]>
Commit | Line | Data |
---|---|---|
1 | GNU C Library NEWS -- history of user-visible changes. | |
2 | Copyright (C) 1992-2024 Free Software Foundation, Inc. | |
3 | See the end for copying conditions. | |
4 | ||
5 | Please send GNU C library bug reports via <https://sourceware.org/bugzilla/> | |
6 | using `glibc' in the "product" field. | |
7 | \f | |
8 | Version 2.41 | |
9 | ||
10 | Major new features: | |
11 | ||
12 | * In /etc/resolv.conf and the RES_OPTIONS, option flags can now be | |
13 | prefixed with “-” to clear previously set flags. For example, if | |
14 | /etc/resolv.conf contains “options no-aaaa”, a process running with | |
15 | the RES_OPTIONS=-no-aaaa environment variable performs AAAA DNS | |
16 | queries when the glibc DNS stub resolver is used. | |
17 | ||
18 | * The DNS stub resolver now supports the strict-error option. If | |
19 | activated, getaddrinfo for the AF_UNSPEC address family (with dual | |
20 | A/AAAA DNS lookups) attemps to obtain an A/AAAA response pair from | |
21 | another DNS server if one of the responses indicates failure. Without | |
22 | the strict-error option, getaddrinfo returns the A record data it has | |
23 | obtained even if the AAAA query failed. The new strict error mode is | |
24 | incompatible with some DNS environments which do not follow the RFCs, | |
25 | which is why this mode is not enabled by default. A future version | |
26 | of the library may turn it on by default, however. | |
27 | ||
28 | * On Linux, the sched_setattr and sched_getattr have been added, for | |
29 | supporting parameterized scheduling policies such as SCHED_DEADLINE. | |
30 | ||
31 | * The iconv program now supports converting files in place. The program | |
32 | automatically uses a temporary file if required. | |
33 | ||
34 | Deprecated and removed features, and other changes affecting compatibility: | |
35 | ||
36 | * The big-endian ARC port (arceb-linux-gnu) has been removed. | |
37 | ||
38 | Changes to build and runtime requirements: | |
39 | ||
40 | [Add changes to build and runtime requirements here] | |
41 | ||
42 | Security related changes: | |
43 | ||
44 | The following CVEs were fixed in this release, details of which can be | |
45 | found in the advisories directory of the release tarball: | |
46 | ||
47 | [The release manager will add the list generated by | |
48 | scripts/process-advisories.sh just before the release.] | |
49 | ||
50 | The following bugs were resolved with this release: | |
51 | ||
52 | [The release manager will add the list generated by | |
53 | scripts/list-fixed-bugs.py just before the release.] | |
54 | \f | |
55 | Version 2.40 | |
56 | ||
57 | Major new features: | |
58 | ||
59 | * The <stdbit.h> header type-generic macros have been changed when using | |
60 | GCC 14.1 or later to use __builtin_stdc_bit_ceil etc. built-in functions | |
61 | in order to support unsigned __int128 and/or unsigned _BitInt(N) operands | |
62 | with arbitrary precisions when supported by the target. | |
63 | ||
64 | * The GNU C Library now supports a feature test macro _ISOC23_SOURCE to | |
65 | enable features from the ISO C23 standard. Only some features from | |
66 | this standard are supported by the GNU C Library. The older name | |
67 | _ISOC2X_SOURCE is still supported. Features from C23 are also enabled | |
68 | by _GNU_SOURCE, or by compiling with the GCC options -std=c23, | |
69 | -std=gnu23, -std=c2x or -std=gnu2x. | |
70 | ||
71 | * The following ISO C23 function families (introduced in TS | |
72 | 18661-4:2015) are now supported in <math.h>. Each family includes | |
73 | functions for float, double, long double, _FloatN and _FloatNx, and a | |
74 | type-generic macro in <tgmath.h>. | |
75 | ||
76 | - Exponential functions: exp2m1, exp10m1. | |
77 | ||
78 | - Logarithmic functions: log2p1, log10p1, logp1. | |
79 | ||
80 | * A new tunable, glibc.rtld.enable_secure, can be used to run a program | |
81 | as if it were a setuid process. This is currently a testing tool to allow | |
82 | more extensive verification tests for AT_SECURE programs and not meant to | |
83 | be a security feature. | |
84 | ||
85 | * On Linux, the epoll header was updated to include epoll ioctl definitions | |
86 | and the related structure added in Linux kernel 6.9. | |
87 | ||
88 | * The fortify functionality has been significantly enhanced for building | |
89 | programs with clang against the GNU C Library. | |
90 | ||
91 | * Many functions have been added to the vector library for aarch64: | |
92 | acosh, asinh, atanh, cbrt, cosh, erf, erfc, hypot, pow, sinh, tanh | |
93 | ||
94 | * On x86, memset can now use non-temporal stores to improve the performance | |
95 | of large writes. This behaviour is controlled by a new tunable | |
96 | x86_memset_non_temporal_threshold. | |
97 | ||
98 | Deprecated and removed features, and other changes affecting compatibility: | |
99 | ||
100 | * Architectures which use a 32-bit seconds-since-epoch field in struct | |
101 | lastlog, struct utmp, struct utmpx (such as i386, powerpc64le, rv32, | |
102 | rv64, x86-64) switched from a signed to an unsigned type for that | |
103 | field. This allows these fields to store timestamps beyond the year | |
104 | 2038, until the year 2106. Please note that applications are still | |
105 | expected to migrate off the interfaces declared in <utmp.h> and | |
106 | <utmpx.h> (except for login_tty) due to locking and session management | |
107 | problems. | |
108 | ||
109 | * __rseq_size now denotes the size of the active rseq area (20 bytes | |
110 | initially), not the size of struct rseq (32 bytes initially). | |
111 | ||
112 | Security related changes: | |
113 | ||
114 | The following CVEs were fixed in this release, details of which can be | |
115 | found in the advisories directory of the release tarball: | |
116 | ||
117 | GLIBC-SA-2024-0004: | |
118 | ISO-2022-CN-EXT: fix out-of-bound writes when writing escape | |
119 | sequence (CVE-2024-2961) | |
120 | ||
121 | GLIBC-SA-2024-0005: | |
122 | nscd: Stack-based buffer overflow in netgroup cache (CVE-2024-33599) | |
123 | ||
124 | GLIBC-SA-2024-0006: | |
125 | nscd: Null pointer crash after notfound response (CVE-2024-33600) | |
126 | ||
127 | GLIBC-SA-2024-0007: | |
128 | nscd: netgroup cache may terminate daemon on memory allocation | |
129 | failure (CVE-2024-33601) | |
130 | ||
131 | GLIBC-SA-2024-0008: | |
132 | nscd: netgroup cache assumes NSS callback uses in-buffer strings | |
133 | (CVE-2024-33602) | |
134 | ||
135 | The following bugs are resolved with this release: | |
136 | ||
137 | [19622] network: Support aliasing with struct sockaddr | |
138 | [21271] localedata: cv_RU: update translations | |
139 | [23774] localedata: lv_LV collates Y/y incorrectly | |
140 | [23865] string: wcsstr is quadratic-time | |
141 | [25119] localedata: Change Czech weekday names to lowercase | |
142 | [27777] stdio: fclose does a linear search, takes ages when many FILE* | |
143 | are opened | |
144 | [29770] libc: prctl does not match manual page ABI on powerpc64le- | |
145 | linux-gnu | |
146 | [29845] localedata: Update hr_HR locale currency to € | |
147 | [30701] time: getutxent misbehaves on 32-bit x86 when _TIME_BITS=64 | |
148 | [31316] build: Fails test misc/tst-dirname "Didn't expect signal from | |
149 | child: got `Illegal instruction'" on non SSE CPUs | |
150 | [31317] dynamic-link: [RISCV] static PIE crashes during self | |
151 | relocation | |
152 | [31325] libc: mips: clone3 is wrong for o32 | |
153 | [31335] math: Compile glibc with -march=x86-64-v3 should disable FMA4 | |
154 | multi-arch version | |
155 | [31339] libc: arm32 loader crash after cleanup in 2.36 | |
156 | [31340] manual: A bad sentence in section 22.3.5 (resource.texi) | |
157 | [31357] dynamic-link: $(objpfx)tst-rtld-list-diagnostics.out rule | |
158 | doesn't work with test wrapper | |
159 | [31370] localedata: wcwidth() does not treat | |
160 | DEFAULT_IGNORABLE_CODE_POINTs as zero-width | |
161 | [31371] dynamic-link: x86-64: APX and Tile registers aren't preserved | |
162 | in ld.so trampoline | |
163 | [31372] dynamic-link: _dl_tlsdesc_dynamic doesn't preserve all caller- | |
164 | saved registers | |
165 | [31383] libc: _FORTIFY_SOURCE=3 and __fortified_attr_access vs size of | |
166 | 0 and zero size types | |
167 | [31385] build: sort-makefile-lines.py doesn't check variable with _ | |
168 | nor with "^# variable" | |
169 | [31402] libc: clone (NULL, NULL, ...) clobbers %r7 register on | |
170 | s390{,x} | |
171 | [31405] libc: Improve dl_iterate_phdr using _dl_find_object | |
172 | [31411] localedata: Add Latgalian locale | |
173 | [31412] build: GCC 6 failed to build i386 glibc on Fedora 39 | |
174 | [31429] build: Glibc failed to build with -march=x86-64-v3 | |
175 | [31468] libc: sigisemptyset returns true when the set contains signals | |
176 | larger than 34 | |
177 | [31476] network: Automatic activation of single-request options break | |
178 | resolv.conf reloading | |
179 | [31479] libc: Missing #include <sys/rseq.h> in sched_getcpu.c may | |
180 | result in a loss of rseq acceleration | |
181 | [31501] dynamic-link: _dl_tlsdesc_dynamic_xsavec may clobber %rbx | |
182 | [31518] manual: documentation: FLT_MAX_10_EXP questionable text, evtl. | |
183 | wrong, | |
184 | [31530] localedata: Locale file for Moksha - mdf_RU | |
185 | [31553] malloc: elf/tst-decorate-maps fails on ppc64el | |
186 | [31596] libc: On the llvm-arm32 platform, dlopen("not_exist.so", -1) | |
187 | triggers segmentation fault | |
188 | [31600] math: math: x86 ceill traps when FE_INEXACT is enabled | |
189 | [31601] math: math: x86 floor traps when FE_INEXACT is enabled | |
190 | [31603] math: math: x86 trunc traps when FE_INEXACT is enabled | |
191 | [31612] libc: arc4random fails to fallback to /dev/urandom if | |
192 | getrandom is not present | |
193 | [31629] build: powerpc64: Configuring with "--with-cpu=power10" and | |
194 | 'CFLAGS=-O2 -mcpu=power9' fails to build glibc | |
195 | [31640] dynamic-link: POWER10 ld.so crashes in | |
196 | elf_machine_load_address with GCC 14 | |
197 | [31661] libc: NPROCESSORS_CONF and NPROCESSORS_ONLN not available in | |
198 | getconf | |
199 | [31676] dynamic-link: Configuring with CC="gcc -march=x86-64-v3" | |
200 | --with-rtld-early-cflags=-march=x86-64 results in linker failure | |
201 | [31677] nscd: nscd: netgroup cache: invalid memcpy under low | |
202 | memory/storage conditions | |
203 | [31678] nscd: nscd: Null pointer dereferences after failed netgroup | |
204 | cache insertion | |
205 | [31679] nscd: nscd: netgroup cache may terminate daemon on memory | |
206 | allocation failure | |
207 | [31680] nscd: nscd: netgroup cache assumes NSS callback uses in-buffer | |
208 | strings | |
209 | [31682] math: [PowerPC] Floating point exception error for math test | |
210 | test-ceil-except-2 test-floor-except-2 test-trunc-except-2 | |
211 | [31686] dynamic-link: Stack-based buffer overflow in | |
212 | parse_tunables_string | |
213 | [31695] libc: pidfd_spawn/pidfd_spawnp leak an fd if clone3 succeeds | |
214 | but execve fails | |
215 | [31719] dynamic-link: --enable-hardcoded-path-in-tests doesn't work | |
216 | with -Wl,--enable-new-dtags | |
217 | [31730] libc: backtrace_symbols_fd prints different strings than | |
218 | backtrace_symbols returns | |
219 | [31753] build: FAIL: link-static-libc with GCC 6/7/8 | |
220 | [31755] libc: procutils_read_file doesn't start with a leading | |
221 | underscore | |
222 | [31756] libc: write_profiling is only in libc.a | |
223 | [31757] build: Should XXXf128_do_not_use functions be excluded? | |
224 | [31759] math: Extra nearbyint symbols in libm.a | |
225 | [31760] math: Missing math functions | |
226 | [31764] build: _res_opcodes should be a compat symbol only | |
227 | [31765] dynamic-link: _dl_mcount_wrapper is exported without prototype | |
228 | [31766] stdio: _IO_stderr_ _IO_stdin_ _IO_stdout should be compat | |
229 | symbols | |
230 | [31768] string: Extra stpncpy symbol in libc.a | |
231 | [31770] libc: clone3 is in libc.a | |
232 | [31774] libc: Missing __isnanf128 in libc.a | |
233 | [31775] math: Missing exp10 exp10f32x exp10f64 fmod fmodf fmodf32 | |
234 | fmodf32x fmodf64 in libm.a | |
235 | [31777] string: Extra memchr strlen symbols in libc.a | |
236 | [31781] math: Missing math functions in libm.a | |
237 | [31782] build: Test build failure with recent GCC trunk (x86/tst-cpu- | |
238 | features-supports.c:69:3: error: parameter to builtin not valid: | |
239 | avx5124fmaps) | |
240 | [31785] string: loongarch: Extra strnlen symbols in libc.a | |
241 | [31786] string: powerpc: Extra strchrnul and strncasecmp_l symbols in | |
242 | libc.a | |
243 | [31787] math: powerpc: Extra llrintf, llrintf, llrintf32, and | |
244 | llrintf32 symbols in libc.a | |
245 | [31788] libc: microblaze: Extra cacheflush symbol in libc.a | |
246 | [31789] libc: powerpc: Extra versionsort symbol in libc.a | |
247 | [31790] libc: s390: Extra getutent32, getutent32_r, getutid32, | |
248 | getutid32_r, getutline32, getutline32_r, getutmp32, getutmpx32, | |
249 | getutxent32, getutxid32, getutxline32, pututline32, pututxline32, | |
250 | updwtmp32, updwtmpx32 in libc.a | |
251 | [31797] build: g++ -static requirement should be able to opt-out | |
252 | [31798] libc: pidfd_getpid.c is miscompiled by GCC 6.4 | |
253 | [31802] time: difftime is pure not const | |
254 | [31808] time: The supported time_t range is not documented. | |
255 | [31840] stdio: Memory leak in _IO_new_fdopen (fdopen) on seek failure | |
256 | [31867] build: "CPU ISA level is lower than required" on SSE2-free | |
257 | CPUs | |
258 | [31876] time: "Date and time" documentation fixes for POSIX.1-2024 etc | |
259 | [31883] build: ISA level support configure check relies on bashism / | |
260 | is otherwise broken for arithmetic | |
261 | [31892] build: Always install mtrace. | |
262 | [31917] libc: clang mq_open fortify wrapper does not handle 4 argument | |
263 | correctly | |
264 | [31927] libc: clang open fortify wrapper does not handle argument | |
265 | correctly | |
266 | [31931] time: tzset may fault on very short TZ string | |
267 | [31934] string: wcsncmp crash on s390x on vlbb instruction | |
268 | [31963] stdio: Crash in _IO_link_in within __gcov_exit | |
269 | [31965] dynamic-link: rseq extension mechanism does not work as | |
270 | intended | |
271 | [31980] build: elf/tst-tunables-enable_secure-env fails on ppc | |
272 | \f | |
273 | Version 2.39 | |
274 | ||
275 | Major new features: | |
276 | ||
277 | * A new tunable, glibc.cpu.plt_rewrite, can be used to enable PLT | |
278 | rewrite on x86-64. When enabled with non-lazy binding, the dynamic | |
279 | linker will rewrite indirect branches in PLT with direct branches. | |
280 | ||
281 | * Sync with Linux kernel 6.6 shadow stack interface. The --enable-cet | |
282 | configure option is only supported on x86-64. | |
283 | ||
284 | * struct statvfs now has an f_type member, equal to the f_type statfs member; | |
285 | on the Hurd this was always available under a reserved name, | |
286 | and under Linux a spare has been allocated: it was always zero | |
287 | in previous versions of glibc, and zero is not a valid result. | |
288 | ||
289 | * On Linux, the functions posix_spawnattr_getcgroup_np and | |
290 | posix_spawnattr_setcgroup_np have been added, along with the | |
291 | POSIX_SPAWN_SETCGROUP flag. They allow posix_spawn and posix_spawnp | |
292 | to set the cgroupv2 in the new process in a race-free manner. These | |
293 | functions are GNU extensions and require a kernel with clone3 support. | |
294 | ||
295 | * On Linux, the pidfd_spawn and pidfd_spawp functions have been added. | |
296 | They have a similar prototype and semantic as posix_spawn, but instead of | |
297 | returning a process ID, they return a file descriptor that can be used | |
298 | along other pidfd functions (like pidfd_send_signal, poll, or waitid). | |
299 | The pidfd functionality avoids the issue of PID reuse with the traditional | |
300 | posix_spawn interface. | |
301 | ||
302 | * On Linux, the pidfd_getpid function has been added. It allows retrieving | |
303 | the process ID associated with the process file descriptor created by | |
304 | pid_spawn, fork_np, or pidfd_open. | |
305 | ||
306 | * scanf-family functions now support the wN format length modifiers for | |
307 | arguments pointing to types intN_t, int_leastN_t, uintN_t or | |
308 | uint_leastN_t (for example, %w32d to read int32_t or int_least32_t in | |
309 | decimal, or %w32x to read uint32_t or uint_least32_t in hexadecimal) | |
310 | and the wfN format length modifiers for arguments pointing to types | |
311 | int_fastN_t or uint_fastN_t, as specified in draft ISO C2X. | |
312 | ||
313 | * A new tunable, glibc.mem.decorate_maps, can be used to add additional | |
314 | information on underlying memory allocated by the glibc (for instance, | |
315 | on thread stack created by pthread_create or memory allocated by | |
316 | malloc). | |
317 | ||
318 | * The <stdbit.h> header has been added from ISO C2X, with | |
319 | stdc_leading_zeros, stdc_leading_ones, stdc_trailing_zeros, | |
320 | stdc_trailing_ones, stdc_first_leading_zero, stdc_first_leading_one, | |
321 | stdc_first_trailing_zero, stdc_first_trailing_one, stdc_count_zeros, | |
322 | stdc_count_ones, stdc_has_single_bit, stdc_bit_width, stdc_bit_floor | |
323 | and stdc_bit_ceil function families, each having functions for | |
324 | unsigned char, unsigned short, unsigned int, unsigned long int and | |
325 | unsigned long long int, and a type-generic macro. | |
326 | ||
327 | * On AArch64 new symbols were added to libmvec and now math.h has | |
328 | annotations to allow GCC 9 or newer to auto-vectorize calls to the | |
329 | following scalar math functions when -ffast-math is specified: | |
330 | acos, acosf, asin, asinf, atan, atanf, atan2, atan2f, cos, cosf, | |
331 | exp, expf, exp10, exp10f, exp2, exp2f, expm1, expm1f, log, logf, | |
332 | log10, log10f, log1p, log1pf, log2, log2f, sin, sinf, tan, tanf. | |
333 | ||
334 | Deprecated and removed features, and other changes affecting compatibility: | |
335 | ||
336 | * The ldconfig program now skips file names containing ';' or ending in | |
337 | ".dpkg.tmp" or ".dpkg.new", to avoid examining temporary files created | |
338 | by the RPM and dpkg package managers. | |
339 | ||
340 | * libcrypt has been removed from the GNU C Library. The configure | |
341 | options "--enable-crypt" and "--enable-nss-crypt" are no longer | |
342 | available. <crypt.h>, libcrypt.a, and libcrypt.so.1 will not be | |
343 | installed. For now <unistd.h> continues to declare the crypt | |
344 | function by default, to avoid introducing vulnerabilities into | |
345 | existing applications due to a missing prototype. This declaration | |
346 | is deprecated and may be removed in a future glibc release. | |
347 | ||
348 | The replacement for libcrypt is libxcrypt, maintained separately from | |
349 | GNU libc, but available under compatible licensing terms, and providing | |
350 | binary backward compatibility with the former libcrypt. It is currently | |
351 | distributed from <https://github.com/besser82/libxcrypt/>. | |
352 | ||
353 | As a consequence of this removal, GNU libc no longer makes any use of | |
354 | the NSS cryptography library (Network Security Services; not to be | |
355 | confused with Name Service Switch). Distributors of binary packages | |
356 | of GNU libc are advised to check whether their build processes can be | |
357 | simplified. | |
358 | ||
359 | * The dynamic linker calls the malloc and free functions in more cases | |
360 | during TLS access if a shared object with dynamic TLS is loaded and | |
361 | unloaded. This can result in an infinite recursion if a malloc | |
362 | replacement library or its dependencies use dynamic TLS instead of | |
363 | initial-exec TLS. | |
364 | ||
365 | * The ia64*-*-linux-gnu configurations are no longer supported. | |
366 | ||
367 | Changes to build and runtime requirements: | |
368 | ||
369 | * Building on LoongArch requires at a minimum binutils 2.41 for vector | |
370 | instructions. | |
371 | ||
372 | Security related changes: | |
373 | ||
374 | The following CVEs were fixed in this release, details of which can be | |
375 | found in the advisories directory of the release tarball: | |
376 | ||
377 | GLIBC-SA-2023-0002: | |
378 | getaddrinfo: Stack read overflow in no-aaaa mode (CVE-2023-4527) | |
379 | ||
380 | GLIBC-SA-2023-0003: | |
381 | getaddrinfo: Potential use-after-free (CVE-2023-4806) | |
382 | ||
383 | GLIBC-SA-2023-0004: | |
384 | tunables: local privilege escalation through buffer overflow | |
385 | (CVE-2023-4911) | |
386 | ||
387 | GLIBC-SA-2024-0001: | |
388 | syslog: Heap buffer overflow in __vsyslog_internal (CVE-2023-6246) | |
389 | ||
390 | GLIBC-SA-2024-0002: | |
391 | syslog: Heap buffer overflow in __vsyslog_internal (CVE-2023-6779) | |
392 | ||
393 | GLIBC-SA-2024-0003: | |
394 | syslog: Integer overflow in __vsyslog_internal (CVE-2023-6780) | |
395 | ||
396 | GLIBC-SA-2024-0004: | |
397 | ISO-2022-CN-EXT: fix out-of-bound writes when writing escape | |
398 | sequence (CVE-2024-2961) | |
399 | ||
400 | GLIBC-SA-2024-0005: | |
401 | nscd: Stack-based buffer overflow in netgroup cache (CVE-2024-33599) | |
402 | ||
403 | GLIBC-SA-2024-0006: | |
404 | nscd: Null pointer crashes after notfound response | |
405 | (CVE-2024-33600) | |
406 | ||
407 | GLIBC-SA-2024-0007: | |
408 | nscd: netgroup cache may terminate daemon on memory allocation | |
409 | failure (CVE-2024-33601) | |
410 | ||
411 | GLIBC-SA-2024-0008: | |
412 | nscd: netgroup cache assumes NSS callback uses in-buffer strings | |
413 | (CVE-2024-33602) | |
414 | ||
415 | The following bugs are resolved with this release: | |
416 | ||
417 | [14522] localedata: fy_DE: LC_IDENTIFICATION data looks weird | |
418 | [19305] libc: qsort() should return early if (nmemb <= 1) | |
419 | [19479] localedata: gbm_IN: new Garhwali Locale | |
420 | [19924] dynamic-link: TLS performance degradation after dlopen | |
421 | [19956] localedata: ssy_ER: rename from aa_ER@saaho | |
422 | [21719] libc: stdlib/msort : optimizing merge sort | |
423 | [22526] localedata: th_TH LC_COLLATE does not use copy "iso14651_t1" | |
424 | [23012] localedata: el_GR: Greece now uses the 24h format for time | |
425 | [23172] localedata: miq_NI: Provide actually abbreviated month names | |
426 | [24006] localedata: Cyclic dependencies via copy in locales | |
427 | [24013] localedata: am_pm definitions for es_ES | |
428 | [24386] localedata: crh_RU: new locale | |
429 | [24877] localedata: [Redundant Data] Remove redundant data between | |
430 | en_NZ and en_AU | |
431 | [25868] localedata: Incorrect trailing spaces in weekday names for | |
432 | nn_NO | |
433 | [26752] localedata: Please add the new locale zgh_MA | |
434 | [27069] dynamic-link: Need a way to tell if a tunable is set by user | |
435 | [27163] localedata: Error on test glk_IR with localedef | |
436 | [27312] localedata: su_ID: new Sundanese locale | |
437 | [27547] manual: "Summary of malloc-Related Functions" shows wrong | |
438 | argument order for `aligned_alloc` and `memalign` | |
439 | [27574] libc: glibc should probably not define __WORDSIZE=64 for | |
440 | __sparcv9 | |
441 | [27601] localedata: License information update in | |
442 | localedata/locales/ast_ES | |
443 | [28558] localedata: it_IT LC_MONETARY outdated p_cs_precedes and | |
444 | n_cs_precedes | |
445 | [28787] localedata: Add information for Occitan | |
446 | [29039] dynamic-link: Corrupt DTV after reuse of a TLS module ID | |
447 | following dlclose with unused TLS | |
448 | [29486] localedata: New Zealand locales (en_NZ & mi_NZ) first day of | |
449 | week should be Monday | |
450 | [29504] localedata: Incorrect/misleading Time Format For ms_MY (AM/PM) | |
451 | [29506] localedata: UTF-8 HANGUL SYLLABLE bugs | |
452 | [30349] libc: Support returning a pidfd from posix_spawn() | |
453 | [30412] localedata: d_t_fmt in id_ID uses %r placeholder but am_pm and | |
454 | t_fmt_ampm are undefined | |
455 | [30605] localedata: New locale for Komi language | |
456 | [30649] localedata: [PATCH] Add transliteration of common emojis to | |
457 | smileys | |
458 | [30694] locale: The iconv program no longer tells the user which given | |
459 | encoding name was wrong | |
460 | [30709] nscd: nscd fails to build with cleanup handler if built with | |
461 | -fexceptions | |
462 | [30737] libc: fdopendir() is not robust - returns bogus DIR* instead | |
463 | of flagging an error | |
464 | [30740] build: [m68k] undefined reference to | |
465 | `_wordcopy_fwd_dest_aligned' | |
466 | [30745] libc: Slight bug in cache info codes for x86 | |
467 | [30750] network: Unaligned accesses in resolver | |
468 | [30773] math: [m68k] busybox awk is broken (lshift.S related) | |
469 | [30789] libc: [2.38 Regression] sem_open will fail on multithreaded | |
470 | scenarios when semaphore file doesn't exist (O_CREAT) | |
471 | [30800] nscd: Improper assert in prune_cache triggers if clock jumps | |
472 | backwards | |
473 | [30804] libc: F_GETLK, F_SETLK, and F_SETLKW value change for | |
474 | powerpc64 with -D_FILE_OFFSET_BITS=64 | |
475 | [30842] network: Stack read overflow in getaddrinfo in no-aaaa mode | |
476 | (CVE-2023-4527) | |
477 | [30843] network: potential use-after-free in getcanonname | |
478 | (CVE-2023-4806) | |
479 | [30854] localedata: Update locale data to Unicode 15.1.0 | |
480 | [30884] network: Memory leak in getaddrinfo after fix for bug 30843 | |
481 | (CVE-2023-5156) | |
482 | [30932] libc: Fortify Source has false-positives when too many files | |
483 | are open | |
484 | [30945] malloc: Core affinity setting incurs lock contentions between | |
485 | threads | |
486 | [30960] math: signed integer overflow in | |
487 | glibc/sysdeps/s390/fpu/feenablxcpt.c | |
488 | [30964] locale: Number grouping check mishandles multibyte thousands | |
489 | separator | |
490 | [30981] dynamic-link: dlclose does not properly implement force-first | |
491 | handling | |
492 | [30988] math: fesetexcept raises floating-point exception traps on | |
493 | ppc, ppc64, ppc64le | |
494 | [30989] math: fesetexcept raises floating-point exception traps on | |
495 | i386 | |
496 | [30990] libc: fesetexceptflag raises floating-point exception traps on | |
497 | i386, x86_64 | |
498 | [30998] math: fesetexceptflag clears too many floating-point exception | |
499 | flags on alpha | |
500 | [31019] manual: The documentation of feenableexcept is incomplete | |
501 | [31022] math: feupdateenv (FE_DFL_ENV) crashes on riscv | |
502 | [31035] libc: Library search path terminates on relative non-directory | |
503 | name | |
504 | [31042] libc: [s390x] .init and .fini padding | |
505 | [31068] libc: sysdeps: sparc: invalid data access in memset due to | |
506 | regression | |
507 | [31078] manual: Code example in "Noncanonical Mode Example" has unused | |
508 | 'char *name;' | |
509 | [31086] localedata: Errors in Tibetan, Dzongkha data | |
510 | [31113] string: Wrong unwind information for rawmemchr on aarch64 | |
511 | [31151] libc: [RISC-V] missing support for profile/audit PLT setup | |
512 | [31163] nss: getaddrinfo returns EAI_NONAME in oom situation | |
513 | [31183] stdio: Wide stream buffer size reduced MB_LEN_MAX bytes after | |
514 | bug 17522 fix | |
515 | [31184] dynamic-link: FAIL: elf/tst-tlsgap | |
516 | [31185] dynamic-link: Incorrect thread point access in | |
517 | _dl_tlsdesc_undefweak and _dl_tlsdesc_dynamic | |
518 | [31187] dynamic-link: Some CET tests fail with GCC 14 | |
519 | [31204] localedata: Fix decimal point and thousands separator for | |
520 | uz_UZ | |
521 | [31205] localedata: Inconsistent (mon_)grouping formats | |
522 | [31218] dynamic-link: PLT rewrite overflows large displacement on x32 | |
523 | [31221] localedata: Add localedata for ISO code "tok" (Toki Pona) | |
524 | [31230] dynamic-link: PLT rewrite failed without SELinux | |
525 | [31239] localedata: anp_IN locale: abbreviated month names are the | |
526 | same as the full month names | |
527 | [31244] nptl: pthread_cancel hangs on sparc32 | |
528 | [31257] localedata: Sync with CLDR: “Turkey” -> “Türkiye” | |
529 | [31266] string: sparc: string/tst-memmove-overflow fails on 32-bit | |
530 | sparcv9 | |
531 | [31276] libc: Wrong condition for heap allocation in qsort_r | |
532 | \f | |
533 | Version 2.38 | |
534 | ||
535 | Major new features: | |
536 | ||
537 | * When C2X features are enabled and the base argument is 0 or 2, the | |
538 | following functions support binary integers prefixed by 0b or 0B as | |
539 | input: strtol, strtoll, strtoul, strtoull, strtol_l, strtoll_l, | |
540 | strtoul_l, strtoull_l, strtoimax, strtoumax, strtoq, strtouq, wcstol, | |
541 | wcstoll, wcstoul, wcstoull, wcstol_l, wcstoll_l, wcstoul_l, | |
542 | wcstoull_l, wcstoimax, wcstoumax, wcstoq, wcstouq. Similarly, the | |
543 | following functions support binary integers prefixed by 0b or 0B as | |
544 | input to the %i format: fscanf, scanf, sscanf, vscanf, vsscanf, | |
545 | vfscanf, fwscanf, wscanf, swscanf, vfwscanf, vwscanf, vswscanf; those | |
546 | functions also support the %b format for binary integers, with or | |
547 | without such a prefix and independent of standards mode. | |
548 | ||
549 | * PRIb*, PRIB* and SCNb* macros from C2X have been added to | |
550 | <inttypes.h>. | |
551 | ||
552 | * printf-family functions now support the wN format length modifiers for | |
553 | arguments of type intN_t, int_leastN_t, uintN_t or uint_leastN_t (for | |
554 | example, %w32d to print int32_t or int_least32_t in decimal, or %w32x | |
555 | to print uint32_t or uint_least32_t in hexadecimal) and the wfN format | |
556 | length modifiers for arguments of type int_fastN_t or uint_fastN_t, as | |
557 | specified in draft ISO C2X. | |
558 | ||
559 | * A new tunable, glibc.pthread.stack_hugetlb, can be used to disable | |
560 | Transparent Huge Pages (THP) in stack allocation at pthread_create. | |
561 | ||
562 | * Support for x86_64 running on Hurd has been added. This port requires | |
563 | as least binutils 2.40 and GCC 13: | |
564 | ||
565 | - x86_64-gnu | |
566 | ||
567 | * Vector math library libmvec support has been added to AArch64. It | |
568 | requires GCC version >= 10.1.0. It can be disabled via | |
569 | "--disable-mathvec", however that is not a supported configuration as | |
570 | it changes the ABI. The symbol names follow the AArch64 vector ABI, | |
571 | they are declared in math.h and have to be called manually at this point. | |
572 | ||
573 | * The strlcpy and strlcat functions have been added. They are derived | |
574 | from OpenBSD, and are expected to be added to a future POSIX version. | |
575 | ||
576 | * A new configure option, "--enable-fortify-source", can be used to build the | |
577 | GNU C Library with _FORTIFY_SOURCE. The level of fortification can either be | |
578 | provided, or is set to the highest value supported by the compiler. If not | |
579 | explicitly enabled, then fortify source is forcibly disabled so to keep | |
580 | original behavior unchanged. | |
581 | ||
582 | Deprecated and removed features, and other changes affecting compatibility: | |
583 | ||
584 | * libcrypt is no longer built by default; one may use the "--enable-crypt" | |
585 | option to build libcrypt. libcrypt is likely to be removed from the | |
586 | GNU C Library in a future release, so it is recommended that | |
587 | applications port away from it to an alternative such as libxcrypt. | |
588 | ||
589 | * In the Linux kernel for the hppa/parisc architecture some of the | |
590 | MADV_XXX constants were changed to have the same values as the other | |
591 | architectures. New programs compiled with this glibc version and which | |
592 | use the madvise call will require at least Linux kernel version 6.2, | |
593 | alternatively stable kernels from versions 6.1.6, 5.15.87, 5.10.163, | |
594 | 5.4.228, 4.19.270 or 4.14.303. | |
595 | ||
596 | * The "--disable-experimental-malloc" option is no longer available. The | |
597 | per-thread cache can still be disabled per-application using tunables | |
598 | (glibc.malloc.tcache_count set to zero). | |
599 | ||
600 | * The configure option "--enable-tunables" has been removed. The tunable | |
601 | feature is now always enabled. | |
602 | ||
603 | Changes to build and runtime requirements: | |
604 | ||
605 | * Building libmvec on AArch64 requires at a minimum GCC 10.1.0 for SVE | |
606 | ACLE. | |
607 | ||
608 | Security related changes: | |
609 | ||
610 | CVE-2023-25139: When the printf family of functions is called with a | |
611 | format specifier that uses an <apostrophe> (enable grouping) and a | |
612 | minimum width specifier, the resulting output could be larger than | |
613 | reasonably expected by a caller that computed a tight bound on the | |
614 | buffer size. The resulting larger than expected output could result | |
615 | in a buffer overflow in the printf family of functions. | |
616 | ||
617 | The following bugs are resolved with this release: | |
618 | ||
619 | [178] string: Please add strlcpy and strlcat (attached) | |
620 | [14697] nptl: Behavior of exit is nonconformant with respect to | |
621 | threads and stdio | |
622 | [15142] stdio: Missing locking in _IO_cleanup | |
623 | [18096] glob: null deref in wordexp/parse_dollars/parse_arith | |
624 | [18906] stdio: fopen: ccs value may affect open mode | |
625 | [24466] stdio: Feature request: provide special printf formats for | |
626 | intXX_t | |
627 | [25457] nss: hosts lookup fails for ipv4mapped ipv6 addresses | |
628 | [28519] libc: system and popen should pass "--" between /bin/sh and | |
629 | argument | |
630 | [29016] stdio: popen() sets errno to ENOMEM when shell does not exist | |
631 | [29591] string: wcsnlen length can overflow in page cross case. | |
632 | [30053] time: strftime %s returns -1 after 2038 on 32 bits systems | |
633 | [30068] stdio: incorrect printf output for integers with thousands | |
634 | separator and width field (CVE-2023-25139) | |
635 | [30111] time: support_descriptors_list fails after 2038 on 32 bits | |
636 | systems | |
637 | [30125] dynamic-link: [regression, bisected] glibc-2.37 creates new | |
638 | symlink for libraries without soname | |
639 | [30130] math: [s390] The _FPU_SETCW macro yields compile error with | |
640 | Clang | |
641 | [30156] time: Potential ntp_gettime abi break | |
642 | [30235] libc: Missing fallback in getlogin if loginuid is unset | |
643 | [30258] dynamic-link: sprof cannot read and display shared object | |
644 | profiling data correctly | |
645 | [30263] libc: Add test coverage for abs(), labs(), and llabs(). | |
646 | [30305] math: Incorrect asm constraint in feraiseexcept on x86-64 | |
647 | [30402] libc: FAIL: elf/tst-glibcelf | |
