local-prefix=@var{dir}} if you want the compiler to
350 search directory @file{@var{dir}/include} for locally installed header
351 files @emph{instead} of @file{/usr/local/include}.
353 You should specify @samp{--with-local-prefix} @strong{only} if your site has
354 a different convention (not @file{/usr/local}) for where to put
357 The default value for @samp{--with-local-prefix} is @file{/usr/local}
358 regardless of the value of @samp{--prefix}. Specifying @samp{--prefix}
359 has no effect on which directory GNU CC searches for local header files.
360 This may seem counterintuitive, but actually it is logical.
362 The purpose of @samp{--prefix} is to specify where to @emph{install GNU
363 CC}. The local header files in @file{/usr/local/include}---if you put
364 any in that directory---are not part of GNU CC. They are part of other
365 programs---perhaps many others. (GNU CC installs its own header files
366 in another directory which is based on the @samp{--prefix} value.)
368 @strong{Do not} specify @file{/usr} as the @samp{--with-local-prefix}! The
369 directory you use for @samp{--with-local-prefix} @strong{must not} contain
370 any of the system's standard header files. If it did contain them,
371 certain programs would be miscompiled (including GNU Emacs, on certain
372 targets), because this would override and nullify the header file
373 corrections made by the @code{fixincludes} script.
375 Indications are that people who use this option use it based on
376 mistaken ideas of what it is for. People use it as if it specified
377 where to install part of GNU CC. Perhaps they make this assumption
378 because installing GNU CC creates the directory.
382 Build the compiler. Just type @samp{make LANGUAGES=c} in the compiler
385 @samp{LANGUAGES=c} specifies that only the C compiler should be
386 compiled. The makefile normally builds compilers for all the supported
387 languages; currently, C, C++, Objective C, Java, FORTRAN, and CHILL.
388 However, C is the only language that is sure to work when you build with
389 other non-GNU C compilers. In addition, building anything but C at this
390 stage is a waste of time.
392 In general, you can specify the languages to build by typing the
393 argument @samp{LANGUAGES="@var{list}"}, where @var{list} is one or more
394 words from the list @samp{c}, @samp{c++}, @samp{objective-c},
395 @samp{java}, @samp{f77}, and @samp{CHILL}. If you have any additional
396 GNU compilers as subdirectories of the GNU CC source directory, you may
397 also specify their names in this list.
399 Ignore any warnings you may see about ``statement not reached'' in
400 @file{insn-emit.c}; they are normal. Also, warnings about ``unknown
401 escape sequence'' are normal in @file{genopinit.c} and perhaps some
402 other files. Likewise, you should ignore warnings about ``constant is
403 so large that it is unsigned'' in @file{insn-emit.c} and
404 @file{insn-recog.c}, and a warning about a comparison always being zero
405 in @file{enquire.o}. Any other compilation errors may represent bugs in
406 the port to your machine or operating system, and
408 should be investigated and reported (@pxref{Bugs}).
411 should be investigated and reported.
414 Some compilers fail to compile GNU CC because they have bugs or
415 limitations. For example, the Microsoft compiler is said to run out of
416 macro space. Some Ultrix compilers run out of expression space; then
417 you need to break up the statement where the problem happens.
420 If you are building a cross-compiler, stop here. @xref{Cross-Compiler}.
424 Move the first-stage object files and executables into a subdirectory
431 The files are moved into a subdirectory named @file{stage1}.
432 Once installation is complete, you may wish to delete these files
433 with @code{rm -r stage1}.
436 If you have chosen a configuration for GNU CC which requires other GNU
437 tools (such as GAS or the GNU linker) instead of the standard system
438 tools, install the required tools in the @file{stage1} subdirectory
439 under the names @file{as}, @file{ld} or whatever is appropriate. This
440 will enable the stage 1 compiler to find the proper tools in the
443 Alternatively, you can do subsequent compilation using a value of the
444 @code{PATH} environment variable such that the necessary GNU tools come
445 before the standard system tools.
448 Recompile the compiler with itself, with this command:
451 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2"
454 This is called making the stage 2 compiler.
456 The command shown above builds compilers for all the supported
457 languages. If you don't want them all, you can specify the languages to
458 build by typing the argument @samp{LANGUAGES="@var{list}"}. @var{list}
459 should contain one or more words from the list @samp{c}, @samp{c++},
460 @samp{objective-c}, and @samp{proto}. Separate the words with spaces.
461 @samp{proto} stands for the programs @code{protoize} and
462 @code{unprotoize}; they are not a separate language, but you use
463 @code{LANGUAGES} to enable or disable their installation.
465 If you are going to build the stage 3 compiler, then you might want to
466 build only the C language in stage 2.
468 Once you have built the stage 2 compiler, if you are short of disk
469 space, you can delete the subdirectory @file{stage1}.
471 On a 68000 or 68020 system lacking floating point hardware,
472 unless you have selected a @file{tm.h} file that expects by default
473 that there is no such hardware, do this instead:
476 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2 -msoft-float"
480 If you wish to test the compiler by compiling it with itself one more
481 time, install any other necessary GNU tools (such as GAS or the GNU
482 linker) in the @file{stage2} subdirectory as you did in the
483 @file{stage1} subdirectory, then do this:
487 make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2"
491 This is called making the stage 3 compiler. Aside from the @samp{-B}
492 option, the compiler options should be the same as when you made the
493 stage 2 compiler. But the @code{LANGUAGES} option need not be the
494 same. The command shown above builds compilers for all the supported
495 languages; if you don't want them all, you can specify the languages to
496 build by typing the argument @samp{LANGUAGES="@var{list}"}, as described
499 If you do not have to install any additional GNU tools, you may use the
503 make bootstrap LANGUAGES=@var{language-list} BOOT_CFLAGS=@var{option-list}
507 instead of making @file{stage1}, @file{stage2}, and performing
508 the two compiler builds.
511 Compare the latest object files with the stage 2 object files---they
512 ought to be identical, aside from time stamps (if any).
514 On some systems, meaningful comparison of object files is impossible;
515 they always appear ``different.'' This is currently true on Solaris and
516 some systems that use ELF object file format. On some versions of Irix
517 on SGI machines and DEC Unix (OSF/1) on Alpha systems, you will not be
518 able to compare the files without specifying @file{-save-temps}; see the
519 description of individual systems above to see if you get comparison
520 failures. You may have similar problems on other systems.
522 Use this command to compare the files:
528 This will mention any object files that differ between stage 2 and stage
529 3. Any difference, no matter how innocuous, indicates that the stage 2
530 compiler has compiled GNU CC incorrectly, and is therefore a potentially
532 serious bug which you should investigate and report (@pxref{Bugs}).
535 serious bug which you should investigate and report.
538 If your system does not put time stamps in the object files, then this
539 is a faster way to compare them (using the Bourne shell):
543 cmp $file stage2/$file
547 If you have built the compiler with the @samp{-mno-mips-tfile} option on
548 MIPS machines, you will not be able to compare the files.
551 Install the compiler driver, the compiler's passes and run-time support
552 with @samp{make install}. Use the same value for @code{CC},
553 @code{CFLAGS} and @code{LANGUAGES} that you used when compiling the
554 files that are being installed. One reason this is necessary is that
555 some versions of Make have bugs and recompile files gratuitously when
556 you do this step. If you use the same variable values, those files will
557 be recompiled properly.
559 For example, if you have built the stage 2 compiler, you can use the
563 make install CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O" LANGUAGES="@var{list}"
567 This copies the files @file{cc1}, @file{cpp} and @file{libgcc.a} to
568 files @file{cc1}, @file{cpp} and @file{libgcc.a} in the directory
569 @file{/usr/local/lib/gcc-lib/@var{target}/@var{version}}, which is where
570 the compiler driver program looks for them. Here @var{target} is the
571 canonicalized form of target machine type specified when you ran
572 @file{configure}, and @var{version} is the version number of GNU CC.
573 This naming scheme permits various versions and/or cross-compilers to
574 coexist. It also copies the executables for compilers for other
575 languages (e.g., @file{cc1plus} for C++) to the same directory.
577 This also copies the driver program @file{xgcc} into
578 @file{/usr/local/bin/gcc}, so that it appears in typical execution
579 search paths. It also copies @file{gcc.1} into
580 @file{/usr/local/man/man1} and info pages into @file{/usr/local/info}.
582 On some systems, this command causes recompilation of some files. This
583 is usually due to bugs in @code{make}. You should either ignore this
584 problem, or use GNU Make.
586 @cindex @code{alloca} and SunOS
587 @strong{Warning: there is a bug in @code{alloca} in the Sun library. To
588 avoid this bug, be sure to install the executables of GNU CC that were
589 compiled by GNU CC. (That is, the executables from stage 2 or 3, not
590 stage 1.) They use @code{alloca} as a built-in function and never the
593 (It is usually better to install GNU CC executables from stage 2 or 3,
594 since they usually run faster than the ones compiled with some other
598 @cindex C++ runtime library
599 @cindex @code{libstdc++}
600 If you're going to use C++, you need to install the C++ runtime library.
601 This includes all I/O functionality, special class libraries, etc.
603 The standard C++ runtime library for GNU CC is called @samp{libstdc++}.
604 An obsolescent library @samp{libg++} may also be available, but it's
605 necessary only for older software that hasn't been converted yet; if
606 you don't know whether you need @samp{libg++} then you probably don't
609 Here's one way to build and install @samp{libstdc++} for GNU CC:
613 Build and install GNU CC, so that invoking @samp{gcc} obtains the GNU CC
617 Obtain a copy of a compatible @samp{libstdc++} distribution. For
618 example, the @samp{libstdc++-2.8.0.tar.gz} distribution should be
619 compatible with GCC 2.8.0. GCC distributors normally distribute
620 @samp{libstdc++} as well.
