Frequently Asked Question on GNU C Library As every FAQ this one also tries to answer questions the user might have when using the package. Please make sure you read this before sending questions or bug reports to the maintainers. The GNU C Library is very complex. The building process exploits the features available in tools generally available. But many things can only be done using GNU tools. Also the code is sometimes hard to understand because it has to be portable but on the other hand must be fast. But you need not understand the details to use GNU C Library. This will only be necessary if you intend to contribute or change it. If you have any questions you think should be answered in this document, please let me know. --drepper@cygnus.com ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q1] ``What systems does the GNU C Library run on?'' [Q2] ``What compiler do I need to build GNU libc?'' [Q3] ``When starting make I get only error messages. What's wrong?'' [Q4] ``After I changed configure.in I get `Autoconf version X.Y. or higher is required for this script'. What can I do?'' [Q5] ``Do I need a special linker or archiver?'' [Q6] ``Do I need some more things to compile GNU C Library?'' [Q7] ``When I run `nm -u libc.so' on the produced library I still find unresolved symbols? Can this be ok?'' [Q8] ``Can I replace the libc on my Linux system with GNU libc?'' [Q9] ``I expect GNU libc to be 100% source code compatible with the old Linux based GNU libc. Why isn't it like this?'' [Q10] ``Why does getlogin() always return NULL on my Linux box?'' [Q11] ``Where are the DST_* constants found in on many systems?'' [Q12] ``The `gencat' utility cannot process the input which are successfully used on my Linux libc based system. Why?'' [Q13] ``How do I configure GNU libc so that the essential libraries like libc.so go into /lib and the other into /usr/lib?'' [Q14] ``When linking with the new libc I get unresolved symbols `crypt' and `setkey'. Why aren't these functions in the libc anymore?'' [Q15] ``What are these `add-ons'?'' [Q16] ``When I use GNU libc on my Linux system by linking against to libc.so which comes with glibc all I get is a core dump.'' [Q17] ``Looking through the shared libc file I haven't found the functions `stat', `lstat', `fstat', and `mknod' and while linking on my Linux system I get error messages. How is this supposed to work?'' [Q18] ``The prototypes for `connect', `accept', `getsockopt', `setsockopt', `getsockname', `getpeername', `send', `sendto', and `recvfrom' are different in GNU libc than on any other system I saw. This is a bug, isn't it?'' [Q19] ``My XXX kernel emulates a floating-point coprocessor for me. Should I enable --with-fp?'' [Q20] ``How can I compile gcc 2.7.2.1 from the gcc source code using glibc 2.x? [Q21] ``On Linux I've got problems with the declarations in Linux kernel headers.'' [Q22] ``When I try to compile code which uses IPv6 header and definitions on my Linux 2.x.y system I am in trouble. Nothing seems to work.'' [Q23] ``When compiling GNU libc I get lots of errors saying functions in glibc are duplicated in libgcc.'' [Q24] ``I have set up /etc/nis.conf, and the Linux libc 5 with NYS works great. But the glibc NIS+ doesn't seem to work.'' [Q25] ``After installing glibc name resolving doesn't work properly.'' [Q26] ``I have /usr/include/net and /usr/include/scsi as symlinks into my Linux source tree. Is that wrong?'' [Q27] ``Programs like `logname', `top', `uptime' `users', `w' and `who', show incorrect information about the (number of) users on my system. Why?'' [Q28] ``After upgrading to a glibc 2.1 with symbol versioning I get errors about undefined symbols. What went wrong?'' [Q29] ``I don't include any kernel header myself but still the compiler complains about type redeclarations of types in the kernel headers.'' [Q30] ``When I start the program XXX after upgrading the library I get XXX: Symbol `_sys_errlist' has different size in shared object, consider re-linking Why? What to do?'' [Q31] ``What's the problem with configure --enable-omitfp?'' [Q32] ``Why don't signals interrupt system calls anymore?'' [Q33] ``I've got errors compiling code that uses certain string functions. Why?'' ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q1] ``What systems does the GNU C Library run on?'' [A1] {UD} This is difficult to answer. The file `README' lists the architectures GNU libc is known to run *at some time*. This does not mean that it still can be compiled and run on them in the moment. The systems glibc is known to work on in the moment and most probably in the future are: *-*-gnu GNU Hurd i[3456]86-*-linux-gnu Linux-2.0 on Intel m68k-*-linux-gnu Linux-2.0 on Motorola 680x0 alpha-*-linux-gnu Linux-2.0 on DEC Alpha powerpc-*-linux-gnu Linux and MkLinux on PowerPC systems sparc-*-linux-gnu Linux-2.0 on SPARC sparc64-*-linux-gnu Linux-2.0 on UltraSPARC Other Linux platforms are also on the way to be supported but I need some success reports first. If you have a system not listed above (or in the `README' file) and you are really