]>
Commit | Line | Data |
---|---|---|
c85f7c16 | 1 | @c Copyright (C) 1988,89,92,93,94,95,96,97,1998 Free Software Foundation, Inc. |
2284f91b DE |
2 | @c This is part of the GCC manual. |
3 | @c For copying conditions, see the file gcc.texi. | |
4 | ||
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. | |
7 | ||
f2d76545 JL |
8 | Note most of this information is out of date and superceded by the EGCS |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 It is provided for historical reference only. | |
10 | ||
2284f91b DE |
11 | @ifclear INSTALLONLY |
12 | @node Installation | |
13 | @chapter Installing GNU CC | |
14 | @end ifclear | |
15 | @cindex installing GNU CC | |
16 | ||
17 | @menu | |
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. | |
25 | @end menu | |
26 | ||
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 | |
31 | systems. | |
32 | ||
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. | |
37 | ||
38 | @enumerate | |
39 | @item | |
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 | |
45 | clean. | |
46 | ||
47 | @item | |
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. | |
51 | ||
52 | @item | |
53 | Specify the host, build and target machine configurations. You do this | |
54 | by running the file @file{configure}. | |
55 | ||
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. | |
60 | ||
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 | |
67 | wrong. | |
68 | ||
69 | In those cases, specify the build machine's @dfn{configuration name} | |
e5e809f4 JL |
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, | |
2284f91b DE |
72 | see @ref{Cross-Compiler}.) |
73 | ||
74 | Here is an example: | |
75 | ||
76 | @smallexample | |
77 | ./configure --build=sparc-sun-sunos4.1 | |
78 | @end smallexample | |
79 | ||
80 | A configuration name may be canonical or it may be more or less | |
81 | abbreviated. | |
82 | ||
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. | |
88 | ||
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. | |
94 | ||
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. | |
98 | ||
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. | |
102 | ||
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 | |
106 | @samp{--nfp}. | |
107 | ||
108 | @table @samp | |
109 | @item --with-gnu-as | |
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 | |
112 | @file{configure}. | |
113 | ||
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. | |
116 | ||
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. | |
123 | ||
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. | |
133 | ||
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}). | |
137 | ||
138 | @item --with-gnu-ld | |
139 | Specify the option @samp{--with-gnu-ld} if you plan to use the GNU | |
140 | linker with GNU CC. | |
141 | ||
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. | |
ace62b49 JL |
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 | |
146 | @c configurations. | |
2284f91b DE |
147 | |
148 | @item --with-stabs | |
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. | |
154 | ||
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 | |
157 | CC. | |
158 | ||
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. | |
162 | ||
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. | |
167 | ||
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. | |
174 | ||
175 | @item --nfp | |
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 | |
180 | make a difference. | |
181 | ||
9101297d DL |
182 | @cindex Haifa scheduler |
183 | @cindex scheduler, experimental | |
184 | @item --enable-haifa | |
04afd9d6 | 185 | @itemx --disable-haifa |
9101297d DL |
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. | |
191 | ||
2284f91b DE |
192 | @cindex Objective C threads |
193 | @cindex threads, Objective C | |
6d8ccdbb | 194 | @item --enable-threads=@var{type} |
861bb6c1 JL |
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}. | |
2284f91b DE |
202 | @end table |
203 | ||
204 | The @file{configure} script searches subdirectories of the source | |
205 | directory for other compilers that are to be integrated into GNU CC. | |
206 | The GNU compiler for C++, called G++ is in a subdirectory named | |
207 | @file{cp}. @file{configure} inserts rules into @file{Makefile} to build | |
208 | all of those compilers. | |
209 | ||
210 | Here we spell out what files will be set up by @code{configure}. Normally | |
211 | you need not be concerned with these files. | |
212 | ||
213 | @itemize @bullet | |
214 | @item | |
215 | @ifset INTERNALS | |
216 | A file named @file{config.h} is created that contains a @samp{#include} | |
217 | of the top-level config file for the machine you will run the compiler | |
218 | on (@pxref{Config}). This file is responsible for defining information | |
219 | about the host machine. It includes @file{tm.h}. | |
220 | @end ifset | |
221 | @ifclear INTERNALS | |
222 | A file named @file{config.h} is created that contains a @samp{#include} | |
223 | of the top-level config file for the machine you will run the compiler | |
224 | on (@pxref{Config,,The Configuration File, gcc.info, Using and Porting | |
225 | GCC}). This file is responsible for defining information about the host | |
226 | machine. It includes @file{tm.h}. | |
227 | @end ifclear | |
228 | ||
229 | The top-level config file is located in the subdirectory @file{config}. | |
230 | Its name is always @file{xm-@var{something}.h}; usually | |
231 | @file{xm-@var{machine}.h}, but there are some exceptions. | |
232 | ||
233 | If your system does not support symbolic links, you might want to | |
234 | set up @file{config.h} to contain a @samp{#include} command which | |
235 | refers to the appropriate file. | |
236 | ||
237 | @item | |
238 | A file named @file{tconfig.h} is created which includes the top-level config | |
239 | file for your target machine. This is used for compiling certain | |
240 | programs to run on that machine. | |
241 | ||
242 | @item | |
243 | A file named @file{tm.h} is created which includes the | |
244 | machine-description macro file for your target machine. It should be in | |
245 | the subdirectory @file{config} and its name is often | |
246 | @file{@var{machine}.h}. | |
247 | ||
248 | @item | |
249 | The command file @file{configure} also constructs the file | |
250 | @file{Makefile} by adding some text to the template file | |
251 | @file{Makefile.in}. The additional text comes from files in the | |
252 | @file{config} directory, named @file{t-@var{target}} and | |
253 | @file{x-@var{host}}. If these files do not exist, it means nothing | |
254 | needs to be added for a given target or host. | |
255 | @end itemize | |
256 | ||
257 | @item | |
258 | The standard directory for installing GNU CC is @file{/usr/local/lib}. | |
259 | If you want to install its files somewhere else, specify | |
260 | @samp{--prefix=@var{dir}} when you run @file{configure}. Here @var{dir} | |
261 | is a directory name to use instead of @file{/usr/local} for all purposes | |
262 | with one exception: the directory @file{/usr/local/include} is searched | |
263 | for header files no matter where you install the compiler. To override | |
e5e809f4 JL |
264 | this name, use the @code{--with-local-prefix} option below. The directory |
265 | you specify need not exist, but its parent directory must exist. | |
2284f91b DE |
266 | |
267 | @item | |
e5e809f4 | 268 | Specify @samp{--with-local-prefix=@var{dir}} if you want the compiler to |
2284f91b DE |
269 | search directory @file{@var{dir}/include} for locally installed header |
270 | files @emph{instead} of @file{/usr/local/include}. | |
271 | ||
e5e809f4 | 272 | You should specify @samp{--with-local-prefix} @strong{only} if your site has |
2284f91b DE |
273 | a different convention (not @file{/usr/local}) for where to put |
274 | site-specific files. | |
275 | ||
e5e809f4 | 276 | The default value for @samp{--with-local-prefix} is @file{/usr/local} |
2284f91b DE |
277 | regardless of the value of @samp{--prefix}. Specifying @samp{--prefix} |
278 | has no effect on which directory GNU CC searches for local header files. | |
279 | This may seem counterintuitive, but actually it is logical. | |
280 | ||
281 | The purpose of @samp{--prefix} is to specify where to @emph{install GNU | |
282 | CC}. The local header files in @file{/usr/local/include}---if you put | |
283 | any in that directory---are not part of GNU CC. They are part of other | |
284 | programs---perhaps many others. (GNU CC installs its own header files | |
285 | in another directory which is based on the @samp{--prefix} value.) | |
286 | ||
e5e809f4 JL |
287 | @strong{Do not} specify @file{/usr} as the @samp{--with-local-prefix}! The |
288 | directory you use for @samp{--with-local-prefix} @strong{must not} contain | |
2284f91b DE |
289 | any of the system's standard header files. If it did contain them, |
290 | certain programs would be miscompiled (including GNU Emacs, on certain | |
291 | targets), because this would override and nullify the header file | |
292 | corrections made by the @code{fixincludes} script. | |
293 | ||
294 | Indications are that people who use this option use it based on | |
295 | mistaken ideas of what it is for. People use it as if it specified | |
296 | where to install part of GNU CC. Perhaps they make this assumption | |
297 | because installing GNU CC creates the directory. | |
298 | ||
299 | @cindex Bison parser generator | |
300 | @cindex parser generator, Bison | |
301 | @item | |
302 | Make sure the Bison parser generator is installed. (This is | |
303 | unnecessary if the Bison output files @file{c-parse.c} and | |
304 | @file{cexp.c} are more recent than @file{c-parse.y} and @file{cexp.y} | |
305 | and you do not plan to change the @samp{.y} files.) | |
306 | ||
307 | Bison versions older than Sept 8, 1988 will produce incorrect output | |
308 | for @file{c-parse.c}. | |
309 | ||
310 | @item | |
311 | If you have chosen a configuration for GNU CC which requires other GNU | |
312 | tools (such as GAS or the GNU linker) instead of the standard system | |
313 | tools, install the required tools in the build directory under the names | |
314 | @file{as}, @file{ld} or whatever is appropriate. This will enable the | |
315 | compiler to find the proper tools for compilation of the program | |
316 | @file{enquire}. | |
317 | ||
318 | Alternatively, you can do subsequent compilation using a value of the | |
319 | @code{PATH} environment variable such that the necessary GNU tools come | |
320 | before the standard system tools. | |
321 | ||
322 | @item | |
323 | Build the compiler. Just type @samp{make LANGUAGES=c} in the compiler | |
324 | directory. | |
325 | ||
326 | @samp{LANGUAGES=c} specifies that only the C compiler should be | |
327 | compiled. The makefile normally builds compilers for all the supported | |
328 | languages; currently, C, C++ and Objective C. However, C is the only | |
329 | language that is sure to work when you build with other non-GNU C | |
330 | compilers. In addition, building anything but C at this stage is a | |
331 | waste of time. | |
332 | ||
333 | In general, you can specify the languages to build by typing the | |
334 | argument @samp{LANGUAGES="@var{list}"}, where @var{list} is one or more | |
335 | words from the list @samp{c}, @samp{c++}, and @samp{objective-c}. If | |
336 | you have any additional GNU compilers as subdirectories of the GNU CC | |
337 | source directory, you may also specify their names in this list. | |
338 | ||
339 | Ignore any warnings you may see about ``statement not reached'' in | |
340 | @file{insn-emit.c}; they are normal. Also, warnings about ``unknown | |
341 | escape sequence'' are normal in @file{genopinit.c} and perhaps some | |
342 | other files. Likewise, you should ignore warnings about ``constant is | |
343 | so large that it is unsigned'' in @file{insn-emit.c} and | |
344 | @file{insn-recog.c} and a warning about a comparison always being zero | |
345 | in @file{enquire.o}. Any other compilation errors may represent bugs in | |
346 | the port to your machine or operating system, and | |
347 | @ifclear INSTALLONLY | |
348 | should be investigated and reported (@pxref{Bugs}). | |
349 | @end ifclear | |
350 | @ifset INSTALLONLY | |
351 | should be investigated and reported. | |
352 | @end ifset | |
353 | ||
354 | Some commercial compilers fail to compile GNU CC because they have bugs | |
355 | or limitations. For example, the Microsoft compiler is said to run out | |
356 | of macro space. Some Ultrix compilers run out of expression space; then | |
357 | you need to break up the statement where the problem happens. | |
358 | ||
359 | @item | |
360 | If you are building a cross-compiler, stop here. @xref{Cross-Compiler}. | |
361 | ||
362 | @cindex stage1 | |
363 | @item | |
364 | Move the first-stage object files and executables into a subdirectory | |
365 | with this command: | |
366 | ||
367 | @smallexample | |
368 | make stage1 | |
369 | @end smallexample | |
370 | ||
371 | The files are moved into a subdirectory named @file{stage1}. | |
372 | Once installation is complete, you may wish to delete these files | |
373 | with @code{rm -r stage1}. | |
374 | ||
375 | @item | |
376 | If you have chosen a configuration for GNU CC which requires other GNU | |
377 | tools (such as GAS or the GNU linker) instead of the standard system | |
378 | tools, install the required tools in the @file{stage1} subdirectory | |
379 | under the names @file{as}, @file{ld} or whatever is appropriate. This | |
380 | will enable the stage 1 compiler to find the proper tools in the | |
381 | following stage. | |
382 | ||
383 | Alternatively, you can do subsequent compilation using a value of the | |
384 | @code{PATH} environment variable such that the necessary GNU tools come | |
385 | before the standard system tools. | |
386 | ||
387 | @item | |
388 | Recompile the compiler with itself, with this command: | |
389 | ||
390 | @smallexample | |
391 | make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2" | |
392 | @end smallexample | |
393 | ||