623 Set the @samp{CXX} environment variable to @samp{gcc} while running the
624 @samp{libstdc++} distribution's @file{configure} command. Use the same
625 @file{configure} options that you used when you invoked GCC's
626 @file{configure} command.
629 Invoke @samp{make} to build the C++ runtime.
632 Invoke @samp{make install} to install the C++ runtime.
636 To summarize, after building and installing GNU CC, invoke the following
637 shell commands in the topmost directory of the C++ library distribution.
638 For @var{configure-options}, use the same options that
639 you used to configure GNU CC.
642 $ CXX=gcc ./configure @var{configure-options}
648 GNU CC includes a runtime library for Objective-C because it is an
649 integral part of the language. You can find the files associated with
650 the library in the subdirectory @file{objc}. The GNU Objective-C
651 Runtime Library requires header files for the target's C library in
652 order to be compiled,and also requires the header files for the target's
653 thread library if you want thread support. @xref{Cross Headers,
654 Cross-Compilers and Header Files, Cross-Compilers and Header Files}, for
655 discussion about header files issues for cross-compilation.
657 When you run @file{configure}, it picks the appropriate Objective-C
658 thread implementation file for the target platform. In some situations,
659 you may wish to choose a different back-end as some platforms support
660 multiple thread implementations or you may wish to disable thread
661 support completely. You do this by specifying a value for the
662 @var{OBJC_THREAD_FILE} makefile variable on the command line when you
663 run make, for example:
666 make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2" OBJC_THREAD_FILE=thr-single
670 Below is a list of the currently available back-ends.
674 Disable thread support, should work for all platforms.
676 DEC OSF/1 thread support.
678 SGI IRIX thread support.
680 Generic MACH thread support, known to work on NEXTSTEP.
682 IBM OS/2 thread support.
684 Generix POSIX thread support.
686 PCThreads on Linux-based GNU systems.
688 SUN Solaris thread support.
690 Microsoft Win32 API thread support.
694 @node Configuration Files
695 @section Files Created by @code{configure}
697 Here we spell out what files will be set up by @code{configure}. Normally
698 you need not be concerned with these files.
703 A file named @file{config.h} is created that contains a @samp{#include}
704 of the top-level config file for the machine you will run the compiler
705 on (@pxref{Config}). This file is responsible for defining information
706 about the host machine. It includes @file{tm.h}.
709 A file named @file{config.h} is created that contains a @samp{#include}
710 of the top-level config file for the machine you will run the compiler
711 on (@pxref{Config,,The Configuration File, gcc.info, Using and Porting
712 GCC}). This file is responsible for defining information about the host
713 machine. It includes @file{tm.h}.
716 The top-level config file is located in the subdirectory @file{config}.
717 Its name is always @file{xm-@var{something}.h}; usually
718 @file{xm-@var{machine}.h}, but there are some exceptions.
720 If your system does not support symbolic links, you might want to
721 set up @file{config.h} to contain a @samp{#include} command which
722 refers to the appropriate file.
725 A file named @file{tconfig.h} is created which includes the top-level config
726 file for your target machine. This is used for compiling certain
727 programs to run on that machine.
730 A file named @file{tm.h} is created which includes the
731 machine-description macro file for your target machine. It should be in
732 the subdirectory @file{config} and its name is often
733 @file{@var{machine}.h}.
736 The command file @file{configure} also constructs the file
737 @file{Makefile} by adding some text to the template file
738 @file{Makefile.in}. The additional text comes from files in the
739 @file{config} directory, named @file{t-@var{target}} and
740 @file{x-@var{host}}. If these files do not exist, it means nothing
741 needs to be added for a given target or host.
745 @section Configurations Supported by GNU CC
746 @cindex configurations supported by GNU CC
748 Here are the possible CPU types:
751 @c gmicro, fx80, spur and tahoe omitted since they don't work.
752 1750a, a29k, alpha, arm, avr, c@var{n}, clipper, dsp16xx, elxsi, fr30, h8300,
753 hppa1.0, hppa1.1, i370, i386, i486, i586, i686, i786, i860, i960, m32r,
754 m68000, m68k, m88k, mcore, mips, mipsel, mips64, mips64el, mn10200, mn10300,
755 ns32k, pdp11, powerpc, powerpcle, romp, rs6000, sh, sparc, sparclite,
756 sparc64, v850, vax, we32k.
759 Here are the recognized company names. As you can see, customary
760 abbreviations are used rather than the longer official names.
762 @c What should be done about merlin, tek*, dolphin?
764 acorn, alliant, altos, apollo, apple, att, bull,
765 cbm, convergent, convex, crds, dec, dg, dolphin,
766 elxsi, encore, harris, hitachi, hp, ibm, intergraph, isi,
767 mips, motorola, ncr, next, ns, omron, plexus,
768 sequent, sgi, sony, sun, tti, unicom, wrs.
771 The company name is meaningful only to disambiguate when the rest of
772 the information supplied is insufficient. You can omit it, writing
773 just @samp{@var{cpu}-@var{system}}, if it is not needed. For example,
774 @samp{vax-ultrix4.2} is equivalent to @samp{vax-dec-ultrix4.2}.
776 Here is a list of system types:
779 386bsd, aix, acis, amigaos, aos, aout, aux, bosx, bsd, clix, coff, ctix, cxux,
780 dgux, dynix, ebmon, ecoff, elf, esix, freebsd, hms, genix, gnu, linux,
781 linux-gnu, hiux, hpux, iris, irix, isc, luna, lynxos, mach, minix, msdos, mvs,
782 netbsd, newsos, nindy, ns, osf, osfrose, ptx, riscix, riscos, rtu, sco, sim,
783 solaris, sunos, sym, sysv, udi, ultrix, unicos, uniplus, unos, vms, vsta,
784 vxworks, winnt, xenix.
788 You can omit the system type; then @file{configure} guesses the
789 operating system from the CPU and company.
791 You can add a version number to the system type; this may or may not
792 make a difference. For example, you can write @samp{bsd4.3} or
793 @samp{bsd4.4} to distinguish versions of BSD. In practice, the version
794 number is most needed for @samp{sysv3} and @samp{sysv4}, which are often
797 @samp{linux-gnu} is the canonical name for the GNU/Linux target; however
798 GNU CC will also accept @samp{linux}. The version of the kernel in use is
799 not relevant on these systems. A suffix such as @samp{libc1} or @samp{aout}
800 distinguishes major versions of the C library; all of the suffixed versions
803 If you specify an impossible combination such as @samp{i860-dg-vms},
804 then you may get an error message from @file{configure}, or it may
805 ignore part of the information and do the best it can with the rest.
806 @file{configure} always prints the canonical name for the alternative
807 that it used. GNU CC does not support all possible alternatives.
809 Often a particular model of machine has a name. Many machine names are
810 recognized as aliases for CPU/company combinations. Thus, the machine
811 name @samp{sun3}, mentioned above, is an alias for @samp{m68k-sun}.
812 Sometimes we accept a company name as a machine name, when the name is
813 popularly used for a particular machine. Here is a table of the known
817 3300, 3b1, 3b@var{n}, 7300, altos3068, altos,
818 apollo68, att-7300, balance,
819 convex-c@var{n}, crds, decstation-3100,
820 decstation, delta, encore,
821 fx2800, gmicro, hp7@var{nn}, hp8@var{nn},
822 hp9k2@var{nn}, hp9k3@var{nn}, hp9k7@var{nn},
823 hp9k8@var{nn}, iris4d, iris, isi68,
824 m3230, magnum, merlin, miniframe,
825 mmax, news-3600, news800, news, next,
826 pbd, pc532, pmax, powerpc, powerpcle, ps2, risc-news,
827 rtpc, sun2, sun386i, sun386, sun3,
828 sun4, symmetry, tower-32, tower.
832 Remember that a machine name specifies both the cpu type and the company
834 If you want to install your own homemade configuration files, you can
835 use @samp{local} as the company name to access them. If you use
836 configuration @samp{@var{cpu}-local}, the configuration name
837 without the cpu prefix
838 is used to form the configuration file names.
840 Thus, if you specify @samp{m68k-local}, configuration uses
841 files @file{m68k.md}, @file{local.h}, @file{m68k.c},
842 @file{xm-local.h}, @file{t-local}, and @file{x-local}, all in the
843 directory @file{config/m68k}.
845 Here is a list of configurations that have special treatment or special
846 things you must know:
850 MIL-STD-1750A processors.
852 The MIL-STD-1750A cross configuration produces output for
853 @code{as1750}, an assembler/linker available under the GNU Public
854 License for the 1750A. @code{as1750} can be obtained at
855 @emph{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}.
856 A similarly licensed simulator for
857 the 1750A is available from same address.
859 You should ignore a fatal error during the building of libgcc (libgcc is
860 not yet implemented for the 1750A.)
862 The @code{as1750} assembler requires the file @file{ms1750.inc}, which is
863 found in the directory @file{config/1750a}.
865 GNU CC produced the same sections as the Fairchild F9450 C Compiler,
870 The program code section.
873 The read/write (RAM) data section.
876 The read-only (ROM) constants section.
879 Initialization section (code to copy KREL to SREL).
882 The smallest addressable unit is 16 bits (BITS_PER_UNIT is 16). This
883 means that type `char' is represented with a 16-bit word per character.
884 The 1750A's "Load/Store Upper/Lower Byte" instructions are not used by
888 Systems using processors that implement the DEC Alpha architecture and
889 are running the DEC Unix (OSF/1) operating system, for example the DEC
890 Alpha AXP systems.CC.)
892 GNU CC writes a @samp{.verstamp} directive to the assembler output file
893 unless it is built as a cross-compiler. It gets the version to use from
894 the system header file @file{/usr/include/stamp.h}. If you install a
895 new version of DEC Unix, you should rebuild GCC to pick up the new version
898 Note that since the Alpha is a 64-bit architecture, cross-compilers from
899 32-bit machines will not generate code as efficient as that generated
900 when the compiler is running on a 64-bit machine because many
901 optimizations that depend on being able to represent a word on the
902 target in an integral value on the host cannot be performed. Building
903 cross-compilers on the Alpha for 32-bit machines has only been tested in
904 a few cases and may not work properly.
