1 @c This is for making the `INSTALL' file for the distribution.
2 @c Makeinfo ignores it when processing the file from the include.
6 @node Installation, Maintenance, Library Summary, Top
7 @c %MENU% How to install the GNU C Library
8 @appendix Installing @theglibc{}
10 Before you do anything else, you should read the FAQ at
11 @url{http://sourceware.org/glibc/wiki/FAQ}. It answers common
12 questions and describes problems you may experience with compilation
15 Features can be added to @theglibc{} via @dfn{add-on} bundles. These are
16 separate tar files, which you unpack into the top level of the source
17 tree. Then you give @code{configure} the @samp{--enable-add-ons} option
18 to activate them, and they will be compiled into the library.
20 You will need recent versions of several GNU tools: definitely GCC and
21 GNU Make, and possibly others. @xref{Tools for Compilation}, below.
24 * Configuring and compiling:: How to compile and test GNU libc.
25 * Running make install:: How to install it once you've got it
27 * Tools for Compilation:: You'll need these first.
28 * Linux:: Specific advice for GNU/Linux systems.
29 * Reporting Bugs:: So they'll get fixed.
32 @node Configuring and compiling
33 @appendixsec Configuring and compiling @theglibc{}
37 @Theglibc{} cannot be compiled in the source directory. You must build
38 it in a separate build directory. For example, if you have unpacked
39 the @glibcadj{} sources in @file{/src/gnu/glibc-@var{version}},
41 @file{/src/gnu/glibc-build} to put the object files in. This allows
42 removing the whole build directory in case an error occurs, which is
43 the safest way to get a fresh start and should always be done.
45 From your object directory, run the shell script @file{configure} located
46 at the top level of the source tree. In the scenario above, you'd type
49 $ ../glibc-@var{version}/configure @var{args@dots{}}
52 Please note that even though you're building in a separate build
53 directory, the compilation needs to modify a few files in the source
54 directory, especially some files in the manual subdirectory.
57 @code{configure} takes many options, but the only one that is usually
58 mandatory is @samp{--prefix}. This option tells @code{configure}
59 where you want @theglibc{} installed. This defaults to @file{/usr/local},
60 but the normal setting to install as the standard system library is
61 @samp{--prefix=/usr} for @gnulinuxsystems{} and @samp{--prefix=} (an
62 empty prefix) for @gnuhurdsystems{}.
64 It may also be useful to set the @var{CC} and @var{CFLAGS} variables in
65 the environment when running @code{configure}. @var{CC} selects the C
66 compiler that will be used, and @var{CFLAGS} sets optimization options
69 The following list describes all of the available options for
73 @item --prefix=@var{directory}
74 Install machine-independent data files in subdirectories of
75 @file{@var{directory}}. The default is to install in @file{/usr/local}.
77 @item --exec-prefix=@var{directory}
78 Install the library and other machine-dependent files in subdirectories
79 of @file{@var{directory}}. The default is to the @samp{--prefix}
80 directory if that option is specified, or @file{/usr/local} otherwise.
82 @item --with-headers=@var{directory}
83 Look for kernel header files in @var{directory}, not
84 @file{/usr/include}. @Theglibc{} needs information from the kernel's header
85 files describing the interface to the kernel. @Theglibc{} will normally
86 look in @file{/usr/include} for them,
87 but if you specify this option, it will look in @var{DIRECTORY} instead.
89 This option is primarily of use on a system where the headers in
90 @file{/usr/include} come from an older version of @theglibc{}. Conflicts can
91 occasionally happen in this case. You can also use this option if you want to
92 compile @theglibc{} with a newer set of kernel headers than the ones found in
95 @item --enable-add-ons[=@var{list}]
96 Specify add-on packages to include in the build. If this option is
97 specified with no list, it enables all the add-on packages it finds in
98 the main source directory; this is the default behavior. You may
99 specify an explicit list of add-ons to use in @var{list}, separated by
100 spaces or commas (if you use spaces, remember to quote them from the
101 shell). Each add-on in @var{list} can be an absolute directory name
102 or can be a directory name relative to the main source directory, or
103 relative to the build directory (that is, the current working directory).
