1 @c Copyright (C) 1988,89,92,93,94,95,96,97,1998 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8 Note most of this information is out of date and superceded by the EGCS
9 install procedures. It is provided for historical reference only.
13 @chapter Installing GNU CC
15 @cindex installing GNU CC
18 * Configurations:: Configurations Supported by GNU CC.
19 * Other Dir:: Compiling in a separate directory (not where the source is).
20 * Cross-Compiler:: Building and installing a cross-compiler.
21 * Sun Install:: See below for installation on the Sun.
22 * VMS Install:: See below for installation on VMS.
23 * Collect2:: How @code{collect2} works; how it finds @code{ld}.
24 * Header Dirs:: Understanding the standard header file directories.
27 Here is the procedure for installing GNU CC on a Unix system. See
28 @ref{VMS Install}, for VMS systems. In this section we assume you
29 compile in the same directory that contains the source files; see
30 @ref{Other Dir}, to find out how to compile in a separate directory on Unix
33 You cannot install GNU C by itself on MSDOS; it will not compile under
34 any MSDOS compiler except itself. You need to get the complete
35 compilation package DJGPP, which includes binaries as well as sources,
36 and includes all the necessary compilation tools and libraries.
40 If you have built GNU CC previously in the same directory for a
41 different target machine, do @samp{make distclean} to delete all files
42 that might be invalid. One of the files this deletes is
43 @file{Makefile}; if @samp{make distclean} complains that @file{Makefile}
44 does not exist, it probably means that the directory is already suitably
48 On a System V release 4 system, make sure @file{/usr/bin} precedes
49 @file{/usr/ucb} in @code{PATH}. The @code{cc} command in
50 @file{/usr/ucb} uses libraries which have bugs.
53 Specify the host, build and target machine configurations. You do this
54 by running the file @file{configure}.
56 The @dfn{build} machine is the system which you are using, the
57 @dfn{host} machine is the system where you want to run the resulting
58 compiler (normally the build machine), and the @dfn{target} machine is
59 the system for which you want the compiler to generate code.
61 If you are building a compiler to produce code for the machine it runs
62 on (a native compiler), you normally do not need to specify any operands
63 to @file{configure}; it will try to guess the type of machine you are on
64 and use that as the build, host and target machines. So you don't need
65 to specify a configuration when building a native compiler unless
66 @file{configure} cannot figure out what your configuration is or guesses
69 In those cases, specify the build machine's @dfn{configuration name}
70 with the @samp{--host} option; the host and target will default to be
71 the same as the host machine. (If you are building a cross-compiler,
72 see @ref{Cross-Compiler}.)
77 ./configure --host=sparc-sun-sunos4.1
80 A configuration name may be canonical or it may be more or less
83 A canonical configuration name has three parts, separated by dashes.
84 It looks like this: @samp{@var{cpu}-@var{company}-@var{system}}.
85 (The three parts may themselves contain dashes; @file{configure}
86 can figure out which dashes serve which purpose.) For example,
87 @samp{m68k-sun-sunos4.1} specifies a Sun 3.
89 You can also replace parts of the configuration by nicknames or aliases.
90 For example, @samp{sun3} stands for @samp{m68k-sun}, so
91 @samp{sun3-sunos4.1} is another way to specify a Sun 3. You can also
92 use simply @samp{sun3-sunos}, since the version of SunOS is assumed by
93 default to be version 4.
95 You can specify a version number after any of the system types, and some
96 of the CPU types. In most cases, the version is irrelevant, and will be
97 ignored. So you might as well specify the version if you know it.
99 See @ref{Configurations}, for a list of supported configuration names and
100 notes on many of the configurations. You should check the notes in that
101 section before proceeding any further with the installation of GNU CC.
103 There are four additional options you can specify independently to
104 describe variant hardware and software configurations. These are
105 @samp{--with-gnu-as}, @samp{--with-gnu-ld}, @samp{--with-stabs} and
110 If you will use GNU CC with the GNU assembler (GAS), you should declare
111 this by using the @samp{--with-gnu-as} option when you run
114 Using this option does not install GAS. It only modifies the output of
115 GNU CC to work with GAS. Building and installing GAS is up to you.
117 Conversely, if you @emph{do not} wish to use GAS and do not specify
118 @samp{--with-gnu-as} when building GNU CC, it is up to you to make sure
119 that GAS is not installed. GNU CC searches for a program named
120 @code{as} in various directories; if the program it finds is GAS, then
121 it runs GAS. If you are not sure where GNU CC finds the assembler it is
122 using, try specifying @samp{-v} when you run it.
124 The systems where it makes a difference whether you use GAS are@*
125 @samp{hppa1.0-@var{any}-@var{any}}, @samp{hppa1.1-@var{any}-@var{any}},
126 @samp{i386-@var{any}-sysv}, @samp{i386-@var{any}-isc},@*
127 @samp{i860-@var{any}-bsd}, @samp{m68k-bull-sysv},@*
128 @samp{m68k-hp-hpux}, @samp{m68k-sony-bsd},@*
129 @samp{m68k-altos-sysv}, @samp{m68000-hp-hpux},@*
130 @samp{m68000-att-sysv}, @samp{@var{any}-lynx-lynxos},
131 and @samp{mips-@var{any}}).
132 On any other system, @samp{--with-gnu-as} has no effect.
134 On the systems listed above (except for the HP-PA, for ISC on the
135 386, and for @samp{mips-sgi-irix5.*}), if you use GAS, you should also
136 use the GNU linker (and specify @samp{--with-gnu-ld}).
139 Specify the option @samp{--with-gnu-ld} if you plan to use the GNU
142 This option does not cause the GNU linker to be installed; it just
143 modifies the behavior of GNU CC to work with the GNU linker.
144 @c Specifically, it inhibits the installation of @code{collect2}, a program
145 @c which otherwise serves as a front-end for the system's linker on most
149 On MIPS based systems and on Alphas, you must specify whether you want
150 GNU CC to create the normal ECOFF debugging format, or to use BSD-style
151 stabs passed through the ECOFF symbol table. The normal ECOFF debug
152 format cannot fully handle languages other than C. BSD stabs format can
153 handle other languages, but it only works with the GNU debugger GDB.
155 Normally, GNU CC uses the ECOFF debugging format by default; if you
156 prefer BSD stabs, specify @samp{--with-stabs} when you configure GNU
159 No matter which default you choose when you configure GNU CC, the user
160 can use the @samp{-gcoff} and @samp{-gstabs+} options to specify explicitly
161 the debug format for a particular compilation.
163 @samp{--with-stabs} is meaningful on the ISC system on the 386, also, if
164 @samp{--with-gas} is used. It selects use of stabs debugging
165 information embedded in COFF output. This kind of debugging information
166 supports C++ well; ordinary COFF debugging information does not.
168 @samp{--with-stabs} is also meaningful on 386 systems running SVR4. It
169 selects use of stabs debugging information embedded in ELF output. The
170 C++ compiler currently (2.6.0) does not support the DWARF debugging
171 information normally used on 386 SVR4 platforms; stabs provide a
172 workable alternative. This requires gas and gdb, as the normal SVR4
173 tools can not generate or interpret stabs.
176 On certain systems, you must specify whether the machine has a floating
177 point unit. These systems include @samp{m68k-sun-sunos@var{n}} and
178 @samp{m68k-isi-bsd}. On any other system, @samp{--nfp} currently has no
179 effect, though perhaps there are other systems where it could usefully
182 @cindex Haifa scheduler
183 @cindex scheduler, experimental
185 @itemx --disable-haifa
186 Use @samp{--enable-haifa} to enable use of an experimental instruction
187 scheduler (from IBM Haifa). This may or may not produce better code.
188 Some targets on which it is known to be a win enable it by default; use
189 @samp{--disable-haifa} to disable it in these cases. @code{configure}
190 will print out whether the Haifa scheduler is enabled when it is run.
192 @cindex Objective C threads
193 @cindex threads, Objective C
194 @item --enable-threads=@var{type}
195 Certain systems, notably Linux-based GNU systems, can't be relied on to
196 supply a threads facility for the Objective C runtime and so will
197 default to single-threaded runtime. They may, however, have a library
198 threads implementation available, in which case threads can be enabled
199 with this option by supplying a suitable @var{type}, probably
200 @samp{posix}. The possibilities for @var{type} are @samp{single},
201 @samp{posix}, @samp{win32}, @samp{solaris}, @samp{irix} and @samp{mach}.
203 @cindex Internal Compiler Checking
204 @item --enable-checking
205 When you specify this option, the compiler is built to perform checking
206 of tree node types when referencing fields of that node. This does not
207 change the generated code, but adds error checking within the compiler.
208 This will slow down the compiler and may only work properly if you
209 are building the compiler with GNU C.
212 The @file{configure} script searches subdirectories of the source
213 directory for other compilers that are to be integrated into GNU CC.
214 The GNU compiler for C++, called G++ is in a subdirectory named
215 @file{cp}. @file{configure} inserts rules into @file{Makefile} to build
216 all of those compilers.
218 Here we spell out what files will be set up by @code{configure}. Normally
219 you need not be concerned with these files.
224 A file named @file{config.h} is created that contains a @samp{#include}
225 of the top-level config file for the machine you will run the compiler
226 on (@pxref{Config}). This file is responsible for defining information
227 about the host machine. It includes @file{tm.h}.
230 A file named @file{config.h} is created that contains a @samp{#include}
231 of the top-level config file for the machine you will run the compiler
232 on (@pxref{Config,,The Configuration File, gcc.info, Using and Porting
233 GCC}). This file is responsible for defining information about the host
234 machine. It includes @file{tm.h}.
237 The top-level config file is located in the subdirectory @file{config}.
238 Its name is always @file{xm-@var{something}.h}; usually
239 @file{xm-@var{machine}.h}, but there are some exceptions.
241 If your system does not support symbolic links, you might want to
242 set up @file{config.h} to contain a @samp{#include} command which
243 refers to the appropriate file.
246 A file named @file{tconfig.h} is created which includes the top-level config
247 file for your target machine. This is used for compiling certain
248 programs to run on that machine.
251 A file named @file{tm.h} is created which includes the
252 machine-description macro file for your target machine. It should be in
253 the subdirectory @file{config} and its name is often
254 @file{@var{machine}.h}.
257 The command file @file{configure} also constructs the file
258 @file{Makefile} by adding some text to the template file
259 @file{Makefile.in}. The additional text comes from files in the
260 @file{config} directory, named @file{t-@var{target}} and
261 @file{x-@var{host}}. If these files do not exist, it means nothing
262 needs to be added for a given target or host.
266 The standard directory for installing GNU CC is @file{/usr/local/lib}.
267 If you want to install its files somewhere else, specify
268 @samp{--prefix=@var{dir}} when you run @file{configure}. Here @var{dir}
269 is a directory name to use instead of @file{/usr/local} for all purposes
270 with one exception: the directory @file{/usr/local/include} is searched
271 for header files no matter where you install the compiler. To override
272 this name, use the @code{--with-local-prefix} option below. The directory
273 you specify need not exist, but its parent directory must exist.