648 | [30425] dynamic-link: Symbol lookup during dlclose may fail | |
649 | unnecessarily | |
650 | [30435] dynamic-link: Root dir wrongly marked as nonexist in open_path | |
651 | [30477] libc: [RISCV]: time64 does not work on riscv32 | |
652 | [30515] dynamic-link: _dl_find_object incorrectly returns 1 during | |
653 | early startup | |
654 | [30527] network: resolv_conf lock not unlocked on allocation failure | |
655 | [30550] math: powerpc64le: GCC-specific code for isinf() is being used | |
656 | on clang | |
657 | [30555] string: strerror can incorrectly return NULL | |
658 | [30579] malloc: trim_threshold in realloc lead to high memory usage | |
659 | [30662] nscd: Group and password cache use errno in place of errval | |
660 | \f | |
661 | Version 2.37 | |
662 | ||
663 | Major new features: | |
664 | ||
665 | * The getent tool now supports the --no-addrconfig option. The output of | |
666 | getent with --no-addrconfig may contain addresses of families not | |
667 | configured on the current host i.e. as-if you had not passed | |
668 | AI_ADDRCONFIG to getaddrinfo calls. | |
669 | ||
670 | Deprecated and removed features, and other changes affecting compatibility: | |
671 | ||
672 | * The dynamic linker no longer loads shared objects from the "tls" | |
673 | subdirectories on the library search path or the subdirectory that | |
674 | corresponds to the AT_PLATFORM system name, or employs the legacy AT_HWCAP | |
675 | search mechanism, which was deprecated in version 2.33. | |
676 | ||
677 | Security related changes: | |
678 | ||
679 | CVE-2022-39046: When the syslog function is passed a crafted input | |
680 | string larger than 1024 bytes, it reads uninitialized memory from the | |
681 | heap and prints it to the target log file, potentially revealing a | |
682 | portion of the contents of the heap. | |
683 | ||
684 | The following bugs are resolved with this release: | |
685 | ||
686 | [12154] network: Cannot resolve hosts which have wildcard aliases | |
687 | [12165] libc: readdir: Do not skip entries with zero d_ino values | |
688 | [19444] build: build failures with -O1 due to -Wmaybe-uninitialized | |
689 | [24774] nptl: pthread_rwlock_timedwrlock stalls on ARM | |
690 | [24816] nss: nss/tst-nss-files-hosts-long fails when no interface has | |
691 | AF_INET6 address (ie docker) | |
692 | [27087] stdio: PowerPC: Redefinition error with Clang from IEEE | |
693 | redirection headers | |
694 | [28846] network: CMSG_NXTHDR may trigger -Wstrict-overflow warning | |
695 | [28937] dynamic-link: New DSO dependency sorter does not put new map | |
696 | first if in a cycle | |
697 | [29249] libc: csu/libc-tls.c:202: undefined reference to | |
698 | `_startup_fatal_not_constant' | |
699 | [29305] network: Inefficient buffer space usage in nss_dns for | |
700 | gethostbyname and other functions | |
701 | [29375] libc: don't hide MAP_ANONYMOUS behind _GNU_SOURCE | |
702 | [29402] nscd: nscd: No such file or directory | |
703 | [29415] nscd: getaddrinfo with AI_ADDRCONFIG returns addresses with | |
704 | wrong family | |
705 | [29427] dynamic-link: Inconsistency detected by ld.so: dl-printf.c: | |
706 | 200: _dl_debug_vdprintf: Assertion `! "invalid format specifier"' | |
707 | failed! | |
708 | [29463] math: math/test-float128-y1 fails on x86_64 | |
709 | [29485] build: Make hangs when the test misc/tst-pidfile returns | |
710 | FAIL_UNSUPPORTED | |
711 | [29490] dynamic-link: [bisected] new __brk_call causes dynamic loader | |
712 | segfault on alpha | |
713 | [29499] build: Check failed on misc/tst-glibcsyscalls while building | |
714 | for RISCV64 on a unmatched hardware | |
715 | [29501] build: Check failed on stdlib/tst-strfrom while building for | |
716 | RISCV64 on a unmatched hardware | |
717 | [29502] libc: alpha sys/acct.h out of date | |
718 | [29514] build: Need to use -fPIE not -fpie | |
719 | [29528] dynamic-link: __libc_early_init not called after dlmopen that | |
720 | reuses namespace | |
721 | [29536] libc: syslog fail to create large messages (CVE-2022-39046) | |
722 | [29537] libc: [2.34 regression]: Alignment issue on m68k when using | |
723 | futexes on qemu-user | |
724 | [29539] libc: LD_TRACE_LOADED_OBJECTS changed how vDSO library are | |
725 | printed | |
726 | [29544] libc: Regression in syslog(3) calls breaks RFC due to extra | |
727 | whitespace | |
728 | [29564] build: Incorrect way to change MAKEFLAGS in Makerules | |
729 | [29576] build: librtld.os: in function `_dl_start_profile': | |
730 | (.text+0x9444): undefined reference to `strcpy' | |
731 | [29578] libc: Definition of SUN_LEN() is wrong | |
732 | [29583] build: iconv failures on 32bit platform due to missing large | |
733 | file support | |
734 | [29600] dynamic-link: dlmopen hangs after loading certain libraries | |
735 | [29604] localedata: Update locale data to Unicode 15.0.0 | |
736 | [29605] nscd: Regression in NSCD backend of getaddrinfo | |
737 | [29607] nscd: nscd repeatably crashes calling __strlen_avx2 when hosts | |
738 | cache is enabled | |
739 | [29611] string: Optimized AVX2 string functions unconditionally use | |
740 | BMI2 instructions | |
741 | [29624] malloc: errno is not cleared when entering main | |
742 | [29638] libc: stdlib: arc4random fallback is never used | |
743 | [29657] libc: Incorrect struct stat for 64-bit time on linux/generic | |
744 | platforms | |
745 | [29698] build: Configuring for AArch32 on ARMv8+ disables | |
746 | optimizations | |
747 | [29727] locale: __strtol_internal out-of-bounds read when parsing | |
748 | thousands grouping | |
749 | [29730] libc: broken y2038 support in fstatat on MIPS N64 | |
750 | [29746] libc: ppoll() does not switch to __ppoll64 when | |
751 | -D_TIME_BITS=64 and -D_FORTIFY_SOURCE=2 is given on 32bit | |
752 | [29771] libc: Restore IPC_64 support in sysvipc *ctl functions | |
753 | [29780] build: possible parallel make issue in glibc-2.36 (siglist- | |
754 | aux.S: No such file or directory) | |
755 | [29864] libc: __libc_start_main() should obtain program headers | |
756 | address (_dl_phdr) from the auxv, not the ELF header. | |
757 | [29951] time: daylight variable not set correctly if last DST change | |
758 | coincides with offset change | |
759 | [30039] stdio: __vsprintf_internal does not handle unspecified buffer | |
760 | length in fortify mode | |
761 | \f | |
762 | Version 2.36 | |
763 | ||
764 | Major new features: | |
765 | ||
766 | * Support for DT_RELR relative relocation format has been added to | |
767 | glibc. This is a new ELF dynamic tag that improves the size of | |
768 | relative relocations in shared object files and position independent | |
769 | executables (PIE). DT_RELR generation requires linker support for | |
770 | -z pack-relative-relocs option, which is supported for some targets | |
771 | in recent binutils versions. Lazy binding doesn't apply to DT_RELR. | |
772 | ||
773 | * On Linux, the pidfd_open, pidfd_getfd, and pidfd_send_signal functions | |
774 | have been added. The pidfd functionality provides access to a process | |
775 | while avoiding the issue of PID reuse on traditional Unix systems. | |
776 | ||
777 | * On Linux, the process_madvise function has been added. It has the | |
778 | same functionality as madvise but alters the target process identified | |
779 | by the pidfd. | |
780 | ||
781 | * On Linux, the process_mrelease function has been added. It allows a | |
782 | caller to release the memory of a dying process. The release of the | |
783 | memory is carried out in the context of the caller, using the caller's | |
784 | CPU affinity, and priority with CPU usage accounted to the caller. | |
785 | ||
786 | * The “no-aaaa” DNS stub resolver option has been added. System | |
787 | administrators can use it to suppress AAAA queries made by the stub | |
788 | resolver, including AAAA lookups triggered by NSS-based interfaces | |
789 | such as getaddrinfo. Only DNS lookups are affected: IPv6 data in | |
790 | /etc/hosts is still used, getaddrinfo with AI_PASSIVE will still | |
791 | produce IPv6 addresses, and configured IPv6 name servers are still | |
792 | used. To produce correct Name Error (NXDOMAIN) results, AAAA queries | |
793 | are translated to A queries. The new resolver option is intended | |
794 | primarily for diagnostic purposes, to rule out that AAAA DNS queries | |
795 | have adverse impact. It is incompatible with EDNS0 usage and DNSSEC | |
796 | validation by applications. | |
797 | ||
798 | * On Linux, the fsopen, fsmount, move_mount, fsconfig, fspick, open_tree, | |
799 | and mount_setattr have been added. They are part of the new Linux kernel | |
800 | mount APIs that allow applications to more flexibly configure and operate | |
801 | on filesystem mounts. The new mount APIs are specifically designed to work | |
802 | with namespaces. | |
803 | ||
804 | * localedef now accepts locale definition files encoded in UTF-8. | |
805 | Previously, input bytes not within the ASCII range resulted in | |
806 | unpredictable output. | |
807 | ||
808 | * Support for the mbrtoc8 and c8rtomb multibyte/UTF-8 character conversion | |
809 | functions has been added per the ISO C2X N2653 and C++20 P0482R6 proposals. | |
810 | Support for the char8_t typedef has been added per the ISO C2X N2653 | |
811 | proposal. The functions are declared in uchar.h in C2X mode or when the | |
812 | _GNU_SOURCE macro or C++20 __cpp_char8_t feature test macro is defined. | |
813 | The char8_t typedef is declared in uchar.h in C2X mode or when the | |
814 | _GNU_SOURCE macro is defined and the C++20 __cpp_char8_t feature test macro | |
815 | is not defined (if __cpp_char8_t is defined, then char8_t is a builtin type). | |
816 | ||
817 | * The functions arc4random, arc4random_buf, and arc4random_uniform have been | |
818 | added. The functions wrap getrandom and/or /dev/urandom to return high- | |
819 | quality randomness from the kernel. | |
820 | ||
821 | * Support for LoongArch running on Linux has been added. This port requires | |
822 | as least binutils 2.38, GCC 12, and Linux 5.19. Currently only hard-float | |
823 | ABI is supported: | |
824 | ||
825 | - loongarch64-linux-gnu | |
826 | ||
827 | The LoongArch ABI is 64-bit little-endian. | |
828 | ||
829 | Deprecated and removed features, and other changes affecting compatibility: | |
830 | ||
831 | * Support for prelink will be removed in the next release; this includes | |
832 | removal of the LD_TRACE_PRELINKING, and LD_USE_LOAD_BIAS, environment | |
833 | variables and their functionality in the dynamic loader. | |
834 | ||
835 | * The Linux kernel version check has been removed along with the | |
836 | LD_ASSUME_KERNEL environment variable. The minimum kernel used to built | |
837 | glibc is still provided through NT_GNU_ABI_TAG ELF note and also printed | |
838 | when libc.so is issued directly. | |
839 | ||
840 | * On Linux, The LD_LIBRARY_VERSION environment variable has been removed. | |
841 | ||
842 | The following bugs are resolved with this release: | |
843 | ||
844 | [14932] dynamic-link: dlsym(handle, "foo") and dlsym(RTLD_NEXT, "foo") | |
845 | return different result with versioned "foo" | |
846 | [16355] libc: syslog.h's SYSLOG_NAMES namespace violation and utter | |
847 | mess | |
848 | [23293] dynamic-link: aarch64: getauxval is broken when run as ld.so | |
849 | ./exe and ld.so adjusts argv on the stack | |
850 | [24595] nptl: [2.28 Regression]: Deadlock in atfork handler which | |
851 | calls dlclose | |
852 | [25744] locale: mbrtowc with Big5-HKSCS returns 2 instead of 1 when | |
853 | consuming the second byte of certain double byte characters | |
854 | [25812] stdio: Libio vtable protection is sometimes only partially | |
855 | enforced | |
856 | [27054] libc: pthread_atfork handlers that call pthread_atfork | |
857 | deadlock | |
858 | [27924] dynamic-link: ld.so: Support DT_RELR relative relocation | |
859 | format | |
860 | [28128] build: declare_symbol_alias doesn't work for assembly codes | |
861 | [28566] network: getnameinfo with NI_NOFQDN is not thread safe | |
862 | [28752] nss: Segfault in getpwuid when stat fails | |
863 | [28815] libc: realpath should not copy to resolved buffer on error | |
864 | [28828] stdio: fputwc crashes | |
865 | [28838] libc: FAIL: elf/tst-p_align3 | |
866 | [28845] locale: ld-monetary.c should be updated to match ISO C and | |
867 | other standards. | |
868 | [28850] libc: linux: __get_nprocs_sched reads uninitialized memory | |
869 | from the stack | |
870 | [28852] libc: getaddrinfo leaks memory with AI_ALL | |
871 | [28853] libc: tst-spawn6 changes current foreground process group | |
872 | (breaks test isolation) | |
873 | [28857] libc: FAIL: elf/tst-audit24a | |
874 | [28860] build: --enable-kernel=5.1.0 build fails because of missing | |
875 | __convert_scm_timestamps | |
876 | [28865] libc: linux: _SC_NPROCESSORS_CONF and _SC_NPROCESSORS_ONLN are | |
877 | inaccurate without /sys and /proc | |
878 | [28868] dynamic-link: Dynamic loader DFS algorithm segfaults on | |
879 | missing libraries | |
880 | [28880] libc: Program crashes if date beyond 2038 | |
881 | [28883] libc: sysdeps/unix/sysv/linux/select.c: __select64 | |
882 | !__ASSUME_TIME64_SYSCALLS && !__ASSUME_PSELECT fails on Microblaze | |
883 | [28896] string: strncmp-avx2-rtm and wcsncmp-avx2-rtm fallback on non- | |
884 | rtm variants when avoiding overflow | |
885 | [28922] build: The .d dependency files aren't always generated | |
886 | [28931] libc: hosts lookup broken for SUCCESS=CONTINUE and | |
887 | SUCCESS=MERGE | |
888 | [28936] build: nm: No such file | |
889 | [28950] localedata: Add locale for ISO code "tok" (Toki Pona) | |
890 | [28953] nss: NSS lookup result can be incorrect if function lookup | |
891 | clobbers errno | |
892 | [28970] math: benchtest: libmvec benchmark doesn't build with make | |
893 | bench. | |
894 | [28991] libc: sysconf(_SC_NPROCESSORS_CONF) should read | |
895 | /sys/devices/system/cpu/possible | |
896 | [28993] libc: closefrom() iterates until max int if no access to | |
897 | /proc/self/fd/ | |
898 | [28996] libc: realpath fails to copy partial result to resolved buffer | |
899 | on ENOENT and EACCES | |
900 | [29027] math: [ia64] fabs fails with sNAN input | |
901 | [29029] nptl: poll() spuriously returns EINTR during thread | |
902 | cancellation and with cancellation disabled | |
903 | [29030] string: GLIBC 2.35 regression - Fortify crash on certain valid | |
904 | uses of mbsrtowcs (*** buffer overflow detected ***: terminated) | |
905 | [29062] dynamic-link: Memory leak in _dl_find_object_update if object | |
906 | is promoted to global scope | |
907 | [29069] libc: fstatat64_time64_statx wrapper broken on MIPS N32 with | |
908 | -D_FILE_OFFSET_BITS=64 and -D_TIME_BITS=64 | |
909 | [29071] dynamic-link: m68k: Removal of ELF_DURING_STARTUP optimization | |
910 | broke ld.so | |
911 | [29097] time: fchmodat does not handle 64 bit time_t for | |
912 | AT_SYMLINK_NOFOLLOW | |
913 | [29109] libc: posix_spawn() always returns 1 (EPERM) on clone() | |
914 | failure | |
915 | [29141] libc: _FORTIFY_SOURCE=3 fail for gcc 12/glibc 2.35 | |
916 | [29162] string: [PATCH] string.h syntactic error: | |
917 | include/bits/string_fortified.h:110: error: expected ',' or ';' | |
918 | before '__fortified_attr_access' | |
919 | [29165] libc: [Regression] broken argv adjustment | |
920 | [29187] dynamic-link: [regression] broken argv adjustment for nios2 | |
921 | [29193] math: sincos produces a different output than sin/cos | |
922 | [29197] string: __strncpy_power9() uses uninitialised register vs18 | |
923 | value for filling after \0 | |
924 | [29203] libc: daemon is not y2038 aware | |
925 | [29204] libc: getusershell is not 2038 aware | |
926 | [29207] libc: posix_fallocate fallback implementation is not y2038 | |
927 | aware | |
928 | [29208] libc: fpathconf(_PC_ASYNC_IO) is not y2038 aware | |
929 | [29209] libc: isfdtype is not y2038 aware | |
930 | [29210] network: ruserpass is not y2038 aware | |
931 | [29211] libc: __open_catalog is not y2038 aware | |
932 | [29213] libc: gconv_parseconfdir is not y2038 aware | |
933 | [29214] nptl: pthread_setcanceltype fails to set type | |
934 | [29225] network: Mistyped define statement in socket/sys/socket.h in | |
935 | line 184 | |
936 | [29274] nptl: __read_chk is not a cancellation point | |
937 | [29279] libc: undefined reference to `mbstowcs_chk' after | |
938 | 464d189b9622932a75302290625de84931656ec0 | |
939 | [29304] libc: mq_timedreceive does not handle 64 bit syscall return | |
940 | correct for !__ASSUME_TIME64_SYSCALLS | |
941 | [29403] libc: st_atim, st_mtim, st_ctim stat struct members are | |
942 | missing on microblaze with largefile | |
943 | \f | |
944 | Version 2.35 | |
945 | ||
946 | Major new features: | |
947 | ||
948 | * Unicode 14.0.0 Support: Character encoding, character type info, and | |
949 | transliteration tables are all updated to Unicode 14.0.0, using | |
950 | generator scripts contributed by Mike FABIAN (Red Hat). | |
951 | ||
952 | * Bump r_version in the debugger interface to 2 and add a new field, | |
953 | r_next, support multiple namespaces. | |
954 | ||
955 | * Support for the C.UTF-8 locale has been added to glibc. The locale | |
956 | supports full code-point sorting for all valid Unicode code points. A | |
957 | limitation in the framework for fnmatch, regexec, and regcomp requires | |
958 | a compromise to save space and only ASCII-based range expressions are | |
959 | supported for now (see bug 28255). The full size of the locale is | |
960 | only ~400KiB, with 346KiB coming from LC_CTYPE information for | |
961 | Unicode. This locale harmonizes downstream C.UTF-8 already shipping | |
962 | in various downstream distributions. The locale is not built into | |
963 | glibc, and must be installed. | |
964 | ||
965 | * <math.h> functions that round their results to a narrower type, and | |
966 | corresponding <tgmath.h> macros, are added from TS 18661-1:2014, TS | |
967 | 18661-3:2015 and draft ISO C2X: | |
968 | ||
969 | - fsqrt, fsqrtl, dsqrtl and corresponding fMsqrtfN, fMsqrtfNx, | |
970 | fMxsqrtfN and fMxsqrtfNx functions. | |
971 | ||
972 | - ffma, ffmal, dfmal and corresponding fMfmafN, fMfmafNx, fMxfmafN and | |
973 | fMxfmafNx functions. | |
974 | ||
975 | * <math.h> functions for floating-point maximum and minimum, | |
976 | corresponding to new operations in IEEE 754-2019, and corresponding | |
977 | <tgmath.h> macros, are added from draft ISO C2X: fmaximum, | |
978 | fmaximum_num, fmaximum_mag, fmaximum_mag_num, fminimum, fminimum_num, | |
979 | fminimum_mag, fminimum_mag_num and corresponding functions for float, | |
980 | long double, _FloatN and _FloatNx. | |
981 | ||
982 | * <math.h> macros for single-precision float constants are added as a | |
983 | GNU extension: M_Ef, M_LOG2Ef, M_LOG10Ef, M_LN2f, M_LN10f, M_PIf, | |
984 | M_PI_2f, M_PI_4f, M_1_PIf, M_2_PIf, M_2_SQRTPIf, M_SQRT2f and | |
985 | M_SQRT1_2f. | |
986 | ||
987 | * The __STDC_IEC_60559_BFP__ and __STDC_IEC_60559_COMPLEX__ macros are | |
988 | predefined as specified in TS 18661-1:2014. | |
989 | ||
990 | * The exp10 functions in <math.h> now have a corresponding type-generic | |
991 | macro in <tgmath.h>. | |
992 | ||
993 | * The ISO C2X macro _PRINTF_NAN_LEN_MAX has been added to <stdio.h>. | |
994 | ||
995 | * printf-family functions now support the %b format for output of | |
996 | integers in binary, as specified in draft ISO C2X, and the %B variant | |
997 | of that format recommended by draft ISO C2X. | |
998 | ||
999 | * A new DSO sorting algorithm has been added in the dynamic linker that uses | |
1000 | topological sorting by depth-first search (DFS), solving performance issues | |
1001 | of the existing sorting algorithm when encountering particular circular | |
1002 | object dependency cases. | |
1003 | ||
1004 | * A new tunable, glibc.rtld.dynamic_sort, can be used to select between | |
1005 | the two DSO sorting algorithms. The default setting of '2' uses the | |
1006 | new DFS-based algorithm. The setting '1' switches to the old | |
1007 | algorithm used in glibc 2.33 and earlier. | |
1008 | ||
1009 | * ABI support for a new function '__memcmpeq'. '__memcmpeq' is meant | |
1010 | to be used by compilers for optimizing usage of 'memcmp' when its | |
1011 | return value is only used for its boolean status. | |
1012 | ||
1013 | * Support for automatically registering threads with the Linux rseq | |
1014 | system call has been added. This system call is implemented starting | |
1015 | from Linux 4.18. The Restartable Sequences ABI accelerates user-space | |
1016 | operations on per-cpu data. It allows user-space to perform updates | |
1017 | on per-cpu data without requiring heavy-weight atomic operations. | |
1018 | Automatically registering threads allows all libraries, including | |
1019 | libc, to make immediate use of the rseq support by using the | |
1020 | documented ABI, via the __rseq_flags, __rseq_offset, and __rseq_size | |
1021 | variables. The GNU C Library manual has details on integration of | |
1022 | Restartable Sequences. | |
1023 | ||
1024 | * A symbolic link to the dynamic linker is now installed under | |
1025 | /usr/bin/ld.so (or more precisely, '${bindir}/ld.so'). | |
1026 | ||
1027 | * All programs and the testsuite in glibc are now built as position independent | |
1028 | executables (PIE) by default on toolchains and architectures that support it. | |
1029 | Further, if the toolchain and architecture supports it, even static programs | |
1030 | are built as PIE and the resultant glibc can be used to build static PIE | |
1031 | executables. A new option --disable-default-pie has been added to disable | |
1032 | this behavior and get a non-PIE build. This option replaces | |
1033 | --enable-static-pie, which no longer has any effect on the build | |
1034 | configuration. | |
1035 | ||
1036 | * On Linux, a new tunable, glibc.malloc.hugetlb, can be used to | |
1037 | either make malloc issue madvise plus MADV_HUGEPAGE on mmap and sbrk | |
1038 | or to use huge pages directly with mmap calls with the MAP_HUGETLB | |
1039 | flags). The former can improve performance when Transparent Huge Pages | |
1040 | is set to 'madvise' mode while the latter uses the system reserved | |
1041 | huge pages. | |
1042 | ||
1043 | * The printf family of functions now handles the flagged %#m conversion | |
1044 | specifier, printing errno as an error constant (similar to strerrorname_np). | |
1045 | ||
1046 | * The function _dl_find_object has been added. In-process unwinders | |
1047 | can use it to efficiently locate unwinding information for a code | |
1048 | address. | |
1049 | ||
1050 | * Support for OpenRISC running on Linux has been added. This port requires | |
1051 | as least binutils 2.35, GCC 11, and Linux 5.4. Currently only soft-float | |
1052 | ABI is supported: | |
1053 | ||
1054 | - or1k-linux-gnu | |
1055 | ||
1056 | The OpenRISC ABI is 32-bit big-endian and uses 64-bit time (y2038 safe) and | |
1057 | 64-bit file offsets (LFS default). | |
1058 | ||
1059 | * A new configure option, --with-rtld-early-cflags, can be used to | |
1060 | specify additional compiler flags for building the early startup code | |
1061 | of the dynamic linker. On targets which have CPU compatibility | |
1062 | checks, this can help to ensure that proper diagnostics are printed if | |
1063 | the dynamic loader runs on an incompatible CPU. | |
1064 | ||
1065 | * On Linux, the epoll_pwait2 function has been added. It is similar to | |
1066 | epoll_wait with the difference the timeout has nanoseconds resolution. | |
1067 | ||
1068 | * The function posix_spawn_file_actions_addtcsetpgrp_np has been added, | |
1069 | enabling posix_spawn and posix_spawnp to set the controlling terminal in | |
1070 | the new process in a race free manner. This function is a GNU extension. | |
1071 | ||
1072 | * Source fortification (_FORTIFY_SOURCE) level 3 is now available for | |
1073 | applications compiling with glibc and gcc 12 and later. Level 3 leverages | |
1074 | the __builtin_dynamic_object_size function to deliver additional | |
1075 | fortification balanced against additional runtime cost (checking non-constant | |
1076 | bounds). | |
1077 | ||
1078 | * The audit libraries will avoid unnecessary slowdown if it is not required | |
1079 | PLT tracking (by not implementing the la_pltenter or la_pltexit callbacks). | |
1080 | ||
1081 | Deprecated and removed features, and other changes affecting compatibility: | |
1082 | ||
1083 | * On x86-64, the LD_PREFER_MAP_32BIT_EXEC environment variable support | |
1084 | has been removed since the first PT_LOAD segment is no longer executable | |
1085 | due to defaulting to -z separate-code. | |
1086 | ||
1087 | * The r_version update in the debugger interface makes the glibc binary | |
1088 | incompatible with GDB binaries built without the following commits: | |
1089 | ||
1090 | c0154a4a21a gdb: Don't assume r_ldsomap when r_version > 1 on Linux | |
1091 | 4eb629d50d4 gdbserver: Check r_version < 1 for Linux debugger interface | |
1092 | ||
1093 | when audit modules or dlmopen are used. | |
1094 | ||
1095 | * Intel MPX support (lazy PLT, ld.so profile, and LD_AUDIT) has been removed. | |
1096 | ||
1097 | * The --enable-static-pie option is no longer available. The glibc build | |
1098 | configuration script now automatically detects static-pie support in the | |
1099 | toolchain and architecture and enables it if available. | |
1100 | ||
1101 | * The catchsegv script and associated libSegFault.so shared object have | |
1102 | been removed. There are widely-deployed out-of-process alternatives for | |
1103 | catching coredumps and backtraces. | |
1104 | ||
1105 | * Support for prelink will be removed in the next release; this includes | |
1106 | removal of the LD_TRACE_PRELINKING, and LD_USE_LOAD_BIAS, environment | |
1107 | variables and their functionality in the dynamic loader. | |
1108 | ||
1109 | * The LD_TRACE_PRELINKING environment variable has been removed. Similar | |
1110 | functionality to obtain the program mapping address can be achieved by | |
1111 | using LD_TRACE_LOADED_OBJECTS to value of 2. | |
1112 | ||
1113 | * The LD_USE_LOAD_BIAS has been removed. The variable was mainly used to | |
1114 | support prelink PIE binaries. | |
1115 | ||
1116 | Changes to build and runtime requirements: | |
1117 | ||
1118 | * The audit module interface version LAV_CURRENT is increased to enable | |
1119 | proper bind-now support. The loader now advertises via the la_symbind | |
1120 | flags that PLT trace is not possible. New audit modules require the | |
1121 | new dynamic loader supporting the latest LAV_CURRENT version. Old audit | |
1122 | modules are still loaded for all targets except aarch64. | |
1123 | ||
1124 | * The audit interface on aarch64 is extended to support both the indirect | |
1125 | result location register (x8) and NEON Q register. Old audit modules are | |
1126 | rejected by the loader. Audit modules must be rebuilt to use the newer | |
1127 | structure sizes and the latest module interface version for LAV_CURRENT. | |
1128 | ||
1129 | Security related changes: | |
1130 | ||
1131 | CVE-2022-23219: Passing an overlong file name to the clnt_create | |
1132 | legacy function could result in a stack-based buffer overflow when | |
1133 | using the "unix" protocol. Reported by Martin Sebor. | |
1134 | ||
1135 | CVE-2022-23218: Passing an overlong file name to the svcunix_create | |
1136 | legacy function could result in a stack-based buffer overflow. | |
1137 | ||
1138 | CVE-2021-3998: Passing a path longer than PATH_MAX to the realpath | |
1139 | function could result in a memory leak and potential access of | |
1140 | uninitialized memory. Reported by Qualys. | |
1141 | ||
1142 | CVE-2021-3999: Passing a buffer of size exactly 1 byte to the getcwd | |
1143 | function may result in an off-by-one buffer underflow and overflow | |
1144 | when the current working directory is longer than PATH_MAX and also | |
1145 | corresponds to the / directory through an unprivileged mount | |
1146 | namespace. Reported by Qualys. | |
1147 | ||
1148 | The following bugs are resolved with this release: | |
1149 | ||
1150 | [12889] nptl: Race condition in pthread_kill | |
1151 | [14232] nptl: tst-cancel7 and tst-cancelx7 race condition | |
1152 | [14913] libc: [mips] Clean up MIPS 64-bit register-dump.h output | |
1153 | [15310] dynamic-link: _dl_sort_fini is O(n^3) causing slow exit when | |
1154 | many dsos | |
1155 | [15333] libc: Use 64-bit stat functions in installed programs | |
1156 | [15533] dynamic-link: LD_AUDIT introduces an avoidable performance | |
1157 | degradation | |
1158 | [15971] dynamic-link: No interface for debugger access to libraries | |
1159 | loaded with dlmopen | |
1160 | [17318] locale: [RFE] Provide a C.UTF-8 locale by default | |
1161 | [17645] dynamic-link: RFE: Improve performance of dynamic loader for | |
1162 | deeply nested DSO dependencies. | |
1163 | [19193] nptl: pthread_kill, pthread_cancel return ESRCH for a thread | |
1164 | ID whose lifetime has not ended | |
1165 | [22542] network: buffer overflow in sunrpc clnt_create | |
1166 | (CVE-2022-23219) | |
1167 | [22716] malloc: [PATCH] mtrace.pl: use TRACE_PRELINKING instead of | |
1168 | TRACE_LOADED_OBJECTS | |
1169 | [25947] malloc: memory leak in muntrace | |
1170 | [26045] math: fmaxf(inf, nan) does not always work | |
1171 | [26108] math: exp10() has problems with <tgmath.h> | |
1172 | [26779] build: benign use after realloc at localealias.c:329 | |
1173 | [27609] dynamic-link: [2.32/2.33/2.34 Regression] In elf/dl-open.c | |
1174 | (_dl_open) we might use __LM_ID_CALLER to index GL(dl_ns)[] | |
1175 | [27945] build: build-many-glibcs.py doesn't configure GCC with | |
1176 | --enable-initfini-array | |
1177 | [27991] build: x86: sysdeps/x86/configure.ac breaks when | |
1178 | libc_cv_include_x86_isa_level is loaded from cache | |
1179 | [28036] nptl: Incorrect types for pthread_mutexattr_set/getrobust_np | |
1180 | in __REDIRECT_NTH macro | |
1181 | [28061] dynamic-link: A failing dlmopen called by an auditor crashed | |
1182 | [28062] dynamic-link: Suppress audit calls when a (new) namespace is | |
1183 | empty | |
1184 | [28126] libc: nftw aborts for paths longer than PATH_MAX | |
1185 | [28129] dynamic-link: Unnecessary check DT_DEBUG in ld.so | |
1186 | [28153] libc: [test] gmon/tst-gmon-gprof* may have a f3 line when | |
1187 | built with ld.lld | |
1188 | [28182] libc: _TIME_BITS=64 in C++ has issues with fcntl, ioctl, prctl | |
1189 | [28185] math: Inaccurate j0f function (again) | |
1190 | [28199] locale: iconvconfig prefix flag behaves differently in glibc | |
1191 | 2.34 | |
1192 | [28203] dynamic-link: aarch64: elf_machine_{load_address,dynamic} | |
1193 | should drop _GLOBAL_OFFSET_TABLE_[0] in favor of __ehdr_start for | |
1194 | robustness | |
1195 | [28213] librt: NULL pointer dereference in mq_notify (CVE-2021-38604) | |
1196 | [28223] libc: mips: clone does not align stack | |
1197 | [28253] dynamic-link: Missing colon in LD_SHOW_AUXV output after | |
1198 | AT_MINSIGSTKSZ | |
1199 | [28256] malloc: Conditional jump or move depends on uninitialised | |
1200 | value(s) in __GI___tunables_init | |
1201 | [28260] build: io/tst-closefrom, misc/tst-close_range, posix/tst- | |
1202 | spawn5 fail if stray fds are open | |
1203 | [28310] libc: Do not use affinity mask for sysconf | |
1204 | (_SC_NPROCESSORS_CONF) | |
1205 | [28338] time: undefined behavior in __tzfile_compute with oddball TZif | |
1206 | file | |
1207 | [28340] dynamic-link: ld.so crashes while loading a DSO with a read- | |
1208 | only dynamic section | |
1209 | [28349] libc: Segfault for ping -R on qemux86 caused by recvmsg() | |
1210 | [28350] libc: ping receives SIGABRT on lib32-qemux86-64 caused by | |
1211 | recvmsg() | |
1212 | [28353] network: Race condition on __opensock | |
1213 | [28357] dynamic-link: deadlock between pthread_create and ctors | |
1214 | [28358] math: f64xdivf128 and f64xmulf128 spurious underflows | |
1215 | [28361] nptl: Fix for bug 12889 causes setxid deadlock | |
1216 | [28368] build: -Waddress instances in stdio-common/vfprintf-internal.c | |
1217 | [28390] localedata: Update locale data to Unicode 14.0.0 | |
1218 | [28397] math: tgmath.h should not define fmaxmag, fminmag macros for | |
1219 | C2X | |
1220 | [28400] libc: [2.35 Regression] string/test-strncasecmp: cannot set | |
1221 | locale "en_US.UTF-8" | |
1222 | [28407] nptl: pthread_kill assumes that kill (getpid ()) is equivalent | |
1223 | to tgkill (getpid (), gettid()) | |
1224 | [28455] dynamic-link: -Wl,--enable-new-dtags doesn't work | |
1225 | [28457] dynamic-link: Missing reldepmod4.so dependency for | |
1226 | globalmod1.so | |
1227 | [28469] time: linux: struct timex is not correctly set for 32-bit | |
1228 | systems with TIMESIZE=64 | |
1229 | [28470] regex: Buffer read overrun in regular expression searching | |
1230 | [28475] string: Incorrect access attribute on memfrob | |
1231 | [28524] libc: Conversion from ISO-2022-JP-3 with iconv may emit | |
1232 | spurious NUL character on state reset | |
1233 | [28532] libc: powerpc64[le]: CFI for assembly templated syscalls is | |
1234 | incorrect | |
1235 | [28550] dynamic-link: FAIL: tst-dso- | |
1236 | ordering9_112-ecbda(GLIBC_TUNABLES=glibc.rtld.dynamic_sort=1) | |
1237 | execution test | |
1238 | [28554] build: Undefined generate-md5 | |
1239 | [28572] libc: Misaligned accesses in test-memcpy and test-mempcpy on | |
1240 | hppa | |
1241 | [28607] nptl: Masked signals are delivered on thread exit | |
1242 | [28624] libc: openjdk 8/9 assume uni processor and gets stuck due to | |
1243 | lack of cpu counting /proc fallback with glibc 2.34 | |
1244 | [28646] string: [2.35 Regression] mock -r fedora-36-x86_64 | |
1245 | /tmp/java-1.8.0-openjdk-1.8.0.312.b07-2.fc36.src.rpm& fails to build | |
1246 | [28648] dynamic-link: Running ld.so on statically linked binaries | |
1247 | crashes | |
1248 | [28656] dynamic-link: LD_PREFER_MAP_32BIT_EXEC no longer works due to | |
1249 | binutils changes | |
1250 | [28676] dynamic-link: p_align on PT_LOAD segment in DSO isn't honored | |
1251 | [28678] nptl: nptl/tst-create1 hangs sporadically | |
1252 | [28688] dynamic-link: PT_LOAD p_align check is too strict | |
1253 | [28700] nss: "dns [!UNAVAIL=return] files" default for hosts database | |
1254 | is not useful | |
1255 | [28707] time: assert in tzfile.c __tzfile_read striking with truncated | |
1256 | timezones generated by tzcode-2021d and later | |
1257 | [28713] math: GCC 12 miscompiles libm | |
1258 | [28732] dynamic-link: FAIL: elf/tst-dl_find_object | |
1259 | [28738] build: LIBC_LINKER_FEATURE doesn't work on linker -z option | |
1260 | [28745] dynamic-link: _dl_find_object miscompilation on powerpc64le | |
1261 | [28746] libc: _FORTIFY_SOURCE does not work for stpcpy | |
1262 | [28749] libc: Inconsistency detected by ld.so: rtld.c: 1632: dl_main: | |