104 For example, @samp{--enable-add-ons=nptl,../glibc-libidn-@var{version}}.
106 @item --enable-kernel=@var{version}
107 This option is currently only useful on @gnulinuxsystems{}. The
108 @var{version} parameter should have the form X.Y.Z and describes the
109 smallest version of the Linux kernel the generated library is expected
110 to support. The higher the @var{version} number is, the less
111 compatibility code is added, and the faster the code gets.
113 @item --with-binutils=@var{directory}
114 Use the binutils (assembler and linker) in @file{@var{directory}}, not
115 the ones the C compiler would default to. You can use this option if
116 the default binutils on your system cannot deal with all the constructs
117 in @theglibc{}. In that case, @code{configure} will detect the
118 problem and suppress these constructs, so that the library will still be
119 usable, but functionality may be lost---for example, you can't build a
120 shared libc with old binutils.
123 Use this option if your computer lacks hardware floating-point support
124 and your operating system does not emulate an FPU.
126 @c disable static doesn't work currently
127 @c @item --disable-static
128 @c Don't build static libraries. Static libraries aren't that useful these
129 @c days, but we recommend you build them in case you need them.
131 @item --disable-shared
132 Don't build shared libraries even if it is possible. Not all systems
133 support shared libraries; you need ELF support and (currently) the GNU
136 @item --disable-profile
137 Don't build libraries with profiling information. You may want to use
138 this option if you don't plan to do profiling.
140 @item --disable-versioning
141 Don't compile the shared libraries with symbol version information.
142 Doing this will make the resulting library incompatible with old
143 binaries, so it's not recommended.
145 @item --enable-static-nss
146 Compile static versions of the NSS (Name Service Switch) libraries.
147 This is not recommended because it defeats the purpose of NSS; a program
148 linked statically with the NSS libraries cannot be dynamically
149 reconfigured to use a different name database.
152 By default the C library is built with support for thread-local storage
153 if the used tools support it. By using @samp{--without-tls} this can be
154 prevented though there generally is no reason since it creates
155 compatibility problems.
157 @item --build=@var{build-system}
158 @itemx --host=@var{host-system}
159 These options are for cross-compiling. If you specify both options and
160 @var{build-system} is different from @var{host-system}, @code{configure}
161 will prepare to cross-compile @theglibc{} from @var{build-system} to be used
162 on @var{host-system}. You'll probably need the @samp{--with-headers}
163 option too, and you may have to override @var{configure}'s selection of
164 the compiler and/or binutils.
166 If you only specify @samp{--host}, @code{configure} will prepare for a
167 native compile but use what you specify instead of guessing what your
168 system is. This is most useful to change the CPU submodel. For example,
169 if @code{configure} guesses your machine as @code{i586-pc-linux-gnu} but
170 you want to compile a library for 386es, give
171 @samp{--host=i386-pc-linux-gnu} or just @samp{--host=i386-linux} and add
172 the appropriate compiler flags (@samp{-mcpu=i386} will do the trick) to
175 If you specify just @samp{--build}, @code{configure} will get confused.
178 To build the library and related programs, type @code{make}. This will
179 produce a lot of output, some of which may look like errors from
180 @code{make} but isn't. Look for error messages from @code{make}
181 containing @samp{***}. Those indicate that something is seriously wrong.
183 The compilation process can take a long time, depending on the
184 configuration and the speed of your machine. Some complex modules may
185 take a very long time to compile, as much as several minutes on slower
186 machines. Do not panic if the compiler appears to hang.
188 If you want to run a parallel make, simply pass the @samp{-j} option
189 with an appropriate numeric parameter to @code{make}. You need a recent
190 GNU @code{make} version, though.
192 To build and run test programs which exercise some of the library
193 facilities, type @code{make check}. If it does not complete
194 successfully, do not use the built library, and report a bug after
195 verifying that the problem is not already known. @xref{Reporting Bugs},
196 for instructions on reporting bugs. Note that some of the tests assume
197 they are not being run by @code{root}. We recommend you compile and
198 test @theglibc{} as an unprivileged user.