276 Specify @samp{--with-local-prefix=@var{dir}} if you want the compiler to
277 search directory @file{@var{dir}/include} for locally installed header
278 files @emph{instead} of @file{/usr/local/include}.
280 You should specify @samp{--with-local-prefix} @strong{only} if your site has
281 a different convention (not @file{/usr/local}) for where to put
284 The default value for @samp{--with-local-prefix} is @file{/usr/local}
285 regardless of the value of @samp{--prefix}. Specifying @samp{--prefix}
286 has no effect on which directory GNU CC searches for local header files.
287 This may seem counterintuitive, but actually it is logical.
289 The purpose of @samp{--prefix} is to specify where to @emph{install GNU
290 CC}. The local header files in @file{/usr/local/include}---if you put
291 any in that directory---are not part of GNU CC. They are part of other
292 programs---perhaps many others. (GNU CC installs its own header files
293 in another directory which is based on the @samp{--prefix} value.)
295 @strong{Do not} specify @file{/usr} as the @samp{--with-local-prefix}! The
296 directory you use for @samp{--with-local-prefix} @strong{must not} contain
297 any of the system's standard header files. If it did contain them,
298 certain programs would be miscompiled (including GNU Emacs, on certain
299 targets), because this would override and nullify the header file
300 corrections made by the @code{fixincludes} script.
302 Indications are that people who use this option use it based on
303 mistaken ideas of what it is for. People use it as if it specified
304 where to install part of GNU CC. Perhaps they make this assumption
305 because installing GNU CC creates the directory.
307 @cindex Bison parser generator
308 @cindex parser generator, Bison
310 Make sure the Bison parser generator is installed. (This is
311 unnecessary if the Bison output files @file{c-parse.c} and
312 @file{cexp.c} are more recent than @file{c-parse.y} and @file{cexp.y}
313 and you do not plan to change the @samp{.y} files.)
315 Bison versions older than Sept 8, 1988 will produce incorrect output
316 for @file{c-parse.c}.
319 If you have chosen a configuration for GNU CC which requires other GNU
320 tools (such as GAS or the GNU linker) instead of the standard system
321 tools, install the required tools in the build directory under the names
322 @file{as}, @file{ld} or whatever is appropriate. This will enable the
323 compiler to find the proper tools for compilation of the program
326 Alternatively, you can do subsequent compilation using a value of the
327 @code{PATH} environment variable such that the necessary GNU tools come
328 before the standard system tools.
331 Build the compiler. Just type @samp{make LANGUAGES=c} in the compiler
334 @samp{LANGUAGES=c} specifies that only the C compiler should be
335 compiled. The makefile normally builds compilers for all the supported
336 languages; currently, C, C++ and Objective C. However, C is the only
337 language that is sure to work when you build with other non-GNU C
338 compilers. In addition, building anything but C at this stage is a
341 In general, you can specify the languages to build by typing the
342 argument @samp{LANGUAGES="@var{list}"}, where @var{list} is one or more
343 words from the list @samp{c}, @samp{c++}, and @samp{objective-c}. If
344 you have any additional GNU compilers as subdirectories of the GNU CC
345 source directory, you may also specify their names in this list.
347 Ignore any warnings you may see about ``statement not reached'' in
348 @file{insn-emit.c}; they are normal. Also, warnings about ``unknown
349 escape sequence'' are normal in @file{genopinit.c} and perhaps some
350 other files. Likewise, you should ignore warnings about ``constant is
351 so large that it is unsigned'' in @file{insn-emit.c} and
352 @file{insn-recog.c}, a warning about a comparison always being zero
353 in @file{enquire.o}, and warnings about shift counts exceeding type
354 widths in @file{cexp.y}. Any other compilation errors may represent bugs in
355 the port to your machine or operating system, and
357 should be investigated and reported (@pxref{Bugs}).
360 should be investigated and reported.
363 Some commercial compilers fail to compile GNU CC because they have bugs
364 or limitations. For example, the Microsoft compiler is said to run out
365 of macro space. Some Ultrix compilers run out of expression space; then
366 you need to break up the statement where the problem happens.
369 If you are building a cross-compiler, stop here. @xref{Cross-Compiler}.
373 Move the first-stage object files and executables into a subdirectory
380 The files are moved into a subdirectory named @file{stage1}.
381 Once installation is complete, you may wish to delete these files
382 with @code{rm -r stage1}.
385 If you have chosen a configuration for GNU CC which requires other GNU
386 tools (such as GAS or the GNU linker) instead of the standard system
387 tools, install the required tools in the @file{stage1} subdirectory
388 under the names @file{as}, @file{ld} or whatever is appropriate. This
389 will enable the stage 1 compiler to find the proper tools in the
392 Alternatively, you can do subsequent compilation using a value of the
393 @code{PATH} environment variable such that the necessary GNU tools come
394 before the standard system tools.
397 Recompile the compiler with itself, with this command:
400 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2"
403 This is called making the stage 2 compiler.
405 The command shown above builds compilers for all the supported
406 languages. If you don't want them all, you can specify the languages to
407 build by typing the argument @samp{LANGUAGES="@var{list}"}. @var{list}
408 should contain one or more words from the list @samp{c}, @samp{c++},
409 @samp{objective-c}, and @samp{proto}. Separate the words with spaces.
410 @samp{proto} stands for the programs @code{protoize} and
411 @code{unprotoize}; they are not a separate language, but you use
412 @code{LANGUAGES} to enable or disable their installation.
414 If you are going to build the stage 3 compiler, then you might want to
415 build only the C language in stage 2.
417 Once you have built the stage 2 compiler, if you are short of disk
418 space, you can delete the subdirectory @file{stage1}.
420 On a 68000 or 68020 system lacking floating point hardware,
421 unless you have selected a @file{tm.h} file that expects by default
422 that there is no such hardware, do this instead:
425 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2 -msoft-float"
429 If you wish to test the compiler by compiling it with itself one more
430 time, install any other necessary GNU tools (such as GAS or the GNU
431 linker) in the @file{stage2} subdirectory as you did in the
432 @file{stage1} subdirectory, then do this:
436 make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2"
440 This is called making the stage 3 compiler. Aside from the @samp{-B}
441 option, the compiler options should be the same as when you made the
442 stage 2 compiler. But the @code{LANGUAGES} option need not be the
443 same. The command shown above builds compilers for all the supported
444 languages; if you don't want them all, you can specify the languages to
445 build by typing the argument @samp{LANGUAGES="@var{list}"}, as described
448 If you do not have to install any additional GNU tools, you may use the
452 make bootstrap LANGUAGES=@var{language-list} BOOT_CFLAGS=@var{option-list}
456 instead of making @file{stage1}, @file{stage2}, and performing
457 the two compiler builds.
460 Then compare the latest object files with the stage 2 object
461 files---they ought to be identical, aside from time stamps (if any).
463 On some systems, meaningful comparison of object files is impossible;
464 they always appear ``different.'' This is currently true on Solaris and
465 some systems that use ELF object file format. On some versions of Irix
466 on SGI machines and DEC Unix (OSF/1) on Alpha systems, you will not be
467 able to compare the files without specifying @file{-save-temps}; see the
468 description of individual systems above to see if you get comparison
469 failures. You may have similar problems on other systems.
471 Use this command to compare the files:
477 This will mention any object files that differ between stage 2 and stage
478 3. Any difference, no matter how innocuous, indicates that the stage 2
479 compiler has compiled GNU CC incorrectly, and is therefore a potentially
481 serious bug which you should investigate and report (@pxref{Bugs}).
484 serious bug which you should investigate and report.
487 If your system does not put time stamps in the object files, then this
488 is a faster way to compare them (using the Bourne shell):
492 cmp $file stage2/$file
496 If you have built the compiler with the @samp{-mno-mips-tfile} option on
497 MIPS machines, you will not be able to compare the files.
500 Install the compiler driver, the compiler's passes and run-time support
501 with @samp{make install}. Use the same value for @code{CC},
502 @code{CFLAGS} and @code{LANGUAGES} that you used when compiling the
503 files that are being installed. One reason this is necessary is that
504 some versions of Make have bugs and recompile files gratuitously when
505 you do this step. If you use the same variable values, those files will
506 be recompiled properly.
508 For example, if you have built the stage 2 compiler, you can use the
512 make install CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O" LANGUAGES="@var{list}"
516 This copies the files @file{cc1}, @file{cpp} and @file{libgcc.a} to
517 files @file{cc1}, @file{cpp} and @file{libgcc.a} in the directory
518 @file{/usr/local/lib/gcc-lib/@var{target}/@var{version}}, which is where
519 the compiler driver program looks for them. Here @var{target} is the
520 canonicalized form of target machine type specified when you ran
521 @file{configure}, and @var{version} is the version number of GNU CC.
522 This naming scheme permits various versions and/or cross-compilers to
523 coexist. It also copies the executables for compilers for other
524 languages (e.g., @file{cc1plus} for C++) to the same directory.
526 This also copies the driver program @file{xgcc} into
527 @file{/usr/local/bin/gcc}, so that it appears in typical execution
528 search paths. It also copies @file{gcc.1} into
529 @file{/usr/local/man/man1} and info pages into @file{/usr/local/info}.
531 On some systems, this command causes recompilation of some files. This
532 is usually due to bugs in @code{make}. You should either ignore this
533 problem, or use GNU Make.
535 @cindex @code{alloca} and SunOS
536 @strong{Warning: there is a bug in @code{alloca} in the Sun library. To
537 avoid this bug, be sure to install the executables of GNU CC that were
538 compiled by GNU CC. (That is, the executables from stage 2 or 3, not
539 stage 1.) They use @code{alloca} as a built-in function and never the
542 (It is usually better to install GNU CC executables from stage 2 or 3,
543 since they usually run faster than the ones compiled with some other
547 @cindex C++ runtime library
548 @cindex @code{libstdc++}
549 If you're going to use C++, it's likely that you need to also install
550 a C++ runtime library. Just as GNU C does not
551 distribute a C runtime library, it also does not include a C++ runtime
552 library. All I/O functionality, special class libraries, etc., are
553 provided by the C++ runtime library.
555 The standard C++ runtime library for GNU CC is called @samp{libstdc++}.
556 An obsolescent library @samp{libg++} may also be available, but it's
557 necessary only for older software that hasn't been converted yet; if
558 you don't know whether you need @samp{libg++} then you probably don't
561 Here's one way to build and install @samp{libstdc++} for GNU CC:
565 Build and install GNU CC, so that invoking @samp{gcc} obtains the GNU CC
569 Obtain a copy of a compatible @samp{libstdc++} distribution. For
570 example, the @samp{libstdc++-2.8.0.tar.gz} distribution should be
571 compatible with GCC 2.8.0. GCC distributors normally distribute
572 @samp{libstdc++} as well.
575 Set the @samp{CXX} environment variable to @samp{gcc} while running the
576 @samp{libstdc++} distribution's @file{configure} command. Use the same
577 @file{configure} options that you used when you invoked GCC's
578 @file{configure} command.
581 Invoke @samp{make} to build the C++ runtime.
584 Invoke @samp{make install} to install the C++ runtime.
588 To summarize, after building and installing GNU CC, invoke the following
589 shell commands in the topmost directory of the C++ library distribution.
590 For @var{configure-options}, use the same options that
591 you used to configure GNU CC.
594 $ CXX=gcc ./configure @var{configure-options}
600 GNU CC includes a runtime library for Objective-C because it is an
601 integral part of the language. You can find the files associated with
602 the library in the subdirectory @file{objc}. The GNU Objective-C
603 Runtime Library requires header files for the target's C library in
604 order to be compiled,and also requires the header files for the target's
605 thread library if you want thread support. @xref{Cross Headers,
606 Cross-Compilers and Header Files, Cross-Compilers and Header Files}, for
607 discussion about header files issues for cross-compilation.
609 When you run @file{configure}, it picks the appropriate Objective-C
610 thread implementation file for the target platform. In some situations,
611 you may wish to choose a different back-end as some platforms support
612 multiple thread implementations or you may wish to disable thread
613 support completely. You do this by specifying a value for the
614 @var{OBJC_THREAD_FILE} makefile variable on the command line when you
615 run make, for example:
618 make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2" OBJC_THREAD_FILE=thr-single
622 Below is a list of the currently available back-ends.
626 Disable thread support, should work for all platforms.
628 DEC OSF/1 thread support.
630 SGI IRIX thread support.
632 Generic MACH thread support, known to work on NEXTSTEP.
634 IBM OS/2 thread support.
636 Generix POSIX thread support.
638 PCThreads on Linux-based GNU systems.
640 SUN Solaris thread support.
642 Microsoft Win32 API thread support.
647 @section Configurations Supported by GNU CC
648 @cindex configurations supported by GNU CC
650 Here are the possible CPU types:
653 @c gmicro, alliant, spur and tahoe omitted since they don't work.
654 1750a, a29k, alpha, arm, c@var{n}, clipper, dsp16xx, elxsi, h8300,
655 hppa1.0, hppa1.1, i370, i386, i486, i586, i860, i960, m32r, m68000, m68k,
656 m88k, mips, mipsel, mips64, mips64el, ns32k, powerpc, powerpcle,
657 pyramid, romp, rs6000, sh, sparc, sparclite, sparc64, vax, we32k.
660 Here are the recognized company names. As you can see, customary
661 abbreviations are used rather than the longer official names.
663 @c What should be done about merlin, tek*, dolphin?
665 acorn, alliant, altos, apollo, apple, att, bull,
666 cbm, convergent, convex, crds, dec, dg, dolphin,
667 elxsi, encore, harris, hitachi, hp, ibm, intergraph, isi,
668 mips, motorola, ncr, next, ns, omron, plexus,
669 sequent, sgi, sony, sun, tti, unicom, wrs.
672 The company name is meaningful only to disambiguate when the rest of
673 the information supplied is insufficient. You can omit it, writing
674 just @samp{@var{cpu}-@var{system}}, if it is not needed. For example,
675 @samp{vax-ultrix4.2} is equivalent to @samp{vax-dec-ultrix4.2}.
677 Here is a list of system types:
680 386bsd, aix, acis, amigaos, aos, aout, aux, bosx, bsd, clix, coff, ctix, cxux,
681 dgux, dynix, ebmon, ecoff, elf, esix, freebsd, hms, genix, gnu, linux-gnu,
682 hiux, hpux, iris, irix, isc, luna, lynxos, mach, minix, msdos, mvs,
683 netbsd, newsos, nindy, ns, osf, osfrose, ptx, riscix, riscos, rtu, sco, sim,
684 solaris, sunos, sym, sysv, udi, ultrix, unicos, uniplus, unos, vms, vsta,
685 vxworks, winnt, xenix.
689 You can omit the system type; then @file{configure} guesses the
690 operating system from the CPU and company.
692 You can add a version number to the system type; this may or may not
693 make a difference. For example, you can write @samp{bsd4.3} or
694 @samp{bsd4.4} to distinguish versions of BSD. In practice, the version
695 number is most needed for @samp{sysv3} and @samp{sysv4}, which are often
698 If you specify an impossible combination such as @samp{i860-dg-vms},
699 then you may get an error message from @file{configure}, or it may
700 ignore part of the information and do the best it can with the rest.
701 @file{configure} always prints the canonical name for the alternative
702 that it used. GNU CC does not support all possible alternatives.
704 Often a particular model of machine has a name. Many machine names are
705 recognized as aliases for CPU/company combinations. Thus, the machine
706 name @samp{sun3}, mentioned above, is an alias for @samp{m68k-sun}.
707 Sometimes we accept a company name as a machine name, when the name is
708 popularly used for a particular machine. Here is a table of the known
712 3300, 3b1, 3b@var{n}, 7300, altos3068, altos,
713 apollo68, att-7300, balance,
714 convex-c@var{n}, crds, decstation-3100,
715 decstation, delta, encore,
716 fx2800, gmicro, hp7@var{nn}, hp8@var{nn},
717 hp9k2@var{nn}, hp9k3@var{nn}, hp9k7@var{nn},
718 hp9k8@var{nn}, iris4d, iris, isi68,
719 m3230, magnum, merlin, miniframe,
720 mmax, news-3600, news800, news, next,
721 pbd, pc532, pmax, powerpc, powerpcle, ps2, risc-news,
722 rtpc, sun2, sun386i, sun386, sun3,
723 sun4, symmetry, tower-32, tower.
727 Remember that a machine name specifies both the cpu type and the company
729 If you want to install your own homemade configuration files, you can
730 use @samp{local} as the company name to access them. If you use
731 configuration @samp{@var{cpu}-local}, the configuration name
732 without the cpu prefix
733 is used to form the configuration file names.
735 Thus, if you specify @samp{m68k-local}, configuration uses
736 files @file{m68k.md}, @file{local.h}, @file{m68k.c},
737 @file{xm-local.h}, @file{t-local}, and @file{x-local}, all in the
738 directory @file{config/m68k}.
740 Here is a list of configurations that have special treatment or special
741 things you must know:
745 MIL-STD-1750A processors.
747 The MIL-STD-1750A cross configuration produces output for
748 @code{as1750}, an assembler/linker available under the GNU Public
749 License for the 1750A. @code{as1750} can be obtained at
750 @emph{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}.
751 A similarly licensed simulator for
752 the 1750A is available from same address.
754 You should ignore a fatal error during the building of libgcc (libgcc is
755 not yet implemented for the 1750A.)
757 The @code{as1750} assembler requires the file @file{ms1750.inc}, which is
758 found in the directory @file{config/1750a}.
760 GNU CC produced the same sections as the Fairchild F9450 C Compiler,
765 The program code section.
768 The read/write (RAM) data section.
771 The read-only (ROM) constants section.
774 Initialization section (code to copy KREL to SREL).
777 The smallest addressable unit is 16 bits (BITS_PER_UNIT is 16). This
778 means that type `char' is represented with a 16-bit word per character.
779 The 1750A's "Load/Store Upper/Lower Byte" instructions are not used by
783 Systems using processors that implement the DEC Alpha architecture and
784 are running the DEC Unix (OSF/1) operating system, for example the DEC
785 Alpha AXP systems.CC.)
787 GNU CC writes a @samp{.verstamp} directive to the assembler output file
788 unless it is built as a cross-compiler. It gets the version to use from
789 the system header file @file{/usr/include/stamp.h}. If you install a
790 new version of DEC Unix, you should rebuild GCC to pick up the new version
793 Note that since the Alpha is a 64-bit architecture, cross-compilers from
794 32-bit machines will not generate code as efficient as that generated
795 when the compiler is running on a 64-bit machine because many
796 optimizations that depend on being able to represent a word on the
797 target in an integral value on the host cannot be performed. Building
798 cross-compilers on the Alpha for 32-bit machines has only been tested in
799 a few cases and may not work properly.
801 @code{make compare} may fail on old versions of DEC Unix unless you add
802 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
803 assembler input file is stored in the object file, and that makes
804 comparison fail if it differs between the @code{stage1} and
805 @code{stage2} compilations. The option @samp{-save-temps} forces a
806 fixed name to be used for the assembler input file, instead of a
807 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
808 unless the comparisons fail without that option. If you add
809 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
810 @samp{.s} files after each series of compilations.
812 GNU CC now supports both the native (ECOFF) debugging format used by DBX
813 and GDB and an encapsulated STABS format for use only with GDB. See the
814 discussion of the @samp{--with-stabs} option of @file{configure} above
815 for more information on these formats and how to select them.
817 There is a bug in DEC's assembler that produces incorrect line numbers
818 for ECOFF format when the @samp{.align} directive is used. To work
819 around this problem, GNU CC will not emit such alignment directives
820 while writing ECOFF format debugging information even if optimization is
821 being performed. Unfortunately, this has the very undesirable
822 side-effect that code addresses when @samp{-O} is specified are
823 different depending on whether or not @samp{-g} is also specified.
825 To avoid this behavior, specify @samp{-gstabs+} and use GDB instead of
826 DBX. DEC is now aware of this problem with the assembler and hopes to
827 provide a fix shortly.
830 Argonaut ARC processor.
831 This configuration is intended for embedded systems.
834 Advanced RISC Machines ARM-family processors. These are often used in
835 embedded applications. There are no standard Unix configurations.
836 This configuration corresponds to the basic instruction sequences and will
837 produce @file{a.out} format object modules.
839 You may need to make a variant of the file @file{arm.h} for your particular
842 @item arm-*-linuxaout
843 Any of the ARM family processors running the Linux-based GNU system with
844 the @file{a.out} binary format (ELF is not yet supported). You must use
845 version 2.8.1.0.7 or later of the GNU/Linux binutils, which you can download
846 from @file{sunsite.unc.edu:/pub/Linux/GCC} and other mirror sites for
847 Linux-based GNU systems.
850 The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix.
851 If you are running a version of RISC iX prior to 1.2 then you must
852 specify the version number during configuration. Note that the
853 assembler shipped with RISC iX does not support stabs debugging
854 information; a new version of the assembler, with stabs support
855 included, is now available from Acorn and via ftp
856 @file{ftp.acorn.com:/pub/riscix/as+xterm.tar.Z}. To enable stabs
857 debugging, pass @samp{--with-gnu-as} to configure.
859 You will need to install GNU @file{sed} before you can run configure.
862 AMD Am29k-family processors. These are normally used in embedded
863 applications. There are no standard Unix configurations.
865 corresponds to AMD's standard calling sequence and binary interface
866 and is compatible with other 29k tools.
868 You may need to make a variant of the file @file{a29k.h} for your
869 particular configuration.
872 AMD Am29050 used in a system running a variant of BSD Unix.
875 MIPS-based DECstations can support three different personalities:
876 Ultrix, DEC OSF/1, and OSF/rose. (Alpha-based DECstation products have
877 a configuration name beginning with @samp{alpha-dec}.) To configure GCC
878 for these platforms use the following configurations:
881 @item decstation-ultrix
882 Ultrix configuration.
884 @item decstation-osf1
885 Dec's version of OSF/1.
887 @item decstation-osfrose
888 Open Software Foundation reference port of OSF/1 which uses the
889 OSF/rose object file format instead of ECOFF. Normally, you
890 would not select this configuration.
893 The MIPS C compiler needs to be told to increase its table size
894 for switch statements with the @samp{-Wf,-XNg1500} option in
895 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
896 optimization option, you also need to use @samp{-Olimit 3000}.
897 Both of these options are automatically generated in the
898 @file{Makefile} that the shell script @file{configure} builds.
899 If you override the @code{CC} make variable and use the MIPS
900 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
902 @item elxsi-elxsi-bsd
903 The Elxsi's C compiler has known limitations that prevent it from
904 compiling GNU C. Please contact @code{mrs@@cygnus.com} for more details.
907 A port to the AT&T DSP1610 family of processors.
911 Alliant FX/8 computer. Note that the standard installed C compiler in
912 Concentrix 5.0 has a bug which prevent it from compiling GNU CC
913 correctly. You can patch the compiler bug as follows:
917 adb -w ./pcc - << EOF
922 Then you must use the @samp{-ip12} option when compiling GNU CC
923 with the patched compiler, as shown here:
926 make CC="./pcc -ip12" CFLAGS=-w
929 Note also that Alliant's version of DBX does not manage to work with the
934 Hitachi H8/300 series of processors.
936 The calling convention and structure layout has changed in release 2.6.
937 All code must be recompiled. The calling convention now passes the
938 first three arguments in function calls in registers. Structures are no
939 longer a multiple of 2 bytes.
942 There are several variants of the HP-PA processor which run a variety
943 of operating systems. GNU CC must be configured to use the correct
944 processor type and operating system, or GNU CC will not function correctly.
945 The easiest way to handle this problem is to @emph{not} specify a target
946 when configuring GNU CC, the @file{configure} script will try to automatically
947 determine the right processor type and operating system.
949 @samp{-g} does not work on HP-UX, since that system uses a peculiar
950 debugging format which GNU CC does not know about. However, @samp{-g}
951 will work if you also use GAS and GDB in conjunction with GCC. We
952 highly recommend using GAS for all HP-PA configurations.
954 You should be using GAS-2.6 (or later) along with GDB-4.16 (or later). These
955 can be retrieved from all the traditional GNU ftp archive sites.
957 On some versions of HP-UX, you will need to install GNU @file{sed}.
959 You will need to be install GAS into a directory before @code{/bin},
960 @code{/usr/bin}, and @code{/usr/ccs/bin} in your search path. You
961 should install GAS before you build GNU CC.
963 To enable debugging, you must configure GNU CC with the @samp{--with-gnu-as}
964 option before building.
967 This port is very preliminary and has many known bugs. We hope to
968 have a higher-quality port for this machine soon.
970 @item i386-*-linux-gnuoldld
971 Use this configuration to generate @file{a.out} binaries on Linux-based
972 GNU systems if you do not have gas/binutils version 2.5.2 or later
973 installed. This is an obsolete configuration.
975 @item i386-*-linux-gnuaout
976 Use this configuration to generate @file{a.out} binaries on Linux-based
977 GNU systems. This configuration is being superseded. You must use
978 gas/binutils version 2.5.2 or later.
980 @item i386-*-linux-gnu
981 Use this configuration to generate ELF binaries on Linux-based GNU
982 systems. You must use gas/binutils version 2.5.2 or later.
985 Compilation with RCC is recommended. Also, it may be a good idea to
986 link with GNU malloc instead of the malloc that comes with the system.
988 @item i386-*-sco3.2v4
989 Use this configuration for SCO release 3.2 version 4.
991 @item i386-*-sco3.2v5*
992 Use this for the SCO OpenServer Release family including 5.0.0, 5.0.2,
993 5.0.4, 5.0.5, Internet FastStart 1.0, and Internet FastStart 1.1.
995 GNU CC can generate either ELF or COFF binaries. ELF is the default.
996 To get COFF output, you must specify @samp{-mcoff}) on the command line.
998 For 5.0.0 and 5.0.2, you must install TLS597 from ftp.sco.com/TLS.
999 5.0.4 and later do not require this patch.
1001 The native SCO assembler that is provided with the OS at no charge
1002 is normally required. If, however, you must be able to use the GNU
1003 assembler (perhaps you have complex asms) you must configure this
1004 package @samp{--with-gnu-as}. To do this, install (cp or symlink)
1005 gcc/as to your copy of the GNU assembler. You must use a recent version
1006 of GNU binutils; version 2.9.1 seems to work well. If you select this
1007 option, you will be unable to build COFF images. Trying to do so will
1008 result in non-obvious failures. In general, the "--with-gnu-as" option
1009 isn't as well tested as the native assembler.
1011 @emph{NOTE:} You must follow the instructions about invoking
1012 @samp{make bootstrap} because the native OpenServer compiler builds
1013 a @file{cc1plus} that will not correctly parse many valid C++ programs.
1014 You must do a @samp{make bootstrap} if you are building with the native
1018 It may be a good idea to link with GNU malloc instead of the malloc that
1019 comes with the system.
1021 In ISC version 4.1, @file{sed} core dumps when building
1022 @file{deduced.h}. Use the version of @file{sed} from version 4.0.
1025 It may be good idea to link with GNU malloc instead of the malloc that
1026 comes with the system.
1029 You need to use GAS version 2.1 or later, and LD from
1030 GNU binutils version 2.2 or later.
1032 @item i386-sequent-bsd
1033 Go to the Berkeley universe before compiling.
1035 @item i386-sequent-ptx1*
1036 @itemx i386-sequent-ptx2*
1037 You must install GNU @file{sed} before running @file{configure}.
1039 @item i386-sun-sunos4
1040 You may find that you need another version of GNU CC to begin
1041 bootstrapping with, since the current version when built with the
1042 system's own compiler seems to get an infinite loop compiling part of
1043 @file{libgcc2.c}. GNU CC version 2 compiled with GNU CC (any version)
1044 seems not to have this problem.
1046 See @ref{Sun Install}, for information on installing GNU CC on Sun
1049 @item i[345]86-*-winnt3.5
1050 This version requires a GAS that has not yet been released. Until it
1051 is, you can get a prebuilt binary version via anonymous ftp from
1052 @file{cs.washington.edu:pub/gnat} or @file{cs.nyu.edu:pub/gnat}. You
1053 must also use the Microsoft header files from the Windows NT 3.5 SDK.
1054 Find these on the CDROM in the @file{/mstools/h} directory dated 9/4/94. You
1055 must use a fixed version of Microsoft linker made especially for NT 3.5,
1056 which is also is available on the NT 3.5 SDK CDROM. If you do not have
1057 this linker, can you also use the linker from Visual C/C++ 1.0 or 2.0.
1059 Installing GNU CC for NT builds a wrapper linker, called @file{ld.exe},
1060 which mimics the behaviour of Unix @file{ld} in the specification of
1061 libraries (@samp{-L} and @samp{-l}). @file{ld.exe} looks for both Unix
1062 and Microsoft named libraries. For example, if you specify
1063 @samp{-lfoo}, @file{ld.exe} will look first for @file{libfoo.a}
1064 and then for @file{foo.lib}.
1066 You may install GNU CC for Windows NT in one of two ways, depending on
1067 whether or not you have a Unix-like shell and various Unix-like
1072 If you do not have a Unix-like shell and few Unix-like utilities, you
1073 will use a DOS style batch script called @file{configure.bat}. Invoke
1074 it as @code{configure winnt} from an MSDOS console window or from the
1075 program manager dialog box. @file{configure.bat} assumes you have
1076 already installed and have in your path a Unix-like @file{sed} program
1077 which is used to create a working @file{Makefile} from @file{Makefile.in}.
1079 @file{Makefile} uses the Microsoft Nmake program maintenance utility and
1080 the Visual C/C++ V8.00 compiler to build GNU CC. You need only have the
1081 utilities @file{sed} and @file{touch} to use this installation method,
1082 which only automatically builds the compiler itself. You must then
1083 examine what @file{fixinc.winnt} does, edit the header files by hand and
1084 build @file{libgcc.a} manually.
1087 The second type of installation assumes you are running a Unix-like
1088 shell, have a complete suite of Unix-like utilities in your path, and
1089 have a previous version of GNU CC already installed, either through
1090 building it via the above installation method or acquiring a pre-built
1091 binary. In this case, use the @file{configure} script in the normal
1095 @item i860-intel-osf1
1096 This is the Paragon.
1098 If you have version 1.0 of the operating system, you need to take
1099 special steps to build GNU CC due to peculiarities of the system. Newer
1100 system versions have no problem. See the section `Installation Problems'
1101 in the GNU CC Manual.
1103 @ifclear INSTALLONLY
1104 If you have version 1.0 of the operating system,
1105 see @ref{Installation Problems}, for special things you need to do to
1106 compensate for peculiarities in the system.
1110 LynxOS 2.2 and earlier comes with GNU CC 1.x already installed as
1111 @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}.
1112 You can tell GNU CC to use the GNU assembler and linker, by specifying
1113 @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce
1114 COFF format object files and executables; otherwise GNU CC will use the
1115 installed tools, which produce @file{a.out} format executables.
1118 Mitsubishi M32R processor.
1119 This configuration is intended for embedded systems.
1122 HP 9000 series 200 running BSD. Note that the C compiler that comes
1123 with this system cannot compile GNU CC; contact @code{law@@cygnus.com}
1124 to get binaries of GNU CC for bootstrapping.
1127 Altos 3068. You must use the GNU assembler, linker and debugger.
1128 Also, you must fix a kernel bug. Details in the file @file{README.ALTOS}.
1130 @item m68k-apple-aux
1131 Apple Macintosh running A/UX.
1132 You may configure GCC to use either the system assembler and
1133 linker or the GNU assembler and linker. You should use the GNU configuration
1134 if you can, especially if you also want to use GNU C++. You enabled
1135 that configuration with + the @samp{--with-gnu-as} and @samp{--with-gnu-ld}
1136 options to @code{configure}.
1138 Note the C compiler that comes
1139 with this system cannot compile GNU CC. You can find binaries of GNU CC
1140 for bootstrapping on @code{jagubox.gsfc.nasa.gov}.
1141 You will also a patched version of @file{/bin/ld} there that
1142 raises some of the arbitrary limits found in the original.
1145 AT&T 3b1, a.k.a. 7300 PC. Special procedures are needed to compile GNU
1146 CC with this machine's standard C compiler, due to bugs in that
1147 compiler. You can bootstrap it more easily with
1148 previous versions of GNU CC if you have them.
1150 Installing GNU CC on the 3b1 is difficult if you do not already have
1151 GNU CC running, due to bugs in the installed C compiler. However,
1152 the following procedure might work. We are unable to test it.
1156 Comment out the @samp{#include "config.h"} line near the start of
1157 @file{cccp.c} and do @samp{make cpp}. This makes a preliminary version
1161 Save the old @file{/lib/cpp} and copy the preliminary GNU cpp to that
1165 Undo your change in @file{cccp.c}, or reinstall the original version,
1166 and do @samp{make cpp} again.
