1 Frequently Asked Questions about the GNU C Library
3 This document tries to answer questions a user might have when installing
4 and using glibc. Please make sure you read this before sending questions or
5 bug reports to the maintainers.
7 The GNU C library is very complex. The installation process has not been
8 completely automated; there are too many variables. You can do substantial
9 damage to your system by installing the library incorrectly. Make sure you
10 understand what you are undertaking before you begin.
12 If you have any questions you think should be answered in this document,
19 ?? What systems does the GNU C Library run on?
21 {UD} This is difficult to answer. The file `README' lists the architectures
22 GNU libc was known to run on *at some time*. This does not mean that it
23 still can be compiled and run on them now.
25 The systems glibc is known to work on as of this release, and most probably
29 i[3456]86-*-linux-gnu Linux-2.x on Intel
30 m68k-*-linux-gnu Linux-2.x on Motorola 680x0
31 alpha-*-linux-gnu Linux-2.x on DEC Alpha
32 powerpc-*-linux-gnu Linux and MkLinux on PowerPC systems
33 sparc-*-linux-gnu Linux-2.x on SPARC
34 sparc64-*-linux-gnu Linux-2.x on UltraSPARC
35 arm-*-none ARM standalone systems
36 arm-*-linux Linux-2.x on ARM
37 arm-*-linuxaout Linux-2.x on ARM using a.out binaries
39 Ports to other Linux platforms are in development, and may in fact work
40 already, but no one has sent us success reports for them. Currently no
41 ports to other operating systems are underway, although a few people have
44 If you have a system not listed above (or in the `README' file) and you are
45 really interested in porting it, contact
49 ??binsize What compiler do I need to build GNU libc?
51 {UD} You must use GNU CC to compile GNU libc. A lot of extensions of GNU CC
52 are used to increase portability and speed.
54 GNU CC is found, like all other GNU packages, on
56 ftp://ftp.gnu.org/pub/gnu
58 and the many mirror sites. ftp.gnu.org is always overloaded, so try to find
61 You should always try to use the latest official release. Older versions
62 may not have all the features GNU libc requires. The current releases of
63 egcs (1.0.3 and 1.1.1) should work with the GNU C library (for powerpc see
64 ?powerpc; for ARM see ?arm).
66 While the GNU CC should be able to compile glibc it is nevertheless adviced
67 to use EGCS. Comparing the sizes of glibc on Intel compiled with a recent
68 EGCS and gcc 2.8.1 shows this:
70 text data bss dec hex filename
71 egcs-2.93.10 862897 15944 12824 891665 d9b11 libc.so
72 gcc-2.8.1 959965 16468 12152 988585 f15a9 libc.so
74 Make up your own decision.
76 ?? When I try to compile glibc I get only error messages.
79 {UD} You definitely need GNU make to build GNU libc. No other make
80 program has the needed functionality.
82 We recommend version GNU make version 3.75 or 3.77. Versions before 3.75
83 have bugs and/or are missing features. Version 3.76 has bugs which
84 appear when building big projects like GNU libc. 3.76.1 appears to work but
85 some people have reported problems. If you build GNU make 3.77 from source,
86 please read ?make first.
88 ?? Do I need a special linker or assembler?
90 {ZW} If you want a shared library, you need a linker and assembler that
91 understand all the features of ELF, including weak and versioned symbols.
92 The static library can be compiled with less featureful tools, but lacks key
95 For Linux or Hurd, you want binutils 2.8.1.0.23, 2.9.1, or 2.9.1.0.15 or
96 higher. These are the only versions we've tested and found reliable. Other
97 versions after 2.8.1.0.23 may work but we don't recommend them, especially
98 not when C++ is involved. Earlier versions do not work at all.
100 Other operating systems may come with system tools that have all the
101 necessary features, but this is moot because glibc hasn't been ported to
104 ??powerpc Which compiler should I use for powerpc?
106 {GK} You want to use egcs 1.1 or later (together with the right versions
107 of all the other tools, of course).
109 In fact, egcs 1.1 has a bug that causes linuxthreads to be
110 miscompiled, resulting in segmentation faults when using condition
111 variables. There is a temporary patch at:
113 <http://discus.anu.edu.au/~geoffk/egcs-3.diff>
115 Later versions of egcs may fix this problem.
117 ??arm Which tools should I use for ARM?
119 {PB} You should use egcs 1.1 or a later version. For ELF systems some
120 changes are needed to the compiler; a patch against egcs-1.1.x can be found
123 <ftp://ftp.netwinder.org/users/p/philb/egcs-1.1.1pre2-diff-981126>
125 Binutils 2.9.1.0.16 or later is also required.
127 ?? Do I need some more things to compile the GNU C Library?
129 {UD} Yes, there are some more :-).
131 * GNU gettext. This package contains the tools needed to construct
132 `message catalog' files containing translated versions of system
133 messages. See ftp://ftp.gnu.org/pub/gnu or better any mirror
134 site. (We distribute compiled message catalogs, but they may not be
135 updated in patches.) Please note that the required minimal version
136 (0.10.35) of gettext is alpha software and available from
137 ftp://alpha.gnu.org/gnu .
139 * Some files are built with special tools. E.g., files ending in .gperf
140 need a `gperf' program. The GNU version (now available in a separate
141 package, formerly only as part of libg++) is known to work while some
142 vendor versions do not.
144 You should not need these tools unless you change the source files.
146 * Perl 5 is needed if you wish to test an installation of GNU libc
147 as the primary C library.
149 * When compiling for Linux, the header files of the Linux kernel must
150 be available to the compiler as <linux/*.h> and <asm/*.h>.
152 * lots of disk space (~170MB for i?86-linux; more for RISC platforms,
155 * plenty of time. Compiling just the shared and static libraries for
156 i?86-linux takes approximately 1h on an AMD-K6@225MHz w/ 96MB of RAM,
157 45mins on a Celeron@400MHz w/ 128MB, and 55mins on a Alpha@533MHz w/ 256MB.
158 Multiply this by 1.5 or 2.0 if you build profiling and/or the highly
159 optimized version as well. For Hurd systems times are much higher.
161 You should avoid compiling in a NFS mounted filesystem. This is
164 James Troup <J.J.Troup@comp.brad.ac.uk> reports a compile time of
165 45h34m for a full build (shared, static, and profiled) on Atari
166 Falcon (Motorola 68030 @ 16 Mhz, 14 Mb memory) and Jan Barte
167 <yann@plato.uni-paderborn.de> reports 22h48m on Atari TT030
168 (Motorola 68030 @ 32 Mhz, 34 Mb memory)
170 If you have some more measurements let me know.
172 ?? What version of the Linux kernel headers should be used?
174 {AJ,UD} The headers from the most recent Linux kernel should be used. The
175 headers used while compiling the GNU C library and the kernel binary used
176 when using the library do not need to match. The GNU C library runs without
177 problems on kernels that are older than the kernel headers used. The other
178 way round (compiling the GNU C library with old kernel headers and running
179 on a recent kernel) does not necessarily work. For example you can't use
180 new kernel features if you used old kernel headers to compile the GNU C
183 {ZW} Even if you are using a 2.0 kernel on your machine, we recommend you
184 compile GNU libc with 2.2 kernel headers. That way you won't have to
185 recompile libc if you ever upgrade to kernel 2.2. To tell libc which
186 headers to use, give configure the --with-headers switch
187 (e.g. --with-headers=/usr/src/linux-2.2.0/include).
189 Note that you must configure the 2.2 kernel if you do this, otherwise libc
190 will be unable to find <linux/version.h>. Just change the current directory
191 to the root of the 2.2 tree and do `make include/linux/version.h'.
193 ?? The compiler hangs while building iconvdata modules. What's
196 {ZW} This is a problem with old versions of GCC. Initialization of large
197 static arrays is very slow. The compiler will eventually finish; give it
200 The problem is fixed in egcs 1.1.
202 ?? When I run `nm -u libc.so' on the produced library I still
203 find unresolved symbols. Can this be ok?
205 {UD} Yes, this is ok. There can be several kinds of unresolved symbols:
207 * magic symbols automatically generated by the linker. These have names
208 like __start_* and __stop_*
210 * symbols starting with _dl_* come from the dynamic linker
212 * weak symbols, which need not be resolved at all (fabs for example)
214 Generally, you should make sure you find a real program which produces
215 errors while linking before deciding there is a problem.
217 ??addon What are these `add-ons'?
219 {UD} To avoid complications with export rules or external source code some
220 optional parts of the libc are distributed as separate packages (e.g., the
221 crypt package, see ?crypt).
223 To use these packages as part of GNU libc, just unpack the tarfiles in the
224 libc source directory and tell the configuration script about them using the
225 --enable-add-ons option. If you give just --enable-add-ons configure tries
226 to find all the add-on packages in your source tree. This may not work. If
227 it doesn't, or if you want to select only a subset of the add-ons, give a
228 comma-separated list of the add-ons to enable:
230 configure --enable-add-ons=crypt,linuxthreads
234 Add-ons can add features (including entirely new shared libraries), override
235 files, provide support for additional architectures, and just about anything
236 else. The existing makefiles do most of the work; only some few stub rules
237 must be written to get everything running.
239 ?? My XXX kernel emulates a floating-point coprocessor for me.
240 Should I enable --with-fp?
242 {ZW} An emulated FPU is just as good as a real one, as far as the C library
243 is concerned. You only need to say --without-fp if your machine has no way
244 to execute floating-point instructions.
246 People who are interested in squeezing the last drop of performance
247 out of their machine may wish to avoid the trap overhead, but this is
248 far more trouble than it's worth: you then have to compile
249 *everything* this way, including the compiler's internal libraries
250 (libgcc.a for GNU C), because the calling conventions change.
252 ?? When compiling GNU libc I get lots of errors saying functions
253 in glibc are duplicated in libgcc.
255 {EY} This is *exactly* the same problem that I was having. The problem was
256 due to the fact that configure didn't correctly detect that the linker flag
257 --no-whole-archive was supported in my linker. In my case it was because I
258 had run ./configure with bogus CFLAGS, and the test failed.
260 One thing that is particularly annoying about this problem is that once this
261 is misdetected, running configure again won't fix it unless you first delete
264 {UD} Starting with glibc-2.0.3 there should be a better test to avoid some
265 problems of this kind. The setting of CFLAGS is checked at the very
266 beginning and if it is not usable `configure' will bark.
268 ?? Why do I get messages about missing thread functions when I use
269 librt? I don't even use threads.
271 {UD} In this case you probably mixed up your installation. librt uses
272 threads internally and has implicit references to the thread library.
273 Normally these references are satisfied automatically but if the thread
274 library is not in the expected place you must tell the linker where it is.
275 When using GNU ld it works like this:
277 gcc -o foo foo.c -Wl,-rpath-link=/some/other/dir -lrt
279 The `/some/other/dir' should contain the thread library. `ld' will use the
280 given path to find the implicitly referenced library while not disturbing
283 ?? What's the problem with configure --enable-omitfp?
285 {AJ} When --enable-omitfp is set the libraries are built without frame
286 pointers. Some compilers produce buggy code for this model and therefore we
287 don't advise using it at the moment.
289 If you use --enable-omitfp, you're on your own. If you encounter problems
290 with a library that was build this way, we advise you to rebuild the library
291 without --enable-omitfp. If the problem vanishes consider tracking the
292 problem down and report it as compiler failure.
294 Since a library built with --enable-omitfp is undebuggable on most systems,
295 debuggable libraries are also built - you can use them by appending "_g" to
298 The compilation of these extra libraries and the compiler optimizations slow
299 down the build process and need more disk space.
301 ?? I get failures during `make check'. What should I do?
303 {AJ} The testsuite should compile and run cleanly on your system; every
304 failure should be looked into. Depending on the failures, you probably
305 should not install the library at all.
307 You should consider using the `glibcbug' script to report the failure,
308 providing as much detail as possible. If you run a test directly, please
309 remember to set up the environment correctly. You want to test the compiled
310 library - and not your installed one. The best way is to copy the exact
311 command line which failed and run the test from the subdirectory for this
314 There are some failures which are not directly related to the GNU libc:
315 - Some compilers produce buggy code. No compiler gets single precision
316 complex numbers correct on Alpha. Otherwise, the egcs 1.1 release should be
317 ok; gcc 2.8.1 might cause some failures; gcc 2.7.2.x is so buggy that
318 explicit checks have been used so that you can't build with it.
319 - The kernel might have bugs. For example on Linux/Alpha 2.0.34 the
320 floating point handling has quite a number of bugs and therefore most of
321 the test cases in the math subdirectory will fail. Linux 2.2 has
322 fixes for the floating point support on Alpha. The Linux/SPARC kernel has
323 also some bugs in the FPU emulation code (as of Linux 2.2.0).
324 - Other tools might have problems. For example bash 2.03 gives a
325 segmentation fault running the tst-rpmatch.sh test script.
327 ?? What is symbol versioning good for? Do I need it?
329 {AJ} Symbol versioning solves problems that are related to interface
330 changes. One version of an interface might have been introduced in a
331 previous version of the GNU C library but the interface or the semantics of
332 the function has been changed in the meantime. For binary compatibility
333 with the old library, a newer library needs to still have the old interface
334 for old programs. On the other hand, new programs should use the new
335 interface. Symbol versioning is the solution for this problem. The GNU
336 libc version 2.1 uses symbol versioning by default if the installed binutils
339 We don't advise building without symbol versioning, since you lose binary
340 compatibility - forever! The binary compatibility you lose is not only
341 against the previous version of the GNU libc (version 2.0) but also against
345 ? Installation and configuration issues
347 ?? Can I replace the libc on my Linux system with GNU libc?
349 {UD} You cannot replace any existing libc for Linux with GNU libc. It is
350 binary incompatible and therefore has a different major version. You can,
351 however, install it alongside your existing libc.
353 For Linux there are three major libc versions:
355 libc-5 original ELF libc
358 You can have any combination of these three installed. For more information
359 consult documentation for shared library handling. The Makefiles of GNU
360 libc will automatically generate the needed symbolic links which the linker
363 ?? How do I configure GNU libc so that the essential libraries
364 like libc.so go into /lib and the other into /usr/lib?
366 {UD,AJ} Like all other GNU packages GNU libc is designed to use a base
367 directory and install all files relative to this. The default is
368 /usr/local, because this is safe (it will not damage the system if installed
369 there). If you wish to install GNU libc as the primary C library on your
370 system, set the base directory to /usr (i.e. run configure --prefix=/usr
371 <other_options>). Note that this can damage your system; see ?safety for
374 Some systems like Linux have a filesystem standard which makes a difference
375 between essential libraries and others. Essential libraries are placed in
376 /lib because this directory is required to be located on the same disk
377 partition as /. The /usr subtree might be found on another
378 partition/disk. If you configure for Linux with --prefix=/usr, then this
379 will be done automatically.
381 To install the essential libraries which come with GNU libc in /lib on
382 systems other than Linux one must explicitly request it. Autoconf has no
383 option for this so you have to use a `configparms' file (see the `INSTALL'
384 file for details). It should contain:
389 The first line specifies the directory for the essential libraries, the
390 second line the directory for system configuration files.
392 ??safety How should I avoid damaging my system when I install GNU libc?
394 {ZW} If you wish to be cautious, do not configure with --prefix=/usr. If
395 you don't specify a prefix, glibc will be installed in /usr/local, where it
396 will probably not break anything. (If you wish to be certain, set the
397 prefix to something like /usr/local/glibc2 which is not used for anything.)
399 The dangers when installing glibc in /usr are twofold:
401 * glibc will overwrite the headers in /usr/include. Other C libraries
402 install a different but overlapping set of headers there, so the effect
403 will probably be that you can't compile anything. You need to rename
404 /usr/include out of the way before running `make install'. (Do not throw
405 it away; you will then lose the ability to compile programs against your
408 * None of your old libraries, static or shared, can be used with a
409 different C library major version. For shared libraries this is not a
410 problem, because the filenames are different and the dynamic linker
411 will enforce the restriction. But static libraries have no version
412 information. You have to evacuate all the static libraries in
413 /usr/lib to a safe location.
415 The situation is rather similar to the move from a.out to ELF which
416 long-time Linux users will remember.
418 ?? Do I need to use GNU CC to compile programs that will use the
421 {ZW} In theory, no; the linker does not care, and the headers are supposed
422 to check for GNU CC before using its extensions to the C language.
424 However, there are currently no ports of glibc to systems where another
425 compiler is the default, so no one has tested the headers extensively
426 against another compiler. You may therefore encounter difficulties. If you
427 do, please report them as bugs.
429 Also, in several places GNU extensions provide large benefits in code
430 quality. For example, the library has hand-optimized, inline assembly
431 versions of some string functions. These can only be used with GCC. See
434 ??crypt When linking with the new libc I get unresolved symbols
435 `crypt' and `setkey'. Why aren't these functions in the
438 {UD} The US places restrictions on exporting cryptographic programs and
439 source code. Until this law gets abolished we cannot ship the cryptographic
440 functions together with glibc.
442 The functions are available, as an add-on (see ?addon). People in the US
443 may get it from the same place they got GNU libc from. People outside the
444 US should get the code from ftp.funet.fi [128.214.248.6] in the directory
445 pub/gnu/funet, or another archive site outside the USA. The README explains
446 how to install the sources.
448 If you already have the crypt code on your system the reason for the failure
449 is probably that you did not link with -lcrypt. The crypto functions are in
450 a separate library to make it possible to export GNU libc binaries from the
453 ?? When I use GNU libc on my Linux system by linking against
454 the libc.so which comes with glibc all I get is a core dump.
456 {UD} On Linux, gcc sets the dynamic linker to /lib/ld-linux.so.1 unless the
457 user specifies a --dynamic-linker argument. This is the name of the libc5
458 dynamic linker, which does not work with glibc.
460 For casual use of GNU libc you can just specify to the linker
461 --dynamic-linker=/lib/ld-linux.so.2
463 which is the glibc dynamic linker, on Linux systems. On other systems the
464 name is /lib/ld.so.1. When linking via gcc, you've got to add
465 -Wl,--dynamic-linker=/lib/ld-linux.so.2
467 to the gcc command line.
469 To change your environment to use GNU libc for compiling you need to change
470 the `specs' file of your gcc. This file is normally found at
472 /usr/lib/gcc-lib/<arch>/<version>/specs
474 In this file you have to change a few things:
476 - change `ld-linux.so.1' to `ld-linux.so.2'
478 - remove all expression `%{...:-lgmon}'; there is no libgmon in glibc
480 - fix a minor bug by changing %{pipe:-} to %|
482 Here is what the gcc-2.7.2 specs file should look like when GNU libc is
485 -----------------------------------------------------------------------
487 %{V} %{v:%{!V:-V}} %{Qy:} %{!Qn:-Qy} %{n} %{T} %{Ym,*} %{Yd,*} %{Wa,*:%*}
493 %{fPIC:-D__PIC__ -D__pic__} %{fpic:-D__PIC__ -D__pic__} %{!m386:-D__i486__} %{posix:-D_POSIX_SOURCE} %{pthread:-D_REENTRANT}
502 %{!shared:crtend.o%s} %{shared:crtendS.o%s} crtn.o%s
505 -m elf_i386 %{shared:-shared} %{!shared: %{!ibcs: %{!static: %{rdynamic:-export-dynamic} %{!dynamic-linker:-dynamic-linker /lib/ld-linux.so.2}} %{static:-static}}}
508 %{!shared: %{pthread:-lpthread} %{profile:-lc_p} %{!profile: -lc}}
514 %{!shared: %{pg:gcrt1.o%s} %{!pg:%{p:gcrt1.o%s} %{!p:%{profile:gcrt1.o%s} %{!profile:crt1.o%s}}}} crti.o%s %{!shared:crtbegin.o%s} %{shared:crtbeginS.o%s}
516 *switches_need_spaces:
520 %{funsigned-char:-D__CHAR_UNSIGNED__}
523 -D__ELF__ -Dunix -Di386 -Dlinux -Asystem(unix) -Asystem(posix) -Acpu(i386) -Amachine(i386)
531 -----------------------------------------------------------------------
533 Things get a bit more complicated if you have GNU libc installed in some
534 other place than /usr, i.e., if you do not want to use it instead of the old
535 libc. In this case the needed startup files and libraries are not found in
536 the regular places. So the specs file must tell the compiler and linker
539 Version 2.7.2.3 does and future versions of GCC will automatically
540 provide the correct specs.
542 ?? Looking through the shared libc file I haven't found the
543 functions `stat', `lstat', `fstat', and `mknod' and while
544 linking on my Linux system I get error messages. How is
545 this supposed to work?
547 {RM} Believe it or not, stat and lstat (and fstat, and mknod) are supposed
548 to be undefined references in libc.so.6! Your problem is probably a missing
549 or incorrect /usr/lib/libc.so file; note that this is a small text file now,
550 not a symlink to libc.so.6. It should look something like this:
552 GROUP ( libc.so.6 libc_nonshared.a )
554 ?? When I run an executable on one system which I compiled on
555 another, I get dynamic linker errors. Both systems have the same
556 version of glibc installed. What's wrong?
558 {ZW} Glibc on one of these systems was compiled with gcc 2.7 or 2.8, the
559 other with egcs (any version). Egcs has functions in its internal
560 `libgcc.a' to support exception handling with C++. They are linked into
561 any program or dynamic library compiled with egcs, whether it needs them or
562 not. Dynamic libraries then turn around and export those functions again
563 unless special steps are taken to prevent them.
565 When you link your program, it resolves its references to the exception
566 functions to the ones exported accidentally by libc.so. That works fine as
567 long as libc has those functions. On the other system, libc doesn't have
568 those functions because it was compiled by gcc 2.8, and you get undefined
569 symbol errors. The symbols in question are named things like
570 `__register_frame_info'.
572 For glibc 2.0, the workaround is to not compile libc with egcs. We've also
573 incorporated a patch which should prevent the EH functions sneaking into
574 libc. It doesn't matter what compiler you use to compile your program.
576 For glibc 2.1, we've chosen to do it the other way around: libc.so
577 explicitly provides the EH functions. This is to prevent other shared
578 libraries from doing it.
580 {UD} Starting with glibc 2.1.1 you can compile glibc with gcc 2.8.1 or
581 newer since we have explicitly add references to the functions causing the
582 problem. But you nevertheless should use EGCS for other reasons
585 ?? How can I compile gcc 2.7.2.1 from the gcc source code using
588 {AJ} There's only correct support for glibc 2.0.x in gcc 2.7.2.3 or later.
589 But you should get at least gcc 2.8.1 or egcs 1.1 (or later versions)
592 ?? The `gencat' utility cannot process the catalog sources which
593 were used on my Linux libc5 based system. Why?
595 {UD} The `gencat' utility provided with glibc complies to the XPG standard.
596 The older Linux version did not obey the standard, so they are not
599 To ease the transition from the Linux version some of the non-standard
600 features are also present in the `gencat' program of GNU libc. This mainly
601 includes the use of symbols for the message number and the automatic
602 generation of header files which contain the needed #defines to map the
605 Here is a simple SED script to convert at least some Linux specific catalog
606 files to the XPG4 form:
608 -----------------------------------------------------------------------
609 # Change catalog source in Linux specific format to standard XPG format.
610 # Ulrich Drepper <drepper@cygnus.com>, 1996.
614 s/\$ #\([^ ]*\).*/\1/
616 s/\$ #[^ ]* *\(.*\)/\$ \1/
622 s/\(.*\)\n\(.*\)/\2 \1/
624 -----------------------------------------------------------------------
626 ?? Programs using libc have their messages translated, but other
627 behavior is not localized (e.g. collating order); why?
629 {ZW} Translated messages are automatically installed, but the locale
630 database that controls other behaviors is not. You need to run localedef to
631 install this database, after you have run `make install'. For example, to
632 set up the French Canadian locale, simply issue the command
634 localedef -i fr_CA -f ISO-8859-1 fr_CA
636 Please see localedata/README in the source tree for further details.
638 ?? I have set up /etc/nis.conf, and the Linux libc 5 with NYS
639 works great. But the glibc NIS+ doesn't seem to work.
641 {TK} The glibc NIS+ implementation uses a /var/nis/NIS_COLD_START file for
642 storing information about the NIS+ server and their public keys, because the
643 nis.conf file does not contain all the necessary information. You have to
644 copy a NIS_COLD_START file from a Solaris client (the NIS_COLD_START file is
645 byte order independent) or generate it with nisinit from the nis-tools
646 package; available at
648 http://www.suse.de/~kukuk/linux/nisplus.html
650 ?? I have killed ypbind to stop using NIS, but glibc
653 {TK} For faster NIS lookups, glibc uses the /var/yp/binding/ files from
654 ypbind. ypbind 3.3 and older versions don't always remove these files, so
655 glibc will continue to use them. Other BSD versions seem to work correctly.
656 Until ypbind 3.4 is released, you can find a patch at
658 <ftp://ftp.kernel.org/pub/linux/utils/net/NIS/ypbind-3.3-glibc4.diff.gz>
660 ?? Under Linux/Alpha, I always get "do_ypcall: clnt_call:
661 RPC: Unable to receive; errno = Connection refused" when using NIS.
663 {TK} You need a ypbind version which is 64bit clean. Some versions are not
664 64bit clean. A 64bit clean implementation is ypbind-mt. For ypbind 3.3,
665 you need the patch from ftp.kernel.org (See the previous question). I don't
666 know about other versions.
669 ?? After installing glibc name resolving doesn't work properly.
671 {AJ} You probably should read the manual section describing nsswitch.conf
672 (just type `info libc "NSS Configuration File"'). The NSS configuration
673 file is usually the culprit.
676 ?? How do I create the databases for NSS?
678 {AJ} If you have an entry "db" in /etc/nsswitch.conf you should also create
679 the database files. The glibc sources contain a Makefile which does the
680 necessary conversion and calls to create those files. The file is
681 `db-Makefile' in the subdirectory `nss' and you can call it with `make -f
682 db-Makefile'. Please note that not all services are capable of using a
683 database. Currently passwd, group, ethers, protocol, rpc, services shadow
684 and netgroup are implemented.
686 ?? I have /usr/include/net and /usr/include/scsi as symlinks
687 into my Linux source tree. Is that wrong?
689 {PB} This was necessary for libc5, but is not correct when using glibc.
690 Including the kernel header files directly in user programs usually does not
691 work (see ?kerhdr). glibc provides its own <net/*> and <scsi/*> header
692 files to replace them, and you may have to remove any symlink that you have
693 in place before you install glibc. However, /usr/include/asm and
694 /usr/include/linux should remain as they were.
696 ?? Programs like `logname', `top', `uptime' `users', `w' and
697 `who', show incorrect information about the (number of)
698 users on my system. Why?
702 ?? After upgrading to glibc 2.1 with symbol versioning I get
703 errors about undefined symbols. What went wrong?
705 {AJ} The problem is caused either by wrong program code or tools. In the
706 versioned libc a lot of symbols are now local that were global symbols in
707 previous versions. It seems that programs linked against older versions
708 often accidentally used libc global variables -- something that should not
711 The only way to fix this is to recompile your program. Sorry, that's the
712 price you might have to pay once for quite a number of advantages with
715 ?? When I start the program XXX after upgrading the library
717 XXX: Symbol `_sys_errlist' has different size in shared
718 object, consider re-linking
719 Why? What should I do?
721 {UD} As the message says, relink the binary. The problem is that a few
722 symbols from the library can change in size and there is no way to avoid
723 this. _sys_errlist is a good example. Occasionally there are new error
724 numbers added to the kernel and this must be reflected at user level,
725 breaking programs that refer to them directly.
727 Such symbols should normally not be used at all. There are mechanisms to
728 avoid using them. In the case of _sys_errlist, there is the strerror()
729 function which should _always_ be used instead. So the correct fix is to
730 rewrite that part of the application.
732 In some situations (especially when testing a new library release) it might
733 be possible that a symbol changed size when that should not have happened.
734 So in case of doubt report such a warning message as a problem.
736 ?? What do I need for C++ development?
738 {HJ,AJ} You need either egcs 1.1 which comes directly with libstdc++ or
739 gcc-2.8.1 together with libstdc++ 2.8.1.1. egcs 1.1 has the better C++
740 support and works directly with glibc 2.1. If you use gcc-2.8.1 with
741 libstdc++ 2.8.1.1, you need to modify libstdc++ a bit. A patch is available
743 <ftp://alpha.gnu.org/gnu/libstdc++-2.8.1.1-glibc2.1-diff.gz>
745 Please note that libg++ 2.7.2 (and the Linux Versions 2.7.2.x) doesn't work
746 very well with the GNU C library due to vtable thunks. If you're upgrading
747 from glibc 2.0.x to 2.1 you have to recompile libstdc++ since the library
748 compiled for 2.0 is not compatible due to the new Large File Support (LFS)
751 {UD} But since in the case of a shared libstdc++ the version numbers should
752 be different existing programs will continue to work.
754 ?? Even statically linked programs need some shared libraries
755 which is not acceptable for me. What can I do?
757 {AJ} NSS (for details just type `info libc "Name Service Switch"') won't
758 work properly without shared libraries. NSS allows using different services
759 (e.g. NIS, files, db, hesiod) by just changing one configuration file
760 (/etc/nsswitch.conf) without relinking any programs. The only disadvantage
761 is that now static libraries need to access shared libraries. This is
762 handled transparently by the GNU C library.
764 A solution is to configure glibc with --enable-static-nss. In this case you
765 can create a static binary that will use only the services dns and files
766 (change /etc/nsswitch.conf for this). You need to link explicitly against
767 all these services. For example:
769 gcc -static test-netdb.c -o test-netdb.c \
770 -lc -lnss_files -lnss_dns -lresolv
772 The problem with this approach is that you've got to link every static
773 program that uses NSS routines with all those libraries.
775 {UD} In fact, one cannot say anymore that a libc compiled with this
776 option is using NSS. There is no switch anymore. Therefore it is
777 *highly* recommended *not* to use --enable-static-nss since this makes
778 the behaviour of the programs on the system inconsistent.
780 ?? I just upgraded my Linux system to glibc and now I get
781 errors whenever I try to link any program.
783 {ZW} This happens when you have installed glibc as the primary C library but
784 have stray symbolic links pointing at your old C library. If the first
785 `libc.so' the linker finds is libc 5, it will use that. Your program
786 expects to be linked with glibc, so the link fails.
788 The most common case is that glibc put its `libc.so' in /usr/lib, but there
789 was a `libc.so' from libc 5 in /lib, which gets searched first. To fix the
790 problem, just delete /lib/libc.so. You may also need to delete other
791 symbolic links in /lib, such as /lib/libm.so if it points to libm.so.5.
793 {AJ} The perl script test-installation.pl which is run as last step during
794 an installation of glibc that is configured with --prefix=/usr should help
795 detect these situations. If the script reports problems, something is
798 ?? When I use nscd the machine freezes.
800 {UD} You cannot use nscd with Linux 2.0.*. There is functionality missing
801 in the kernel and work-arounds are not suitable. Besides, some parts of the
802 kernel are too buggy when it comes to using threads.
804 If you need nscd, you have to use at least a 2.1 kernel.
806 Note that I have at this point no information about any other platform.
808 ?? I need lots of open files. What do I have to do?
810 {AJ} This is at first a kernel issue. The kernel defines limits with
811 OPEN_MAX the number of simultaneous open files and with FD_SETSIZE the
812 number of used file descriptors. You need to change these values in your
813 kernel and recompile the kernel so that the kernel allows to use more open
814 files. You don't necessarily need to recompile the GNU C library since the
815 only place where OPEN_MAX and FD_SETSIZE is really needed in the library
816 itself is the size of fd_set which is used by select.
818 The GNU C library is now (nearly) select free. This means it internally has
819 no limits imposed by the `fd_set' type. Instead almost all places where the
820 functionality is needed the `poll' function is used.
822 If you increase the number of file descriptors in the kernel you don't need
823 to recompile the C library. The remaining select calls are in the RPC code.
824 If your RPC daemons don't need more than FD_SETSIZE file descriptors, you
825 don't need to change anything at all.
827 {UD} You can always get the maximum number of file descriptors a process is
828 allowed to have open at any time using
830 number = sysconf (_SC_OPEN_MAX);
832 This will work even if the kernel limits change.
834 ?? How do I get the same behavior on parsing /etc/passwd and
835 /etc/group as I have with libc5 ?
837 {TK} The name switch setup in /etc/nsswitch.conf selected by most Linux
838 distributions does not support +/- and netgroup entries in the files like
839 /etc/passwd. Though this is the preferred setup some people might have
840 setups coming over from the libc5 days where it was the default to recognize
841 lines like this. To get back to the old behaviour one simply has to change
842 the rules for passwd, group, and shadow in the nsswitch.conf file as
853 ??libs What needs to be recompiled when upgrading from glibc 2.0 to glibc
856 {AJ,CG} If you just upgrade the glibc from 2.0.x (x <= 7) to 2.1, binaries
857 that have been linked against glibc 2.0 will continue to work.
859 If you compile your own binaries against glibc 2.1, you also need to
860 recompile some other libraries. The problem is that libio had to be
861 changed and therefore libraries that are based or depend on the libio
862 of glibc, e.g. ncurses or slang, need to be recompiled. If you
863 experience strange segmentation faults in your programs linked against
864 glibc 2.1, you might need to recompile your libraries.
866 Another problem is that older binaries that were linked statically against
867 glibc 2.0 will reference the older nss modules (libnss_files.so.1 instead of
868 libnss_files.so.2), so don't remove them. Also, the old glibc-2.0 compiled
869 static libraries (libfoo.a) which happen to depend on the older libio
870 behavior will be broken by the glibc 2.1 upgrade. We plan to produce a
871 compatibility library that people will be able to link in if they want
872 to compile a static library generated against glibc 2.0 into a program
873 on a glibc 2.1 system. You just add -lcompat and you should be fine.
875 The glibc-compat add-on will provide the libcompat.a library, the older
876 nss modules, and a few other files. Together, they should make it
877 possible to do development with old static libraries on a glibc 2.1
878 system. This add-on is still in development. You can get it from
879 <ftp://alpha.gnu.org/gnu/glibc-compat-2.1.tar.gz>
880 but please keep in mind that it is experimental.
882 ?? Why is extracting files via tar so slow?
884 {AJ} Extracting of tar archives might be quite slow since tar has to look up
885 userid and groupids and doesn't cache negative results. If you have nis or
886 nisplus in your /etc/nsswitch.conf for the passwd and/or group database,
887 each file extractions needs a network connection. There are two possible
890 - do you really need NIS/NIS+ (some Linux distributions add by default
891 nis/nisplus even if it's not needed)? If not, just remove the entries.
893 - if you need NIS/NIS+, use the Name Service Cache Daemon nscd that comes
896 ?? Compiling programs I get parse errors in libio.h (e.g. "parse error
897 before `_IO_seekoff'"). How should I fix this?
899 {AJ} You might get the following errors when upgrading to glibc 2.1:
901 In file included from /usr/include/stdio.h:57,
903 /usr/include/libio.h:335: parse error before `_IO_seekoff'
904 /usr/include/libio.h:335: parse error before `_G_off64_t'
905 /usr/include/libio.h:336: parse error before `_IO_seekpos'
906 /usr/include/libio.h:336: parse error before `_G_fpos64_t'
908 The problem is a wrong _G_config.h file in your include path. The
909 _G_config.h file that comes with glibc 2.1 should be used and not one from
910 libc5 or from a compiler directory. To check which _G_config.h file the
911 compiler uses, compile your program with `gcc -E ...|grep G_config.h' and
912 remove that file. Your compiler should pick up the file that has been
913 installed by glibc 2.1 in your include directory.
915 ?? After upgrading to glibc 2.1, libraries that were compiled against
916 glibc 2.0.x don't work anymore.
921 ? Source and binary incompatibilities, and what to do about them
923 ?? I expect GNU libc to be 100% source code compatible with
924 the old Linux based GNU libc. Why isn't it like this?
926 {DMT,UD} Not every extension in Linux libc's history was well thought-out.
927 In fact it had a lot of problems with standards compliance and with
928 cleanliness. With the introduction of a new version number these errors can
929 now be corrected. Here is a list of the known source code
932 * _GNU_SOURCE: glibc does not make the GNU extensions available
933 automatically. If a program depends on GNU extensions or some
934 other non-standard functionality, it is necessary to compile it
935 with the C compiler option -D_GNU_SOURCE, or better, to put
936 `#define _GNU_SOURCE' at the beginning of your source files, before
937 any C library header files are included. This difference normally
938 manifests itself in the form of missing prototypes and/or data type
939 definitions. Thus, if you get such errors, the first thing you
940 should do is try defining _GNU_SOURCE and see if that makes the
943 For more information consult the file `NOTES' in the GNU C library
946 * reboot(): GNU libc sanitizes the interface of reboot() to be more
947 compatible with the interface used on other OSes. reboot() as
948 implemented in glibc takes just one argument. This argument
949 corresponds to the third argument of the Linux reboot system call.
950 That is, a call of the form reboot(a, b, c) needs to be changed into
951 reboot(c). Beside this the header <sys/reboot.h> defines the needed
952 constants for the argument. These RB_* constants should be used
953 instead of the cryptic magic numbers.
955 * swapon(): the interface of this function didn't change, but the
956 prototype is in a separate header file <sys/swap.h>. This header
957 file also provides the SWAP_* constants defined by <linux/swap.h>;
958 you should use them for the second argument to swapon().
960 * errno: If a program uses the variable "errno", then it _must_
961 include <errno.h>. The old libc often (erroneously) declared this
962 variable implicitly as a side-effect of including other libc header
963 files. glibc is careful to avoid such namespace pollution, which,
964 in turn, means that you really need to include the header files that
965 you depend on. This difference normally manifests itself in the
966 form of the compiler complaining about references to an undeclared
969 * Linux-specific syscalls: All Linux system calls now have appropriate
970 library wrappers and corresponding declarations in various header files.
971 This is because the syscall() macro that was traditionally used to
972 work around missing syscall wrappers are inherently non-portable and
973 error-prone. The following table lists all the new syscall stubs,
974 the header-file declaring their interface and the system call name.
976 syscall name: wrapper name: declaring header file:
977 ------------- ------------- ----------------------
978 bdflush bdflush <sys/kdaemon.h>
979 syslog ksyslog_ctl <sys/klog.h>
981 * lpd: Older versions of lpd depend on a routine called _validuser().
982 The library does not provide this function, but instead provides
983 __ivaliduser() which has a slightly different interface. Simply
984 upgrading to a newer lpd should fix this problem (e.g., the 4.4BSD
985 lpd is known to be working).
987 * resolver functions/BIND: like on many other systems the functions of
988 the resolver library are not included in libc itself. There is a
989 separate library libresolv. If you get undefined symbol errors for
990 symbols starting with `res_*' simply add -lresolv to your linker
993 * the `signal' function's behavior corresponds to the BSD semantic and
994 not the SysV semantic as it was in libc-5. The interface on all GNU
995 systems shall be the same and BSD is the semantic of choice. To use
996 the SysV behavior simply use `sysv_signal', or define _XOPEN_SOURCE.
997 See ?signal for details.
999 ??getlog Why does getlogin() always return NULL on my Linux box?
1001 {UD} The GNU C library has a format for the UTMP and WTMP file which differs
1002 from what your system currently has. It was extended to fulfill the needs
1003 of the next years when IPv6 is introduced. The record size is different and
1004 some fields have different positions. The files written by functions from
1005 the one library cannot be read by functions from the other library. Sorry,
1006 but this is what a major release is for. It's better to have a cut now than
1007 having no means to support the new techniques later.
1009 {MK} There is however a (partial) solution for this problem. Please take a
1010 look at the file `login/README.utmpd'.
1012 ?? Where are the DST_* constants found in <sys/time.h> on many
1015 {UD} These constants come from the old BSD days and are not used anymore
1016 (libc5 does not actually implement the handling although the constants are
1019 Instead GNU libc contains zone database support and compatibility code for
1020 POSIX TZ environment variable handling. For former is very much preferred
1023 ?? The prototypes for `connect', `accept', `getsockopt',
1024 `setsockopt', `getsockname', `getpeername', `send',
1025 `sendto', and `recvfrom' are different in GNU libc from
1026 any other system I saw. This is a bug, isn't it?
1028 {UD} No, this is no bug. This version of GNU libc already follows the new
1029 Single Unix specifications (and I think the POSIX.1g draft which adopted the
1030 solution). The type for a parameter describing a size is now `socklen_t', a
1033 ??kerhdr On Linux I've got problems with the declarations in Linux
1036 {UD,AJ} On Linux, the use of kernel headers is reduced to the minimum. This
1037 gives Linus the ability to change the headers more freely. Also, user
1038 programs are now insulated from changes in the size of kernel data
1041 For example, the sigset_t type is 32 or 64 bits wide in the kernel. In
1042 glibc it is 1024 bits wide. This guarantees that when the kernel gets a
1043 bigger sigset_t (for POSIX.1e realtime support, say) user programs will not
1044 have to be recompiled. Consult the header files for more information about
1047 Therefore you shouldn't include Linux kernel header files directly if glibc
1048 has defined a replacement. Otherwise you might get undefined results because
1051 ?? I don't include any kernel headers myself but the compiler
1052 still complains about redeclarations of types in the kernel
1055 {UD} The kernel headers before Linux 2.1.61 and 2.0.32 don't work correctly
1056 with glibc. Compiling C programs is possible in most cases but C++ programs
1057 have (due to the change of the name lookups for `struct's) problems. One
1058 prominent example is `struct fd_set'.
1060 There might be some problems left but 2.1.61/2.0.32 fix most of the known
1061 ones. See the BUGS file for other known problems.
1063 ??signal Why don't signals interrupt system calls anymore?
1065 {ZW} By default GNU libc uses the BSD semantics for signal(), unlike Linux
1066 libc 5 which used System V semantics. This is partially for compatibility
1067 with other systems and partially because the BSD semantics tend to make
1068 programming with signals easier.
1070 There are three differences:
1072 * BSD-style signals that occur in the middle of a system call do not
1073 affect the system call; System V signals cause the system call to
1074 fail and set errno to EINTR.
1076 * BSD signal handlers remain installed once triggered. System V signal
1077 handlers work only once, so one must reinstall them each time.
1079 * A BSD signal is blocked during the execution of its handler. In other
1080 words, a handler for SIGCHLD (for example) does not need to worry about
1081 being interrupted by another SIGCHLD. It may, however, be interrupted
1084 There is general consensus that for `casual' programming with signals, the
1085 BSD semantics are preferable. You don't need to worry about system calls
1086 returning EINTR, and you don't need to worry about the race conditions
1087 associated with one-shot signal handlers.
1089 If you are porting an old program that relies on the old semantics, you can
1090 quickly fix the problem by changing signal() to sysv_signal() throughout.
1091 Alternatively, define _XOPEN_SOURCE before including <signal.h>.
1093 For new programs, the sigaction() function allows you to specify precisely
1094 how you want your signals to behave. All three differences listed above are
1095 individually switchable on a per-signal basis with this function.
1097 If all you want is for one specific signal to cause system calls to fail and
1098 return EINTR (for example, to implement a timeout) you can do this with
1102 ??string I've got errors compiling code that uses certain string
1105 {AJ} glibc 2.1 has special string functions that are faster than the normal
1106 library functions. Some of the functions are additionally implemented as
1107 inline functions and others as macros. This might lead to problems with
1108 existing codes but it is explicitly allowed by ISO C.
1110 The optimized string functions are only used when compiling with
1111 optimizations (-O1 or higher). The behavior can be changed with two feature
1114 * __NO_STRING_INLINES: Don't do any string optimizations.
1115 * __USE_STRING_INLINES: Use assembly language inline functions (might
1116 increase code size dramatically).
1118 Since some of these string functions are now additionally defined as macros,
1119 code like "char *strncpy();" doesn't work anymore (and is unnecessary, since
1120 <string.h> has the necessary declarations). Either change your code or
1121 define __NO_STRING_INLINES.
1123 {UD} Another problem in this area is that gcc still has problems on machines
1124 with very few registers (e.g., ix86). The inline assembler code can require
1125 almost all the registers and the register allocator cannot always handle
1128 One can disable the string optimizations selectively. Instead of writing
1130 cp = strcpy (foo, "lkj");
1134 cp = (strcpy) (foo, "lkj");
1136 This disables the optimization for that specific call.
1138 ?? I get compiler messages "Initializer element not constant" with
1139 stdin/stdout/stderr. Why?
1141 {RM,AJ} Constructs like:
1142 static FILE *InPtr = stdin;
1144 lead to this message. This is correct behaviour with glibc since stdin is
1145 not a constant expression. Please note that a strict reading of ISO C does
1146 not allow above constructs.
1148 One of the advantages of this is that you can assign to stdin, stdout, and
1149 stderr just like any other global variable (e.g. `stdout = my_stream;'),
1150 which can be very useful with custom streams that you can write with libio
1151 (but beware this is not necessarily portable). The reason to implement it
1152 this way were versioning problems with the size of the FILE structure.
1154 To fix those programs you've got to initialize the variable at run time.
1155 This can be done, e.g. in main, like:
1163 or by constructors (beware this is gcc specific):
1166 static void inPtr_construct (void) __attribute__((constructor));
1167 static void inPtr_construct (void) { InPtr = stdin; }
1170 ?? I can't compile with gcc -traditional (or
1171 -traditional-cpp). Why?
1173 {AJ} glibc2 does break -traditional and -traditonal-cpp - and will continue
1174 to do so. For example constructs of the form:
1180 are useful for debugging purposes (you can use foo with your debugger that's
1181 why we need the enum) and for compatibility (other systems use defines and
1184 ?? I get some errors with `gcc -ansi'. Isn't glibc ANSI compatible?
1186 {AJ} The GNU C library is compatible with the ANSI/ISO C standard. If
1187 you're using `gcc -ansi', the glibc includes which are specified in the
1188 standard follow the standard. The ANSI/ISO C standard defines what has to be
1189 in the include files - and also states that nothing else should be in the
1190 include files (btw. you can still enable additional standards with feature
1193 The GNU C library is conforming to ANSI/ISO C - if and only if you're only
1194 using the headers and library functions defined in the standard.
1196 ?? I can't access some functions anymore. nm shows that they do
1197 exist but linking fails nevertheless.
1199 {AJ} With the introduction of versioning in glibc 2.1 it is possible to
1200 export only those identifiers (functions, variables) that are really needed
1201 by application programs and by other parts of glibc. This way a lot of
1202 internal interfaces are now hidden. nm will still show those identifiers
1203 but marking them as internal. ISO C states that identifiers beginning with
1204 an underscore are internal to the libc. An application program normally
1205 shouldn't use those internal interfaces (there are exceptions,
1206 e.g. __ivaliduser). If a program uses these interfaces, it's broken. These
1207 internal interfaces might change between glibc releases or dropped
1210 ?? When using the db-2 library which comes with glibc is used in
1211 the Perl db modules the testsuite is not passed. This did not
1212 happen with db-1, gdbm, or ndbm.
1214 {MK} Db-2 does not support zero-sized keys. The Perl testsuite
1215 tests the support for zero-sized keys and therefore fails when db-2 is
1216 used. The Perl folks are looking for a solution, but thus far have
1217 not found a satisfactory one.
1219 ?? The pow() inline function I get when including <math.h> is broken.
1220 I get segmentation faults when I run the program.
1222 {UD} Nope, the implementation is correct. The problem is with egcs version
1223 prior to 1.1. I.e., egcs 1.0 to 1.0.3 are all broken (at least on Intel).
1224 If you have to use this compiler you must define __NO_MATH_INLINES before
1225 including <math.h> to prevent the inline functions from being used. egcs 1.1
1226 fixes the problem. I don't know about gcc 2.8 and 2.8.1.
1228 ?? The sys/sem.h file lacks the definition of `union semun'.
1230 {UD} Nope. This union has to be provided by the user program. Former glibc
1231 versions defined this but it was an error since it does not make much sense
1232 when thinking about it. The standards describing the System V IPC functions
1233 define it this way and therefore programs must be adopted.
1235 ?? Why has <netinet/ip_fw.h> disappeared?
1237 {AJ} The corresponding Linux kernel data structures and constants are
1238 totally different in Linux 2.0 and Linux 2.2. This situation has to be
1239 taken care in user programs using the firewall structures and therefore
1240 those programs (ipfw is AFAIK the only one) should deal with this problem
1243 ?? I get floods of warnings when I use -Wconversion and include
1244 <string.h> or <math.h>.
1246 {ZW} <string.h> and <math.h> intentionally use prototypes to override
1247 argument promotion. -Wconversion warns about all these. You can safely
1248 ignore the warnings.
1250 -Wconversion isn't really intended for production use, only for shakedown
1251 compiles after converting an old program to standard C.
1254 ?? After upgrading to glibc 2.1, I receive errors about
1255 unresolved symbols, like `_dl_initial_searchlist' and can not
1256 execute any binaries. What went wrong?
1258 {AJ} This normally happens if your libc and ld (dynamic linker) are from
1259 different releases of glibc. For example, the dynamic linker
1260 /lib/ld-linux.so.2 comes from glibc 2.0.x, but the version of libc.so.6 is
1263 The path /lib/ld-linux.so.2 is hardcoded in every glibc2 binary but
1264 libc.so.6 is searched via /etc/ld.so.cache and in some special directories
1265 like /lib and /usr/lib. If you run configure with another prefix than /usr
1266 and put this prefix before /lib in /etc/ld.so.conf, your system will break.
1268 So what can you do? Either of the following should work:
1270 * Run `configure' with the same prefix argument you've used for glibc 2.0.x
1271 so that the same paths are used.
1272 * Replace /lib/ld-linux.so.2 with a link to the dynamic linker from glibc
1275 You can even call the dynamic linker by hand if everything fails. You've
1276 got to set LD_LIBRARY_PATH so that the corresponding libc is found and also
1277 need to provide an absolute path to your binary:
1279 LD_LIBRARY_PATH=<path-where-libc.so.6-lives> \
1280 <path-where-corresponding-dynamic-linker-lives>/ld-linux.so.2 \
1281 <path-to-binary>/binary
1283 For example `LD_LIBRARY_PATH=/libold /libold/ld-linux.so.2 /bin/mv ...'
1284 might be useful in fixing a broken system (if /libold contains dynamic
1285 linker and corresponding libc).
1287 With that command line no path is used. To further debug problems with the
1288 dynamic linker, use the LD_DEBUG environment variable, e.g.
1289 `LD_DEBUG=help echo' for the help text.
1291 If you just want to test this release, don't put the lib directory in
1292 /etc/ld.so.conf. You can call programs directly with full paths (as above).
1293 When compiling new programs against glibc 2.1, you've got to specify the
1294 correct paths to the compiler (option -I with gcc) and linker (options
1295 --dynamic-linker, -L and --rpath).
1297 ?? bonnie reports that char i/o with glibc 2 is much slower than with
1298 libc5. What can be done?
1300 {AJ} The GNU C library uses thread safe functions by default and libc5 used
1301 non thread safe versions. The non thread safe functions have in glibc the
1302 suffix `_unlocked', for details check <stdio.h>. Using `putc_unlocked' etc.
1303 instead of `putc' should give nearly the same speed with bonnie (bonnie is a
1304 benchmark program for measuring disk access).
1306 ?? Programs compiled with glibc 2.1 can't read db files made with glibc
1307 2.0. What has changed that programs like rpm break?
1309 {AJ} The GNU C library 2.1 uses db2 instead of db1 which was used in version
1310 2.0. The internal formats of the actual db files are different. To convert
1311 the db files from db1 format to db2 format, you can use the programs
1312 `db_dump185' and `db_load'. Alternativly programs can be linked with db1
1313 using `-ldb1' instead of linking with db2 which uses `-ldb'. Linking with
1314 db1 might be preferable if older programs need to access the db file.
1316 db2 supports the old db1 programming interface and also a new programming
1317 interface. For compilation with the old API, <db_185.h> has to be included
1318 (and not <db.h>) and you can link with either `-ldb1' or `-ldb' for either
1321 ?? Autoconf's AC_CHECK_FUNC macro reports that a function exists, but
1322 when I try to use it, it always returns -1 and sets errno to ENOSYS.
1324 {ZW} You are using a 2.0 Linux kernel, and the function you are trying to
1325 use is only implemented in 2.1/2.2. Libc considers this to be a function
1326 which exists, because if you upgrade to a 2.2 kernel, it will work. One
1327 such function is sigaltstack.
1329 Your program should check at runtime whether the function works, and
1330 implement a fallback. Note that Autoconf cannot detect unimplemented
1331 functions in other systems' C libraries, so you need to do this anyway.
1333 ?? My program segfaults when I call fclose() on the FILE* returned
1334 from setmntent(). Is this a glibc bug?
1336 {GK} No. Don't do this. Use endmntent(), that's what it's for.
1338 In general, you should use the correct deallocation routine. For instance,
1339 if you open a file using fopen(), you should deallocate the FILE * using
1340 fclose(), not free(), even though the FILE * is also a pointer.
1342 In the case of setmntent(), it may appear to work in most cases, but it
1343 won't always work. Unfortunately, for compatibility reasons, we can't
1344 change the return type of setmntent() to something other than FILE *.
1349 ?? After I changed configure.in I get `Autoconf version X.Y.
1350 or higher is required for this script'. What can I do?
1352 {UD} You have to get the specified autoconf version (or a later one)
1353 from your favorite mirror of ftp.gnu.org.
1355 ?? When I try to compile code which uses IPv6 headers and
1356 definitions on my Linux 2.x.y system I am in trouble.
1357 Nothing seems to work.
1359 {UD} The problem is that IPv6 development still has not reached a point
1360 where the headers are stable. There are still lots of incompatible changes
1361 made and the libc headers have to follow.
1363 {PB} The 2.1 release of GNU libc aims to comply with the current versions of
1364 all the relevant standards. The IPv6 support libraries for older Linux
1365 systems used a different naming convention and so code written to work with
1366 them may need to be modified. If the standards make incompatible changes in
1367 the future then the libc may need to change again.
1369 IPv6 will not work with a 2.0.x kernel. When kernel 2.2 is released it
1370 should contain all the necessary support; until then you should use the
1371 latest 2.1.x release you can find. As of 98/11/26 the currently recommended
1372 kernel for IPv6 is 2.1.129.
1374 Also, as of the 2.1 release the IPv6 API provided by GNU libc is not
1375 100% complete. In particular the getipnodebyname and getipnodebyaddr
1376 functions are not implemented.
1378 ??tzdb When I set the timezone by setting the TZ environment variable
1379 to EST5EDT things go wrong since glibc computes the wrong time
1380 from this information.
1382 {UD} The problem is that people still use the braindamaged POSIX method to
1383 select the timezone using the TZ environment variable with a format EST5EDT
1384 or whatever. People, if you insist on using TZ instead of the timezone
1385 database (see below), read the POSIX standard, the implemented behaviour is
1386 correct! What you see is in fact the result of the decisions made while
1387 POSIX.1 was created. We've only implemented the handling of TZ this way to
1388 be POSIX compliant. It is not really meant to be used.
1390 The alternative approach to handle timezones which is implemented is the
1391 correct one to use: use the timezone database. This avoids all the problems
1392 the POSIX method has plus it is much easier to use. Simply run the tzselect
1393 shell script, answer the question and use the name printed in the end by
1394 making a symlink /etc/localtime pointing to /usr/share/zoneinfo/NAME (NAME
1395 is the returned value from tzselect). That's all. You never again have to
1398 So, please avoid sending bug reports about time related problems if you use
1399 the POSIX method and you have not verified something is really broken by
1400 reading the POSIX standards.
1402 ?? What other sources of documentation about glibc are available?
1404 {AJ} The FSF has a page about the GNU C library at
1405 <http://www.gnu.org/software/libc/>. The problem data base of open and
1406 solved bugs in GNU libc is available at
1407 <http://www-gnats.gnu.org:8080/cgi-bin/wwwgnats.pl>. Eric Green has written
1408 a HowTo for converting from Linux libc5 to glibc2. The HowTo is accessable
1409 via the FSF page and at <http://www.imaxx.net/~thrytis/glibc>. Frodo
1410 Looijaard describes a different way installing glibc2 as secondary libc at
1411 <http://huizen.dds.nl/~frodol/glibc>.
1413 Please note that this is not a complete list.
1415 ?? The timezone string for Sydney/Australia is wrong since even when
1416 daylight saving time is in effect the timezone string is EST.
1418 {UD} The problem for some timezones is that the local authorities decided
1419 to use the term "summer time" instead of "daylight saving time". In this
1420 case the abbreviation character `S' is the same as the standard one. So,
1423 Eastern Standard Time = EST
1424 Eastern Summer Time = EST
1426 Great! To get this bug fixed convince the authorities to change the laws
1427 and regulations of the country this effects. glibc behaves correctly.
1429 ??make I've build make 3.77 against glibc 2.1 and now make gets
1430 segmentation faults.
1432 {AJ} GNU make 3.77 has support for 64 bit filesystems which is slightly
1433 broken (and one of the new features in the GNU C library 2.1 is 64 bit
1434 filesystem support :-( ). To get a working make you can use either make
1435 3.75 or patch 3.77. A working patch is available via RedHat's Rawhide server
1436 (ftp://rawhide.redhat.com/SRPMS/SRPMS/make-3.77-*src.rpm).
1439 Answers were given by:
1440 {UD} Ulrich Drepper, <drepper@cygnus.com>
1441 {DMT} David Mosberger-Tang, <davidm@AZStarNet.com>
1442 {RM} Roland McGrath, <roland@gnu.org>
1443 {AJ} Andreas Jaeger, <aj@arthur.rhein-neckar.de>
1444 {EY} Eric Youngdale, <eric@andante.jic.com>
1445 {PB} Phil Blundell, <Philip.Blundell@pobox.com>
1446 {MK} Mark Kettenis, <kettenis@phys.uva.nl>
1447 {ZW} Zack Weinberg, <zack@rabi.phys.columbia.edu>
1448 {TK} Thorsten Kukuk, <kukuk@suse.de>
1449 {GK} Geoffrey Keating, <geoffk@ozemail.com.au>
1450 {HJ} H.J. Lu, <hjl@gnu.org>
1451 {CG} Cristian Gafton, <gafton@redhat.com>
1455 outline-regexp:"\\?"