200 Before reporting bugs make sure there is no problem with your system.
201 The tests (and later installation) use some pre-existing files of the
202 system such as @file{/etc/passwd}, @file{/etc/nsswitch.conf} and others.
203 These files must all contain correct and sensible content.
205 To format the @cite{GNU C Library Reference Manual} for printing, type
206 @w{@code{make dvi}}. You need a working @TeX{} installation to do
207 this. The distribution builds the on-line formatted version of the
208 manual, as Info files, as part of the build process. You can build
209 them manually with @w{@code{make info}}.
211 The library has a number of special-purpose configuration parameters
212 which you can find in @file{Makeconfig}. These can be overwritten with
213 the file @file{configparms}. To change them, create a
214 @file{configparms} in your build directory and add values as appropriate
215 for your system. The file is included and parsed by @code{make} and has
216 to follow the conventions for makefiles.
218 It is easy to configure @theglibc{} for cross-compilation by
219 setting a few variables in @file{configparms}. Set @code{CC} to the
220 cross-compiler for the target you configured the library for; it is
221 important to use this same @code{CC} value when running
222 @code{configure}, like this: @samp{CC=@var{target}-gcc configure
223 @var{target}}. Set @code{BUILD_CC} to the compiler to use for programs
224 run on the build system as part of compiling the library. You may need to
225 set @code{AR} to cross-compiling versions of @code{ar}
226 if the native tools are not configured to work with
227 object files for the target you configured for.
230 @node Running make install
231 @appendixsec Installing the C Library
234 To install the library and its header files, and the Info files of the
235 manual, type @code{env LANGUAGE=C LC_ALL=C make install}. This will
236 build things, if necessary, before installing them; however, you should
237 still compile everything first. If you are installing @theglibc{} as your
238 primary C library, we recommend that you shut the system down to
239 single-user mode first, and reboot afterward. This minimizes the risk
240 of breaking things when the library changes out from underneath.
242 @samp{make install} will do the entire job of upgrading from a
243 previous installation of @theglibc{} version 2.x. There may sometimes
245 left behind from the previous installation, but those are generally
246 harmless. If you want to avoid leaving headers behind you can do
247 things in the following order.
249 You must first build the library (@samp{make}), optionally check it
250 (@samp{make check}), switch the include directories and then install
251 (@samp{make install}). The steps must be done in this order. Not moving
252 the directory before install will result in an unusable mixture of header
253 files from both libraries, but configuring, building, and checking the
254 library requires the ability to compile and run programs against the old
255 library. The new @file{/usr/include}, after switching the include
256 directories and before installing the library should contain the Linux
257 headers, but nothing else. If you do this, you will need to restore
258 any headers from libraries other than @theglibc{} yourself after installing the
261 You can install @theglibc{} somewhere other than where you configured it to go
262 by setting the @code{install_root} variable on the command line for
263 @samp{make install}. The value of this variable is prepended to all the
264 paths for installation. This is useful when setting up a chroot
265 environment or preparing a binary distribution. The directory should be
266 specified with an absolute file name.
268 @Theglibc{} includes a daemon called @code{nscd}, which you
269 may or may not want to run. @code{nscd} caches name service lookups; it
270 can dramatically improve performance with NIS+, and may help with DNS as
273 One auxiliary program, @file{/usr/libexec/pt_chown}, is installed setuid
274 @code{root}. This program is invoked by the @code{grantpt} function; it
275 sets the permissions on a pseudoterminal so it can be used by the
276 calling process. This means programs like @code{xterm} and
277 @code{screen} do not have to be setuid to get a pty. (There may be
278 other reasons why they need privileges.) If you are using a 2.1 or
279 newer Linux kernel with the @code{devptsfs} or @code{devfs} filesystems
280 providing pty slaves, you don't need this program; otherwise you do.
281 The source for @file{pt_chown} is in @file{login/programs/pt_chown.c}.
283 After installation you might want to configure the timezone and locale
284 installation of your system. @Theglibc{} comes with a locale
285 database which gets configured with @code{localedef}. For example, to
286 set up a German locale with name @code{de_DE}, simply issue the command
287 @samp{localedef -i de_DE -f ISO-8859-1 de_DE}. To configure all locales
288 that are supported by @theglibc{}, you can issue from your build directory the
289 command @samp{make localedata/install-locales}.
291 To configure the locally used timezone, set the @code{TZ} environment
292 variable. The script @code{tzselect} helps you to select the right value.
293 As an example, for Germany, @code{tzselect} would tell you to use
294 @samp{TZ='Europe/Berlin'}. For a system wide installation (the given
295 paths are for an installation with @samp{--prefix=/usr}), link the
296 timezone file which is in @file{/usr/share/zoneinfo} to the file
297 @file{/etc/localtime}. For Germany, you might execute @samp{ln -s
298 /usr/share/zoneinfo/Europe/Berlin /etc/localtime}.
300 @node Tools for Compilation
301 @appendixsec Recommended Tools for Compilation
302 @cindex installation tools
303 @cindex tools, for installing library
305 We recommend installing the following GNU tools before attempting to
310 GNU @code{make} 3.79 or newer
312 You need the latest version of GNU @code{make}. Modifying @theglibc{}
313 to work with other @code{make} programs would be so difficult that
314 we recommend you port GNU @code{make} instead. @strong{Really.} We
315 recommend GNU @code{make} version 3.79. All earlier versions have severe
316 bugs or lack features.
319 GCC 4.3 or newer, GCC 4.6 recommended
321 GCC 4.3 or higher is required; as of this writing, GCC 4.6 is the
322 compiler we advise to use to build @theglibc{}.
324 You can use whatever compiler you like to compile programs that use
327 Check the FAQ for any special compiler issues on particular platforms.
330 GNU @code{binutils} 2.20 or later
332 You must use GNU @code{binutils} (as and ld) to build @theglibc{}.
333 No other assembler or linker has the necessary functionality at the
337 GNU @code{texinfo} 4.5 or later
339 To correctly translate and install the Texinfo documentation you need
340 this version of the @code{texinfo} package. Earlier versions do not
341 understand all the tags used in the document, and the installation
342 mechanism for the info files is not present or works differently.
345 GNU @code{awk} 3.0, or higher
347 @code{Awk} is used in several places to generate files.
348 @code{gawk} 3.0 is known to work.
353 Perl is not required, but it is used if present to test the
354 installation. We may decide to use it elsewhere in the future.
357 GNU @code{sed} 3.02 or newer
359 @code{Sed} is used in several places to generate files. Most scripts work
360 with any version of @code{sed}. The known exception is the script
361 @code{po2test.sed} in the @code{intl} subdirectory which is used to
362 generate @code{msgs.h} for the test suite. This script works correctly
363 only with GNU @code{sed} 3.02. If you like to run the test suite, you
364 should definitely upgrade @code{sed}.
369 If you change any of the @file{configure.in} files you will also need
373 GNU @code{autoconf} 2.53 or higher
377 and if you change any of the message translation files you will need
381 GNU @code{gettext} 0.10.36 or later
385 You may also need these packages if you upgrade your source tree using
386 patches, although we try to avoid this.
389 @appendixsec Specific advice for @gnulinuxsystems{}
390 @cindex kernel header files
392 If you are installing @theglibc{} on @gnulinuxsystems{}, you need to have
393 the header files from a 2.6.19.1 or newer kernel around for reference.
394 These headers must be installed using @samp{make headers_install}; the
395 headers present in the kernel source directory are not suitable for
396 direct use by @theglibc{}. You do not need to use that kernel, just have
397 its headers installed where @theglibc{} can access them, referred to here as
398 @var{install-directory}. The easiest way to do this is to unpack it
399 in a directory such as @file{/usr/src/linux-@var{version}}. In that
400 directory, run @samp{make headers_install
401 INSTALL_HDR_PATH=@var{install-directory}}. Finally, configure @theglibc{}
402 with the option @samp{--with-headers=@var{install-directory}/include}.
403 Use the most recent kernel you can get your hands on. (If you are
404 cross-compiling @theglibc{}, you need to specify
405 @samp{ARCH=@var{architecture}} in the @samp{make headers_install}
406 command, where @var{architecture} is the architecture name used by the
407 Linux kernel, such as @samp{x86} or @samp{powerpc}.)
409 After installing @theglibc{}, you may need to remove or rename
410 directories such as @file{/usr/include/linux} and
411 @file{/usr/include/asm}, and replace them with copies of directories
412 such as @file{linux} and @file{asm} from
413 @file{@var{install-directory}/include}. All directories present in
414 @file{@var{install-directory}/include} should be copied, except that
415 @theglibc{} provides its own version of @file{/usr/include/scsi}; the
416 files provided by the kernel should be copied without replacing those
417 provided by @theglibc{}. The @file{linux}, @file{asm} and
418 @file{asm-generic} directories are required to compile programs using
419 @theglibc{}; the other directories describe interfaces to the kernel but
420 are not required if not compiling programs using those interfaces.
421 You do not need to copy kernel headers if you did not specify an
422 alternate kernel header source using @samp{--with-headers}.
424 The Filesystem Hierarchy Standard for @gnulinuxsystems{} expects some
425 components of the @glibcadj{} installation to be in
426 @file{/lib} and some in @file{/usr/lib}. This is handled automatically
427 if you configure @theglibc{} with @samp{--prefix=/usr}. If you set some other
428 prefix or allow it to default to @file{/usr/local}, then all the
429 components are installed there.
431 You cannot use @code{nscd} with 2.0 kernels, due to bugs in the
432 kernel-side thread support. @code{nscd} happens to hit these bugs
433 particularly hard, but you might have problems with any threaded
437 @appendixsec Reporting Bugs
438 @cindex reporting bugs
439 @cindex bugs, reporting
441 There are probably bugs in @theglibc{}. There are certainly
442 errors and omissions in this manual. If you report them, they will get
443 fixed. If you don't, no one will ever know about them and they will
444 remain unfixed for all eternity, if not longer.
446 It is a good idea to verify that the problem has not already been
447 reported. Bugs are documented in two places: The file @file{BUGS}
448 describes a number of well known bugs and the bug tracking system has a
450 @url{http://sourceware.org/bugzilla/}. The WWW
451 interface gives you access to open and closed reports. A closed report
452 normally includes a patch or a hint on solving the problem.
454 To report a bug, first you must find it. With any luck, this will be the
455 hard part. Once you've found a bug, make sure it's really a bug. A
456 good way to do this is to see if @theglibc{} behaves the same way
457 some other C library does. If so, probably you are wrong and the
458 libraries are right (but not necessarily). If not, one of the libraries
459 is probably wrong. It might not be @theglibc{}. Many historical
460 Unix C libraries permit things that we don't, such as closing a file
463 If you think you have found some way in which @theglibc{} does not
464 conform to the ISO and POSIX standards (@pxref{Standards and
465 Portability}), that is definitely a bug. Report it!
467 Once you're sure you've found a bug, try to narrow it down to the
468 smallest test case that reproduces the problem. In the case of a C
469 library, you really only need to narrow it down to one library
470 function call, if possible. This should not be too difficult.
472 The final step when you have a simple test case is to report the bug.
473 Do this using the WWW interface to the bug database.
475 If you are not sure how a function should behave, and this manual
476 doesn't tell you, that's a bug in the manual. Report that too! If the
477 function's behavior disagrees with the manual, then either the library
478 or the manual has a bug, so report the disagreement. If you find any
479 errors or omissions in this manual, please report them to the
480 bug database. If you refer to specific
481 sections of the manual, please include the section names for easier