906 @code{make compare} may fail on old versions of DEC Unix unless you add
907 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
908 assembler input file is stored in the object file, and that makes
909 comparison fail if it differs between the @code{stage1} and
910 @code{stage2} compilations. The option @samp{-save-temps} forces a
911 fixed name to be used for the assembler input file, instead of a
912 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
913 unless the comparisons fail without that option. If you add
914 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
915 @samp{.s} files after each series of compilations.
917 GNU CC now supports both the native (ECOFF) debugging format used by DBX
918 and GDB and an encapsulated STABS format for use only with GDB. See the
919 discussion of the @samp{--with-stabs} option of @file{configure} above
920 for more information on these formats and how to select them.
922 There is a bug in DEC's assembler that produces incorrect line numbers
923 for ECOFF format when the @samp{.align} directive is used. To work
924 around this problem, GNU CC will not emit such alignment directives
925 while writing ECOFF format debugging information even if optimization is
926 being performed. Unfortunately, this has the very undesirable
927 side-effect that code addresses when @samp{-O} is specified are
928 different depending on whether or not @samp{-g} is also specified.
930 To avoid this behavior, specify @samp{-gstabs+} and use GDB instead of
931 DBX. DEC is now aware of this problem with the assembler and hopes to
932 provide a fix shortly.
935 Argonaut ARC processor.
936 This configuration is intended for embedded systems.
939 Advanced RISC Machines ARM-family processors. These are often used in
940 embedded applications. There are no standard Unix configurations.
941 This configuration corresponds to the basic instruction sequences and will
942 produce @file{a.out} format object modules.
944 You may need to make a variant of the file @file{arm.h} for your particular
948 This configuration is intended for embedded systems.
950 @item arm-*-linux*aout
951 Any of the ARM-family processors running the Linux-based GNU system with
952 the @file{a.out} binary format. This is an obsolete configuration.
955 @itemx arm-*-linux-gnu
956 @itemx arm-*-linux*oldld
957 Any of the ARM-family processors running the Linux-based GNU system with
958 the @file{ELF} binary format. You must use version 2.9.1.0.22 or later
959 of the GNU/Linux binutils, which you can download from
960 @file{ftp.varesearch.com:/pub/support/hjl/binutils}.
962 These two configurations differ only in the required version of GNU
963 binutils. For binutils 2.9.1.0.x, use @samp{arm-*-linux-gnuoldld}. For
964 newer versions of binutils, use @samp{arm-*-linux-gnu}.
967 The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix.
968 If you are running a version of RISC iX prior to 1.2 then you must
969 specify the version number during configuration. Note that the
970 assembler shipped with RISC iX does not support stabs debugging
971 information; a new version of the assembler, with stabs support
972 included, is now available from Acorn and via ftp
973 @file{ftp.acorn.com:/pub/riscix/as+xterm.tar.Z}. To enable stabs
974 debugging, pass @samp{--with-gnu-as} to configure.
976 You will need to install GNU @file{sed} before you can run configure.
979 AMD Am29k-family processors. These are normally used in embedded
980 applications. There are no standard Unix configurations.
982 corresponds to AMD's standard calling sequence and binary interface
983 and is compatible with other 29k tools.
985 You may need to make a variant of the file @file{a29k.h} for your
986 particular configuration.
989 AMD Am29050 used in a system running a variant of BSD Unix.
992 ATMEL AVR-family micro controllers. These are used in embedded
993 applications. There are no standard Unix configurations.
994 Supports following MCU's:
1003 MIPS-based DECstations can support three different personalities:
1004 Ultrix, DEC OSF/1, and OSF/rose. (Alpha-based DECstation products have
1005 a configuration name beginning with @samp{alpha-dec}.) To configure GCC
1006 for these platforms use the following configurations:
1009 @item decstation-ultrix
1010 Ultrix configuration.
1012 @item decstation-osf1
1013 Dec's version of OSF/1.
1015 @item decstation-osfrose
1016 Open Software Foundation reference port of OSF/1 which uses the
1017 OSF/rose object file format instead of ECOFF. Normally, you
1018 would not select this configuration.
1021 The MIPS C compiler needs to be told to increase its table size
1022 for switch statements with the @samp{-Wf,-XNg1500} option in
1023 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1024 optimization option, you also need to use @samp{-Olimit 3000}.
1025 Both of these options are automatically generated in the
1026 @file{Makefile} that the shell script @file{configure} builds.
1027 If you override the @code{CC} make variable and use the MIPS
1028 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1030 @item elxsi-elxsi-bsd
1031 The Elxsi's C compiler has known limitations that prevent it from
1032 compiling GNU C. Please contact @code{mrs@@cygnus.com} for more details.
1035 A port to the AT&T DSP1610 family of processors.
1039 Alliant FX/8 computer. Note that the standard installed C compiler in
1040 Concentrix 5.0 has a bug which prevent it from compiling GNU CC
1041 correctly. You can patch the compiler bug as follows:
1045 adb -w ./pcc - << EOF
1050 Then you must use the @samp{-ip12} option when compiling GNU CC
1051 with the patched compiler, as shown here:
1054 make CC="./pcc -ip12" CFLAGS=-w
1057 Note also that Alliant's version of DBX does not manage to work with the
1062 Hitachi H8/300 series of processors.
1064 The calling convention and structure layout has changed in release 2.6.
1065 All code must be recompiled. The calling convention now passes the
1066 first three arguments in function calls in registers. Structures are no
1067 longer a multiple of 2 bytes.
1070 There are several variants of the HP-PA processor which run a variety
1071 of operating systems. GNU CC must be configured to use the correct
1072 processor type and operating system, or GNU CC will not function correctly.
1073 The easiest way to handle this problem is to @emph{not} specify a target
1074 when configuring GNU CC, the @file{configure} script will try to automatically
1075 determine the right processor type and operating system.
1077 @samp{-g} does not work on HP-UX, since that system uses a peculiar
1078 debugging format which GNU CC does not know about. However, @samp{-g}
1079 will work if you also use GAS and GDB in conjunction with GCC. We
1080 highly recommend using GAS for all HP-PA configurations.
1082 You should be using GAS-2.6 (or later) along with GDB-4.16 (or later). These
1083 can be retrieved from all the traditional GNU ftp archive sites.
1085 On some versions of HP-UX, you will need to install GNU @file{sed}.
1087 You will need to be install GAS into a directory before @code{/bin},
1088 @code{/usr/bin}, and @code{/usr/ccs/bin} in your search path. You
1089 should install GAS before you build GNU CC.
1091 To enable debugging, you must configure GNU CC with the @samp{--with-gnu-as}
1092 option before building.
1095 This port is very preliminary and has many known bugs. We hope to
1096 have a higher-quality port for this machine soon.
1098 @item i386-*-linux*oldld
1099 Use this configuration to generate @file{a.out} binaries on Linux-based
1100 GNU systems if you do not have gas/binutils version 2.5.2 or later
1101 installed. This is an obsolete configuration.
1103 @item i386-*-linux*aout
1104 Use this configuration to generate @file{a.out} binaries on Linux-based
1105 GNU systems. This configuration is being superseded. You must use
1106 gas/binutils version 2.5.2 or later.
1109 @itemx i386-*-linux-gnu
1110 Use this configuration to generate ELF binaries on Linux-based GNU
1111 systems. You must use gas/binutils version 2.5.2 or later.
1114 Compilation with RCC is recommended. Also, it may be a good idea to
1115 link with GNU malloc instead of the malloc that comes with the system.
1117 @item i386-*-sco3.2v4
1118 Use this configuration for SCO release 3.2 version 4.
1120 @item i386-*-sco3.2v5*
1121 Use this for the SCO OpenServer Release family including 5.0.0, 5.0.2,
1122 5.0.4, 5.0.5, Internet FastStart 1.0, and Internet FastStart 1.1.
1124 GNU CC can generate COFF binaries if you specify @samp{-mcoff} or ELF
1125 binaries, the default. A full @samp{make bootstrap} is recommended
1126 so that an ELF compiler that builds ELF is generated.
1128 You must have TLS597 from @uref{ftp://ftp.sco.com/TLS} installed for ELF
1129 C++ binaries to work correctly on releases before 5.0.4.
1131 The native SCO assembler that is provided with the OS at no charge
1132 is normally required. If, however, you must be able to use the GNU
1133 assembler (perhaps you have complex asms) you must configure this
1134 package @samp{--with-gnu-as}. To do this, install (cp or symlink)
1135 gcc/as to your copy of the GNU assembler. You must use a recent version
1136 of GNU binutils; version 2.9.1 seems to work well. If you select this
1137 option, you will be unable to build COFF images. Trying to do so will
1138 result in non-obvious failures. In general, the "--with-gnu-as" option
1139 isn't as well tested as the native assembler.
1141 @emph{NOTE:} If you are building C++, you must follow the instructions
1142 about invoking @samp{make bootstrap} because the native OpenServer
1143 compiler may build a @file{cc1plus} that will not correctly parse many
1144 valid C++ programs. You must do a @samp{make bootstrap} if you are
1145 building with the native compiler.
1148 It may be a good idea to link with GNU malloc instead of the malloc that
1149 comes with the system.
1151 In ISC version 4.1, @file{sed} core dumps when building
1152 @file{deduced.h}. Use the version of @file{sed} from version 4.0.
1155 It may be good idea to link with GNU malloc instead of the malloc that
1156 comes with the system.