394 | This is called making the stage 2 compiler. | |
395 | ||
396 | The command shown above builds compilers for all the supported | |
397 | languages. If you don't want them all, you can specify the languages to | |
398 | build by typing the argument @samp{LANGUAGES="@var{list}"}. @var{list} | |
399 | should contain one or more words from the list @samp{c}, @samp{c++}, | |
400 | @samp{objective-c}, and @samp{proto}. Separate the words with spaces. | |
401 | @samp{proto} stands for the programs @code{protoize} and | |
402 | @code{unprotoize}; they are not a separate language, but you use | |
403 | @code{LANGUAGES} to enable or disable their installation. | |
404 | ||
405 | If you are going to build the stage 3 compiler, then you might want to | |
406 | build only the C language in stage 2. | |
407 | ||
408 | Once you have built the stage 2 compiler, if you are short of disk | |
409 | space, you can delete the subdirectory @file{stage1}. | |
410 | ||
411 | On a 68000 or 68020 system lacking floating point hardware, | |
412 | unless you have selected a @file{tm.h} file that expects by default | |
413 | that there is no such hardware, do this instead: | |
414 | ||
415 | @smallexample | |
416 | make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2 -msoft-float" | |
417 | @end smallexample | |
418 | ||
419 | @item | |
420 | If you wish to test the compiler by compiling it with itself one more | |
421 | time, install any other necessary GNU tools (such as GAS or the GNU | |
422 | linker) in the @file{stage2} subdirectory as you did in the | |
423 | @file{stage1} subdirectory, then do this: | |
424 | ||
425 | @smallexample | |
426 | make stage2 | |
427 | make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2" | |
428 | @end smallexample | |
429 | ||
430 | @noindent | |
431 | This is called making the stage 3 compiler. Aside from the @samp{-B} | |
432 | option, the compiler options should be the same as when you made the | |
433 | stage 2 compiler. But the @code{LANGUAGES} option need not be the | |
434 | same. The command shown above builds compilers for all the supported | |
435 | languages; if you don't want them all, you can specify the languages to | |
436 | build by typing the argument @samp{LANGUAGES="@var{list}"}, as described | |
437 | above. | |
438 | ||
439 | If you do not have to install any additional GNU tools, you may use the | |
440 | command | |
441 | ||
442 | @smallexample | |
443 | make bootstrap LANGUAGES=@var{language-list} BOOT_CFLAGS=@var{option-list} | |
444 | @end smallexample | |
445 | ||
446 | @noindent | |
447 | instead of making @file{stage1}, @file{stage2}, and performing | |
448 | the two compiler builds. | |
449 | ||
450 | @item | |
451 | Then compare the latest object files with the stage 2 object | |
452 | files---they ought to be identical, aside from time stamps (if any). | |
453 | ||
454 | On some systems, meaningful comparison of object files is impossible; | |
455 | they always appear ``different.'' This is currently true on Solaris and | |
456 | some systems that use ELF object file format. On some versions of Irix | |
457 | on SGI machines and DEC Unix (OSF/1) on Alpha systems, you will not be | |
458 | able to compare the files without specifying @file{-save-temps}; see the | |
459 | description of individual systems above to see if you get comparison | |
460 | failures. You may have similar problems on other systems. | |
461 | ||
462 | Use this command to compare the files: | |
463 | ||
464 | @smallexample | |
465 | make compare | |
466 | @end smallexample | |
467 | ||
468 | This will mention any object files that differ between stage 2 and stage | |
469 | 3. Any difference, no matter how innocuous, indicates that the stage 2 | |
470 | compiler has compiled GNU CC incorrectly, and is therefore a potentially | |
471 | @ifclear INSTALLONLY | |
472 | serious bug which you should investigate and report (@pxref{Bugs}). | |
473 | @end ifclear | |
474 | @ifset INSTALLONLY | |
475 | serious bug which you should investigate and report. | |
476 | @end ifset | |
477 | ||
478 | If your system does not put time stamps in the object files, then this | |
479 | is a faster way to compare them (using the Bourne shell): | |
480 | ||
481 | @smallexample | |
482 | for file in *.o; do | |
483 | cmp $file stage2/$file | |
484 | done | |
485 | @end smallexample | |
486 | ||
487 | If you have built the compiler with the @samp{-mno-mips-tfile} option on | |
488 | MIPS machines, you will not be able to compare the files. | |
489 | ||
490 | @item | |
491 | Install the compiler driver, the compiler's passes and run-time support | |
492 | with @samp{make install}. Use the same value for @code{CC}, | |
493 | @code{CFLAGS} and @code{LANGUAGES} that you used when compiling the | |
494 | files that are being installed. One reason this is necessary is that | |
495 | some versions of Make have bugs and recompile files gratuitously when | |
496 | you do this step. If you use the same variable values, those files will | |
497 | be recompiled properly. | |
498 | ||
499 | For example, if you have built the stage 2 compiler, you can use the | |
500 | following command: | |
501 | ||
502 | @smallexample | |
503 | make install CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O" LANGUAGES="@var{list}" | |
504 | @end smallexample | |
505 | ||
506 | @noindent | |
507 | This copies the files @file{cc1}, @file{cpp} and @file{libgcc.a} to | |
508 | files @file{cc1}, @file{cpp} and @file{libgcc.a} in the directory | |
509 | @file{/usr/local/lib/gcc-lib/@var{target}/@var{version}}, which is where | |
510 | the compiler driver program looks for them. Here @var{target} is the | |
e5e809f4 JL |
511 | canonicalized form of target machine type specified when you ran |
512 | @file{configure}, and @var{version} is the version number of GNU CC. | |
513 | This naming scheme permits various versions and/or cross-compilers to | |
514 | coexist. It also copies the executables for compilers for other | |
515 | languages (e.g., @file{cc1plus} for C++) to the same directory. | |
2284f91b DE |
516 | |
517 | This also copies the driver program @file{xgcc} into | |
518 | @file{/usr/local/bin/gcc}, so that it appears in typical execution | |
519 | search paths. It also copies @file{gcc.1} into | |
520 | @file{/usr/local/man/man1} and info pages into @file{/usr/local/info}. | |
521 | ||
522 | On some systems, this command causes recompilation of some files. This | |
523 | is usually due to bugs in @code{make}. You should either ignore this | |
524 | problem, or use GNU Make. | |
525 | ||
526 | @cindex @code{alloca} and SunOS | |
527 | @strong{Warning: there is a bug in @code{alloca} in the Sun library. To | |
528 | avoid this bug, be sure to install the executables of GNU CC that were | |
529 | compiled by GNU CC. (That is, the executables from stage 2 or 3, not | |
530 | stage 1.) They use @code{alloca} as a built-in function and never the | |
531 | one in the library.} | |
532 | ||
533 | (It is usually better to install GNU CC executables from stage 2 or 3, | |
534 | since they usually run faster than the ones compiled with some other | |
535 | compiler.) | |
536 | ||
537 | @item | |
c85f7c16 JL |
538 | @cindex C++ runtime library |
539 | @cindex @code{libstdc++} | |
2284f91b | 540 | If you're going to use C++, it's likely that you need to also install |
c85f7c16 JL |
541 | a C++ runtime library. Just as GNU C does not |
542 | distribute a C runtime library, it also does not include a C++ runtime | |
2284f91b | 543 | library. All I/O functionality, special class libraries, etc., are |
c85f7c16 JL |
544 | provided by the C++ runtime library. |
545 | ||
e5e809f4 JL |
546 | The standard C++ runtime library for GNU CC is called @samp{libstdc++}. |
547 | An obsolescent library @samp{libg++} may also be available, but it's | |
548 | necessary only for older software that hasn't been converted yet; if | |
549 | you don't know whether you need @samp{libg++} then you probably don't | |
550 | need it. | |
551 | ||
552 | Here's one way to build and install @samp{libstdc++} for GNU CC: | |
c85f7c16 JL |
553 | |
554 | @itemize @bullet | |
555 | @item | |
556 | Build and install GNU CC, so that invoking @samp{gcc} obtains the GNU CC | |
557 | that was just built. | |
558 | ||
559 | @item | |
560 | Obtain a copy of a compatible @samp{libstdc++} distribution. For | |
561 | example, the @samp{libstdc++-2.8.0.tar.gz} distribution should be | |
562 | compatible with GCC 2.8.0. GCC distributors normally distribute | |
563 | @samp{libstdc++} as well. | |
564 | ||
565 | @item | |
566 | Set the @samp{CXX} environment variable to @samp{gcc} while running the | |
567 | @samp{libstdc++} distribution's @file{configure} command. Use the same | |
568 | @file{configure} options that you used when you invoked GCC's | |
569 | @file{configure} command. | |
570 | ||
571 | @item | |
572 | Invoke @samp{make} to build the C++ runtime. | |
573 | ||
574 | @item | |
575 | Invoke @samp{make install} to install the C++ runtime. | |
576 | ||
577 | @end itemize | |
578 | ||
579 | To summarize, after building and installing GNU CC, invoke the following | |
580 | shell commands in the topmost directory of the C++ library distribution. | |
581 | For @var{configure-options}, use the same options that | |
582 | you used to configure GNU CC. | |
583 | ||
584 | @example | |
585 | $ CXX=gcc ./configure @var{configure-options} | |
586 | $ make | |
587 | $ make install | |
588 | @end example | |
2284f91b DE |
589 | |
590 | @item | |
591 | GNU CC includes a runtime library for Objective-C because it is an | |
592 | integral part of the language. You can find the files associated with | |
593 | the library in the subdirectory @file{objc}. The GNU Objective-C | |
594 | Runtime Library requires header files for the target's C library in | |
595 | order to be compiled,and also requires the header files for the target's | |
596 | thread library if you want thread support. @xref{Cross Headers, | |
597 | Cross-Compilers and Header Files, Cross-Compilers and Header Files}, for | |
598 | discussion about header files issues for cross-compilation. | |
599 | ||
600 | When you run @file{configure}, it picks the appropriate Objective-C | |
601 | thread implementation file for the target platform. In some situations, | |
602 | you may wish to choose a different back-end as some platforms support | |
603 | multiple thread implementations or you may wish to disable thread | |
604 | support completely. You do this by specifying a value for the | |
605 | @var{OBJC_THREAD_FILE} makefile variable on the command line when you | |
606 | run make, for example: | |
607 | ||
608 | @smallexample | |
609 | make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2" OBJC_THREAD_FILE=thr-single | |
610 | @end smallexample | |
611 | ||
612 | @noindent | |
613 | Below is a list of the currently available back-ends. | |
614 | ||
615 | @itemize @bullet | |
616 | @item thr-single | |
617 | Disable thread support, should work for all platforms. | |
618 | @item thr-decosf1 | |
619 | DEC OSF/1 thread support. | |
620 | @item thr-irix | |
621 | SGI IRIX thread support. | |
622 | @item thr-mach | |
623 | Generic MACH thread support, known to work on NEXTSTEP. | |
624 | @item thr-os2 | |
625 | IBM OS/2 thread support. | |
626 | @item thr-posix | |
627 | Generix POSIX thread support. | |
628 | @item thr-pthreads | |
629 | PCThreads on Linux-based GNU systems. | |
630 | @item thr-solaris | |
631 | SUN Solaris thread support. | |
632 | @item thr-win32 | |
633 | Microsoft Win32 API thread support. | |
634 | @end itemize | |
635 | @end enumerate | |
636 | ||
637 | @node Configurations | |
638 | @section Configurations Supported by GNU CC | |
639 | @cindex configurations supported by GNU CC | |
640 | ||
641 | Here are the possible CPU types: | |
642 | ||
643 | @quotation | |
644 | @c gmicro, alliant, spur and tahoe omitted since they don't work. | |
645 | 1750a, a29k, alpha, arm, c@var{n}, clipper, dsp16xx, elxsi, h8300, | |
65376d28 | 646 | hppa1.0, hppa1.1, i370, i386, i486, i586, i860, i960, m32r, m68000, m68k, |
2284f91b DE |
647 | m88k, mips, mipsel, mips64, mips64el, ns32k, powerpc, powerpcle, |
648 | pyramid, romp, rs6000, sh, sparc, sparclite, sparc64, vax, we32k. | |
649 | @end quotation | |
650 | ||
651 | Here are the recognized company names. As you can see, customary | |
652 | abbreviations are used rather than the longer official names. | |
653 | ||
654 | @c What should be done about merlin, tek*, dolphin? | |
655 | @quotation | |
656 | acorn, alliant, altos, apollo, apple, att, bull, | |
657 | cbm, convergent, convex, crds, dec, dg, dolphin, | |
658 | elxsi, encore, harris, hitachi, hp, ibm, intergraph, isi, | |
659 | mips, motorola, ncr, next, ns, omron, plexus, | |
660 | sequent, sgi, sony, sun, tti, unicom, wrs. | |
661 | @end quotation | |
662 | ||
663 | The company name is meaningful only to disambiguate when the rest of | |
664 | the information supplied is insufficient. You can omit it, writing | |
665 | just @samp{@var{cpu}-@var{system}}, if it is not needed. For example, | |
666 | @samp{vax-ultrix4.2} is equivalent to @samp{vax-dec-ultrix4.2}. | |
667 | ||
668 | Here is a list of system types: | |
669 | ||
670 | @quotation | |
0c82f6bf | 671 | 386bsd, aix, acis, amigaos, aos, aout, aux, bosx, bsd, clix, coff, ctix, cxux, |
861bb6c1 | 672 | dgux, dynix, ebmon, ecoff, elf, esix, freebsd, hms, genix, gnu, linux-gnu, |
2284f91b DE |
673 | hiux, hpux, iris, irix, isc, luna, lynxos, mach, minix, msdos, mvs, |
674 | netbsd, newsos, nindy, ns, osf, osfrose, ptx, riscix, riscos, rtu, sco, sim, | |
675 | solaris, sunos, sym, sysv, udi, ultrix, unicos, uniplus, unos, vms, vsta, | |
676 | vxworks, winnt, xenix. | |
677 | @end quotation | |
678 | ||
679 | @noindent | |
680 | You can omit the system type; then @file{configure} guesses the | |
681 | operating system from the CPU and company. | |
682 | ||
683 | You can add a version number to the system type; this may or may not | |
684 | make a difference. For example, you can write @samp{bsd4.3} or | |
685 | @samp{bsd4.4} to distinguish versions of BSD. In practice, the version | |
686 | number is most needed for @samp{sysv3} and @samp{sysv4}, which are often | |
687 | treated differently. | |
688 | ||
689 | If you specify an impossible combination such as @samp{i860-dg-vms}, | |
690 | then you may get an error message from @file{configure}, or it may | |
691 | ignore part of the information and do the best it can with the rest. | |
692 | @file{configure} always prints the canonical name for the alternative | |
693 | that it used. GNU CC does not support all possible alternatives. | |
694 | ||
695 | Often a particular model of machine has a name. Many machine names are | |
696 | recognized as aliases for CPU/company combinations. Thus, the machine | |
697 | name @samp{sun3}, mentioned above, is an alias for @samp{m68k-sun}. | |
698 | Sometimes we accept a company name as a machine name, when the name is | |
699 | popularly used for a particular machine. Here is a table of the known | |
700 | machine names: | |
701 | ||
702 | @quotation | |
703 | 3300, 3b1, 3b@var{n}, 7300, altos3068, altos, | |
704 | apollo68, att-7300, balance, | |
705 | convex-c@var{n}, crds, decstation-3100, | |
706 | decstation, delta, encore, | |
707 | fx2800, gmicro, hp7@var{nn}, hp8@var{nn}, | |
708 | hp9k2@var{nn}, hp9k3@var{nn}, hp9k7@var{nn}, | |
709 | hp9k8@var{nn}, iris4d, iris, isi68, | |
710 | m3230, magnum, merlin, miniframe, | |
711 | mmax, news-3600, news800, news, next, | |
712 | pbd, pc532, pmax, powerpc, powerpcle, ps2, risc-news, | |
713 | rtpc, sun2, sun386i, sun386, sun3, | |
714 | sun4, symmetry, tower-32, tower. | |
715 | @end quotation | |
716 | ||
717 | @noindent | |
718 | Remember that a machine name specifies both the cpu type and the company | |
719 | name. | |
720 | If you want to install your own homemade configuration files, you can | |
721 | use @samp{local} as the company name to access them. If you use | |
722 | configuration @samp{@var{cpu}-local}, the configuration name | |
723 | without the cpu prefix | |
724 | is used to form the configuration file names. | |
725 | ||
726 | Thus, if you specify @samp{m68k-local}, configuration uses | |
727 | files @file{m68k.md}, @file{local.h}, @file{m68k.c}, | |
728 | @file{xm-local.h}, @file{t-local}, and @file{x-local}, all in the | |
729 | directory @file{config/m68k}. | |
730 | ||
731 | Here is a list of configurations that have special treatment or special | |
732 | things you must know: | |
733 | ||
734 | @table @samp | |
735 | @item 1750a-*-* | |
736 | MIL-STD-1750A processors. | |
737 | ||
738 | The MIL-STD-1750A cross configuration produces output for | |
739 | @code{as1750}, an assembler/linker available under the GNU Public | |
740 | License for the 1750A. @code{as1750} can be obtained at | |
741 | @emph{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}. | |
742 | A similarly licensed simulator for | |
743 | the 1750A is available from same address. | |
744 | ||
745 | You should ignore a fatal error during the building of libgcc (libgcc is | |
746 | not yet implemented for the 1750A.) | |
747 | ||
748 | The @code{as1750} assembler requires the file @file{ms1750.inc}, which is | |
749 | found in the directory @file{config/1750a}. | |
750 | ||
751 | GNU CC produced the same sections as the Fairchild F9450 C Compiler, | |
752 | namely: | |
753 | ||
754 | @table @code | |
755 | @item Normal | |
756 | The program code section. | |
757 | ||
758 | @item Static | |
759 | The read/write (RAM) data section. | |
760 | ||
761 | @item Konst | |
762 | The read-only (ROM) constants section. | |
763 | ||
764 | @item Init | |
765 | Initialization section (code to copy KREL to SREL). | |
766 | @end table | |
767 | ||
768 | The smallest addressable unit is 16 bits (BITS_PER_UNIT is 16). This | |
769 | means that type `char' is represented with a 16-bit word per character. | |
770 | The 1750A's "Load/Store Upper/Lower Byte" instructions are not used by | |
771 | GNU CC. | |
772 | ||
773 | @item alpha-*-osf1 | |
774 | Systems using processors that implement the DEC Alpha architecture and | |
775 | are running the DEC Unix (OSF/1) operating system, for example the DEC | |
956d6950 | 776 | Alpha AXP systems.CC.) |
2284f91b DE |
777 | |
778 | GNU CC writes a @samp{.verstamp} directive to the assembler output file | |
779 | unless it is built as a cross-compiler. It gets the version to use from | |
780 | the system header file @file{/usr/include/stamp.h}. If you install a | |
781 | new version of DEC Unix, you should rebuild GCC to pick up the new version | |
782 | stamp. | |
783 | ||
784 | Note that since the Alpha is a 64-bit architecture, cross-compilers from | |
785 | 32-bit machines will not generate code as efficient as that generated | |
786 | when the compiler is running on a 64-bit machine because many | |
787 | optimizations that depend on being able to represent a word on the | |
788 | target in an integral value on the host cannot be performed. Building | |
789 | cross-compilers on the Alpha for 32-bit machines has only been tested in | |
790 | a few cases and may not work properly. | |
791 | ||
792 | @code{make compare} may fail on old versions of DEC Unix unless you add | |
793 | @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the | |
794 | assembler input file is stored in the object file, and that makes | |
795 | comparison fail if it differs between the @code{stage1} and | |
796 | @code{stage2} compilations. The option @samp{-save-temps} forces a | |
797 | fixed name to be used for the assembler input file, instead of a | |
798 | randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps} | |
799 | unless the comparisons fail without that option. If you add | |
800 | @samp{-save-temps}, you will have to manually delete the @samp{.i} and | |
801 | @samp{.s} files after each series of compilations. | |
802 | ||
803 | GNU CC now supports both the native (ECOFF) debugging format used by DBX | |
804 | and GDB and an encapsulated STABS format for use only with GDB. See the | |
805 | discussion of the @samp{--with-stabs} option of @file{configure} above | |
806 | for more information on these formats and how to select them. | |
807 | ||
808 | There is a bug in DEC's assembler that produces incorrect line numbers | |
809 | for ECOFF format when the @samp{.align} directive is used. To work | |
810 | around this problem, GNU CC will not emit such alignment directives | |
811 | while writing ECOFF format debugging information even if optimization is | |
812 | being performed. Unfortunately, this has the very undesirable | |
813 | side-effect that code addresses when @samp{-O} is specified are | |
814 | different depending on whether or not @samp{-g} is also specified. | |
815 | ||
816 | To avoid this behavior, specify @samp{-gstabs+} and use GDB instead of | |
817 | DBX. DEC is now aware of this problem with the assembler and hopes to | |
818 | provide a fix shortly. | |
819 | ||
e9a25f70 JL |
820 | @item arc-*-elf |
821 | Argonaut ARC processor. | |
822 | This configuration is intended for embedded systems. | |
823 | ||
861bb6c1 | 824 | @item arm-*-aout |
2284f91b DE |
825 | Advanced RISC Machines ARM-family processors. These are often used in |
826 | embedded applications. There are no standard Unix configurations. | |
827 | This configuration corresponds to the basic instruction sequences and will | |
861bb6c1 | 828 | produce @file{a.out} format object modules. |
2284f91b DE |
829 | |
830 | You may need to make a variant of the file @file{arm.h} for your particular | |
831 | configuration. | |
832 | ||
861bb6c1 JL |
833 | @item arm-*-linuxaout |
834 | Any of the ARM family processors running the Linux-based GNU system with | |
835 | the @file{a.out} binary format (ELF is not yet supported). You must use | |
956d6950 | 836 | version 2.8.1.0.7 or later of the GNU/Linux binutils, which you can download |
861bb6c1 JL |
837 | from @file{sunsite.unc.edu:/pub/Linux/GCC} and other mirror sites for |
838 | Linux-based GNU systems. | |
839 | ||
2284f91b | 840 | @item arm-*-riscix |
14cd4d23 JL |
841 | The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix. |
842 | If you are running a version of RISC iX prior to 1.2 then you must | |
843 | specify the version number during configuration. Note that the | |
844 | assembler shipped with RISC iX does not support stabs debugging | |
845 | information; a new version of the assembler, with stabs support | |
846 | included, is now available from Acorn and via ftp | |
847 | @file{ftp.acorn.com:/pub/riscix/as+xterm.tar.Z}. To enable stabs | |
848 | debugging, pass @samp{--with-gnu-as} to configure. | |
849 | ||
850 | You will need to install GNU @file{sed} before you can run configure. | |
2284f91b DE |
851 | |
852 | @item a29k | |
853 | AMD Am29k-family processors. These are normally used in embedded | |
854 | applications. There are no standard Unix configurations. | |
855 | This configuration | |
856 | corresponds to AMD's standard calling sequence and binary interface | |
857 | and is compatible with other 29k tools. | |
858 | ||
859 | You may need to make a variant of the file @file{a29k.h} for your | |
860 | particular configuration. | |
861 | ||
862 | @item a29k-*-bsd | |
863 | AMD Am29050 used in a system running a variant of BSD Unix. | |
864 | ||
865 | @item decstation-* | |
e5e809f4 JL |
866 | MIPS-based DECstations can support three different personalities: |
867 | Ultrix, DEC OSF/1, and OSF/rose. (Alpha-based DECstation products have | |
868 | a configuration name beginning with @samp{alpha-dec}.) To configure GCC | |
869 | for these platforms use the following configurations: | |
2284f91b DE |
870 | |
871 | @table @samp | |
872 | @item decstation-ultrix | |
873 | Ultrix configuration. | |
874 | ||
875 | @item decstation-osf1 | |
876 | Dec's version of OSF/1. | |
877 | ||
878 | @item decstation-osfrose | |
879 | Open Software Foundation reference port of OSF/1 which uses the | |
880 | OSF/rose object file format instead of ECOFF. Normally, you | |
881 | would not select this configuration. | |
882 | @end table | |
883 | ||
884 | The MIPS C compiler needs to be told to increase its table size | |
885 | for switch statements with the @samp{-Wf,-XNg1500} option in | |
886 | order to compile @file{cp/parse.c}. If you use the @samp{-O2} | |
887 | optimization option, you also need to use @samp{-Olimit 3000}. | |
888 | Both of these options are automatically generated in the | |
889 | @file{Makefile} that the shell script @file{configure} builds. | |
890 | If you override the @code{CC} make variable and use the MIPS | |
891 | compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}. | |
892 | ||
893 | @item elxsi-elxsi-bsd | |
894 | The Elxsi's C compiler has known limitations that prevent it from | |
895 | compiling GNU C. Please contact @code{mrs@@cygnus.com} for more details. | |
896 | ||
897 | @item dsp16xx | |
898 | A port to the AT&T DSP1610 family of processors. | |
899 | ||
900 | @ignore | |
901 | @item fx80 | |
902 | Alliant FX/8 computer. Note that the standard installed C compiler in | |
903 | Concentrix 5.0 has a bug which prevent it from compiling GNU CC | |
904 | correctly. You can patch the compiler bug as follows: | |
905 | ||
906 | @smallexample | |
907 | cp /bin/pcc ./pcc | |
908 | adb -w ./pcc - << EOF | |
909 | 15f6?w 6610 | |
910 | EOF | |
911 | @end smallexample | |
912 | ||
913 | Then you must use the @samp{-ip12} option when compiling GNU CC | |
914 | with the patched compiler, as shown here: | |
915 | ||
916 | @smallexample | |
917 | make CC="./pcc -ip12" CFLAGS=-w | |
918 | @end smallexample | |
919 | ||
920 | Note also that Alliant's version of DBX does not manage to work with the | |
921 | output from GNU CC. | |
922 | @end ignore | |
923 | ||
924 | @item h8300-*-* | |
e9a25f70 JL |
925 | Hitachi H8/300 series of processors. |
926 | ||
2284f91b DE |
927 | The calling convention and structure layout has changed in release 2.6. |
928 | All code must be recompiled. The calling convention now passes the | |
929 | first three arguments in function calls in registers. Structures are no | |
930 | longer a multiple of 2 bytes. | |
931 | ||
932 | @item hppa*-*-* | |
933 | There are several variants of the HP-PA processor which run a variety | |
934 | of operating systems. GNU CC must be configured to use the correct | |
935 | processor type and operating system, or GNU CC will not function correctly. | |
936 | The easiest way to handle this problem is to @emph{not} specify a target | |
937 | when configuring GNU CC, the @file{configure} script will try to automatically | |
938 | determine the right processor type and operating system. | |
939 | ||
940 | @samp{-g} does not work on HP-UX, since that system uses a peculiar | |
941 | debugging format which GNU CC does not know about. However, @samp{-g} | |
942 | will work if you also use GAS and GDB in conjunction with GCC. We | |
943 | highly recommend using GAS for all HP-PA configurations. | |
944 | ||
945 | You should be using GAS-2.6 (or later) along with GDB-4.16 (or later). These | |
946 | can be retrieved from all the traditional GNU ftp archive sites. | |
947 | ||
948 | GAS will need to be installed into a directory before @code{/bin}, | |
949 | @code{/usr/bin}, and @code{/usr/ccs/bin} in your search path. You | |
950 | should install GAS before you build GNU CC. | |
951 | ||
952 | To enable debugging, you must configure GNU CC with the @samp{--with-gnu-as} | |
953 | option before building. | |
954 | ||
955 | @item i370-*-* | |
956 | This port is very preliminary and has many known bugs. We hope to | |
957 | have a higher-quality port for this machine soon. | |
958 | ||
861bb6c1 JL |
959 | @item i386-*-linux-gnuoldld |
960 | Use this configuration to generate @file{a.out} binaries on Linux-based | |
961 | GNU systems if you do not have gas/binutils version 2.5.2 or later | |
962 | installed. This is an obsolete configuration. | |
2284f91b | 963 | |
861bb6c1 JL |
964 | @item i386-*-linux-gnuaout |
965 | Use this configuration to generate @file{a.out} binaries on Linux-based | |
966 | GNU systems. This configuration is being superseded. You must use | |