1263 | Assertion `GL(dl_rtld_map).l_libname' failed! | |
1264 | [28755] string: overflow bug in wcsncmp_avx2 and wcsncmp_evex | |
1265 | [28757] nptl: GDB printer tests failed with new GDB | |
1266 | [28765] math: x86_64 libmvec atan2 accuracy | |
1267 | [28766] manual: Document libmvec accuracy | |
1268 | [28768] network: Buffer overflow in svcunix_create with long pathnames | |
1269 | (CVE-2022-23218) | |
1270 | [28769] libc: Off-by-one buffer overflow/underflow in getcwd() | |
1271 | (CVE-2021-3999) | |
1272 | [28770] libc: Unexpected return value from realpath() for too long | |
1273 | results (CVE-2021-3998) | |
1274 | [28771] libc: %ebx optimization macros are incompatible with .altmacro | |
1275 | in Systemtap probes | |
1276 | [28780] build: --disable-default-pie doesn't work on static programs | |
1277 | [28782] libc: x86-64 ISA level for glibc itself is always | |
1278 | x86-64-baseline | |
1279 | [28792] glob: possible wrong behaviour with patterns with double [ | |
1280 | with no closing ] | |
1281 | [28837] libc: FAIL: socket/tst-socket-timestamp-compat | |
1282 | [28847] locale: Empty mon_decimal_point in LC_MONETARY results in non- | |
1283 | empty mon_decimal_point_wc | |
1284 | ||
1285 | \f | |
1286 | Version 2.34 | |
1287 | ||
1288 | Major new features: | |
1289 | ||
1290 | * In order to support smoother in-place-upgrades and to simplify | |
1291 | the implementation of the runtime all functionality formerly | |
1292 | implemented in the libraries libpthread, libdl, libutil, libanl has | |
1293 | been integrated into libc. New applications do not need to link with | |
1294 | -lpthread, -ldl, -lutil, -lanl anymore. For backwards compatibility, | |
1295 | empty static archives libpthread.a, libdl.a, libutil.a, libanl.a are | |
1296 | provided, so that the linker options keep working. Applications which | |
1297 | have been linked against glibc 2.33 or earlier continue to load the | |
1298 | corresponding shared objects (which are now empty). The integration | |
1299 | of those libraries into libc means that additional symbols become | |
1300 | available by default. This can cause applications that contain weak | |
1301 | references to take unexpected code paths that would only have been | |
1302 | used in previous glibc versions when e.g. preloading libpthread.so.0, | |
1303 | potentially exposing application bugs. | |
1304 | ||
1305 | * When _DYNAMIC_STACK_SIZE_SOURCE or _GNU_SOURCE are defined, | |
1306 | PTHREAD_STACK_MIN is no longer constant and is redefined to | |
1307 | sysconf(_SC_THREAD_STACK_MIN). This supports dynamic sized register | |
1308 | sets for modern architectural features like Arm SVE. | |
1309 | ||
1310 | * Add _SC_MINSIGSTKSZ and _SC_SIGSTKSZ. When _DYNAMIC_STACK_SIZE_SOURCE | |
1311 | or _GNU_SOURCE are defined, MINSIGSTKSZ and SIGSTKSZ are no longer | |
1312 | constant on Linux. MINSIGSTKSZ is redefined to sysconf(_SC_MINSIGSTKSZ) | |
1313 | and SIGSTKSZ is redefined to sysconf (_SC_SIGSTKSZ). This supports | |
1314 | dynamic sized register sets for modern architectural features like | |
1315 | Arm SVE. | |
1316 | ||
1317 | * The dynamic linker implements the --list-diagnostics option, printing | |
1318 | a dump of information related to IFUNC resolver operation and | |
1319 | glibc-hwcaps subdirectory selection. | |
1320 | ||
1321 | * On Linux, the function execveat has been added. It operates similar to | |
1322 | execve and it is is already used to implement fexecve without requiring | |
1323 | /proc to be mounted. However, different than fexecve, if the syscall is not | |
1324 | supported by the kernel an error is returned instead of trying a fallback. | |
1325 | ||
1326 | * The ISO C2X function timespec_getres has been added. | |
1327 | ||
1328 | * The feature test macro __STDC_WANT_IEC_60559_EXT__, from draft ISO | |
1329 | C2X, is supported to enable declarations of functions defined in Annex F | |
1330 | of C2X. Those declarations are also enabled when | |
1331 | __STDC_WANT_IEC_60559_BFP_EXT__, as specified in TS 18661-1, is | |
1332 | defined, and when _GNU_SOURCE is defined. | |
1333 | ||
1334 | * On powerpc64*, glibc can now be compiled without scv support using the | |
1335 | --disable-scv configure option. | |
1336 | ||
1337 | * Add support for 64-bit time_t on configurations like x86 where time_t | |
1338 | is traditionally 32-bit. Although time_t still defaults to 32-bit on | |
1339 | these configurations, this default may change in future versions. | |
1340 | This is enabled with the _TIME_BITS preprocessor macro set to 64 and is | |
1341 | only supported when LFS (_FILE_OFFSET_BITS=64) is also enabled. It is | |
1342 | only enabled for Linux and the full support requires a minimum kernel | |
1343 | version of 5.1. | |
1344 | ||
1345 | * The main gconv-modules file in glibc now contains only a small set of | |
1346 | essential converter modules and the rest have been moved into a supplementary | |
1347 | configuration file gconv-modules-extra.conf in the gconv-modules.d directory | |
1348 | in the same GCONV_PATH. Similarly, external converter modules directories | |
1349 | may have supplementary configuration files in a gconv-modules.d directory | |
1350 | with names ending with .conf to logically classify the converter modules in | |
1351 | that directory. | |
1352 | ||
1353 | * On Linux, a new tunable, glibc.pthread.stack_cache_size, can be used | |
1354 | to configure the size of the thread stack cache. | |
1355 | ||
1356 | * The function _Fork has been added as an async-signal-safe fork replacement | |
1357 | since Austin Group issue 62 dropped the async-signal-safe requirement for | |
1358 | fork (and it will be included in the future POSIX standard). The new _Fork | |
1359 | function does not run any atfork function neither resets any internal state | |
1360 | or lock (such as the malloc one), and only sets up a minimal state required | |
1361 | to call async-signal-safe functions (such as raise or execve). This function | |
1362 | is currently a GNU extension. | |
1363 | ||
1364 | * On Linux, the close_range function has been added. It allows efficiently | |
1365 | closing a range of file descriptors on recent kernels (version 5.9). | |
1366 | ||
1367 | * The function closefrom has been added. It closes all file descriptors | |
1368 | greater than or equal to a given integer. This function is a GNU extension, | |
1369 | although it is also present in other systems. | |
1370 | ||
1371 | * The posix_spawn_file_actions_addclosefrom_np function has been added, | |
1372 | enabling posix_spawn and posix_spawnp to close all file descriptors greater | |
1373 | than or equal to a given integer. This function is a GNU extension, | |
1374 | although Solaris also provides a similar function. | |
1375 | ||
1376 | * When invoked explicitly, the dynamic linker now uses the kernel to | |
1377 | execute programs that do not have any dynamic dependency (that is, | |
1378 | they are statically linked). This feature is Linux-specific. | |
1379 | ||
1380 | Deprecated and removed features, and other changes affecting compatibility: | |
1381 | ||
1382 | * The function pthread_mutex_consistent_np has been deprecated; programs | |
1383 | should use the equivalent standard function pthread_mutex_consistent | |
1384 | instead. | |
1385 | ||
1386 | * The function pthread_mutexattr_getrobust_np has been deprecated; | |
1387 | programs should use the equivalent standard function | |
1388 | pthread_mutexattr_getrobust instead. | |
1389 | ||
1390 | * The function pthread_mutexattr_setrobust_np has been deprecated; | |
1391 | programs should use the equivalent standard function | |
1392 | pthread_mutexattr_setrobust instead. | |
1393 | ||
1394 | * The function pthread_yield has been deprecated; programs should use | |
1395 | the equivalent standard function sched_yield instead. | |
1396 | ||
1397 | * The function inet_neta declared in <arpa/inet.h> has been deprecated. | |
1398 | ||
1399 | * Various rarely-used functions declared in <resolv.h> and | |
1400 | <arpa/nameser.h> have been deprecated. Applications are encouraged to | |
1401 | use dedicated DNS processing libraries if applicable. For <resolv.h>, | |
1402 | this affects the functions dn_count_labels, fp_nquery, fp_query, | |
1403 | fp_resstat, hostalias, loc_aton, loc_ntoa, p_cdname, p_cdnname, | |
1404 | p_class, p_fqname, p_fqnname, p_option, p_query, p_rcode, p_time, | |
1405 | p_type, putlong, putshort, res_hostalias, res_isourserver, | |
1406 | res_nameinquery, res_queriesmatch, res_randomid, sym_ntop, sym_ntos, | |
1407 | sym_ston. For <arpa/nameser.h>, the functions ns_datetosecs, | |
1408 | ns_format_ttl, ns_makecanon, ns_parse_ttl, ns_samedomain, ns_samename, | |
1409 | ns_sprintrr, ns_sprintrrf, ns_subdomain have been deprecated. | |
1410 | ||
1411 | * Various symbols previously defined in libresolv have been moved to libc | |
1412 | in order to prepare for libresolv moving entirely into libc (see earlier | |
1413 | entry for merging libraries into libc). The symbols __dn_comp, | |
1414 | __dn_expand, __dn_skipname, __res_dnok, __res_hnok, __res_mailok, | |
1415 | __res_mkquery, __res_nmkquery, __res_nquery, __res_nquerydomain, | |
1416 | __res_nsearch, __res_nsend, __res_ownok, __res_query, __res_querydomain, | |
1417 | __res_search, __res_send formerly in libresolv have been renamed and no | |
1418 | longer have a __ prefix. They are now available in libc. | |
1419 | ||
1420 | * The pthread cancellation handler is now installed with SA_RESTART and | |
1421 | pthread_cancel will always send the internal SIGCANCEL on a cancellation | |
1422 | request. It should not be visible to applications since the cancellation | |
1423 | handler should either act upon cancellation (if asynchronous cancellation | |
1424 | is enabled) or ignore the cancellation internal signal. However there are | |
1425 | buggy kernel interfaces (for instance some CIFS versions) that could still | |
1426 | see a spurious EINTR error when cancellation interrupts a blocking syscall. | |
1427 | ||
1428 | * Previously, glibc installed its various shared objects under versioned | |
1429 | file names such as libc-2.33.so. The ABI sonames (e.g., libc.so.6) | |
1430 | were provided as symbolic links. Starting with glibc 2.34, the shared | |
1431 | objects are installed under their ABI sonames directly, without | |
1432 | symbolic links. This increases compatibility with distribution | |
1433 | package managers that delete removed files late during the package | |
1434 | upgrade or downgrade process. | |
1435 | ||
1436 | * The symbols mallwatch and tr_break are now deprecated and no longer used in | |
1437 | mtrace. Similar functionality can be achieved by using conditional | |
1438 | breakpoints within mtrace functions from within gdb. | |
1439 | ||
1440 | * The __morecore and __after_morecore_hook malloc hooks and the default | |
1441 | implementation __default_morecore have been removed from the API. Existing | |
1442 | applications will continue to link against these symbols but the interfaces | |
1443 | no longer have any effect on malloc. | |
1444 | ||
1445 | * Debugging features in malloc such as the MALLOC_CHECK_ environment variable | |
1446 | (or the glibc.malloc.check tunable), mtrace() and mcheck() have now been | |
1447 | disabled by default in the main C library. Users looking to use these | |
1448 | features now need to preload a new debugging DSO libc_malloc_debug.so to get | |
1449 | this functionality back. | |
1450 | ||
1451 | * The deprecated functions malloc_get_state and malloc_set_state have been | |
1452 | moved from the core C library into libc_malloc_debug.so. Legacy applications | |
1453 | that still use these functions will now need to preload libc_malloc_debug.so | |
1454 | in their environment using the LD_PRELOAD environment variable. | |
1455 | ||
1456 | * The deprecated memory allocation hooks __malloc_hook, __realloc_hook, | |
1457 | __memalign_hook and __free_hook are now removed from the API. Compatibility | |
1458 | symbols are present to support legacy programs but new applications can no | |
1459 | longer link to these symbols. These hooks no longer have any effect on glibc | |
1460 | functionality. The malloc debugging DSO libc_malloc_debug.so currently | |
1461 | supports hooks and can be preloaded to get this functionality back for older | |
1462 | programs. However this is a transitional measure and may be removed in a | |
1463 | future release of the GNU C Library. Users may port away from these hooks by | |
1464 | writing and preloading their own malloc interposition library. | |
1465 | ||
1466 | Changes to build and runtime requirements: | |
1467 | ||
1468 | * On Linux, the shm_open, sem_open, and related functions now expect the | |
1469 | file shared memory file system to be mounted at /dev/shm. These functions | |
1470 | no longer search among the system's mount points for a suitable | |
1471 | replacement if /dev/shm is not available. | |
1472 | ||
1473 | Security related changes: | |
1474 | ||
1475 | CVE-2021-27645: The nameserver caching daemon (nscd), when processing | |
1476 | a request for netgroup lookup, may crash due to a double-free, | |
1477 | potentially resulting in degraded service or Denial of Service on the | |
1478 | local system. Reported by Chris Schanzle. | |
1479 | ||
1480 | CVE-2021-33574: The mq_notify function has a potential use-after-free | |
1481 | issue when using a notification type of SIGEV_THREAD and a thread | |
1482 | attribute with a non-default affinity mask. | |
1483 | ||
1484 | CVE-2021-35942: The wordexp function may overflow the positional | |
1485 | parameter number when processing the expansion resulting in a crash. | |
1486 | Reported by Philippe Antoine. | |
1487 | ||
1488 | The following bugs are resolved with this release: | |
1489 | ||
1490 | [4737] libc: fork is not async-signal-safe | |
1491 | [5781] math: Slow dbl-64 sin/cos/sincos for special values | |
1492 | [10353] libc: Methods for deleting all file descriptors greater than | |
1493 | given integer (closefrom) | |
1494 | [14185] glob: fnmatch() fails when '*' wildcard is applied on the file | |
1495 | name containing multi-byte character(s) | |
1496 | [14469] math: Inaccurate j0f function | |
1497 | [14470] math: Inaccurate j1f function | |
1498 | [14471] math: Inaccurate y0f function | |
1499 | [14472] math: Inaccurate y1f function | |
1500 | [14744] nptl: kill -32 $pid or kill -33 $pid on a process cancels a | |
1501 | random thread | |
1502 | [15271] dynamic-link: dlmopen()ed shared library with LM_ID_NEWLM | |
1503 | crashes if it fails dlsym() twice | |
1504 | [15648] nptl: multiple definition of `__lll_lock_wait_private' | |
1505 | [16063] nptl: Provide a pthread_once variant in libc directly | |
1506 | [17144] libc: syslog is not thread-safe if NO_SIGPIPE is not defined | |
1507 | [17145] libc: syslog with LOG_CONS leaks console file descriptor | |
1508 | [17183] manual: description of ENTRY struct in <search.h> in glibc | |
1509 | manual is incorrect | |
1510 | [18435] nptl: pthread_once hangs when init routine throws an exception | |
1511 | [18524] nptl: Missing calloc error checking in | |
1512 | __cxa_thread_atexit_impl | |
1513 | [19329] dynamic-link: dl-tls.c assert failure at concurrent | |
1514 | pthread_create and dlopen | |
1515 | [19366] nptl: returning from a thread should disable cancellation | |
1516 | [19511] nptl: 8MB memory leak in pthread_create in case of failure | |
1517 | when non-root user changes priority | |
1518 | [20802] dynamic-link: getauxval NULL pointer dereference after static | |
1519 | dlopen | |
1520 | [20813] nptl: pthread_exit is inconsistent between libc and libpthread | |
1521 | [22057] malloc: malloc_usable_size is broken with mcheck | |
1522 | [22668] locale: LC_COLLATE: the last character of ellipsis is not | |
1523 | ordered correctly | |
1524 | [23323] libc: [RFE] CSU startup hardening. | |
1525 | [23328] malloc: Remove malloc hooks and ensure related APIs return no | |
1526 | data. | |
1527 | [23462] dynamic-link: Static binary with dynamic string tokens ($LIB, | |
1528 | $PLATFORM, $ORIGIN) crashes | |
1529 | [23489] libc: "gcc -lmcheck" aborts on free when using posix_memalign | |
1530 | [23554] nptl: pthread_getattr_np reports wrong stack size with | |
1531 | MULTI_PAGE_ALIASING | |
1532 | [24106] libc: Bash interpreter in ldd script is taken from host | |
1533 | [24773] dynamic-link: dlerror in an secondary namespace does not use | |
1534 | the right free implementation | |
1535 | [25036] localedata: Update collation order for Swedish | |
1536 | [25383] libc: where_is_shmfs/__shm_directory/SHM_GET_NAME may cause | |
1537 | shm_open to pick wrong directory | |
1538 | [25680] dynamic-link: ifuncmain9picstatic and ifuncmain9picstatic | |
1539 | crash in IFUNC resolver due to stack canary (--enable-stack- | |
1540 | protector=all) | |
1541 | [26874] build: -Warray-bounds in _IO_wdefault_doallocate | |
1542 | [26983] math: [x86_64] x86_64 tgamma has too large ULP error | |
1543 | [27111] dynamic-link: pthread_create and tls access use link_map | |
1544 | objects that may be concurrently freed by dlclose | |
1545 | [27132] malloc: memusagestat is linked to system librt, leading to | |
1546 | undefined symbols on major version upgrade | |
1547 | [27136] dynamic-link: dtv setup at thread creation may leave an entry | |
1548 | uninitialized | |
1549 | [27249] libc: libSegFault.so does not output signal number properly | |
1550 | [27304] nptl: pthread_cond_destroy does not pass private flag to futex | |
1551 | system calls | |
1552 | [27318] dynamic-link: glibc fails to load binaries when built with | |
1553 | -march=sandybridge: CPU ISA level is lower than required | |
1554 | [27343] nss: initgroups() SIGSEGVs when called on a system without | |
1555 | nsswich.conf (in a chroot) | |
1556 | [27346] dynamic-link: x86: PTWRITE feature check is missing | |
1557 | [27389] network: NSS chroot hardening causes regressions in chroot | |
1558 | deployments | |
1559 | [27403] dynamic-link: aarch64: tlsdesc htab is not freed on dlclose | |
1560 | [27444] libc: sysconf reports unsupported option (-1) for | |
1561 | _SC_LEVEL1_ICACHE_LINESIZE on X86 since v2.33 | |
1562 | [27462] nscd: double-free in nscd (CVE-2021-27645) | |
1563 | [27468] malloc: aarch64: realloc crash with heap tagging: FAIL: | |
1564 | malloc/tst-malloc-thread-fail | |
1565 | [27498] dynamic-link: __dl_iterate_phdr lacks unwinding information | |
1566 | [27511] libc: S390 memmove assumes Vector Facility when MIE Facility 3 | |
1567 | is present | |
1568 | [27522] glob: glob, glob64 incorrectly marked as __THROW | |
1569 | [27555] dynamic-link: Static tests fail with --enable-stack- | |
1570 | protector=all | |
1571 | [27559] libc: fstat(AT_FDCWD) succeeds (it shouldn't) and returns | |
1572 | information for the current directory | |
1573 | [27577] dynamic-link: elf/ld.so --help doesn't work | |
1574 | [27605] libc: tunables can't control xsave/xsavec selection in | |
1575 | dl_runtime_resolve_* | |
1576 | [27623] libc: powerpc: Missing registers in sc[v] clobbers list | |
1577 | [27645] libc: [linux] sysconf(_SC_NPROCESSOR...) breaks down on | |
1578 | containers | |
1579 | [27646] dynamic-link: Linker error for non-existing NSS symbols (e.g. | |
1580 | _nss_files_getcanonname_r) from within a dlmopen namespace. | |
1581 | [27648] libc: FAIL: misc/tst-select | |
1582 | [27650] stdio: vfscanf returns too early if a match is longer than | |
1583 | INT_MAX | |
1584 | [27651] libc: Performance regression after updating to 2.33 | |
1585 | [27655] string: Wrong size calculation in string/test-strnlen.c | |
1586 | [27706] libc: select fails to update timeout on error | |
1587 | [27709] libc: arm: FAIL: debug/tst-longjmp_chk2 | |
1588 | [27721] dynamic-link: x86: ld_audit ignores bind now for TLSDESC and | |
1589 | tries resolving them lazily | |
1590 | [27744] nptl: Support different libpthread/ld.so load orders in | |
1591 | libthread_db | |
1592 | [27749] libc: Data race __run_exit_handlers | |
1593 | [27761] libc: getconf: Segmentation fault when passing '-vq' as | |
1594 | argument | |
1595 | [27832] nss: makedb.c:797:7: error: 'writev' specified size 4294967295 | |
1596 | exceeds maximum object size 2147483647 | |
1597 | [27870] malloc: MALLOC_CHECK_ causes realloc(valid_ptr, TOO_LARGE) to | |
1598 | not set ENOMEM | |
1599 | [27872] build: Obsolete configure option --enable-stackguard- | |
1600 | randomization | |
1601 | [27873] build: tst-cpu-features-cpuinfo fail when building on AMD cpu | |
1602 | [27882] localedata: Use U+00AF MACRON in more EBCDIC charsets | |
1603 | [27892] libc: powerpc: scv ABI error handling fails to check | |
1604 | IS_ERR_VALUE | |
1605 | [27896] nptl: mq_notify does not handle separately allocated thread | |
1606 | attributes (CVE-2021-33574) | |
1607 | [27901] libc: TEST_STACK_ALIGN doesn't work | |
1608 | [27902] libc: The x86-64 clone wrapper fails to align child stack | |
1609 | [27914] nptl: Install SIGSETXID handler with SA_ONSTACK | |
1610 | [27939] libc: aarch64: clone does not align the stack | |
1611 | [27968] libc: s390x: clone does not align the stack | |
1612 | [28011] libc: Wild read in wordexp (parse_param) (CVE-2021-35942) | |
1613 | [28024] string: s390(31bit): Wrong result of memchr (MEMCHR_Z900_G5) | |
1614 | with n >= 0x80000000 | |
1615 | [28028] malloc: malloc: tcache shutdown sequence does not work if the | |
1616 | thread never allocated anything | |
1617 | [28033] libc: Need to check RTM_ALWAYS_ABORT for RTM | |
1618 | [28064] string: x86_64:wcslen implementation list has wcsnlen | |
1619 | [28067] libc: FAIL: posix/tst-spawn5 | |
1620 | [28068] malloc: FAIL: malloc/tst-mallocalign1-mcheck | |
1621 | [28071] time: clock_gettime, gettimeofday, time lost vDSO acceleration | |
1622 | on older kernels | |
1623 | [28075] nis: Out-of-bounds static buffer read in nis_local_domain | |
1624 | [28089] build: tst-tls20 fails when linker defaults to --as-needed | |
1625 | [28090] build: elf/tst-cpu-features-cpuinfo-static fails on certain | |
1626 | AMD64 cpus | |
1627 | [28091] network: ns_name_skip may return 0 for domain names without | |
1628 | terminator | |
1629 | ||
1630 | \f | |
1631 | Version 2.33 | |
1632 | ||
1633 | Major new features: | |
1634 | ||
1635 | * The dynamic linker accepts the --list-tunables argument which prints | |
1636 | all the supported tunables. This option is disable if glibc is | |
1637 | configured with tunables disabled (--enable-tunables=no). | |
1638 | ||
1639 | * The dynamic linker accepts the --argv0 argument and provides opportunity | |
1640 | to change argv[0] string. | |
1641 | ||
1642 | * The dynamic linker loads optimized implementations of shared objects | |
1643 | from subdirectories under the glibc-hwcaps directory on the library | |
1644 | search path if the system's capabilities meet the requirements for | |
1645 | that subdirectory. Initially supported subdirectories include | |
1646 | "power9" and "power10" for the powerpc64le-linux-gnu architecture, | |
1647 | "z13", "z14", "z15" for s390x-linux-gnu, and "x86-64-v2", "x86-64-v3", | |
1648 | "x86-64-v4" for x86_64-linux-gnu. In the x86_64-linux-gnu case, the | |
1649 | subdirectory names correspond to the vendor-independent x86-64 | |
1650 | microarchitecture levels defined in the x86-64 psABI supplement. | |
1651 | ||
1652 | * The new --help option of the dynamic linker provides usage and | |
1653 | information and library search path diagnostics. | |
1654 | ||
1655 | * The mallinfo2 function is added to report statistics as per mallinfo, | |
1656 | but with larger field widths to accurately report values that are | |
1657 | larger than fit in an integer. | |
1658 | ||
1659 | * Add <sys/platform/x86.h> to provide query macros for x86 CPU features. | |
1660 | ||
1661 | * Support for the RISC-V ISA running on Linux has been expanded to run on | |
1662 | 32-bit hardware. This is supported for the following ISA and ABI pairs: | |
1663 | ||
1664 | - rv32imac ilp32 | |
1665 | - rv32imafdc ilp32 | |
1666 | - rv32imafdc ilp32d | |
1667 | ||
1668 | The 32-bit RISC-V port requires at least Linux 5.4, GCC 7.1 and binutils | |
1669 | 2.28. | |
1670 | ||
1671 | * A new fortification level _FORTIFY_SOURCE=3 is available. At this level, | |
1672 | glibc may use additional checks that may have an additional performance | |
1673 | overhead. At present these checks are available only on LLVM 9 and later. | |
1674 | The latest GCC available at this time (10.2) does not support this level of | |
1675 | fortification. | |
1676 | ||
1677 | Deprecated and removed features, and other changes affecting compatibility: | |
1678 | ||
1679 | * The mallinfo function is marked deprecated. Callers should call | |
1680 | mallinfo2 instead. | |
1681 | ||
1682 | * When dlopen is used in statically linked programs, alternative library | |
1683 | implementations from HWCAP subdirectories are no longer loaded. | |
1684 | Instead, the default implementation is used. | |
1685 | ||
1686 | * The deprecated <sys/vtimes.h> header and the function vtimes have been | |
1687 | removed. To support old binaries, the vtimes function continues to exist | |
1688 | as a compatibility symbol. Applications should use the getrlimit or | |
1689 | prlimit. | |
1690 | ||
1691 | * Following a change in the tzdata 2018a release upstream, the zdump | |
1692 | program is now installed in the /usr/bin subdirectory. Previously, | |
1693 | the /usr/sbin subdirectory was used. | |
1694 | ||
1695 | * On s390(x), the type float_t is now derived from the macro | |
1696 | __FLT_EVAL_METHOD__ that is defined by the compiler, instead of being | |
1697 | hardcoded to double. This does not affect the ABI of any libraries | |
1698 | that are part of the GNU C Library, but may affect the ABI of other | |
1699 | libraries that use this type in their interfaces. The new definition | |
1700 | improves consistency with compiler behavior in many scenarios. | |
1701 | ||
1702 | * A future version of glibc will stop loading shared objects from the | |
1703 | "tls" subdirectories on the library search path, the subdirectory that | |
1704 | corresponds to the AT_PLATFORM system name, and also stop employing | |
1705 | the legacy AT_HWCAP search mechanism. Applications should switch to | |
1706 | the new glibc-hwcaps mechanism instead; if they do not do that, only | |
1707 | the baseline version (directly from the search path directory) will be | |
1708 | loaded. | |
1709 | ||
1710 | Changes to build and runtime requirements: | |
1711 | ||
1712 | * On Linux, the system administrator needs to configure /dev/pts with | |
1713 | the intended access modes for pseudo-terminals. glibc no longer | |
1714 | attempts to adjust permissions of terminal devices. The previous glibc | |
1715 | defaults ("tty" group, user read/write and group write) already | |
1716 | corresponded to what most systems used, so that grantpt did not | |
1717 | perform any adjustments. | |
1718 | ||
1719 | * On Linux, the posix_openpt and getpt functions no longer attempt to | |
1720 | use legacy (BSD) pseudo-terminals and assume that if /dev/ptmx exists | |
1721 | (and pseudo-terminals are supported), a devpts file system is mounted | |
1722 | on /dev/pts. Current systems already meet these requirements. | |
1723 | ||
1724 | * s390x requires GCC 7.1 or newer. See gcc Bug 98269. | |
1725 | ||
1726 | Security related changes: | |
1727 | ||
1728 | CVE-2021-3326: An assertion failure during conversion from the | |
1729 | ISO-20220-JP-3 character set using the iconv function has been fixed. | |
1730 | This assertion was triggered by certain valid inputs in which the | |
1731 | converted output contains a combined sequence of two wide characters | |
1732 | crossing a buffer boundary. Reported by Tavis Ormandy. | |
1733 | ||
1734 | CVE-2020-27618: An infinite loop has been fixed in the iconv program when | |
1735 | invoked with input containing redundant shift sequences in the IBM1364, | |
1736 | IBM1371, IBM1388, IBM1390, or IBM1399 character sets. | |
1737 | ||
1738 | CVE-2020-29562: An assertion failure has been fixed in the iconv function | |
1739 | when invoked with UCS4 input containing an invalid character. | |
1740 | ||
1741 | CVE-2019-25013: A buffer overflow has been fixed in the iconv function when | |
1742 | invoked with EUC-KR input containing invalid multibyte input sequences. | |
1743 | ||
1744 | The following bugs are resolved with this release: | |
1745 | ||
1746 | [10635] libc: realpath portability patches | |
1747 | [16124] dynamic-link: ld.so should allow to change argv[0] | |
1748 | [17924] malloc: 'free' should not set errno | |
1749 | [18683] libc: Linux faccessat implementation can incorrectly ignore | |
1750 | AT_EACCESS | |
1751 | [22899] libc: Use 64-bit readdir() in generic POSIX getcwd() | |
1752 | [23091] hurd: missing waitid support | |
1753 | [23249] libc: Epyc and other current AMD CPUs do not select the | |
1754 | "haswell" platform subdirectory | |
1755 | [24080] dynamic-link: Definition of "haswell" platform is inconsistent | |
1756 | with GCC | |
1757 | [24202] libc: m68k setjmp() saves incorrect 'a5' register in --enable- | |
1758 | stack-protector=all | |
1759 | [24941] libc: Make grantpt usable after multi-threaded fork in more | |
1760 | cases | |
1761 | [24970] libc: realpath mishandles EOVERFLOW; stat not needed anyway | |
1762 | [24973] locale: iconv encounters segmentation fault when converting | |
1763 | 0x00 0xfe in EUC-KR to UTF-8 (CVE-2019-25013) | |
1764 | [25399] string: undefined reference to `__warn_memset_zero_len' when | |
1765 | changing gnuc version | |
1766 | [25859] libc: glibc parser for /sys/devices/system/cpu/online is | |
1767 | incorrect | |
1768 | [25938] dynamic-link: ld.so.cache should store meaning of hwcap mask | |
1769 | bits | |
1770 | [25971] libc: s390 bits/hwcap.h out of sync with kernel | |
1771 | [26053] libc: unlockpt fails with ENOTTY for non-ptmx descriptors | |
1772 | [26100] libc: Race in syslog(3) with regards to tag printing. | |
1773 | [26124] libc: Export <cpu-features.h> | |
1774 | [26130] nscd: Inconsistent nscd cache during pruning | |
1775 | [26203] libc: GLRO(dl_x86_cpu_features) may not be initialized | |
1776 | [26224] locale: iconv hangs when converting some invalid inputs from | |
1777 | several IBM character sets (CVE-2020-27618) | |
1778 | [26341] libc: realpath cyclically call __alloca(path_max) to consume | |
1779 | too much stack space | |
1780 | [26343] manual: invalid documented return type for strerrorname_np(), | |
1781 | strerrordesc_np(), sigdescr_np(), sigabbrev_np() | |
1782 | [26376] libc: Namespace violation in stdio.h and sys/stat.h if build | |
1783 | with optimization. | |
1784 | [26383] locale: bind_textdomain_codeset doesn't accept //TRANSLIT | |
1785 | anymore | |
1786 | [26394] time: [2.33 Regression] FAIL: nptl/tst-join14 | |
1787 | [26534] math: libm.so 2.32 SIGILL in pow() due to FMA4 instruction on | |
1788 | non-FMA4 system | |
1789 | [26552] dynamic-link: CPU_FEATURE_USABLE_P should be more conservative | |
1790 | [26553] libc: mtx_init allows type set to "mtx_recursive" only | |
1791 | [26555] string: strerrorname_np does not return the documented value | |
1792 | [26592] libc: pointer arithmetic overflows in realpath | |
1793 | [26600] network: Transaction ID collisions cause slow DNS lookups in | |
1794 | getaddrinfo | |
1795 | [26606] libc: [2.33 Regression] pselect is broken on x32 | |
1796 | [26615] libc: powerpc: libc segfaults when LD_PRELOADed with libgcc | |
1797 | [26620] glob: fnmatch with collating symbols results in segmentation | |
1798 | fault | |
1799 | [26625] libc: [2.33 Regression] CET is disabled | |
1800 | [26636] libc: 32-bit shmctl(IPC_INFO) crashes when shminfo struct is | |
1801 | at the end of a memory mapping | |
1802 | [26637] libc: semctl SEM_STAT_ANY fails to pass the buffer specified | |
1803 | by the caller to the kernel | |
1804 | [26639] libc: msgctl IPC_INFO and MSG_INFO return garbage | |
1805 | [26647] build: [-Werror=array-parameter=] due to different | |
1806 | declarations for __sigsetjmp | |
1807 | [26648] libc: mkstemp is likely to fail on systems with non-stricly- | |
1808 | monotonic clocks | |
1809 | [26649] stdio: printf should handle non-normal x86 long double numbers | |
1810 | gracefully (CVE-2020-29573) | |
1811 | [26686] build: -Warray-parameter instances building with GCC 11 | |
1812 | [26687] build: -Warray-bounds instances building with GCC 11 | |
1813 | [26690] stdio: Aliasing violation in __vfscanf_internal | |
1814 | [26691] nptl: Use a minimum guard size of 64 KiB on aarch64 | |
1815 | [26726] build: GCC warning calling new_composite_name with an array of | |
1816 | one element | |
1817 | [26736] libc: FAIL: misc/tst-sysvshm-linux | |
1818 | [26737] libc: Random FAIL: rt/tst-shm | |
1819 | [26791] libc: Missing O_CLOEXEC in sysconf.c | |
1820 | [26798] dynamic-link: aarch64: variant PCS symbols may be incorrectly | |
1821 | lazy bound | |
1822 | [26801] nptl: pthread_mutex_clocklock with CLOCK_MONOTONIC can fail on | |
1823 | PI mutexes | |
1824 | [26818] string: aarch64: string tests may run ifunc variants that are | |
1825 | not safe | |
1826 | [26821] libc: Memory leak test failures on Fedora 33 | |
1827 | [26824] libc: FAIL: elf/tst-cpu-features-supports with recent trunk: | |
1828 | FSGSBASE/LM/RDRAND check failure | |
1829 | [26833] time: adjtime() with delta == NULL segfaults on armv7 32bit | |
1830 | platform | |
1831 | [26853] libc: aarch64: Missing unwind information in statically linked | |
1832 | startup code | |
1833 | [26923] locale: Assertion failure in iconv when converting invalid | |
1834 | UCS4 (CVE-2020-29562) | |
1835 | [26926] dynamic-link: aarch64: library dependencies are not bti | |
1836 | protected | |
1837 | [26932] libc: sh: Multiple floating point functions defined as stubs | |
1838 | only since 2.31 | |
1839 | [26964] nptl: pthread_mutex_timedlock returning EAGAIN after futex is | |
1840 | locked | |
1841 | [26988] dynamic-link: aarch64: BTI mprotect address is not page | |
1842 | aligned | |
1843 | [27002] build: libc_freeres_fn build failure with GCC 11 | |
1844 | [27004] dynamic-link: ld.so is miscompiled by GCC 11 | |
1845 | [27008] dynamic-link: ld.so.cache should have endianness markup | |
1846 | [27042] libc: [alpha] anonymous union in struct stat confuses | |
1847 | detection logic | |
1848 | [27053] libc: Conformance regression in system(3) (and probably also | |
1849 | pclose(3)) | |
1850 | [27072] dynamic-link: static pie ifunc resolvers run before hwcap is | |
1851 | setup | |
1852 | [27077] network: Do not reload /etc/nsswitch.conf from chroot | |
1853 | [27083] libc: Unsafe unbounded alloca in addmntent | |
1854 | [27104] dynamic-link: The COMMON_CPUID_INDEX_MAX handshake does not | |
1855 | work | |
1856 | [27130] string: "rep movsb" performance issue | |
1857 | [27150] libc: alpha: wait4() is unavailable in static linking | |
1858 | [27177] dynamic-link: | |
1859 | GLIBC_TUNABLES=glibc.cpu.x86_ibt=on:glibc.cpu.x86_shstk=on doesn't | |
1860 | work | |
1861 | [27222] dynamic-link: Incorrect sysdeps/x86/tst-cpu-features-cpuinfo.c | |
1862 | [27237] malloc: deadlock in malloc/tst-malloc-stats-cancellation | |
1863 | [27256] locale: Assertion failure in ISO-2022-JP-3 gconv module | |
1864 | related to combining characters (CVE-2021-3326) | |
1865 | ||
1866 | \f | |
1867 | Version 2.32 | |
1868 | ||
1869 | Major new features: | |
1870 | ||
1871 | * Unicode 13.0.0 Support: Character encoding, character type info, and | |
1872 | transliteration tables are all updated to Unicode 13.0.0, using | |
1873 | generator scripts contributed by Mike FABIAN (Red Hat). | |
1874 | ||
1875 | * New locale added: ckb_IQ (Kurdish/Sorani spoken in Iraq) | |
1876 | ||
1877 | * Support for Synopsys ARC HS cores (ARCv2 ISA) running Linux has been | |
1878 | added. This port requires at least binutils-2.32, gcc-8.3 and Linux-5.1. | |
1879 | Three ABIs are supported: | |
1880 | ||
1881 | - arc-linux-gnu | |
1882 | - arc-linux-gnuhf | |
1883 | - arceb-linux-gnu | |
1884 | ||
1885 | The arc* ABIs are little-endian while arceb is big-endian. All ABIs use | |
1886 | 64-bit time (y2038 safe) and 64-bit file offsets (LFS default). | |
1887 | ||
1888 | * The GNU C Library now loads audit modules listed in the DT_AUDIT and | |
1889 | DT_DEPAUDIT dynamic section entries of the main executable. | |
1890 | ||
1891 | * powerpc64le supports IEEE128 long double libm/libc redirects when | |
1892 | using -mabi=ieeelongdouble to compile C code on supported GCC | |
1893 | toolchains. It is recommended to use GCC 8 or newer when testing | |
1894 | this option. | |
1895 | ||
1896 | * To help detect buffer overflows and other out-of-bounds accesses | |
1897 | several APIs have been annotated with GCC 'access' attribute. This | |
1898 | should help GCC 10 issue better warnings. | |
1899 | ||
1900 | * On Linux, functions pthread_attr_setsigmask_np and | |
1901 | pthread_attr_getsigmask_np have been added. They allow applications | |
1902 | to specify the signal mask of a thread created with pthread_create. | |
1903 | ||
1904 | * The GNU C Library now provides the header file <sys/single_threaded.h> | |
1905 | which declares the variable __libc_single_threaded. Applications are | |
1906 | encouraged to use this variable for single-thread optimizations, | |
1907 | instead of weak references to symbols historically defined in | |
1908 | libpthread. | |
1909 | ||
1910 | * The functions sigabbrev_np and sigdescr_np have been added. The | |
1911 | sigabbrev_np function returns the abbreviated signal name (e.g. "HUP" for | |
1912 | SIGHUP) while sigdescr_np returns a string describing the signal number | |
1913 | (e.g "Hangup" for SIGHUP). Different than strsignal, sigdescr_np does not | |
1914 | attempt to translate the return description, both functions return | |
1915 | NULL for an invalid signal number. | |
1916 | ||
1917 | They should be used instead of sys_siglist or sys_sigabbrev and they | |
1918 | are both thread and async-signal safe. These functions are GNU extensions. | |
1919 | ||
1920 | * The functions strerrorname_np and strerrordesc_np have been added. The | |
1921 | strerroname_np function returns error number name (e.g. "EINVAL" for EINVAL) | |
1922 | while strerrordesc_np returns a string describing the error number | |
1923 | (e.g "Invalid argument" for EINVAL). Different than strerror, | |
1924 | strerrordesc_np does not attempt to translate the return description, both | |
1925 | functions return NULL for an invalid error number. | |
1926 | ||
1927 | They should be used instead of sys_errlist and sys_nerr, both are | |
1928 | thread and async-signal safe. These functions are GNU extensions. | |
1929 | ||
1930 | * AArch64 now supports standard branch protection security hardening | |
1931 | in glibc when it is built with a GCC that is configured with | |
1932 | --enable-standard-branch-protection (or if -mbranch-protection=standard | |
1933 | flag is passed when building both GCC target libraries and glibc, | |
1934 | in either case a custom GCC is needed). This includes branch target | |
1935 | identification (BTI) and pointer authentication for return addresses | |
1936 | (PAC-RET). They require armv8.5-a and armv8.3-a architecture | |
1937 | extensions respectively for the protection to be effective, | |
1938 | otherwise the used instructions are nops. User code can use PAC-RET | |
1939 | without libc support, but BTI requires a libc that is built with BTI | |
1940 | support, otherwise runtime objects linked into user code will not be | |
1941 | BTI compatible. | |
1942 | ||
1943 | Deprecated and removed features, and other changes affecting compatibility: | |
1944 | ||
1945 | * Remove configure option --enable-obsolete-rpc. Sun RPC is removed | |
1946 | from glibc. This includes the rpcgen program, librpcsvc, and the Sun | |
1947 | RPC header files. Backward compatibility for old programs is kept | |
1948 | only for architectures and ABIs that have been added in or before | |
1949 | glibc 2.31. New programs need to use TI-RPC | |
1950 | <http://git.linux-nfs.org/?p=steved/libtirpc.git;a=summary> and | |
1951 | rpcsvc-proto <https://github.com/thkukuk/rpcsvc-proto>. | |
1952 | ||
1953 | * Remove configure option --enable-obsolete-nsl. libnsl is only built | |
1954 | as shared library for backward compatibility and the NSS modules "nis" | |
1955 | and "nisplus" are not built at all and libnsl's headers aren't | |
1956 | installed. This compatibility is kept only for architectures and ABIs | |
1957 | that have been added in or before version 2.28. Replacement | |
1958 | implementations based on TI-RPC, which additionally support IPv6, are | |
1959 | available from <https://github.com/thkukuk/>. This change does not | |
1960 | affect the "compat" NSS module, which does not depend on libnsl | |
1961 | since 2.27 and thus can be used without NIS. | |
1962 | ||
1963 | * The deprecated <sys/sysctl.h> header and the sysctl function have been | |
1964 | removed. To support old binaries, the sysctl function continues to | |
1965 | exist as a compatibility symbol (on those architectures which had it), | |
1966 | but always fails with ENOSYS. This reflects the removal of the system | |
1967 | call from all architectures, starting with Linux 5.5. | |
1968 | ||
1969 | * The sstk function is no longer available to newly linked binaries. | |
1970 | Its implementation always returned with a failure, and the function | |
1971 | was not declared in any header file. | |
1972 | ||
1973 | * The legacy signal handling functions siginterrupt, sigpause, sighold, | |
1974 | sigrelse, sigignore and sigset, and the sigmask macro have been | |
1975 | deprecated. Applications should use the sigsuspend, sigprocmask and | |
1976 | sigaction functions instead. | |
1977 | ||
1978 | * ldconfig now defaults to the new format for ld.so.cache. glibc has | |
1979 | already supported this format for almost 20 years. | |
1980 | ||
1981 | * The deprecated arrays sys_siglist, _sys_siglist, and sys_sigabbrev | |
1982 | are no longer available to newly linked binaries, and their declarations | |
1983 | have been removed from <string.h>. They are exported solely as | |
1984 | compatibility symbols to support old binaries. All programs should use | |
1985 | strsignal instead. | |
1986 | ||
1987 | * The deprecated symbols sys_errlist, _sys_errlist, sys_nerr, and _sys_nerr | |
1988 | are no longer available to newly linked binaries, and their declarations | |
1989 | have been removed from <stdio.h>. They are exported solely as | |
1990 | compatibility symbols to support old binaries. All programs should use | |
1991 | strerror or strerror_r instead. | |
1992 | ||
1993 | * Both strerror and strerror_l now share the same internal buffer in the | |
1994 | calling thread, meaning that the returned string pointer may be invalided | |
1995 | or contents might be overwritten on subsequent calls in the same thread or | |
1996 | if the thread is terminated. It makes strerror MT-safe. | |
1997 | ||
1998 | * Using weak references to libpthread functions such as pthread_create | |
1999 | or pthread_key_create to detect the singled-threaded nature of a | |
2000 | program is an obsolescent feature. Future versions of glibc will | |
2001 | define pthread_create within libc.so.6 itself, so such checks will | |
2002 | always flag the program as multi-threaded. Applications should check | |
2003 | the __libc_single_threaded variable declared in | |
2004 | <sys/single_threaded.h> instead. | |
2005 | ||
2006 | * The "files" NSS module no longer supports the "key" database (used for | |
2007 | secure RPC). The contents of the /etc/publickey file will be ignored, | |
2008 | regardless of the settings in /etc/nsswitch.conf. (This method of | |
2009 | storing RPC keys only supported the obsolete and insecure AUTH_DES | |
2010 | flavor of secure RPC.) | |
2011 | ||
2012 | * The __morecore and __after_morecore_hook malloc hooks and the default | |
2013 | implementation __default_morecore have been deprecated. Applications | |
2014 | should use malloc interposition to change malloc behavior, and mmap to | |
2015 | allocate anonymous memory. A future version of glibc may require that | |
2016 | applications which use the malloc hooks must preload a special shared | |
2017 | object, to enable the hooks. | |
2018 | ||
2019 | * The hesiod NSS module has been deprecated and will be removed in a | |
2020 | future version of glibc. System administrators are encouraged to | |
2021 | switch to other approaches for networked account databases, such as | |
2022 | LDAP. | |
2023 | ||
2024 | Changes to build and runtime requirements: | |
2025 | ||
2026 | * powerpc64le requires GCC 7.4 or newer. This is required for supporting | |
2027 | long double redirects. | |
2028 | ||
2029 | Security related changes: | |
2030 | ||
2031 | CVE-2016-10228: An infinite loop has been fixed in the iconv program when | |
2032 | invoked with the -c option and when processing invalid multi-byte input | |
2033 | sequences. Reported by Jan Engelhardt. | |
2034 | ||
2035 | CVE-2020-10029: Trigonometric functions on x86 targets suffered from stack | |
2036 | corruption when they were passed a pseudo-zero argument. Reported by Guido | |
2037 | Vranken / ForAllSecure Mayhem. | |
2038 | ||
2039 | CVE-2020-1752: A use-after-free vulnerability in the glob function when | |
2040 | expanding ~user has been fixed. | |
2041 | ||
2042 | CVE-2020-6096: A signed comparison vulnerability in the ARMv7 memcpy and | |
2043 | memmove functions has been fixed. Discovered by Jason Royes and Samual | |
2044 | Dytrych of the Cisco Security Assessment and Penetration Team (See | |
2045 | TALOS-2020-1019). | |
2046 | ||
2047 | The following bugs are resolved with this release: | |
2048 | ||
2049 | [9809] localedata: ckb_IQ: new Kurdish Sorani locale | |
2050 | [10441] manual: Backtraces code example lacks error checking | |
2051 | [10815] librt: [timer_create / SIGEV_THREAD] signalmask of | |
2052 | timer_sigev_thread dangerous | |
2053 | [14231] stdio: stdio-common tests memory requirements | |
2054 | [14578] libc: /proc-based emulation for lchmod, fchmodat | |
2055 | [16272] dynamic-link: dlopen()ing a DT_FILTER library crashes if | |
2056 | filtee has constructor | |
2057 | [19519] locale: iconv(1) with -c option hangs on illegal multi-byte | |
2058 | sequences (CVE-2016-10228) | |
2059 | [19737] admin: Doc page “20.5.2 Infinity and NaN” has incorrect HTML | |
2060 | character entities for infinity & pi | |
2061 | [20338] libc: Parsing of /etc/gshadow can return bad pointers causing | |
2062 | segfaults in applications | |
2063 | [20543] libc: Please move from .gnu.linkonce to comdat | |
2064 | [22489] network: gcc warns about implicit conversion in | |
2065 | ICMP6_FILTER_SETPASS with -Wsign-conversion | |
2066 | [22525] localedata: or_IN LC_COLLATE does not use copy "iso14651_t1" | |
2067 | [23294] math: Complex _FloatN functions are redirected to the wrong | |
2068 | function with -mlong-double-64 | |
2069 | [23296] libc: Data race in setting function descriptor during lazy | |
2070 | binding | |
2071 | [23668] dynamic-link: ldconfig: Default to the new format for | |
2072 | ld.so.cache | |
2073 | [23819] hurd: hurd: Add C11 thread support | |
2074 | [23990] build: test-container error out on failure to exec child. | |
2075 | [23991] build: shell-container typo in run_command_array | |
2076 | [24638] manual: Error in example of parsing a template string | |
2077 | [24654] manual: Wrong declaration of wcschr in libc manual | |
2078 | [24943] dynamic-link: Support DT_AUDIT, DT_DEPAUDIT in the dynamic | |
2079 | linker | |
2080 | [25051] dynamic-link: aarch64, powerpc64 uses surplus static tls for | |
2081 | dynamically loaded dsos | |
2082 | [25098] nptl: nptl: ctype classification functions are not AS-Safe | |
2083 | [25219] libc: improve out-of-bounds checking with GCC 10 attribute | |
2084 | access | |
2085 | [25262] libc: getcontext/setcontext/swapcontext unnecessarily save and | |
2086 | restore EAX, ECX and EDX | |
2087 | [25397] dynamic-link: Legacy bitmap doesn't cover jitted code | |
2088 | [25414] glob: 'glob' use-after-free bug (CVE-2020-1752) | |
2089 | [25420] network: Race condition in resolv_conf.c can result in caching | |
2090 | stale configuration forever | |
2091 | [25487] math: sinl() stack corruption from crafted input | |
2092 | (CVE-2020-10029) | |
2093 | [25506] build: configure: broken detection of STT_GNU_IFUNC when GCC | |
2094 | defaults to PIE | |
2095 | [25523] libc: MIPS/Linux inline syscall template is miscompiled | |
2096 | [25620] libc: Signed comparison vulnerability in the ARMv7 memcpy() | |
2097 | (CVE-2020-6096) | |
2098 | [25623] libc: test-sysvmsg, test-sysvsem, test-sysvshm fail with 2.31 | |
2099 | on 32 bit and old kernel | |
2100 | [25635] libc: arm: Wrong sysdep order selection for soft-fp | |
2101 | [25639] localedata: Some names of days and months wrongly spelt in | |
2102 | Occitan | |
2103 | [25657] libc: sigprocmask() and sigisemptyset() manipulate different | |
2104 | amount of sigset_t bytes | |
2105 | [25691] stdio: printf: memory leak when printing long multibyte | |
2106 | strings | |
2107 | [25715] libc: system() returns wrong errors when posix_spawn fails | |
2108 | [25733] malloc: mallopt(M_MXFAST) can set global_max_fast to 0 | |
2109 | [25734] locale: mbrtowc with Big5-HKSCS fails to reset conversion | |
2110 | state for conversions that produce two Unicode code points | |
2111 | [25765] nptl: Incorrect futex syscall in __pthread_disable_asynccancel | |
2112 | for linux x86_64 leads to livelock | |
2113 | [25788] dynamic-link: [i386] -fno-omit-frame-pointer in CFLAGS causes | |
2114 | test failures, invalid instruction in ld.so | |
2115 | [25790] glob: Typo in tst-fnmatch.input | |
2116 | [25810] libc: x32: Incorrect syscall entries with pointer, off_t and | |
2117 | size_t | |
2118 | [25819] localedata: Update locale data to Unicode 13.0.0 | |
2119 | [25824] libc: Abnormal function of strnlen in aarch64 | |
2120 | [25887] dynamic-link: Wasted space in _dl_x86_feature_1[1] | |
2121 | [25896] libc: Incorrect prctl | |
2122 | [25902] libc: Bad LOADARGS_N | |
2123 | [25905] dynamic-link: VSX registers are corrupted during PLT | |
2124 | resolution when glibc is built with --disable-multi-arch and --with- | |
2125 | cpu=power9 | |
2126 | [25933] string: Off by one error in __strncmp_avx2 when | |
2127 | length=VEC_SIZE*4 and strings are at page boundaries can cause a | |
2128 | segfault | |
2129 | [25942] nptl: Deadlock on stack_cache_lock between __nptl_setxid and | |
2130 | exiting detached thread | |
2131 | [25966] libc: Incorrect access of __x86_shared_non_temporal_threshold | |
2132 | for x32 | |
2133 | [25976] nss: internal_end*ent in nss_compat may clobber errno, hiding | |
2134 | ERANGE | |
2135 | [25999] nptl: Use-after-free issue in pthread_getaddr_default_np | |
2136 | [26073] math: getpayload() has wrong return value | |
2137 | [26076] dynamic-link: dlmopen crashes after failing to load | |
2138 | dependencies in audit mode | |
2139 | [26120] localedata: column width of of some Korean | |
2140 | JUNGSEONG/JONGSEONG characters wrong (should be 0) | |
2141 | [26128] libc: Incorrect bit_cpu_CLFLUSHOPT | |
2142 | [26133] libc: Incorrect need_arch_feature_F16C | |
2143 | [26137] libc: strtod() triggers exception FE_INEXACT on reasonable | |
2144 | input | |
2145 | [26149] libc: PKU is usable only if OSPKE is set | |
2146 | [26173] libc: powerpc64*: Add @notoc to calls to functions that do not | |
2147 | preserve r2 | |
2148 | [26208] libc: Incorrect bit_cpu_CLFSH | |
2149 | [26210] network: Incorrect use of hidden symbols for global sunrpc | |
2150 | variables | |
2151 | [26211] stdio: printf integer overflow calculating allocation size | |
2152 | [26214] stdio: printf_fp double free | |
2153 | [26215] stdio: printf_fp memory leak | |
2154 | [26232] time: FAIL: support/tst-timespec for 32-bit targets | |
2155 | [26258] nss: nss_compat should not read input files with mmap | |
2156 | [26332] string: Incorrect cache line size load causes memory | |
2157 | corruption in memset | |
2158 | ||
2159 | \f | |
2160 | Version 2.31 | |
2161 | ||
2162 | Major new features: | |
2163 | ||
2164 | * The GNU C Library now supports a feature test macro _ISOC2X_SOURCE to | |
2165 | enable features from the draft ISO C2X standard. Only some features from | |
2166 | this draft standard are supported by the GNU C Library, and as the draft | |
2167 | is under active development, the set of features enabled by this macro is | |
2168 | liable to change. Features from C2X are also enabled by _GNU_SOURCE, or | |
2169 | by compiling with "gcc -std=gnu2x". | |
2170 | ||
2171 | * The <math.h> functions that round their results to a narrower type now | |
2172 | have corresponding type-generic macros in <tgmath.h>, as defined in TS | |
2173 | 18661-1:2014 and TS 18661-3:2015 as amended by the resolution of | |
2174 | Clarification Request 13 to TS 18661-3. | |
2175 | ||
2176 | * The function pthread_clockjoin_np has been added, enabling join with a | |
2177 | terminated thread with a specific clock. It allows waiting against | |
2178 | CLOCK_MONOTONIC and CLOCK_REALTIME. This function is a GNU extension. | |
2179 | ||
2180 | * New locale added: mnw_MM (Mon language spoken in Myanmar). | |
2181 | ||
2182 | * The DNS stub resolver will optionally send the AD (authenticated data) bit | |
2183 | in queries if the trust-ad option is set via the options directive in | |
2184 | /etc/resolv.conf (or if RES_TRUSTAD is set in _res.options). In this | |
2185 | mode, the AD bit, as provided by the name server, is available to | |
2186 | applications which call res_search and related functions. In the default | |
2187 | mode, the AD bit is not set in queries, and it is automatically cleared in | |
2188 | responses, indicating a lack of DNSSEC validation. (Therefore, the name | |
2189 | servers and the network path to them are treated as untrusted.) | |
2190 | ||
2191 | Deprecated and removed features, and other changes affecting compatibility: | |
2192 | ||
2193 | * The totalorder and totalordermag functions, and the corresponding | |
2194 | functions for other floating-point types, now take pointer arguments to | |
2195 | avoid signaling NaNs possibly being converted to quiet NaNs in argument | |
2196 | passing. This is in accordance with the resolution of Clarification | |
2197 | Request 25 to TS 18661-1, as applied for C2X. Existing binaries that pass | |
2198 | floating-point arguments directly will continue to work. | |
2199 | ||
2200 | * The obsolete function stime is no longer available to newly linked | |
2201 | binaries, and its declaration has been removed from <time.h>. | |
2202 | Programs that set the system time should use clock_settime instead. | |
2203 | ||
2204 | * We plan to remove the obsolete function ftime, and the header <sys/timeb.h>, | |
2205 | in a future version of glibc. In this release, the header still exists | |
2206 | but calling ftime will cause a compiler warning. All programs should use | |
2207 | gettimeofday or clock_gettime instead. | |
2208 | ||
2209 | * The gettimeofday function no longer reports information about a | |
2210 | system-wide time zone. This 4.2-BSD-era feature has been deprecated for | |
2211 | many years, as it cannot handle the full complexity of the world's | |
2212 | timezones, but hitherto we have supported it on a best-effort basis. | |
2213 | Changes required to support 64-bit time_t on 32-bit architectures have | |
2214 | made this no longer practical. | |
2215 | ||
2216 | As of this release, callers of gettimeofday with a non-null 'tzp' argument | |
2217 | should expect to receive a 'struct timezone' whose tz_minuteswest and | |
2218 | tz_dsttime fields are zero. (For efficiency reasons, this does not always | |
2219 | happen on a few Linux-based ports. This will be corrected in a future | |
2220 | release.) | |
2221 | ||
2222 | All callers should supply a null pointer for the 'tzp' argument to | |
2223 | gettimeofday. For accurate information about the time zone associated | |
2224 | with the current time, use the localtime function. | |
2225 | ||
2226 | gettimeofday itself is obsolescent according to POSIX. We have no plans | |
2227 | to remove access to this function, but portable programs should consider | |
2228 | using clock_gettime instead. | |
2229 | ||
2230 | * The settimeofday function can still be used to set a system-wide time | |
2231 | zone when the operating system supports it. This is because the Linux | |
2232 | kernel reused the API, on some architectures, to describe a system-wide | |
2233 | time-zone-like offset between the software clock maintained by the kernel, | |
2234 | and the "RTC" clock that keeps time when the system is shut down. | |
2235 | ||
2236 | However, to reduce the odds of this offset being set by accident, | |
2237 | settimeofday can no longer be used to set the time and the offset | |
2238 | simultaneously. If both of its two arguments are non-null, the call | |
2239 | will fail (setting errno to EINVAL). | |
2240 | ||
2241 | Callers attempting to set this offset should also be prepared for the call | |
2242 | to fail and set errno to ENOSYS; this already happens on the Hurd and on | |
2243 | some Linux architectures. The Linux kernel maintainers are discussing a | |
2244 | more principled replacement for the reused API. After a replacement | |
2245 | becomes available, we will change settimeofday to fail with ENOSYS on all | |
2246 | platforms when its 'tzp' argument is not a null pointer. | |
2247 | ||
2248 | settimeofday itself is obsolescent according to POSIX. Programs that set | |
2249 | the system time should use clock_settime and/or the adjtime family of | |
2250 | functions instead. We may cease to make settimeofday available to newly | |
2251 | linked binaries after there is a replacement for Linux's time-zone-like | |
2252 | offset API. | |
2253 | ||
2254 | * SPARC ISA v7 is no longer supported. v8 is still supported, but only if | |
2255 | the optional CAS instruction is implemented (for instance, LEON processors | |
2256 | are still supported, but SuperSPARC processors are not). | |
2257 | ||
2258 | As the oldest 64-bit SPARC ISA is v9, this only affects 32-bit | |
2259 | configurations. | |
2260 | ||
2261 | * If a lazy binding failure happens during dlopen, during the execution of | |
2262 | an ELF constructor, the process is now terminated. Previously, the | |
2263 | dynamic loader would return NULL from dlopen, with the lazy binding error | |
2264 | captured in a dlerror message. In general, this is unsafe because | |
2265 | resetting the stack in an arbitrary function call is not possible. | |
2266 | ||
2267 | * For MIPS hard-float ABIs, the GNU C Library will be configured to need an | |
2268 | executable stack unless explicitly configured at build time to require | |
2269 | minimum kernel version 4.8 or newer. This is because executing | |
2270 | floating-point branches on a non-executable stack on Linux kernels prior to | |
2271 | 4.8 can lead to application crashes for some MIPS configurations. While | |
2272 | currently PT_GNU_STACK is not widely used on MIPS, future releases of GCC are | |
2273 | expected to enable non-executable stack by default with PT_GNU_STACK by | |
2274 | default and is thus likely to trigger a crash on older kernels. | |
2275 | ||
2276 | The GNU C Library can be built with --enable-kernel=4.8.0 in order to keep a | |
2277 | non-executable stack while dropping support for older kernels. | |
2278 | ||
2279 | * System call wrappers for time system calls now use the new time64 system | |
2280 | calls when available. On 32-bit targets, these wrappers attempt to call | |
2281 | the new system calls first and fall back to the older 32-bit time system | |
2282 | calls if they are not present. This may cause issues in environments | |
2283 | that cannot handle unsupported system calls gracefully by returning | |
2284 | -ENOSYS. Seccomp sandboxes are affected by this issue. | |
2285 | ||
2286 | Changes to build and runtime requirements: | |
2287 | ||
2288 | * It is no longer necessary to have recent Linux kernel headers to build | |
2289 | working (non-stub) system call wrappers on all architectures except 64-bit | |
2290 | RISC-V. 64-bit RISC-V requires a minimum kernel headers version of 5.0. | |
2291 | ||
2292 | * The ChangeLog file is no longer present in the toplevel directory of the | |
2293 | source tree. ChangeLog files are located in the ChangeLog.old directory as | |
2294 | ChangeLog.N where the highest N has the latest entries. | |
2295 | ||
2296 | Security related changes: | |
2297 | ||
2298 | CVE-2020-1751: A defect in the PowerPC backtrace function could cause an | |
2299 | out-of-bounds write when executed in a signal frame context. | |
2300 | ||
2301 | CVE-2019-19126: ld.so failed to ignore the LD_PREFER_MAP_32BIT_EXEC | |
2302 | environment variable during program execution after a security | |
2303 | transition, allowing local attackers to restrict the possible mapping | |
2304 | addresses for loaded libraries and thus bypass ASLR for a setuid | |
2305 | program. Reported by Marcin Kościelnicki. | |
2306 | ||
2307 | The following bugs are resolved with this release: | |
2308 | ||
2309 | [12031] localedata: iconv -t ascii//translit with Greek characters | |
2310 | [15813] libc: Multiple issues in __gen_tempname | |
2311 | [17726] libc: [arm, sparc] profil_counter should be compat symbol | |
2312 | [18231] libc: ipc_perm struct's mode member has wrong type in sys/ipc.h | |
2313 | [19767] libc: vdso is not used with static linking | |
2314 | [19903] hurd: Shared mappings not being inherited by children processes | |
2315 | [20358] network: RES_USE_DNSSEC sets DO; should also have a way to set AD | |
2316 | [20839] dynamic-link: Incomplete rollback of dynamic linker state on | |
2317 | linking failure | |
2318 | [23132] localedata: Missing transliterations in Miscellaneous Mathematical | |
2319 | Symbols-A/B Unicode blocks | |
2320 | [23518] libc: Eliminate __libc_utmp_jump_table | |
2321 | [24026] malloc: malloc_info() returns wrong numbers | |
2322 | [24054] localedata: Many locales are missing date_fmt | |
2323 | [24214] dynamic-link: user defined ifunc resolvers may run in ldd mode | |
2324 | [24304] dynamic-link: Lazy binding failure during ELF | |
2325 | constructors/destructors is not fatal | |
2326 | [24376] libc: RISC-V symbol size confusion with _start | |
2327 | [24682] localedata: zh_CN first weekday should be Monday per GB/T | |
2328 | 7408-2005 | |
2329 | [24824] libc: test-in-container does not install charmap files compatible | |
2330 | with localedef | |
2331 | [24844] regex: regex bad pointer / leakage if malloc fails | |
2332 | [24867] malloc: Unintended malloc_info formatting changes | |
2333 | [24879] libc: login: utmp alarm timer can arrive after lock acquisition | |
2334 | [24880] libc: login: utmp implementation uses struct flock with fcntl64 | |
2335 | [24882] libc: login: pututline uses potentially outdated cache | |
2336 | [24899] libc: Missing nonstring attributes in <utmp.h>, <utmpx.h> | |
2337 | [24902] libc: login: Repeating pututxline on EINTR/EAGAIN causes stale | |
2338 | utmp entries | |
2339 | [24916] dynamic-link: [MIPS] Highest EI_ABIVERSION value not raised to | |
2340 | ABSOLUTE ABI | |
2341 | [24930] dynamic-link: dlopen of PIE executable can result in | |
2342 | _dl_allocate_tls_init assertion failure | |
2343 | [24950] localedata: Top-of-tree glibc does not build with top-of-tree GCC | |
2344 | (stringop-overflow error) | |
2345 | [24959] time: librt IFUNC resolvers for clock_gettime and clock_* | |
2346 | functions other can lead to crashes | |
2347 | [24967] libc: jemalloc static linking causes runtime failure | |
2348 | [24986] libc: alpha: new getegid, geteuid and getppid syscalls used | |
2349 | unconditionally | |
2350 | [25035] libc: sbrk() failure handled poorly in tunables_strdup | |
2351 | [25087] dynamic-link: ldconfig mishandles unusual .dynstr placement | |
2352 | [25097] libc: new -Warray-bounds with GCC 10 | |
2353 | [25112] dynamic-link: dlopen must not make new objects accessible when it | |
2354 | still can fail with an error | |
2355 | [25139] localedata: Please add the new mnw_MM locale | |
2356 | [25149] regex: Array bounds violation in proceed_next_node | |
2357 | [25157] dynamic-link: Audit cookie for the dynamic loader is not | |
2358 | initialized correctly | |
2359 | [25189] libc: glibc's __glibc_has_include causes issues with clang | |
2360 | -frewrite-includes | |
2361 | [25194] malloc: malloc.c: do_set_mxfast incorrectly casts the mallopt | |
2362 | value to an unsigned | |
2363 | [25204] dynamic-link: LD_PREFER_MAP_32BIT_EXEC not ignored in setuid | |
2364 | binaries (CVE-2019-19126) | |
2365 | [25225] libc: ld.so fails to link on x86 if GCC defaults to -fcf- | |
2366 | protection | |
2367 | [25226] string: strstr: Invalid result if needle crosses page on s390-z15 | |
2368 | ifunc variant. | |
2369 | [25232] string: <string.h> does not enable const correctness for strchr et | |
2370 | al. for Clang++ | |
2371 | [25233] localedata: Consider "." as the thousands separator for sl_SI | |
2372 | (Slovenian) | |
2373 | [25241] nptl: __SIZEOF_PTHREAD_MUTEX_T defined twice for x86 | |
2374 | [25251] build: Failure to run tests when CFLAGS contains -DNDEBUG. | |
2375 | [25271] libc: undeclared identifier PTHREAD_MUTEX_DEFAULT when compiling | |
2376 | with -std=c11 | |
2377 | [25323] localedata: km_KH: d_t_fmt contains "m" instead of "%M" | |
2378 | [25324] localedata: lv_LV: d_t_fmt contains suspicious words in the time | |
2379 | part | |
2380 | [25396] dynamic-link: Failing dlopen can leave behind dangling GL | |
2381 | (dl_initfirst) link map pointer | |
2382 | [25401] malloc: pvalloc must not have __attribute_alloc_size__ | |
2383 | [25423] libc: Array overflow in backtrace on powerpc | |
2384 | [25425] network: Missing call to __resolv_context_put in | |
2385 | getaddrinfo.c:gethosts | |
2386 | ||
2387 | \f | |
2388 | Version 2.30 | |
2389 | ||
2390 | Major new features: | |
2391 | ||
2392 | * Unicode 12.1.0 Support: Character encoding, character type info, and | |
2393 | transliteration tables are all updated to Unicode 12.1.0, using | |
2394 | generator scripts contributed by Mike FABIAN (Red Hat). | |
2395 | ||
2396 | * The dynamic linker accepts the --preload argument to preload shared | |
2397 | objects, in addition to the LD_PRELOAD environment variable. | |
2398 | ||
2399 | * The twalk_r function has been added. It is similar to the existing | |
2400 | twalk function, but it passes an additional caller-supplied argument | |
2401 | to the callback function. | |
2402 | ||
2403 | * On Linux, the getdents64, gettid, and tgkill functions have been added. | |
2404 | ||
2405 | * Minguo (Republic of China) calendar support has been added as an | |
2406 | alternative calendar for the following locales: zh_TW, cmn_TW, hak_TW, | |
2407 | nan_TW, lzh_TW. | |
2408 | ||
2409 | * The entry for the new Japanese era has been added for ja_JP locale. | |
2410 | ||
2411 | * Memory allocation functions malloc, calloc, realloc, reallocarray, valloc, | |
2412 | pvalloc, memalign, and posix_memalign fail now with total object size | |
2413 | larger than PTRDIFF_MAX. This is to avoid potential undefined behavior with | |
2414 | pointer subtraction within the allocated object, where results might | |
2415 | overflow the ptrdiff_t type. | |
2416 | ||
2417 | * The dynamic linker no longer refuses to load objects which reference | |
2418 | versioned symbols whose implementation has moved to a different soname | |
2419 | since the object has been linked. The old error message, symbol | |
2420 | FUNCTION-NAME, version SYMBOL-VERSION not defined in file DSO-NAME with | |
2421 | link time reference, is gone. | |
2422 | ||
2423 | * Add new POSIX-proposed pthread_cond_clockwait, pthread_mutex_clocklock, | |
2424 | pthread_rwlock_clockrdlock, pthread_rwlock_clockwrlock and sem_clockwait | |
2425 | functions. These behave similarly to their "timed" equivalents, but also | |
2426 | accept a clockid_t parameter to determine which clock their timeout should | |
2427 | be measured against. All functions allow waiting against CLOCK_MONOTONIC | |
2428 | and CLOCK_REALTIME. The decision of which clock to be used is made at the | |
2429 | time of the wait (unlike with pthread_condattr_setclock, which requires | |
2430 | the clock choice at initialization time). | |
2431 | ||
2432 | * On AArch64 the GNU IFUNC resolver call ABI changed: old resolvers still | |
2433 | work, new resolvers can use a second argument which can be extended in | |
2434 | the future, currently it contains the AT_HWCAP2 value. | |
2435 | ||
2436 | Deprecated and removed features, and other changes affecting compatibility: | |
2437 | ||
2438 | * The copy_file_range function fails with ENOSYS if the kernel does not | |
2439 | support the system call of the same name. Previously, user space | |
2440 | emulation was performed, but its behavior did not match the kernel | |
2441 | behavior, which was deemed too confusing. Applications which use the | |
2442 | copy_file_range function can no longer rely on glibc to provide a fallback | |
2443 | on kernels that do not support the copy_file_range system call, and if | |
2444 | this function returns ENOSYS, they will need to use their own fallback. | |
2445 | Support for copy_file_range for most architectures was added in version | |
2446 | 4.5 of the mainline Linux kernel. | |
2447 | ||
2448 | * The functions clock_gettime, clock_getres, clock_settime, | |
2449 | clock_getcpuclockid, clock_nanosleep were removed from the librt library | |
2450 | for new applications (on architectures which had them). Instead, the | |
2451 | definitions in libc will be used automatically, which have been available | |
2452 | since glibc 2.17. | |
2453 | ||
2454 | * The obsolete and never-implemented XSI STREAMS header files <stropts.h> | |
2455 | and <sys/stropts.h> have been removed. | |
2456 | ||
2457 | * Support for the "inet6" option in /etc/resolv.conf and the RES_USE_INET6 | |
2458 | resolver flag (deprecated in glibc 2.25) have been removed. | |
2459 | ||
2460 | * The obsolete RES_INSECURE1 and RES_INSECURE2 option flags for the DNS stub | |
2461 | resolver have been removed from <resolv.h>. | |
2462 | ||
2463 | * With --enable-bind-now, installed programs are now linked with the | |
2464 | BIND_NOW flag. | |
2465 | ||
2466 | * Support for the PowerPC SPE ISA extension (powerpc-*-*gnuspe* | |
2467 | configurations) has been removed, following the deprecation of this | |
2468 | subarchitecture in version 8 of GCC, and its removal in version 9. | |
2469 | ||
2470 | * On 32-bit Arm, support for the port-based I/O emulation and the <sys/io.h> | |
2471 | header have been removed. | |
2472 | ||
2473 | * The Linux-specific <sys/sysctl.h> header and the sysctl function have been | |
2474 | deprecated and will be removed from a future version of glibc. | |
2475 | Application should directly access /proc instead. For obtaining random | |
2476 | bits, the getentropy function can be used. | |
2477 | ||
2478 | Changes to build and runtime requirements: | |
2479 | ||
2480 | * GCC 6.2 or later is required to build the GNU C Library. | |
2481 | ||
2482 | Older GCC versions and non-GNU compilers are still supported when | |
2483 | compiling programs that use the GNU C Library. | |
2484 | ||
2485 | Security related changes: | |
2486 | ||
2487 | CVE-2019-7309: x86-64 memcmp used signed Jcc instructions to check | |
2488 | size. For x86-64, memcmp on an object size larger than SSIZE_MAX | |
2489 | has undefined behavior. On x32, the size_t argument may be passed | |
2490 | in the lower 32 bits of the 64-bit RDX register with non-zero upper | |
2491 | 32 bits. When it happened with the sign bit of RDX register set, | |
2492 | memcmp gave the wrong result since it treated the size argument as | |
2493 | zero. Reported by H.J. Lu. | |
2494 | ||
2495 | CVE-2019-9169: Attempted case-insensitive regular-expression match | |
2496 | via proceed_next_node in posix/regexec.c leads to heap-based buffer | |
2497 | over-read. Reported by Hongxu Chen. | |
2498 | ||
2499 | The following bugs are resolved with this release: | |
2500 | ||
2501 | [2872] locale: Transliteration Cyrillic -> ASCII fails | |
2502 | [6399] libc: gettid() should have a wrapper | |
2503 | [16573] malloc: mtrace hangs when MALLOC_TRACE is defined | |
2504 | [16976] glob: fnmatch unbounded stack VLA for collating symbols | |
2505 | [17396] localedata: globbing for locale by [[.collating-element.]] | |
2506 | [18035] dynamic-link: pldd does no longer work, enters infinite loop | |
2507 | [18465] malloc: memusagestat is built using system C library | |
2508 | [18830] locale: iconv -c -f ascii with >buffer size worth of input before | |
2509 | invalid input drops valid char | |
2510 | [20188] nptl: libpthread IFUNC resolver for vfork can lead to crash | |
2511 | [20568] locale: Segfault with wide characters and setlocale/fgetwc/UTF-8 | |
2512 | [21897] localedata: Afar locales: Fix mon, abmon, and abday | |
2513 | [22964] localedata: The Japanese Era name will be changed on May 1, 2019 | |
2514 | [23352] malloc: __malloc_check_init still defined in public header | |
2515 | malloc.h. | |
2516 | [23403] nptl: Wrong alignment of TLS variables | |
2517 | [23501] libc: nftw() doesn't return dangling symlink's inode | |
2518 | [23733] malloc: Check the count before calling tcache_get() | |
2519 | [23741] malloc: Missing __attribute_alloc_size__ in many allocation | |
2520 | functions | |
2521 | [23831] localedata: nl_NL missing LC_NUMERIC thousands_sep | |
2522 | [23844] nptl: pthread_rwlock_trywrlock results in hang | |
2523 | [23983] argparse: Missing compat versions of argp_failure and argp_error | |
2524 | for long double = double | |
2525 | [23984] libc: Missing compat versions of err.h and error.h functions for | |
2526 | long double = double | |
2527 | [23996] localedata: Dutch salutations | |
2528 | [24040] libc: riscv64: unterminated call chain in __thread_start | |
2529 | [24047] network: libresolv should use IP_RECVERR/IPV6_RECVERR to avoid | |
2530 | long timeouts | |
2531 | [24051] stdio: puts and putchar output to _IO_stdout instead of stdout | |
2532 | [24059] nss: nss_files: get_next_alias calls fgets_unlocked without | |
2533 | checking for NULL. | |
2534 | [24114] regex: regexec buffer read overrun in "grep -i | |
2535 | '\(\(\)*.\)*\(\)\(\)\1'" | |
2536 | [24122] libc: Segfaults if 0 returned from la_version | |
2537 | [24153] stdio: Some input functions do not react to stdin assignment | |
2538 | [24155] string: x32 memcmp can treat positive length as 0 (if sign bit in | |
2539 | RDX is set) (CVE-2019-7309) | |
2540 | [24161] nptl: __run_fork_handlers self-deadlocks in malloc/tst-mallocfork2 | |
2541 | [24164] libc: Systemtap probes need to use "nr" constraint on 32-bit Arm, | |
2542 | not the default "nor" | |
2543 | [24166] dynamic-link: Dl_serinfo.dls_serpath[1] in dlfcn.h causes UBSAN | |
2544 | false positives, change to modern flexible array | |
2545 | [24180] nptl: pthread_mutex_trylock does not use the correct order of | |
2546 | instructions while maintaining the robust mutex list due to missing | |
2547 | compiler barriers. | |
2548 | [24194] librt: Non-compatibility symbols for clock_gettime etc. cause | |
2549 | unnecessary librt dependencies | |
2550 | [24200] localedata: Revert first_weekday removal in en_IE locale | |
2551 | [24211] nptl: Use-after-free in Systemtap probe in pthread_join | |
2552 | [24215] nptl: pthread_timedjoin_np should be a cancellation point | |
2553 | [24216] malloc: Check for large bin list corruption when inserting | |
2554 | unsorted chunk | |
2555 | [24228] stdio: old x86 applications that use legacy libio crash on exit | |
2556 | [24231] dynamic-link: [sparc64] R_SPARC_H34 implementation falls through | |
2557 | to R_SPARC_H44 | |
2558 | [24293] localedata: Missing Minguo calendar support for TW locales | |
2559 | [24296] localedata: Orthographic mistakes in 'day' and 'abday' sections in | |
2560 | tt_RU (Tatar) locale | |
2561 | [24307] localedata: Update locale data to Unicode 12.0.0 | |
2562 | [24323] dynamic-link: dlopen should not be able open PIE objects | |
2563 | [24335] build: "Obsolete types detected" with Linux 5.0 headers | |
2564 | [24369] localedata: Orthographic mistakes in 'mon' and 'abmon' sections in | |
2565 | tt_RU (Tatar) locale | |
2566 | [24370] localedata: Add lang_name for tt_RU locale | |
2567 | [24372] locale: Binary locale files are not architecture independent | |
2568 | [24394] time: strptime %Ey mis-parses final year of era | |
2569 | [24476] dynamic-link: __libc_freeres triggers bad free in libdl if dlerror | |
2570 | was not used | |
2571 | [24506] dynamic-link: FAIL: elf/tst-pldd with --enable-hardcoded-path-in- | |
2572 | tests | |
2573 | [24531] malloc: Malloc tunables give tcache assertion failures | |
2574 | [24532] libc: conform/arpa/inet.h failures due to linux kernel 64-bit | |
2575 | time_t changes | |
2576 | [24535] localedata: Update locale data to Unicode 12.1.0 | |
2577 | [24537] build: nptl/tst-eintr1 test case can hit task limits on some | |
2578 | kernels and break testing | |
2579 | [24544] build: elf/tst-pldd doesn't work if you install with a --prefix | |
2580 | [24556] build: [GCC 9] error: ‘%s’ directive argument is null | |
2581 | [-Werror=format-overflow=] | |
2582 | [24570] libc: alpha: compat msgctl uses __IPC_64 | |
2583 | [24584] locale: Data race in __wcsmbs_clone_conv | |
2584 | [24588] stdio: Remove codecvt vtables from libio | |
2585 | [24603] math: sysdeps/ieee754/dbl-64/branred.c is slow when compiled with | |
2586 | -O3 -march=skylake | |
2587 | [24614] localedata: nl_NL LC_MONETARY doesn't match CLDR 35 | |
2588 | [24632] stdio: Old binaries which use freopen with default stdio handles | |
2589 | crash | |
2590 | [24640] libc: __ppc_get_timebase_freq() always return 0 when using static | |
2591 | linked glibc | |
2592 | [24652] localedata: szl_PL spelling correction | |
2593 | [24695] nss: nss_db: calling getpwent after endpwent crashes | |
2594 | [24696] nss: endgrent() clobbers errno=ERRNO for 'group: db files' entry | |
2595 | in /etc/nsswitch.conf | |
2596 | [24699] libc: mmap64 with very large offset broken on MIPS64 n32 | |
2597 | [24740] libc: getdents64 type confusion | |
2598 | [24741] dynamic-link: ld.so should not require that a versioned symbol is | |
2599 | always implemented in the same library | |
2600 | [24744] libc: Remove copy_file_range emulation | |
2601 | [24757] malloc: memusagestat is linked against system libpthread | |
2602 | [24794] libc: Partial test suite run builds corrupt test-in-container | |
2603 | testroot | |
2604 | ||
2605 | \f | |
2606 | Version 2.29 | |
2607 | ||
2608 | Major new features: | |
2609 | ||
2610 | * The getcpu wrapper function has been added, which returns the currently | |
2611 | used CPU and NUMA node. This function is Linux-specific. | |
2612 | ||
2613 | * A new convenience target has been added for distribution maintainers | |
2614 | to build and install all locales as directories with files. The new | |
2615 | target is run by issuing the following command in your build tree: | |
2616 | 'make localedata/install-locale-files', with an optional DESTDIR | |
2617 | to set the install root if you wish to install into a non-default | |
2618 | configured location. | |
2619 | ||
2620 | * Optimized generic exp, exp2, log, log2, pow, sinf, cosf, sincosf and tanf. | |
2621 | ||
2622 | * The reallocarray function is now declared under _DEFAULT_SOURCE, not just | |
2623 | for _GNU_SOURCE, to match BSD environments. | |
2624 | ||
2625 | * For powercp64le ABI, Transactional Lock Elision is now enabled iff kernel | |
2626 | indicates that it will abort the transaction prior to entering the kernel | |
2627 | (PPC_FEATURE2_HTM_NOSC on hwcap2). On older kernels the transaction is | |
2628 | suspended, and this caused some undefined side-effects issues by aborting | |
2629 | transactions manually. Glibc avoided it by abort transactions manually on | |
2630 | each syscall, but it lead to performance issues on newer kernels where the | |
2631 | HTM state is saved and restore lazily (the state being saved even when the | |
2632 | process actually does not use HTM). | |
2633 | ||
2634 | * The functions posix_spawn_file_actions_addchdir_np and | |
2635 | posix_spawn_file_actions_addfchdir_np have been added, enabling | |
2636 | posix_spawn and posix_spawnp to run the new process in a different | |
2637 | directory. These functions are GNU extensions. The function | |
2638 | posix_spawn_file_actions_addchdir_np is similar to the Solaris function | |
2639 | of the same name. | |
2640 | ||
2641 | * The popen and system do not run atfork handlers anymore (BZ#17490). | |
2642 | Although it is a possible POSIX violation, the POSIX rationale in | |
2643 | pthread_atfork documentation regarding atfork handlers is to handle | |
2644 | inconsistent mutex state after a fork call in a multi-threaded process. | |
2645 | In both popen and system there is no direct access to user-defined mutexes. | |
2646 | ||
2647 | * Support for the C-SKY ABIV2 running on Linux has been added. This port | |
2648 | requires at least binutils-2.32, gcc-9.0, and linux-4.20. Two ABIs are | |
2649 | supported: | |
2650 | - C-SKY ABIV2 soft-float little-endian | |
2651 | - C-SKY ABIV2 hard-float little-endian | |
2652 | ||
2653 | * strftime's default formatting of a locale's alternative year (%Ey) | |
2654 | has been changed to zero-pad the year to a minimum of two digits, | |
2655 | like "%y". This improves the display of Japanese era years during | |
2656 | the first nine years of a new era, and is expected to be harmless | |
2657 | for all other locales (only Japanese locales regularly have | |
2658 | alternative year numbers less than 10). Zero-padding can be | |
2659 | overridden with the '_' or '-' flags (which are GNU extensions). | |
2660 | ||
2661 | * As a GNU extension, the '_' and '-' flags can now be applied to | |
2662 | "%EY" to control how the year number is formatted; they have the | |
2663 | same effect that they would on "%Ey". | |
2664 | ||
2665 | Deprecated and removed features, and other changes affecting compatibility: | |
2666 | ||
2667 | * The glibc.tune tunable namespace has been renamed to glibc.cpu and the | |
2668 | tunable glibc.tune.cpu has been renamed to glibc.cpu.name. | |
2669 | ||
2670 | * The type of the pr_uid and pr_gid members of struct elf_prpsinfo, defined | |
2671 | in <sys/procfs.h>, has been corrected to match the type actually used by | |
2672 | the Linux kernel. This affects the size and layout of that structure on | |
2673 | MicroBlaze, MIPS (n64 ABI only), Nios II and RISC-V. | |
2674 | ||
2675 | * For the MIPS n32 ABI, the type of the pr_sigpend and pr_sighold members of | |
2676 | struct elf_prstatus, and the pr_flag member of struct elf_prpsinfo, | |
2677 | defined in <sys/procfs.h>, has been corrected to match the type actually | |
2678 | used by the Linux kernel. This affects the size and layout of those | |
2679 | structures. | |
2680 | ||
2681 | * An archaic GNU extension to scanf, under which '%as', '%aS', and '%a[...]' | |
2682 | meant to scan a string and allocate space for it with malloc, is now | |
2683 | restricted to programs compiled in C89 or C++98 mode with _GNU_SOURCE | |
2684 | defined. This extension conflicts with C99's use of '%a' to scan a | |
2685 | hexadecimal floating-point number, which is now available to programs | |
2686 | compiled as C99 or C++11 or higher, regardless of _GNU_SOURCE. | |
2687 | ||
2688 | POSIX.1-2008 includes the feature of allocating a buffer for string input | |
2689 | with malloc, using the modifier letter 'm' instead. Programs using | |
2690 | '%as', '%aS', or '%a[...]' with the old GNU meaning should change to | |
2691 | '%ms', '%mS', or '%m[...]' respectively. Programs that wish to use the | |
2692 | C99 '%a' no longer need to avoid _GNU_SOURCE. | |
2693 | ||
2694 | GCC's -Wformat warnings can detect most uses of this extension, as long | |
2695 | as all functions that call vscanf, vfscanf, or vsscanf are annotated with | |
2696 | __attribute__ ((format (scanf, ...))). | |
2697 | ||
2698 | Changes to build and runtime requirements: | |
2699 | ||
2700 | * Python 3.4 or later is required to build the GNU C Library. | |
2701 | ||
2702 | * On most architectures, GCC 5 or later is required to build the GNU C | |
2703 | Library. (On powerpc64le, GCC 6.2 or later is still required, as before.) | |
2704 | ||
2705 | Older GCC versions and non-GNU compilers are still supported when | |
2706 | compiling programs that use the GNU C Library. | |
2707 | ||
2708 | Security related changes: | |
2709 | ||
2710 | CVE-2018-19591: A file descriptor leak in if_nametoindex can lead to a | |
2711 | denial of service due to resource exhaustion when processing getaddrinfo | |
2712 | calls with crafted host names. Reported by Guido Vranken. | |
2713 | ||
2714 | CVE-2019-6488: On x32, the size_t parameter may be passed in the lower | |
2715 | 32 bits of a 64-bit register with with non-zero upper 32 bit. When it | |
2716 | happened, accessing the 32-bit size_t value as the full 64-bit register | |
2717 | in the assembly string/memory functions would cause a buffer overflow. | |
2718 | Reported by H.J. Lu. | |
2719 | ||
2720 | CVE-2016-10739: The getaddrinfo function could successfully parse IPv4 | |
2721 | addresses with arbitrary trailing characters, potentially leading to data | |
2722 | or command injection issues in applications. | |
2723 | ||
2724 | The following bugs are resolved with this release: | |
2725 | ||
2726 | [10425] localedata: it_IT/it_CH: LC_TIME format is wrong | |
2727 | [10496] localedata: 12h time representation in multiple locales faulty | |
2728 | [10797] localedata: it_IT locale numeric does not have a separator for | |
2729 | thousands | |
2730 | [11319] libc: dprintf doesn't handle errors properly | |
2731 | [16346] time: mktime: potentially unsafe use of localtime_offset | |
2732 | [17248] build: glibc should not sort CFLAGS (support gcc plugins and | |
2733 | --param options) | |
2734 | [17405] libc: Implement posix_spawn_file_actions_addchdir_np, | |
2735 | posix_spawn_file_actions_addfchdir_np | |
2736 | [17426] localedata: Indian locales: set the correct date format | |
2737 | [17490] stdio: popen should not invoke atfork handlers | |
2738 | [17783] libc: TIOCSER_TEMT conditions inconsistent | |
2739 | [18040] regex: use-after-free in regexec/get_subexp | |
2740 | [18093] libc: Corrupted aux-cache causes ldconfig to segfault | |
2741 | [20018] network: getaddrinfo should reject IP addresses with trailing | |
2742 | characters (CVE-2016-10739) | |
2743 | [20209] localedata: Spelling mistake for Sunday in Greenlandic kl_GL | |
2744 | [20271] libc: Missing "\n" in __libc_fatal calls | |
2745 | [20480] dynamic-link: Patch: ifunc not executable, crashes sudo qemu | |
2746 | [20544] libc: RFE: atexit, __cxa_atexit, on_exit should assert function | |
2747 | pointer argument is non-NULL | |
2748 | [21037] stdio: open_memstream and freopen | |
2749 | [21286] libc: bits/siginfo.h is missing enum definition for TRAP_HWBKPT | |
2750 | [21716] time: Crash in glibc's mktime in low-memory situations | |
2751 | [22834] stdio: Subprocess forked by popen may crash in Linux when | |
2752 | multithreads call popen | |
2753 | [22927] network: crash in vn_gai_enqueue_request if requests_tail was NULL | |
2754 | and pthread_create fails. | |
2755 | [23032] hurd: sysdeps/htl/pt-barrier-init.c:39: bad call to memcmp ? | |
2756 | [23125] libc: riscv64: endless loop when throwing an exception from a | |
2757 | constructor | |
2758 | [23275] nptl: Race in pthread_mutex_lock while promoting to | |
2759 | PTHREAD_MUTEX_ELISION_NP. | |
2760 | [23400] libc: stdlib/test-bz22786.c creates temporary files in glibc | |
2761 | source tree | |
2762 | [23479] math: [mips] bits/fenv.h should not define some macros for soft- | |
2763 | float | |
2764 | [23490] libc: sysdeps/unix/sysv/linux/x86/tst-cet-property-2.c:49: off by | |
2765 | one error | |
2766 | [23497] libc: readdir64@GLIBC_2.1 cannot parse the kernel directory stream | |
2767 | [23509] dynamic-link: CET enabled glibc is incompatible with the older | |
2768 | linker | |
2769 | [23520] nscd: nscd: Use-after-free in addgetnetgrentX and its callers | |
2770 | [23521] nss: get_next_alias nss_files file stream leak | |
2771 | [23538] nptl: Hang in pthread_cond_broadcast | |
2772 | [23562] libc: Wrong type for si_band in Linux-specific siginfo_t | |
2773 | [23578] regex: Invalid memory access if regex pattern contains NUL byte | |
2774 | [23579] libc: Errors misreported in preadv2 | |
2775 | [23597] build: support/test-container.c doesn't work with different | |
2776 | filesystems | |
2777 | [23603] time: mktime signed integer overflow on large timestamps | |
2778 | [23606] libc: Missing ENDBR32 in sysdeps/i386/start.S | |
2779 | [23614] libc: powerpc: missing CFI register information in __mpn_* | |
2780 | functions | |
2781 | [23637] string: Generic strstr/strcasestr fails with huge needles | |
2782 | [23640] libc: no way to easily clear FD_CLOEXEC in | |
2783 | posix_spawn_file_actions_adddup2() | |
2784 | [23649] libc: [microblaze/mips/nios2/riscv] sys/procfs.h pr_uid, pr_gid | |
2785 | have wrong type | |
2786 | [23656] libc: [mips n32] sys/procfs.h pr_sigpend, pr_sighold, pr_flag have | |
2787 | wrong type | |
2788 | [23679] libc: gethostid: Missing NULL check for gethostbyname_r result | |
2789 | [23689] libc: Bug in documentation for rusage.ru_ixrss in | |
2790 | bits/types/struct_rusage.h | |
2791 | [23690] dynamic-link: Segfault in _dl_profile_fixup with a high number of | |
2792 | threads | |
2793 | [23707] dynamic-link: Missing unwind info in sysdeps/powerpc/powerpc32/dl- | |
2794 | start.S | |
2795 | [23709] string: glibc 2.25 lacks sse2 optimized strstr() | |
2796 | [23716] dynamic-link: _dl_runtime_resolve_shstk isn't selected properly | |
2797 | [23717] libc: glibc: stdlib/tst-setcontext9 test suite failure on | |
2798 | powerpc64le | |
2799 | [23724] localedata: Albanian date formats are incorrect | |
2800 | [23735] math: libnldbl_nonshared.a references internal libm symbols | |
2801 | [23740] localedata: kl_GL: Month names and date formats need update | |
2802 | [23744] regex: regex refactorings to remove BE, avoid duplication | |
2803 | [23745] time: mktime fix for Gnulib + coreutils | |
2804 | [23758] time: Improve the width of alternate representation for year in | |
2805 | strftime | |
2806 | [23783] libc: [mips] Missing CMSPAR bits/termios.h | |
2807 | [23789] time: mktime does not set errno on failure | |
2808 | [23791] localedata: Wrong monetary format for ca_ES locale | |
2809 | [23793] locale: c32rtomb and mbrtoc32 should not alias wcrtomb and mbrtowc | |
2810 | [23794] locale: c16rtomb does not handle surrogate pairs | |
2811 | [23821] libc: si_band in siginfo_t has wrong type long int on sparc64 | |
2812 | [23822] math: ia64 static libm.a is missing exp2f, log2f and powf symbols | |
2813 | [23836] time: time/tst-mktime2 test failure on Arm (32-bit) | |
2814 | [23848] libc: [sparc] Some socket syscalls wrongly assumed to be present | |
2815 | [23861] nptl: rdlock stalls indefinitely on an unlocked pthread rwlock | |
2816 | [23862] libc: [sh] missing kernel-features.h undefines | |
2817 | [23864] libc: [riscv] missing kernel-features.h undefines | |
2818 | [23867] libc: [arm/microblaze] __ASSUME_MLOCK2 incorrect | |
2819 | [23907] malloc: Incorrect double-free malloc tcache check disregards | |
2820 | tcache size | |
2821 | [23913] libc: off-by-one in function maybe_script_execute in | |
2822 | sysdeps/posix/spawni.c | |
2823 | [23915] libc: [arm] __ASSUME_COPY_FILE_RANGE incorrect | |
2824 | [23923] locale: Add --no-hard-links option to localedef | |
2825 | [23927] network: Linux if_nametoindex() does not close descriptor | |
2826 | (CVE-2018-19591) | |
2827 | [23961] math: powf can overflow to inf without setting errno in non- | |
2828 | nearest rounding mode | |
2829 | [23967] libc: [2.28 Regression]: New sigaction implementation breaks m68k | |
2830 | [23972] libc: __old_getdents64 uses wrong d_off value on overflow | |
2831 | [23993] libc: glibc 2.29 doesn't build with gcc 4.9 | |
2832 | [23995] localedata: Remove execution flags from localedata/locales/bi_VU | |
2833 | [24011] localedata: Fixed small type in comment for locale bs_BA | |
2834 | [24018] libc: gettext() may return NULL | |
2835 | [24022] build: riscv build failure with Linux kernel 4.20-rc7 | |
2836 | [24023] build: [2.29 Regression] FAIL: elf/check-localplt | |
2837 | [24024] string: strerror() might set errno to ENOMEM due to -fno-math- | |
2838 | error | |
2839 | [24027] malloc: glibc: realloc() ncopies 32-bit integer overflow | |
2840 | [24034] libc: tst-cancel21-static fails with SIGBUS on pre-ARMv7 when | |
2841 | using GCC 8 | |
2842 | [24046] localedata: en_US locale doesn't define date_fmt | |
2843 | [24063] manual: @var{errno} should be @code{errno} | |
2844 | [24066] soft-fp: Inconsistent _FP_W_TYPE_SIZE check | |
2845 | [24088] libc: VSCR field is not being correctly read in ucontext_t on | |
2846 | ppc64le | |
2847 | [24096] time: Specifying '_' or '-' flag for "%EY" does not produce the | |
2848 | expected result | |
2849 | [24097] string: Can't use 64-bit register for size_t in assembly codes for | |
2850 | x32 (CVE-2019-6488) | |
2851 | [24110] hurd: SS_DISABLE never set in stack_t value returned by | |
2852 | sigaltstack | |
2853 | [24112] network: Do not send DNS queries for non-host names (where all | |
2854 | answers will be rejected) | |
2855 | [24130] libc: alpha __remqu corrupts $f3 register | |
2856 | ||
2857 | \f | |
2858 | Version 2.28 | |
2859 | ||
2860 | Major new features: | |
2861 | ||
2862 | * The localization data for ISO 14651 is updated to match the 2016 | |
2863 | Edition 4 release of the standard, this matches data provided by | |
2864 | Unicode 9.0.0. This update introduces significant improvements to the | |
2865 | collation of Unicode characters. This release deviates slightly from | |
2866 | the standard in that the collation element ordering for lowercase and | |
2867 | uppercase LATIN script characters is adjusted to ensure that regular | |
2868 | expressions with ranges like [a-z] and [A-Z] don't interleave e.g. A | |
2869 | is not matched by [a-z]. With the update many locales have been | |
2870 | updated to take advantage of the new collation information. The new | |
2871 | collation information has increased the size of the compiled locale | |
2872 | archive or binary locales. | |
2873 | ||
2874 | * The GNU C Library can now be compiled with support for Intel CET, AKA | |
2875 | Intel Control-flow Enforcement Technology. When the library is built | |
2876 | with --enable-cet, the resulting glibc is protected with indirect | |
2877 | branch tracking (IBT) and shadow stack (SHSTK). CET-enabled glibc is | |
2878 | compatible with all existing executables and shared libraries. This | |
2879 | feature is currently supported on i386, x86_64 and x32 with GCC 8 and | |
2880 | binutils 2.29 or later. Note that CET-enabled glibc requires CPUs | |
2881 | capable of multi-byte NOPs, like x86-64 processors as well as Intel | |
2882 | Pentium Pro or newer. NOTE: --enable-cet has been tested for i686, | |
2883 | x86_64 and x32 on non-CET processors. --enable-cet has been tested | |
2884 | for x86_64 and x32 on CET SDVs, but Intel CET support hasn't been | |
2885 | validated for i686. | |
2886 | ||
2887 | * The GNU C Library now has correct support for ABSOLUTE symbols | |
2888 | (SHN_ABS-relative symbols). Previously such ABSOLUTE symbols were | |
2889 | relocated incorrectly or in some cases discarded. The GNU linker can | |
2890 | make use of the newer semantics, but it must communicate it to the | |
2891 | dynamic loader by setting the ELF file's identification (EI_ABIVERSION | |
2892 | field) to indicate such support is required. | |
2893 | ||
2894 | * Unicode 11.0.0 Support: Character encoding, character type info, and | |
2895 | transliteration tables are all updated to Unicode 11.0.0, using | |
2896 | generator scripts contributed by Mike FABIAN (Red Hat). | |
2897 | ||
2898 | * <math.h> functions that round their results to a narrower type are added | |
2899 | from TS 18661-1:2014 and TS 18661-3:2015: | |
2900 | ||
2901 | - fadd, faddl, daddl and corresponding fMaddfN, fMaddfNx, fMxaddfN and | |
2902 | fMxaddfNx functions. | |
2903 | ||
2904 | - fsub, fsubl, dsubl and corresponding fMsubfN, fMsubfNx, fMxsubfN and | |
2905 | fMxsubfNx functions. | |
2906 | ||
2907 | - fmul, fmull, dmull and corresponding fMmulfN, fMmulfNx, fMxmulfN and | |
2908 | fMxmulfNx functions. | |
2909 | ||
2910 | - fdiv, fdivl, ddivl and corresponding fMdivfN, fMdivfNx, fMxdivfN and | |
2911 | fMxdivfNx functions. | |
2912 | ||
2913 | * Two grammatical forms of month names are now supported for the following | |
2914 | languages: Armenian, Asturian, Catalan, Czech, Kashubian, Occitan, Ossetian, | |
2915 | Scottish Gaelic, Upper Sorbian, and Walloon. The following languages now | |
2916 | support two grammatical forms in abbreviated month names: Catalan, Greek, | |
2917 | and Kashubian. | |
2918 | ||
2919 | * Newly added locales: Lower Sorbian (dsb_DE) and Yakut (sah_RU) also | |
2920 | include the support for two grammatical forms of month names. | |
2921 | ||
2922 | * Building and running on GNU/Hurd systems now works without out-of-tree | |
2923 | patches. | |
2924 | ||
2925 | * The renameat2 function has been added, a variant of the renameat function | |
2926 | which has a flags argument. If the flags are zero, the renameat2 function | |
2927 | acts like renameat. If the flag is not zero and there is no kernel | |
2928 | support for renameat2, the function will fail with an errno value of | |
2929 | EINVAL. This is different from the existing gnulib function renameatu, | |
2930 | which performs a plain rename operation in case of a RENAME_NOREPLACE | |
2931 | flags and a non-existing destination (and therefore has a race condition | |
2932 | that can clobber the destination inadvertently). | |
2933 | ||
2934 | * The statx function has been added, a variant of the fstatat64 | |
2935 | function with an additional flags argument. If there is no direct | |
2936 | kernel support for statx, glibc provides basic stat support based on | |
2937 | the fstatat64 function. | |
2938 | ||
2939 | * IDN domain names in getaddrinfo and getnameinfo now use the system libidn2 | |
2940 | library if installed. libidn2 version 2.0.5 or later is recommended. If | |
2941 | libidn2 is not available, internationalized domain names are not encoded | |
2942 | or decoded even if the AI_IDN or NI_IDN flags are passed to getaddrinfo or | |
2943 | getnameinfo. (getaddrinfo calls with non-ASCII names and AI_IDN will fail | |
2944 | with an encoding error.) Flags which used to change the IDN encoding and | |
2945 | decoding behavior (AI_IDN_ALLOW_UNASSIGNED, AI_IDN_USE_STD3_ASCII_RULES, | |
2946 | NI_IDN_ALLOW_UNASSIGNED, NI_IDN_USE_STD3_ASCII_RULES) have been | |
2947 | deprecated. They no longer have any effect. | |
2948 | ||
2949 | * Parsing of dynamic string tokens in DT_RPATH, DT_RUNPATH, DT_NEEDED, | |
2950 | DT_AUXILIARY, and DT_FILTER has been expanded to support the full | |
2951 | range of ELF gABI expressions including such constructs as | |
2952 | '$ORIGIN$ORIGIN' (if valid). For SUID/GUID applications the rules | |
2953 | have been further restricted, and where in the past a dynamic string | |
2954 | token sequence may have been interpreted as a literal string it will | |
2955 | now cause a load failure. These load failures were always considered | |
2956 | unspecified behaviour from the perspective of the dynamic loader, and | |
2957 | for safety are now load errors e.g. /foo/${ORIGIN}.so in DT_NEEDED | |
2958 | results in a load failure now. | |
2959 | ||
2960 | * Support for ISO C threads (ISO/IEC 9899:2011) has been added. The | |
2961 | implementation includes all the standard functions provided by | |
2962 | <threads.h>: | |
2963 | ||
2964 | - thrd_current, thrd_equal, thrd_sleep, thrd_yield, thrd_create, | |
2965 | thrd_detach, thrd_exit, and thrd_join for thread management. | |
2966 | ||
2967 | - mtx_init, mtx_lock, mtx_timedlock, mtx_trylock, mtx_unlock, and | |
2968 | mtx_destroy for mutual exclusion. | |
2969 | ||
2970 | - call_once for function call synchronization. | |
2971 | ||
2972 | - cnd_broadcast, cnd_destroy, cnd_init, cnd_signal, cnd_timedwait, and | |
2973 | cnd_wait for conditional variables. | |
2974 | ||
2975 | - tss_create, tss_delete, tss_get, and tss_set for thread-local storage. | |
2976 | ||
2977 | Application developers must link against libpthread to use ISO C threads. | |
2978 | ||
2979 | Deprecated and removed features, and other changes affecting compatibility: | |
2980 | ||
2981 | * The nonstandard header files <libio.h> and <_G_config.h> are no longer | |
2982 | installed. Software that was using either header should be updated to | |
2983 | use standard <stdio.h> interfaces instead. | |
2984 | ||
2985 | * The stdio functions 'getc' and 'putc' are no longer defined as macros. | |
2986 | This was never required by the C standard, and the macros just expanded | |
2987 | to call alternative names for the same functions. If you hoped getc and | |
2988 | putc would provide performance improvements over fgetc and fputc, instead | |
2989 | investigate using (f)getc_unlocked and (f)putc_unlocked, and, if | |
2990 | necessary, flockfile and funlockfile. | |
2991 | ||
2992 | * All stdio functions now treat end-of-file as a sticky condition. If you | |
2993 | read from a file until EOF, and then the file is enlarged by another | |
2994 | process, you must call clearerr or another function with the same effect | |
2995 | (e.g. fseek, rewind) before you can read the additional data. This | |
2996 | corrects a longstanding C99 conformance bug. It is most likely to affect | |
2997 | programs that use stdio to read interactive input from a terminal. | |
2998 | (Bug #1190.) | |
2999 | ||
3000 | * The macros 'major', 'minor', and 'makedev' are now only available from | |
3001 | the header <sys/sysmacros.h>; not from <sys/types.h> or various other | |
3002 | headers that happen to include <sys/types.h>. These macros are rarely | |
3003 | used, not part of POSIX nor XSI, and their names frequently collide with | |
3004 | user code; see https://sourceware.org/bugzilla/show_bug.cgi?id=19239 for | |
3005 | further explanation. | |
3006 | ||
3007 | <sys/sysmacros.h> is a GNU extension. Portable programs that require | |
3008 | these macros should first include <sys/types.h>, and then include | |
3009 | <sys/sysmacros.h> if __GNU_LIBRARY__ is defined. | |
3010 | ||
3011 | * The tilegx*-*-linux-gnu configurations are no longer supported. | |
3012 | ||
3013 | * The obsolete function ustat is no longer available to newly linked | |
3014 | binaries; the headers <ustat.h> and <sys/ustat.h> have been removed. This | |
3015 | function has been deprecated in favor of fstatfs and statfs. | |
3016 | ||
3017 | * The obsolete function nfsservctl is no longer available to newly linked | |
3018 | binaries. This function was specific to systems using the Linux kernel | |
3019 | and could not usefully be used with the GNU C Library on systems with | |
3020 | version 3.1 or later of the Linux kernel. | |
3021 | ||
3022 | * The obsolete function name llseek is no longer available to newly linked | |
3023 | binaries. This function was specific to systems using the Linux kernel | |
3024 | and was not declared in a header. Programs should use the lseek64 name | |
3025 | for this function instead. | |
3026 | ||
3027 | * The AI_IDN_ALLOW_UNASSIGNED and NI_IDN_ALLOW_UNASSIGNED flags for the | |
3028 | getaddrinfo and getnameinfo functions have been deprecated. The behavior | |
3029 | previously selected by them is now always enabled. | |
3030 | ||
3031 | * The AI_IDN_USE_STD3_ASCII_RULES and NI_IDN_USE_STD3_ASCII_RULES flags for | |
3032 | the getaddrinfo and getnameinfo functions have been deprecated. The STD3 | |
3033 | restriction (rejecting '_' in host names, among other things) has been | |
3034 | removed, for increased compatibility with non-IDN name resolution. | |
3035 | ||
3036 | * The fcntl function now have a Long File Support variant named fcntl64. It | |
3037 | is added to fix some Linux Open File Description (OFD) locks usage on non | |
3038 | LFS mode. As for others *64 functions, fcntl64 semantics are analogous with | |
3039 | fcntl and LFS support is handled transparently. Also for Linux, the OFD | |
3040 | locks act as a cancellation entrypoint. | |
3041 | ||
3042 | * The obsolete functions encrypt, encrypt_r, setkey, setkey_r, cbc_crypt, | |
3043 | ecb_crypt, and des_setparity are no longer available to newly linked | |
3044 | binaries, and the headers <rpc/des_crypt.h> and <rpc/rpc_des.h> are no | |
3045 | longer installed. These functions encrypted and decrypted data with the | |
3046 | DES block cipher, which is no longer considered secure. Software that | |
3047 | still uses these functions should switch to a modern cryptography library, | |
3048 | such as libgcrypt. | |
3049 | ||
3050 | * Reflecting the removal of the encrypt and setkey functions above, the | |
3051 | macro _XOPEN_CRYPT is no longer defined. As a consequence, the crypt | |
3052 | function is no longer declared unless _DEFAULT_SOURCE or _GNU_SOURCE is | |
3053 | enabled. | |
3054 | ||
3055 | * The obsolete function fcrypt is no longer available to newly linked | |
3056 | binaries. It was just another name for the standard function crypt, | |
3057 | and it has not appeared in any header file in many years. | |
3058 | ||
3059 | * We have tentative plans to hand off maintenance of the passphrase-hashing | |
3060 | library, libcrypt, to a separate development project that will, we hope, | |
3061 | keep up better with new passphrase-hashing algorithms. We will continue | |
3062 | to declare 'crypt' in <unistd.h>, and programs that use 'crypt' or | |
3063 | 'crypt_r' should not need to change at all; however, distributions will | |
3064 | need to install <crypt.h> and libcrypt from a separate project. | |
3065 | ||
3066 | In this release, if the configure option --disable-crypt is used, glibc | |
3067 | will not install <crypt.h> or libcrypt, making room for the separate | |
3068 | project's versions of these files. The plan is to make this the default | |
3069 | behavior in a future release. | |
3070 | ||
3071 | Changes to build and runtime requirements: | |
3072 | ||
3073 | GNU make 4.0 or later is now required to build glibc. | |
3074 | ||
3075 | Security related changes: | |
3076 | ||
3077 | CVE-2016-6261, CVE-2016-6263, CVE-2017-14062: Various vulnerabilities have | |
3078 | been fixed by removing the glibc-internal IDNA implementation and using | |
3079 | the system-provided libidn2 library instead. Originally reported by Hanno | |
3080 | Böck and Christian Weisgerber. | |
3081 | ||
3082 | CVE-2017-18269: An SSE2-based memmove implementation for the i386 | |
3083 | architecture could corrupt memory. Reported by Max Horn. | |
3084 | ||
3085 | CVE-2018-11236: Very long pathname arguments to realpath function could | |
3086 | result in an integer overflow and buffer overflow. Reported by Alexey | |
3087 | Izbyshev. | |
3088 | ||
3089 | CVE-2018-11237: The mempcpy implementation for the Intel Xeon Phi | |
3090 | architecture could write beyond the target buffer, resulting in a buffer | |
3091 | overflow. Reported by Andreas Schwab. | |
3092 | ||
3093 | The following bugs are resolved with this release: | |
3094 | ||
3095 | [1190] stdio: fgetc()/fread() behaviour is not POSIX compliant | |
3096 | [6889] manual: 'PWD' mentioned but not specified | |
3097 | [13575] libc: SSIZE_MAX defined as LONG_MAX is inconsistent with ssize_t, | |
3098 | when __WORDSIZE != 64 | |
3099 | [13762] regex: re_search etc. should return -2 on memory exhaustion | |
3100 | [13888] build: /tmp usage during testing | |
3101 | [13932] math: dbl-64 pow unexpectedly slow for some inputs | |
3102 | [14092] nptl: Support C11 threads | |
3103 | [14095] localedata: Review / update collation data from Unicode / ISO | |
3104 | 14651 | |
3105 | [14508] libc: -Wformat warnings | |
3106 | [14553] libc: Namespace pollution loff_t in sys/types.h | |
3107 | [14890] libc: Make NT_PRFPREG canonical. | |
3108 | [15105] libc: Extra PLT references with -Os | |
3109 | [15512] libc: __bswap_constant_16 not compiled when -Werror -Wsign- | |
3110 | conversion is given | |
3111 | [16335] manual: Feature test macro documentation incomplete and out of | |
3112 | date | |
3113 | [16552] libc: Unify umount implementations in terms of umount2 | |
3114 | [17082] libc: htons et al.: statement-expressions prevent use on global | |
3115 | scope with -O1 and higher | |
3116 | [17343] libc: Signed integer overflow in /stdlib/random_r.c | |
3117 | [17438] localedata: pt_BR: wrong d_fmt delimiter | |
3118 | [17662] libc: please implement binding for the new renameat2 syscall | |
3119 | [17721] libc: __restrict defined as /* Ignore */ even in c11 | |
3120 | [17979] libc: inconsistency between uchar.h and stdint.h | |
3121 | [18018] dynamic-link: Additional $ORIGIN handling issues (CVE-2011-0536) | |
3122 | [18023] libc: extend_alloca is broken (questionable pointer comparison, | |
3123 | horrible machine code) | |
3124 | [18124] libc: hppa: setcontext erroneously returns -1 as exit code for | |
3125 | last constant. | |
3126 | [18471] libc: llseek should be a compat symbol | |
3127 | [18473] soft-fp: [powerpc-nofpu] __sqrtsf2, __sqrtdf2 should be compat | |
3128 | symbols | |
3129 | [18991] nss: nss_files skips large entry in database | |
3130 | [19239] libc: Including stdlib.h ends up with macros major and minor being | |
3131 | defined | |
3132 | [19463] libc: linknamespace failures when compiled with -Os | |
3133 | [19485] localedata: csb_PL: Update month translations + add yesstr/nostr | |
3134 | [19527] locale: Normalized charset name not recognized by setlocale | |
3135 | [19667] string: Missing Sanity Check for malloc calls in file 'testcopy.c' | |
3136 | [19668] libc: Missing Sanity Check for malloc() in file 'tst-setcontext- | |
3137 | fpscr.c' | |
3138 | [19728] network: out of bounds stack read in libidn function | |
3139 | idna_to_ascii_4i (CVE-2016-6261) | |
3140 | [19729] network: out of bounds heap read on invalid utf-8 inputs in | |
3141 | stringprep_utf8_nfkc_normalize (CVE-2016-6263) | |
3142 | [19818] dynamic-link: Absolute (SHN_ABS) symbols incorrectly relocated by | |
3143 | the base address | |
3144 | [20079] libc: Add SHT_X86_64_UNWIND to elf.h | |
3145 | [20251] libc: 32bit programs pass garbage in struct flock for OFD locks | |
3146 | [20419] dynamic-link: files with large allocated notes crash in | |
3147 | open_verify | |
3148 | [20530] libc: bswap_16 should use __builtin_bswap16() when available | |
3149 | [20890] dynamic-link: ldconfig: fsync the files before atomic rename | |
3150 | [20980] manual: CFLAGS environment variable replaces vital options | |
3151 | [21163] regex: Assertion failure in pop_fail_stack when executing a | |
3152 | malformed regexp (CVE-2015-8985) | |
3153 | [21234] manual: use of CFLAGS makes glibc detect no optimization | |
3154 | [21269] dynamic-link: i386 sigaction sa_restorer handling is wrong | |
3155 | [21313] build: Compile Error GCC 5.4.0 MIPS with -0S | |
3156 | [21314] build: Compile Error GCC 5.2.0 MIPS with -0s | |
3157 | [21508] locale: intl/tst-gettext failure with latest msgfmt | |
3158 | [21547] localedata: Tibetan script collation broken (Dzongkha and Tibetan) | |
3159 | [21812] network: getifaddrs() returns entries with ifa_name == NULL | |
3160 | [21895] libc: ppc64 setjmp/longjmp not fully interoperable with static | |
3161 | dlopen | |
3162 | [21942] dynamic-link: _dl_dst_substitute incorrectly handles $ORIGIN: with | |
3163 | AT_SECURE=1 | |
3164 | [22241] localedata: New locale: Yakut (Sakha) locale for Russia (sah_RU) | |
3165 | [22247] network: Integer overflow in the decode_digit function in | |
3166 | puny_decode.c in libidn (CVE-2017-14062) | |
3167 | [22342] nscd: NSCD not properly caching netgroup | |
3168 | [22391] nptl: Signal function clear NPTL internal symbols inconsistently | |
3169 | [22550] localedata: es_ES locale (and other es_* locales): collation | |
3170 | should treat ñ as a primary different character, sync the collation | |
3171 | for Spanish with CLDR | |
3172 | [22638] dynamic-link: sparc: static binaries are broken if glibc is built | |
3173 | by gcc configured with --enable-default-pie | |
3174 | [22639] time: year 2039 bug for localtime etc. on 64-bit platforms | |
3175 | [22644] string: memmove-sse2-unaligned on 32bit x86 produces garbage when | |
3176 | crossing 2GB threshold (CVE-2017-18269) | |
3177 | [22646] localedata: redundant data (LC_TIME) for es_CL, es_CU, es_EC and | |
3178 | es_BO | |
3179 | [22735] time: Misleading typo in time.h source comment regarding | |
3180 | CLOCKS_PER_SECOND | |
3181 | [22753] libc: preadv2/pwritev2 fallback code should handle offset=-1 | |
3182 | [22761] libc: No trailing `%n' conversion specifier in FMT passed from | |
3183 | `__assert_perror_fail ()' to `__assert_fail_base ()' | |
3184 | [22766] libc: all glibc internal dlopen should use RTLD_NOW for robust | |
3185 | dlopen failures | |
3186 | [22786] libc: Stack buffer overflow in realpath() if input size is close | |
3187 | to SSIZE_MAX (CVE-2018-11236) | |
3188 | [22787] dynamic-link: _dl_check_caller returns false when libc is linked | |
3189 | through an absolute DT_NEEDED path | |
3190 | [22792] build: tcb-offsets.h dependency dropped | |
3191 | [22797] libc: pkey_get() uses non-reserved name of argument | |
3192 | [22807] libc: PTRACE_* constants missing for powerpc | |
3193 | [22818] glob: posix/tst-glob_lstat_compat failure on alpha | |
3194 | [22827] dynamic-link: RISC-V ELF64 parser mis-reads flag in ldconfig | |
3195 | [22830] malloc: malloc_stats doesn't restore cancellation state on stderr | |
3196 | [22848] localedata: ca_ES: update date definitions from CLDR | |
3197 | [22862] build: _DEFAULT_SOURCE is defined even when _ISOC11_SOURCE is | |
3198 | [22884] math: RISCV fmax/fmin handle signalling NANs incorrectly | |
3199 | [22896] localedata: Update locale data for an_ES | |
3200 | [22902] math: float128 test failures with GCC 8 | |
3201 | [22918] libc: multiple common of `__nss_shadow_database' | |
3202 | [22919] libc: sparc32: backtrace yields infinite backtrace with | |
3203 | makecontext | |
3204 | [22926] libc: FTBFS on powerpcspe | |
3205 | [22932] localedata: lt_LT: Update of abbreviated month names from CLDR | |
3206 | required | |
3207 | [22937] localedata: Greek (el_GR, el_CY) locales actually need ab_alt_mon | |
3208 | [22947] libc: FAIL: misc/tst-preadvwritev2 | |
3209 | [22963] localedata: cs_CZ: Add alternative month names | |
3210 | [22987] math: [powerpc/sparc] fdim inlines errno, exceptions handling | |
3211 | [22996] localedata: change LC_PAPER to en_US in es_BO locale | |
3212 | [22998] dynamic-link: execstack tests are disabled when SELinux is | |
3213 | disabled | |
3214 | [23005] network: Crash in __res_context_send after memory allocation | |
3215 | failure | |
3216 | [23007] math: strtod cannot handle -nan | |
3217 | [23024] nss: getlogin_r is performing NSS lookups when loginid isn't set | |
3218 | [23036] regex: regex equivalence class regression | |
3219 | [23037] libc: initialize msg_flags to zero for sendmmsg() calls | |
3220 | [23069] libc: sigaction broken on riscv64-linux-gnu | |
3221 | [23094] localedata: hr_HR: wrong thousands_sep and mon_thousands_sep | |
3222 | [23102] dynamic-link: Incorrect parsing of multiple consecutive $variable | |
3223 | patterns in runpath entries (e.g. $ORIGIN$ORIGIN) | |
3224 | [23137] nptl: s390: pthread_join sometimes block indefinitely (on 31bit | |
3225 | and libc build with -Os) | |
3226 | [23140] localedata: More languages need two forms of month names | |
3227 | [23145] libc: _init/_fini aren't marked as hidden | |
3228 | [23152] localedata: gd_GB: Fix typo in "May" (abbreviated) | |
3229 | [23171] math: C++ iseqsig for long double converts arguments to double | |
3230 | [23178] nscd: sudo will fail when it is run in concurrent with commands | |
3231 | that changes /etc/passwd | |
3232 | [23196] string: __mempcpy_avx512_no_vzeroupper mishandles large copies | |
3233 | (CVE-2018-11237) | |
3234 | [23206] dynamic-link: static-pie + dlopen breaks debugger interaction | |
3235 | [23208] localedata: New locale - Lower Sorbian (dsb) | |
3236 | [23233] regex: Memory leak in build_charclass_op function in file | |
3237 | posix/regcomp.c | |
3238 | [23236] stdio: Harden function pointers in _IO_str_fields | |
3239 | [23250] nptl: Offset of __private_ss differs from GCC | |
3240 | [23253] math: tgamma test suite failures on i686 with -march=x86-64 | |
3241 | -mtune=generic -mfpmath=sse | |
3242 | [23259] dynamic-link: Unsubstituted ${ORIGIN} remains in DT_NEEDED for | |
3243 | AT_SECURE | |
3244 | [23264] libc: posix_spawnp wrongly executes ENOEXEC in non compat mode | |
3245 | [23266] nis: stringop-truncation warning with new gcc8.1 in nisplus- | |
3246 | parser.c | |
3247 | [23272] math: fma(INFINITY,INFIITY,0.0) should be INFINITY | |
3248 | [23277] math: nan function should not have const attribute | |
3249 | [23279] math: scanf and strtod wrong for some hex floating-point | |
3250 | [23280] math: wscanf rounds wrong; wcstod is ok for negative numbers and | |
3251 | directed rounding | |
3252 | [23290] localedata: IBM273 is not equivalent to ISO-8859-1 | |
3253 | [23303] build: undefined reference to symbol | |
3254 | '__parse_hwcap_and_convert_at_platform@@GLIBC_2.23' | |
3255 | [23307] dynamic-link: Absolute symbols whose value is zero ignored in | |
3256 | lookup | |
3257 | [23313] stdio: libio vtables validation and standard file object | |
3258 | interposition | |
3259 | [23329] libc: The __libc_freeres infrastructure is not properly run across | |
3260 | DSO boundaries. | |
3261 | [23349] libc: Various glibc headers no longer compatible with | |
3262 | <linux/time.h> | |
3263 | [23351] malloc: Remove unused code related to heap dumps and malloc | |
3264 | checking | |
3265 | [23363] stdio: stdio-common/tst-printf.c has non-free license | |
3266 | [23396] regex: Regex equivalence regression in single-byte locales | |
3267 | [23422] localedata: oc_FR: More updates of locale data | |
3268 | [23442] build: New warning with GCC 8 | |
3269 | [23448] libc: Out of bounds access in IBM-1390 converter | |
3270 | [23456] libc: Wrong index_cpu_LZCNT | |
3271 | [23458] build: tst-get-cpu-features-static isn't added to tests | |
3272 | [23459] libc: COMMON_CPUID_INDEX_80000001 isn't populated for Intel | |
3273 | processors | |
3274 | [23467] dynamic-link: x86/CET: A property note parser bug | |
3275 | ||
3276 | \f | |
3277 | Version 2.27 | |
3278 | ||
3279 | Major new features: | |
3280 | ||
3281 | * The GNU C Library can now be compiled with support for building static | |
3282 | PIE executables (See --enable-static-pie in INSTALL). These static PIE | |
3283 | executables are like static executables but can be loaded at any address | |
3284 | and provide additional security hardening benefits at the cost of some | |
3285 | memory and performance. When the library is built with --enable-static-pie | |
3286 | the resulting libc.a is usable with GCC 8 and above to create static PIE | |
3287 | executables using the GCC option '-static-pie'. This feature is currently | |
3288 | supported on i386, x86_64 and x32 with binutils 2.29 or later, and on | |
3289 | aarch64 with binutils 2.30 or later. | |
3290 | ||
3291 | * Optimized x86-64 asin, atan2, exp, expf, log, pow, atan, sin, cosf, | |
3292 | sinf, sincosf and tan with FMA, contributed by Arjan van de Ven and | |
3293 | H.J. Lu from Intel. | |
3294 | ||
3295 | * Optimized x86-64 trunc and truncf for processors with SSE4.1. | |
3296 | ||
3297 | * Optimized generic expf, exp2f, logf, log2f, powf, sinf, cosf and sincosf. | |
3298 | ||
3299 | * In order to support faster and safer process termination the malloc API | |
3300 | family of functions will no longer print a failure address and stack | |
3301 | backtrace after detecting heap corruption. The goal is to minimize the | |
3302 | amount of work done after corruption is detected and to avoid potential | |
3303 | security issues in continued process execution. Reducing shutdown time | |
3304 | leads to lower overall process restart latency, so there is benefit both | |
3305 | from a security and performance perspective. | |
3306 | ||
3307 | * The abort function terminates the process immediately, without flushing | |
3308 | stdio streams. Previous glibc versions used to flush streams, resulting | |
3309 | in deadlocks and further data corruption. This change also affects | |
3310 | process aborts as the result of assertion failures. | |
3311 | ||
3312 | * On platforms where long double has the IEEE binary128 format (aarch64, | |
3313 | alpha, mips64, riscv, s390 and sparc), the math library now implements | |
3314 | _Float128 interfaces for that type, as defined by ISO/IEC TS 18661-3:2015. | |
3315 | These are the same interfaces added in version 2.26 for some platforms where | |
3316 | this format is supported but is not the format of long double. | |
3317 | ||
3318 | * On platforms with support for _Float64x (aarch64, alpha, i386, ia64, | |
3319 | mips64, powerpc64le, riscv, s390, sparc and x86_64), the math library now | |
3320 | implements interfaces for that type, as defined by ISO/IEC TS | |
3321 | 18661-3:2015. These are corresponding interfaces to those supported for | |
3322 | _Float128. | |
3323 | ||
3324 | * The math library now implements interfaces for the _Float32, _Float64 and | |
3325 | _Float32x types, as defined by ISO/IEC TS 18661-3:2015. These are | |
3326 | corresponding interfaces to those supported for _Float128. | |
3327 | ||
3328 | * glibc now implements the memfd_create and mlock2 functions on Linux. | |
3329 | ||
3330 | * Support for memory protection keys was added. The <sys/mman.h> header now | |
3331 | declares the functions pkey_alloc, pkey_free, pkey_mprotect, pkey_set, | |
3332 | pkey_get. | |
3333 | ||
3334 | * The copy_file_range function was added. | |
3335 | ||
3336 | * Optimized memcpy, mempcpy, memmove, and memset for sparc M7. | |
3337 | ||
3338 | * The ldconfig utility now processes `include' directives using the C/POSIX | |
3339 | collation ordering. Previous glibc versions used locale-specific | |
3340 | ordering, the change might break systems that relied on that. | |
3341 | ||
3342 | * Support for two grammatical forms of month names has been added. | |
3343 | In a call to strftime, the "%B" and "%b" format specifiers will now | |
3344 | produce the grammatical form required when the month is used as part | |
3345 | of a complete date. New "%OB" and "%Ob" specifiers produce the form | |
3346 | required when the month is named by itself. For instance, in Greek | |
3347 | and in many Slavic and Baltic languages, "%B" will produce the month | |
3348 | in genitive case, and "%OB" will produce the month in nominative case. | |
3349 | ||
3350 | In a call to strptime, "%B", "%b", "%h", "%OB", "%Ob", and "%Oh" | |
3351 | are all valid and will all accept any known form of month | |
3352 | name---standalone or complete, abbreviated or full. In a call to | |
3353 | nl_langinfo, the query constants MON_1..12 and ABMON_1..12 return | |
3354 | the strings used by "%B" and "%b", respectively. New query | |
3355 | constants ALTMON_1..12 and _NL_ABALTMON_1..12 return the strings | |
3356 | used by "%OB" and "%Ob", respectively. | |
3357 | ||
3358 | In a locale definition file, use "alt_mon" and "ab_alt_mon" to | |
3359 | define the strings for %OB and %Ob, respectively; these have the | |
3360 | same syntax as "mon" and "abmon". These arrays are optional; if they | |
3361 | are not provided then they have the same content as "mon" and "abmon", | |
3362 | respectively. | |
3363 | ||
3364 | These features are provided for locales which define "alt_mon" and/or | |
3365 | "ab_alt_mon" in their locale source data. This release includes such | |
3366 | alternative month name data for the following languages: Belarusian, | |
3367 | Croatian, Greek, Lithuanian, Polish, Russian, and Ukrainian. | |
3368 | ||
3369 | This feature is currently a GNU extension, but it is expected to | |
3370 | be added to the next revision of POSIX, and it is also already | |
3371 | available on some BSD-derived operating systems. | |
3372 | ||
3373 | This feature will cause existing statically compiled applications | |
3374 | to fail to load locales and fall back to the builtin C/POSIX locales. | |
3375 | See notes below for other changes affecting compatibility. | |
3376 | ||
3377 | * Support for the RISC-V ISA running on Linux has been added. This port | |
3378 | requires at least binutils-2.30, gcc-7.3.0, and linux-4.15; and is supported | |
3379 | for the following ISA and ABI pairs: | |
3380 | ||
3381 | - rv64imac lp64 | |
3382 | - rv64imafdc lp64 | |
3383 | - rv64imafdc lp64d | |
3384 | ||
3385 | Deprecated and removed features, and other changes affecting compatibility: | |
3386 | ||
3387 | * Statically compiled applications attempting to load locales compiled for the | |
3388 | GNU C Library version 2.27 will fail and fall back to the builtin C/POSIX | |
3389 | locale. The reason for this is that the addition of the new "%OB" and "%Ob", | |
3390 | support for two grammatical forms of the month names, also extends the locale | |
3391 | data binary format. Static applications needing locale support must be | |
3392 | recompiled to match the runtime and data they are deployed with. In some | |
3393 | distributions there is an upgrade window where dynamically linked applications | |
3394 | may use a new library but the old locale data and also fall back to the | |
3395 | builtin C/POSIX locales; restarting the application process is sufficient to | |
3396 | fix this. | |
3397 | ||
3398 | * Support for statically linked applications which call dlopen is deprecated | |
3399 | and will be removed in a future version of glibc. Applications which call | |
3400 | dlopen need to be linked dynamically instead. | |
3401 | ||
3402 | * Support for old programs which use internal stdio data structures and | |
3403 | functions is deprecated. This includes programs which use the C++ streams | |
3404 | provided by libstdc++ in GCC 2.95. Programs which use the internal | |
3405 | symbols _IO_adjust_wcolumn, _IO_default_doallocate, _IO_default_finish, | |
3406 | _IO_default_pbackfail, _IO_default_uflow, _IO_default_xsgetn, | |
3407 | _IO_default_xsputn, _IO_doallocbuf, _IO_do_write, _IO_file_attach, | |
3408 | _IO_file_close, _IO_file_close_it, _IO_file_doallocate, _IO_file_fopen, | |
3409 | _IO_file_init, _IO_file_jumps, _IO_fileno, _IO_file_open, | |
3410 | _IO_file_overflow, _IO_file_read, _IO_file_seek, _IO_file_seekoff, | |
3411 | _IO_file_setbuf, _IO_file_stat, _IO_file_sync, _IO_file_underflow, | |
3412 | _IO_file_write, _IO_file_xsputn, _IO_flockfile, _IO_flush_all, | |
3413 | _IO_flush_all_linebuffered, _IO_free_backup_area, _IO_free_wbackup_area, | |
3414 | _IO_init, _IO_init_marker, _IO_init_wmarker, _IO_iter_begin, _IO_iter_end, | |
3415 | _IO_iter_file, _IO_iter_next, _IO_least_wmarker, _IO_link_in, | |
3416 | _IO_list_all, _IO_list_lock, _IO_list_resetlock, _IO_list_unlock, | |
3417 | _IO_marker_delta, _IO_marker_difference, _IO_remove_marker, _IO_seekmark, | |
3418 | _IO_seekwmark, _IO_str_init_readonly, _IO_str_init_static, | |
3419 | _IO_str_overflow, _IO_str_pbackfail, _IO_str_seekoff, _IO_str_underflow, | |
3420 | _IO_switch_to_main_wget_area, _IO_switch_to_wget_mode, | |
3421 | _IO_unsave_wmarkers, _IO_wdefault_doallocate, _IO_wdefault_finish, | |
3422 | _IO_wdefault_pbackfail, _IO_wdefault_setbuf, _IO_wdefault_uflow, | |
3423 | _IO_wdefault_xsgetn, _IO_wdefault_xsputn, _IO_wdoallocbuf, _IO_wdo_write, | |
3424 | _IO_wfile_jumps, _IO_wfile_overflow, _IO_wfile_sync, _IO_wfile_underflow, | |
3425 | _IO_wfile_xsputn, _IO_wmarker_delta, or _IO_wsetb may stop working with a | |
3426 | future version of glibc. Unlike other symbol removals, these old | |
3427 | applications will not be supported using compatibility symbols. | |
3428 | ||
3429 | * On GNU/Linux, the obsolete Linux constant PTRACE_SEIZE_DEVEL is no longer | |
3430 | defined by <sys/ptrace.h>. | |
3431 | ||
3432 | * libm no longer supports SVID error handling (calling a user-provided | |
3433 | matherr function on error) or the _LIB_VERSION variable to control error | |
3434 | handling. (SVID error handling and the _LIB_VERSION variable still work | |
3435 | for binaries linked against older versions of the GNU C Library.) The | |
3436 | libieee.a library is no longer provided. math.h no longer defines struct | |
3437 | exception, or the macros X_TLOSS, DOMAIN, SING, OVERFLOW, UNDERFLOW, | |
3438 | TLOSS, PLOSS and HUGE. | |
3439 | ||
3440 | * The libm functions pow10, pow10f and pow10l are no longer supported for | |
3441 | new programs. Programs should use the standard names exp10, exp10f and | |
3442 | exp10l for these functions instead. | |
3443 | ||
3444 | * The mcontext_t type is no longer the same as struct sigcontext. On | |
3445 | platforms where it was previously the same, this changes the C++ name | |
3446 | mangling for interfaces involving this type. | |
3447 | ||
3448 | * The add-ons mechanism for building additional packages at the same time as | |
3449 | glibc has been removed. The --enable-add-ons configure option is now | |
3450 | ignored. | |
3451 | ||
3452 | * The --without-fp configure option is now ignored. Whether hardware | |
3453 | floating-point instructions are used is now configured based on whether | |
3454 | the compiler used at configure time (without any options implied by a | |
3455 | --with-cpu= configure option) uses such instructions. | |
3456 | ||
3457 | * The res_hnok, res_dnok, res_mailok and res_ownok functions now check that | |
3458 | the specified string can be parsed as a domain name. | |
3459 | ||
3460 | * In the malloc_info output, the <heap> element may contain another <aspace> | |
3461 | element, "subheaps", which contains the number of sub-heaps. | |
3462 | ||
3463 | * The libresolv function p_secstodate is no longer supported for new | |
3464 | programs. | |
3465 | ||
3466 | * The tilepro-*-linux-gnu configuration is no longer supported. | |
3467 | ||
3468 | * The nonstandard header files <libio.h> and <_G_config.h> are deprecated | |
3469 | and will be removed in a future release. Software that is still using | |
3470 | either header should be updated to use standard <stdio.h> interfaces | |
3471 | instead. | |
3472 | ||
3473 | libio.h was originally the header for a set of supported GNU extensions, | |
3474 | but they have not been maintained as such in many years, they are now | |
3475 | standing in the way of improvements to stdio, and we don't think there are | |
3476 | any remaining external users. _G_config.h was never intended for public | |
3477 | use, but predates the bits convention. | |
3478 | ||
3479 | Changes to build and runtime requirements: | |
3480 | ||
3481 | * bison version 2.7 or later is required to generate code in the 'intl' | |
3482 | subdirectory. | |
3483 | ||
3484 | Security related changes: | |
3485 | ||
3486 | CVE-2009-5064: The ldd script would sometimes run the program under | |
3487 | examination directly, without preventing code execution through the | |
3488 | dynamic linker. (The glibc project disputes that this is a security | |
3489 | vulnerability; only trusted binaries must be examined using the ldd | |
3490 | script.) | |
3491 | ||
3492 | CVE-2017-15670: The glob function, when invoked with GLOB_TILDE, | |
3493 | suffered from a one-byte overflow during ~ operator processing (either | |
3494 | on the stack or the heap, depending on the length of the user name). | |
3495 | Reported by Tim Rühsen. | |
3496 | ||
3497 | CVE-2017-15671: The glob function, when invoked with GLOB_TILDE, | |
3498 | would sometimes fail to free memory allocated during ~ operator | |
3499 | processing, leading to a memory leak and, potentially, to a denial | |
3500 | of service. | |
3501 | ||
3502 | CVE-2017-15804: The glob function, when invoked with GLOB_TILDE and | |
3503 | without GLOB_NOESCAPE, could write past the end of a buffer while | |
3504 | unescaping user names. Reported by Tim Rühsen. | |
3505 | ||
3506 | CVE-2017-17426: The malloc function, when called with an object size near | |
3507 | the value SIZE_MAX, would return a pointer to a buffer which is too small, | |
3508 | instead of NULL. This was a regression introduced with the new malloc | |
3509 | thread cache in glibc 2.26. Reported by Iain Buclaw. | |
3510 | ||
3511 | CVE-2017-1000408: Incorrect array size computation in _dl_init_paths leads | |
3512 | to the allocation of too much memory. (This is not a security bug per se, | |
3513 | it is mentioned here only because of the CVE assignment.) Reported by | |
3514 | Qualys. | |
3515 | ||
3516 | CVE-2017-1000409: Buffer overflow in _dl_init_paths due to miscomputation | |
3517 | of the number of search path components. (This is not a security | |
3518 | vulnerability per se because no trust boundary is crossed if the fix for | |
3519 | CVE-2017-1000366 has been applied, but it is mentioned here only because | |
3520 | of the CVE assignment.) Reported by Qualys. | |
3521 | ||
3522 | CVE-2017-16997: Incorrect handling of RPATH or RUNPATH containing $ORIGIN | |
3523 | for AT_SECURE or SUID binaries could be used to load libraries from the | |
3524 | current directory. | |
3525 | ||
3526 | CVE-2018-1000001: Buffer underflow in realpath function when getcwd function | |
3527 | succeeds without returning an absolute path due to unexpected behaviour | |
3528 | of the Linux kernel getcwd syscall. Reported by halfdog. | |
3529 | ||
3530 | CVE-2018-6485: The posix_memalign and memalign functions, when called with | |
3531 | an object size near the value of SIZE_MAX, would return a pointer to a | |
3532 | buffer which is too small, instead of NULL. Reported by Jakub Wilk. | |
3533 | ||
3534 | CVE-2018-6551: The malloc function, when called with an object size near | |
3535 | the value of SIZE_MAX, would return a pointer to a buffer which is too | |
3536 | small, instead of NULL. | |
3537 | ||
3538 | The following bugs are resolved with this release: | |
3539 | ||
3540 | [866] glob: glob should match dangling symlinks | |
3541 | [1062] glob: sysdeps/generic/glob.c merge from gnulib (part 3 of 3) | |
3542 | [2522] localedata: ca_ES@valencia: new Valencian (meridional Catalan) | |
3543 | locale | |
3544 | [5997] math: Very slow execution of sinf function | |
3545 | [10580] localedata: hr_HR: updated locale | |
3546 | [10871] locale: 'mon' array should contain both nominative and genitive | |
3547 | cases | |
3548 | [12349] localedata: eu_ES: incorrect thousands separator | |
3549 | [13605] localedata: shn_MM: new Shan locale | |
3550 | [13805] localedata: ru_RU: currency should use ',' as radix point | |
3551 | [13953] localedata: km_KH: locale update | |
3552 | [13994] localedata: mjw_IN: new locale | |
3553 | [14121] build: make writes .mo files in po directory | |
3554 | [14333] libc: Fix the race between atexit() and exit() | |
3555 | [14681] dynamic-link: _dl_get_origin leaks memory via executable link map. | |
3556 | [14925] localedata: bn_*: LC_IDENTIFICATION.language key should be | |
3557 | "Bangla" | |
3558 | [15260] localedata: LC_MESSAGES.{yes,no}{str,expr}: various errors | |
3559 | [15261] localedata: LC_MESSAGES.yesexpr/noexpr: inconsistent use of full- | |
3560 | width Latin characters | |
3561 | [15332] localedata: es_CU: locale update | |
3562 | [15436] stdio: Don't close or flush stdio streams on abort | |
3563 | [15537] localedata: lv_LV: invalid collation for Latvian diacritical | |
3564 | letters | |
3565 | [16148] localedata: ca_ES: incorrect thousands separator | |
3566 | [16750] dynamic-link: ldd should not try to execute the binaries | |
3567 | (CVE-2009-5064) | |
3568 | [16777] localedata: pl_PL: incorrect thousands separator in locale | |
3569 | [16905] localedata: hanzi: new collation | |
3570 | [17563] localedata: cmn_TW: add hanzi collation | |
3571 | [17750] localedata: wrong collation order of diacritics in most locales | |
3572 | [17804] libc: scandirat fails with ENOMEM because it checks for errno even | |
3573 | if malloc succeeded | |
3574 | [17956] build: Build fails on missing definitions from header file | |
3575 | nss/nss.h when Mozilla NSS is used for cryptography | |
3576 | [18203] libc: realpath() does not handle unreachable paths correctly | |
3577 | [18572] dynamic-link: [arm] Lazy TLSDESC relocation has data race | |
3578 | [18812] localedata: kab_DZ: new Kabyle Algeria locale | |
3579 | [18822] libc: Internal functions are called via PLT | |
3580 | [18858] string: _HAVE_STRING_ARCH_xxx aren't defined for i386 nor x86_64 | |
3581 | [19170] libc: __gmon_start__ defined in hppa in crtn.S | |
3582 | [19574] libc: glibc should support building static PIE binaries | |
3583 | [19852] localedata: charmaps/UTF-8: incorrect wcwidth for U+3099 and | |
3584 | U+309A | |
3585 | [19971] glob: glob: Do not skip entries with zero d_ino values | |
3586 | [19982] localedata: fr.po: spelling mistake for error code EXDEV | |
3587 | [20008] localedata: km_KH: convert to translit_neutral | |
3588 | [20009] localedata: tr_TR: convert LC_CTYPE to i18n | |
3589 | [20142] math: [x86_64] Add SSE4.1 trunc, truncf | |
3590 | [20204] dynamic-link: _dl_open_hook and _dlfcn_hook hardening | |
3591 | [20482] localedata: de_CH: abbreviated weekdays should be two letters | |
3592 | [20498] localedata: miq_NI: new Mískitu / Miskito (miq) language locale | |
3593 | for Nicaragua | |
3594 | [20532] nss: getaddrinfo uses errno and h_errno without guaranteeing | |
3595 | they're set, wrong errors returned by gaih_inet when lookup functions | |
3596 | are not found. | |
3597 | [20756] localedata: [PATCH] Use Unicode wise thousands separator | |
3598 | [20826] network: posix/tst-getaddrinfo5 fails on hosts without network | |
3599 | access | |
3600 | [20952] localedata: yuw_PG: new locale | |
3601 | [21084] localedata: charmaps/IBM858: new codepage | |
3602 | [21161] manual: [PATCH] fix typo in manual/arith.texi on strtoul prototype | |
3603 | [21242] libc: assert gives pedantic warning in old gcc versions | |
3604 | [21265] dynamic-link: _dl_runtime_resolve isn't compatible with Intel C++ | |
3605 | __regcall calling convention | |
3606 | [21309] math: signed integer overflow in sysdeps/ieee754/dbl-64/e_pow.c | |
3607 | [21326] libc: C99 functions are not declared for C++11 and later if | |
3608 | _GNU_SOURCE is not predefined by g++ | |
3609 | [21457] libc: sys/ucontext.h namespace | |
3610 | [21530] libc: tmpfile() should be implemented using O_TMPFILE | |
3611 | [21660] math: GCC fails to compile a formula with tgmath.h | |
3612 | [21672] nptl: sys-libs/glibc on ia64 crashes on thread exit: signal | |
3613 | SIGSEGV, Segmentation fault: pthread_create.c:432: __madvise | |
3614 | (pd->stackblock, freesize - PTHREAD_STACK_MIN, MADV_DONTNEED); | |
3615 | [21684] math: tgmath.h handling of complex integers | |
3616 | [21685] math: tgmath.h handling of bit-fields | |
3617 | [21686] math: tgmath.h handling of __int128 | |
3618 | [21706] localedata: yesstr and nostr are missing for Breton [LC_MESSAGES] | |
3619 | locale | |
3620 | [21745] libc: [powerpc64le] Extra PLT reference with --enable-stack- | |
3621 | protector=all | |
3622 | [21750] localedata: column width of characters incompatible with classical | |
3623 | wcwidth | |
3624 | [21754] malloc: malloc: Perform as little work as possible after heap | |
3625 | consistency check failures | |
3626 | [21780] libc: hppa: p{read,write}v2 does not set ENOSUP on invalid flag | |
3627 | [21790] libc: Missing __memset_zero_constant_len_parameter in libc.so | |
3628 | [21791] string: Unused XXX_chk_XXX functions in libc.a | |
3629 | [21815] dynamic-link: FAIL: elf/tst-prelink-cmp with GCC is defaulted to | |
3630 | PIE | |
3631 | [21836] localedata: Removed redundant data (LC_MONETARY) in various Indian | |
3632 | locales | |
3633 | [21845] localedata: Added new Locale bho_NP | |
3634 | [21853] localedata: Fix abday Which looks same as day in zh_SG | |
3635 | [21854] localedata: Added New Locale en_SC | |
3636 | [21864] libc: xmalloc.o is compiled with -DMODULE_NAME=libc | |
3637 | [21871] dynamic-link: _dl_runtime_resolve_avx_opt is slower than | |
3638 | _dl_runtime_resolve_avx_slow | |
3639 | [21885] network: getaddrinfo: gethosts does not release resolver context | |
3640 | on memory allocation failure | |
3641 | [21899] libc: XPG4.2 sigaction namespace | |
3642 | [21908] dynamic-link: dynamic linker broke on ia64 (mmap2 consolidation is | |
3643 | the suspect) | |
3644 | [21913] libc: static binaries SIGSEGV in __brk when host's gcc is pie-by- | |
3645 | default (i386) | |
3646 | [21915] nss: nss_files can return with NSS_STATUS_SUCCESS and a clobbered | |
3647 | errno value, causing getaddrinfo to fail | |
3648 | [21920] localedata: Fix p_cs_precedes/n_cs_precedes for mt_MT | |
3649 | [21922] network: getaddrinfo with AF_INET/AF_INET6 returns EAI_NONAME | |
3650 | instead of EAI_NODATA | |
3651 | [21928] libc: sys/ptrace.h: remove obsolete temporary development Linux | |
3652 | constant PTRACE_SEIZE_DEVEL | |
3653 | [21930] math: C-only gcc builtins used in <math.h> isinf | |
3654 | [21932] network: Unpaired __resolv_context_get in generic get*_r | |
3655 | implementation | |
3656 | [21941] math: powerpc: Wrong register constraint for xssqrtqp in sqrtf128 | |
3657 | [21944] libc: sigval namespace | |
3658 | [21951] localedata: Update hanzi collation by stroke | |
3659 | [21955] math: Wrong alignment of L(SP_RANGE)/L(SP_INF_0) in | |
3660 | sysdeps/x86_64/fpu/e_expf.S | |
3661 | [21956] libc: Stack allocation in MIPS syscall impl (ubounded stack | |
3662 | allocation in syscall loops) | |
3663 | [21959] localedata: Fix Country name for xh_ZA | |
3664 | [21960] localedata: Fix abmon for bem_ZM | |
3665 | [21966] math: AVX2 mathvec functions use FMA without checking | |
3666 | [21967] math: When 512-bit AVX2 wrapper functions in mathvec are used? | |
3667 | [21971] localedata: Added New Locale for mfe_MU | |
3668 | [21972] libc: assert macro requires operator== (int) for its argument type | |
3669 | [21973] math: [sparc] libm missing sqrtl compat symbol | |
3670 | [21974] libc: Remove __bb_init_func and __bb_exit_func | |
3671 | [21982] string: stratcliff.c: error: assuming signed overflow does not | |
3672 | occur with -O3 | |
3673 | [21986] stdio: __guess_grouping is called incorrectly | |
3674 | [21987] math: [sparc32] wrong bits/long-double.h installed | |
3675 | [22019] localedata: Wrong placement of monetary symbol in el_GR (negative | |
3676 | amounts) | |
3677 | [22022] localedata: Missing country_name for mni_IN | |
3678 | [22023] localedata: Removed redundant data (LC_TIME and LC_MESSAGES) for | |
3679 | niu_NZ | |
3680 | [22025] locale: iconv: Inconsistency between pointer mangling and NULL | |
3681 | checks | |
3682 | [22026] locale: iconv_open: heap overflow on gconv_init failure | |
3683 | [22028] math: bits/math-finite.h _MSUF_ expansion namespace | |
3684 | [22035] math: [m68k] bits/math-inline.h macro namespace | |
3685 | [22038] localedata: Fix abbreviated weeks and months for Somali | |
3686 | [22044] localedata: Remove redundant data for Limburgish Language | |
3687 | [22050] malloc: Linking with -lmcheck does not hook | |
3688 | __malloc_initialize_hook correctly | |
3689 | [22051] libc: zero terminator in the middle of glibc's .eh_frame | |
3690 | [22052] malloc: malloc failed to compile with GCC 7 and -O3 | |
3691 | [22070] localedata: charmaps/UTF-8: wcwidth for | |
3692 | Prepended_Concatenation_Mark codepoints set to 0 (should be 1) | |
3693 | [22074] localedata: charmaps/UTF-8: wcwidth for U+1160-U+11FF (Hangul | |
3694 | Jungseong and Jongseong) should be 0 | |
3695 | [22078] nss: nss_files performance issue in multi mode | |
3696 | [22082] math: bits/math-finite.h exp10 condition | |
3697 | [22086] libc: pcprofiledump incorrect cross-endian condition | |
3698 | [22093] dynamic-link: ld.so no longer searches in .../x86_64 | |
3699 | [22095] network: Name server address allocation memory leak in resolv.conf | |
3700 | parsing after OOM | |
3701 | [22096] network: __resolv_conf_attach can incorrectly free passed conf | |
3702 | object | |
3703 | [22100] localedata: om_KE: LC_TIME: copy redundant data from om_ET | |
3704 | [22101] dynamic-link: Dynamic loader must ignore "debug" shared objects | |
3705 | e.g. ET_GNU_DEBUG_* | |
3706 | [22111] malloc: malloc: per thread cache is not returned when thread exits | |
3707 | [22112] localedata: Fix LC_TELEPHONE/LC_NAME for az_AZ | |
3708 | [22134] libc: [linux] implement fexecve with execveat | |
3709 | [22142] libc: [powerpc] printf outputs a wrong value of DBL_MAX on ppc64 and | |
3710 | ppc64le | |
3711 | [22145] libc: ttyname() gives up too early in the face of namespaces | |
3712 | [22146] math: C++ build issue with float128 on x86_64 | |
3713 | [22153] nptl: nptl: save error code before process termination | |
3714 | [22156] libc: [hppa,ia64,microblaze] Executable stack default | |
3715 | [22159] malloc: malloc: MALLOC_CHECK_ broken with --enable-tunables=no | |
3716 | [22161] nscd: nscd cache prune for netgroups hangs after timeout bump | |
3717 | [22165] libc: [hppa] Text relocations in libc.so | |
3718 | [22180] libc: destructor registered via __cxa_atexit is called twice | |
3719 | [22183] glob: commit 5554304f0ddd ("posix: Allow glob to match dangling | |
3720 | symlinks") cause "make" segfaults | |
3721 | [22189] math: [powerpc] math_private.h definitions of math_opt_barrier and | |
3722 | math_force_eval | |
3723 | [22207] libc: FAIL: stdlib/test-atexit-race | |
3724 | [22225] math: nearbyint arithmetic moved before feholdexcept | |
3725 | [22229] math: [sparc32] missing copysignl, fabsl, fmal compat symbols | |
3726 | [22235] math: iscanonical in C++ and float128 | |
3727 | [22243] math: log2(0) and log10(0) are wrong in downward rounding without | |
3728 | the svid compat wrapper | |
3729 | [22244] math: ynf and yn are wrong without the svid compat wrapper | |
3730 | [22273] libc: Improper assert in Linux posix_spawn implementation | |
3731 | [22284] libc: -pg -pie doesn't work | |
3732 | [22292] locale: localedef exits with error 4 when it should be error 1 | |
3733 | [22294] locale: Allow "" for int_currency_symbol definition in locales. | |
3734 | [22295] locale: Don't warn on non-symbolic characters in locale sources in | |
3735 | --verbose. | |
3736 | [22296] math: glibc 2.26: signbit build issue with Gcc 5.5.0 on x86_64 | |
3737 | [22298] nptl: x32: lockups on recursive pthread_mutex_lock after upgrade | |
3738 | to 2.26 | |
3739 | [22299] dynamic-link: Problem with $PLATFORM on x86_64 platform | |
3740 | [22320] glob: Buffer overflow in glob with GLOB_TILDE (CVE-2017-15670) | |
3741 | [22321] libc: sysconf(_SC_IOV_MAX) returns -1 on Linux | |
3742 | [22322] libc: [mips64] wrong bits/long-double.h installed | |
3743 | [22325] glob: Memory leak in glob with GLOB_TILDE (CVE-2017-15671) | |
3744 | [22332] glob: Buffer overflow in glob with GLOB_TILDE in unescaping | |
3745 | (CVE-2017-15804) | |
3746 | [22336] localedata: cs_CZ LC_COLLATE does not use i18n | |
3747 | [22343] malloc: Integer overflow in posix_memalign | |
3748 | [22347] libc: getrandom() returns the number of bytes that were copied to | |
3749 | the buffer even though the comments say "Return 0 on success and -1 on | |
3750 | failure." | |
3751 | [22353] string: sysdeps/i386/i586/strcpy.S isn't maintainable | |
3752 | [22362] libc: Installed crt1.o, crti,.o and crtn.o files are used with | |
3753 | -m32 | |
3754 | [22370] dynamic-link: Incorrect note padding check | |
3755 | [22375] libc: malloc returns pointer from tcache_get when should return | |
3756 | NULL (CVE-2017-17426) | |
3757 | [22377] math: iseqsig, float128 and C++ | |
3758 | [22382] localedata: Error in tpi_PG locale | |
3759 | [22387] localedata: Replace unicode sequences <Uxxxx> for characters | |
3760 | inside the ASCII printable range | |
3761 | [22402] math: [powerpc64le] __MATH_TG does not support _Float128 for | |
3762 | -mlong-double-64 | |
3763 | [22403] localedata: Slash needs escaping in some locales | |
3764 | [22408] malloc: malloc_info access heaps without arena lock, ignores heaps | |
3765 | [22409] network: res_hnok does not accept some host names used on the | |
3766 | Internet | |
3767 | [22412] network: res_dnok, res_hnok should perform syntax checks | |
3768 | [22413] network: ns_name_pton ignores syntactically invalid trailing | |
3769 | backslash | |
3770 | [22415] stdio: setvbuf can lead to invalid free/segfault | |
3771 | [22432] build: Non-deterministic build | |
3772 | [22439] malloc: malloc_info should compute summary statistics for all sub- | |
3773 | heaps in an arena | |
3774 | [22442] network: if_nametoindex could report index for the wrong | |
3775 | networking interface | |
3776 | [22446] build: aliasing violation calling readlink in handle_request | |
3777 | [22447] build: unsafe call to strlen with a non-string in getlogin_r.c | |
3778 | [22457] libc: Generic preadv/pwritev incorrectly calls __posix_memalign | |
3779 | [22459] libc: FAIL: elf/check-localplt with __stack_chk_fail related to | |
3780 | __nscd_hash/__nss_hash | |
3781 | [22463] network: p_secstodate overflow handling | |
3782 | [22469] localedata: pl_PL LC_COLLATE does not use i18n | |
3783 | [22478] libc: sigwait can fail with EINTR | |
3784 | [22505] libc: ldconfig processes include directive in locale-specific | |
3785 | order | |
3786 | [22515] localedata: hsb_DE LC_COLLATE does not use copy "iso14651_t1" | |
3787 | [22517] localedata: et_EE LC_COLLATE does not use copy "iso14651_t1" | |
3788 | [22519] localedata: is_IS LC_COLLATE does not use copy "iso14651_t1" | |
3789 | [22524] localedata: lt_LT LC_COLLATE does not use copy "iso14651_t1" | |
3790 | [22527] localedata: tr_TR LC_COLLATE does not use copy "iso14651_t1" | |
3791 | [22534] localedata: Collation rules for Serbian and Bosnian should be the | |
3792 | same as for Croatian | |
3793 | [22561] math: [DR#471] cacosh (0 + iNaN) should return NaN +/- i pi/2 | |
3794 | [22568] math: [DR#471] ctanh (0 + iNaN), ctanh (0 + i Inf) | |
3795 | [22577] libc: missing newline after "cannot allocate TLS data structures | |
3796 | for initial thread" | |
3797 | [22588] manual: manual/conf.texi: missing underscore in front of | |
3798 | SC_SSIZE_MAX | |
3799 | [22593] math: nextafter and nexttoward are declared with const attribute | |
3800 | [22596] manual: manual: finite(nan) wrongly described as returning nonzero | |
3801 | [22603] string: ia64 memchr overflows internal pointer check | |
3802 | [22605] libc: SH clone does not set the exit code correctly | |
3803 | [22606] dynamic-link: Incorrect array size computation in _dl_init_paths | |
3804 | (CVE-2017-1000408) | |
3805 | [22607] dynamic-link: Buffer Overflow in _dl_init_paths (CVE-2017-1000409) | |
3806 | [22611] malloc: malloc/tst-realloc wrongly assumes that errno must not be | |
3807 | modified in case of success | |
3808 | [22614] build: gcc: error: unrecognized command line option ‘-no-pie’ | |
3809 | [22615] manual: manual: ambiguous wording about errno value in case of | |
3810 | success | |
3811 | [22624] libc: MIPS setjmp() saves incorrect 'o0' register in --enable- | |
3812 | stack-protector=all | |
3813 | [22625] dynamic-link: RPATH $ORIGIN replaced by PWD for AT_SECURE/SUID | |
3814 | binaries or if /proc is not mounted (CVE-2017-16997) | |
3815 | [22627] dynamic-link: $ORIGIN in $LD_LIBRARY_PATH is substituted twice | |
3816 | [22630] build: $(no-pie-ldflag) is no longer effective | |
3817 | [22631] math: [m68k] Bad const attributes in bits/mathinline.h | |
3818 | [22635] nptl: pthread_self returns NULL before libpthread is loaded | |
3819 | [22636] nptl: PTHREAD_STACK_MIN is too small on x86-64 | |
3820 | [22637] nptl: guard size is subtracted from thread stack size instead of | |
3821 | adding it on top | |
3822 | [22648] libc: getrlimit/setrlimit with RLIM_INFINITY broken on alpha | |
3823 | [22657] localedata: hu_HU: Avoid double space in date | |
3824 | [22660] math: fmax, fmin sNaN handling on alpha | |
3825 | [22664] libc: New warning of GCC8 | |
3826 | [22665] math: alpha: ceil and floor raise inexact exceptions | |
3827 | [22666] math: alpha: trunc raise inexact exceptions | |
3828 | [22667] libc: makecontext lacks stack alignment on i386 | |
3829 | [22678] libc: prlimit fails for RLIM_INFINITY values on 32-bit machines | |
3830 | [22679] libc: getcwd(3) can succeed without returning an absolute path | |
3831 | (CVE-2018-1000001) | |
3832 | [22685] libc: PowerPC: Static AT_SECURE binaries segfault with lock- | |
3833 | elision and tunables | |
3834 | [22687] math: [powerpc-nofpu] complex long double functions spurious | |
3835 | "invalid" exception | |
3836 | [22688] math: [powerpc-nofpu] remainderl wrong sign of zero result | |
3837 | [22690] math: [ldbl-128ibm] lrintl, lroundl missing "invalid" exceptions | |
3838 | [22691] math: [powerpc-nofpu] fmaxmagl, fminmagl spurious "invalid" | |
3839 | exception | |
3840 | [22693] math: [ldbl-128ibm] log1pl (-qNaN) spurious "invalid" exception | |
3841 | [22697] math: [powerpc] llround spurious "inexact" exceptions on 32-bit | |
3842 | power4 | |
3843 | [22701] nis: Incomplete removal of libnsl | |
3844 | [22702] math: [powerpc-nofpu] nearbyintl traps with trapping "inexact" | |
3845 | [22707] libc: Missing defines in elf.h for DF_1_STUB and DF_1_PIE. | |
3846 | [22715] dynamic-link: FAIL: elf/tst-audit10 | |
3847 | [22719] libc: Backtrace tests fail on hppa | |
3848 | [22742] libc: [aarch64] mcontext_t __reserved field got renamed | |
3849 | [22743] nptl: __pthread_register_cancel corrupts stack after f81ddabffd | |
3850 | [22765] crypt: (struct crypt_data *data)->initialized is not set to zero | |
3851 | before the first call to crypt_r () in crypt/badsalttest.c | |
3852 | ||
3853 | \f | |
3854 | Version 2.26 | |
3855 | ||
3856 | Major new features: | |
3857 | ||
3858 | * A per-thread cache has been added to malloc. Access to the cache requires | |
3859 | no locks and therefore significantly accelerates the fast path to allocate | |
3860 | and free small amounts of memory. Refilling an empty cache requires locking | |
3861 | the underlying arena. Performance measurements show significant gains in a | |
3862 | wide variety of user workloads. Workloads were captured using a special | |
3863 | instrumented malloc and analyzed with a malloc simulator. Contributed by | |
3864 | DJ Delorie with the help of Florian Weimer, and Carlos O'Donell. | |
3865 | ||
3866 | * Unicode 10.0.0 Support: Character encoding, character type info, and | |
3867 | transliteration tables are all updated to Unicode 10.0.0, using | |
3868 | generator scripts contributed by Mike FABIAN (Red Hat). | |
3869 | These updates cause user visible changes, especially the changes in | |
3870 | wcwidth for many emoji characters cause problems when emoji sequences | |
3871 | are rendered with pango, see for example: | |
3872 | https://bugzilla.gnome.org/show_bug.cgi?id=780669#c5 | |
3873 | ||
3874 | * Collation of Hungarian has been overhauled and is now consistent with "The | |
3875 | Rules of Hungarian Orthography, 12th edition" (Bug 18934). Contributed by | |
3876 | Egmont Koblinger. | |
3877 | ||
3878 | * Improvements to the DNS stub resolver, contributed by Florian Weimer: | |
3879 | ||
3880 | - The GNU C Library will now detect when /etc/resolv.conf has been | |
3881 | modified and reload the changed configuration. The new resolver option | |
3882 | “no-reload” (RES_NORELOAD) disables this behavior. | |
3883 | ||
3884 | - The GNU C Library now supports an arbitrary number of search domains | |
3885 | (configured using the “search” directive in /etc/resolv.conf); | |
3886 | previously, there was a hard limit of six domains. For backward | |
3887 | compatibility, applications that directly modify the ‘_res’ global | |
3888 | object are still limited to six search domains. | |
3889 | ||
3890 | - When the “rotate” (RES_ROTATE) resolver option is active, the GNU C | |
3891 | Library will now randomly pick a name server from the configuration as a | |
3892 | starting point. (Previously, the second name server was always used.) | |
3893 | ||
3894 | * The tunables feature is now enabled by default. This allows users to tweak | |
3895 | behavior of the GNU C Library using the GLIBC_TUNABLES environment variable. | |
3896 | ||
3897 | * New function reallocarray, which resizes an allocated block (like realloc) | |
3898 | to the product of two sizes, with a guaranteed clean failure upon integer | |
3899 | overflow in the multiplication. Originally from OpenBSD, contributed by | |
3900 | Dennis Wölfing and Rüdiger Sonderfeld. | |
3901 | ||
3902 | * New wrappers for the Linux-specific system calls preadv2 and pwritev2. | |
3903 | These are extended versions of preadv and pwritev, respectively, taking an | |
3904 | additional flags argument. The set of supported flags depends on the | |
3905 | running kernel; full support currently requires kernel 4.7 or later. | |
3906 | ||
3907 | * posix_spawnattr_setflags now supports the flag POSIX_SPAWN_SETSID, to | |
3908 | create a new session ID for the spawned process. This feature is | |
3909 | scheduled to be added to the next major revision of POSIX; for the time | |
3910 | being, it is available under _GNU_SOURCE. | |
3911 | ||
3912 | * errno.h is now safe to use from C-preprocessed assembly language on all | |
3913 | supported operating systems. In this context, it will only define the | |
3914 | Exxxx constants, as preprocessor macros expanding to integer literals. | |
3915 | ||
3916 | * On ia64, powerpc64le, x86-32, and x86-64, the math library now implements | |
3917 | 128-bit floating point as defined by ISO/IEC/IEEE 60559:2011 (IEEE | |
3918 | 754-2008) and ISO/IEC TS 18661-3:2015. Contributed by Paul E. Murphy, | |
3919 | Gabriel F. T. Gomes, Tulio Magno Quites Machado Filho, and Joseph Myers. | |
3920 | ||
3921 | To compile programs that use this feature, the compiler must support | |
3922 | 128-bit floating point with the type name _Float128 (as defined by TS | |
3923 | 18661-3) or __float128 (the nonstandard name used by GCC for C++, and for | |
3924 | C prior to version 7). _GNU_SOURCE or __STDC_WANT_IEC_60559_TYPES_EXT__ | |
3925 | must be defined to make the new interfaces visible. | |
3926 | ||
3927 | The new functions and macros correspond to those present for other | |
3928 | floating-point types (except for a few obsolescent interfaces not | |
3929 | supported for the new type), with F128 or f128 suffixes; for example, | |
3930 | strtof128, HUGE_VAL_F128 and cosf128. Following TS 18661-3, there are no | |
3931 | printf or scanf formats for the new type; the strfromf128 and strtof128 | |
3932 | interfaces should be used instead. | |
3933 | ||
3934 | Deprecated and removed features, and other changes affecting compatibility: | |
3935 | ||
3936 | * The synchronization that pthread_spin_unlock performs has been changed to | |
3937 | now be equivalent to a C11 atomic store with release memory order to the | |
3938 | spin lock's memory location. Previously, several (but not all) | |
3939 | architectures used stronger synchronization (e.g., containing what is | |
3940 | often called a full barrier). This change can improve performance, but | |
3941 | may affect odd fringe uses of spin locks that depend on the previous | |
3942 | behavior (e.g., using spin locks as atomic variables to try to implement | |
3943 | Dekker's mutual exclusion algorithm). | |
3944 | ||
3945 | * The port to Native Client running on ARMv7-A (--host=arm-nacl) has been | |
3946 | removed. | |
3947 | ||
3948 | * Sun RPC is deprecated. The rpcgen program, librpcsvc, and Sun RPC headers | |
3949 | will only be built and installed when the GNU C Library is configured with | |
3950 | --enable-obsolete-rpc. This allows alternative RPC implementations, such | |
3951 | as TIRPC or rpcsvc-proto, to be used. | |
3952 | ||
3953 | * The NIS(+) name service modules, libnss_nis, libnss_nisplus, and | |
3954 | libnss_compat, are deprecated, and will not be built or installed by | |
3955 | default. | |
3956 | ||
3957 | The NIS(+) support library, libnsl, is also deprecated. By default, a | |
3958 | compatibility shared library will be built and installed, but not headers | |
3959 | or development libraries. Only a few NIS-related programs require this | |
3960 | library. (In particular, the GNU C Library has never required programs | |
3961 | that use 'gethostbyname' to be linked with libnsl.) | |
3962 | ||
3963 | Replacement implementations based on TIRPC, which additionally support | |
3964 | IPv6, are available from <https://github.com/thkukuk/>. The configure | |
3965 | option --enable-obsolete-nsl will cause libnsl's headers, and the NIS(+) | |
3966 | name service modules, to be built and installed. | |
3967 | ||
3968 | * The DNS stub resolver no longer performs EDNS fallback. If EDNS or DNSSEC | |
3969 | support is enabled, the configured recursive resolver must support EDNS. | |
3970 | (Responding to EDNS-enabled queries with responses which are not | |
3971 | EDNS-enabled is fine, but FORMERR responses are not.) | |
3972 | ||
3973 | * res_mkquery and res_nmkquery no longer support the IQUERY opcode. DNS | |
3974 | servers have not supported this opcode for a long time. | |
3975 | ||
3976 | * The _res_opcodes variable has been removed from libresolv. It had been | |
3977 | exported by accident. | |
3978 | ||
3979 | * <string.h> no longer includes inline versions of any string functions, | |
3980 | as this kind of optimization is better done by the compiler. The macros | |
3981 | __USE_STRING_INLINES and __NO_STRING_INLINES no longer have any effect. | |
3982 | ||
3983 | * The nonstandard header <xlocale.h> has been removed. Most programs should | |
3984 | use <locale.h> instead. If you have a specific need for the definition of | |
3985 | locale_t with no other declarations, please contact | |
3986 | libc-alpha@sourceware.org and explain. | |
3987 | ||
3988 | * The obsolete header <sys/ultrasound.h> has been removed. | |
3989 | ||
3990 | * The obsolete signal constant SIGUNUSED is no longer defined by <signal.h>. | |
3991 | ||
3992 | * The obsolete function cfree has been removed. Applications should use | |
3993 | free instead. | |
3994 | ||
3995 | * The stack_t type no longer has the name struct sigaltstack. This changes | |
3996 | the C++ name mangling for interfaces involving this type. | |
3997 | ||
3998 | * The ucontext_t type no longer has the name struct ucontext. This changes | |
3999 | the C++ name mangling for interfaces involving this type. | |
4000 | ||
4001 | * On M68k GNU/Linux and MIPS GNU/Linux, the fpregset_t type no longer has | |
4002 | the name struct fpregset. On Nios II GNU/Linux, the mcontext_t type no | |
4003 | longer has the name struct mcontext. On SPARC GNU/Linux, the struct | |
4004 | mc_fq, struct rwindow, struct fpq and struct fq types are no longer | |
4005 | defined in sys/ucontext.h, the mc_fpu_t type no longer has the name struct | |
4006 | mc_fpu, the gwindows_t type no longer has the name struct gwindows and the | |
4007 | fpregset_t type no longer has the name struct fpu. This changes the C++ | |
4008 | name mangling for interfaces involving those types. | |
4009 | ||
4010 | * On S/390 GNU/Linux, the constants defined by <sys/ptrace.h> have been | |
4011 | synced with the kernel: | |
4012 | ||
4013 | - PTRACE_GETREGS, PTRACE_SETREGS, PTRACE_GETFPREGS and PTRACE_SETFPREGS | |
4014 | are not supported on this architecture and have been removed. | |
4015 | ||
4016 | - PTRACE_SINGLEBLOCK, PTRACE_SECCOMP_GET_FILTER, PTRACE_PEEKUSR_AREA, | |
4017 | PTRACE_POKEUSR_AREA, PTRACE_GET_LAST_BREAK, PTRACE_ENABLE_TE, | |
4018 | PTRACE_DISABLE_TE and PTRACE_TE_ABORT_RAND have been added. | |
4019 | ||
4020 | Programs that assume the GET/SETREGS ptrace requests are universally | |
4021 | available will now fail to build, instead of malfunctioning at runtime. | |
4022 | ||
4023 | Changes to build and runtime requirements: | |
4024 | ||
4025 | * Linux kernel 3.2 or later is required at runtime, on all architectures | |
4026 | supported by that kernel. (This is a change from version 2.25 only for | |
4027 | x86-32 and x86-64.) | |
4028 | ||
4029 | * GNU Binutils 2.25 or later is now required to build the GNU C Library. | |
4030 | ||
4031 | * On most architectures, GCC 4.9 or later is required to build the GNU C | |
4032 | Library. On powerpc64le, GCC 6.2 or later is required. | |
4033 | ||
4034 | Older GCC versions and non-GNU compilers are still supported when | |
4035 | compiling programs that use the GNU C Library. (We do not know exactly | |
4036 | how old, and some GNU extensions to C may be _de facto_ required. If you | |
4037 | are interested in helping us make this statement less vague, please | |
4038 | contact libc-alpha@sourceware.org.) | |
4039 | ||
4040 | Security related changes: | |
4041 | ||
4042 | * The DNS stub resolver limits the advertised UDP buffer size to 1200 bytes, | |
4043 | to avoid fragmentation-based spoofing attacks (CVE-2017-12132). | |
4044 | ||
4045 | * LD_LIBRARY_PATH is now ignored in binaries running in privileged AT_SECURE | |
4046 | mode to guard against local privilege escalation attacks (CVE-2017-1000366). | |
4047 | ||
4048 | * Avoid printing a backtrace from the __stack_chk_fail function since it is | |
4049 | called on a corrupt stack and a backtrace is unreliable on a corrupt stack | |
4050 | (CVE-2010-3192). | |
4051 | ||
4052 | * A use-after-free vulnerability in clntudp_call in the Sun RPC system has been | |
4053 | fixed (CVE-2017-12133). | |
4054 | ||
4055 | The following bugs are resolved with this release: | |
4056 | ||
4057 | [984] network: Respond to changed resolv.conf in gethostbyname | |
4058 | [5010] network: sunrpc service cleanup causes unwanted port mapper traffic | |
4059 | [12068] localedata: sc_IT: misspelled yesexpr/day/abday/mon/abmon/date_fmt | |
4060 | fields | |
4061 | [12189] libc: __stack_chk_fail should not attempt a backtrace | |
4062 | (CVE-2010-3192) | |
4063 | [14096] time: Race condition on timezone/tst-timezone.out | |
4064 | [14172] localedata: az_IR: new locale | |
4065 | [14995] build: glibc fails to build if gold is the default linker, even if | |
4066 | ld.bfd is available | |
4067 | [15998] build: [powerpc] Set arch_minimum_kernel for powerpc LE | |
4068 | [16637] network: inet_pton function is accepting IPv6 with bad format | |
4069 | [16640] string: string/strtok.c: undefined behaviour inconsistent between | |
4070 | x86 and other generic code | |
4071 | [16875] localedata: ko_KR: fix lang_name | |
4072 | [17225] localedata: ar_SY: localized month names for May and June are | |
4073 | incorrect | |
4074 | [17297] localedata: da_DK: wrong date_fmt string | |
4075 | [18907] stdio: Incorrect order of __wur __THROW in <printf.h> | |
4076 | [18934] localedata: hu_HU: collate: fix multiple bugs and add tests | |
4077 | [18988] nptl: pthread wastes memory with mlockall(MCL_FUTURE) | |
4078 | [19066] localedata: ar_SA abbreviated day and month names are in English | |
4079 | [19569] network: resolv: Support an arbitrary number of search domains | |
4080 | [19570] network: Implement random DNS server selection in the stub | |
4081 | resolver | |
4082 | [19838] locale: localedef fails on PA-RISC | |
4083 | [19919] localedata: iso14651_t1_common: Correct the Malayalam sorting | |
4084 | order of 0D36 and 0D37 | |
4085 | [19922] localedata: iso14651_t1_common: Define collation for Malayalam | |
4086 | chillu characters | |
4087 | [20098] libc: FAIL: debug/backtrace-tst on hppa | |
4088 | [20257] network: sunrpc: clntudp_call does not enforce timeout when | |
4089 | receiving data | |
4090 | [20275] localedata: locale day/abday/mon/abmon should not have trailing | |
4091 | whitespace | |
4092 | [20313] localedata: Update locale data to Unicode 9.0 | |
4093 | [20424] manual: Document how to provide a malloc replacement | |
4094 | [20496] localedata: agr_PE: new language locale Awajún / Aguaruna (agr) | |
4095 | for Peru | |
4096 | [20686] locale: Add el_GR@euro to SUPPORTED. | |
4097 | [20831] dynamic-link: _dl_map_segments does not test for __mprotect | |
4098 | failures consistently | |
4099 | [21015] dynamic-link: Document and fix --enable-bind-now | |
4100 | [21016] nptl: pthread_cond support is broken on hppa | |
4101 | [21029] libc: glibc-2.23 (and later) fails to compile with -fno-omit- | |
4102 | frame-pointer on i386 | |
4103 | [21049] libc: segfault in longjmp_chk() due to clobbered processor | |
4104 | register | |
4105 | [21075] libc: unused assignment to %g4 in sparc/sparc{64,32}/clone.S | |
4106 | [21088] libc: Build fails with --enable-static-nss | |
4107 | [21094] math: cosf(1.57079697) has 3 ulp error on targets where the | |
4108 | generic c code is used | |
4109 | [21109] libc: Tunables broken on big-endian | |
4110 | [21112] math: powf has large ulp errors with base close to 1 and exponent | |
4111 | around 4000 | |
4112 | [21115] network: sunrpc: Use-after-free in error path in clntudp_call | |
4113 | (CVE-2017-12133) | |
4114 | [21120] malloc: glibc malloc is incompatible with GCC 7 | |
4115 | [21130] math: Incorrect return from y0l (-inf) and y1l (-inf) when linking | |
4116 | with -lieee | |
4117 | [21134] math: Exception (divide by zero) not set for y0/y1 (0.0) and y0/y1 | |
4118 | (-0.0) when linking with -lieee | |
4119 | [21171] math: log10, log2 and lgamma return incorrect results | |
4120 | [21179] libc: handle R_SPARC_DISP64 and R_SPARC_REGISTER relocs | |
4121 | [21182] libc: __memchr_sse2: regression in glibc-2.25 on i686 | |
4122 | [21207] localedata: ce_RU: update weekdays from CLDR | |
4123 | [21209] dynamic-link: LD_HWCAP_MASK read in setuid binaries | |
4124 | [21217] localedata: Update months from CLDR-31 | |
4125 | [21232] libc: miss posix_fadvise64 on MIPS64 when static linking | |
4126 | [21243] libc: support_delete_temp_file should issue warning for failed | |
4127 | remove() | |
4128 | [21244] libc: support resolv_test_start() socket fd close should be | |
4129 | checked for errors. | |
4130 | [21253] libc: localedef randomly segfaults when using -fstack-check due to | |
4131 | new posix_spawn implementation | |
4132 | [21258] dynamic-link: Branch predication in _dl_runtime_resolve_avx512_opt | |
4133 | leads to lower CPU frequency | |
4134 | [21259] libc: [alpha] termios.h missing IXANY for POSIX | |
4135 | [21261] libc: [sparc64] bits/setjmp.h namespace | |
4136 | [21267] network: [mips] bits/socket.h IOC* namespace | |
4137 | [21268] libc: [alpha] termios.h NL2, NL3 namespace | |
4138 | [21270] libc: mmap64 silently truncates large offset values | |
4139 | [21275] libc: posix_spawn always crashes on ia64 now | |
4140 | [21277] libc: [alpha] termios.h missing IUCLC for UNIX98 and older | |
4141 | [21280] math: [powerpc] logbl for POWER7 return incorrect results | |
4142 | [21289] libc: Incorrect declaration for 32-bit platforms with | |
4143 | _FILE_OFFSET_BITS=64 causes build error | |
4144 | [21295] network: GETAI(AF_UNSPEC) drops IPv6 addresses if nss module does | |
4145 | not support gethostbyname4_r | |
4146 | [21298] nptl: rwlock can deadlock on frequent reader/writer phase | |
4147 | switching | |
4148 | [21338] malloc: mallopt M_ARENA_MAX doesn't set the maximum number of | |
4149 | arenas | |
4150 | [21340] libc: Support POSIX_SPAWN_SETSID | |
4151 | [21357] libc: unwind-dw2-fde deadlock when using AddressSanitizer | |
4152 | [21359] network: ns_name_pack needs additional byte in destination buffer | |
4153 | [21361] network: resolv: Reduce advertised EDNS0 buffer size to guard | |
4154 | against fragmentation attacks (CVE-2017-12132) | |
4155 | [21369] network: resolv: Remove EDNS fallback | |
4156 | [21371] libc: Missing timespec definition when compiled with _XOPEN_SOURCE | |
4157 | and _POSIX_C_SOURCE | |
4158 | [21386] nptl: Assertion in fork for distinct parent PID is incorrect | |
4159 | [21391] dynamic-link: x86: Set dl_platform and dl_hwcap from CPU features | |
4160 | [21393] stdio: Missing dup3 error check in freopen, freopen64 | |
4161 | [21396] libc: Use AVX2 memcpy/memset on Skylake server | |
4162 | [21399] localedata: Bad description for U00EC in | |
4163 | localedata/charmaps/CP1254 | |
4164 | [21411] malloc: realloc documentation error | |
4165 | [21426] network: sys/socket.h uio.h namespace | |
4166 | [21428] libc: [aarch64] tst-backtrace5 testsuite failure | |
4167 | [21445] libc: signal.h bsd_signal namespace | |
4168 | [21455] network: Network headers stdint.h namespace | |
4169 | [21474] network: resolv: res_init does not use RES_DFLRETRY (2) but 4 for | |
4170 | retry value | |
4171 | [21475] network: resolv: Overlong search path is truncated mid-label | |
4172 | [21511] libc: sigstack namespace | |
4173 | [21512] libc: clone() ends up calling exit_group() through _exit() wrapper | |
4174 | [21514] libc: sysdeps/unix/sysv/linux/sys/syscall.h:31:27: fatal error: | |
4175 | bits/syscall.h: No such file or directory | |
4176 | [21517] libc: struct sigaltstack namespace | |
4177 | [21528] dynamic-link: Duplicated minimal strtoul implementations in ld.so | |
4178 | [21533] localedata: Update locale data to Unicode 10.0 | |
4179 | [21537] libc: | |
4180 | ../sysdeps/unix/sysv/linux/s390/s390-32/__makecontext_ret.S:44: Error: | |
4181 | junk at end of line, first unrecognized character is `@' | |
4182 | [21538] libc: SIG_HOLD missing for XPG4 | |