1159 You need to use GAS version 2.1 or later, and LD from
1160 GNU binutils version 2.2 or later.
1162 @item i386-sequent-bsd
1163 Go to the Berkeley universe before compiling.
1165 @item i386-sequent-ptx1*
1166 @itemx i386-sequent-ptx2*
1167 You must install GNU @file{sed} before running @file{configure}.
1169 @item i386-sun-sunos4
1170 You may find that you need another version of GNU CC to begin
1171 bootstrapping with, since the current version when built with the
1172 system's own compiler seems to get an infinite loop compiling part of
1173 @file{libgcc2.c}. GNU CC version 2 compiled with GNU CC (any version)
1174 seems not to have this problem.
1176 See @ref{Sun Install}, for information on installing GNU CC on Sun
1179 @item i[345]86-*-winnt3.5
1180 This version requires a GAS that has not yet been released. Until it
1181 is, you can get a prebuilt binary version via anonymous ftp from
1182 @file{cs.washington.edu:pub/gnat} or @file{cs.nyu.edu:pub/gnat}. You
1183 must also use the Microsoft header files from the Windows NT 3.5 SDK.
1184 Find these on the CDROM in the @file{/mstools/h} directory dated 9/4/94. You
1185 must use a fixed version of Microsoft linker made especially for NT 3.5,
1186 which is also is available on the NT 3.5 SDK CDROM. If you do not have
1187 this linker, can you also use the linker from Visual C/C++ 1.0 or 2.0.
1189 Installing GNU CC for NT builds a wrapper linker, called @file{ld.exe},
1190 which mimics the behaviour of Unix @file{ld} in the specification of
1191 libraries (@samp{-L} and @samp{-l}). @file{ld.exe} looks for both Unix
1192 and Microsoft named libraries. For example, if you specify
1193 @samp{-lfoo}, @file{ld.exe} will look first for @file{libfoo.a}
1194 and then for @file{foo.lib}.
1196 You may install GNU CC for Windows NT in one of two ways, depending on
1197 whether or not you have a Unix-like shell and various Unix-like
1202 If you do not have a Unix-like shell and few Unix-like utilities, you
1203 will use a DOS style batch script called @file{configure.bat}. Invoke
1204 it as @code{configure winnt} from an MSDOS console window or from the
1205 program manager dialog box. @file{configure.bat} assumes you have
1206 already installed and have in your path a Unix-like @file{sed} program
1207 which is used to create a working @file{Makefile} from @file{Makefile.in}.
1209 @file{Makefile} uses the Microsoft Nmake program maintenance utility and
1210 the Visual C/C++ V8.00 compiler to build GNU CC. You need only have the
1211 utilities @file{sed} and @file{touch} to use this installation method,
1212 which only automatically builds the compiler itself. You must then
1213 examine what @file{fixinc.winnt} does, edit the header files by hand and
1214 build @file{libgcc.a} manually.
1217 The second type of installation assumes you are running a Unix-like
1218 shell, have a complete suite of Unix-like utilities in your path, and
1219 have a previous version of GNU CC already installed, either through
1220 building it via the above installation method or acquiring a pre-built
1221 binary. In this case, use the @file{configure} script in the normal
1225 @item i860-intel-osf1
1226 This is the Paragon.
1228 If you have version 1.0 of the operating system, you need to take
1229 special steps to build GNU CC due to peculiarities of the system. Newer
1230 system versions have no problem. See the section `Installation Problems'
1231 in the GNU CC Manual.
1233 @ifclear INSTALLONLY
1234 If you have version 1.0 of the operating system,
1235 see @ref{Installation Problems}, for special things you need to do to
1236 compensate for peculiarities in the system.
1240 LynxOS 2.2 and earlier comes with GNU CC 1.x already installed as
1241 @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}.
1242 You can tell GNU CC to use the GNU assembler and linker, by specifying
1243 @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce
1244 COFF format object files and executables; otherwise GNU CC will use the
1245 installed tools, which produce @file{a.out} format executables.
1248 Mitsubishi M32R processor.
1249 This configuration is intended for embedded systems.
1252 HP 9000 series 200 running BSD. Note that the C compiler that comes
1253 with this system cannot compile GNU CC; contact @code{law@@cygnus.com}
1254 to get binaries of GNU CC for bootstrapping.
1257 Altos 3068. You must use the GNU assembler, linker and debugger.
1258 Also, you must fix a kernel bug. Details in the file @file{README.ALTOS}.
1260 @item m68k-apple-aux
1261 Apple Macintosh running A/UX.
1262 You may configure GCC to use either the system assembler and
1263 linker or the GNU assembler and linker. You should use the GNU configuration
1264 if you can, especially if you also want to use GNU C++. You enabled
1265 that configuration with + the @samp{--with-gnu-as} and @samp{--with-gnu-ld}
1266 options to @code{configure}.
1268 Note the C compiler that comes
1269 with this system cannot compile GNU CC. You can find binaries of GNU CC
1270 for bootstrapping on @code{jagubox.gsfc.nasa.gov}.
1271 You will also a patched version of @file{/bin/ld} there that
1272 raises some of the arbitrary limits found in the original.
1275 AT&T 3b1, a.k.a. 7300 PC. This version of GNU CC cannot
1276 be compiled with the system C compiler, which is too buggy.
1277 You will need to get a previous version of GCC and use it to
1278 bootstrap. Binaries are available from the OSU-CIS archive, at
1279 @url{ftp://archive.cis.ohio-state.edu/pub/att7300/}.
1281 @item m68k-bull-sysv
1282 Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GNU CC works
1283 either with native assembler or GNU assembler. You can use
1284 GNU assembler with native coff generation by providing @samp{--with-gnu-as} to
1285 the configure script or use GNU assembler with dbx-in-coff encapsulation
1286 by providing @samp{--with-gnu-as --stabs}. For any problem with native
1287 assembler or for availability of the DPX/2 port of GAS, contact
1288 @code{F.Pierresteguy@@frcl.bull.fr}.
1290 @item m68k-crds-unox
1291 Use @samp{configure unos} for building on Unos.
1293 The Unos assembler is named @code{casm} instead of @code{as}. For some
1294 strange reason linking @file{/bin/as} to @file{/bin/casm} changes the
1295 behavior, and does not work. So, when installing GNU CC, you should
1296 install the following script as @file{as} in the subdirectory where
1297 the passes of GCC are installed:
1304 The default Unos library is named @file{libunos.a} instead of
1305 @file{libc.a}. To allow GNU CC to function, either change all
1306 references to @samp{-lc} in @file{gcc.c} to @samp{-lunos} or link
1307 @file{/lib/libc.a} to @file{/lib/libunos.a}.
1309 @cindex @code{alloca}, for Unos
1310 When compiling GNU CC with the standard compiler, to overcome bugs in
1311 the support of @code{alloca}, do not use @samp{-O} when making stage 2.
1312 Then use the stage 2 compiler with @samp{-O} to make the stage 3
1313 compiler. This compiler will have the same characteristics as the usual
1314 stage 2 compiler on other systems. Use it to make a stage 4 compiler
1315 and compare that with stage 3 to verify proper compilation.
1317 (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in
1318 the comments there will make the above paragraph superfluous. Please
1319 inform us of whether this works.)
1321 Unos uses memory segmentation instead of demand paging, so you will need
1322 a lot of memory. 5 Mb is barely enough if no other tasks are running.
1323 If linking @file{cc1} fails, try putting the object files into a library
1324 and linking from that library.
1327 HP 9000 series 300 or 400 running HP-UX. HP-UX version 8.0 has a bug in
1328 the assembler that prevents compilation of GNU CC. To fix it, get patch
1331 In addition, if you wish to use gas @samp{--with-gnu-as} you must use
1332 gas version 2.1 or later, and you must use the GNU linker version 2.1 or
1333 later. Earlier versions of gas relied upon a program which converted the
1334 gas output into the native HP-UX format, but that program has not been
1335 kept up to date. gdb does not understand that native HP-UX format, so
1336 you must use gas if you wish to use gdb.
1339 Sun 3. We do not provide a configuration file to use the Sun FPA by
1340 default, because programs that establish signal handlers for floating
1341 point traps inherently cannot work with the FPA.
1343 See @ref{Sun Install}, for information on installing GNU CC on Sun
1347 Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port.
1348 These systems tend to use the Green Hills C, revision 1.8.5, as the
1349 standard C compiler. There are apparently bugs in this compiler that
1350 result in object files differences between stage 2 and stage 3. If this
1351 happens, make the stage 4 compiler and compare it to the stage 3
1352 compiler. If the stage 3 and stage 4 object files are identical, this
1353 suggests you encountered a problem with the standard C compiler; the
1354 stage 3 and 4 compilers may be usable.
1356 It is best, however, to use an older version of GNU CC for bootstrapping
1360 Motorola m88k running DG/UX. To build 88open BCS native or cross
1361 compilers on DG/UX, specify the configuration name as
1362 @samp{m88k-*-dguxbcs} and build in the 88open BCS software development
1363 environment. To build ELF native or cross compilers on DG/UX, specify
1364 @samp{m88k-*-dgux} and build in the DG/UX ELF development environment.
1365 You set the software development environment by issuing
1366 @samp{sde-target} command and specifying either @samp{m88kbcs} or
1367 @samp{m88kdguxelf} as the operand.
1369 If you do not specify a configuration name, @file{configure} guesses the
1370 configuration based on the current software development environment.
1372 @item m88k-tektronix-sysv3
1373 Tektronix XD88 running UTekV 3.2e. Do not turn on
1374 optimization while building stage1 if you bootstrap with
1375 the buggy Green Hills compiler. Also, The bundled LAI
1376 System V NFS is buggy so if you build in an NFS mounted
1377 directory, start from a fresh reboot, or avoid NFS all together.
1378 Otherwise you may have trouble getting clean comparisons
1382 MIPS machines running the MIPS operating system in BSD mode. It's
1383 possible that some old versions of the system lack the functions
1384 @code{memcpy}, @code{memcmp}, and @code{memset}. If your system lacks
1385 these, you must remove or undo the definition of
1386 @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}.