967 | gas/binutils version 2.5.2 or later. | |
2284f91b | 968 | |
861bb6c1 JL |
969 | @item i386-*-linux-gnu |
970 | Use this configuration to generate ELF binaries on Linux-based GNU | |
971 | systems. You must use gas/binutils version 2.5.2 or later. | |
2284f91b DE |
972 | |
973 | @item i386-*-sco | |
974 | Compilation with RCC is recommended. Also, it may be a good idea to | |
975 | link with GNU malloc instead of the malloc that comes with the system. | |
976 | ||
977 | @item i386-*-sco3.2v4 | |
978 | Use this configuration for SCO release 3.2 version 4. | |
979 | ||
980 | @item i386-*-sco3.2v5* | |
e9a25f70 JL |
981 | Use this for the SCO OpenServer Release family including 5.0.0, 5.0.2, |
982 | 5.0.4, Internet FastStart 1.0, and Internet FastStart 1.1. | |
983 | ||
3a0120cf JL |
984 | GNU CC can generate either ELF or COFF binaries. ELF is the default. |
985 | To get COFF output, you must specify @samp{-mcoff}) on the command line. | |
986 | ||
987 | For 5.0.0 and 5.0.2, you must install TLS597 from ftp.sco.com/TLS. | |
988 | 5.0.4 and later do not require this patch. | |
2284f91b | 989 | |
e9a25f70 JL |
990 | @emph{NOTE:} You must follow the instructions about invoking |
991 | @samp{make bootstrap} because the native OpenServer compiler builds | |
992 | a @file{cc1plus} that will not correctly parse many valid C++ programs. | |
993 | You must do a @samp{make bootstrap} if you are building with the native | |
994 | compiler. | |
995 | ||
2284f91b DE |
996 | @item i386-*-isc |
997 | It may be a good idea to link with GNU malloc instead of the malloc that | |
998 | comes with the system. | |
999 | ||
1000 | In ISC version 4.1, @file{sed} core dumps when building | |
1001 | @file{deduced.h}. Use the version of @file{sed} from version 4.0. | |
1002 | ||
1003 | @item i386-*-esix | |
1004 | It may be good idea to link with GNU malloc instead of the malloc that | |
1005 | comes with the system. | |
1006 | ||
1007 | @item i386-ibm-aix | |
1008 | You need to use GAS version 2.1 or later, and LD from | |
1009 | GNU binutils version 2.2 or later. | |
1010 | ||
1011 | @item i386-sequent-bsd | |
956d6950 | 1012 | Go to the Berkeley universe before compiling. |
2284f91b DE |
1013 | |
1014 | @item i386-sequent-ptx1* | |
1015 | Sequent DYNIX/ptx 1.x. | |
1016 | ||
1017 | @item i386-sequent-ptx2* | |
1018 | Sequent DYNIX/ptx 2.x. | |
1019 | ||
1020 | @item i386-sun-sunos4 | |
1021 | You may find that you need another version of GNU CC to begin | |
1022 | bootstrapping with, since the current version when built with the | |
1023 | system's own compiler seems to get an infinite loop compiling part of | |
1024 | @file{libgcc2.c}. GNU CC version 2 compiled with GNU CC (any version) | |
1025 | seems not to have this problem. | |
1026 | ||
1027 | See @ref{Sun Install}, for information on installing GNU CC on Sun | |
1028 | systems. | |
1029 | ||
1030 | @item i[345]86-*-winnt3.5 | |
e9a25f70 | 1031 | This version requires a GAS that has not yet been released. Until it |
2284f91b DE |
1032 | is, you can get a prebuilt binary version via anonymous ftp from |
1033 | @file{cs.washington.edu:pub/gnat} or @file{cs.nyu.edu:pub/gnat}. You | |
1034 | must also use the Microsoft header files from the Windows NT 3.5 SDK. | |
1035 | Find these on the CDROM in the @file{/mstools/h} directory dated 9/4/94. You | |
1036 | must use a fixed version of Microsoft linker made especially for NT 3.5, | |
1037 | which is also is available on the NT 3.5 SDK CDROM. If you do not have | |
1038 | this linker, can you also use the linker from Visual C/C++ 1.0 or 2.0. | |
1039 | ||
1040 | Installing GNU CC for NT builds a wrapper linker, called @file{ld.exe}, | |
1041 | which mimics the behaviour of Unix @file{ld} in the specification of | |
1042 | libraries (@samp{-L} and @samp{-l}). @file{ld.exe} looks for both Unix | |
1043 | and Microsoft named libraries. For example, if you specify | |
1044 | @samp{-lfoo}, @file{ld.exe} will look first for @file{libfoo.a} | |
1045 | and then for @file{foo.lib}. | |
1046 | ||
1047 | You may install GNU CC for Windows NT in one of two ways, depending on | |
1048 | whether or not you have a Unix-like shell and various Unix-like | |
1049 | utilities. | |
1050 | ||
1051 | @enumerate | |
1052 | @item | |
1053 | If you do not have a Unix-like shell and few Unix-like utilities, you | |
1054 | will use a DOS style batch script called @file{configure.bat}. Invoke | |
1055 | it as @code{configure winnt} from an MSDOS console window or from the | |
1056 | program manager dialog box. @file{configure.bat} assumes you have | |
1057 | already installed and have in your path a Unix-like @file{sed} program | |
1058 | which is used to create a working @file{Makefile} from @file{Makefile.in}. | |
1059 | ||
1060 | @file{Makefile} uses the Microsoft Nmake program maintenance utility and | |
1061 | the Visual C/C++ V8.00 compiler to build GNU CC. You need only have the | |
1062 | utilities @file{sed} and @file{touch} to use this installation method, | |
1063 | which only automatically builds the compiler itself. You must then | |
1064 | examine what @file{fixinc.winnt} does, edit the header files by hand and | |
1065 | build @file{libgcc.a} manually. | |
1066 | ||
1067 | @item | |
1068 | The second type of installation assumes you are running a Unix-like | |
1069 | shell, have a complete suite of Unix-like utilities in your path, and | |
1070 | have a previous version of GNU CC already installed, either through | |
1071 | building it via the above installation method or acquiring a pre-built | |
1072 | binary. In this case, use the @file{configure} script in the normal | |
1073 | fashion. | |
1074 | @end enumerate | |
1075 | ||
1076 | @item i860-intel-osf1 | |
1077 | This is the Paragon. | |
1078 | @ifset INSTALLONLY | |
1079 | If you have version 1.0 of the operating system, you need to take | |
1080 | special steps to build GNU CC due to peculiarities of the system. Newer | |
1081 | system versions have no problem. See the section `Installation Problems' | |
1082 | in the GNU CC Manual. | |
1083 | @end ifset | |
1084 | @ifclear INSTALLONLY | |
1085 | If you have version 1.0 of the operating system, | |
1086 | see @ref{Installation Problems}, for special things you need to do to | |
1087 | compensate for peculiarities in the system. | |
1088 | @end ifclear | |
1089 | ||
1090 | @item *-lynx-lynxos | |
1091 | LynxOS 2.2 and earlier comes with GNU CC 1.x already installed as | |
1092 | @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}. | |
1093 | You can tell GNU CC to use the GNU assembler and linker, by specifying | |
1094 | @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce | |
1095 | COFF format object files and executables; otherwise GNU CC will use the | |
861bb6c1 | 1096 | installed tools, which produce @file{a.out} format executables. |
2284f91b | 1097 | |
65376d28 | 1098 | @item m32r-*-elf |
e9a25f70 JL |
1099 | Mitsubishi M32R processor. |
1100 | This configuration is intended for embedded systems. | |
65376d28 | 1101 | |
2284f91b DE |
1102 | @item m68000-hp-bsd |
1103 | HP 9000 series 200 running BSD. Note that the C compiler that comes | |
03340df4 | 1104 | with this system cannot compile GNU CC; contact @code{law@@cygnus.com} |
2284f91b DE |
1105 | to get binaries of GNU CC for bootstrapping. |
1106 | ||
1107 | @item m68k-altos | |
1108 | Altos 3068. You must use the GNU assembler, linker and debugger. | |
1109 | Also, you must fix a kernel bug. Details in the file @file{README.ALTOS}. | |
1110 | ||
1111 | @item m68k-apple-aux | |
1112 | Apple Macintosh running A/UX. | |
1113 | You may configure GCC to use either the system assembler and | |
1114 | linker or the GNU assembler and linker. You should use the GNU configuration | |
1115 | if you can, especially if you also want to use GNU C++. You enabled | |
1116 | that configuration with + the @samp{--with-gnu-as} and @samp{--with-gnu-ld} | |
1117 | options to @code{configure}. | |
1118 | ||
1119 | Note the C compiler that comes | |
e5e809f4 | 1120 | with this system cannot compile GNU CC. You can find binaries of GNU CC |
2284f91b DE |
1121 | for bootstrapping on @code{jagubox.gsfc.nasa.gov}. |
1122 | You will also a patched version of @file{/bin/ld} there that | |
1123 | raises some of the arbitrary limits found in the original. | |
1124 | ||
1125 | @item m68k-att-sysv | |
1126 | AT&T 3b1, a.k.a. 7300 PC. Special procedures are needed to compile GNU | |
1127 | CC with this machine's standard C compiler, due to bugs in that | |
1128 | compiler. You can bootstrap it more easily with | |
1129 | previous versions of GNU CC if you have them. | |
1130 | ||
1131 | Installing GNU CC on the 3b1 is difficult if you do not already have | |
1132 | GNU CC running, due to bugs in the installed C compiler. However, | |
1133 | the following procedure might work. We are unable to test it. | |
1134 | ||
1135 | @enumerate | |
1136 | @item | |
e9a25f70 | 1137 | Comment out the @samp{#include "config.h"} line near the start of |
2284f91b DE |
1138 | @file{cccp.c} and do @samp{make cpp}. This makes a preliminary version |
1139 | of GNU cpp. | |
1140 | ||
1141 | @item | |
1142 | Save the old @file{/lib/cpp} and copy the preliminary GNU cpp to that | |
1143 | file name. | |
1144 | ||
1145 | @item | |
1146 | Undo your change in @file{cccp.c}, or reinstall the original version, | |
1147 | and do @samp{make cpp} again. | |
1148 | ||
1149 | @item | |
1150 | Copy this final version of GNU cpp into @file{/lib/cpp}. | |
1151 | ||
1152 | @findex obstack_free | |
1153 | @item | |
1154 | Replace every occurrence of @code{obstack_free} in the file | |
1155 | @file{tree.c} with @code{_obstack_free}. | |
1156 | ||
1157 | @item | |
1158 | Run @code{make} to get the first-stage GNU CC. | |
1159 | ||
1160 | @item | |
1161 | Reinstall the original version of @file{/lib/cpp}. | |
1162 | ||
1163 | @item | |
1164 | Now you can compile GNU CC with itself and install it in the normal | |
1165 | fashion. | |
1166 | @end enumerate | |
1167 | ||
1168 | @item m68k-bull-sysv | |
1169 | Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GNU CC works | |
1170 | either with native assembler or GNU assembler. You can use | |
1171 | GNU assembler with native coff generation by providing @samp{--with-gnu-as} to | |
1172 | the configure script or use GNU assembler with dbx-in-coff encapsulation | |
1173 | by providing @samp{--with-gnu-as --stabs}. For any problem with native | |
1174 | assembler or for availability of the DPX/2 port of GAS, contact | |
1175 | @code{F.Pierresteguy@@frcl.bull.fr}. | |
1176 | ||
1177 | @item m68k-crds-unox | |
1178 | Use @samp{configure unos} for building on Unos. | |
1179 | ||
1180 | The Unos assembler is named @code{casm} instead of @code{as}. For some | |
1181 | strange reason linking @file{/bin/as} to @file{/bin/casm} changes the | |
1182 | behavior, and does not work. So, when installing GNU CC, you should | |
1183 | install the following script as @file{as} in the subdirectory where | |
1184 | the passes of GCC are installed: | |
1185 | ||
1186 | @example | |
1187 | #!/bin/sh | |
1188 | casm $* | |
1189 | @end example | |
1190 | ||
1191 | The default Unos library is named @file{libunos.a} instead of | |
1192 | @file{libc.a}. To allow GNU CC to function, either change all | |
1193 | references to @samp{-lc} in @file{gcc.c} to @samp{-lunos} or link | |
1194 | @file{/lib/libc.a} to @file{/lib/libunos.a}. | |
1195 | ||
1196 | @cindex @code{alloca}, for Unos | |
1197 | When compiling GNU CC with the standard compiler, to overcome bugs in | |
1198 | the support of @code{alloca}, do not use @samp{-O} when making stage 2. | |
1199 | Then use the stage 2 compiler with @samp{-O} to make the stage 3 | |
1200 | compiler. This compiler will have the same characteristics as the usual | |
1201 | stage 2 compiler on other systems. Use it to make a stage 4 compiler | |
1202 | and compare that with stage 3 to verify proper compilation. | |
1203 | ||
1204 | (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in | |
1205 | the comments there will make the above paragraph superfluous. Please | |
1206 | inform us of whether this works.) | |
1207 | ||
1208 | Unos uses memory segmentation instead of demand paging, so you will need | |
1209 | a lot of memory. 5 Mb is barely enough if no other tasks are running. | |
1210 | If linking @file{cc1} fails, try putting the object files into a library | |
1211 | and linking from that library. | |
1212 | ||
1213 | @item m68k-hp-hpux | |
1214 | HP 9000 series 300 or 400 running HP-UX. HP-UX version 8.0 has a bug in | |
1215 | the assembler that prevents compilation of GNU CC. To fix it, get patch | |
1216 | PHCO_4484 from HP. | |
1217 | ||
1218 | In addition, if you wish to use gas @samp{--with-gnu-as} you must use | |
1219 | gas version 2.1 or later, and you must use the GNU linker version 2.1 or | |
1220 | later. Earlier versions of gas relied upon a program which converted the | |
1221 | gas output into the native HP/UX format, but that program has not been | |
1222 | kept up to date. gdb does not understand that native HP/UX format, so | |
1223 | you must use gas if you wish to use gdb. | |
1224 | ||
1225 | @item m68k-sun | |
1226 | Sun 3. We do not provide a configuration file to use the Sun FPA by | |
1227 | default, because programs that establish signal handlers for floating | |
1228 | point traps inherently cannot work with the FPA. | |
1229 | ||
1230 | See @ref{Sun Install}, for information on installing GNU CC on Sun | |
1231 | systems. | |
1232 | ||
1233 | @item m88k-*-svr3 | |
1234 | Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port. | |
1235 | These systems tend to use the Green Hills C, revision 1.8.5, as the | |
1236 | standard C compiler. There are apparently bugs in this compiler that | |
1237 | result in object files differences between stage 2 and stage 3. If this | |
1238 | happens, make the stage 4 compiler and compare it to the stage 3 | |
1239 | compiler. If the stage 3 and stage 4 object files are identical, this | |
1240 | suggests you encountered a problem with the standard C compiler; the | |
1241 | stage 3 and 4 compilers may be usable. | |
1242 | ||
1243 | It is best, however, to use an older version of GNU CC for bootstrapping | |
1244 | if you have one. | |
1245 | ||
1246 | @item m88k-*-dgux | |
1247 | Motorola m88k running DG/UX. To build 88open BCS native or cross | |
1248 | compilers on DG/UX, specify the configuration name as | |
1249 | @samp{m88k-*-dguxbcs} and build in the 88open BCS software development | |
1250 | environment. To build ELF native or cross compilers on DG/UX, specify | |
1251 | @samp{m88k-*-dgux} and build in the DG/UX ELF development environment. | |
1252 | You set the software development environment by issuing | |
1253 | @samp{sde-target} command and specifying either @samp{m88kbcs} or | |
1254 | @samp{m88kdguxelf} as the operand. | |
1255 | ||
1256 | If you do not specify a configuration name, @file{configure} guesses the | |
1257 | configuration based on the current software development environment. | |
1258 | ||
1259 | @item m88k-tektronix-sysv3 | |
1260 | Tektronix XD88 running UTekV 3.2e. Do not turn on | |
1261 | optimization while building stage1 if you bootstrap with | |
1262 | the buggy Green Hills compiler. Also, The bundled LAI | |
1263 | System V NFS is buggy so if you build in an NFS mounted | |
1264 | directory, start from a fresh reboot, or avoid NFS all together. | |
1265 | Otherwise you may have trouble getting clean comparisons | |
1266 | between stages. | |
1267 | ||
1268 | @item mips-mips-bsd | |
1269 | MIPS machines running the MIPS operating system in BSD mode. It's | |
1270 | possible that some old versions of the system lack the functions | |
1271 | @code{memcpy}, @code{memcmp}, and @code{memset}. If your system lacks | |
1272 | these, you must remove or undo the definition of | |
1273 | @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}. | |
1274 | ||
1275 | The MIPS C compiler needs to be told to increase its table size | |
1276 | for switch statements with the @samp{-Wf,-XNg1500} option in | |
1277 | order to compile @file{cp/parse.c}. If you use the @samp{-O2} | |
1278 | optimization option, you also need to use @samp{-Olimit 3000}. | |
1279 | Both of these options are automatically generated in the | |
1280 | @file{Makefile} that the shell script @file{configure} builds. | |
1281 | If you override the @code{CC} make variable and use the MIPS | |
1282 | compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}. | |
1283 | ||
1284 | @item mips-mips-riscos* | |
1285 | The MIPS C compiler needs to be told to increase its table size | |
1286 | for switch statements with the @samp{-Wf,-XNg1500} option in | |
1287 | order to compile @file{cp/parse.c}. If you use the @samp{-O2} | |
1288 | optimization option, you also need to use @samp{-Olimit 3000}. | |
1289 | Both of these options are automatically generated in the | |
1290 | @file{Makefile} that the shell script @file{configure} builds. | |
1291 | If you override the @code{CC} make variable and use the MIPS | |
1292 | compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}. | |
1293 | ||
1294 | MIPS computers running RISC-OS can support four different | |
1295 | personalities: default, BSD 4.3, System V.3, and System V.4 | |
1296 | (older versions of RISC-OS don't support V.4). To configure GCC | |
1297 | for these platforms use the following configurations: | |
1298 | ||
1299 | @table @samp | |
1300 | @item mips-mips-riscos@code{rev} | |
1301 | Default configuration for RISC-OS, revision @code{rev}. | |
1302 | ||
1303 | @item mips-mips-riscos@code{rev}bsd | |
1304 | BSD 4.3 configuration for RISC-OS, revision @code{rev}. | |
1305 | ||
1306 | @item mips-mips-riscos@code{rev}sysv4 | |
1307 | System V.4 configuration for RISC-OS, revision @code{rev}. | |
1308 | ||
1309 | @item mips-mips-riscos@code{rev}sysv | |
1310 | System V.3 configuration for RISC-OS, revision @code{rev}. | |
1311 | @end table | |
1312 | ||
1313 | The revision @code{rev} mentioned above is the revision of | |
1314 | RISC-OS to use. You must reconfigure GCC when going from a | |
1315 | RISC-OS revision 4 to RISC-OS revision 5. This has the effect of | |
1316 | avoiding a linker | |
1317 | @ifclear INSTALLONLY | |
1318 | bug (see @ref{Installation Problems}, for more details). | |
1319 | @end ifclear | |
1320 | @ifset INSTALLONLY | |
1321 | bug. | |
1322 | @end ifset | |
1323 | ||
1324 | @item mips-sgi-* | |
1325 | In order to compile GCC on an SGI running IRIX 4, the "c.hdr.lib" | |
1326 | option must be installed from the CD-ROM supplied from Silicon Graphics. | |
1327 | This is found on the 2nd CD in release 4.0.1. | |
1328 | ||
1329 | In order to compile GCC on an SGI running IRIX 5, the "compiler_dev.hdr" | |
1330 | subsystem must be installed from the IDO CD-ROM supplied by Silicon | |
1331 | Graphics. | |
1332 | ||
1333 | @code{make compare} may fail on version 5 of IRIX unless you add | |
1334 | @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the | |
1335 | assembler input file is stored in the object file, and that makes | |
1336 | comparison fail if it differs between the @code{stage1} and | |
1337 | @code{stage2} compilations. The option @samp{-save-temps} forces a | |
1338 | fixed name to be used for the assembler input file, instead of a | |
1339 | randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps} | |
1340 | unless the comparisons fail without that option. If you do you | |
1341 | @samp{-save-temps}, you will have to manually delete the @samp{.i} and | |
1342 | @samp{.s} files after each series of compilations. | |
1343 | ||
1344 | The MIPS C compiler needs to be told to increase its table size | |
1345 | for switch statements with the @samp{-Wf,-XNg1500} option in | |
1346 | order to compile @file{cp/parse.c}. If you use the @samp{-O2} | |
1347 | optimization option, you also need to use @samp{-Olimit 3000}. | |
1348 | Both of these options are automatically generated in the | |
1349 | @file{Makefile} that the shell script @file{configure} builds. | |
1350 | If you override the @code{CC} make variable and use the MIPS | |
1351 | compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}. | |
1352 | ||
1353 | On Irix version 4.0.5F, and perhaps on some other versions as well, | |
1354 | there is an assembler bug that reorders instructions incorrectly. To | |
1355 | work around it, specify the target configuration | |
1356 | @samp{mips-sgi-irix4loser}. This configuration inhibits assembler | |
1357 | optimization. | |
1358 | ||
1359 | In a compiler configured with target @samp{mips-sgi-irix4}, you can turn | |
1360 | off assembler optimization by using the @samp{-noasmopt} option. This | |
1361 | compiler option passes the option @samp{-O0} to the assembler, to | |
1362 | inhibit reordering. | |
1363 | ||
1364 | The @samp{-noasmopt} option can be useful for testing whether a problem | |
1365 | is due to erroneous assembler reordering. Even if a problem does not go | |
1366 | away with @samp{-noasmopt}, it may still be due to assembler | |
1367 | reordering---perhaps GNU CC itself was miscompiled as a result. | |
1368 | ||
1369 | To enable debugging under Irix 5, you must use GNU as 2.5 or later, | |
1370 | and use the @samp{--with-gnu-as} configure option when configuring gcc. | |
1371 | GNU as is distributed as part of the binutils package. | |
1372 | ||
1373 | @item mips-sony-sysv | |
1374 | Sony MIPS NEWS. This works in NEWSOS 5.0.1, but not in 5.0.2 (which | |
1375 | uses ELF instead of COFF). Support for 5.0.2 will probably be provided | |
1376 | soon by volunteers. In particular, the linker does not like the | |
1377 | code generated by GCC when shared libraries are linked in. | |
1378 | ||
1379 | @item ns32k-encore | |
1380 | Encore ns32000 system. Encore systems are supported only under BSD. | |
1381 | ||
1382 | @item ns32k-*-genix | |
1383 | National Semiconductor ns32000 system. Genix has bugs in @code{alloca} | |
1384 | and @code{malloc}; you must get the compiled versions of these from GNU | |
1385 | Emacs. | |
1386 | ||
1387 | @item ns32k-sequent | |
956d6950 | 1388 | Go to the Berkeley universe before compiling. |
2284f91b DE |
1389 | |
1390 | @item ns32k-utek | |
1391 | UTEK ns32000 system (``merlin''). The C compiler that comes with this | |
1392 | system cannot compile GNU CC; contact @samp{tektronix!reed!mason} to get | |
1393 | binaries of GNU CC for bootstrapping. | |
1394 | ||
1395 | @item romp-*-aos | |
1396 | @itemx romp-*-mach | |
1397 | The only operating systems supported for the IBM RT PC are AOS and | |
1398 | MACH. GNU CC does not support AIX running on the RT. We recommend you | |
1399 | compile GNU CC with an earlier version of itself; if you compile GNU CC | |
1400 | with @code{hc}, the Metaware compiler, it will work, but you will get | |
1401 | mismatches between the stage 2 and stage 3 compilers in various files. | |
1402 | These errors are minor differences in some floating-point constants and | |
1403 | can be safely ignored; the stage 3 compiler is correct. | |
1404 | ||
1405 | @item rs6000-*-aix | |
1406 | @itemx powerpc-*-aix | |
1407 | Various early versions of each release of the IBM XLC compiler will not | |
1408 | bootstrap GNU CC. Symptoms include differences between the stage2 and | |
1409 | stage3 object files, and errors when compiling @file{libgcc.a} or | |
1410 | @file{enquire}. Known problematic releases include: xlc-1.2.1.8, | |
1411 | xlc-1.3.0.0 (distributed with AIX 3.2.5), and xlc-1.3.0.19. Both | |
1412 | xlc-1.2.1.28 and xlc-1.3.0.24 (PTF 432238) are known to produce working | |
1413 | versions of GNU CC, but most other recent releases correctly bootstrap | |
1da4b26c DE |
1414 | GNU CC. |
1415 | ||
1416 | Release 4.3.0 of AIX and ones prior to AIX 3.2.4 include a version of | |
2284f91b DE |
1417 | the IBM assembler which does not accept debugging directives: assembler |
1418 | updates are available as PTFs. Also, if you are using AIX 3.2.5 or | |
1419 | greater and the GNU assembler, you must have a version modified after | |
1420 | October 16th, 1995 in order for the GNU C compiler to build. See the | |
1da4b26c | 1421 | file @file{README.RS6000} for more details on any of these problems. |
2284f91b DE |
1422 | |
1423 | GNU CC does not yet support the 64-bit PowerPC instructions. | |
1424 | ||
1425 | Objective C does not work on this architecture because it makes assumptions | |
1426 | that are incompatible with the calling conventions. | |
1427 | ||
1428 | AIX on the RS/6000 provides support (NLS) for environments outside of | |
1429 | the United States. Compilers and assemblers use NLS to support | |
1430 | locale-specific representations of various objects including | |
1431 | floating-point numbers ("." vs "," for separating decimal fractions). | |
1432 | There have been problems reported where the library linked with GNU CC | |
1433 | does not produce the same floating-point formats that the assembler | |
1434 | accepts. If you have this problem, set the LANG environment variable to | |
1435 | "C" or "En_US". | |
1436 | ||
1437 | Due to changes in the way that GNU CC invokes the binder (linker) for AIX | |
1438 | 4.1, you may now receive warnings of duplicate symbols from the link step | |
1439 | that were not reported before. The assembly files generated by GNU CC for | |
1440 | AIX have always included multiple symbol definitions for certain global | |
1441 | variable and function declarations in the original program. The warnings | |
1442 | should not prevent the linker from producing a correct library or runnable | |
1443 | executable. | |
1444 | ||
1445 | By default, AIX 4.1 produces code that can be used on either Power or | |
1446 | PowerPC processors. | |
1447 | ||
1448 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1449 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1450 | ||
1451 | @item powerpc-*-elf | |
1452 | @itemx powerpc-*-sysv4 | |
1453 | PowerPC system in big endian mode, running System V.4. | |
1454 | ||
1455 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1456 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1457 | ||
861bb6c1 JL |
1458 | @item powerpc-*-linux-gnu |
1459 | PowerPC system in big endian mode, running the Linux-based GNU system. | |
2284f91b DE |
1460 | |
1461 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1462 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1463 | ||
1464 | @item powerpc-*-eabiaix | |
1465 | Embedded PowerPC system in big endian mode with -mcall-aix selected as | |
1466 | the default. | |
1467 | ||
1468 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1469 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1470 | ||
1471 | @item powerpc-*-eabisim | |
1472 | Embedded PowerPC system in big endian mode for use in running under the | |
1473 | PSIM simulator. | |
1474 | ||
1475 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1476 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1477 | ||
1478 | @item powerpc-*-eabi | |
1479 | Embedded PowerPC system in big endian mode. | |
1480 | ||
1481 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1482 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1483 | ||
1484 | @item powerpcle-*-elf | |
1485 | @itemx powerpcle-*-sysv4 | |
1486 | PowerPC system in little endian mode, running System V.4. | |
1487 | ||
1488 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1489 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1490 | ||
1491 | @item powerpcle-*-solaris2* | |
1492 | PowerPC system in little endian mode, running Solaris 2.5.1 or higher. | |
1493 | ||
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}. | |
1496 | Beta versions of the Sun 4.0 compiler do not seem to be able to build | |
1497 | GNU CC correctly. There are also problems with the host assembler and | |
1498 | linker that are fixed by using the GNU versions of these tools. | |
1499 | ||
1500 | @item powerpcle-*-eabisim | |
1501 | Embedded PowerPC system in little endian mode for use in running under | |
1502 | the PSIM simulator. | |
1503 | ||
1504 | @itemx powerpcle-*-eabi | |
1505 | Embedded PowerPC system in little endian mode. | |
1506 | ||
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}. | |