interested in porting it, contact ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q2] ``What compiler do I need to build GNU libc?'' [A2] {UD} It is (almost) impossible to compile GNU C Library using a different compiler than GNU CC. A lot of extensions of GNU CC are used to increase the portability and speed. But this does not mean you have to use GNU CC for using the GNU C Library. In fact you should be able to use the native C compiler because the success only depends on the binutils: the linker and archiver. The GNU CC is found like all other GNU packages on ftp://prep.ai.mit.edu/pub/gnu or better one of the many mirror sites. You always should try to use the latest official release. Older versions might not have all the features GNU libc could use. It is known that on most platforms compilers earlier than 2.7.2.3 fail so at least use this version. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q3] ``When starting `make' I get only errors messages. What's wrong?'' [A3] {UD} You definitely need GNU make to translate GNU libc. No other make program has the needed functionality. Versions before 3.74 have bugs which prevent correct execution so you should upgrade to the latest version before starting the compilation. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q4] ``After I changed configure.in I get `Autoconf version X.Y. or higher is required for this script'. What can I do?'' [A4] {UD} You have to get the specified autoconf version (or a later) from your favourite mirror of prep.ai.mit.edu. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q5] ``Do I need a special linker or archiver?'' [A5] {UD} If your native versions are not too buggy you can probably work with them. But GNU libc works best with GNU binutils. On systems where the native linker does not support weak symbols you will not get a really ISO C compliant C library. Generally speaking you should use the GNU binutils if they provide at least the same functionality as your system's tools. Always get the newest release of GNU binutils available. Older releases are known to have bugs that affect building the GNU C Library. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q6] ``Do I need some more things to compile GNU C Library?'' [A6] {UD} Yes, there are some more :-). * GNU gettext; the GNU libc is internationalized and partly localized. For bringing the messages for the different languages in the needed form the tools from the GNU gettext package are necessary. See ftp://prep.ai.mit.edu/pub/gnu or better any mirror site. * lots of diskspace (for i?86-linux this means, e.g., ~170MB; for ppc-linux even ~200MB). You should avoid compiling on a NFS mounted device. This is very slow. * plenty of time (approx 1h for i?86-linux on i586@133 or 2.5h on i486@66 or 4.5h on i486@33), both for shared and static only). Multiply this by 1.5 or 2.0 if you build profiling and/or the highly optimized version as well. For Hurd systems times are much higher. For Atari Falcon (Motorola 68030 @ 16 Mhz, 14 Mb memory) James Troup reports for a full build (shared, static, and profiled) a compile time of 45h34m. For Atari TT030 (Motorola 68030 @ 32 Mhz, 34 Mb memory) (full build) a compile time of 22h48m. If you have some more measurements let me know. * When compiling for Linux: + the header files of the Linux kernel must be available in the search path of the CPP as and . * Some files depend on special tools. E.g., files ending in .gperf need a `gperf' program. The GNU version (part of libg++) is known to work while some vendor versions do not. You should not need these tools unless you change the source files. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q7] ``When I run `nm -u libc.so' on the produced library I still find unresolved symbols? Can this be ok?'' [A7] {UD} Yes, this is ok. There can be several kinds of unresolved symbols: * magic symbols automatically generated by the linker. Names are often like __start_* and __stop_* * symbols starting with _dl_* come from the dynamic linker * symbols resolved by using libgcc.a (__udivdi3, __umoddi3, or similar) * weak symbols, which need not be resolved at all (currently fabs among others; this gets resolved if the program is linked against libm, too.) Generally, you should make sure you find a real program which produces errors while linking before deciding there is a problem. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q8] ``Can I replace the libc on my Linux system with GNU libc?'' [A8] {UD} You cannot replace any existing libc for Linux with GNU libc. There are different versions of C libraries and you can run libcs with different major version independently. For Linux there are today two libc versions: libc-4 old a.out libc libc-5 current ELF libc GNU libc will have the major number 6 and therefore you can have this additionally installed. For more information consult documentation for shared library handling. The Makefiles of GNU libc will automatically generate the needed symbolic links which the linker will use. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q9] ``I expect GNU libc to be 100% source code compatible with the old Linux based GNU libc. Why isn't it like this?'' [A9] {DMT,UD} Not every extension in Linux libc's history was well thought-out. In fact it had a lot of problems with standards compliance and with cleanliness. With the introduction of a new version number these errors now can be corrected. Here is a list of the known source code incompatibilities: * _GNU_SOURCE: glibc does not automatically define _GNU_SOURCE. Thus, if a program depends on GNU extensions or some other non-standard functionality, it is necessary to compile it with C compiler option -D_GNU_SOURCE, or better, to put `#define _GNU_SOURCE' at the beginning of your source files, before any C library header files are included. This difference normally manifests itself in the form of missing prototypes and/or data type definitions. Thus, if you get such errors, the first thing you should do is try defining _GNU_SOURCE and see if that makes the problem go away. For more information consult the file `NOTES' part of the GNU C library sources. * reboot(): GNU libc sanitizes the interface of reboot() to be more compatible with the interface used on other OSes. In particular, reboot() as implemented in glibc takes just one argument. This argument corresponds to the third argument of the Linux reboot system call. That is, a call of the form reboot(a, b, c) needs to be changed into reboot(c). Beside this the header defines the needed constants for the argument. These RB_* constants should be used instead of the cryptic magic numbers. * swapon(): the interface of this function didn't changed, but the prototype is in a separate header file . For the additional argument of swapon() you should use the SWAP_* constants from , which get defined when is included. * errno: If a program uses variable "errno", then it _must_ include header file . The old libc often (erroneously) declared this variable implicitly as a side-effect of including other libc header files. glibc is careful to avoid such namespace pollution, which, in turn, means that you really need to include the header files that you depend on. This difference normally manifests itself in the form of the compiler complaining about the references of the undeclared symbol "errno". * Linux-specific syscalls: All Linux system calls now have appropriate library wrappers and corresponding declarations in various header files. This is because the syscall() macro that was traditionally used to work around missing syscall wrappers are inherently non-portable and error-prone. The following tables lists all the new syscall stubs, the header-file declaring their interface and the system call name. syscall name: wrapper name: declaring header file: ------------- ------------- ---------------------- bdflush bdflush create_module create_module delete_module delete_module get_kernel_syms get_kernel_syms init_module init_module syslog ksyslog_ctl * lpd: Older versions of lpd depend on a routine called _validuser(). The library does not provide this function, but instead provides __ivaliduser() which has a slightly different interfaces. Simply upgrading to a newer lpd should fix this problem (e.g., the 4.4BSD lpd is known to be working). * resolver functions/BIND: like on many other systems the functions of the resolver library are not included in the libc itself. There is a separate library libresolv. If you find some symbols starting with `res_*' undefined simply add -lresolv to your call of the linker. * the `signal' function's behaviour corresponds to the BSD semantic and not the SysV semantic as it was in libc-5. The interface on all GNU systems shall be the same and BSD is the semantic of choice. To use the SysV behaviour simply use `sysv_signal', or define _XOPEN_SOURCE. See question 32 for details. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q10] ``Why does getlogin() always return NULL on my Linux box?'' [A10] {UD} The GNU C library has a format for the UTMP and WTMP file which differs from what your system currently has. It was extended to fulfill the needs of the next years when IPv6 is introduced. So the record size is different, fields might have a different position and so reading the files written by functions from the one library cannot be read by functions from the other library. Sorry, but this is what a major release is for. It's better to have a cut now than having no means to support the new techniques later. {MK} There is however a (partial) solution for this problem. Please take a look at the file `README.utmpd'. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q11] ``Where are the DST_* constants found in on many systems?'' [A11] {UD} These constants come from the old BSD days and are not used today anymore (even the Linux based glibc does not implement the handling although the constants are defined). Instead GNU libc contains the zone database handling and compatibility code for POSIX TZ environment variable handling. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q12] ``The `gencat' utility cannot process the input which are successfully used on my Linux libc based system. Why?'' [A12] {UD} Unlike the author of the `gencat' program which is distributed with Linux libc I have read the underlying standards before writing the code. It is completely compatible with the specification given in X/Open Portability Guide. To ease the transition from the Linux version some of the non-standard features are also present in the `gencat' program of GNU libc. This mainly includes the use of symbols for the message number and the automatic generation of header files which contain the needed #defines to map the symbols to integers. Here is a simple SED script to convert at least some Linux specific catalog files to the XPG4 form: ----------------------------------------------------------------------- # Change catalog source in Linux specific format to standard XPG format. # Ulrich Drepper , 1996. # /^\$ #/ { h s/\$ #\([^ ]*\).*/\1/ x s/\$ #[^ ]* *\(.*\)/\$ \1/ } /^# / { s/^# \(.*\)/\1/ G s/\(.*\)\n\(.*\)/\2 \1/ } ----------------------------------------------------------------------- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q13] ``How do I configure GNU libc so that the essential libraries like libc.so go into /lib and the other into /usr/lib?'' [A13] {UD,AJ} Like all other GNU packages GNU libc is configured to use a base directory and install all files relative to this. If you intend to really use GNU libc on your system this base directory is /usr. I.e., you run configure --prefix=/usr Some systems like Linux have a filesystem standard which makes a difference between essential libraries and others. Essential libraries are placed in /lib because this directory is required to be located on the same disk partition as /. The /usr subtree might be found on another partition/disk. To install the essential libraries which come with GNU libc in /lib one must explicitly tell this (except on Linux, see below). Autoconf has no option for this so you have to use the file where all user supplied additional information should go in: `configparms' (see the `INSTALL' file). Therefore the `configparms' file should contain: slibdir=/lib sysconfdir=/etc The first line specifies the directory for the essential libraries, the second line the directory for file which are by tradition placed in a directory named /etc. No rule without an exception: If you configure for Linux with --prefix=/usr, then slibdir and sysconfdir will automatically be defined as stated above. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q14] ``When linking with the new libc I get unresolved symbols `crypt' and `setkey'. Why aren't these functions in the libc anymore?'' [A14] {UD} Remember the US restrictions of exporting cryptographic programs and source code. Until this law gets abolished we cannot ship the cryptographic function together with the libc. But of course we provide the code and there is an very easy way to use this code. First get the extra package. People in the US may get it from the same place they got the GNU libc from. People outside the US should get the code from ftp://ftp.ifi.uio.no/pub/gnu, or another archive site outside the USA. The README explains how to install the sources. If you already have the crypt code on your system the reason for the failure is probably that you failed to link with -lcrypt. The crypto functions are in a separate library to make it possible to export GNU libc binaries from the US. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q15] ``What are these `add-ons'?'' [A15] {UD} To avoid complications with export rules or external source code some optional parts of the libc are distributed as separate packages (e.g., the crypt package, see Q14). To ease the use as part of GNU libc the installer just has to unpack the package and tell the configuration script about these additional subdirectories using the --enable-add-ons option. When you add the crypt add-on you just have to use configure --enable-add-ons=crypt,XXX ... where XXX are possible other add-ons and ... means the rest of the normal option list. You can use add-ons also to overwrite some files in glibc. The add-on system dependent subdirs are search first. It is also possible to add banner files (use a file named `Banner') or create shared libraries. Using add-ons has the big advantage that the makefiles of the GNU libc can be used. Only some few stub rules must be written to get everything running. Even handling of architecture dependent compilation is provided. The GNU libc's sysdeps/ directory shows how to use this feature. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q16] ``When I use GNU libc on my Linux system by linking against to libc.so which comes with glibc all I get is a core dump.'' [A16] {UD} It is not enough to simply link against the GNU libc library itself. The GNU C library comes with its own dynamic linker which really conforms to the ELF API standard. This dynamic linker must be used. Normally this is done by the compiler. The gcc will use -dynamic-linker /lib/ld-linux.so.1 unless the user specifies her/himself a -dynamic-linker argument. But this is not the correct name for the GNU dynamic linker. The correct name is /lib/ld.so.1 which is the name specified in the SVr4 ABi. To change your environment to use GNU libc for compiling you need to change the `specs' file of your gcc. This file is normally found at /usr/lib/gcc-lib///specs In this file you have to change a few things: - change `ld-linux.so.1' to `ld.so.1' (or to ld-linux.so.2, see below) - remove all expression `%{...:-lgmon}'; there is no libgmon in glibc - fix a minor bug by changing %{pipe:-} to %| Things are getting a bit more complicated if you have GNU libc installed in some other place than /usr, i.e., if you do not want to use it instead of the old libc. In this case the needed startup files and libraries are not found in the regular places. So the specs file must tell the compiler and linker exactly what to use. Here is what the gcc-2.7.2 specs file should look like when GNU libc is installed at /usr: ----------------------------------------------------------------------- *asm: %{V} %{v:%{!V:-V}} %{Qy:} %{!Qn:-Qy} %{n} %{T} %{Ym,*} %{Yd,*} %{Wa,*:%*} *asm_final: %| *cpp: %{fPIC:-D__PIC__ -D__pic__} %{fpic:-D__PIC__ -D__pic__} %{!m386:-D__i486__} %{posix:-D_POSIX_SOURCE} %{pthread:-D_REENTRANT} *cc1: %{profile:-p} *cc1plus: *endfile: %{!shared:crtend.o%s} %{shared:crtendS.o%s} crtn.o%s *link: -m elf_i386 %{shared:-shared} %{!shared: %{!ibcs: %{!static: %{rdynamic:-export-dynamic} %{!dynamic-linker:-dynamic-linker /lib/ld-linux.so.2}} %{static:-static}}} *lib: %{!shared: %{pthread:-lpthread} %{profile:-lc_p} %{!profile: -lc}} *libgcc: -lgcc *startfile: %{!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} *switches_need_spaces: *signed_char: %{funsigned-char:-D__CHAR_UNSIGNED__} *predefines: -D__ELF__ -Dunix -Di386 -Dlinux -Asystem(unix) -Asystem(posix) -Acpu(i386) -Amachine(i386) *cross_compile: 0 *multilib: . ; ----------------------------------------------------------------------- The above is currently correct for ix86/Linux. Because of compatibility issues on this platform the dynamic linker must have a different name: ld-linux.so.2. So you have to replace %{!dynamic-linker:-dynamic-linker=/home/gnu/lib/ld-linux.so.2} by %{!dynamic-linker:-dynamic-linker=/home/gnu/lib/ld.so.1} in the above example specs file to make it work for other systems. Version 2.7.2.3 does and future versions of GCC will automatically provide the correct specs. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q17] ``Looking through the shared libc file I haven't found the functions `stat', `lstat', `fstat', and `mknod' and while linking on my Linux system I get error messages. How is this supposed to work?'' [A17] {RM} Believe it or not, stat and lstat (and fstat, and mknod) are supposed to be undefined references in libc.so.6! Your problem is probably a missing or incorrect /usr/lib/libc.so file; note that this is a small text file now, not a symlink to libc.so.6. It should look something like this: GROUP ( libc.so.6 ld.so.1 libc.a ) or in ix86/Linux and alpha/Linux: GROUP ( libc.so.6 ld-linux.so.2 libc.a ) ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q18] ``The prototypes for `connect', `accept', `getsockopt', `setsockopt', `getsockname', `getpeername', `send', `sendto', and `recvfrom' are different in GNU libc from any other system I saw. This is a bug, isn't it?'' [A18] {UD} No, this is no bug. This version of the GNU libc already follows the Single Unix specifications (and I think the POSIX.1g draft which adopted the solution). The type for parameter describing a size is now `socklen_t', a new type. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q19] ``My XXX kernel emulates a floating-point coprocessor for me. Should I enable --with-fp?'' [A19] {UD} As `configure --help' shows the default value is `yes' and this should not be changed unless the FPU instructions would be invalid. I.e., an emulated FPU is for the libc as good as a real one. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q20] ``How can I compile gcc 2.7.2.1 from the gcc source code using glibc 2.x? [A20] {AJ} There's only correct support for glibc 2.0.x in gcc 2.7.2.3 or later. You should get at least gcc 2.7.2.3. All previous versions had problems with glibc support. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q21] ``On Linux I've got problems with the declarations in Linux kernel headers.'' [A21] {UD,AJ} On Linux, the use of kernel headers is reduced to a very minimum. Besides giving Linus the possibility to change the headers more freely it has another reason: user level programs now do not always use the same types like the kernel does. I.e., the libc abstracts the use of types. E.g., the sigset_t type is in the kernel 32 or 64 bits wide. In glibc it is 1024 bits wide, in preparation for future development. The reasons are obvious: we don't want to have a new major release when the Linux kernel gets these functionality. Consult the headers for more information about the changes. Therefore you shouldn't include Linux kernel header files directly if glibc has defined a replacement. Otherwise you might get undefined results because of type conflicts. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q22] ``When I try to compile code which uses IPv6 header and definitions on my Linux 2.x.y system I am in trouble. Nothing seems to work.'' [A22] {UD} The problem is that the IPv6 development still has not reached a point where it is stable. There are still lots of incompatible changes made and the libc headers have to follow. Currently (as of 970401) according to Philip Blundell the required kernel version is 2.1.30. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q23] ``When compiling GNU libc I get lots of errors saying functions in glibc are duplicated in libgcc.'' [A23] {EY} This is *exactly* the same problem that I was having. The problem was due to the fact that the autoconfigure didn't correctly detect that linker flag --no-whole-archive was supported in my linker. In my case it was because I had run ./configure with bogus CFLAGS, and the test failed. One thing that is particularly annoying about this problem is that once this is misdetected, running configure again won't fix it unless you first delete config.cache. {UD} Starting with glibc-2.0.3 there should be a better test to avoid some problems of this kind. The setting of CFLAGS is checked at the very beginning and if it is not usable `configure' will bark. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q24] ``I have set up /etc/nis.conf, and the Linux libc 5 with NYS works great. But the glibc NIS+ doesn't seem to work.'' [A24] The glibc NIS+ implementation uses a /var/nis/NIS_COLD_START file for storing information about the NIS+ server and their public keys, because the nis.conf file do not contain all necessary information. You have to copy a NIS_COLD_START file from a Solaris client (the NIS_COLD_START file is byte order independend) or generate it new with nisinit from the nis-tools (look at http://www-vt.uni-paderborn.de/~kukuk/linux/nisplus.html). ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q25] ``After installing glibc name resolving doesn't work properly.'' [A25] {AJ} You probable should read the manual section describing ``nsswitch.conf'' (just type `info libc "NSS Configuration File"'). The NSS configuration file is usually the culprit. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q26] ``I have /usr/include/net and /usr/include/scsi as symlinks into my Linux source tree. Is that wrong?'' [A26] {PB} This was necessary for libc5, but is not correct when using glibc. Including the kernel header files directly in user programs usually does not work (see Q21). glibc provides its own and header files to replace them, and you may have to remove any symlink that you have in place before you install glibc. However, /usr/include/asm and /usr/include/linux should remain as they were. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q27] ``Programs like `logname', `top', `uptime' `users', `w' and `who', show incorrect information about the (number of) users on my system. Why?'' [A27] {MK} See Q10. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q28] ``After upgrading to a glibc 2.1 with symbol versioning I get errors about undefined symbols. What went wrong?'' [A28] {AJ} In a versioned libc a lot of symbols are now local that have been global symbols in previous versions. When defining a extern variable both in a user program and extern in the libc the links resolves this to only one reference - the one in the library. The problem is caused by either wrong program code or tools. In no case the global variables from libc should be used by any program. Since these reference are now local, you might see a message like: "msgfmt: error in loading shared libraries: : undefined symbol: _nl_domain_bindings" The only way to fix this is to recompile your program. Sorry, that's the price you might have to pay once for quite a number of advantages with symbol versioning. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q29] ``I don't include any kernel header myself but still the compiler complains about type redeclarations of types in the kernel headers.'' [A29] {UD} The kernel headers before Linux 2.1.61 don't work correctly with glibc since they pollute the name space in a not acceptable way. Compiling C programs is possible in most cases but especially C++ programs have (due to the change of the name lookups for `struct's) problem. One prominent example is `struct fd_set'. There might be some more problems left but 2.1.61 fixes some of the known ones. See the BUGS file for other known problems. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q30] ``When I start the program XXX after upgrading the library I get XXX: Symbol `_sys_errlist' has different size in shared object, consider re-linking Why? What to do?'' [A30] {UD} As the message says, relink the binary. The problem is that very few symbols from the library can change in size and there is no way to avoid this. _sys_errlist is a good example. Occasionally there are new error numbers added to the kernel and this must be reflected at user level. But this does not mean all programs are doomed once such a change is necessary. Such symbols should normally not be used at all. There are mechanisms to avoid using them. In the case of _sys_errlist, there is the strerror() function which should _always_ be used instead. So the correct fix is to rewrite that part of the application. In some situations (especially when testing a new library release) it might be possible that such a symbol size change slipped in though it must not happen. So in case of doubt report such a warning message as a problem. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q31] ``What's the problem with configure --enable-omitfp?'' [A31] {AJ} When configuring with --enable-omitfp the libraries are build without frame pointers. Some compilers produce in this situation buggy code and therefore we don't advise using it at the moment. If you use --enable-omitfp, you're on your own. If you encounter problems with a library that was build this way, I'll advise you to rebuild the library without --enable-omitfp. If the problem vanishes consider tracking the problem down and report it as compiler failure. Since a library build with --enable-omitfp is undebuggable, a debuggable library is also build - you can recognize it by the suffix "_g" to the library names. The compilation of this extra libraries and the compiler optimizations slow down the build process and need more disk space. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q32] ``Why don't signals interrupt system calls anymore?'' [A32] {ZW} By default GNU libc uses the BSD semantics for signal(), unlike Linux libc 5 which used System V semantics. This is partially for compatibility with other systems and partially because the BSD semantics tend to make programming with signals easier. There are three differences: * BSD-style signals that occur in the middle of a system call do not affect the system call; System V signals cause the system call to fail and set errno to EINTR. * BSD signal handlers remain installed once triggered. System V signal handlers work only once, so one must reinstall them each time. * A BSD signal is blocked during the execution of its handler. In other words, a handler for SIGCHLD (for example) does not need to worry about being interrupted by another SIGCHLD. It may, however, be interrrupted by other signals. There is general consensus that for `casual' programming with signals, the BSD semantics are preferable. You don't need to worry about system calls returning EINTR, and you don't need to worry about the race conditions associated with one-shot signal handlers. If you are porting an old program that relies on the old semantics, you can quickly fix the problem by changing signal() to sysv_signal() throughout. Alternatively, define _XOPEN_SOURCE before including . For new programs, the sigaction() function allows you to specify precisely how you want your signals to behave. All three differences listed above are individually switchable on a per-signal basis with this function. If all you want is for one specific signal to cause system calls to fail and return EINTR (for example, to implement a timeout) you can do this with siginterrupt(). ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ [Q33] ``I've got errors compiling code that uses certain string functions. Why?'' [A33] {AJ} glibc 2.1 has the much asked for optimized string functions that are faster than the normal library functions. Some of the functions are implemented as inline functions and others as macros. The optimized string functions are only used when compiling with optimizations (-O1 or higher). The behaviour can be changed with two feature macros: * __NO_STRING_INLINES: Don't use string optimizations. * __USE_STRING_INLINES: Use also assembler inline functions (might increase code use dramatically). Since some of these string functions are now additionally defined as macros, code like "char *strncpy();" doesn't work anymore (and is even unneccessary since has the necessary declarations). Either change your code or define __NO_STRING_INLINES. {UD} Another problem in this area is that the gcc still has problems on machines with very few registers (e.g., ix86). The inline assembler code sometimes requires many/all registers and the register allocator cannot handle these situation in all cases. If a function is also defined as a macro in the libc headers one can prevent the use of the macro easily. E.g., instead of cp = strcpy (foo, "lkj"); one can write cp = (strcpy) (foo, "lkj"); Using this method one can avoid using the optimizations for selected function calls. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Answers were given by: {UD} Ulrich Drepper, {DMT} David Mosberger-Tang, {RM} Roland McGrath, {HJL} H.J. Lu, {AJ} Andreas Jaeger, {EY} Eric Youngdale, {PB} Phil Blundell, {MK} Mark Kettenis, {ZW} Zack Weinberg, Local Variables: mode:text End: