]>
Commit | Line | Data |
---|---|---|
61952351 UD |
1 | Frequently Asked Questions about the GNU C Library |
2 | ||
f12944ec UD |
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. | |
61952351 | 6 | |
f12944ec | 7 | The GNU C library is very complex. The installation process has not been |
fdacb17d | 8 | completely automated; there are too many variables. You can do substantial |
f12944ec UD |
9 | damage to your system by installing the library incorrectly. Make sure you |
10 | understand what you are undertaking before you begin. | |
61952351 UD |
11 | |
12 | If you have any questions you think should be answered in this document, | |
13 | please let me know. | |
14 | ||
15 | --drepper@cygnus.com | |
16 | \f | |
17 | ? Compiling glibc | |
18 | ||
19 | ?? What systems does the GNU C Library run on? | |
20 | ||
f12944ec UD |
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. | |
61952351 | 24 | |
f12944ec UD |
25 | The systems glibc is known to work on as of this release, and most probably |
26 | in the future, are: | |
61952351 UD |
27 | |
28 | *-*-gnu GNU Hurd | |
bd355af0 UD |
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 | |
61952351 | 32 | powerpc-*-linux-gnu Linux and MkLinux on PowerPC systems |
bd355af0 UD |
33 | sparc-*-linux-gnu Linux-2.x on SPARC |
34 | sparc64-*-linux-gnu Linux-2.x on UltraSPARC | |
a35cb74d | 35 | arm-*-none ARM standalone systems |
cb0509a8 | 36 | arm-*-linux Linux-2.x on ARM |
a35cb74d | 37 | arm-*-linuxaout Linux-2.x on ARM using a.out binaries |
61952351 | 38 | |
f12944ec UD |
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 | |
42 | expressed interest. | |
61952351 | 43 | |
f12944ec UD |
44 | If you have a system not listed above (or in the `README' file) and you are |
45 | really interested in porting it, contact | |
61952351 | 46 | |
b9b49b44 | 47 | <bug-glibc@gnu.org> |
61952351 | 48 | |
57b4b78a | 49 | ??binsize What compiler do I need to build GNU libc? |
61952351 | 50 | |
f12944ec UD |
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. | |
61952351 UD |
53 | |
54 | GNU CC is found, like all other GNU packages, on | |
f12944ec | 55 | |
2eb45444 | 56 | ftp://ftp.gnu.org/pub/gnu |
f12944ec | 57 | |
2eb45444 | 58 | and the many mirror sites. ftp.gnu.org is always overloaded, so try to find |
61952351 UD |
59 | a local mirror first. |
60 | ||
ceb27555 | 61 | You should always try to use the latest official release. Older versions |
f12944ec | 62 | may not have all the features GNU libc requires. The current releases of |
b8f558b7 UD |
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). | |
61952351 | 65 | |
57b4b78a UD |
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: | |
69 | ||
70 | text data bss dec hex filename | |
66f6a52b UD |
71 | egcs-2.93.10 862897 15944 12824 891665 d9b11 libc.so |
72 | gcc-2.8.1 959965 16468 12152 988585 f15a9 libc.so | |
57b4b78a UD |
73 | |
74 | Make up your own decision. | |
d89e7a96 | 75 | |
83f6a990 UD |
76 | GNU CC versions 2.95 and above are derived from egcs, and they may do even |
77 | better. | |
78 | ||
6e8afc1c | 79 | Please note that gcc 2.95 and 2.95.x cannot compile glibc on Alpha due to |
a125d9b4 UD |
80 | problems in the complex float support. |
81 | ||
61952351 UD |
82 | ?? When I try to compile glibc I get only error messages. |
83 | What's wrong? | |
84 | ||
b1418d8f | 85 | {UD} You definitely need GNU make to build GNU libc. No other make |
f12944ec | 86 | program has the needed functionality. |
61952351 | 87 | |
d89e7a96 UD |
88 | We recommend version GNU make version 3.75 or 3.77. Versions before 3.75 |
89 | have bugs and/or are missing features. Version 3.76 has bugs which | |
90 | appear when building big projects like GNU libc. 3.76.1 appears to work but | |
d8a167a5 | 91 | some people have reported problems. If you build GNU make 3.77 from source, |
c882585f | 92 | please read ?make first. |
61952351 | 93 | |
d89e7a96 | 94 | ?? Do I need a special linker or assembler? |
61952351 | 95 | |
d89e7a96 UD |
96 | {ZW} If you want a shared library, you need a linker and assembler that |
97 | understand all the features of ELF, including weak and versioned symbols. | |
98 | The static library can be compiled with less featureful tools, but lacks key | |
99 | features such as NSS. | |
61952351 | 100 | |
d89e7a96 UD |
101 | For Linux or Hurd, you want binutils 2.8.1.0.23, 2.9.1, or 2.9.1.0.15 or |
102 | higher. These are the only versions we've tested and found reliable. Other | |
103 | versions after 2.8.1.0.23 may work but we don't recommend them, especially | |
104 | not when C++ is involved. Earlier versions do not work at all. | |
7fd18ea2 | 105 | |
d89e7a96 UD |
106 | Other operating systems may come with system tools that have all the |
107 | necessary features, but this is moot because glibc hasn't been ported to | |
108 | them. | |
61952351 | 109 | |
8619129f | 110 | ??powerpc Which compiler should I use for powerpc? |
4775243a | 111 | |
83f6a990 UD |
112 | {GK} You want to use at least gcc 2.95 (together with the right versions |
113 | of all the other tools, of course). See also question ?excpt. | |
4775243a | 114 | |
cb0509a8 UD |
115 | ??arm Which tools should I use for ARM? |
116 | ||
117 | {PB} You should use egcs 1.1 or a later version. For ELF systems some | |
118 | changes are needed to the compiler; a patch against egcs-1.1.x can be found | |
119 | at: | |
120 | ||
121 | <ftp://ftp.netwinder.org/users/p/philb/egcs-1.1.1pre2-diff-981126> | |
122 | ||
123 | Binutils 2.9.1.0.16 or later is also required. | |
124 | ||
d89e7a96 | 125 | ?? Do I need some more things to compile the GNU C Library? |
61952351 UD |
126 | |
127 | {UD} Yes, there are some more :-). | |
128 | ||
129 | * GNU gettext. This package contains the tools needed to construct | |
130 | `message catalog' files containing translated versions of system | |
2eb45444 | 131 | messages. See ftp://ftp.gnu.org/pub/gnu or better any mirror |
61952351 | 132 | site. (We distribute compiled message catalogs, but they may not be |
2f2f52f7 UD |
133 | updated in patches.) Please note that the required minimal version |
134 | (0.10.35) of gettext is alpha software and available from | |
135 | ftp://alpha.gnu.org/gnu . | |
61952351 | 136 | |
d89e7a96 UD |
137 | * Some files are built with special tools. E.g., files ending in .gperf |
138 | need a `gperf' program. The GNU version (now available in a separate | |
139 | package, formerly only as part of libg++) is known to work while some | |
140 | vendor versions do not. | |
61952351 UD |
141 | |
142 | You should not need these tools unless you change the source files. | |
143 | ||
d89e7a96 UD |
144 | * Perl 5 is needed if you wish to test an installation of GNU libc |
145 | as the primary C library. | |
bd355af0 | 146 | |
61952351 UD |
147 | * When compiling for Linux, the header files of the Linux kernel must |
148 | be available to the compiler as <linux/*.h> and <asm/*.h>. | |
149 | ||
02228370 | 150 | * lots of disk space (~400MB for i?86-linux; more for RISC platforms). |
61952351 UD |
151 | |
152 | * plenty of time. Compiling just the shared and static libraries for | |
16b0f634 UD |
153 | i?86-linux takes approximately 1h on an AMD-K6@225MHz w/ 96MB of RAM, |
154 | 45mins on a Celeron@400MHz w/ 128MB, and 55mins on a Alpha@533MHz w/ 256MB. | |
7c2b945e UD |
155 | Multiply this by 1.5 or 2.0 if you build profiling and/or the highly |
156 | optimized version as well. For Hurd systems times are much higher. | |
61952351 UD |
157 | |
158 | You should avoid compiling in a NFS mounted filesystem. This is | |
159 | very slow. | |
160 | ||
161 | James Troup <J.J.Troup@comp.brad.ac.uk> reports a compile time of | |
bd355af0 UD |
162 | 45h34m for a full build (shared, static, and profiled) on Atari |
163 | Falcon (Motorola 68030 @ 16 Mhz, 14 Mb memory) and Jan Barte | |
164 | <yann@plato.uni-paderborn.de> reports 22h48m on Atari TT030 | |
165 | (Motorola 68030 @ 32 Mhz, 34 Mb memory) | |
61952351 | 166 | |
83f6a990 | 167 | A full build of the PowerPC library took 1h on a PowerPC 750@400Mhz w/ |
6e8afc1c | 168 | 64MB of RAM, and about 9h on a 601@60Mhz w/ 72Mb. |
83f6a990 | 169 | |
61952351 UD |
170 | If you have some more measurements let me know. |
171 | ||
d111572f UD |
172 | ?? What version of the Linux kernel headers should be used? |
173 | ||
f12944ec UD |
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 | |
b1418d8f | 180 | new kernel features if you used old kernel headers to compile the GNU C |
f12944ec UD |
181 | library. |
182 | ||
ceb27555 | 183 | {ZW} Even if you are using a 2.0 kernel on your machine, we recommend you |
440d13e2 UD |
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 | |
ceb27555 | 186 | headers to use, give configure the --with-headers switch |
440d13e2 | 187 | (e.g. --with-headers=/usr/src/linux-2.2.0/include). |
ceb27555 | 188 | |
440d13e2 | 189 | Note that you must configure the 2.2 kernel if you do this, otherwise libc |
62595351 | 190 | will be unable to find <linux/version.h>. Just change the current directory |
440d13e2 | 191 | to the root of the 2.2 tree and do `make include/linux/version.h'. |
ceb27555 | 192 | |
f12944ec UD |
193 | ?? The compiler hangs while building iconvdata modules. What's |
194 | wrong? | |
195 | ||
d89e7a96 UD |
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 | |
198 | time. | |
f12944ec | 199 | |
b8f558b7 | 200 | The problem is fixed in egcs 1.1. |
d111572f | 201 | |
61952351 UD |
202 | ?? When I run `nm -u libc.so' on the produced library I still |
203 | find unresolved symbols. Can this be ok? | |
204 | ||
f12944ec | 205 | {UD} Yes, this is ok. There can be several kinds of unresolved symbols: |
61952351 UD |
206 | |
207 | * magic symbols automatically generated by the linker. These have names | |
208 | like __start_* and __stop_* | |
209 | ||
210 | * symbols starting with _dl_* come from the dynamic linker | |
211 | ||
61952351 UD |
212 | * weak symbols, which need not be resolved at all (fabs for example) |
213 | ||
214 | Generally, you should make sure you find a real program which produces | |
215 | errors while linking before deciding there is a problem. | |
216 | ||
217 | ??addon What are these `add-ons'? | |
218 | ||
f12944ec UD |
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). | |
61952351 | 222 | |
f12944ec UD |
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: | |
61952351 UD |
229 | |
230 | configure --enable-add-ons=crypt,linuxthreads | |
231 | ||
232 | for example. | |
233 | ||
f12944ec UD |
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. | |
61952351 | 238 | |
5bb17dca UD |
239 | Most add-ons are tightly coupled to a specific GNU libc version. Please |
240 | check that the add-ons work with the GNU libc. For example the crypt and | |
241 | linuxthreads add-ons have the same numbering scheme as the libc and will in | |
242 | general only work with the corresponding libc. | |
243 | ||
61952351 UD |
244 | ?? My XXX kernel emulates a floating-point coprocessor for me. |
245 | Should I enable --with-fp? | |
246 | ||
f12944ec UD |
247 | {ZW} An emulated FPU is just as good as a real one, as far as the C library |
248 | is concerned. You only need to say --without-fp if your machine has no way | |
249 | to execute floating-point instructions. | |
61952351 UD |
250 | |
251 | People who are interested in squeezing the last drop of performance | |
252 | out of their machine may wish to avoid the trap overhead, but this is | |
253 | far more trouble than it's worth: you then have to compile | |
254 | *everything* this way, including the compiler's internal libraries | |
255 | (libgcc.a for GNU C), because the calling conventions change. | |
256 | ||
257 | ?? When compiling GNU libc I get lots of errors saying functions | |
258 | in glibc are duplicated in libgcc. | |
259 | ||
f12944ec UD |
260 | {EY} This is *exactly* the same problem that I was having. The problem was |
261 | due to the fact that configure didn't correctly detect that the linker flag | |
262 | --no-whole-archive was supported in my linker. In my case it was because I | |
263 | had run ./configure with bogus CFLAGS, and the test failed. | |
61952351 | 264 | |
f12944ec UD |
265 | One thing that is particularly annoying about this problem is that once this |
266 | is misdetected, running configure again won't fix it unless you first delete | |
267 | config.cache. | |
61952351 | 268 | |
f12944ec UD |
269 | {UD} Starting with glibc-2.0.3 there should be a better test to avoid some |
270 | problems of this kind. The setting of CFLAGS is checked at the very | |
271 | beginning and if it is not usable `configure' will bark. | |
61952351 | 272 | |
74015205 | 273 | ?? Why do I get messages about missing thread functions when I use |
da2d1bc5 | 274 | librt? I don't even use threads. |
74015205 | 275 | |
da2d1bc5 UD |
276 | {UD} In this case you probably mixed up your installation. librt uses |
277 | threads internally and has implicit references to the thread library. | |
f12944ec UD |
278 | Normally these references are satisfied automatically but if the thread |
279 | library is not in the expected place you must tell the linker where it is. | |
280 | When using GNU ld it works like this: | |
74015205 UD |
281 | |
282 | gcc -o foo foo.c -Wl,-rpath-link=/some/other/dir -lrt | |
283 | ||
f12944ec UD |
284 | The `/some/other/dir' should contain the thread library. `ld' will use the |
285 | given path to find the implicitly referenced library while not disturbing | |
286 | any other link path. | |
74015205 | 287 | |
61952351 UD |
288 | ?? What's the problem with configure --enable-omitfp? |
289 | ||
290 | {AJ} When --enable-omitfp is set the libraries are built without frame | |
fdacb17d | 291 | pointers. Some compilers produce buggy code for this model and therefore we |
f12944ec | 292 | don't advise using it at the moment. |
61952351 | 293 | |
fdacb17d | 294 | If you use --enable-omitfp, you're on your own. If you encounter problems |
f12944ec UD |
295 | with a library that was build this way, we advise you to rebuild the library |
296 | without --enable-omitfp. If the problem vanishes consider tracking the | |
297 | problem down and report it as compiler failure. | |
61952351 | 298 | |
b1418d8f UD |
299 | Since a library built with --enable-omitfp is undebuggable on most systems, |
300 | debuggable libraries are also built - you can use them by appending "_g" to | |
f12944ec | 301 | the library names. |
61952351 | 302 | |
f12944ec UD |
303 | The compilation of these extra libraries and the compiler optimizations slow |
304 | down the build process and need more disk space. | |
61952351 | 305 | |
b1418d8f | 306 | ?? I get failures during `make check'. What should I do? |
b0610668 | 307 | |
b1418d8f UD |
308 | {AJ} The testsuite should compile and run cleanly on your system; every |
309 | failure should be looked into. Depending on the failures, you probably | |
310 | should not install the library at all. | |
b0610668 UD |
311 | |
312 | You should consider using the `glibcbug' script to report the failure, | |
313 | providing as much detail as possible. If you run a test directly, please | |
314 | remember to set up the environment correctly. You want to test the compiled | |
315 | library - and not your installed one. The best way is to copy the exact | |
316 | command line which failed and run the test from the subdirectory for this | |
317 | test in the sources. | |
318 | ||
319 | There are some failures which are not directly related to the GNU libc: | |
b1418d8f UD |
320 | - Some compilers produce buggy code. No compiler gets single precision |
321 | complex numbers correct on Alpha. Otherwise, the egcs 1.1 release should be | |
322 | ok; gcc 2.8.1 might cause some failures; gcc 2.7.2.x is so buggy that | |
323 | explicit checks have been used so that you can't build with it. | |
b0610668 UD |
324 | - The kernel might have bugs. For example on Linux/Alpha 2.0.34 the |
325 | floating point handling has quite a number of bugs and therefore most of | |
440d13e2 | 326 | the test cases in the math subdirectory will fail. Linux 2.2 has |
b1418d8f UD |
327 | fixes for the floating point support on Alpha. The Linux/SPARC kernel has |
328 | also some bugs in the FPU emulation code (as of Linux 2.2.0). | |
d32a4020 UD |
329 | - Other tools might have problems. For example bash 2.03 gives a |
330 | segmentation fault running the tst-rpmatch.sh test script. | |
b0610668 | 331 | |
7fd18ea2 UD |
332 | ?? What is symbol versioning good for? Do I need it? |
333 | ||
334 | {AJ} Symbol versioning solves problems that are related to interface | |
335 | changes. One version of an interface might have been introduced in a | |
336 | previous version of the GNU C library but the interface or the semantics of | |
337 | the function has been changed in the meantime. For binary compatibility | |
338 | with the old library, a newer library needs to still have the old interface | |
b1418d8f | 339 | for old programs. On the other hand, new programs should use the new |
7fd18ea2 | 340 | interface. Symbol versioning is the solution for this problem. The GNU |
b1418d8f UD |
341 | libc version 2.1 uses symbol versioning by default if the installed binutils |
342 | supports it. | |
7fd18ea2 | 343 | |
b1418d8f UD |
344 | We don't advise building without symbol versioning, since you lose binary |
345 | compatibility - forever! The binary compatibility you lose is not only | |
346 | against the previous version of the GNU libc (version 2.0) but also against | |
347 | all future versions. | |
7fd18ea2 | 348 | |
bee1e289 UD |
349 | ?? How can I compile on my fast ix86 machine a working libc for my slow |
350 | i386? After installing libc, programs abort with "Illegal | |
351 | Instruction". | |
352 | ||
353 | {AJ} glibc and gcc might generate some instructions on your machine that | |
354 | aren't available on i386. You've got to tell glibc that you're configuring | |
355 | for i386 with adding i386 as your machine, for example: | |
356 | ||
357 | ../configure --prefix=/usr i386-pc-linux-gnu | |
358 | ||
359 | And you need to tell gcc to only generate i386 code, just add `-mcpu=i386' | |
360 | (just -m386 doesn't work) to your CFLAGS. | |
361 | ||
362 | {UD} This applies not only to the i386. Compiling on a i686 for any older | |
363 | model will also fail if the above methods are not used. | |
364 | ||
b2b28911 UD |
365 | ?? `make' complains about a missing dlfcn/libdl.so when building |
366 | malloc/libmemprof.so. How can I fix this? | |
367 | ||
368 | {AJ} Older make version (<= 3.78.90) have a bug which was hidden by a bug in | |
369 | glibc (<= 2.1.2). You need to upgrade make to a newer or fixed version. A | |
370 | patch is available via | |
371 | <http://sourceware.cygnus.com//ml/libc-alpha/2000-02/msg00067.html>. | |
372 | ||
373 | After upgrading make, you should remove the file sysd-sorted in your build | |
374 | directory. The problem is that the broken make creates a wrong order for | |
375 | one list in that file. The list has to be recreated with the new make - | |
376 | which happens if you remove the file. | |
377 | ||
378 | You might encounter this bug also in other situations where make scans | |
379 | directories. I strongly advise to upgrade your make version. | |
380 | ||
bee1e289 | 381 | |
61952351 UD |
382 | ? Installation and configuration issues |
383 | ||
384 | ?? Can I replace the libc on my Linux system with GNU libc? | |
385 | ||
f12944ec UD |
386 | {UD} You cannot replace any existing libc for Linux with GNU libc. It is |
387 | binary incompatible and therefore has a different major version. You can, | |
388 | however, install it alongside your existing libc. | |
61952351 UD |
389 | |
390 | For Linux there are three major libc versions: | |
391 | libc-4 a.out libc | |
392 | libc-5 original ELF libc | |
393 | libc-6 GNU libc | |
394 | ||
f12944ec UD |
395 | You can have any combination of these three installed. For more information |
396 | consult documentation for shared library handling. The Makefiles of GNU | |
397 | libc will automatically generate the needed symbolic links which the linker | |
398 | will use. | |
61952351 UD |
399 | |
400 | ?? How do I configure GNU libc so that the essential libraries | |
401 | like libc.so go into /lib and the other into /usr/lib? | |
402 | ||
403 | {UD,AJ} Like all other GNU packages GNU libc is designed to use a base | |
404 | directory and install all files relative to this. The default is | |
f12944ec UD |
405 | /usr/local, because this is safe (it will not damage the system if installed |
406 | there). If you wish to install GNU libc as the primary C library on your | |
407 | system, set the base directory to /usr (i.e. run configure --prefix=/usr | |
408 | <other_options>). Note that this can damage your system; see ?safety for | |
409 | details. | |
410 | ||
411 | Some systems like Linux have a filesystem standard which makes a difference | |
412 | between essential libraries and others. Essential libraries are placed in | |
413 | /lib because this directory is required to be located on the same disk | |
414 | partition as /. The /usr subtree might be found on another | |
415 | partition/disk. If you configure for Linux with --prefix=/usr, then this | |
416 | will be done automatically. | |
61952351 UD |
417 | |
418 | To install the essential libraries which come with GNU libc in /lib on | |
f12944ec UD |
419 | systems other than Linux one must explicitly request it. Autoconf has no |
420 | option for this so you have to use a `configparms' file (see the `INSTALL' | |
421 | file for details). It should contain: | |
61952351 UD |
422 | |
423 | slibdir=/lib | |
424 | sysconfdir=/etc | |
425 | ||
f12944ec UD |
426 | The first line specifies the directory for the essential libraries, the |
427 | second line the directory for system configuration files. | |
61952351 UD |
428 | |
429 | ??safety How should I avoid damaging my system when I install GNU libc? | |
430 | ||
f12944ec UD |
431 | {ZW} If you wish to be cautious, do not configure with --prefix=/usr. If |
432 | you don't specify a prefix, glibc will be installed in /usr/local, where it | |
433 | will probably not break anything. (If you wish to be certain, set the | |
434 | prefix to something like /usr/local/glibc2 which is not used for anything.) | |
61952351 UD |
435 | |
436 | The dangers when installing glibc in /usr are twofold: | |
437 | ||
438 | * glibc will overwrite the headers in /usr/include. Other C libraries | |
27e309c1 UD |
439 | install a different but overlapping set of headers there, so the effect |
440 | will probably be that you can't compile anything. You need to rename | |
441 | /usr/include out of the way before running `make install'. (Do not throw | |
442 | it away; you will then lose the ability to compile programs against your | |
443 | old libc.) | |
61952351 UD |
444 | |
445 | * None of your old libraries, static or shared, can be used with a | |
446 | different C library major version. For shared libraries this is not a | |
447 | problem, because the filenames are different and the dynamic linker | |
448 | will enforce the restriction. But static libraries have no version | |
449 | information. You have to evacuate all the static libraries in | |
450 | /usr/lib to a safe location. | |
451 | ||
452 | The situation is rather similar to the move from a.out to ELF which | |
453 | long-time Linux users will remember. | |
454 | ||
455 | ?? Do I need to use GNU CC to compile programs that will use the | |
456 | GNU C Library? | |
457 | ||
f12944ec UD |
458 | {ZW} In theory, no; the linker does not care, and the headers are supposed |
459 | to check for GNU CC before using its extensions to the C language. | |
61952351 | 460 | |
f12944ec UD |
461 | However, there are currently no ports of glibc to systems where another |
462 | compiler is the default, so no one has tested the headers extensively | |
463 | against another compiler. You may therefore encounter difficulties. If you | |
464 | do, please report them as bugs. | |
61952351 UD |
465 | |
466 | Also, in several places GNU extensions provide large benefits in code | |
467 | quality. For example, the library has hand-optimized, inline assembly | |
f12944ec UD |
468 | versions of some string functions. These can only be used with GCC. See |
469 | ?string for details. | |
61952351 UD |
470 | |
471 | ??crypt When linking with the new libc I get unresolved symbols | |
472 | `crypt' and `setkey'. Why aren't these functions in the | |
473 | libc anymore? | |
474 | ||
f12944ec UD |
475 | {UD} The US places restrictions on exporting cryptographic programs and |
476 | source code. Until this law gets abolished we cannot ship the cryptographic | |
477 | functions together with glibc. | |
61952351 | 478 | |
f12944ec UD |
479 | The functions are available, as an add-on (see ?addon). People in the US |
480 | may get it from the same place they got GNU libc from. People outside the | |
f05f5ca3 UD |
481 | US should get the code from ftp.gwdg.de [134.76.11.100] in the directory |
482 | pub/linux/glibc, or another archive site outside the USA. The README explains | |
2f512715 | 483 | how to install the sources. |
61952351 | 484 | |
f12944ec UD |
485 | If you already have the crypt code on your system the reason for the failure |
486 | is probably that you did not link with -lcrypt. The crypto functions are in | |
487 | a separate library to make it possible to export GNU libc binaries from the | |
488 | US. | |
61952351 UD |
489 | |
490 | ?? When I use GNU libc on my Linux system by linking against | |
491 | the libc.so which comes with glibc all I get is a core dump. | |
492 | ||
f12944ec | 493 | {UD} On Linux, gcc sets the dynamic linker to /lib/ld-linux.so.1 unless the |
b3864d70 | 494 | user specifies a --dynamic-linker argument. This is the name of the libc5 |
f12944ec | 495 | dynamic linker, which does not work with glibc. |
61952351 | 496 | |
a379e56a UD |
497 | For casual use of GNU libc you can just specify to the linker |
498 | --dynamic-linker=/lib/ld-linux.so.2 | |
61952351 | 499 | |
f12944ec | 500 | which is the glibc dynamic linker, on Linux systems. On other systems the |
a379e56a UD |
501 | name is /lib/ld.so.1. When linking via gcc, you've got to add |
502 | -Wl,--dynamic-linker=/lib/ld-linux.so.2 | |
503 | ||
504 | to the gcc command line. | |
61952351 | 505 | |
f12944ec UD |
506 | To change your environment to use GNU libc for compiling you need to change |
507 | the `specs' file of your gcc. This file is normally found at | |
61952351 UD |
508 | |
509 | /usr/lib/gcc-lib/<arch>/<version>/specs | |
510 | ||
511 | In this file you have to change a few things: | |
512 | ||
513 | - change `ld-linux.so.1' to `ld-linux.so.2' | |
514 | ||
515 | - remove all expression `%{...:-lgmon}'; there is no libgmon in glibc | |
516 | ||
517 | - fix a minor bug by changing %{pipe:-} to %| | |
518 | ||
f12944ec UD |
519 | Here is what the gcc-2.7.2 specs file should look like when GNU libc is |
520 | installed at /usr: | |
61952351 UD |
521 | |
522 | ----------------------------------------------------------------------- | |
523 | *asm: | |
524 | %{V} %{v:%{!V:-V}} %{Qy:} %{!Qn:-Qy} %{n} %{T} %{Ym,*} %{Yd,*} %{Wa,*:%*} | |
525 | ||
526 | *asm_final: | |
527 | %| | |
528 | ||
529 | *cpp: | |
530 | %{fPIC:-D__PIC__ -D__pic__} %{fpic:-D__PIC__ -D__pic__} %{!m386:-D__i486__} %{posix:-D_POSIX_SOURCE} %{pthread:-D_REENTRANT} | |
531 | ||
532 | *cc1: | |
533 | %{profile:-p} | |
534 | ||
535 | *cc1plus: | |
536 | ||
537 | ||
538 | *endfile: | |
539 | %{!shared:crtend.o%s} %{shared:crtendS.o%s} crtn.o%s | |
540 | ||
541 | *link: | |
542 | -m elf_i386 %{shared:-shared} %{!shared: %{!ibcs: %{!static: %{rdynamic:-export-dynamic} %{!dynamic-linker:-dynamic-linker /lib/ld-linux.so.2}} %{static:-static}}} | |
543 | ||
544 | *lib: | |
545 | %{!shared: %{pthread:-lpthread} %{profile:-lc_p} %{!profile: -lc}} | |
546 | ||
547 | *libgcc: | |
548 | -lgcc | |
549 | ||
550 | *startfile: | |
551 | %{!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} | |
552 | ||
553 | *switches_need_spaces: | |
554 | ||
555 | ||
556 | *signed_char: | |
557 | %{funsigned-char:-D__CHAR_UNSIGNED__} | |
558 | ||
559 | *predefines: | |
560 | -D__ELF__ -Dunix -Di386 -Dlinux -Asystem(unix) -Asystem(posix) -Acpu(i386) -Amachine(i386) | |
561 | ||
562 | *cross_compile: | |
563 | 0 | |
564 | ||
565 | *multilib: | |
566 | . ; | |
567 | ||
568 | ----------------------------------------------------------------------- | |
569 | ||
f12944ec UD |
570 | Things get a bit more complicated if you have GNU libc installed in some |
571 | other place than /usr, i.e., if you do not want to use it instead of the old | |
572 | libc. In this case the needed startup files and libraries are not found in | |
573 | the regular places. So the specs file must tell the compiler and linker | |
574 | exactly what to use. | |
61952351 UD |
575 | |
576 | Version 2.7.2.3 does and future versions of GCC will automatically | |
577 | provide the correct specs. | |
578 | ||
579 | ?? Looking through the shared libc file I haven't found the | |
580 | functions `stat', `lstat', `fstat', and `mknod' and while | |
581 | linking on my Linux system I get error messages. How is | |
582 | this supposed to work? | |
583 | ||
f12944ec UD |
584 | {RM} Believe it or not, stat and lstat (and fstat, and mknod) are supposed |
585 | to be undefined references in libc.so.6! Your problem is probably a missing | |
586 | or incorrect /usr/lib/libc.so file; note that this is a small text file now, | |
587 | not a symlink to libc.so.6. It should look something like this: | |
61952351 | 588 | |
71bedb76 | 589 | GROUP ( libc.so.6 libc_nonshared.a ) |
61952351 | 590 | |
83f6a990 | 591 | ??excpt When I run an executable on one system which I compiled on |
d89e7a96 UD |
592 | another, I get dynamic linker errors. Both systems have the same |
593 | version of glibc installed. What's wrong? | |
594 | ||
595 | {ZW} Glibc on one of these systems was compiled with gcc 2.7 or 2.8, the | |
596 | other with egcs (any version). Egcs has functions in its internal | |
597 | `libgcc.a' to support exception handling with C++. They are linked into | |
598 | any program or dynamic library compiled with egcs, whether it needs them or | |
599 | not. Dynamic libraries then turn around and export those functions again | |
600 | unless special steps are taken to prevent them. | |
601 | ||
602 | When you link your program, it resolves its references to the exception | |
603 | functions to the ones exported accidentally by libc.so. That works fine as | |
604 | long as libc has those functions. On the other system, libc doesn't have | |
605 | those functions because it was compiled by gcc 2.8, and you get undefined | |
606 | symbol errors. The symbols in question are named things like | |
607 | `__register_frame_info'. | |
608 | ||
609 | For glibc 2.0, the workaround is to not compile libc with egcs. We've also | |
610 | incorporated a patch which should prevent the EH functions sneaking into | |
611 | libc. It doesn't matter what compiler you use to compile your program. | |
612 | ||
613 | For glibc 2.1, we've chosen to do it the other way around: libc.so | |
614 | explicitly provides the EH functions. This is to prevent other shared | |
57b4b78a UD |
615 | libraries from doing it. |
616 | ||
617 | {UD} Starting with glibc 2.1.1 you can compile glibc with gcc 2.8.1 or | |
618 | newer since we have explicitly add references to the functions causing the | |
619 | problem. But you nevertheless should use EGCS for other reasons | |
620 | (see ?binsize). | |
d89e7a96 | 621 | |
83f6a990 UD |
622 | {GK} On some Linux distributions for PowerPC, you can see this when you have |
623 | built gcc or egcs from the Web sources (gcc versions 2.95 or earlier), then | |
624 | re-built glibc. This happens because in these versions of gcc, exception | |
625 | handling is implemented using an older method; the people making the | |
626 | distributions are a little ahead of their time. | |
627 | ||
628 | A quick solution to this is to find the libgcc.a file that came with the | |
6e8afc1c | 629 | distribution (it would have been installed under /usr/lib/gcc-lib), do |
83f6a990 UD |
630 | `ar x libgcc.a frame.o' to get the frame.o file out, and add a line saying |
631 | `LDLIBS-c.so += frame.o' to the file `configparms' in the directory you're | |
632 | building in. You can check you've got the right `frame.o' file by running | |
633 | `nm frame.o' and checking that it has the symbols defined that you're | |
634 | missing. | |
635 | ||
636 | This will let you build glibc with the C compiler. The C++ compiler | |
637 | will still be binary incompatible with any C++ shared libraries that | |
638 | you got with your distribution. | |
639 | ||
61952351 UD |
640 | ?? How can I compile gcc 2.7.2.1 from the gcc source code using |
641 | glibc 2.x? | |
642 | ||
f12944ec | 643 | {AJ} There's only correct support for glibc 2.0.x in gcc 2.7.2.3 or later. |
57b4b78a | 644 | But you should get at least gcc 2.8.1 or egcs 1.1 (or later versions) |
f12944ec | 645 | instead. |
61952351 UD |
646 | |
647 | ?? The `gencat' utility cannot process the catalog sources which | |
648 | were used on my Linux libc5 based system. Why? | |
649 | ||
f12944ec UD |
650 | {UD} The `gencat' utility provided with glibc complies to the XPG standard. |
651 | The older Linux version did not obey the standard, so they are not | |
652 | compatible. | |
61952351 UD |
653 | |
654 | To ease the transition from the Linux version some of the non-standard | |
f12944ec UD |
655 | features are also present in the `gencat' program of GNU libc. This mainly |
656 | includes the use of symbols for the message number and the automatic | |
61952351 UD |
657 | generation of header files which contain the needed #defines to map the |
658 | symbols to integers. | |
659 | ||
f12944ec UD |
660 | Here is a simple SED script to convert at least some Linux specific catalog |
661 | files to the XPG4 form: | |
61952351 UD |
662 | |
663 | ----------------------------------------------------------------------- | |
664 | # Change catalog source in Linux specific format to standard XPG format. | |
665 | # Ulrich Drepper <drepper@cygnus.com>, 1996. | |
666 | # | |
667 | /^\$ #/ { | |
668 | h | |
669 | s/\$ #\([^ ]*\).*/\1/ | |
670 | x | |
671 | s/\$ #[^ ]* *\(.*\)/\$ \1/ | |
672 | } | |
673 | ||
674 | /^# / { | |
675 | s/^# \(.*\)/\1/ | |
676 | G | |
677 | s/\(.*\)\n\(.*\)/\2 \1/ | |
678 | } | |
679 | ----------------------------------------------------------------------- | |
680 | ||
da2d1bc5 UD |
681 | ?? Programs using libc have their messages translated, but other |
682 | behavior is not localized (e.g. collating order); why? | |
683 | ||
684 | {ZW} Translated messages are automatically installed, but the locale | |
f12944ec UD |
685 | database that controls other behaviors is not. You need to run localedef to |
686 | install this database, after you have run `make install'. For example, to | |
687 | set up the French Canadian locale, simply issue the command | |
da2d1bc5 UD |
688 | |
689 | localedef -i fr_CA -f ISO-8859-1 fr_CA | |
690 | ||
691 | Please see localedata/README in the source tree for further details. | |
692 | ||
61952351 UD |
693 | ?? I have set up /etc/nis.conf, and the Linux libc 5 with NYS |
694 | works great. But the glibc NIS+ doesn't seem to work. | |
695 | ||
f12944ec UD |
696 | {TK} The glibc NIS+ implementation uses a /var/nis/NIS_COLD_START file for |
697 | storing information about the NIS+ server and their public keys, because the | |
698 | nis.conf file does not contain all the necessary information. You have to | |
699 | copy a NIS_COLD_START file from a Solaris client (the NIS_COLD_START file is | |
700 | byte order independent) or generate it with nisinit from the nis-tools | |
701 | package; available at | |
702 | ||
50f301a8 | 703 | http://www.suse.de/~kukuk/linux/nisplus.html |
61952351 | 704 | |
da2d1bc5 | 705 | ?? I have killed ypbind to stop using NIS, but glibc |
3dcf8ea6 | 706 | continues using NIS. |
4d06461a | 707 | |
f12944ec UD |
708 | {TK} For faster NIS lookups, glibc uses the /var/yp/binding/ files from |
709 | ypbind. ypbind 3.3 and older versions don't always remove these files, so | |
710 | glibc will continue to use them. Other BSD versions seem to work correctly. | |
711 | Until ypbind 3.4 is released, you can find a patch at | |
712 | ||
66f6a52b | 713 | <ftp://ftp.kernel.org/pub/linux/utils/net/NIS/ypbind-3.3-glibc4.diff.gz> |
a788b6c2 | 714 | |
3dcf8ea6 UD |
715 | ?? Under Linux/Alpha, I always get "do_ypcall: clnt_call: |
716 | RPC: Unable to receive; errno = Connection refused" when using NIS. | |
a788b6c2 | 717 | |
f12944ec UD |
718 | {TK} You need a ypbind version which is 64bit clean. Some versions are not |
719 | 64bit clean. A 64bit clean implementation is ypbind-mt. For ypbind 3.3, | |
720 | you need the patch from ftp.kernel.org (See the previous question). I don't | |
721 | know about other versions. | |
a788b6c2 | 722 | |
4d06461a | 723 | |
61952351 UD |
724 | ?? After installing glibc name resolving doesn't work properly. |
725 | ||
f12944ec UD |
726 | {AJ} You probably should read the manual section describing nsswitch.conf |
727 | (just type `info libc "NSS Configuration File"'). The NSS configuration | |
728 | file is usually the culprit. | |
61952351 | 729 | |
3dcf8ea6 UD |
730 | |
731 | ?? How do I create the databases for NSS? | |
732 | ||
733 | {AJ} If you have an entry "db" in /etc/nsswitch.conf you should also create | |
734 | the database files. The glibc sources contain a Makefile which does the | |
7fd18ea2 | 735 | necessary conversion and calls to create those files. The file is |
3dcf8ea6 UD |
736 | `db-Makefile' in the subdirectory `nss' and you can call it with `make -f |
737 | db-Makefile'. Please note that not all services are capable of using a | |
738 | database. Currently passwd, group, ethers, protocol, rpc, services shadow | |
14a6b4e4 | 739 | and netgroup are implemented. See also question ?nssdb. |
3dcf8ea6 | 740 | |
61952351 UD |
741 | ?? I have /usr/include/net and /usr/include/scsi as symlinks |
742 | into my Linux source tree. Is that wrong? | |
743 | ||
f12944ec UD |
744 | {PB} This was necessary for libc5, but is not correct when using glibc. |
745 | Including the kernel header files directly in user programs usually does not | |
746 | work (see ?kerhdr). glibc provides its own <net/*> and <scsi/*> header | |
747 | files to replace them, and you may have to remove any symlink that you have | |
748 | in place before you install glibc. However, /usr/include/asm and | |
749 | /usr/include/linux should remain as they were. | |
61952351 UD |
750 | |
751 | ?? Programs like `logname', `top', `uptime' `users', `w' and | |
752 | `who', show incorrect information about the (number of) | |
753 | users on my system. Why? | |
754 | ||
755 | {MK} See ?getlog. | |
756 | ||
757 | ?? After upgrading to glibc 2.1 with symbol versioning I get | |
758 | errors about undefined symbols. What went wrong? | |
759 | ||
f12944ec UD |
760 | {AJ} The problem is caused either by wrong program code or tools. In the |
761 | versioned libc a lot of symbols are now local that were global symbols in | |
762 | previous versions. It seems that programs linked against older versions | |
763 | often accidentally used libc global variables -- something that should not | |
764 | happen. | |
61952351 | 765 | |
f12944ec UD |
766 | The only way to fix this is to recompile your program. Sorry, that's the |
767 | price you might have to pay once for quite a number of advantages with | |
768 | symbol versioning. | |
61952351 UD |
769 | |
770 | ?? When I start the program XXX after upgrading the library | |
771 | I get | |
772 | XXX: Symbol `_sys_errlist' has different size in shared | |
773 | object, consider re-linking | |
774 | Why? What should I do? | |
775 | ||
f12944ec UD |
776 | {UD} As the message says, relink the binary. The problem is that a few |
777 | symbols from the library can change in size and there is no way to avoid | |
778 | this. _sys_errlist is a good example. Occasionally there are new error | |
779 | numbers added to the kernel and this must be reflected at user level, | |
780 | breaking programs that refer to them directly. | |
61952351 | 781 | |
f12944ec UD |
782 | Such symbols should normally not be used at all. There are mechanisms to |
783 | avoid using them. In the case of _sys_errlist, there is the strerror() | |
784 | function which should _always_ be used instead. So the correct fix is to | |
785 | rewrite that part of the application. | |
61952351 | 786 | |
f12944ec UD |
787 | In some situations (especially when testing a new library release) it might |
788 | be possible that a symbol changed size when that should not have happened. | |
789 | So in case of doubt report such a warning message as a problem. | |
61952351 | 790 | |
da2d1bc5 UD |
791 | ?? What do I need for C++ development? |
792 | ||
d89e7a96 UD |
793 | {HJ,AJ} You need either egcs 1.1 which comes directly with libstdc++ or |
794 | gcc-2.8.1 together with libstdc++ 2.8.1.1. egcs 1.1 has the better C++ | |
795 | support and works directly with glibc 2.1. If you use gcc-2.8.1 with | |
796 | libstdc++ 2.8.1.1, you need to modify libstdc++ a bit. A patch is available | |
797 | as: | |
66f6a52b | 798 | <ftp://alpha.gnu.org/gnu/libstdc++-2.8.1.1-glibc2.1-diff.gz> |
d89e7a96 UD |
799 | |
800 | Please note that libg++ 2.7.2 (and the Linux Versions 2.7.2.x) doesn't work | |
801 | very well with the GNU C library due to vtable thunks. If you're upgrading | |
802 | from glibc 2.0.x to 2.1 you have to recompile libstdc++ since the library | |
803 | compiled for 2.0 is not compatible due to the new Large File Support (LFS) | |
804 | in version 2.1. | |
fb98e2bf UD |
805 | |
806 | {UD} But since in the case of a shared libstdc++ the version numbers should | |
807 | be different existing programs will continue to work. | |
da2d1bc5 | 808 | |
6ca96fe2 UD |
809 | ?? Even statically linked programs need some shared libraries |
810 | which is not acceptable for me. What can I do? | |
811 | ||
f12944ec UD |
812 | {AJ} NSS (for details just type `info libc "Name Service Switch"') won't |
813 | work properly without shared libraries. NSS allows using different services | |
814 | (e.g. NIS, files, db, hesiod) by just changing one configuration file | |
815 | (/etc/nsswitch.conf) without relinking any programs. The only disadvantage | |
816 | is that now static libraries need to access shared libraries. This is | |
817 | handled transparently by the GNU C library. | |
6ca96fe2 | 818 | |
f12944ec UD |
819 | A solution is to configure glibc with --enable-static-nss. In this case you |
820 | can create a static binary that will use only the services dns and files | |
821 | (change /etc/nsswitch.conf for this). You need to link explicitly against | |
822 | all these services. For example: | |
6ca96fe2 UD |
823 | |
824 | gcc -static test-netdb.c -o test-netdb.c \ | |
825 | -lc -lnss_files -lnss_dns -lresolv | |
826 | ||
827 | The problem with this approach is that you've got to link every static | |
828 | program that uses NSS routines with all those libraries. | |
829 | ||
830 | {UD} In fact, one cannot say anymore that a libc compiled with this | |
831 | option is using NSS. There is no switch anymore. Therefore it is | |
832 | *highly* recommended *not* to use --enable-static-nss since this makes | |
833 | the behaviour of the programs on the system inconsistent. | |
834 | ||
bf47fa23 UD |
835 | ?? I just upgraded my Linux system to glibc and now I get |
836 | errors whenever I try to link any program. | |
837 | ||
838 | {ZW} This happens when you have installed glibc as the primary C library but | |
839 | have stray symbolic links pointing at your old C library. If the first | |
840 | `libc.so' the linker finds is libc 5, it will use that. Your program | |
841 | expects to be linked with glibc, so the link fails. | |
842 | ||
843 | The most common case is that glibc put its `libc.so' in /usr/lib, but there | |
844 | was a `libc.so' from libc 5 in /lib, which gets searched first. To fix the | |
845 | problem, just delete /lib/libc.so. You may also need to delete other | |
846 | symbolic links in /lib, such as /lib/libm.so if it points to libm.so.5. | |
847 | ||
0e0316f4 UD |
848 | {AJ} The perl script test-installation.pl which is run as last step during |
849 | an installation of glibc that is configured with --prefix=/usr should help | |
850 | detect these situations. If the script reports problems, something is | |
851 | really screwed up. | |
852 | ||
48244d09 UD |
853 | ?? When I use nscd the machine freezes. |
854 | ||
d89e7a96 UD |
855 | {UD} You cannot use nscd with Linux 2.0.*. There is functionality missing |
856 | in the kernel and work-arounds are not suitable. Besides, some parts of the | |
857 | kernel are too buggy when it comes to using threads. | |
48244d09 | 858 | |
440d13e2 | 859 | If you need nscd, you have to use at least a 2.1 kernel. |
48244d09 UD |
860 | |
861 | Note that I have at this point no information about any other platform. | |
862 | ||
0155a773 UD |
863 | ?? I need lots of open files. What do I have to do? |
864 | ||
865 | {AJ} This is at first a kernel issue. The kernel defines limits with | |
866 | OPEN_MAX the number of simultaneous open files and with FD_SETSIZE the | |
867 | number of used file descriptors. You need to change these values in your | |
e8b1163e | 868 | kernel and recompile the kernel so that the kernel allows more open |
0155a773 UD |
869 | files. You don't necessarily need to recompile the GNU C library since the |
870 | only place where OPEN_MAX and FD_SETSIZE is really needed in the library | |
871 | itself is the size of fd_set which is used by select. | |
872 | ||
7b19af68 UD |
873 | The GNU C library is now select free. This means it internally has no |
874 | limits imposed by the `fd_set' type. Instead all places where the | |
0155a773 UD |
875 | functionality is needed the `poll' function is used. |
876 | ||
877 | If you increase the number of file descriptors in the kernel you don't need | |
6e8afc1c | 878 | to recompile the C library. |
0155a773 UD |
879 | |
880 | {UD} You can always get the maximum number of file descriptors a process is | |
881 | allowed to have open at any time using | |
882 | ||
883 | number = sysconf (_SC_OPEN_MAX); | |
884 | ||
885 | This will work even if the kernel limits change. | |
886 | ||
d8a167a5 UD |
887 | ?? How do I get the same behavior on parsing /etc/passwd and |
888 | /etc/group as I have with libc5 ? | |
889 | ||
890 | {TK} The name switch setup in /etc/nsswitch.conf selected by most Linux | |
891 | distributions does not support +/- and netgroup entries in the files like | |
892 | /etc/passwd. Though this is the preferred setup some people might have | |
893 | setups coming over from the libc5 days where it was the default to recognize | |
894 | lines like this. To get back to the old behaviour one simply has to change | |
895 | the rules for passwd, group, and shadow in the nsswitch.conf file as | |
896 | follows: | |
897 | ||
898 | passwd: compat | |
899 | group: compat | |
900 | shadow: compat | |
901 | ||
902 | passwd_compat: nis | |
903 | group_compat: nis | |
904 | shadow_compat: nis | |
905 | ||
4f7ea427 | 906 | ??libs What needs to be recompiled when upgrading from glibc 2.0 to glibc |
0f6052a8 UD |
907 | 2.1? |
908 | ||
909 | {AJ,CG} If you just upgrade the glibc from 2.0.x (x <= 7) to 2.1, binaries | |
910 | that have been linked against glibc 2.0 will continue to work. | |
911 | ||
912 | If you compile your own binaries against glibc 2.1, you also need to | |
70cafe50 UD |
913 | recompile some other libraries. The problem is that libio had to be changed |
914 | and therefore libraries that are based or depend on the libio of glibc, | |
915 | e.g. ncurses, slang and most C++ libraries, need to be recompiled. If you | |
916 | experience strange segmentation faults in your programs linked against glibc | |
917 | 2.1, you might need to recompile your libraries. | |
0f6052a8 UD |
918 | |
919 | Another problem is that older binaries that were linked statically against | |
920 | glibc 2.0 will reference the older nss modules (libnss_files.so.1 instead of | |
921 | libnss_files.so.2), so don't remove them. Also, the old glibc-2.0 compiled | |
922 | static libraries (libfoo.a) which happen to depend on the older libio | |
50b65db1 UD |
923 | behavior will be broken by the glibc 2.1 upgrade. We plan to produce a |
924 | compatibility library that people will be able to link in if they want | |
925 | to compile a static library generated against glibc 2.0 into a program | |
926 | on a glibc 2.1 system. You just add -lcompat and you should be fine. | |
927 | ||
928 | The glibc-compat add-on will provide the libcompat.a library, the older | |
929 | nss modules, and a few other files. Together, they should make it | |
930 | possible to do development with old static libraries on a glibc 2.1 | |
8d8c6efa | 931 | system. This add-on is still in development. You can get it from |
df08cc56 | 932 | <ftp://alpha.gnu.org/gnu/glibc/glibc-compat-2.1.tar.gz> |
50b65db1 | 933 | but please keep in mind that it is experimental. |
0155a773 | 934 | |
b7398be5 UD |
935 | ?? Why is extracting files via tar so slow? |
936 | ||
937 | {AJ} Extracting of tar archives might be quite slow since tar has to look up | |
938 | userid and groupids and doesn't cache negative results. If you have nis or | |
939 | nisplus in your /etc/nsswitch.conf for the passwd and/or group database, | |
940 | each file extractions needs a network connection. There are two possible | |
941 | solutions: | |
942 | ||
943 | - do you really need NIS/NIS+ (some Linux distributions add by default | |
944 | nis/nisplus even if it's not needed)? If not, just remove the entries. | |
945 | ||
946 | - if you need NIS/NIS+, use the Name Service Cache Daemon nscd that comes | |
947 | with glibc 2.1. | |
948 | ||
2ee511d9 UD |
949 | ?? Compiling programs I get parse errors in libio.h (e.g. "parse error |
950 | before `_IO_seekoff'"). How should I fix this? | |
951 | ||
952 | {AJ} You might get the following errors when upgrading to glibc 2.1: | |
953 | ||
954 | In file included from /usr/include/stdio.h:57, | |
955 | from ... | |
956 | /usr/include/libio.h:335: parse error before `_IO_seekoff' | |
957 | /usr/include/libio.h:335: parse error before `_G_off64_t' | |
958 | /usr/include/libio.h:336: parse error before `_IO_seekpos' | |
959 | /usr/include/libio.h:336: parse error before `_G_fpos64_t' | |
960 | ||
961 | The problem is a wrong _G_config.h file in your include path. The | |
962 | _G_config.h file that comes with glibc 2.1 should be used and not one from | |
963 | libc5 or from a compiler directory. To check which _G_config.h file the | |
964 | compiler uses, compile your program with `gcc -E ...|grep G_config.h' and | |
965 | remove that file. Your compiler should pick up the file that has been | |
966 | installed by glibc 2.1 in your include directory. | |
967 | ||
4f7ea427 UD |
968 | ?? After upgrading to glibc 2.1, libraries that were compiled against |
969 | glibc 2.0.x don't work anymore. | |
970 | ||
971 | {AJ} See ?libs. | |
972 | ||
14a6b4e4 UD |
973 | ??nssdb What happened to the Berkeley DB libraries? Can I still use db |
974 | in /etc/nsswitch.conf? | |
975 | ||
976 | {AJ} Due to too many incompatible changes in disk layout and API of Berkeley | |
977 | DB and a too tight coupling of libc and libdb, the db library has been | |
978 | removed completely from glibc 2.2. The only place that really used the | |
979 | Berkeley DB was the NSS db module. | |
980 | ||
981 | The NSS db module has been rewritten to support a number of different | |
982 | versions of Berkeley DB for the NSS db module. Currently the releases 2.x | |
983 | and 3.x of Berkeley DB are supported. The older db 1.85 library is not | |
984 | supported. You can use the version from glibc 2.1.x or download a version | |
985 | from Sleepycat Software (http://www.sleepycat.com). The library has to be | |
986 | compiled as shared library and installed in the system lib directory | |
987 | (normally /lib). The library needs to have a special soname to be found by | |
988 | the NSS module. | |
989 | ||
990 | If public structures change in a new Berkeley db release, this needs to be | |
991 | reflected in glibc. | |
992 | ||
993 | Currently the code searches for libraries with a soname of "libdb.so.3" | |
994 | (that's the name from db 2.4.14 which comes with glibc 2.1.x) and | |
995 | "libdb-3.0.so" (the name used by db 3.0.55 as default). | |
996 | ||
2ee511d9 | 997 | |
61952351 UD |
998 | ? Source and binary incompatibilities, and what to do about them |
999 | ||
1000 | ?? I expect GNU libc to be 100% source code compatible with | |
1001 | the old Linux based GNU libc. Why isn't it like this? | |
1002 | ||
f12944ec UD |
1003 | {DMT,UD} Not every extension in Linux libc's history was well thought-out. |
1004 | In fact it had a lot of problems with standards compliance and with | |
1005 | cleanliness. With the introduction of a new version number these errors can | |
1006 | now be corrected. Here is a list of the known source code | |
61952351 UD |
1007 | incompatibilities: |
1008 | ||
1009 | * _GNU_SOURCE: glibc does not make the GNU extensions available | |
1010 | automatically. If a program depends on GNU extensions or some | |
1011 | other non-standard functionality, it is necessary to compile it | |
1012 | with the C compiler option -D_GNU_SOURCE, or better, to put | |
1013 | `#define _GNU_SOURCE' at the beginning of your source files, before | |
1014 | any C library header files are included. This difference normally | |
1015 | manifests itself in the form of missing prototypes and/or data type | |
1016 | definitions. Thus, if you get such errors, the first thing you | |
1017 | should do is try defining _GNU_SOURCE and see if that makes the | |
1018 | problem go away. | |
1019 | ||
1020 | For more information consult the file `NOTES' in the GNU C library | |
1021 | sources. | |
1022 | ||
1023 | * reboot(): GNU libc sanitizes the interface of reboot() to be more | |
1024 | compatible with the interface used on other OSes. reboot() as | |
1025 | implemented in glibc takes just one argument. This argument | |
1026 | corresponds to the third argument of the Linux reboot system call. | |
1027 | That is, a call of the form reboot(a, b, c) needs to be changed into | |
1028 | reboot(c). Beside this the header <sys/reboot.h> defines the needed | |
1029 | constants for the argument. These RB_* constants should be used | |
1030 | instead of the cryptic magic numbers. | |
1031 | ||
1032 | * swapon(): the interface of this function didn't change, but the | |
1033 | prototype is in a separate header file <sys/swap.h>. This header | |
1034 | file also provides the SWAP_* constants defined by <linux/swap.h>; | |
1035 | you should use them for the second argument to swapon(). | |
1036 | ||
1037 | * errno: If a program uses the variable "errno", then it _must_ | |
1038 | include <errno.h>. The old libc often (erroneously) declared this | |
1039 | variable implicitly as a side-effect of including other libc header | |
1040 | files. glibc is careful to avoid such namespace pollution, which, | |
1041 | in turn, means that you really need to include the header files that | |
1042 | you depend on. This difference normally manifests itself in the | |
1043 | form of the compiler complaining about references to an undeclared | |
1044 | symbol "errno". | |
1045 | ||
1046 | * Linux-specific syscalls: All Linux system calls now have appropriate | |
1047 | library wrappers and corresponding declarations in various header files. | |
1048 | This is because the syscall() macro that was traditionally used to | |
1049 | work around missing syscall wrappers are inherently non-portable and | |
1050 | error-prone. The following table lists all the new syscall stubs, | |
1051 | the header-file declaring their interface and the system call name. | |
1052 | ||
1053 | syscall name: wrapper name: declaring header file: | |
1054 | ------------- ------------- ---------------------- | |
1055 | bdflush bdflush <sys/kdaemon.h> | |
1056 | syslog ksyslog_ctl <sys/klog.h> | |
1057 | ||
1058 | * lpd: Older versions of lpd depend on a routine called _validuser(). | |
1059 | The library does not provide this function, but instead provides | |
1060 | __ivaliduser() which has a slightly different interface. Simply | |
1061 | upgrading to a newer lpd should fix this problem (e.g., the 4.4BSD | |
1062 | lpd is known to be working). | |
1063 | ||
1064 | * resolver functions/BIND: like on many other systems the functions of | |
1065 | the resolver library are not included in libc itself. There is a | |
1066 | separate library libresolv. If you get undefined symbol errors for | |
1067 | symbols starting with `res_*' simply add -lresolv to your linker | |
1068 | command line. | |
1069 | ||
1070 | * the `signal' function's behavior corresponds to the BSD semantic and | |
1071 | not the SysV semantic as it was in libc-5. The interface on all GNU | |
1072 | systems shall be the same and BSD is the semantic of choice. To use | |
1073 | the SysV behavior simply use `sysv_signal', or define _XOPEN_SOURCE. | |
1074 | See ?signal for details. | |
1075 | ||
1076 | ??getlog Why does getlogin() always return NULL on my Linux box? | |
1077 | ||
f12944ec UD |
1078 | {UD} The GNU C library has a format for the UTMP and WTMP file which differs |
1079 | from what your system currently has. It was extended to fulfill the needs | |
1080 | of the next years when IPv6 is introduced. The record size is different and | |
1081 | some fields have different positions. The files written by functions from | |
1082 | the one library cannot be read by functions from the other library. Sorry, | |
1083 | but this is what a major release is for. It's better to have a cut now than | |
1084 | having no means to support the new techniques later. | |
61952351 | 1085 | |
f12944ec UD |
1086 | {MK} There is however a (partial) solution for this problem. Please take a |
1087 | look at the file `login/README.utmpd'. | |
61952351 UD |
1088 | |
1089 | ?? Where are the DST_* constants found in <sys/time.h> on many | |
1090 | systems? | |
1091 | ||
f12944ec UD |
1092 | {UD} These constants come from the old BSD days and are not used anymore |
1093 | (libc5 does not actually implement the handling although the constants are | |
1094 | defined). | |
61952351 | 1095 | |
f12944ec | 1096 | Instead GNU libc contains zone database support and compatibility code for |
8b4a4715 UD |
1097 | POSIX TZ environment variable handling. For former is very much preferred |
1098 | (see ?tzdb). | |
61952351 UD |
1099 | |
1100 | ?? The prototypes for `connect', `accept', `getsockopt', | |
1101 | `setsockopt', `getsockname', `getpeername', `send', | |
1102 | `sendto', and `recvfrom' are different in GNU libc from | |
1103 | any other system I saw. This is a bug, isn't it? | |
1104 | ||
f12944ec UD |
1105 | {UD} No, this is no bug. This version of GNU libc already follows the new |
1106 | Single Unix specifications (and I think the POSIX.1g draft which adopted the | |
1107 | solution). The type for a parameter describing a size is now `socklen_t', a | |
1108 | new type. | |
61952351 UD |
1109 | |
1110 | ??kerhdr On Linux I've got problems with the declarations in Linux | |
1111 | kernel headers. | |
1112 | ||
f12944ec UD |
1113 | {UD,AJ} On Linux, the use of kernel headers is reduced to the minimum. This |
1114 | gives Linus the ability to change the headers more freely. Also, user | |
a9ddb793 | 1115 | programs are now insulated from changes in the size of kernel data |
f12944ec | 1116 | structures. |
61952351 | 1117 | |
f12944ec UD |
1118 | For example, the sigset_t type is 32 or 64 bits wide in the kernel. In |
1119 | glibc it is 1024 bits wide. This guarantees that when the kernel gets a | |
1120 | bigger sigset_t (for POSIX.1e realtime support, say) user programs will not | |
1121 | have to be recompiled. Consult the header files for more information about | |
1122 | the changes. | |
61952351 | 1123 | |
f12944ec UD |
1124 | Therefore you shouldn't include Linux kernel header files directly if glibc |
1125 | has defined a replacement. Otherwise you might get undefined results because | |
1126 | of type conflicts. | |
61952351 UD |
1127 | |
1128 | ?? I don't include any kernel headers myself but the compiler | |
1129 | still complains about redeclarations of types in the kernel | |
1130 | headers. | |
1131 | ||
f12944ec UD |
1132 | {UD} The kernel headers before Linux 2.1.61 and 2.0.32 don't work correctly |
1133 | with glibc. Compiling C programs is possible in most cases but C++ programs | |
1134 | have (due to the change of the name lookups for `struct's) problems. One | |
1135 | prominent example is `struct fd_set'. | |
61952351 | 1136 | |
f12944ec UD |
1137 | There might be some problems left but 2.1.61/2.0.32 fix most of the known |
1138 | ones. See the BUGS file for other known problems. | |
61952351 UD |
1139 | |
1140 | ??signal Why don't signals interrupt system calls anymore? | |
1141 | ||
f12944ec UD |
1142 | {ZW} By default GNU libc uses the BSD semantics for signal(), unlike Linux |
1143 | libc 5 which used System V semantics. This is partially for compatibility | |
1144 | with other systems and partially because the BSD semantics tend to make | |
1145 | programming with signals easier. | |
61952351 UD |
1146 | |
1147 | There are three differences: | |
1148 | ||
1149 | * BSD-style signals that occur in the middle of a system call do not | |
1150 | affect the system call; System V signals cause the system call to | |
1151 | fail and set errno to EINTR. | |
1152 | ||
1153 | * BSD signal handlers remain installed once triggered. System V signal | |
1154 | handlers work only once, so one must reinstall them each time. | |
1155 | ||
1156 | * A BSD signal is blocked during the execution of its handler. In other | |
1157 | words, a handler for SIGCHLD (for example) does not need to worry about | |
1158 | being interrupted by another SIGCHLD. It may, however, be interrupted | |
1159 | by other signals. | |
1160 | ||
1161 | There is general consensus that for `casual' programming with signals, the | |
1162 | BSD semantics are preferable. You don't need to worry about system calls | |
1163 | returning EINTR, and you don't need to worry about the race conditions | |
1164 | associated with one-shot signal handlers. | |
1165 | ||
1166 | If you are porting an old program that relies on the old semantics, you can | |
1167 | quickly fix the problem by changing signal() to sysv_signal() throughout. | |
1168 | Alternatively, define _XOPEN_SOURCE before including <signal.h>. | |
1169 | ||
1170 | For new programs, the sigaction() function allows you to specify precisely | |
1171 | how you want your signals to behave. All three differences listed above are | |
1172 | individually switchable on a per-signal basis with this function. | |
1173 | ||
f12944ec UD |
1174 | If all you want is for one specific signal to cause system calls to fail and |
1175 | return EINTR (for example, to implement a timeout) you can do this with | |
61952351 UD |
1176 | siginterrupt(). |
1177 | ||
1178 | ||
1179 | ??string I've got errors compiling code that uses certain string | |
1180 | functions. Why? | |
1181 | ||
f12944ec | 1182 | {AJ} glibc 2.1 has special string functions that are faster than the normal |
fdacb17d | 1183 | library functions. Some of the functions are additionally implemented as |
a9d75566 UD |
1184 | inline functions and others as macros. This might lead to problems with |
1185 | existing codes but it is explicitly allowed by ISO C. | |
61952351 UD |
1186 | |
1187 | The optimized string functions are only used when compiling with | |
fdacb17d | 1188 | optimizations (-O1 or higher). The behavior can be changed with two feature |
f12944ec | 1189 | macros: |
61952351 UD |
1190 | |
1191 | * __NO_STRING_INLINES: Don't do any string optimizations. | |
1192 | * __USE_STRING_INLINES: Use assembly language inline functions (might | |
1193 | increase code size dramatically). | |
1194 | ||
f12944ec UD |
1195 | Since some of these string functions are now additionally defined as macros, |
1196 | code like "char *strncpy();" doesn't work anymore (and is unnecessary, since | |
fdacb17d | 1197 | <string.h> has the necessary declarations). Either change your code or |
f12944ec | 1198 | define __NO_STRING_INLINES. |
61952351 | 1199 | |
f12944ec UD |
1200 | {UD} Another problem in this area is that gcc still has problems on machines |
1201 | with very few registers (e.g., ix86). The inline assembler code can require | |
1202 | almost all the registers and the register allocator cannot always handle | |
1203 | this situation. | |
61952351 UD |
1204 | |
1205 | One can disable the string optimizations selectively. Instead of writing | |
1206 | ||
1207 | cp = strcpy (foo, "lkj"); | |
1208 | ||
1209 | one can write | |
1210 | ||
1211 | cp = (strcpy) (foo, "lkj"); | |
1212 | ||
1213 | This disables the optimization for that specific call. | |
1214 | ||
4775243a UD |
1215 | ?? I get compiler messages "Initializer element not constant" with |
1216 | stdin/stdout/stderr. Why? | |
1217 | ||
1218 | {RM,AJ} Constructs like: | |
66f6a52b | 1219 | static FILE *InPtr = stdin; |
4775243a | 1220 | |
fdacb17d UD |
1221 | lead to this message. This is correct behaviour with glibc since stdin is |
1222 | not a constant expression. Please note that a strict reading of ISO C does | |
f12944ec | 1223 | not allow above constructs. |
4775243a | 1224 | |
f12944ec UD |
1225 | One of the advantages of this is that you can assign to stdin, stdout, and |
1226 | stderr just like any other global variable (e.g. `stdout = my_stream;'), | |
1227 | which can be very useful with custom streams that you can write with libio | |
fdacb17d | 1228 | (but beware this is not necessarily portable). The reason to implement it |
f12944ec | 1229 | this way were versioning problems with the size of the FILE structure. |
4775243a | 1230 | |
fdacb17d UD |
1231 | To fix those programs you've got to initialize the variable at run time. |
1232 | This can be done, e.g. in main, like: | |
1233 | ||
66f6a52b UD |
1234 | static FILE *InPtr; |
1235 | int main(void) | |
1236 | { | |
1237 | InPtr = stdin; | |
1238 | } | |
fdacb17d UD |
1239 | |
1240 | or by constructors (beware this is gcc specific): | |
1241 | ||
66f6a52b UD |
1242 | static FILE *InPtr; |
1243 | static void inPtr_construct (void) __attribute__((constructor)); | |
1244 | static void inPtr_construct (void) { InPtr = stdin; } | |
fdacb17d | 1245 | |
4775243a UD |
1246 | |
1247 | ?? I can't compile with gcc -traditional (or | |
1248 | -traditional-cpp). Why? | |
1249 | ||
1250 | {AJ} glibc2 does break -traditional and -traditonal-cpp - and will continue | |
fdacb17d | 1251 | to do so. For example constructs of the form: |
f12944ec | 1252 | |
66f6a52b UD |
1253 | enum {foo |
1254 | #define foo foo | |
1255 | } | |
f12944ec UD |
1256 | |
1257 | are useful for debugging purposes (you can use foo with your debugger that's | |
1258 | why we need the enum) and for compatibility (other systems use defines and | |
1259 | check with #ifdef). | |
4775243a UD |
1260 | |
1261 | ?? I get some errors with `gcc -ansi'. Isn't glibc ANSI compatible? | |
1262 | ||
1263 | {AJ} The GNU C library is compatible with the ANSI/ISO C standard. If | |
f12944ec | 1264 | you're using `gcc -ansi', the glibc includes which are specified in the |
fdacb17d | 1265 | standard follow the standard. The ANSI/ISO C standard defines what has to be |
f12944ec UD |
1266 | in the include files - and also states that nothing else should be in the |
1267 | include files (btw. you can still enable additional standards with feature | |
1268 | flags). | |
4775243a | 1269 | |
f12944ec UD |
1270 | The GNU C library is conforming to ANSI/ISO C - if and only if you're only |
1271 | using the headers and library functions defined in the standard. | |
4775243a | 1272 | |
4d42000c UD |
1273 | ?? I can't access some functions anymore. nm shows that they do |
1274 | exist but linking fails nevertheless. | |
1275 | ||
f12944ec UD |
1276 | {AJ} With the introduction of versioning in glibc 2.1 it is possible to |
1277 | export only those identifiers (functions, variables) that are really needed | |
1278 | by application programs and by other parts of glibc. This way a lot of | |
1279 | internal interfaces are now hidden. nm will still show those identifiers | |
1280 | but marking them as internal. ISO C states that identifiers beginning with | |
1281 | an underscore are internal to the libc. An application program normally | |
1282 | shouldn't use those internal interfaces (there are exceptions, | |
1283 | e.g. __ivaliduser). If a program uses these interfaces, it's broken. These | |
1284 | internal interfaces might change between glibc releases or dropped | |
1285 | completely. | |
4d42000c | 1286 | |
9de4e203 UD |
1287 | ?? When using the db-2 library which comes with glibc is used in |
1288 | the Perl db modules the testsuite is not passed. This did not | |
1289 | happen with db-1, gdbm, or ndbm. | |
1290 | ||
1291 | Removed. Does not apply anymore. | |
1292 | ||
5148d49f UD |
1293 | ?? The pow() inline function I get when including <math.h> is broken. |
1294 | I get segmentation faults when I run the program. | |
1295 | ||
1296 | {UD} Nope, the implementation is correct. The problem is with egcs version | |
1297 | prior to 1.1. I.e., egcs 1.0 to 1.0.3 are all broken (at least on Intel). | |
1298 | If you have to use this compiler you must define __NO_MATH_INLINES before | |
1299 | including <math.h> to prevent the inline functions from being used. egcs 1.1 | |
1300 | fixes the problem. I don't know about gcc 2.8 and 2.8.1. | |
1301 | ||
05f732b3 UD |
1302 | ?? The sys/sem.h file lacks the definition of `union semun'. |
1303 | ||
1304 | {UD} Nope. This union has to be provided by the user program. Former glibc | |
1305 | versions defined this but it was an error since it does not make much sense | |
1306 | when thinking about it. The standards describing the System V IPC functions | |
1307 | define it this way and therefore programs must be adopted. | |
1308 | ||
a42134a7 UD |
1309 | ?? Why has <netinet/ip_fw.h> disappeared? |
1310 | ||
1311 | {AJ} The corresponding Linux kernel data structures and constants are | |
440d13e2 | 1312 | totally different in Linux 2.0 and Linux 2.2. This situation has to be |
a42134a7 UD |
1313 | taken care in user programs using the firewall structures and therefore |
1314 | those programs (ipfw is AFAIK the only one) should deal with this problem | |
1315 | themselves. | |
1316 | ||
ee586e0e UD |
1317 | ?? I get floods of warnings when I use -Wconversion and include |
1318 | <string.h> or <math.h>. | |
1319 | ||
1320 | {ZW} <string.h> and <math.h> intentionally use prototypes to override | |
1321 | argument promotion. -Wconversion warns about all these. You can safely | |
1322 | ignore the warnings. | |
1323 | ||
1324 | -Wconversion isn't really intended for production use, only for shakedown | |
1325 | compiles after converting an old program to standard C. | |
1326 | ||
4d42000c | 1327 | |
49b75f5e UD |
1328 | ?? After upgrading to glibc 2.1, I receive errors about |
1329 | unresolved symbols, like `_dl_initial_searchlist' and can not | |
1330 | execute any binaries. What went wrong? | |
1331 | ||
1332 | {AJ} This normally happens if your libc and ld (dynamic linker) are from | |
1333 | different releases of glibc. For example, the dynamic linker | |
1334 | /lib/ld-linux.so.2 comes from glibc 2.0.x, but the version of libc.so.6 is | |
1335 | from glibc 2.1. | |
1336 | ||
1337 | The path /lib/ld-linux.so.2 is hardcoded in every glibc2 binary but | |
1338 | libc.so.6 is searched via /etc/ld.so.cache and in some special directories | |
1339 | like /lib and /usr/lib. If you run configure with another prefix than /usr | |
1340 | and put this prefix before /lib in /etc/ld.so.conf, your system will break. | |
1341 | ||
1342 | So what can you do? Either of the following should work: | |
1343 | ||
1344 | * Run `configure' with the same prefix argument you've used for glibc 2.0.x | |
1345 | so that the same paths are used. | |
1346 | * Replace /lib/ld-linux.so.2 with a link to the dynamic linker from glibc | |
1347 | 2.1. | |
1348 | ||
1349 | You can even call the dynamic linker by hand if everything fails. You've | |
1350 | got to set LD_LIBRARY_PATH so that the corresponding libc is found and also | |
1351 | need to provide an absolute path to your binary: | |
1352 | ||
1353 | LD_LIBRARY_PATH=<path-where-libc.so.6-lives> \ | |
1354 | <path-where-corresponding-dynamic-linker-lives>/ld-linux.so.2 \ | |
1355 | <path-to-binary>/binary | |
1356 | ||
1357 | For example `LD_LIBRARY_PATH=/libold /libold/ld-linux.so.2 /bin/mv ...' | |
1358 | might be useful in fixing a broken system (if /libold contains dynamic | |
1359 | linker and corresponding libc). | |
1360 | ||
1361 | With that command line no path is used. To further debug problems with the | |
1362 | dynamic linker, use the LD_DEBUG environment variable, e.g. | |
1363 | `LD_DEBUG=help echo' for the help text. | |
1364 | ||
1365 | If you just want to test this release, don't put the lib directory in | |
1366 | /etc/ld.so.conf. You can call programs directly with full paths (as above). | |
1367 | When compiling new programs against glibc 2.1, you've got to specify the | |
1368 | correct paths to the compiler (option -I with gcc) and linker (options | |
1369 | --dynamic-linker, -L and --rpath). | |
1370 | ||
b74656f9 | 1371 | ?? bonnie reports that char i/o with glibc 2 is much slower than with |
9f6b6d8d UD |
1372 | libc5. What can be done? |
1373 | ||
1374 | {AJ} The GNU C library uses thread safe functions by default and libc5 used | |
1375 | non thread safe versions. The non thread safe functions have in glibc the | |
1376 | suffix `_unlocked', for details check <stdio.h>. Using `putc_unlocked' etc. | |
1377 | instead of `putc' should give nearly the same speed with bonnie (bonnie is a | |
1378 | benchmark program for measuring disk access). | |
1379 | ||
9de4e203 UD |
1380 | ?? Programs compiled with glibc 2.1 can't read db files made with glibc |
1381 | 2.0. What has changed that programs like rpm break? | |
1382 | ||
1383 | Removed. Does not apply anymore. | |
1384 | ||
8a40ed68 UD |
1385 | ?? Autoconf's AC_CHECK_FUNC macro reports that a function exists, but |
1386 | when I try to use it, it always returns -1 and sets errno to ENOSYS. | |
1387 | ||
1388 | {ZW} You are using a 2.0 Linux kernel, and the function you are trying to | |
1389 | use is only implemented in 2.1/2.2. Libc considers this to be a function | |
1390 | which exists, because if you upgrade to a 2.2 kernel, it will work. One | |
1391 | such function is sigaltstack. | |
1392 | ||
1393 | Your program should check at runtime whether the function works, and | |
1394 | implement a fallback. Note that Autoconf cannot detect unimplemented | |
1395 | functions in other systems' C libraries, so you need to do this anyway. | |
1396 | ||
b5a9efcd UD |
1397 | ?? My program segfaults when I call fclose() on the FILE* returned |
1398 | from setmntent(). Is this a glibc bug? | |
1399 | ||
1400 | {GK} No. Don't do this. Use endmntent(), that's what it's for. | |
1401 | ||
1402 | In general, you should use the correct deallocation routine. For instance, | |
1403 | if you open a file using fopen(), you should deallocate the FILE * using | |
1404 | fclose(), not free(), even though the FILE * is also a pointer. | |
1405 | ||
1406 | In the case of setmntent(), it may appear to work in most cases, but it | |
1407 | won't always work. Unfortunately, for compatibility reasons, we can't | |
1408 | change the return type of setmntent() to something other than FILE *. | |
1409 | ||
49b75f5e | 1410 | |
61952351 UD |
1411 | ? Miscellaneous |
1412 | ||
1413 | ?? After I changed configure.in I get `Autoconf version X.Y. | |
1414 | or higher is required for this script'. What can I do? | |
1415 | ||
1416 | {UD} You have to get the specified autoconf version (or a later one) | |
2eb45444 | 1417 | from your favorite mirror of ftp.gnu.org. |
61952351 UD |
1418 | |
1419 | ?? When I try to compile code which uses IPv6 headers and | |
1420 | definitions on my Linux 2.x.y system I am in trouble. | |
1421 | Nothing seems to work. | |
1422 | ||
f12944ec UD |
1423 | {UD} The problem is that IPv6 development still has not reached a point |
1424 | where the headers are stable. There are still lots of incompatible changes | |
1425 | made and the libc headers have to follow. | |
61952351 | 1426 | |
cb0509a8 UD |
1427 | {PB} The 2.1 release of GNU libc aims to comply with the current versions of |
1428 | all the relevant standards. The IPv6 support libraries for older Linux | |
1429 | systems used a different naming convention and so code written to work with | |
1430 | them may need to be modified. If the standards make incompatible changes in | |
1431 | the future then the libc may need to change again. | |
1432 | ||
1433 | IPv6 will not work with a 2.0.x kernel. When kernel 2.2 is released it | |
1434 | should contain all the necessary support; until then you should use the | |
3f7b3d9b | 1435 | latest 2.1.x release you can find. As of 98/11/26 the currently recommended |
cb0509a8 UD |
1436 | kernel for IPv6 is 2.1.129. |
1437 | ||
1438 | Also, as of the 2.1 release the IPv6 API provided by GNU libc is not | |
1439 | 100% complete. In particular the getipnodebyname and getipnodebyaddr | |
1440 | functions are not implemented. | |
61952351 | 1441 | |
8b4a4715 | 1442 | ??tzdb When I set the timezone by setting the TZ environment variable |
73237de3 UD |
1443 | to EST5EDT things go wrong since glibc computes the wrong time |
1444 | from this information. | |
1445 | ||
f12944ec UD |
1446 | {UD} The problem is that people still use the braindamaged POSIX method to |
1447 | select the timezone using the TZ environment variable with a format EST5EDT | |
8b4a4715 UD |
1448 | or whatever. People, if you insist on using TZ instead of the timezone |
1449 | database (see below), read the POSIX standard, the implemented behaviour is | |
f12944ec UD |
1450 | correct! What you see is in fact the result of the decisions made while |
1451 | POSIX.1 was created. We've only implemented the handling of TZ this way to | |
1452 | be POSIX compliant. It is not really meant to be used. | |
1453 | ||
1454 | The alternative approach to handle timezones which is implemented is the | |
1455 | correct one to use: use the timezone database. This avoids all the problems | |
1456 | the POSIX method has plus it is much easier to use. Simply run the tzselect | |
1457 | shell script, answer the question and use the name printed in the end by | |
8b4a4715 UD |
1458 | making a symlink /etc/localtime pointing to /usr/share/zoneinfo/NAME (NAME |
1459 | is the returned value from tzselect). That's all. You never again have to | |
1460 | worry. | |
f12944ec UD |
1461 | |
1462 | So, please avoid sending bug reports about time related problems if you use | |
1463 | the POSIX method and you have not verified something is really broken by | |
1464 | reading the POSIX standards. | |
73237de3 | 1465 | |
fdacb17d UD |
1466 | ?? What other sources of documentation about glibc are available? |
1467 | ||
1468 | {AJ} The FSF has a page about the GNU C library at | |
1469 | <http://www.gnu.org/software/libc/>. The problem data base of open and | |
1470 | solved bugs in GNU libc is available at | |
1471 | <http://www-gnats.gnu.org:8080/cgi-bin/wwwgnats.pl>. Eric Green has written | |
14a6b4e4 | 1472 | a HowTo for converting from Linux libc5 to glibc2. The HowTo is accessible |
fdacb17d UD |
1473 | via the FSF page and at <http://www.imaxx.net/~thrytis/glibc>. Frodo |
1474 | Looijaard describes a different way installing glibc2 as secondary libc at | |
1475 | <http://huizen.dds.nl/~frodol/glibc>. | |
1476 | ||
1477 | Please note that this is not a complete list. | |
1478 | ||
3f7b3d9b UD |
1479 | ?? The timezone string for Sydney/Australia is wrong since even when |
1480 | daylight saving time is in effect the timezone string is EST. | |
1481 | ||
1482 | {UD} The problem for some timezones is that the local authorities decided | |
1483 | to use the term "summer time" instead of "daylight saving time". In this | |
1484 | case the abbreviation character `S' is the same as the standard one. So, | |
1485 | for Sydney we have | |
1486 | ||
1487 | Eastern Standard Time = EST | |
1488 | Eastern Summer Time = EST | |
1489 | ||
1490 | Great! To get this bug fixed convince the authorities to change the laws | |
1491 | and regulations of the country this effects. glibc behaves correctly. | |
1492 | ||
eeabe877 UD |
1493 | ??make I've build make 3.77 against glibc 2.1 and now make gets |
1494 | segmentation faults. | |
1495 | ||
1496 | {AJ} GNU make 3.77 has support for 64 bit filesystems which is slightly | |
1497 | broken (and one of the new features in the GNU C library 2.1 is 64 bit | |
1498 | filesystem support :-( ). To get a working make you can use either make | |
1499 | 3.75 or patch 3.77. A working patch is available via RedHat's Rawhide server | |
1500 | (ftp://rawhide.redhat.com/SRPMS/SRPMS/make-3.77-*src.rpm). | |
1501 | ||
c63598bf UD |
1502 | ?? Why do so many programs using math functions fail on my AlphaStation? |
1503 | ||
1504 | {AO} The functions floor() and floorf() use an instruction that is not | |
1505 | implemented in some old PALcodes of AlphaStations. This may cause | |
1506 | `Illegal Instruction' core dumps or endless loops in programs that | |
1507 | catch these signals. Updating the firmware to a 1999 release has | |
1508 | fixed the problem on an AlphaStation 200 4/166. | |
1509 | ||
8892c471 UD |
1510 | ?? The conversion table for character set XX does not match with |
1511 | what I expect. | |
1512 | ||
1513 | {UD} I don't doubt for a minute that some of the conversion tables contain | |
1514 | errors. We tried the best we can and relied on automatic generation of the | |
1515 | data to prevent human-introduced errors but this still is no guarantee. If | |
1516 | you think you found a problem please send a bug report describing it and | |
1517 | give an authoritive reference. The latter is important since otherwise | |
1518 | the current behaviour is as good as the proposed one. | |
1519 | ||
1520 | Before doing this look through the list of known problem first: | |
1521 | ||
1522 | - the GBK (simplified Chinese) encoding is based on Unicode tables. This | |
1523 | is good. These tables, however, differ slightly from the tables used | |
1524 | by the M$ people. The differences are these [+ Unicode, - M$]: | |
1525 | ||
1526 | +0xA1AA 0x2015 | |
1527 | +0xA844 0x2014 | |
1528 | -0xA1AA 0x2014 | |
1529 | -0xA844 0x2015 | |
1530 | ||
1531 | In addition the Unicode tables contain mappings for the GBK characters | |
1532 | 0xA8BC, 0xA8BF, 0xA989 to 0xA995, and 0xFE50 to 0xFEA0. | |
1533 | ||
ffa156af UD |
1534 | - when mapping from EUC-CN to GBK and vice versa we ignore the fact that |
1535 | the coded character at position 0xA1A4 maps to different Unicode | |
1536 | characters. Since the iconv() implementation can do whatever it wants | |
1537 | if it cannot directly map a character this is a perfectly good solution | |
1538 | since the semantics and appearance of the character does not change. | |
8892c471 | 1539 | |
61952351 UD |
1540 | \f |
1541 | Answers were given by: | |
1542 | {UD} Ulrich Drepper, <drepper@cygnus.com> | |
1543 | {DMT} David Mosberger-Tang, <davidm@AZStarNet.com> | |
1544 | {RM} Roland McGrath, <roland@gnu.org> | |
14a6b4e4 | 1545 | {AJ} Andreas Jaeger, <aj@suse.de> |
61952351 UD |
1546 | {EY} Eric Youngdale, <eric@andante.jic.com> |
1547 | {PB} Phil Blundell, <Philip.Blundell@pobox.com> | |
1548 | {MK} Mark Kettenis, <kettenis@phys.uva.nl> | |
1549 | {ZW} Zack Weinberg, <zack@rabi.phys.columbia.edu> | |
50f301a8 | 1550 | {TK} Thorsten Kukuk, <kukuk@suse.de> |
8619129f | 1551 | {GK} Geoffrey Keating, <geoffk@ozemail.com.au> |
da2d1bc5 | 1552 | {HJ} H.J. Lu, <hjl@gnu.org> |
0f6052a8 | 1553 | {CG} Cristian Gafton, <gafton@redhat.com> |
c63598bf | 1554 | {AO} Alexandre Oliva, <oliva@lsd.ic.unicamp.br> |
61952351 UD |
1555 | \f |
1556 | Local Variables: | |
1557 | mode:outline | |
1558 | outline-regexp:"\\?" | |
f12944ec | 1559 | fill-column:76 |
61952351 | 1560 | End: |