1509 | ||
1510 | @item powerpcle-*-winnt | |
1511 | @itemx powerpcle-*-pe | |
1512 | PowerPC system in little endian mode running Windows NT. | |
1513 | ||
1514 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1515 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1516 | ||
1517 | @item vax-dec-ultrix | |
1518 | Don't try compiling with Vax C (@code{vcc}). It produces incorrect code | |
1519 | in some cases (for example, when @code{alloca} is used). | |
1520 | ||
1521 | Meanwhile, compiling @file{cp/parse.c} with pcc does not work because of | |
1522 | an internal table size limitation in that compiler. To avoid this | |
1523 | problem, compile just the GNU C compiler first, and use it to recompile | |
1524 | building all the languages that you want to run. | |
1525 | ||
1526 | @item sparc-sun-* | |
2284f91b DE |
1527 | See @ref{Sun Install}, for information on installing GNU CC on Sun |
1528 | systems. | |
1529 | ||
1530 | @item vax-dec-vms | |
1531 | See @ref{VMS Install}, for details on how to install GNU CC on VMS. | |
1532 | ||
1533 | @item we32k-*-* | |
1534 | These computers are also known as the 3b2, 3b5, 3b20 and other similar | |
1535 | names. (However, the 3b1 is actually a 68000; see | |
1536 | @ref{Configurations}.) | |
1537 | ||
1538 | Don't use @samp{-g} when compiling with the system's compiler. The | |
1539 | system's linker seems to be unable to handle such a large program with | |
1540 | debugging information. | |
1541 | ||
1542 | The system's compiler runs out of capacity when compiling @file{stmt.c} | |
1543 | in GNU CC. You can work around this by building @file{cpp} in GNU CC | |
1544 | first, then use that instead of the system's preprocessor with the | |
1545 | system's C compiler to compile @file{stmt.c}. Here is how: | |
1546 | ||
1547 | @smallexample | |
1548 | mv /lib/cpp /lib/cpp.att | |
1549 | cp cpp /lib/cpp.gnu | |
1550 | echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp | |
1551 | chmod +x /lib/cpp | |
1552 | @end smallexample | |
1553 | ||
1554 | The system's compiler produces bad code for some of the GNU CC | |
1555 | optimization files. So you must build the stage 2 compiler without | |
1556 | optimization. Then build a stage 3 compiler with optimization. | |
1557 | That executable should work. Here are the necessary commands: | |
1558 | ||
1559 | @smallexample | |
1560 | make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g" | |
1561 | make stage2 | |
1562 | make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O" | |
1563 | @end smallexample | |
1564 | ||
1565 | You may need to raise the ULIMIT setting to build a C++ compiler, | |
1566 | as the file @file{cc1plus} is larger than one megabyte. | |
1567 | @end table | |
1568 | ||
1569 | @node Other Dir | |
1570 | @section Compilation in a Separate Directory | |
1571 | @cindex other directory, compilation in | |
1572 | @cindex compilation in a separate directory | |
1573 | @cindex separate directory, compilation in | |
1574 | ||
1575 | If you wish to build the object files and executables in a directory | |
1576 | other than the one containing the source files, here is what you must | |
1577 | do differently: | |
1578 | ||
1579 | @enumerate | |
1580 | @item | |
1581 | Make sure you have a version of Make that supports the @code{VPATH} | |
1582 | feature. (GNU Make supports it, as do Make versions on most BSD | |
1583 | systems.) | |
1584 | ||
1585 | @item | |
1586 | If you have ever run @file{configure} in the source directory, you must undo | |
1587 | the configuration. Do this by running: | |
1588 | ||
1589 | @example | |
1590 | make distclean | |
1591 | @end example | |
1592 | ||
1593 | @item | |
1594 | Go to the directory in which you want to build the compiler before | |
1595 | running @file{configure}: | |
1596 | ||
1597 | @example | |
1598 | mkdir gcc-sun3 | |
1599 | cd gcc-sun3 | |
1600 | @end example | |
1601 | ||
1602 | On systems that do not support symbolic links, this directory must be | |
1603 | on the same file system as the source code directory. | |
1604 | ||
1605 | @item | |
1606 | Specify where to find @file{configure} when you run it: | |
1607 | ||
1608 | @example | |
1609 | ../gcc/configure @dots{} | |
1610 | @end example | |
1611 | ||
1612 | This also tells @code{configure} where to find the compiler sources; | |
1613 | @code{configure} takes the directory from the file name that was used to | |
1614 | invoke it. But if you want to be sure, you can specify the source | |
1615 | directory with the @samp{--srcdir} option, like this: | |
1616 | ||
1617 | @example | |
1618 | ../gcc/configure --srcdir=../gcc @var{other options} | |
1619 | @end example | |
1620 | ||
1621 | The directory you specify with @samp{--srcdir} need not be the same | |
1622 | as the one that @code{configure} is found in. | |
1623 | @end enumerate | |
1624 | ||
1625 | Now, you can run @code{make} in that directory. You need not repeat the | |
1626 | configuration steps shown above, when ordinary source files change. You | |
1627 | must, however, run @code{configure} again when the configuration files | |
1628 | change, if your system does not support symbolic links. | |
1629 | ||
1630 | @node Cross-Compiler | |
1631 | @section Building and Installing a Cross-Compiler | |
1632 | @cindex cross-compiler, installation | |
1633 | ||
1634 | GNU CC can function as a cross-compiler for many machines, but not all. | |
1635 | ||
1636 | @itemize @bullet | |
1637 | @item | |
1638 | Cross-compilers for the Mips as target using the Mips assembler | |
1639 | currently do not work, because the auxiliary programs | |
1640 | @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on | |
1641 | anything but a Mips. It does work to cross compile for a Mips | |
1642 | if you use the GNU assembler and linker. | |
1643 | ||
1644 | @item | |
1645 | Cross-compilers between machines with different floating point formats | |
1646 | have not all been made to work. GNU CC now has a floating point | |
1647 | emulator with which these can work, but each target machine description | |
1648 | needs to be updated to take advantage of it. | |
1649 | ||
1650 | @item | |
1651 | Cross-compilation between machines of different word sizes is | |
1652 | somewhat problematic and sometimes does not work. | |
1653 | @end itemize | |
1654 | ||
1655 | Since GNU CC generates assembler code, you probably need a | |
1656 | cross-assembler that GNU CC can run, in order to produce object files. | |
1657 | If you want to link on other than the target machine, you need a | |
1658 | cross-linker as well. You also need header files and libraries suitable | |
1659 | for the target machine that you can install on the host machine. | |
1660 | ||
1661 | @menu | |
1662 | * Steps of Cross:: Using a cross-compiler involves several steps | |
1663 | that may be carried out on different machines. | |
1664 | * Configure Cross:: Configuring a cross-compiler. | |
1665 | * Tools and Libraries:: Where to put the linker and assembler, and the C library. | |
1666 | * Cross Headers:: Finding and installing header files | |
1667 | for a cross-compiler. | |
1668 | * Cross Runtime:: Supplying arithmetic runtime routines (@file{libgcc1.a}). | |
1669 | * Build Cross:: Actually compiling the cross-compiler. | |
1670 | @end menu | |
1671 | ||
1672 | @node Steps of Cross | |
1673 | @subsection Steps of Cross-Compilation | |
1674 | ||
1675 | To compile and run a program using a cross-compiler involves several | |
1676 | steps: | |
1677 | ||
1678 | @itemize @bullet | |
1679 | @item | |
1680 | Run the cross-compiler on the host machine to produce assembler files | |
1681 | for the target machine. This requires header files for the target | |
1682 | machine. | |
1683 | ||
1684 | @item | |
1685 | Assemble the files produced by the cross-compiler. You can do this | |
1686 | either with an assembler on the target machine, or with a | |
1687 | cross-assembler on the host machine. | |
1688 | ||
1689 | @item | |
1690 | Link those files to make an executable. You can do this either with a | |
1691 | linker on the target machine, or with a cross-linker on the host | |
1692 | machine. Whichever machine you use, you need libraries and certain | |
1693 | startup files (typically @file{crt@dots{}.o}) for the target machine. | |
1694 | @end itemize | |
1695 | ||
1696 | It is most convenient to do all of these steps on the same host machine, | |
1697 | since then you can do it all with a single invocation of GNU CC. This | |
1698 | requires a suitable cross-assembler and cross-linker. For some targets, | |
1699 | the GNU assembler and linker are available. | |
1700 | ||
1701 | @node Configure Cross | |
1702 | @subsection Configuring a Cross-Compiler | |
1703 | ||
1704 | To build GNU CC as a cross-compiler, you start out by running | |
1705 | @file{configure}. Use the @samp{--target=@var{target}} to specify the | |
1706 | target type. If @file{configure} was unable to correctly identify the | |
1707 | system you are running on, also specify the @samp{--build=@var{build}} | |
1708 | option. For example, here is how to configure for a cross-compiler that | |
1709 | produces code for an HP 68030 system running BSD on a system that | |
1710 | @file{configure} can correctly identify: | |
1711 | ||
1712 | @smallexample | |
1713 | ./configure --target=m68k-hp-bsd4.3 | |
1714 | @end smallexample | |
1715 | ||
1716 | @node Tools and Libraries | |
1717 | @subsection Tools and Libraries for a Cross-Compiler | |
1718 | ||
1719 | If you have a cross-assembler and cross-linker available, you should | |
1720 | install them now. Put them in the directory | |
1721 | @file{/usr/local/@var{target}/bin}. Here is a table of the tools | |
1722 | you should put in this directory: | |
1723 | ||
1724 | @table @file | |
1725 | @item as | |
1726 | This should be the cross-assembler. | |
1727 | ||
1728 | @item ld | |
1729 | This should be the cross-linker. | |
1730 | ||
1731 | @item ar | |
1732 | This should be the cross-archiver: a program which can manipulate | |
1733 | archive files (linker libraries) in the target machine's format. | |
1734 | ||
1735 | @item ranlib | |
1736 | This should be a program to construct a symbol table in an archive file. | |
1737 | @end table | |
1738 | ||
1739 | The installation of GNU CC will find these programs in that directory, | |
1740 | and copy or link them to the proper place to for the cross-compiler to | |
1741 | find them when run later. | |
1742 | ||
1743 | The easiest way to provide these files is to build the Binutils package | |
1744 | and GAS. Configure them with the same @samp{--host} and @samp{--target} | |
1745 | options that you use for configuring GNU CC, then build and install | |
1746 | them. They install their executables automatically into the proper | |
1747 | directory. Alas, they do not support all the targets that GNU CC | |
1748 | supports. | |
1749 | ||
1750 | If you want to install libraries to use with the cross-compiler, such as | |
1751 | a standard C library, put them in the directory | |
1752 | @file{/usr/local/@var{target}/lib}; installation of GNU CC copies | |
1753 | all the files in that subdirectory into the proper place for GNU CC to | |
1754 | find them and link with them. Here's an example of copying some | |
1755 | libraries from a target machine: | |
1756 | ||
1757 | @example | |
1758 | ftp @var{target-machine} | |
1759 | lcd /usr/local/@var{target}/lib | |
1760 | cd /lib | |
1761 | get libc.a | |
1762 | cd /usr/lib | |
1763 | get libg.a | |
1764 | get libm.a | |
1765 | quit | |
1766 | @end example | |
1767 | ||
1768 | @noindent | |
1769 | The precise set of libraries you'll need, and their locations on | |
1770 | the target machine, vary depending on its operating system. | |
1771 | ||
1772 | @cindex start files | |
1773 | Many targets require ``start files'' such as @file{crt0.o} and | |
1774 | @file{crtn.o} which are linked into each executable; these too should be | |
1775 | placed in @file{/usr/local/@var{target}/lib}. There may be several | |
1776 | alternatives for @file{crt0.o}, for use with profiling or other | |
1777 | compilation options. Check your target's definition of | |
1778 | @code{STARTFILE_SPEC} to find out what start files it uses. | |
1779 | Here's an example of copying these files from a target machine: | |
1780 | ||
1781 | @example | |
1782 | ftp @var{target-machine} | |
1783 | lcd /usr/local/@var{target}/lib | |
1784 | prompt | |
1785 | cd /lib | |
1786 | mget *crt*.o | |
1787 | cd /usr/lib | |
1788 | mget *crt*.o | |
1789 | quit | |
1790 | @end example | |
1791 | ||
1792 | @node Cross Runtime | |
1793 | @subsection @file{libgcc.a} and Cross-Compilers | |
1794 | ||
1795 | Code compiled by GNU CC uses certain runtime support functions | |
1796 | implicitly. Some of these functions can be compiled successfully with | |
1797 | GNU CC itself, but a few cannot be. These problem functions are in the | |
1798 | source file @file{libgcc1.c}; the library made from them is called | |
1799 | @file{libgcc1.a}. | |
1800 | ||
1801 | When you build a native compiler, these functions are compiled with some | |
1802 | other compiler--the one that you use for bootstrapping GNU CC. | |
1803 | Presumably it knows how to open code these operations, or else knows how | |
1804 | to call the run-time emulation facilities that the machine comes with. | |
1805 | But this approach doesn't work for building a cross-compiler. The | |
1806 | compiler that you use for building knows about the host system, not the | |
1807 | target system. | |
1808 | ||
1809 | So, when you build a cross-compiler you have to supply a suitable | |
1810 | library @file{libgcc1.a} that does the job it is expected to do. | |
1811 | ||
1812 | To compile @file{libgcc1.c} with the cross-compiler itself does not | |
1813 | work. The functions in this file are supposed to implement arithmetic | |
1814 | operations that GNU CC does not know how to open code for your target | |
1815 | machine. If these functions are compiled with GNU CC itself, they | |
1816 | will compile into infinite recursion. | |
1817 | ||
1818 | On any given target, most of these functions are not needed. If GNU CC | |
1819 | can open code an arithmetic operation, it will not call these functions | |
1820 | to perform the operation. It is possible that on your target machine, | |
1821 | none of these functions is needed. If so, you can supply an empty | |
1822 | library as @file{libgcc1.a}. | |
1823 | ||
1824 | Many targets need library support only for multiplication and division. | |
1825 | If you are linking with a library that contains functions for | |
1826 | multiplication and division, you can tell GNU CC to call them directly | |
1827 | by defining the macros @code{MULSI3_LIBCALL}, and the like. These | |
1828 | macros need to be defined in the target description macro file. For | |
1829 | some targets, they are defined already. This may be sufficient to | |
1830 | avoid the need for libgcc1.a; if so, you can supply an empty library. | |
1831 | ||
1832 | Some targets do not have floating point instructions; they need other | |
1833 | functions in @file{libgcc1.a}, which do floating arithmetic. | |
1834 | Recent versions of GNU CC have a file which emulates floating point. | |
1835 | With a certain amount of work, you should be able to construct a | |
1836 | floating point emulator that can be used as @file{libgcc1.a}. Perhaps | |
1837 | future versions will contain code to do this automatically and | |
1838 | conveniently. That depends on whether someone wants to implement it. | |
1839 | ||
1840 | Some embedded targets come with all the necessary @file{libgcc1.a} | |
1841 | routines written in C or assembler. These targets build | |
1842 | @file{libgcc1.a} automatically and you do not need to do anything | |
1843 | special for them. Other embedded targets do not need any | |
1844 | @file{libgcc1.a} routines since all the necessary operations are | |
1845 | supported by the hardware. | |
1846 | ||
1847 | If your target system has another C compiler, you can configure GNU CC | |
1848 | as a native compiler on that machine, build just @file{libgcc1.a} with | |
1849 | @samp{make libgcc1.a} on that machine, and use the resulting file with | |
1850 | the cross-compiler. To do this, execute the following on the target | |
1851 | machine: | |
1852 | ||
1853 | @example | |
1854 | cd @var{target-build-dir} | |
1855 | ./configure --host=sparc --target=sun3 | |
1856 | make libgcc1.a | |
1857 | @end example | |
1858 | ||
1859 | @noindent | |
1860 | And then this on the host machine: | |
1861 | ||
1862 | @example | |
1863 | ftp @var{target-machine} | |
1864 | binary | |
1865 | cd @var{target-build-dir} | |
1866 | get libgcc1.a | |
1867 | quit | |
1868 | @end example | |
1869 | ||
1870 | Another way to provide the functions you need in @file{libgcc1.a} is to | |
1871 | define the appropriate @code{perform_@dots{}} macros for those | |
1872 | functions. If these definitions do not use the C arithmetic operators | |
1873 | that they are meant to implement, you should be able to compile them | |
1874 | with the cross-compiler you are building. (If these definitions already | |
1875 | exist for your target file, then you are all set.) | |
1876 | ||
1877 | To build @file{libgcc1.a} using the perform macros, use | |
1878 | @samp{LIBGCC1=libgcc1.a OLDCC=./xgcc} when building the compiler. | |
1879 | Otherwise, you should place your replacement library under the name | |
1880 | @file{libgcc1.a} in the directory in which you will build the | |
1881 | cross-compiler, before you run @code{make}. | |
1882 | ||
1883 | @node Cross Headers | |
1884 | @subsection Cross-Compilers and Header Files | |
1885 | ||
1886 | If you are cross-compiling a standalone program or a program for an | |
1887 | embedded system, then you may not need any header files except the few | |
1888 | that are part of GNU CC (and those of your program). However, if you | |
1889 | intend to link your program with a standard C library such as | |
1890 | @file{libc.a}, then you probably need to compile with the header files | |
1891 | that go with the library you use. | |
1892 | ||
1893 | The GNU C compiler does not come with these files, because (1) they are | |
1894 | system-specific, and (2) they belong in a C library, not in a compiler. | |
1895 | ||
1896 | If the GNU C library supports your target machine, then you can get the | |
1897 | header files from there (assuming you actually use the GNU library when | |
1898 | you link your program). | |
1899 | ||
1900 | If your target machine comes with a C compiler, it probably comes with | |
1901 | suitable header files also. If you make these files accessible from the host | |
1902 | machine, the cross-compiler can use them also. | |
1903 | ||
1904 | Otherwise, you're on your own in finding header files to use when | |
1905 | cross-compiling. | |
1906 | ||
1907 | When you have found suitable header files, put them in the directory | |
1908 | @file{/usr/local/@var{target}/include}, before building the cross | |
1909 | compiler. Then installation will run fixincludes properly and install | |
1910 | the corrected versions of the header files where the compiler will use | |
1911 | them. | |
1912 | ||
1913 | Provide the header files before you build the cross-compiler, because | |
1914 | the build stage actually runs the cross-compiler to produce parts of | |
1915 | @file{libgcc.a}. (These are the parts that @emph{can} be compiled with | |
1916 | GNU CC.) Some of them need suitable header files. | |
1917 | ||
1918 | Here's an example showing how to copy the header files from a target | |
1919 | machine. On the target machine, do this: | |
1920 | ||
1921 | @example | |
1922 | (cd /usr/include; tar cf - .) > tarfile | |
1923 | @end example | |
1924 | ||
1925 | Then, on the host machine, do this: | |
1926 | ||
1927 | @example | |
1928 | ftp @var{target-machine} | |
1929 | lcd /usr/local/@var{target}/include | |
1930 | get tarfile | |
1931 | quit | |
1932 | tar xf tarfile | |
1933 | @end example | |
1934 | ||
1935 | @node Build Cross | |
1936 | @subsection Actually Building the Cross-Compiler | |
1937 | ||
1938 | Now you can proceed just as for compiling a single-machine compiler | |
1939 | through the step of building stage 1. If you have not provided some | |
1940 | sort of @file{libgcc1.a}, then compilation will give up at the point | |
1941 | where it needs that file, printing a suitable error message. If you | |
1942 | do provide @file{libgcc1.a}, then building the compiler will automatically | |
1943 | compile and link a test program called @file{libgcc1-test}; if you get | |
1944 | errors in the linking, it means that not all of the necessary routines | |
1945 | in @file{libgcc1.a} are available. | |
1946 | ||
1947 | You must provide the header file @file{float.h}. One way to do this is | |
1948 | to compile @file{enquire} and run it on your target machine. The job of | |
1949 | @file{enquire} is to run on the target machine and figure out by | |
1950 | experiment the nature of its floating point representation. | |
1951 | @file{enquire} records its findings in the header file @file{float.h}. | |
1952 | If you can't produce this file by running @file{enquire} on the target | |
1953 | machine, then you will need to come up with a suitable @file{float.h} in | |
1954 | some other way (or else, avoid using it in your programs). | |
1955 | ||
1956 | Do not try to build stage 2 for a cross-compiler. It doesn't work to | |
1957 | rebuild GNU CC as a cross-compiler using the cross-compiler, because | |
1958 | that would produce a program that runs on the target machine, not on the | |
1959 | host. For example, if you compile a 386-to-68030 cross-compiler with | |
1960 | itself, the result will not be right either for the 386 (because it was | |
1961 | compiled into 68030 code) or for the 68030 (because it was configured | |
1962 | for a 386 as the host). If you want to compile GNU CC into 68030 code, | |
1963 | whether you compile it on a 68030 or with a cross-compiler on a 386, you | |
1964 | must specify a 68030 as the host when you configure it. | |
1965 | ||
1966 | To install the cross-compiler, use @samp{make install}, as usual. | |
1967 | ||
1968 | @node Sun Install | |
1969 | @section Installing GNU CC on the Sun | |
1970 | @cindex Sun installation | |
1971 | @cindex installing GNU CC on the Sun | |
1972 | ||
e5e809f4 | 1973 | On Solaris, do not use the linker or other tools in |
2284f91b DE |
1974 | @file{/usr/ucb} to build GNU CC. Use @code{/usr/ccs/bin}. |
1975 | ||
e5e809f4 JL |
1976 | If the assembler reports @samp{Error: misaligned data} when bootstrapping, |
1977 | you are probably using an obsolete version of the GNU assembler. Upgrade | |
1978 | to the latest version of GNU @code{binutils}, or use the Solaris assembler. | |
1979 | ||
2284f91b DE |
1980 | Make sure the environment variable @code{FLOAT_OPTION} is not set when |
1981 | you compile @file{libgcc.a}. If this option were set to @code{f68881} | |
1982 | when @file{libgcc.a} is compiled, the resulting code would demand to be | |
1983 | linked with a special startup file and would not link properly without | |
1984 | special pains. | |
1985 | ||
1986 | @cindex @code{alloca}, for SunOS | |
1987 | There is a bug in @code{alloca} in certain versions of the Sun library. | |
1988 | To avoid this bug, install the binaries of GNU CC that were compiled by | |
1989 | GNU CC. They use @code{alloca} as a built-in function and never the one | |
1990 | in the library. | |
1991 | ||
1992 | Some versions of the Sun compiler crash when compiling GNU CC. The | |
1993 | problem is a segmentation fault in cpp. This problem seems to be due to | |
1994 | the bulk of data in the environment variables. You may be able to avoid | |
1995 | it by using the following command to compile GNU CC with Sun CC: | |
1996 | ||
1997 | @example | |
1998 | make CC="TERMCAP=x OBJS=x LIBFUNCS=x STAGESTUFF=x cc" | |
1999 | @end example | |
2000 | ||
2001 | SunOS 4.1.3 and 4.1.3_U1 have bugs that can cause intermittent core | |
2002 | dumps when compiling GNU CC. A common symptom is an | |
2003 | internal compiler error which does not recur if you run it again. | |
2004 | To fix the problem, install Sun recommended patch 100726 (for SunOS 4.1.3) | |
2005 | or 101508 (for SunOS 4.1.3_U1), or upgrade to a later SunOS release. | |
2006 | ||
2007 | @node VMS Install | |
2008 | @section Installing GNU CC on VMS | |
2009 | @cindex VMS installation | |
2010 | @cindex installing GNU CC on VMS | |
2011 | ||
2012 | The VMS version of GNU CC is distributed in a backup saveset containing | |
2013 | both source code and precompiled binaries. | |
2014 | ||
2015 | To install the @file{gcc} command so you can use the compiler easily, in | |
2016 | the same manner as you use the VMS C compiler, you must install the VMS CLD | |
2017 | file for GNU CC as follows: | |
2018 | ||
2019 | @enumerate | |
2020 | @item | |
2021 | Define the VMS logical names @samp{GNU_CC} and @samp{GNU_CC_INCLUDE} | |
2022 | to point to the directories where the GNU CC executables | |
2023 | (@file{gcc-cpp.exe}, @file{gcc-cc1.exe}, etc.) and the C include files are | |
2024 | kept respectively. This should be done with the commands:@refill | |
2025 | ||
2026 | @smallexample | |
2027 | $ assign /system /translation=concealed - | |
2028 | disk:[gcc.] gnu_cc | |
2029 | $ assign /system /translation=concealed - | |
2030 | disk:[gcc.include.] gnu_cc_include | |
2031 | @end smallexample | |
2032 | ||
2033 | @noindent | |
2034 | with the appropriate disk and directory names. These commands can be | |
2035 | placed in your system startup file so they will be executed whenever | |
2036 | the machine is rebooted. You may, if you choose, do this via the | |
2037 | @file{GCC_INSTALL.COM} script in the @file{[GCC]} directory. | |
2038 | ||
2039 | @item | |
2040 | Install the @file{GCC} command with the command line: | |
2041 | ||
2042 | @smallexample | |
2043 | $ set command /table=sys$common:[syslib]dcltables - | |
2044 | /output=sys$common:[syslib]dcltables gnu_cc:[000000]gcc | |
2045 | $ install replace sys$common:[syslib]dcltables | |
2046 | @end smallexample | |
2047 | ||
2048 | @item | |
2049 | To install the help file, do the following: | |
2050 | ||
2051 | @smallexample | |
2052 | $ library/help sys$library:helplib.hlb gcc.hlp | |
2053 | @end smallexample | |
2054 | ||
2055 | @noindent | |
2056 | Now you can invoke the compiler with a command like @samp{gcc /verbose | |
2057 | file.c}, which is equivalent to the command @samp{gcc -v -c file.c} in | |
2058 | Unix. | |
2059 | @end enumerate | |
2060 | ||
2061 | If you wish to use GNU C++ you must first install GNU CC, and then | |
2062 | perform the following steps: | |
2063 | ||
2064 | @enumerate | |
2065 | @item | |
2066 | Define the VMS logical name @samp{GNU_GXX_INCLUDE} to point to the | |
2067 | directory where the preprocessor will search for the C++ header files. | |
2068 | This can be done with the command:@refill | |
2069 | ||
2070 | @smallexample | |
2071 | $ assign /system /translation=concealed - | |
2072 | disk:[gcc.gxx_include.] gnu_gxx_include | |
2073 | @end smallexample | |
2074 | ||
2075 | @noindent | |
2076 | with the appropriate disk and directory name. If you are going to be | |
c85f7c16 JL |
2077 | using a C++ runtime library, this is where its install procedure will install |
2078 | its header files. | |
2284f91b DE |
2079 | |
2080 | @item | |
2081 | Obtain the file @file{gcc-cc1plus.exe}, and place this in the same | |
2082 | directory that @file{gcc-cc1.exe} is kept. | |
2083 | ||
2084 | The GNU C++ compiler can be invoked with a command like @samp{gcc /plus | |
2085 | /verbose file.cc}, which is equivalent to the command @samp{g++ -v -c | |
2086 | file.cc} in Unix. | |
2087 | @end enumerate | |
2088 | ||
2089 | We try to put corresponding binaries and sources on the VMS distribution | |
2090 | tape. But sometimes the binaries will be from an older version than the | |
2091 | sources, because we don't always have time to update them. (Use the | |
2092 | @samp{/version} option to determine the version number of the binaries and | |
2093 | compare it with the source file @file{version.c} to tell whether this is | |
2094 | so.) In this case, you should use the binaries you get to recompile the | |
2095 | sources. If you must recompile, here is how: | |
2096 | ||
2097 | @enumerate | |
2098 | @item | |
2099 | Execute the command procedure @file{vmsconfig.com} to set up the files | |
2100 | @file{tm.h}, @file{config.h}, @file{aux-output.c}, and @file{md.}, and | |
2101 | to create files @file{tconfig.h} and @file{hconfig.h}. This procedure | |
2102 | also creates several linker option files used by @file{make-cc1.com} and | |
2103 | a data file used by @file{make-l2.com}.@refill | |
2104 | ||
2105 | @smallexample | |
2106 | $ @@vmsconfig.com | |
2107 | @end smallexample | |
2108 | ||
2109 | @item | |
2110 | Setup the logical names and command tables as defined above. In | |
2111 | addition, define the VMS logical name @samp{GNU_BISON} to point at the | |
2112 | to the directories where the Bison executable is kept. This should be | |
2113 | done with the command:@refill | |
2114 | ||
2115 | @smallexample | |
2116 | $ assign /system /translation=concealed - | |
2117 | disk:[bison.] gnu_bison | |
2118 | @end smallexample | |
2119 | ||
2120 | You may, if you choose, use the @file{INSTALL_BISON.COM} script in the | |
2121 | @file{[BISON]} directory. | |
2122 | ||
2123 | @item | |
2124 | Install the @samp{BISON} command with the command line:@refill | |
2125 | ||
2126 | @smallexample | |
2127 | $ set command /table=sys$common:[syslib]dcltables - | |
2128 | /output=sys$common:[syslib]dcltables - | |
2129 | gnu_bison:[000000]bison | |
2130 | $ install replace sys$common:[syslib]dcltables | |
2131 | @end smallexample | |
2132 | ||
2133 | @item | |
2134 | Type @samp{@@make-gcc} to recompile everything (alternatively, submit | |
2135 | the file @file{make-gcc.com} to a batch queue). If you wish to build | |
2136 | the GNU C++ compiler as well as the GNU CC compiler, you must first edit | |
2137 | @file{make-gcc.com} and follow the instructions that appear in the | |
2138 | comments.@refill | |
2139 | ||
2140 | @item | |
2141 | In order to use GCC, you need a library of functions which GCC compiled code | |
2142 | will call to perform certain tasks, and these functions are defined in the | |
2143 | file @file{libgcc2.c}. To compile this you should use the command procedure | |
2144 | @file{make-l2.com}, which will generate the library @file{libgcc2.olb}. | |
2145 | @file{libgcc2.olb} should be built using the compiler built from | |
2146 | the same distribution that @file{libgcc2.c} came from, and | |
2147 | @file{make-gcc.com} will automatically do all of this for you. | |
2148 | ||
2149 | To install the library, use the following commands:@refill | |
2150 | ||
2151 | @smallexample | |
2152 | $ library gnu_cc:[000000]gcclib/delete=(new,eprintf) | |
2153 | $ library gnu_cc:[000000]gcclib/delete=L_* | |
2154 | $ library libgcc2/extract=*/output=libgcc2.obj | |
2155 | $ library gnu_cc:[000000]gcclib libgcc2.obj | |
2156 | @end smallexample | |
2157 | ||
2158 | The first command simply removes old modules that will be replaced with | |
2159 | modules from @file{libgcc2} under different module names. The modules | |
2160 | @code{new} and @code{eprintf} may not actually be present in your | |
2161 | @file{gcclib.olb}---if the VMS librarian complains about those modules | |
2162 | not being present, simply ignore the message and continue on with the | |
2163 | next command. The second command removes the modules that came from the | |
2164 | previous version of the library @file{libgcc2.c}. | |
2165 | ||
2166 | Whenever you update the compiler on your system, you should also update the | |
2167 | library with the above procedure. | |
2168 | ||
2169 | @item | |
2170 | You may wish to build GCC in such a way that no files are written to the | |
2171 | directory where the source files reside. An example would be the when | |
2172 | the source files are on a read-only disk. In these cases, execute the | |
2173 | following DCL commands (substituting your actual path names): | |
2174 | ||
2175 | @smallexample | |
2176 | $ assign dua0:[gcc.build_dir.]/translation=concealed, - | |
2177 | dua1:[gcc.source_dir.]/translation=concealed gcc_build | |
2178 | $ set default gcc_build:[000000] | |
2179 | @end smallexample | |
2180 | ||
2181 | @noindent | |
2182 | where the directory @file{dua1:[gcc.source_dir]} contains the source | |
2183 | code, and the directory @file{dua0:[gcc.build_dir]} is meant to contain | |
2184 | all of the generated object files and executables. Once you have done | |
2185 | this, you can proceed building GCC as described above. (Keep in mind | |
2186 | that @file{gcc_build} is a rooted logical name, and thus the device | |
2187 | names in each element of the search list must be an actual physical | |
2188 | device name rather than another rooted logical name). | |
2189 | ||
2190 | @item | |
2191 | @strong{If you are building GNU CC with a previous version of GNU CC, | |
2192 | you also should check to see that you have the newest version of the | |
2193 | assembler}. In particular, GNU CC version 2 treats global constant | |
2194 | variables slightly differently from GNU CC version 1, and GAS version | |
2195 | 1.38.1 does not have the patches required to work with GCC version 2. | |
2196 | If you use GAS 1.38.1, then @code{extern const} variables will not have | |
2197 | the read-only bit set, and the linker will generate warning messages | |
2198 | about mismatched psect attributes for these variables. These warning | |
2199 | messages are merely a nuisance, and can safely be ignored. | |
2200 | ||
2201 | If you are compiling with a version of GNU CC older than 1.33, specify | |
2202 | @samp{/DEFINE=("inline=")} as an option in all the compilations. This | |
2203 | requires editing all the @code{gcc} commands in @file{make-cc1.com}. | |
2204 | (The older versions had problems supporting @code{inline}.) Once you | |
2205 | have a working 1.33 or newer GNU CC, you can change this file back. | |
2206 | ||
2207 | @item | |
2208 | If you want to build GNU CC with the VAX C compiler, you will need to | |
2209 | make minor changes in @file{make-cccp.com} and @file{make-cc1.com} | |
2210 | to choose alternate definitions of @code{CC}, @code{CFLAGS}, and | |
2211 | @code{LIBS}. See comments in those files. However, you must | |
2212 | also have a working version of the GNU assembler (GNU as, aka GAS) as | |
2213 | it is used as the back-end for GNU CC to produce binary object modules | |
2214 | and is not included in the GNU CC sources. GAS is also needed to | |
2215 | compile @file{libgcc2} in order to build @file{gcclib} (see above); | |
2216 | @file{make-l2.com} expects to be able to find it operational in | |
2217 | @file{gnu_cc:[000000]gnu-as.exe}. | |
2218 | ||
2219 | To use GNU CC on VMS, you need the VMS driver programs | |
2220 | @file{gcc.exe}, @file{gcc.com}, and @file{gcc.cld}. They are | |
2221 | distributed with the VMS binaries (@file{gcc-vms}) rather than the | |
2222 | GNU CC sources. GAS is also included in @file{gcc-vms}, as is Bison. | |
2223 | ||
2224 | Once you have successfully built GNU CC with VAX C, you should use the | |
2225 | resulting compiler to rebuild itself. Before doing this, be sure to | |
2226 | restore the @code{CC}, @code{CFLAGS}, and @code{LIBS} definitions in | |
2227 | @file{make-cccp.com} and @file{make-cc1.com}. The second generation | |
2228 | compiler will be able to take advantage of many optimizations that must | |
2229 | be suppressed when building with other compilers. | |
2230 | @end enumerate | |
2231 | ||
2232 | Under previous versions of GNU CC, the generated code would occasionally | |
2233 | give strange results when linked with the sharable @file{VAXCRTL} library. | |
2234 | Now this should work. | |
2235 | ||
2236 | Even with this version, however, GNU CC itself should not be linked with | |
2237 | the sharable @file{VAXCRTL}. The version of @code{qsort} in | |
2238 | @file{VAXCRTL} has a bug (known to be present in VMS versions V4.6 | |
2239 | through V5.5) which causes the compiler to fail. | |
2240 | ||
2241 | The executables are generated by @file{make-cc1.com} and | |
2242 | @file{make-cccp.com} use the object library version of @file{VAXCRTL} in | |
2243 | order to make use of the @code{qsort} routine in @file{gcclib.olb}. If | |
2244 | you wish to link the compiler executables with the shareable image | |
2245 | version of @file{VAXCRTL}, you should edit the file @file{tm.h} (created | |
2246 | by @file{vmsconfig.com}) to define the macro @code{QSORT_WORKAROUND}. | |
2247 | ||
2248 | @code{QSORT_WORKAROUND} is always defined when GNU CC is compiled with | |
2249 | VAX C, to avoid a problem in case @file{gcclib.olb} is not yet | |
2250 | available. | |
2251 | ||
2252 | @node Collect2 | |
2253 | @section @code{collect2} | |
2254 | ||
23851576 JL |
2255 | GNU CC uses a utility called @code{collect2} on nearly all systems to arrange |
2256 | to call various initialization functions at start time. | |
2284f91b DE |
2257 | |
2258 | The program @code{collect2} works by linking the program once and | |
2259 | looking through the linker output file for symbols with particular names | |
2260 | indicating they are constructor functions. If it finds any, it | |
2261 | creates a new temporary @samp{.c} file containing a table of them, | |
2262 | compiles it, and links the program a second time including that file. | |
2263 | ||
2264 | @findex __main | |
2265 | @cindex constructors, automatic calls | |
2266 | The actual calls to the constructors are carried out by a subroutine | |
2267 | called @code{__main}, which is called (automatically) at the beginning | |
2268 | of the body of @code{main} (provided @code{main} was compiled with GNU | |
2269 | CC). Calling @code{__main} is necessary, even when compiling C code, to | |
2270 | allow linking C and C++ object code together. (If you use | |
2271 | @samp{-nostdlib}, you get an unresolved reference to @code{__main}, | |
2272 | since it's defined in the standard GCC library. Include @samp{-lgcc} at | |
2273 | the end of your compiler command line to resolve this reference.) | |
2274 | ||
2275 | The program @code{collect2} is installed as @code{ld} in the directory | |
2276 | where the passes of the compiler are installed. When @code{collect2} | |
2277 | needs to find the @emph{real} @code{ld}, it tries the following file | |
2278 | names: | |
2279 | ||
2280 | @itemize @bullet | |
2281 | @item | |
2282 | @file{real-ld} in the directories listed in the compiler's search | |
2283 | directories. | |
2284 | ||
2285 | @item | |
2286 | @file{real-ld} in the directories listed in the environment variable | |
2287 | @code{PATH}. | |
2288 | ||
2289 | @item | |
2290 | The file specified in the @code{REAL_LD_FILE_NAME} configuration macro, | |
2291 | if specified. | |
2292 | ||
2293 | @item | |
2294 | @file{ld} in the compiler's search directories, except that | |
2295 | @code{collect2} will not execute itself recursively. | |
2296 | ||
2297 | @item | |
2298 | @file{ld} in @code{PATH}. | |
2299 | @end itemize | |
2300 | ||
2301 | ``The compiler's search directories'' means all the directories where | |
2302 | @code{gcc} searches for passes of the compiler. This includes | |
2303 | directories that you specify with @samp{-B}. | |
2304 | ||
2305 | Cross-compilers search a little differently: | |
2306 | ||
2307 | @itemize @bullet | |
2308 | @item | |
2309 | @file{real-ld} in the compiler's search directories. | |
2310 | ||
2311 | @item | |
2312 | @file{@var{target}-real-ld} in @code{PATH}. | |
2313 | ||
2314 | @item | |
2315 | The file specified in the @code{REAL_LD_FILE_NAME} configuration macro, | |
2316 | if specified. | |
2317 | ||
2318 | @item | |
2319 | @file{ld} in the compiler's search directories. | |
2320 | ||
2321 | @item | |
2322 | @file{@var{target}-ld} in @code{PATH}. | |
2323 | @end itemize | |
2324 | ||
2325 | @code{collect2} explicitly avoids running @code{ld} using the file name | |
2326 | under which @code{collect2} itself was invoked. In fact, it remembers | |
2327 | up a list of such names---in case one copy of @code{collect2} finds | |
2328 | another copy (or version) of @code{collect2} installed as @code{ld} in a | |
2329 | second place in the search path. | |
2330 | ||
2331 | @code{collect2} searches for the utilities @code{nm} and @code{strip} | |
2332 | using the same algorithm as above for @code{ld}. | |
2333 | ||
2334 | @node Header Dirs | |
2335 | @section Standard Header File Directories | |
2336 | ||
2337 | @code{GCC_INCLUDE_DIR} means the same thing for native and cross. It is | |
2338 | where GNU CC stores its private include files, and also where GNU CC | |
2339 | stores the fixed include files. A cross compiled GNU CC runs | |
2340 | @code{fixincludes} on the header files in @file{$(tooldir)/include}. | |
2341 | (If the cross compilation header files need to be fixed, they must be | |
2342 | installed before GNU CC is built. If the cross compilation header files | |
2343 | are already suitable for ANSI C and GNU CC, nothing special need be | |
2344 | done). | |
2345 | ||
2346 | @code{GPLUS_INCLUDE_DIR} means the same thing for native and cross. It | |
c85f7c16 | 2347 | is where @code{g++} looks first for header files. The C++ library |
2284f91b DE |
2348 | installs only target independent header files in that directory. |
2349 | ||
2350 | @code{LOCAL_INCLUDE_DIR} is used only for a native compiler. It is | |
2351 | normally @file{/usr/local/include}. GNU CC searches this directory so | |
2352 | that users can install header files in @file{/usr/local/include}. | |
2353 | ||
2354 | @code{CROSS_INCLUDE_DIR} is used only for a cross compiler. GNU CC | |
2355 | doesn't install anything there. | |
2356 | ||
2357 | @code{TOOL_INCLUDE_DIR} is used for both native and cross compilers. It | |
2358 | is the place for other packages to install header files that GNU CC will | |
2359 | use. For a cross-compiler, this is the equivalent of | |
2360 | @file{/usr/include}. When you build a cross-compiler, | |
2361 | @code{fixincludes} processes any header files in this directory. |