1169 Copy this final version of GNU cpp into @file{/lib/cpp}.
1171 @findex obstack_free
1173 Replace every occurrence of @code{obstack_free} in the file
1174 @file{tree.c} with @code{_obstack_free}.
1177 Run @code{make} to get the first-stage GNU CC.
1180 Reinstall the original version of @file{/lib/cpp}.
1183 Now you can compile GNU CC with itself and install it in the normal
1187 @item m68k-bull-sysv
1188 Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GNU CC works
1189 either with native assembler or GNU assembler. You can use
1190 GNU assembler with native coff generation by providing @samp{--with-gnu-as} to
1191 the configure script or use GNU assembler with dbx-in-coff encapsulation
1192 by providing @samp{--with-gnu-as --stabs}. For any problem with native
1193 assembler or for availability of the DPX/2 port of GAS, contact
1194 @code{F.Pierresteguy@@frcl.bull.fr}.
1196 @item m68k-crds-unox
1197 Use @samp{configure unos} for building on Unos.
1199 The Unos assembler is named @code{casm} instead of @code{as}. For some
1200 strange reason linking @file{/bin/as} to @file{/bin/casm} changes the
1201 behavior, and does not work. So, when installing GNU CC, you should
1202 install the following script as @file{as} in the subdirectory where
1203 the passes of GCC are installed:
1210 The default Unos library is named @file{libunos.a} instead of
1211 @file{libc.a}. To allow GNU CC to function, either change all
1212 references to @samp{-lc} in @file{gcc.c} to @samp{-lunos} or link
1213 @file{/lib/libc.a} to @file{/lib/libunos.a}.
1215 @cindex @code{alloca}, for Unos
1216 When compiling GNU CC with the standard compiler, to overcome bugs in
1217 the support of @code{alloca}, do not use @samp{-O} when making stage 2.
1218 Then use the stage 2 compiler with @samp{-O} to make the stage 3
1219 compiler. This compiler will have the same characteristics as the usual
1220 stage 2 compiler on other systems. Use it to make a stage 4 compiler
1221 and compare that with stage 3 to verify proper compilation.
1223 (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in
1224 the comments there will make the above paragraph superfluous. Please
1225 inform us of whether this works.)
1227 Unos uses memory segmentation instead of demand paging, so you will need
1228 a lot of memory. 5 Mb is barely enough if no other tasks are running.
1229 If linking @file{cc1} fails, try putting the object files into a library
1230 and linking from that library.
1233 HP 9000 series 300 or 400 running HP-UX. HP-UX version 8.0 has a bug in
1234 the assembler that prevents compilation of GNU CC. To fix it, get patch
1237 In addition, if you wish to use gas @samp{--with-gnu-as} you must use
1238 gas version 2.1 or later, and you must use the GNU linker version 2.1 or
1239 later. Earlier versions of gas relied upon a program which converted the
1240 gas output into the native HP-UX format, but that program has not been
1241 kept up to date. gdb does not understand that native HP-UX format, so
1242 you must use gas if you wish to use gdb.
1245 Sun 3. We do not provide a configuration file to use the Sun FPA by
1246 default, because programs that establish signal handlers for floating
1247 point traps inherently cannot work with the FPA.
1249 See @ref{Sun Install}, for information on installing GNU CC on Sun
1253 Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port.
1254 These systems tend to use the Green Hills C, revision 1.8.5, as the
1255 standard C compiler. There are apparently bugs in this compiler that
1256 result in object files differences between stage 2 and stage 3. If this
1257 happens, make the stage 4 compiler and compare it to the stage 3
1258 compiler. If the stage 3 and stage 4 object files are identical, this
1259 suggests you encountered a problem with the standard C compiler; the
1260 stage 3 and 4 compilers may be usable.
1262 It is best, however, to use an older version of GNU CC for bootstrapping
1266 Motorola m88k running DG/UX. To build 88open BCS native or cross
1267 compilers on DG/UX, specify the configuration name as
1268 @samp{m88k-*-dguxbcs} and build in the 88open BCS software development
1269 environment. To build ELF native or cross compilers on DG/UX, specify
1270 @samp{m88k-*-dgux} and build in the DG/UX ELF development environment.
1271 You set the software development environment by issuing
1272 @samp{sde-target} command and specifying either @samp{m88kbcs} or
1273 @samp{m88kdguxelf} as the operand.
1275 If you do not specify a configuration name, @file{configure} guesses the
1276 configuration based on the current software development environment.
1278 @item m88k-tektronix-sysv3
1279 Tektronix XD88 running UTekV 3.2e. Do not turn on
1280 optimization while building stage1 if you bootstrap with
1281 the buggy Green Hills compiler. Also, The bundled LAI
1282 System V NFS is buggy so if you build in an NFS mounted
1283 directory, start from a fresh reboot, or avoid NFS all together.
1284 Otherwise you may have trouble getting clean comparisons
1288 MIPS machines running the MIPS operating system in BSD mode. It's
1289 possible that some old versions of the system lack the functions
1290 @code{memcpy}, @code{memcmp}, and @code{memset}. If your system lacks
1291 these, you must remove or undo the definition of
1292 @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}.
1294 The MIPS C compiler needs to be told to increase its table size
1295 for switch statements with the @samp{-Wf,-XNg1500} option in
1296 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1297 optimization option, you also need to use @samp{-Olimit 3000}.
1298 Both of these options are automatically generated in the
1299 @file{Makefile} that the shell script @file{configure} builds.
1300 If you override the @code{CC} make variable and use the MIPS
1301 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1303 @item mips-mips-riscos*
1304 The MIPS C compiler needs to be told to increase its table size
1305 for switch statements with the @samp{-Wf,-XNg1500} option in
1306 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1307 optimization option, you also need to use @samp{-Olimit 3000}.
1308 Both of these options are automatically generated in the
1309 @file{Makefile} that the shell script @file{configure} builds.
1310 If you override the @code{CC} make variable and use the MIPS
1311 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1313 MIPS computers running RISC-OS can support four different
1314 personalities: default, BSD 4.3, System V.3, and System V.4
1315 (older versions of RISC-OS don't support V.4). To configure GCC
1316 for these platforms use the following configurations:
1319 @item mips-mips-riscos@code{rev}
1320 Default configuration for RISC-OS, revision @code{rev}.
1322 @item mips-mips-riscos@code{rev}bsd
1323 BSD 4.3 configuration for RISC-OS, revision @code{rev}.
1325 @item mips-mips-riscos@code{rev}sysv4
1326 System V.4 configuration for RISC-OS, revision @code{rev}.
1328 @item mips-mips-riscos@code{rev}sysv
1329 System V.3 configuration for RISC-OS, revision @code{rev}.
1332 The revision @code{rev} mentioned above is the revision of
1333 RISC-OS to use. You must reconfigure GCC when going from a
1334 RISC-OS revision 4 to RISC-OS revision 5. This has the effect of
1336 @ifclear INSTALLONLY
1337 bug (see @ref{Installation Problems}, for more details).
1344 In order to compile GCC on an SGI running IRIX 4, the "c.hdr.lib"
1345 option must be installed from the CD-ROM supplied from Silicon Graphics.
1346 This is found on the 2nd CD in release 4.0.1.
1348 In order to compile GCC on an SGI running IRIX 5, the "compiler_dev.hdr"
1349 subsystem must be installed from the IDO CD-ROM supplied by Silicon
1352 @code{make compare} may fail on version 5 of IRIX unless you add
1353 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
1354 assembler input file is stored in the object file, and that makes
1355 comparison fail if it differs between the @code{stage1} and
1356 @code{stage2} compilations. The option @samp{-save-temps} forces a
1357 fixed name to be used for the assembler input file, instead of a
1358 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
1359 unless the comparisons fail without that option. If you do you
1360 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
1361 @samp{.s} files after each series of compilations.
1363 The MIPS C compiler needs to be told to increase its table size
1364 for switch statements with the @samp{-Wf,-XNg1500} option in
1365 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1366 optimization option, you also need to use @samp{-Olimit 3000}.
1367 Both of these options are automatically generated in the
1368 @file{Makefile} that the shell script @file{configure} builds.
1369 If you override the @code{CC} make variable and use the MIPS
1370 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1372 On Irix version 4.0.5F, and perhaps on some other versions as well,
1373 there is an assembler bug that reorders instructions incorrectly. To
1374 work around it, specify the target configuration
1375 @samp{mips-sgi-irix4loser}. This configuration inhibits assembler
1378 In a compiler configured with target @samp{mips-sgi-irix4}, you can turn
1379 off assembler optimization by using the @samp{-noasmopt} option. This
1380 compiler option passes the option @samp{-O0} to the assembler, to
1383 The @samp{-noasmopt} option can be useful for testing whether a problem
1384 is due to erroneous assembler reordering. Even if a problem does not go
1385 away with @samp{-noasmopt}, it may still be due to assembler
1386 reordering---perhaps GNU CC itself was miscompiled as a result.
1388 To enable debugging under Irix 5, you must use GNU as 2.5 or later,
1389 and use the @samp{--with-gnu-as} configure option when configuring gcc.
1390 GNU as is distributed as part of the binutils package.
1392 @item mips-sony-sysv
1393 Sony MIPS NEWS. This works in NEWSOS 5.0.1, but not in 5.0.2 (which
1394 uses ELF instead of COFF). Support for 5.0.2 will probably be provided
1395 soon by volunteers. In particular, the linker does not like the
1396 code generated by GCC when shared libraries are linked in.
1399 Encore ns32000 system. Encore systems are supported only under BSD.
1402 National Semiconductor ns32000 system. Genix has bugs in @code{alloca}
1403 and @code{malloc}; you must get the compiled versions of these from GNU
1407 Go to the Berkeley universe before compiling.
1410 UTEK ns32000 system (``merlin''). The C compiler that comes with this
1411 system cannot compile GNU CC; contact @samp{tektronix!reed!mason} to get
1412 binaries of GNU CC for bootstrapping.
1416 The only operating systems supported for the IBM RT PC are AOS and
1417 MACH. GNU CC does not support AIX running on the RT. We recommend you
1418 compile GNU CC with an earlier version of itself; if you compile GNU CC
1419 with @code{hc}, the Metaware compiler, it will work, but you will get
1420 mismatches between the stage 2 and stage 3 compilers in various files.
1421 These errors are minor differences in some floating-point constants and
1422 can be safely ignored; the stage 3 compiler is correct.
1425 @itemx powerpc-*-aix
1426 Various early versions of each release of the IBM XLC compiler will not
1427 bootstrap GNU CC. Symptoms include differences between the stage2 and
1428 stage3 object files, and errors when compiling @file{libgcc.a} or
1429 @file{enquire}. Known problematic releases include: xlc-1.2.1.8,
1430 xlc-1.3.0.0 (distributed with AIX 3.2.5), and xlc-1.3.0.19. Both
1431 xlc-1.2.1.28 and xlc-1.3.0.24 (PTF 432238) are known to produce working
1432 versions of GNU CC, but most other recent releases correctly bootstrap
1435 Release 4.3.0 of AIX and ones prior to AIX 3.2.4 include a version of
1436 the IBM assembler which does not accept debugging directives: assembler
1437 updates are available as PTFs. Also, if you are using AIX 3.2.5 or
1438 greater and the GNU assembler, you must have a version modified after
1439 October 16th, 1995 in order for the GNU C compiler to build. See the
1440 file @file{README.RS6000} for more details on any of these problems.