1388 The MIPS C compiler needs to be told to increase its table size
1389 for switch statements with the @samp{-Wf,-XNg1500} option in
1390 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1391 optimization option, you also need to use @samp{-Olimit 3000}.
1392 Both of these options are automatically generated in the
1393 @file{Makefile} that the shell script @file{configure} builds.
1394 If you override the @code{CC} make variable and use the MIPS
1395 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1397 @item mips-mips-riscos*
1398 The MIPS C compiler needs to be told to increase its table size
1399 for switch statements with the @samp{-Wf,-XNg1500} option in
1400 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1401 optimization option, you also need to use @samp{-Olimit 3000}.
1402 Both of these options are automatically generated in the
1403 @file{Makefile} that the shell script @file{configure} builds.
1404 If you override the @code{CC} make variable and use the MIPS
1405 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1407 MIPS computers running RISC-OS can support four different
1408 personalities: default, BSD 4.3, System V.3, and System V.4
1409 (older versions of RISC-OS don't support V.4). To configure GCC
1410 for these platforms use the following configurations:
1413 @item mips-mips-riscos@code{rev}
1414 Default configuration for RISC-OS, revision @code{rev}.
1416 @item mips-mips-riscos@code{rev}bsd
1417 BSD 4.3 configuration for RISC-OS, revision @code{rev}.
1419 @item mips-mips-riscos@code{rev}sysv4
1420 System V.4 configuration for RISC-OS, revision @code{rev}.
1422 @item mips-mips-riscos@code{rev}sysv
1423 System V.3 configuration for RISC-OS, revision @code{rev}.
1426 The revision @code{rev} mentioned above is the revision of
1427 RISC-OS to use. You must reconfigure GCC when going from a
1428 RISC-OS revision 4 to RISC-OS revision 5. This has the effect of
1430 @ifclear INSTALLONLY
1431 bug (see @ref{Installation Problems}, for more details).
1438 In order to compile GCC on an SGI running IRIX 4, the "c.hdr.lib"
1439 option must be installed from the CD-ROM supplied from Silicon Graphics.
1440 This is found on the 2nd CD in release 4.0.1.
1442 In order to compile GCC on an SGI running IRIX 5, the "compiler_dev.hdr"
1443 subsystem must be installed from the IDO CD-ROM supplied by Silicon
1446 @code{make compare} may fail on version 5 of IRIX unless you add
1447 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
1448 assembler input file is stored in the object file, and that makes
1449 comparison fail if it differs between the @code{stage1} and
1450 @code{stage2} compilations. The option @samp{-save-temps} forces a
1451 fixed name to be used for the assembler input file, instead of a
1452 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
1453 unless the comparisons fail without that option. If you do you
1454 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
1455 @samp{.s} files after each series of compilations.
1457 The MIPS C compiler needs to be told to increase its table size
1458 for switch statements with the @samp{-Wf,-XNg1500} option in
1459 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1460 optimization option, you also need to use @samp{-Olimit 3000}.
1461 Both of these options are automatically generated in the
1462 @file{Makefile} that the shell script @file{configure} builds.
1463 If you override the @code{CC} make variable and use the MIPS
1464 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1466 On Irix version 4.0.5F, and perhaps on some other versions as well,
1467 there is an assembler bug that reorders instructions incorrectly. To
1468 work around it, specify the target configuration
1469 @samp{mips-sgi-irix4loser}. This configuration inhibits assembler
1472 In a compiler configured with target @samp{mips-sgi-irix4}, you can turn
1473 off assembler optimization by using the @samp{-noasmopt} option. This
1474 compiler option passes the option @samp{-O0} to the assembler, to
1477 The @samp{-noasmopt} option can be useful for testing whether a problem
1478 is due to erroneous assembler reordering. Even if a problem does not go
1479 away with @samp{-noasmopt}, it may still be due to assembler
1480 reordering---perhaps GNU CC itself was miscompiled as a result.
1482 To enable debugging under Irix 5, you must use GNU as 2.5 or later,
1483 and use the @samp{--with-gnu-as} configure option when configuring gcc.
1484 GNU as is distributed as part of the binutils package.
1486 @item mips-sony-sysv
1487 Sony MIPS NEWS. This works in NEWSOS 5.0.1, but not in 5.0.2 (which
1488 uses ELF instead of COFF). Support for 5.0.2 will probably be provided
1489 soon by volunteers. In particular, the linker does not like the
1490 code generated by GCC when shared libraries are linked in.
1493 Encore ns32000 system. Encore systems are supported only under BSD.
1496 National Semiconductor ns32000 system. Genix has bugs in @code{alloca}
1497 and @code{malloc}; you must get the compiled versions of these from GNU
1501 Go to the Berkeley universe before compiling.
1504 UTEK ns32000 system (``merlin''). The C compiler that comes with this
1505 system cannot compile GNU CC; contact @samp{tektronix!reed!mason} to get
1506 binaries of GNU CC for bootstrapping.
1510 The only operating systems supported for the IBM RT PC are AOS and
1511 MACH. GNU CC does not support AIX running on the RT. We recommend you
1512 compile GNU CC with an earlier version of itself; if you compile GNU CC
1513 with @code{hc}, the Metaware compiler, it will work, but you will get
1514 mismatches between the stage 2 and stage 3 compilers in various files.
1515 These errors are minor differences in some floating-point constants and
1516 can be safely ignored; the stage 3 compiler is correct.
1519 @itemx powerpc-*-aix
1520 Various early versions of each release of the IBM XLC compiler will not
1521 bootstrap GNU CC. Symptoms include differences between the stage2 and
1522 stage3 object files, and errors when compiling @file{libgcc.a} or
1523 @file{enquire}. Known problematic releases include: xlc-1.2.1.8,
1524 xlc-1.3.0.0 (distributed with AIX 3.2.5), and xlc-1.3.0.19. Both
1525 xlc-1.2.1.28 and xlc-1.3.0.24 (PTF 432238) are known to produce working
1526 versions of GNU CC, but most other recent releases correctly bootstrap
1529 Release 4.3.0 of AIX and ones prior to AIX 3.2.4 include a version of
1530 the IBM assembler which does not accept debugging directives: assembler
1531 updates are available as PTFs. Also, if you are using AIX 3.2.5 or
1532 greater and the GNU assembler, you must have a version modified after
1533 October 16th, 1995 in order for the GNU C compiler to build. See the
1534 file @file{README.RS6000} for more details on any of these problems.
1536 GNU CC does not yet support the 64-bit PowerPC instructions.
1538 Objective C does not work on this architecture because it makes assumptions
1539 that are incompatible with the calling conventions.
1541 AIX on the RS/6000 provides support (NLS) for environments outside of
1542 the United States. Compilers and assemblers use NLS to support
1543 locale-specific representations of various objects including
1544 floating-point numbers ("." vs "," for separating decimal fractions).
1545 There have been problems reported where the library linked with GNU CC
1546 does not produce the same floating-point formats that the assembler
1547 accepts. If you have this problem, set the LANG environment variable to
1550 Due to changes in the way that GNU CC invokes the binder (linker) for AIX
1551 4.1, you may now receive warnings of duplicate symbols from the link step
1552 that were not reported before. The assembly files generated by GNU CC for
1553 AIX have always included multiple symbol definitions for certain global
1554 variable and function declarations in the original program. The warnings
1555 should not prevent the linker from producing a correct library or runnable
1558 By default, AIX 4.1 produces code that can be used on either Power or
1561 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1562 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1565 @itemx powerpc-*-sysv4
1566 PowerPC system in big endian mode, running System V.4.
1568 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1569 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1571 @item powerpc-*-linux
1572 @itemx powerpc-*-linux-gnu
1573 PowerPC system in big endian mode, running the Linux-based GNU system.
1575 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1576 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1578 @item powerpc-*-eabiaix
1579 Embedded PowerPC system in big endian mode with -mcall-aix selected as
1582 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1583 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1585 @item powerpc-*-eabisim
1586 Embedded PowerPC system in big endian mode for use in running under the
1589 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1590 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1592 @item powerpc-*-eabi
1593 Embedded PowerPC system in big endian mode.
1595 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1596 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1598 @item powerpcle-*-elf
1599 @itemx powerpcle-*-sysv4
1600 PowerPC system in little endian mode, running System V.4.
1602 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1603 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1605 @item powerpcle-*-solaris2*
1606 PowerPC system in little endian mode, running Solaris 2.5.1 or higher.
1608 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1609 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1610 Beta versions of the Sun 4.0 compiler do not seem to be able to build
1611 GNU CC correctly. There are also problems with the host assembler and
1612 linker that are fixed by using the GNU versions of these tools.
1614 @item powerpcle-*-eabisim
1615 Embedded PowerPC system in little endian mode for use in running under
1618 @itemx powerpcle-*-eabi
1619 Embedded PowerPC system in little endian mode.
1621 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1622 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1624 @item powerpcle-*-winnt
1625 @itemx powerpcle-*-pe
1626 PowerPC system in little endian mode running Windows NT.
1628 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1629 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1631 @item vax-dec-ultrix
1632 Don't try compiling with Vax C (@code{vcc}). It produces incorrect code
1633 in some cases (for example, when @code{alloca} is used).
1635 Meanwhile, compiling @file{cp/parse.c} with pcc does not work because of
1636 an internal table size limitation in that compiler. To avoid this
1637 problem, compile just the GNU C compiler first, and use it to recompile
1638 building all the languages that you want to run.
1641 See @ref{Sun Install}, for information on installing GNU CC on Sun
1645 See @ref{VMS Install}, for details on how to install GNU CC on VMS.
1648 These computers are also known as the 3b2, 3b5, 3b20 and other similar
1649 names. (However, the 3b1 is actually a 68000; see
1650 @ref{Configurations}.)
1652 Don't use @samp{-g} when compiling with the system's compiler. The
1653 system's linker seems to be unable to handle such a large program with
1654 debugging information.
1656 The system's compiler runs out of capacity when compiling @file{stmt.c}
1657 in GNU CC. You can work around this by building @file{cpp} in GNU CC
1658 first, then use that instead of the system's preprocessor with the
1659 system's C compiler to compile @file{stmt.c}. Here is how:
1662 mv /lib/cpp /lib/cpp.att
1664 echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp
1668 The system's compiler produces bad code for some of the GNU CC
1669 optimization files. So you must build the stage 2 compiler without
1670 optimization. Then build a stage 3 compiler with optimization.
1671 That executable should work. Here are the necessary commands:
1674 make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g"
1676 make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O"
1679 You may need to raise the ULIMIT setting to build a C++ compiler,
1680 as the file @file{cc1plus} is larger than one megabyte.
1684 @section Compilation in a Separate Directory
1685 @cindex other directory, compilation in
1686 @cindex compilation in a separate directory
1687 @cindex separate directory, compilation in
1689 If you wish to build the object files and executables in a directory
1690 other than the one containing the source files, here is what you must
1695 Make sure you have a version of Make that supports the @code{VPATH}
1696 feature. (GNU Make supports it, as do Make versions on most BSD
1700 If you have ever run @file{configure} in the source directory, you must undo
1701 the configuration. Do this by running:
1708 Go to the directory in which you want to build the compiler before
1709 running @file{configure}:
1716 On systems that do not support symbolic links, this directory must be
1717 on the same file system as the source code directory.
1720 Specify where to find @file{configure} when you run it:
1723 ../gcc/configure @dots{}
1726 This also tells @code{configure} where to find the compiler sources;
1727 @code{configure} takes the directory from the file name that was used to
1728 invoke it. But if you want to be sure, you can specify the source
1729 directory with the @samp{--srcdir} option, like this:
1732 ../gcc/configure --srcdir=../gcc @var{other options}
1735 The directory you specify with @samp{--srcdir} need not be the same
1736 as the one that @code{configure} is found in.
1739 Now, you can run @code{make} in that directory. You need not repeat the
1740 configuration steps shown above, when ordinary source files change. You
1741 must, however, run @code{configure} again when the configuration files
1742 change, if your system does not support symbolic links.
1744 @node Cross-Compiler
1745 @section Building and Installing a Cross-Compiler
1746 @cindex cross-compiler, installation
1748 GNU CC can function as a cross-compiler for many machines, but not all.
1752 Cross-compilers for the Mips as target using the Mips assembler
1753 currently do not work, because the auxiliary programs
1754 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
1755 anything but a Mips. It does work to cross compile for a Mips
1756 if you use the GNU assembler and linker.
1759 Cross-compilers between machines with different floating point formats
1760 have not all been made to work. GNU CC now has a floating point
1761 emulator with which these can work, but each target machine description
1762 needs to be updated to take advantage of it.
1765 Cross-compilation between machines of different word sizes is
1766 somewhat problematic and sometimes does not work.
1769 Since GNU CC generates assembler code, you probably need a
1770 cross-assembler that GNU CC can run, in order to produce object files.
1771 If you want to link on other than the target machine, you need a
1772 cross-linker as well. You also need header files and libraries suitable
1773 for the target machine that you can install on the host machine.
1776 * Steps of Cross:: Using a cross-compiler involves several steps
1777 that may be carried out on different machines.
1778 * Configure Cross:: Configuring a cross-compiler.
1779 * Tools and Libraries:: Where to put the linker and assembler, and the C library.
1780 * Cross Headers:: Finding and installing header files
1781 for a cross-compiler.
1782 * Cross Runtime:: Supplying arithmetic runtime routines (@file{libgcc1.a}).
1783 * Build Cross:: Actually compiling the cross-compiler.
1786 @node Steps of Cross
1787 @subsection Steps of Cross-Compilation
1789 To compile and run a program using a cross-compiler involves several
1794 Run the cross-compiler on the host machine to produce assembler files
1795 for the target machine. This requires header files for the target
1799 Assemble the files produced by the cross-compiler. You can do this
1800 either with an assembler on the target machine, or with a
1801 cross-assembler on the host machine.
1804 Link those files to make an executable. You can do this either with a
1805 linker on the target machine, or with a cross-linker on the host
1806 machine. Whichever machine you use, you need libraries and certain
1807 startup files (typically @file{crt@dots{}.o}) for the target machine.
1810 It is most convenient to do all of these steps on the same host machine,
1811 since then you can do it all with a single invocation of GNU CC. This
1812 requires a suitable cross-assembler and cross-linker. For some targets,
1813 the GNU assembler and linker are available.
1815 @node Configure Cross
1816 @subsection Configuring a Cross-Compiler
1818 To build GNU CC as a cross-compiler, you start out by running
1819 @file{configure}. Use the @samp{--target=@var{target}} to specify the
1820 target type. If @file{configure} was unable to correctly identify the
1821 system you are running on, also specify the @samp{--build=@var{build}}
1822 option. For example, here is how to configure for a cross-compiler that
1823 produces code for an HP 68030 system running BSD on a system that
1824 @file{configure} can correctly identify:
1827 ./configure --target=m68k-hp-bsd4.3
1830 @node Tools and Libraries
1831 @subsection Tools and Libraries for a Cross-Compiler
1833 If you have a cross-assembler and cross-linker available, you should
1834 install them now. Put them in the directory
1835 @file{/usr/local/@var{target}/bin}. Here is a table of the tools
1836 you should put in this directory:
1840 This should be the cross-assembler.
1843 This should be the cross-linker.
1846 This should be the cross-archiver: a program which can manipulate
1847 archive files (linker libraries) in the target machine's format.
1850 This should be a program to construct a symbol table in an archive file.
1853 The installation of GNU CC will find these programs in that directory,
1854 and copy or link them to the proper place to for the cross-compiler to
1855 find them when run later.
1857 The easiest way to provide these files is to build the Binutils package
1858 and GAS. Configure them with the same @samp{--host} and @samp{--target}
1859 options that you use for configuring GNU CC, then build and install
1860 them. They install their executables automatically into the proper
1861 directory. Alas, they do not support all the targets that GNU CC
1864 If you want to install libraries to use with the cross-compiler, such as
1865 a standard C library, put them in the directory
1866 @file{/usr/local/@var{target}/lib}; installation of GNU CC copies
1867 all the files in that subdirectory into the proper place for GNU CC to
1868 find them and link with them. Here's an example of copying some
1869 libraries from a target machine:
1872 ftp @var{target-machine}
1873 lcd /usr/local/@var{target}/lib
1883 The precise set of libraries you'll need, and their locations on
1884 the target machine, vary depending on its operating system.
1887 Many targets require ``start files'' such as @file{crt0.o} and
1888 @file{crtn.o} which are linked into each executable; these too should be
1889 placed in @file{/usr/local/@var{target}/lib}. There may be several
1890 alternatives for @file{crt0.o}, for use with profiling or other
1891 compilation options. Check your target's definition of
1892 @code{STARTFILE_SPEC} to find out what start files it uses.
1893 Here's an example of copying these files from a target machine:
1896 ftp @var{target-machine}
1897 lcd /usr/local/@var{target}/lib
1907 @subsection @file{libgcc.a} and Cross-Compilers
1909 Code compiled by GNU CC uses certain runtime support functions
1910 implicitly. Some of these functions can be compiled successfully with
1911 GNU CC itself, but a few cannot be. These problem functions are in the
1912 source file @file{libgcc1.c}; the library made from them is called
1915 When you build a native compiler, these functions are compiled with some
1916 other compiler--the one that you use for bootstrapping GNU CC.
1917 Presumably it knows how to open code these operations, or else knows how
1918 to call the run-time emulation facilities that the machine comes with.
1919 But this approach doesn't work for building a cross-compiler. The
1920 compiler that you use for building knows about the host system, not the
1923 So, when you build a cross-compiler you have to supply a suitable
1924 library @file{libgcc1.a} that does the job it is expected to do.
1926 To compile @file{libgcc1.c} with the cross-compiler itself does not
1927 work. The functions in this file are supposed to implement arithmetic
1928 operations that GNU CC does not know how to open code for your target
1929 machine. If these functions are compiled with GNU CC itself, they
1930 will compile into infinite recursion.
1932 On any given target, most of these functions are not needed. If GNU CC
1933 can open code an arithmetic operation, it will not call these functions
1934 to perform the operation. It is possible that on your target machine,
1935 none of these functions is needed. If so, you can supply an empty
1936 library as @file{libgcc1.a}.
1938 Many targets need library support only for multiplication and division.
1939 If you are linking with a library that contains functions for
1940 multiplication and division, you can tell GNU CC to call them directly
1941 by defining the macros @code{MULSI3_LIBCALL}, and the like. These
1942 macros need to be defined in the target description macro file. For
1943 some targets, they are defined already. This may be sufficient to
1944 avoid the need for libgcc1.a; if so, you can supply an empty library.
1946 Some targets do not have floating point instructions; they need other
1947 functions in @file{libgcc1.a}, which do floating arithmetic.
1948 Recent versions of GNU CC have a file which emulates floating point.
1949 With a certain amount of work, you should be able to construct a
1950 floating point emulator that can be used as @file{libgcc1.a}. Perhaps
1951 future versions will contain code to do this automatically and
1952 conveniently. That depends on whether someone wants to implement it.
1954 Some embedded targets come with all the necessary @file{libgcc1.a}
1955 routines written in C or assembler. These targets build
1956 @file{libgcc1.a} automatically and you do not need to do anything
1957 special for them. Other embedded targets do not need any
1958 @file{libgcc1.a} routines since all the necessary operations are
1959 supported by the hardware.
1961 If your target system has another C compiler, you can configure GNU CC
1962 as a native compiler on that machine, build just @file{libgcc1.a} with
1963 @samp{make libgcc1.a} on that machine, and use the resulting file with
1964 the cross-compiler. To do this, execute the following on the target
1968 cd @var{target-build-dir}
1969 ./configure --host=sparc --target=sun3
1974 And then this on the host machine:
1977 ftp @var{target-machine}
1979 cd @var{target-build-dir}
1984 Another way to provide the functions you need in @file{libgcc1.a} is to
1985 define the appropriate @code{perform_@dots{}} macros for those
1986 functions. If these definitions do not use the C arithmetic operators
1987 that they are meant to implement, you should be able to compile them
1988 with the cross-compiler you are building. (If these definitions already
1989 exist for your target file, then you are all set.)
1991 To build @file{libgcc1.a} using the perform macros, use
1992 @samp{LIBGCC1=libgcc1.a OLDCC=./xgcc} when building the compiler.
1993 Otherwise, you should place your replacement library under the name
1994 @file{libgcc1.a} in the directory in which you will build the
1995 cross-compiler, before you run @code{make}.
1998 @subsection Cross-Compilers and Header Files
2000 If you are cross-compiling a standalone program or a program for an
2001 embedded system, then you may not need any header files except the few
2002 that are part of GNU CC (and those of your program). However, if you
2003 intend to link your program with a standard C library such as
2004 @file{libc.a}, then you probably need to compile with the header files
2005 that go with the library you use.
2007 The GNU C compiler does not come with these files, because (1) they are
2008 system-specific, and (2) they belong in a C library, not in a compiler.
2010 If the GNU C library supports your target machine, then you can get the
2011 header files from there (assuming you actually use the GNU library when
2012 you link your program).
2014 If your target machine comes with a C compiler, it probably comes with
2015 suitable header files also. If you make these files accessible from the host
2016 machine, the cross-compiler can use them also.
2018 Otherwise, you're on your own in finding header files to use when
2021 When you have found suitable header files, put them in the directory
2022 @file{/usr/local/@var{target}/include}, before building the cross
2023 compiler. Then installation will run fixincludes properly and install
2024 the corrected versions of the header files where the compiler will use
2027 Provide the header files before you build the cross-compiler, because
2028 the build stage actually runs the cross-compiler to produce parts of
2029 @file{libgcc.a}. (These are the parts that @emph{can} be compiled with
2030 GNU CC.) Some of them need suitable header files.
2032 Here's an example showing how to copy the header files from a target
2033 machine. On the target machine, do this:
2036 (cd /usr/include; tar cf - .) > tarfile
2039 Then, on the host machine, do this:
2042 ftp @var{target-machine}
2043 lcd /usr/local/@var{target}/include
2050 @subsection Actually Building the Cross-Compiler
2052 Now you can proceed just as for compiling a single-machine compiler
2053 through the step of building stage 1. If you have not provided some
2054 sort of @file{libgcc1.a}, then compilation will give up at the point
2055 where it needs that file, printing a suitable error message. If you
2056 do provide @file{libgcc1.a}, then building the compiler will automatically
2057 compile and link a test program called @file{libgcc1-test}; if you get
2058 errors in the linking, it means that not all of the necessary routines
2059 in @file{libgcc1.a} are available.
2061 You must provide the header file @file{float.h}. One way to do this is
2062 to compile @file{enquire} and run it on your target machine. The job of
2063 @file{enquire} is to run on the target machine and figure out by
2064 experiment the nature of its floating point representation.
2065 @file{enquire} records its findings in the header file @file{float.h}.
2066 If you can't produce this file by running @file{enquire} on the target
2067 machine, then you will need to come up with a suitable @file{float.h} in
2068 some other way (or else, avoid using it in your programs).
2070 Do not try to build stage 2 for a cross-compiler. It doesn't work to
2071 rebuild GNU CC as a cross-compiler using the cross-compiler, because
2072 that would produce a program that runs on the target machine, not on the
2073 host. For example, if you compile a 386-to-68030 cross-compiler with
2074 itself, the result will not be right either for the 386 (because it was
2075 compiled into 68030 code) or for the 68030 (because it was configured
2076 for a 386 as the host). If you want to compile GNU CC into 68030 code,
2077 whether you compile it on a 68030 or with a cross-compiler on a 386, you
2078 must specify a 68030 as the host when you configure it.
2080 To install the cross-compiler, use @samp{make install}, as usual.
2083 @section Installing GNU CC on the Sun
2084 @cindex Sun installation
2085 @cindex installing GNU CC on the Sun
2087 On Solaris, do not use the linker or other tools in
2088 @file{/usr/ucb} to build GNU CC. Use @code{/usr/ccs/bin}.
2090 If the assembler reports @samp{Error: misaligned data} when bootstrapping,
2091 you are probably using an obsolete version of the GNU assembler. Upgrade
2092 to the latest version of GNU @code{binutils}, or use the Solaris assembler.
2094 Make sure the environment variable @code{FLOAT_OPTION} is not set when
2095 you compile @file{libgcc.a}. If this option were set to @code{f68881}
2096 when @file{libgcc.a} is compiled, the resulting code would demand to be
2097 linked with a special startup file and would not link properly without
2100 @cindex @code{alloca}, for SunOS
2101 There is a bug in @code{alloca} in certain versions of the Sun library.
2102 To avoid this bug, install the binaries of GNU CC that were compiled by
2103 GNU CC. They use @code{alloca} as a built-in function and never the one
2106 Some versions of the Sun compiler crash when compiling GNU CC. The
2107 problem is a segmentation fault in cpp. This problem seems to be due to
2108 the bulk of data in the environment variables. You may be able to avoid
2109 it by using the following command to compile GNU CC with Sun CC:
2112 make CC="TERMCAP=x OBJS=x LIBFUNCS=x STAGESTUFF=x cc"
2115 SunOS 4.1.3 and 4.1.3_U1 have bugs that can cause intermittent core
2116 dumps when compiling GNU CC. A common symptom is an
2117 internal compiler error which does not recur if you run it again.
2118 To fix the problem, install Sun recommended patch 100726 (for SunOS 4.1.3)
2119 or 101508 (for SunOS 4.1.3_U1), or upgrade to a later SunOS release.
2122 @section Installing GNU CC on VMS
2123 @cindex VMS installation
2124 @cindex installing GNU CC on VMS
2126 The VMS version of GNU CC is distributed in a backup saveset containing
2127 both source code and precompiled binaries.
2129 To install the @file{gcc} command so you can use the compiler easily, in
2130 the same manner as you use the VMS C compiler, you must install the VMS CLD
2131 file for GNU CC as follows:
2135 Define the VMS logical names @samp{GNU_CC} and @samp{GNU_CC_INCLUDE}
2136 to point to the directories where the GNU CC executables
2137 (@file{gcc-cpp.exe}, @file{gcc-cc1.exe}, etc.) and the C include files are
2138 kept respectively. This should be done with the commands:@refill
2141 $ assign /system /translation=concealed -
2143 $ assign /system /translation=concealed -
2144 disk:[gcc.include.] gnu_cc_include
2148 with the appropriate disk and directory names. These commands can be
2149 placed in your system startup file so they will be executed whenever
2150 the machine is rebooted. You may, if you choose, do this via the
2151 @file{GCC_INSTALL.COM} script in the @file{[GCC]} directory.
2154 Install the @file{GCC} command with the command line:
2157 $ set command /table=sys$common:[syslib]dcltables -
2158 /output=sys$common:[syslib]dcltables gnu_cc:[000000]gcc
2159 $ install replace sys$common:[syslib]dcltables
2163 To install the help file, do the following:
2166 $ library/help sys$library:helplib.hlb gcc.hlp
2170 Now you can invoke the compiler with a command like @samp{gcc /verbose
2171 file.c}, which is equivalent to the command @samp{gcc -v -c file.c} in
2175 If you wish to use GNU C++ you must first install GNU CC, and then
2176 perform the following steps:
2180 Define the VMS logical name @samp{GNU_GXX_INCLUDE} to point to the
2181 directory where the preprocessor will search for the C++ header files.
2182 This can be done with the command:@refill
2185 $ assign /system /translation=concealed -
2186 disk:[gcc.gxx_include.] gnu_gxx_include
2190 with the appropriate disk and directory name. If you are going to be
2191 using a C++ runtime library, this is where its install procedure will install
2195 Obtain the file @file{gcc-cc1plus.exe}, and place this in the same
2196 directory that @file{gcc-cc1.exe} is kept.
2198 The GNU C++ compiler can be invoked with a command like @samp{gcc /plus
2199 /verbose file.cc}, which is equivalent to the command @samp{g++ -v -c
2203 We try to put corresponding binaries and sources on the VMS distribution
2204 tape. But sometimes the binaries will be from an older version than the
2205 sources, because we don't always have time to update them. (Use the
2206 @samp{/version} option to determine the version number of the binaries and
2207 compare it with the source file @file{version.c} to tell whether this is
2208 so.) In this case, you should use the binaries you get to recompile the
2209 sources. If you must recompile, here is how:
2213 Execute the command procedure @file{vmsconfig.com} to set up the files
2214 @file{tm.h}, @file{config.h}, @file{aux-output.c}, and @file{md.}, and
2215 to create files @file{tconfig.h} and @file{hconfig.h}. This procedure
2216 also creates several linker option files used by @file{make-cc1.com} and
2217 a data file used by @file{make-l2.com}.@refill
2224 Setup the logical names and command tables as defined above. In
2225 addition, define the VMS logical name @samp{GNU_BISON} to point at the
2226 to the directories where the Bison executable is kept. This should be
2227 done with the command:@refill
2230 $ assign /system /translation=concealed -
2231 disk:[bison.] gnu_bison
2234 You may, if you choose, use the @file{INSTALL_BISON.COM} script in the
2235 @file{[BISON]} directory.