1442 GNU CC does not yet support the 64-bit PowerPC instructions.
1444 Objective C does not work on this architecture because it makes assumptions
1445 that are incompatible with the calling conventions.
1447 AIX on the RS/6000 provides support (NLS) for environments outside of
1448 the United States. Compilers and assemblers use NLS to support
1449 locale-specific representations of various objects including
1450 floating-point numbers ("." vs "," for separating decimal fractions).
1451 There have been problems reported where the library linked with GNU CC
1452 does not produce the same floating-point formats that the assembler
1453 accepts. If you have this problem, set the LANG environment variable to
1456 Due to changes in the way that GNU CC invokes the binder (linker) for AIX
1457 4.1, you may now receive warnings of duplicate symbols from the link step
1458 that were not reported before. The assembly files generated by GNU CC for
1459 AIX have always included multiple symbol definitions for certain global
1460 variable and function declarations in the original program. The warnings
1461 should not prevent the linker from producing a correct library or runnable
1464 By default, AIX 4.1 produces code that can be used on either Power or
1467 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1468 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1471 @itemx powerpc-*-sysv4
1472 PowerPC system in big endian mode, running System V.4.
1474 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1475 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1477 @item powerpc-*-linux-gnu
1478 PowerPC system in big endian mode, running the Linux-based GNU system.
1480 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1481 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1483 @item powerpc-*-eabiaix
1484 Embedded PowerPC system in big endian mode with -mcall-aix selected as
1487 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1488 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1490 @item powerpc-*-eabisim
1491 Embedded PowerPC system in big endian mode for use in running under the
1494 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1495 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1497 @item powerpc-*-eabi
1498 Embedded PowerPC system in big endian mode.
1500 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1501 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1503 @item powerpcle-*-elf
1504 @itemx powerpcle-*-sysv4
1505 PowerPC system in little endian mode, running System V.4.
1507 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1508 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1510 @item powerpcle-*-solaris2*
1511 PowerPC system in little endian mode, running Solaris 2.5.1 or higher.
1513 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1514 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1515 Beta versions of the Sun 4.0 compiler do not seem to be able to build
1516 GNU CC correctly. There are also problems with the host assembler and
1517 linker that are fixed by using the GNU versions of these tools.
1519 @item powerpcle-*-eabisim
1520 Embedded PowerPC system in little endian mode for use in running under
1523 @itemx powerpcle-*-eabi
1524 Embedded PowerPC system in little endian mode.
1526 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1527 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1529 @item powerpcle-*-winnt
1530 @itemx powerpcle-*-pe
1531 PowerPC system in little endian mode running Windows NT.
1533 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1534 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1536 @item vax-dec-ultrix
1537 Don't try compiling with Vax C (@code{vcc}). It produces incorrect code
1538 in some cases (for example, when @code{alloca} is used).
1540 Meanwhile, compiling @file{cp/parse.c} with pcc does not work because of
1541 an internal table size limitation in that compiler. To avoid this
1542 problem, compile just the GNU C compiler first, and use it to recompile
1543 building all the languages that you want to run.
1546 See @ref{Sun Install}, for information on installing GNU CC on Sun
1550 See @ref{VMS Install}, for details on how to install GNU CC on VMS.
1553 These computers are also known as the 3b2, 3b5, 3b20 and other similar
1554 names. (However, the 3b1 is actually a 68000; see
1555 @ref{Configurations}.)
1557 Don't use @samp{-g} when compiling with the system's compiler. The
1558 system's linker seems to be unable to handle such a large program with
1559 debugging information.
1561 The system's compiler runs out of capacity when compiling @file{stmt.c}
1562 in GNU CC. You can work around this by building @file{cpp} in GNU CC
1563 first, then use that instead of the system's preprocessor with the
1564 system's C compiler to compile @file{stmt.c}. Here is how:
1567 mv /lib/cpp /lib/cpp.att
1569 echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp
1573 The system's compiler produces bad code for some of the GNU CC
1574 optimization files. So you must build the stage 2 compiler without
1575 optimization. Then build a stage 3 compiler with optimization.
1576 That executable should work. Here are the necessary commands:
1579 make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g"
1581 make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O"
1584 You may need to raise the ULIMIT setting to build a C++ compiler,
1585 as the file @file{cc1plus} is larger than one megabyte.
1589 @section Compilation in a Separate Directory
1590 @cindex other directory, compilation in
1591 @cindex compilation in a separate directory
1592 @cindex separate directory, compilation in
1594 If you wish to build the object files and executables in a directory
1595 other than the one containing the source files, here is what you must
1600 Make sure you have a version of Make that supports the @code{VPATH}
1601 feature. (GNU Make supports it, as do Make versions on most BSD
1605 If you have ever run @file{configure} in the source directory, you must undo
1606 the configuration. Do this by running:
1613 Go to the directory in which you want to build the compiler before
1614 running @file{configure}:
1621 On systems that do not support symbolic links, this directory must be
1622 on the same file system as the source code directory.
1625 Specify where to find @file{configure} when you run it:
1628 ../gcc/configure @dots{}
1631 This also tells @code{configure} where to find the compiler sources;
1632 @code{configure} takes the directory from the file name that was used to
1633 invoke it. But if you want to be sure, you can specify the source
1634 directory with the @samp{--srcdir} option, like this:
1637 ../gcc/configure --srcdir=../gcc @var{other options}
1640 The directory you specify with @samp{--srcdir} need not be the same
1641 as the one that @code{configure} is found in.
1644 Now, you can run @code{make} in that directory. You need not repeat the
1645 configuration steps shown above, when ordinary source files change. You
1646 must, however, run @code{configure} again when the configuration files
1647 change, if your system does not support symbolic links.
1649 @node Cross-Compiler
1650 @section Building and Installing a Cross-Compiler
1651 @cindex cross-compiler, installation
1653 GNU CC can function as a cross-compiler for many machines, but not all.
1657 Cross-compilers for the Mips as target using the Mips assembler
1658 currently do not work, because the auxiliary programs
1659 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
1660 anything but a Mips. It does work to cross compile for a Mips
1661 if you use the GNU assembler and linker.
1664 Cross-compilers between machines with different floating point formats
1665 have not all been made to work. GNU CC now has a floating point
1666 emulator with which these can work, but each target machine description
1667 needs to be updated to take advantage of it.
1670 Cross-compilation between machines of different word sizes is
1671 somewhat problematic and sometimes does not work.
1674 Since GNU CC generates assembler code, you probably need a
1675 cross-assembler that GNU CC can run, in order to produce object files.
1676 If you want to link on other than the target machine, you need a
1677 cross-linker as well. You also need header files and libraries suitable
1678 for the target machine that you can install on the host machine.
1681 * Steps of Cross:: Using a cross-compiler involves several steps
1682 that may be carried out on different machines.
1683 * Configure Cross:: Configuring a cross-compiler.
1684 * Tools and Libraries:: Where to put the linker and assembler, and the C library.
1685 * Cross Headers:: Finding and installing header files
1686 for a cross-compiler.
1687 * Cross Runtime:: Supplying arithmetic runtime routines (@file{libgcc1.a}).
1688 * Build Cross:: Actually compiling the cross-compiler.
1691 @node Steps of Cross
1692 @subsection Steps of Cross-Compilation
1694 To compile and run a program using a cross-compiler involves several
1699 Run the cross-compiler on the host machine to produce assembler files
1700 for the target machine. This requires header files for the target
1704 Assemble the files produced by the cross-compiler. You can do this
1705 either with an assembler on the target machine, or with a
1706 cross-assembler on the host machine.
1709 Link those files to make an executable. You can do this either with a
1710 linker on the target machine, or with a cross-linker on the host
1711 machine. Whichever machine you use, you need libraries and certain
1712 startup files (typically @file{crt@dots{}.o}) for the target machine.
1715 It is most convenient to do all of these steps on the same host machine,
1716 since then you can do it all with a single invocation of GNU CC. This
1717 requires a suitable cross-assembler and cross-linker. For some targets,
1718 the GNU assembler and linker are available.
1720 @node Configure Cross
1721 @subsection Configuring a Cross-Compiler
1723 To build GNU CC as a cross-compiler, you start out by running
1724 @file{configure}. Use the @samp{--target=@var{target}} to specify the
1725 target type. If @file{configure} was unable to correctly identify the
1726 system you are running on, also specify the @samp{--build=@var{build}}
1727 option. For example, here is how to configure for a cross-compiler that
1728 produces code for an HP 68030 system running BSD on a system that
1729 @file{configure} can correctly identify:
1732 ./configure --target=m68k-hp-bsd4.3
1735 @node Tools and Libraries
1736 @subsection Tools and Libraries for a Cross-Compiler
1738 If you have a cross-assembler and cross-linker available, you should
1739 install them now. Put them in the directory
1740 @file{/usr/local/@var{target}/bin}. Here is a table of the tools
1741 you should put in this directory:
1745 This should be the cross-assembler.
1748 This should be the cross-linker.
1751 This should be the cross-archiver: a program which can manipulate
1752 archive files (linker libraries) in the target machine's format.
1755 This should be a program to construct a symbol table in an archive file.
1758 The installation of GNU CC will find these programs in that directory,
1759 and copy or link them to the proper place to for the cross-compiler to
1760 find them when run later.
1762 The easiest way to provide these files is to build the Binutils package
1763 and GAS. Configure them with the same @samp{--host} and @samp{--target}
1764 options that you use for configuring GNU CC, then build and install
1765 them. They install their executables automatically into the proper
1766 directory. Alas, they do not support all the targets that GNU CC
1769 If you want to install libraries to use with the cross-compiler, such as
1770 a standard C library, put them in the directory
1771 @file{/usr/local/@var{target}/lib}; installation of GNU CC copies
1772 all the files in that subdirectory into the proper place for GNU CC to
1773 find them and link with them. Here's an example of copying some
1774 libraries from a target machine:
1777 ftp @var{target-machine}
1778 lcd /usr/local/@var{target}/lib
1788 The precise set of libraries you'll need, and their locations on
1789 the target machine, vary depending on its operating system.
1792 Many targets require ``start files'' such as @file{crt0.o} and
1793 @file{crtn.o} which are linked into each executable; these too should be
1794 placed in @file{/usr/local/@var{target}/lib}. There may be several
1795 alternatives for @file{crt0.o}, for use with profiling or other
1796 compilation options. Check your target's definition of
1797 @code{STARTFILE_SPEC} to find out what start files it uses.
1798 Here's an example of copying these files from a target machine:
1801 ftp @var{target-machine}
1802 lcd /usr/local/@var{target}/lib
1812 @subsection @file{libgcc.a} and Cross-Compilers
1814 Code compiled by GNU CC uses certain runtime support functions
1815 implicitly. Some of these functions can be compiled successfully with
1816 GNU CC itself, but a few cannot be. These problem functions are in the
1817 source file @file{libgcc1.c}; the library made from them is called
1820 When you build a native compiler, these functions are compiled with some
1821 other compiler--the one that you use for bootstrapping GNU CC.
1822 Presumably it knows how to open code these operations, or else knows how
1823 to call the run-time emulation facilities that the machine comes with.
1824 But this approach doesn't work for building a cross-compiler. The
1825 compiler that you use for building knows about the host system, not the
1828 So, when you build a cross-compiler you have to supply a suitable
1829 library @file{libgcc1.a} that does the job it is expected to do.
1831 To compile @file{libgcc1.c} with the cross-compiler itself does not
1832 work. The functions in this file are supposed to implement arithmetic
1833 operations that GNU CC does not know how to open code for your target
1834 machine. If these functions are compiled with GNU CC itself, they
1835 will compile into infinite recursion.
1837 On any given target, most of these functions are not needed. If GNU CC
1838 can open code an arithmetic operation, it will not call these functions
1839 to perform the operation. It is possible that on your target machine,
1840 none of these functions is needed. If so, you can supply an empty
1841 library as @file{libgcc1.a}.
1843 Many targets need library support only for multiplication and division.
1844 If you are linking with a library that contains functions for
1845 multiplication and division, you can tell GNU CC to call them directly
1846 by defining the macros @code{MULSI3_LIBCALL}, and the like. These
1847 macros need to be defined in the target description macro file. For
1848 some targets, they are defined already. This may be sufficient to
1849 avoid the need for libgcc1.a; if so, you can supply an empty library.
1851 Some targets do not have floating point instructions; they need other
1852 functions in @file{libgcc1.a}, which do floating arithmetic.
1853 Recent versions of GNU CC have a file which emulates floating point.
1854 With a certain amount of work, you should be able to construct a
1855 floating point emulator that can be used as @file{libgcc1.a}. Perhaps
1856 future versions will contain code to do this automatically and
1857 conveniently. That depends on whether someone wants to implement it.
1859 Some embedded targets come with all the necessary @file{libgcc1.a}
1860 routines written in C or assembler. These targets build
1861 @file{libgcc1.a} automatically and you do not need to do anything
1862 special for them. Other embedded targets do not need any
1863 @file{libgcc1.a} routines since all the necessary operations are
1864 supported by the hardware.
1866 If your target system has another C compiler, you can configure GNU CC
1867 as a native compiler on that machine, build just @file{libgcc1.a} with
1868 @samp{make libgcc1.a} on that machine, and use the resulting file with
1869 the cross-compiler. To do this, execute the following on the target
1873 cd @var{target-build-dir}
1874 ./configure --host=sparc --target=sun3
1879 And then this on the host machine:
1882 ftp @var{target-machine}
1884 cd @var{target-build-dir}
1889 Another way to provide the functions you need in @file{libgcc1.a} is to
1890 define the appropriate @code{perform_@dots{}} macros for those
1891 functions. If these definitions do not use the C arithmetic operators
1892 that they are meant to implement, you should be able to compile them
1893 with the cross-compiler you are building. (If these definitions already
1894 exist for your target file, then you are all set.)
1896 To build @file{libgcc1.a} using the perform macros, use
1897 @samp{LIBGCC1=libgcc1.a OLDCC=./xgcc} when building the compiler.
1898 Otherwise, you should place your replacement library under the name
1899 @file{libgcc1.a} in the directory in which you will build the
1900 cross-compiler, before you run @code{make}.
1903 @subsection Cross-Compilers and Header Files
1905 If you are cross-compiling a standalone program or a program for an
1906 embedded system, then you may not need any header files except the few
1907 that are part of GNU CC (and those of your program). However, if you
1908 intend to link your program with a standard C library such as
1909 @file{libc.a}, then you probably need to compile with the header files
1910 that go with the library you use.
1912 The GNU C compiler does not come with these files, because (1) they are
1913 system-specific, and (2) they belong in a C library, not in a compiler.
1915 If the GNU C library supports your target machine, then you can get the
1916 header files from there (assuming you actually use the GNU library when
1917 you link your program).
1919 If your target machine comes with a C compiler, it probably comes with
1920 suitable header files also. If you make these files accessible from the host
1921 machine, the cross-compiler can use them also.
1923 Otherwise, you're on your own in finding header files to use when
1926 When you have found suitable header files, put them in the directory
1927 @file{/usr/local/@var{target}/include}, before building the cross
1928 compiler. Then installation will run fixincludes properly and install
1929 the corrected versions of the header files where the compiler will use
1932 Provide the header files before you build the cross-compiler, because
1933 the build stage actually runs the cross-compiler to produce parts of
1934 @file{libgcc.a}. (These are the parts that @emph{can} be compiled with
1935 GNU CC.) Some of them need suitable header files.
1937 Here's an example showing how to copy the header files from a target
1938 machine. On the target machine, do this:
1941 (cd /usr/include; tar cf - .) > tarfile
1944 Then, on the host machine, do this:
1947 ftp @var{target-machine}
1948 lcd /usr/local/@var{target}/include
1955 @subsection Actually Building the Cross-Compiler
1957 Now you can proceed just as for compiling a single-machine compiler
1958 through the step of building stage 1. If you have not provided some
1959 sort of @file{libgcc1.a}, then compilation will give up at the point
1960 where it needs that file, printing a suitable error message. If you
1961 do provide @file{libgcc1.a}, then building the compiler will automatically
1962 compile and link a test program called @file{libgcc1-test}; if you get
1963 errors in the linking, it means that not all of the necessary routines
1964 in @file{libgcc1.a} are available.
1966 You must provide the header file @file{float.h}. One way to do this is
1967 to compile @file{enquire} and run it on your target machine. The job of
1968 @file{enquire} is to run on the target machine and figure out by
1969 experiment the nature of its floating point representation.
1970 @file{enquire} records its findings in the header file @file{float.h}.
1971 If you can't produce this file by running @file{enquire} on the target
1972 machine, then you will need to come up with a suitable @file{float.h} in
1973 some other way (or else, avoid using it in your programs).
1975 Do not try to build stage 2 for a cross-compiler. It doesn't work to
1976 rebuild GNU CC as a cross-compiler using the cross-compiler, because
1977 that would produce a program that runs on the target machine, not on the
1978 host. For example, if you compile a 386-to-68030 cross-compiler with
1979 itself, the result will not be right either for the 386 (because it was
1980 compiled into 68030 code) or for the 68030 (because it was configured
1981 for a 386 as the host). If you want to compile GNU CC into 68030 code,
1982 whether you compile it on a 68030 or with a cross-compiler on a 386, you
1983 must specify a 68030 as the host when you configure it.
1985 To install the cross-compiler, use @samp{make install}, as usual.
1988 @section Installing GNU CC on the Sun
1989 @cindex Sun installation
1990 @cindex installing GNU CC on the Sun
1992 On Solaris, do not use the linker or other tools in
1993 @file{/usr/ucb} to build GNU CC. Use @code{/usr/ccs/bin}.
1995 If the assembler reports @samp{Error: misaligned data} when bootstrapping,
1996 you are probably using an obsolete version of the GNU assembler. Upgrade
1997 to the latest version of GNU @code{binutils}, or use the Solaris assembler.
1999 Make sure the environment variable @code{FLOAT_OPTION} is not set when
2000 you compile @file{libgcc.a}. If this option were set to @code{f68881}
2001 when @file{libgcc.a} is compiled, the resulting code would demand to be
2002 linked with a special startup file and would not link properly without
2005 @cindex @code{alloca}, for SunOS
2006 There is a bug in @code{alloca} in certain versions of the Sun library.
2007 To avoid this bug, install the binaries of GNU CC that were compiled by
2008 GNU CC. They use @code{alloca} as a built-in function and never the one
2011 Some versions of the Sun compiler crash when compiling GNU CC. The
2012 problem is a segmentation fault in cpp. This problem seems to be due to
2013 the bulk of data in the environment variables. You may be able to avoid
2014 it by using the following command to compile GNU CC with Sun CC:
2017 make CC="TERMCAP=x OBJS=x LIBFUNCS=x STAGESTUFF=x cc"
2020 SunOS 4.1.3 and 4.1.3_U1 have bugs that can cause intermittent core
2021 dumps when compiling GNU CC. A common symptom is an
2022 internal compiler error which does not recur if you run it again.
2023 To fix the problem, install Sun recommended patch 100726 (for SunOS 4.1.3)
2024 or 101508 (for SunOS 4.1.3_U1), or upgrade to a later SunOS release.
2027 @section Installing GNU CC on VMS
2028 @cindex VMS installation
2029 @cindex installing GNU CC on VMS
2031 The VMS version of GNU CC is distributed in a backup saveset containing
2032 both source code and precompiled binaries.
2034 To install the @file{gcc} command so you can use the compiler easily, in
2035 the same manner as you use the VMS C compiler, you must install the VMS CLD
2036 file for GNU CC as follows:
2040 Define the VMS logical names @samp{GNU_CC} and @samp{GNU_CC_INCLUDE}
2041 to point to the directories where the GNU CC executables
2042 (@file{gcc-cpp.exe}, @file{gcc-cc1.exe}, etc.) and the C include files are
2043 kept respectively. This should be done with the commands:@refill
2046 $ assign /system /translation=concealed -
2048 $ assign /system /translation=concealed -
2049 disk:[gcc.include.] gnu_cc_include
2053 with the appropriate disk and directory names. These commands can be
2054 placed in your system startup file so they will be executed whenever
2055 the machine is rebooted. You may, if you choose, do this via the
2056 @file{GCC_INSTALL.COM} script in the @file{[GCC]} directory.
2059 Install the @file{GCC} command with the command line:
2062 $ set command /table=sys$common:[syslib]dcltables -
2063 /output=sys$common:[syslib]dcltables gnu_cc:[000000]gcc
2064 $ install replace sys$common:[syslib]dcltables
2068 To install the help file, do the following:
2071 $ library/help sys$library:helplib.hlb gcc.hlp
2075 Now you can invoke the compiler with a command like @samp{gcc /verbose
2076 file.c}, which is equivalent to the command @samp{gcc -v -c file.c} in
2080 If you wish to use GNU C++ you must first install GNU CC, and then
2081 perform the following steps:
2085 Define the VMS logical name @samp{GNU_GXX_INCLUDE} to point to the
2086 directory where the preprocessor will search for the C++ header files.
2087 This can be done with the command:@refill
2090 $ assign /system /translation=concealed -
2091 disk:[gcc.gxx_include.] gnu_gxx_include
2095 with the appropriate disk and directory name. If you are going to be
2096 using a C++ runtime library, this is where its install procedure will install
2100 Obtain the file @file{gcc-cc1plus.exe}, and place this in the same
2101 directory that @file{gcc-cc1.exe} is kept.
2103 The GNU C++ compiler can be invoked with a command like @samp{gcc /plus
2104 /verbose file.cc}, which is equivalent to the command @samp{g++ -v -c
2108 We try to put corresponding binaries and sources on the VMS distribution
2109 tape. But sometimes the binaries will be from an older version than the
2110 sources, because we don't always have time to update them. (Use the
2111 @samp{/version} option to determine the version number of the binaries and
2112 compare it with the source file @file{version.c} to tell whether this is
2113 so.) In this case, you should use the binaries you get to recompile the
2114 sources. If you must recompile, here is how:
2118 Execute the command procedure @file{vmsconfig.com} to set up the files
2119 @file{tm.h}, @file{config.h}, @file{aux-output.c}, and @file{md.}, and
2120 to create files @file{tconfig.h} and @file{hconfig.h}. This procedure
2121 also creates several linker option files used by @file{make-cc1.com} and
2122 a data file used by @file{make-l2.com}.@refill
2129 Setup the logical names and command tables as defined above. In
2130 addition, define the VMS logical name @samp{GNU_BISON} to point at the
2131 to the directories where the Bison executable is kept. This should be
2132 done with the command:@refill
2135 $ assign /system /translation=concealed -
2136 disk:[bison.] gnu_bison
2139 You may, if you choose, use the @file{INSTALL_BISON.COM} script in the
2140 @file{[BISON]} directory.
2143 Install the @samp{BISON} command with the command line:@refill
2146 $ set command /table=sys$common:[syslib]dcltables -
2147 /output=sys$common:[syslib]dcltables -
2148 gnu_bison:[000000]bison
2149 $ install replace sys$common:[syslib]dcltables
2153 Type @samp{@@make-gcc} to recompile everything (alternatively, submit
2154 the file @file{make-gcc.com} to a batch queue). If you wish to build
2155 the GNU C++ compiler as well as the GNU CC compiler, you must first edit
2156 @file{make-gcc.com} and follow the instructions that appear in the
2160 In order to use GCC, you need a library of functions which GCC compiled code
2161 will call to perform certain tasks, and these functions are defined in the
2162 file @file{libgcc2.c}. To compile this you should use the command procedure
2163 @file{make-l2.com}, which will generate the library @file{libgcc2.olb}.
2164 @file{libgcc2.olb} should be built using the compiler built from
2165 the same distribution that @file{libgcc2.c} came from, and
2166 @file{make-gcc.com} will automatically do all of this for you.
2168 To install the library, use the following commands:@refill
2171 $ library gnu_cc:[000000]gcclib/delete=(new,eprintf)
2172 $ library gnu_cc:[000000]gcclib/delete=L_*
2173 $ library libgcc2/extract=*/output=libgcc2.obj
2174 $ library gnu_cc:[000000]gcclib libgcc2.obj
2177 The first command simply removes old modules that will be replaced with
2178 modules from @file{libgcc2} under different module names. The modules
2179 @code{new} and @code{eprintf} may not actually be present in your
2180 @file{gcclib.olb}---if the VMS librarian complains about those modules
2181 not being present, simply ignore the message and continue on with the
2182 next command. The second command removes the modules that came from the
2183 previous version of the library @file{libgcc2.c}.
2185 Whenever you update the compiler on your system, you should also update the
2186 library with the above procedure.
2189 You may wish to build GCC in such a way that no files are written to the
2190 directory where the source files reside. An example would be the when
2191 the source files are on a read-only disk. In these cases, execute the
2192 following DCL commands (substituting your actual path names):
2195 $ assign dua0:[gcc.build_dir.]/translation=concealed, -
2196 dua1:[gcc.source_dir.]/translation=concealed gcc_build
2197 $ set default gcc_build:[000000]
2201 where the directory @file{dua1:[gcc.source_dir]} contains the source
2202 code, and the directory @file{dua0:[gcc.build_dir]} is meant to contain
2203 all of the generated object files and executables. Once you have done
2204 this, you can proceed building GCC as described above. (Keep in mind
2205 that @file{gcc_build} is a rooted logical name, and thus the device
2206 names in each element of the search list must be an actual physical
2207 device name rather than another rooted logical name).
2210 @strong{If you are building GNU CC with a previous version of GNU CC,
2211 you also should check to see that you have the newest version of the
2212 assembler}. In particular, GNU CC version 2 treats global constant
2213 variables slightly differently from GNU CC version 1, and GAS version
2214 1.38.1 does not have the patches required to work with GCC version 2.
2215 If you use GAS 1.38.1, then @code{extern const} variables will not have
2216 the read-only bit set, and the linker will generate warning messages
2217 about mismatched psect attributes for these variables. These warning
2218 messages are merely a nuisance, and can safely be ignored.
2220 If you are compiling with a version of GNU CC older than 1.33, specify
2221 @samp{/DEFINE=("inline=")} as an option in all the compilations. This
2222 requires editing all the @code{gcc} commands in @file{make-cc1.com}.
2223 (The older versions had problems supporting @code{inline}.) Once you
2224 have a working 1.33 or newer GNU CC, you can change this file back.
2227 If you want to build GNU CC with the VAX C compiler, you will need to
2228 make minor changes in @file{make-cccp.com} and @file{make-cc1.com}
2229 to choose alternate definitions of @code{CC}, @code{CFLAGS}, and
2230 @code{LIBS}. See comments in those files. However, you must
2231 also have a working version of the GNU assembler (GNU as, aka GAS) as
2232 it is used as the back-end for GNU CC to produce binary object modules
2233 and is not included in the GNU CC sources. GAS is also needed to
2234 compile @file{libgcc2} in order to build @file{gcclib} (see above);
2235 @file{make-l2.com} expects to be able to find it operational in
2236 @file{gnu_cc:[000000]gnu-as.exe}.
2238 To use GNU CC on VMS, you need the VMS driver programs
2239 @file{gcc.exe}, @file{gcc.com}, and @file{gcc.cld}. They are
2240 distributed with the VMS binaries (@file{gcc-vms}) rather than the
2241 GNU CC sources. GAS is also included in @file{gcc-vms}, as is Bison.
2243 Once you have successfully built GNU CC with VAX C, you should use the
2244 resulting compiler to rebuild itself. Before doing this, be sure to
2245 restore the @code{CC}, @code{CFLAGS}, and @code{LIBS} definitions in
2246 @file{make-cccp.com} and @file{make-cc1.com}. The second generation
2247 compiler will be able to take advantage of many optimizations that must
2248 be suppressed when building with other compilers.
2251 Under previous versions of GNU CC, the generated code would occasionally
2252 give strange results when linked with the sharable @file{VAXCRTL} library.
2253 Now this should work.
2255 Even with this version, however, GNU CC itself should not be linked with
2256 the sharable @file{VAXCRTL}. The version of @code{qsort} in
2257 @file{VAXCRTL} has a bug (known to be present in VMS versions V4.6
2258 through V5.5) which causes the compiler to fail.
2260 The executables are generated by @file{make-cc1.com} and
2261 @file{make-cccp.com} use the object library version of @file{VAXCRTL} in
2262 order to make use of the @code{qsort} routine in @file{gcclib.olb}. If
2263 you wish to link the compiler executables with the shareable image
2264 version of @file{VAXCRTL}, you should edit the file @file{tm.h} (created
2265 by @file{vmsconfig.com}) to define the macro @code{QSORT_WORKAROUND}.
2267 @code{QSORT_WORKAROUND} is always defined when GNU CC is compiled with
2268 VAX C, to avoid a problem in case @file{gcclib.olb} is not yet
2272 @section @code{collect2}
2274 GNU CC uses a utility called @code{collect2} on nearly all systems to arrange
2275 to call various initialization functions at start time.
2277 The program @code{collect2} works by linking the program once and
2278 looking through the linker output file for symbols with particular names
2279 indicating they are constructor functions. If it finds any, it
2280 creates a new temporary @samp{.c} file containing a table of them,
2281 compiles it, and links the program a second time including that file.
2284 @cindex constructors, automatic calls
2285 The actual calls to the constructors are carried out by a subroutine
2286 called @code{__main}, which is called (automatically) at the beginning
2287 of the body of @code{main} (provided @code{main} was compiled with GNU
2288 CC). Calling @code{__main} is necessary, even when compiling C code, to
2289 allow linking C and C++ object code together. (If you use
2290 @samp{-nostdlib}, you get an unresolved reference to @code{__main},
2291 since it's defined in the standard GCC library. Include @samp{-lgcc} at
2292 the end of your compiler command line to resolve this reference.)
2294 The program @code{collect2} is installed as @code{ld} in the directory
2295 where the passes of the compiler are installed. When @code{collect2}
2296 needs to find the @emph{real} @code{ld}, it tries the following file
2301 @file{real-ld} in the directories listed in the compiler's search
2305 @file{real-ld} in the directories listed in the environment variable
2309 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2313 @file{ld} in the compiler's search directories, except that
2314 @code{collect2} will not execute itself recursively.
2317 @file{ld} in @code{PATH}.
2320 ``The compiler's search directories'' means all the directories where
2321 @code{gcc} searches for passes of the compiler. This includes
2322 directories that you specify with @samp{-B}.
2324 Cross-compilers search a little differently:
2328 @file{real-ld} in the compiler's search directories.
2331 @file{@var{target}-real-ld} in @code{PATH}.
2334 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2338 @file{ld} in the compiler's search directories.
2341 @file{@var{target}-ld} in @code{PATH}.
2344 @code{collect2} explicitly avoids running @code{ld} using the file name
2345 under which @code{collect2} itself was invoked. In fact, it remembers
2346 up a list of such names---in case one copy of @code{collect2} finds
2347 another copy (or version) of @code{collect2} installed as @code{ld} in a
2348 second place in the search path.
2350 @code{collect2} searches for the utilities @code{nm} and @code{strip}
2351 using the same algorithm as above for @code{ld}.
2354 @section Standard Header File Directories
2356 @code{GCC_INCLUDE_DIR} means the same thing for native and cross. It is
2357 where GNU CC stores its private include files, and also where GNU CC
2358 stores the fixed include files. A cross compiled GNU CC runs
2359 @code{fixincludes} on the header files in @file{$(tooldir)/include}.
2360 (If the cross compilation header files need to be fixed, they must be
2361 installed before GNU CC is built. If the cross compilation header files
2362 are already suitable for ANSI C and GNU CC, nothing special need be
2365 @code{GPLUS_INCLUDE_DIR} means the same thing for native and cross. It
2366 is where @code{g++} looks first for header files. The C++ library
2367 installs only target independent header files in that directory.
2369 @code{LOCAL_INCLUDE_DIR} is used only for a native compiler. It is
2370 normally @file{/usr/local/include}. GNU CC searches this directory so
2371 that users can install header files in @file{/usr/local/include}.
2373 @code{CROSS_INCLUDE_DIR} is used only for a cross compiler. GNU CC
2374 doesn't install anything there.
2376 @code{TOOL_INCLUDE_DIR} is used for both native and cross compilers. It
2377 is the place for other packages to install header files that GNU CC will
2378 use. For a cross-compiler, this is the equivalent of
2379 @file{/usr/include}. When you build a cross-compiler,
2380 @code{fixincludes} processes any header files in this directory.