2238 Install the @samp{BISON} command with the command line:@refill
2241 $ set command /table=sys$common:[syslib]dcltables -
2242 /output=sys$common:[syslib]dcltables -
2243 gnu_bison:[000000]bison
2244 $ install replace sys$common:[syslib]dcltables
2248 Type @samp{@@make-gcc} to recompile everything (alternatively, submit
2249 the file @file{make-gcc.com} to a batch queue). If you wish to build
2250 the GNU C++ compiler as well as the GNU CC compiler, you must first edit
2251 @file{make-gcc.com} and follow the instructions that appear in the
2255 In order to use GCC, you need a library of functions which GCC compiled code
2256 will call to perform certain tasks, and these functions are defined in the
2257 file @file{libgcc2.c}. To compile this you should use the command procedure
2258 @file{make-l2.com}, which will generate the library @file{libgcc2.olb}.
2259 @file{libgcc2.olb} should be built using the compiler built from
2260 the same distribution that @file{libgcc2.c} came from, and
2261 @file{make-gcc.com} will automatically do all of this for you.
2263 To install the library, use the following commands:@refill
2266 $ library gnu_cc:[000000]gcclib/delete=(new,eprintf)
2267 $ library gnu_cc:[000000]gcclib/delete=L_*
2268 $ library libgcc2/extract=*/output=libgcc2.obj
2269 $ library gnu_cc:[000000]gcclib libgcc2.obj
2272 The first command simply removes old modules that will be replaced with
2273 modules from @file{libgcc2} under different module names. The modules
2274 @code{new} and @code{eprintf} may not actually be present in your
2275 @file{gcclib.olb}---if the VMS librarian complains about those modules
2276 not being present, simply ignore the message and continue on with the
2277 next command. The second command removes the modules that came from the
2278 previous version of the library @file{libgcc2.c}.
2280 Whenever you update the compiler on your system, you should also update the
2281 library with the above procedure.
2284 You may wish to build GCC in such a way that no files are written to the
2285 directory where the source files reside. An example would be the when
2286 the source files are on a read-only disk. In these cases, execute the
2287 following DCL commands (substituting your actual path names):
2290 $ assign dua0:[gcc.build_dir.]/translation=concealed, -
2291 dua1:[gcc.source_dir.]/translation=concealed gcc_build
2292 $ set default gcc_build:[000000]
2296 where the directory @file{dua1:[gcc.source_dir]} contains the source
2297 code, and the directory @file{dua0:[gcc.build_dir]} is meant to contain
2298 all of the generated object files and executables. Once you have done
2299 this, you can proceed building GCC as described above. (Keep in mind
2300 that @file{gcc_build} is a rooted logical name, and thus the device
2301 names in each element of the search list must be an actual physical
2302 device name rather than another rooted logical name).
2305 @strong{If you are building GNU CC with a previous version of GNU CC,
2306 you also should check to see that you have the newest version of the
2307 assembler}. In particular, GNU CC version 2 treats global constant
2308 variables slightly differently from GNU CC version 1, and GAS version
2309 1.38.1 does not have the patches required to work with GCC version 2.
2310 If you use GAS 1.38.1, then @code{extern const} variables will not have
2311 the read-only bit set, and the linker will generate warning messages
2312 about mismatched psect attributes for these variables. These warning
2313 messages are merely a nuisance, and can safely be ignored.
2315 If you are compiling with a version of GNU CC older than 1.33, specify
2316 @samp{/DEFINE=("inline=")} as an option in all the compilations. This
2317 requires editing all the @code{gcc} commands in @file{make-cc1.com}.
2318 (The older versions had problems supporting @code{inline}.) Once you
2319 have a working 1.33 or newer GNU CC, you can change this file back.
2322 If you want to build GNU CC with the VAX C compiler, you will need to
2323 make minor changes in @file{make-cccp.com} and @file{make-cc1.com}
2324 to choose alternate definitions of @code{CC}, @code{CFLAGS}, and
2325 @code{LIBS}. See comments in those files. However, you must
2326 also have a working version of the GNU assembler (GNU as, aka GAS) as
2327 it is used as the back-end for GNU CC to produce binary object modules
2328 and is not included in the GNU CC sources. GAS is also needed to
2329 compile @file{libgcc2} in order to build @file{gcclib} (see above);
2330 @file{make-l2.com} expects to be able to find it operational in
2331 @file{gnu_cc:[000000]gnu-as.exe}.
2333 To use GNU CC on VMS, you need the VMS driver programs
2334 @file{gcc.exe}, @file{gcc.com}, and @file{gcc.cld}. They are
2335 distributed with the VMS binaries (@file{gcc-vms}) rather than the
2336 GNU CC sources. GAS is also included in @file{gcc-vms}, as is Bison.
2338 Once you have successfully built GNU CC with VAX C, you should use the
2339 resulting compiler to rebuild itself. Before doing this, be sure to
2340 restore the @code{CC}, @code{CFLAGS}, and @code{LIBS} definitions in
2341 @file{make-cccp.com} and @file{make-cc1.com}. The second generation
2342 compiler will be able to take advantage of many optimizations that must
2343 be suppressed when building with other compilers.
2346 Under previous versions of GNU CC, the generated code would occasionally
2347 give strange results when linked with the sharable @file{VAXCRTL} library.
2348 Now this should work.
2350 Even with this version, however, GNU CC itself should not be linked with
2351 the sharable @file{VAXCRTL}. The version of @code{qsort} in
2352 @file{VAXCRTL} has a bug (known to be present in VMS versions V4.6
2353 through V5.5) which causes the compiler to fail.
2355 The executables are generated by @file{make-cc1.com} and
2356 @file{make-cccp.com} use the object library version of @file{VAXCRTL} in
2357 order to make use of the @code{qsort} routine in @file{gcclib.olb}. If
2358 you wish to link the compiler executables with the shareable image
2359 version of @file{VAXCRTL}, you should edit the file @file{tm.h} (created
2360 by @file{vmsconfig.com}) to define the macro @code{QSORT_WORKAROUND}.
2362 @code{QSORT_WORKAROUND} is always defined when GNU CC is compiled with
2363 VAX C, to avoid a problem in case @file{gcclib.olb} is not yet
2367 @section @code{collect2}
2369 GNU CC uses a utility called @code{collect2} on nearly all systems to arrange
2370 to call various initialization functions at start time.
2372 The program @code{collect2} works by linking the program once and
2373 looking through the linker output file for symbols with particular names
2374 indicating they are constructor functions. If it finds any, it
2375 creates a new temporary @samp{.c} file containing a table of them,
2376 compiles it, and links the program a second time including that file.
2379 @cindex constructors, automatic calls
2380 The actual calls to the constructors are carried out by a subroutine
2381 called @code{__main}, which is called (automatically) at the beginning
2382 of the body of @code{main} (provided @code{main} was compiled with GNU
2383 CC). Calling @code{__main} is necessary, even when compiling C code, to
2384 allow linking C and C++ object code together. (If you use
2385 @samp{-nostdlib}, you get an unresolved reference to @code{__main},
2386 since it's defined in the standard GCC library. Include @samp{-lgcc} at
2387 the end of your compiler command line to resolve this reference.)
2389 The program @code{collect2} is installed as @code{ld} in the directory
2390 where the passes of the compiler are installed. When @code{collect2}
2391 needs to find the @emph{real} @code{ld}, it tries the following file
2396 @file{real-ld} in the directories listed in the compiler's search
2400 @file{real-ld} in the directories listed in the environment variable
2404 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2408 @file{ld} in the compiler's search directories, except that
2409 @code{collect2} will not execute itself recursively.
2412 @file{ld} in @code{PATH}.
2415 ``The compiler's search directories'' means all the directories where
2416 @code{gcc} searches for passes of the compiler. This includes
2417 directories that you specify with @samp{-B}.
2419 Cross-compilers search a little differently:
2423 @file{real-ld} in the compiler's search directories.
2426 @file{@var{target}-real-ld} in @code{PATH}.
2429 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2433 @file{ld} in the compiler's search directories.
2436 @file{@var{target}-ld} in @code{PATH}.
2439 @code{collect2} explicitly avoids running @code{ld} using the file name
2440 under which @code{collect2} itself was invoked. In fact, it remembers
2441 up a list of such names---in case one copy of @code{collect2} finds
2442 another copy (or version) of @code{collect2} installed as @code{ld} in a
2443 second place in the search path.
2445 @code{collect2} searches for the utilities @code{nm} and @code{strip}
2446 using the same algorithm as above for @code{ld}.
2449 @section Standard Header File Directories
2451 @code{GCC_INCLUDE_DIR} means the same thing for native and cross. It is
2452 where GNU CC stores its private include files, and also where GNU CC
2453 stores the fixed include files. A cross compiled GNU CC runs
2454 @code{fixincludes} on the header files in @file{$(tooldir)/include}.
2455 (If the cross compilation header files need to be fixed, they must be
2456 installed before GNU CC is built. If the cross compilation header files
2457 are already suitable for ANSI C and GNU CC, nothing special need be
2460 @code{GPLUSPLUS_INCLUDE_DIR} means the same thing for native and cross. It
2461 is where @code{g++} looks first for header files. The C++ library
2462 installs only target independent header files in that directory.
2464 @code{LOCAL_INCLUDE_DIR} is used only for a native compiler. It is
2465 normally @file{/usr/local/include}. GNU CC searches this directory so
2466 that users can install header files in @file{/usr/local/include}.
2468 @code{CROSS_INCLUDE_DIR} is used only for a cross compiler. GNU CC
2469 doesn't install anything there.
2471 @code{TOOL_INCLUDE_DIR} is used for both native and cross compilers. It
2472 is the place for other packages to install header files that GNU CC will
2473 use. For a cross-compiler, this is the equivalent of
2474 @file{/usr/include}. When you build a cross-compiler,
2475 @code{fixincludes} processes any header files in this directory.