4183 | [21539] libc: S390: Mismatch between kernel and glibc ptrace.h with | |
4184 | request 12: PTRACE_SINGLEBLOCK vs PTRACE_GETREGS. | |
4185 | [21542] libc: Use conservative default for sysconf (_SC_NPROCESSORS_ONLN) | |
4186 | [21543] libc: sigevent namespace | |
4187 | [21548] libc: [mips] get/set/make/swap context for MIPS O32 assume wrong | |
4188 | size for general purpose registers in mcontext_t structure | |
4189 | [21550] libc: sigwait namespace | |
4190 | [21552] libc: XPG4 bsd_signal namespace | |
4191 | [21554] libc: sigpause namespace | |
4192 | [21560] libc: sys/wait.h signal.h namespace | |
4193 | [21561] libc: waitid namespace | |
4194 | [21573] nptl: GCC 7: /usr/bin/install: cannot remove | |
4195 | '/usr/include/stdlib.h': Permission denied | |
4196 | [21575] libc: sys/wait.h missing struct rusage definition | |
4197 | [21584] libc: sigaltstack etc namespace | |
4198 | [21597] libc: siginterrupt namespace | |
4199 | [21607] math: hppa: FAIL: math/test-tgmath | |
4200 | [21609] dynamic-link: Incomplete workaround for GCC __tls_get_addr ABI | |
4201 | issue on x86-64 | |
4202 | [21622] libc: [tile] missing SA_* for POSIX.1:2008 | |
4203 | [21624] dynamic-link: ld.so: Unsafe alloca allows local attackers to alias | |
4204 | stack and heap (CVE-2017-1000366) | |
4205 | [21625] libc: wait3 namespace | |
4206 | [21654] nss: Incorrect pointer alignment in NSS group merge result | |
4207 | construction | |
4208 | [21657] network: Parse interface zone id for node-local multicast | |
4209 | [21662] string: memcmp-avx2-movbe.S lacks saturating subtraction for | |
4210 | between_2_3 | |
4211 | [21666] libc: .symver is used on common symbol | |
4212 | [21668] network: resolv: res_init cross-thread broadcast introduces race | |
4213 | conditions | |
4214 | [21687] math: tgmath.h totalorder, totalordermag return type | |
4215 | [21694] locale: Current Glibc Locale Does Not Support Tok-Pisin and Fiji | |
4216 | Hindi Locale | |
4217 | [21696] libc: Incorrect assumption of of __cpu_mask in | |
4218 | posix/sched_cpucount.c | |
4219 | [21697] libc: sysdeps/posix/spawni.c: 2 * suspicious condition ? | |
4220 | [21706] localedata: yesstr and nostr are missing for Breton [LC_MESSAGES] | |
4221 | locale | |
4222 | [21707] math: ppc64le: Invalid IFUNC resolver from libgcc calls getauxval, | |
4223 | leading to relocation crash | |
4224 | [21709] libc: resolv_conf.c:552: update_from_conf: Assertion | |
4225 | `resolv_conf_matches (resp, conf)' failed. | |
4226 | [21710] localedata: Added Samoan language locale for Samoa | |
4227 | [21711] localedata: Pashto yesstr/nostr locale are missing | |
4228 | [21715] nptl: sysdeps/nptl/bits/pthreadtypes.h: typedef guard | |
4229 | __have_pthread_attr_t can cause redefinition of typedef ‘pthread_attr_t’ | |
4230 | [21721] localedata: Incorrect Full Weekday names for ks_IN@devanagari | |
4231 | [21723] localedata: yesstr/nostr missing for Chinese language locale | |
4232 | [21724] localedata: yesstr and nostr are missing for Xhosa [LC_MESSAGES] | |
4233 | locale | |
4234 | [21727] localedata: yesstr and nostr are missing for Tsonga [LC_MESSAGES] | |
4235 | locale | |
4236 | [21728] localedata: New Locale for Tongan language | |
4237 | [21729] localedata: incorrect LC_NAME fields for hi_IN | |
4238 | [21733] localedata: yesstr and nostr are missing for zh_HK | |
4239 | [21734] localedata: Missing yesstr and nostr are for kw_GB | |
4240 | [21738] libc: misc/tst-preadvwritev2 and misc/tst-preadvwritev64v2 fail | |
4241 | [21741] libc: Undefined __memmove_chk_XXX and __memset_chk_XXX in libc.a | |
4242 | [21742] libc: _dl_num_cache_relocations is undefined in libc.a | |
4243 | [21743] localedata: ks_IN@devanagari: abday strings mismatch the day | |
4244 | strings | |
4245 | [21744] libc: Tests failing on --enable-tunables --enable-stack- | |
4246 | protector=all | |
4247 | [21749] localedata: Wrong abbreviated day name (“abday”) for | |
4248 | ar_JO/ar_LB/ar_SY | |
4249 | [21756] localedata: missing yesstr, nostr for nds_DE and nds_NL | |
4250 | [21757] localedata: missing yesstr, nostr for pap_AW and pap_CW | |
4251 | [21759] localedata: missing yesstr and nostr for Tigrinya | |
4252 | [21760] localedata: Fix LC_MESSAGES and LC_ADDRESS for anp_IN | |
4253 | [21766] localedata: Wrong LC_MESSAGES for om_ET Locale | |
4254 | [21767] localedata: Missing Bislama locales | |
4255 | [21768] localedata: Missing yesstr and nostr for aa_ET | |
4256 | [21770] localedata: Missing Field in li_NL | |
4257 | [21778] nptl: Robust mutex may deadlock | |
4258 | [21779] libc: MicroBlaze segfaults when loading libpthread | |
4259 | [21783] localedata: Fix int_select international_call_prefixes | |
4260 | [21784] localedata: Inconsistency in country_isbn | |
4261 | [21788] localedata: Missing Country Postal Abbreviations | |
4262 | [21794] localedata: Added-country_isbn-for-Italy | |
4263 | [21795] localedata: Add/Fix country_isbn for France | |
4264 | [21796] localedata: Added country_isbn for Republic of Korea | |
4265 | [21797] localedata: Fix inconsistency in country_isbn and missing prefixes | |
4266 | [21799] localedata: Added int_select international_call_prefixes | |
4267 | [21801] localedata: Added int_select international_call_prefixes | |
4268 | [21804] nptl: Double semicolon in thread-shared-types.h | |
4269 | [21807] localedata: LC_ADDRESS fix for pap_CW | |
4270 | [21808] localedata: Fix LC_ADDRESS for pap_AW | |
4271 | [21821] localedata: Added country_name in mai_IN | |
4272 | [21822] localedata: Fix LC_TIME for mai_IN | |
4273 | [21823] localedata: missing yesstr, nostr for sa_IN | |
4274 | [21825] localedata: Fix name_mrs for mag_IN | |
4275 | [21828] localedata: 2.26 changelog should mention user visible changes | |
4276 | with unicode 9.0 | |
4277 | [21835] localedata: Added Maithili language locale for Nepal | |
4278 | [21838] localedata: Removed redundant data for the_NP | |
4279 | [21839] localedata: Fix LC_MONETARY for ta_LK | |
4280 | [21844] localedata: Fix Latin characters and Months Sequence. | |
4281 | [21848] localedata: Fix mai_NP Title Name | |
4282 | ||
4283 | \f | |
4284 | Version 2.25 | |
4285 | ||
4286 | * The feature test macro __STDC_WANT_LIB_EXT2__, from ISO/IEC TR | |
4287 | 24731-2:2010, is supported to enable declarations of functions from that | |
4288 | TR. Note that not all functions from that TR are supported by the GNU C | |
4289 | Library. | |
4290 | ||
4291 | * The feature test macro __STDC_WANT_IEC_60559_BFP_EXT__, from ISO/IEC TS | |
4292 | 18661-1:2014, is supported to enable declarations of functions and macros | |
4293 | from that TS. Note that not all features from that TS are supported by | |
4294 | the GNU C Library. | |
4295 | ||
4296 | * The feature test macro __STDC_WANT_IEC_60559_FUNCS_EXT__, from ISO/IEC TS | |
4297 | 18661-4:2015, is supported to enable declarations of functions and macros | |
4298 | from that TS. Note that most features from that TS are not supported by | |
4299 | the GNU C Library. | |
4300 | ||
4301 | * The nonstandard feature selection macros _REENTRANT and _THREAD_SAFE are | |
4302 | now treated as compatibility synonyms for _POSIX_C_SOURCE=199506L. | |
4303 | Since the GNU C Library defaults to a much newer revision of POSIX, this | |
4304 | will only affect programs that specifically request an old conformance | |
4305 | mode. For instance, a program compiled with -std=c89 -D_REENTRANT will | |
4306 | see a change in the visible declarations, but a program compiled with | |
4307 | just -D_REENTRANT, or -std=c99 -D_POSIX_C_SOURCE=200809L -D_REENTRANT, | |
4308 | will not. | |
4309 | ||
4310 | Some C libraries once required _REENTRANT and/or _THREAD_SAFE to be | |
4311 | defined by all multithreaded code, but glibc has not required this for | |
4312 | many years. | |
4313 | ||
4314 | * The inclusion of <sys/sysmacros.h> by <sys/types.h> is deprecated. This | |
4315 | means that in a future release, the macros “major”, “minor”, and “makedev” | |
4316 | will only be available from <sys/sysmacros.h>. | |
4317 | ||
4318 | These macros are not part of POSIX nor XSI, and their names frequently | |
4319 | collide with user code; see for instance glibc bug 19239 and Red Hat bug | |
4320 | 130601. <stdlib.h> includes <sys/types.h> under _GNU_SOURCE, and C++ code | |
4321 | presently cannot avoid being compiled under _GNU_SOURCE, exacerbating the | |
4322 | problem. | |
4323 | ||
4324 | * New <fenv.h> features from TS 18661-1:2014 are added to libm: the | |
4325 | fesetexcept, fetestexceptflag, fegetmode and fesetmode functions, the | |
4326 | femode_t type and the FE_DFL_MODE and FE_SNANS_ALWAYS_SIGNAL macros. | |
4327 | ||
4328 | * Integer width macros from TS 18661-1:2014 are added to <limits.h>: | |
4329 | CHAR_WIDTH, SCHAR_WIDTH, UCHAR_WIDTH, SHRT_WIDTH, USHRT_WIDTH, INT_WIDTH, | |
4330 | UINT_WIDTH, LONG_WIDTH, ULONG_WIDTH, LLONG_WIDTH, ULLONG_WIDTH; and to | |
4331 | <stdint.h>: INT8_WIDTH, UINT8_WIDTH, INT16_WIDTH, UINT16_WIDTH, | |
4332 | INT32_WIDTH, UINT32_WIDTH, INT64_WIDTH, UINT64_WIDTH, INT_LEAST8_WIDTH, | |
4333 | UINT_LEAST8_WIDTH, INT_LEAST16_WIDTH, UINT_LEAST16_WIDTH, | |
4334 | INT_LEAST32_WIDTH, UINT_LEAST32_WIDTH, INT_LEAST64_WIDTH, | |
4335 | UINT_LEAST64_WIDTH, INT_FAST8_WIDTH, UINT_FAST8_WIDTH, INT_FAST16_WIDTH, | |
4336 | UINT_FAST16_WIDTH, INT_FAST32_WIDTH, UINT_FAST32_WIDTH, INT_FAST64_WIDTH, | |
4337 | UINT_FAST64_WIDTH, INTPTR_WIDTH, UINTPTR_WIDTH, INTMAX_WIDTH, | |
4338 | UINTMAX_WIDTH, PTRDIFF_WIDTH, SIG_ATOMIC_WIDTH, SIZE_WIDTH, WCHAR_WIDTH, | |
4339 | WINT_WIDTH. | |
4340 | ||
4341 | * New <math.h> features are added from TS 18661-1:2014: | |
4342 | ||
4343 | - Signaling NaN macros: SNANF, SNAN, SNANL. | |
4344 | ||
4345 | - Nearest integer functions: roundeven, roundevenf, roundevenl, fromfp, | |
4346 | fromfpf, fromfpl, ufromfp, ufromfpf, ufromfpl, fromfpx, fromfpxf, | |
4347 | fromfpxl, ufromfpx, ufromfpxf, ufromfpxl. | |
4348 | ||
4349 | - llogb functions: the llogb, llogbf and llogbl functions, and the | |
4350 | FP_LLOGB0 and FP_LLOGBNAN macros. | |
4351 | ||
4352 | - Max-min magnitude functions: fmaxmag, fmaxmagf, fmaxmagl, fminmag, | |
4353 | fminmagf, fminmagl. | |
4354 | ||
4355 | - Comparison macros: iseqsig. | |
4356 | ||
4357 | - Classification macros: iscanonical, issubnormal, iszero. | |
4358 | ||
4359 | - Total order functions: totalorder, totalorderf, totalorderl, | |
4360 | totalordermag, totalordermagf, totalordermagl. | |
4361 | ||
4362 | - Canonicalize functions: canonicalize, canonicalizef, canonicalizel. | |
4363 | ||
4364 | - NaN functions: getpayload, getpayloadf, getpayloadl, setpayload, | |
4365 | setpayloadf, setpayloadl, setpayloadsig, setpayloadsigf, setpayloadsigl. | |
4366 | ||
4367 | * The functions strfromd, strfromf, and strfroml, from ISO/IEC TS 18661-1:2014, | |
4368 | are added to libc. They convert a floating-point number into string. | |
4369 | ||
4370 | * Most of glibc can now be built with the stack smashing protector enabled. | |
4371 | It is recommended to build glibc with --enable-stack-protector=strong. | |
4372 | Implemented by Nick Alcock (Oracle). | |
4373 | ||
4374 | * The function explicit_bzero, from OpenBSD, has been added to libc. It is | |
4375 | intended to be used instead of memset() to erase sensitive data after use; | |
4376 | the compiler will not optimize out calls to explicit_bzero even if they | |
4377 | are "unnecessary" (in the sense that no _correct_ program can observe the | |
4378 | effects of the memory clear). | |
4379 | ||
4380 | * On ColdFire, MicroBlaze, Nios II and SH3, the float_t type is now defined | |
4381 | to float instead of double. This does not affect the ABI of any libraries | |
4382 | that are part of the GNU C Library, but may affect the ABI of other | |
4383 | libraries that use this type in their interfaces. | |
4384 | ||
4385 | * On x86_64, when compiling with -mfpmath=387 or -mfpmath=sse+387, the | |
4386 | float_t and double_t types are now defined to long double instead of float | |
4387 | and double. These options are not the default, and this does not affect | |
4388 | the ABI of any libraries that are part of the GNU C Library, but it may | |
4389 | affect the ABI of other libraries that use this type in their interfaces, | |
4390 | if they are compiled or used with those options. | |
4391 | ||
4392 | * The getentropy and getrandom functions, and the <sys/random.h> header file | |
4393 | have been added. | |
4394 | ||
4395 | * The buffer size for byte-oriented stdio streams is now limited to 8192 | |
4396 | bytes by default. Previously, on Linux, the default buffer size on most | |
4397 | file systems was 4096 bytes (and thus remains unchanged), except on | |
4398 | network file systems, where the buffer size was unpredictable and could be | |
4399 | as large as several megabytes. | |
4400 | ||
4401 | * The <sys/quota.h> header now includes the <linux/quota.h> header. Support | |
4402 | for the Linux quota interface which predates kernel version 2.4.22 has | |
4403 | been removed. | |
4404 | ||
4405 | * The malloc_get_state and malloc_set_state functions have been removed. | |
4406 | Already-existing binaries that dynamically link to these functions will | |
4407 | get a hidden implementation in which malloc_get_state is a stub. As far | |
4408 | as we know, these functions are used only by GNU Emacs and this change | |
4409 | will not adversely affect already-built Emacs executables. Any undumped | |
4410 | Emacs executables, which normally exist only during an Emacs build, should | |
4411 | be rebuilt by re-running “./configure; make” in the Emacs build tree. | |
4412 | ||
4413 | * The “ip6-dotint” and “no-ip6-dotint” resolver options, and the | |
4414 | corresponding RES_NOIP6DOTINT flag from <resolv.h> have been removed. | |
4415 | “no-ip6-dotint” had already been the default, and support for the | |
4416 | “ip6-dotint” option was removed from the Internet in 2006. | |
4417 | ||
4418 | * The "ip6-bytestring" resolver option and the corresponding RES_USEBSTRING | |
4419 | flag from <resolv.h> have been removed. The option relied on a | |
4420 | backwards-incompatible DNS extension which was never deployed on the | |
4421 | Internet. | |
4422 | ||
4423 | * The flags RES_AAONLY, RES_PRIMARY, RES_NOCHECKNAME, RES_KEEPTSIG, | |
4424 | RES_BLAST defined in the <resolv.h> header file have been deprecated. | |
4425 | They were already unimplemented. | |
4426 | ||
4427 | * The "inet6" option in /etc/resolv.conf and the RES_USE_INET6 flag for | |
4428 | _res.flags are deprecated. The flag was standardized in RFC 2133, but | |
4429 | removed again from the IETF name lookup interface specification in RFC | |
4430 | 2553. Applications should use getaddrinfo instead. | |
4431 | ||
4432 | * DNSSEC-related declarations and definitions have been removed from the | |
4433 | <arpa/nameser.h> header file, and libresolv will no longer attempt to | |
4434 | decode the data part of DNSSEC record types. Previous versions of glibc | |
4435 | only implemented minimal support for the previous version of DNSSEC, which | |
4436 | is incompatible with the currently deployed version. | |
4437 | ||
4438 | * The resource record type classification macros ns_t_qt_p, ns_t_mrr_p, | |
4439 | ns_t_rr_p, ns_t_udp_p, ns_t_xfr_p have been removed from the | |
4440 | <arpa/nameser.h> header file because the distinction between RR types and | |
4441 | meta-RR types is not officially standardized, subject to revision, and | |
4442 | thus not suitable for encoding in a macro. | |
4443 | ||
4444 | * The types res_sendhookact, res_send_qhook, re_send_rhook, and the qhook | |
4445 | and rhook members of the res_state type in <resolv.h> have been removed. | |
4446 | The glibc stub resolver did not support these hooks, but the header file | |
4447 | did not reflect that. | |
4448 | ||
4449 | * For multi-arch support it is recommended to use a GCC which has | |
4450 | been built with support for GNU indirect functions. This ensures | |
4451 | that correct debugging information is generated for functions | |
4452 | selected by IFUNC resolvers. This support can either be enabled by | |
4453 | configuring GCC with '--enable-gnu-indirect-function', or by | |
4454 | enabling it by default by setting 'default_gnu_indirect_function' | |
4455 | variable for a particular architecture in the GCC source file | |
4456 | 'gcc/config.gcc'. | |
4457 | ||
4458 | * GDB pretty printers have been added for mutex and condition variable | |
4459 | structures in POSIX Threads. When installed and loaded in gdb these pretty | |
4460 | printers show various pthread variables in human-readable form when read | |
4461 | using the 'print' or 'display' commands in gdb. | |
4462 | ||
4463 | * Tunables feature added to allow tweaking of the runtime for an application | |
4464 | program. This feature can be enabled with the '--enable-tunables' configure | |
4465 | flag. The GNU C Library manual has details on usage and README.tunables has | |
4466 | instructions on adding new tunables to the library. | |
4467 | ||
4468 | * A new version of condition variables functions have been implemented in | |
4469 | the NPTL implementation of POSIX Threads to provide stronger ordering | |
4470 | guarantees. | |
4471 | ||
4472 | * A new version of pthread_rwlock functions have been implemented to use a more | |
4473 | scalable algorithm primarily through not using a critical section anymore to | |
4474 | make state changes. | |
4475 | ||
4476 | Security related changes: | |
4477 | ||
4478 | * On ARM EABI (32-bit), generating a backtrace for execution contexts which | |
4479 | have been created with makecontext could fail to terminate due to a | |
4480 | missing .cantunwind annotation. This has been observed to lead to a hang | |
4481 | (denial of service) in some Go applications compiled with gccgo. Reported | |
4482 | by Andreas Schwab. (CVE-2016-6323) | |
4483 | ||
4484 | * The DNS stub resolver functions would crash due to a NULL pointer | |
4485 | dereference when processing a query with a valid DNS question type which | |
4486 | was used internally in the implementation. The stub resolver now uses a | |
4487 | question type which is outside the range of valid question type values. | |
4488 | (CVE-2015-5180) | |
4489 | ||
4490 | The following bugs are resolved with this release: | |
4491 | ||
4492 | [4099] stdio: Overly aggressive caching by stream i/o functions. | |
4493 | [7065] build: Support building glibc with -fstack-protector or -fstack- | |
4494 | protector-all | |
4495 | [9842] localedata: en_CA: incorrect date format | |
4496 | [13165] nptl: pthread_cond_wait() can consume a signal that was sent | |
4497 | before it started waiting | |
4498 | [14139] manual: Do not hardcode platform names in manual/libm-err-tab.pl | |
4499 | [15765] nptl: sem_open is wrongly a cancellation point | |
4500 | [16421] network: IN6_IS_ADDR_UNSPECIFIED can use undefined s6_addr32 | |
4501 | [16458] libc: endian.h and netinet/in.h byte order macros return values of | |
4502 | wrong type | |
4503 | [16628] dynamic-link: Segfault after a binary without pthread dlopen()s a | |
4504 | library linked with pthread | |
4505 | [16630] nptl: Use SYSENTER for pthread_cond_broadcast/signal() (i.e. fix | |
4506 | "FIXME: Ingo" issue) | |
4507 | [16907] libc: <argp.h> compiled with --std=cXX disables __attribute__ | |
4508 | [17252] libc: getrandom and getentropy syscall | |
4509 | [17730] malloc: thread-local storage is sometimes improperly free()'d | |
4510 | after being __libc_memalign()'d | |
4511 | [18241] stdio: failed fseek on memstream does not set errno and can fail | |
4512 | when it shouldn't | |
4513 | [18243] nptl: sem_wait, sem_timedwait are cancellation points shm_open is | |
4514 | not | |
4515 | [18463] nptl: pthread_cond_broadcast issue when surrounded by | |
4516 | PTHREAD_PRIO_INHERIT mutex on ARM | |
4517 | [18784] network: res_query and related function crash for special record | |
4518 | type queries (CVE-2015-5180) | |
4519 | [19380] math: strtod does not raise "inexact" | |
4520 | [19387] string: Integer overflow in memchr | |
4521 | [19390] string: Integer overflow in strncat | |
4522 | [19398] build: linknamespace tests fail with massively parallel build | |
4523 | [19402] nptl: Deadlock with robust shared mutex and asynchronous | |
4524 | termination | |
4525 | [19469] malloc: M_PERTURB in test-skeleton.c invalidates malloc tests | |
4526 | [19473] malloc: Turn malloc_get_state etc. in compatibility symbols | |
4527 | [19514] libc: [PATCH] Fix spelling errors in spelling | |
4528 | "implement"/"implementation" in several places | |
4529 | [19582] network: Deprecate RES_USE_INET6 | |
4530 | [19673] manual: clog10 docs appear to be erroneous | |
4531 | [19810] dynamic-link: dlopen with both RTLD_NOLOAD and RTLD_NODELETE | |
4532 | causes a segmentation fault | |
4533 | [19826] libc: invalid pointer returned from __tls_get_addr with static | |
4534 | linking | |
4535 | [20016] network: resolv: Remove hooks support from the API | |
4536 | [20019] dynamic-link: NULL pointer dereference in libc.so.6 IFUNC due to | |
4537 | uninitialized GOT | |
4538 | [20033] math: [x86_64] vectorized math function don't call the __finite | |
4539 | versions | |
4540 | [20116] nptl: use after free in pthread_create | |
4541 | [20181] stdio: open_memstream(): writes not at end of stream corrupt data | |
4542 | [20292] dynamic-link: Comparison in elf/dl-open.c _dl_addr_inside_object | |
4543 | is always true. | |
4544 | [20311] nptl: please install proc_service.h | |
4545 | [20366] libc: Compilation errors in installed headers in strict-compliance | |
4546 | modes | |
4547 | [20370] malloc: malloc: Arena free list management is still racy | |
4548 | (incorrect fix in bug 19243) | |
4549 | [20386] libc: assert (X = 0) does not result in GCC warning | |
4550 | [20432] malloc: malloc: Minimize interface required for interposition | |
4551 | [20435] libc: Missing unwind info in __startcontext causes infinite loop | |
4552 | in _Unwind_Backtrace (CVE-2016-6323) | |
4553 | [20444] hurd: recvmsg: PF_LOCAL sockets and msg_name lead to SIGLOST | |
4554 | [20452] nptl: Addition of sysdep.o to libpthread.a breaks relinking | |
4555 | libpthread.a | |
4556 | [20455] math: [powerpc] fesetexceptflag fails to clear FE_INVALID | |
4557 | [20459] localedata: et_EE: locale has wrong {p,n}_cs_precedes value | |
4558 | [20477] network: resolv: incorrect double-checked locking related to | |
4559 | _res_hconf | |
4560 | [20478] libc: libc_ifunc macro and similar usages leads to false debug- | |
4561 | information. | |
4562 | [20495] math: x86_64 performance degradation due to AVX/SSE transition | |
4563 | penalty | |
4564 | [20497] localedata: lt_LT: LC_TIME d_fmt used is obsolete | |
4565 | [20508] dynamic-link: _dl_runtime_resolve_avx/_dl_runtime_profile_avx512 | |
4566 | cause transition penalty | |
4567 | [20517] math: sparcv9 missing fdiml compat symbol | |
4568 | [20524] manual: strverscmp is inconsistent | |
4569 | [20525] libc: <sys/quota.h> should be based on kernel headers | |
4570 | [20539] math: GCC 7 -static -lm fails to link at -O3 | |
4571 | [20554] libc: ld: bss-plt forced due to /usr/lib/libc_nonshared.a(ppc- | |
4572 | mcount.oS) | |
4573 | [20558] string: POSIX bcopy/bzero decls do not implement Object Size | |
4574 | Checking | |
4575 | [20591] network: Remove obsolete DNSSEC support | |
4576 | [20592] network: DNS resource record type classification macros in | |
4577 | <arpa/nameser.h> are incorrect | |
4578 | [20593] network: Update DNS RR type definitions | |
4579 | [20611] network: getaddrinfo accepts invalid numeric scope IDs | |
4580 | [20615] build: glibc build fails when using --with-cpu=power9 --enable- | |
4581 | multi-arch | |
4582 | [20629] network: libresolv: Remove support for bitstring labels | |
4583 | (RES_USEBSTRING) | |
4584 | [20647] libc: GLIBC quitting every program - glibc on Pentium-S leads to | |
4585 | assertion: "maxidx >= 2" | |
4586 | [20660] math: [arm] Use VSQRT | |
4587 | [20662] libc: checking whether x86_64-pc-linux-gnu-gcc implicitly enables | |
4588 | -fstack-protector no (32bit gcc 6.2.0 pie and ssp enable) | |
4589 | [20689] libc: Test for FMA should also check for AVX. | |
4590 | [20707] glob: gl_pathv entries not set to NULL with GLOB_DOOFFS | |
4591 | [20715] math: iszero macro breaks existing code | |
4592 | [20718] math: [powerpc] copysignl raises "invalid" for sNaN | |
4593 | [20728] libc: powerpc: Missing TOC stub in clone | |
4594 | [20729] build: glibc-2.24 fails to build for i486 with -Os | |
4595 | [20750] build: Build fails with default PIE enabled gcc-6.x | |
4596 | [20768] math: [alpha] sqrt fegetenv namespace | |
4597 | [20785] libc: binutils 2.28 fails to assemble power6/memset.S file in | |
4598 | glibc | |
4599 | [20787] math: float_t is defined as float incorrectly on x86_64 even if | |
4600 | __FLT_EVAL_METHOD__ is 2 | |
4601 | [20790] network: rpcgen buffer overrun in get_prog_declaration | |
4602 | [20822] nptl: race condition in __lll_unlock_elision on powerpc | |
4603 | [20829] libc: crypt snprintf namespace | |
4604 | [20847] libc: tst-vfork3 failure | |
4605 | [20855] math: Default bits/mathdef.h has inappropriate float_t | |
4606 | [20859] math: [sh4] FP_ILOGB0 invalid | |
4607 | [20864] localedata: iconv: cp936 missing single-byte euro sign (0x80, | |
4608 | U+20AC), not same as GBK | |
4609 | [20915] dynamic-link: global-dynamic TLS broken on aarch64 | |
4610 | [20916] math: pow handling of sNaN arguments | |
4611 | [20918] build: Building with --enable-nss-crypt fails tst-linkall-static. | |
4612 | [20919] math: Bad pow (qNaN, 0) result with -lieee | |
4613 | [20940] math: hypot sNaN handling | |
4614 | [20947] math: fmax, fmin sNaN handling | |
4615 | [20956] libc: debug/tst-backtrace3-6 don't work with -O3 anymore | |
4616 | [20964] network: sunrpc: Stack-based buffer overflow in getrpcport with | |
4617 | RES_USE_INET6 | |
4618 | [20971] string: powerpc64/power7 memchr overflows internal pointer check | |
4619 | [20973] nptl: robust mutexes: Lost wake-ups | |
4620 | [20974] locale: bs_BA: yesexpr/noexpr regexes accept any character | |
4621 | [20978] nis: strlen on null pointer in nss_nisplus | |
4622 | [20985] nptl: robust mutexes: lowlevelrobustlock assembly on x86 blocks on | |
4623 | wrong condition | |
4624 | [21014] string: i686 memchr overflows internal pointer check | |
4625 | [21019] libc: [mips] n32 lseek incorrect on overflow | |
4626 | [21022] libc: [microblaze] __backtrace get_frame_size namespace | |
4627 | [21026] libc: [MIPS] readahead syscall is broken on n64 | |
4628 | [21028] math: Fallback fesetexceptflag should always succeed | |
4629 | [21045] libc: [powerpc-nofpu] swapcontext does not restore signal mask | |
4630 | [21047] math: arm: fpu_control.h: _FPU_GETCW/_FPU_SETCW is rejected by | |
4631 | clang | |
4632 | [21053] libc: [SH] Namespace pollution from sys/ucontext.h | |
4633 | [21061] librt: [microblaze] librt lost clock_* exports | |
4634 | [21073] libc: tunables: insecure environment variables passed to | |
4635 | subprocesses with AT_SECURE | |
4636 | [21081] string: Missing vzeroupper in memset-vec-unaligned-erms.S | |
4637 | \f | |
4638 | Version 2.24 | |
4639 | ||
4640 | * The minimum Linux kernel version that this version of the GNU C Library | |
4641 | can be used with is 3.2, except on i[4567]86 and x86_64, where Linux | |
4642 | kernel version 2.6.32 or later suffices (on architectures that already | |
4643 | required kernel versions more recent than 3.2, those requirements remain | |
4644 | unchanged). Linux 3.2 or later kernel headers are required on all | |
4645 | architectures. | |
4646 | ||
4647 | * The pap_AN locale has been deleted. This has been deprecated for a long | |
4648 | time. It has been replaced by pap_AW & pap_CW, both of which have long | |
4649 | been included in previous releases. | |
4650 | ||
4651 | * The readdir_r and readdir64_r functions have been deprecated. It is | |
4652 | recommended to use readdir and readdir64 instead. | |
4653 | ||
4654 | * The type “union wait” has been removed. It was deprecated in the early | |
4655 | 1990s and never part of POSIX. Application code should use the int type | |
4656 | instead of “union wait”. | |
4657 | ||
4658 | * A new NSS action is added to facilitate large distributed system | |
4659 | administration. The action, MERGE, allows remote user stores like LDAP | |
4660 | to be merged into local user stores like /etc/groups in order to provide | |
4661 | easy to use, updated, and managed sets of merged credentials. The new | |
4662 | action can be used by configuring it in /etc/nsswitch.conf: | |
4663 | group: files [SUCCESS=merge] nis | |
4664 | Implemented by Stephen Gallagher (Red Hat). | |
4665 | ||
4666 | * The deprecated __malloc_initialize_hook variable has been removed from the | |
4667 | API. | |
4668 | ||
4669 | * The long unused localedef --old-style option has been removed. It hasn't | |
4670 | done anything in over 16 years. Scripts using this option can safely | |
4671 | drop it. | |
4672 | ||
4673 | * nextupl, nextup, nextupf, nextdownl, nextdown and nextdownf are added to | |
4674 | libm. They are defined by TS 18661 and IEEE754-2008. The nextup functions | |
4675 | return the next representable value in the direction of positive infinity | |
4676 | and the nextdown functions return the next representable value in the | |
4677 | direction of negative infinity. These are currently enabled as GNU | |
4678 | extensions. | |
4679 | ||
4680 | Security related changes: | |
4681 | ||
4682 | * An unnecessary stack copy in _nss_dns_getnetbyname_r was removed. It | |
4683 | could result in a stack overflow when getnetbyname was called with an | |
4684 | overly long name. (CVE-2016-3075) | |
4685 | ||
4686 | * Previously, getaddrinfo copied large amounts of address data to the stack, | |
4687 | even after the fix for CVE-2013-4458 has been applied, potentially | |
4688 | resulting in a stack overflow. getaddrinfo now uses a heap allocation | |
4689 | instead. Reported by Michael Petlan. (CVE-2016-3706) | |
4690 | ||
4691 | * The glob function suffered from a stack-based buffer overflow when it was | |
4692 | called with the GLOB_ALTDIRFUNC flag and encountered a long file name. | |
4693 | Reported by Alexander Cherepanov. (CVE-2016-1234) | |
4694 | ||
4695 | * The Sun RPC UDP client could exhaust all available stack space when | |
4696 | flooded with crafted ICMP and UDP messages. Reported by Aldy Hernandez' | |
4697 | alloca plugin for GCC. (CVE-2016-4429) | |
4698 | ||
4699 | * The IPv6 name server management code in libresolv could result in a memory | |
4700 | leak for each thread which is created, performs a failing naming lookup, | |
4701 | and exits. Over time, this could result in a denial of service due to | |
4702 | memory exhaustion. Reported by Matthias Schiffer. (CVE-2016-5417) | |
4703 | ||
4704 | The following bugs are resolved with this release: | |
4705 | ||
4706 | [1170] localedata: ne_NP: update Nepali locale definition file | |
4707 | [3629] manual: stpcpy description in string.texi refers to MS-DOG instead | |
4708 | of MS-DOS. | |
4709 | [6527] malloc: [powerpc] Malloc alignment insufficient for PowerPC | |
4710 | [6796] math: fdim() does not set errno on overflow | |
4711 | [10354] libc: posix_spawn should use vfork() in more cases than presently | |
4712 | [11213] localedata: localedata: add copyright disclaimer to locale files | |
4713 | [12143] localedata: chr_US: new Cherokee locale | |
4714 | [12450] localedata: sgs_LT: new locale | |
4715 | [12676] localedata: ln_CD: new locale | |
4716 | [13237] localedata: LC_ADDRESS.country_name: update all locales w/latest | |
4717 | CLDR data | |
4718 | [13304] math: fma, fmaf, fmal produce wrong results | |
4719 | [14259] build: --localedir arg to configure is ignored | |
4720 | [14499] nptl: Does posix_spawn invoke atfork handlers / use vfork? | |
4721 | [14750] libc: Race condition in posix_spawn vfork usage vs signal handlers | |
4722 | [14934] localedata: es_CL: wrong first weekday chilean locale | |
4723 | [15262] localedata: LC_MESSAGES.yesexpr/noexpr: inconsistent use of | |
4724 | romanisation | |
4725 | [15263] localedata: LC_MESSAGES.yesexpr/noexpr: inconsistent use of 1/0 | |
4726 | and +/- | |
4727 | [15264] localedata: LC_MESSAGES.yesstr/nostr: lacking in many locales | |
4728 | [15368] nptl: raise() is not async-signal-safe | |
4729 | [15479] math: ceil, floor, round and trunc raise inexact exception | |
4730 | [15578] localedata: kk_KZ: various updates | |
4731 | [16003] localedata: pap_AN: punt old locale | |
4732 | [16137] localedata: iw_IL: punt old locale | |
4733 | [16190] localedata: eo: new esperanto locale | |
4734 | [16374] localedata: lv_LV: change currency symbol in LC_MONETARY to euro | |
4735 | [16742] malloc: race condition: pthread_atfork() called before first | |
4736 | malloc() results in unexpected locking behaviour/deadlocks | |
4737 | [16975] localedata: LC_MESSAGES.yesexpr/noexpr: revisit capitalization in | |
4738 | all locales | |
4739 | [16983] localedata: postal_fmt does not allow %l and %n modifiers | |
4740 | [17565] localedata: pt_PT: wrong (work-)week start | |
4741 | [17899] math: [powerpc] floorl returns negative zero with FE_DOWNWARD | |
4742 | [17950] build: Build fails with -msse | |
4743 | [18205] localedata: be_BY*: wrong first_weekday and first_workday | |
4744 | [18433] libc: posix_spawn does not return correctly upon failure to | |
4745 | execute | |
4746 | [18453] localedata: charmaps/IBM875: incorrect codes | |
4747 | [18712] string: bits/string2.h incompatible with -O2 -Werror=packed | |
4748 | -Wsystem-headers | |
4749 | [18896] localedata: he_IL: improvements for currency | |
4750 | [18911] localedata: ro_RO: Correcting week day name for "Tuesday" in | |
4751 | Romanian locale data | |
4752 | [18960] locale: s390: _nl_locale_subfreeres uses larl opcode on misaligned | |
4753 | symbol | |
4754 | [19056] libc: Deprecate readdir_r | |
4755 | [19133] localedata: pt_*: days & months should be lowercase in Portuguese | |
4756 | language | |
4757 | [19198] localedata: nl_NL: small improvements for Dutch locales | |
4758 | [19257] network: Per-thread memory leak in __res_vinit with IPv6 | |
4759 | nameservers (CVE-2016-5417) | |
4760 | [19269] build: tst-audit4 and tst-audit10 failures with gcc-6 on non avx | |
4761 | machine | |
4762 | [19400] locale: Language missing in "iso-639.def", trivial fix in | |
4763 | description | |
4764 | [19431] malloc: Deadlock between fflush, getdelim, and fork | |
4765 | [19505] libc: Incorrect file descriptor validity checks in | |
4766 | posix_spawn_file_actions_add{open,close,dup2} | |
4767 | [19509] dynamic-link: dlsym, dlvsym do not report errors through dlerror | |
4768 | when using RTLD_NEXT | |
4769 | [19512] locale: Stale `#ifndef HAVE_BUILTIN_EXPECT' in | |
4770 | `intl/{gettextP,loadinfo}.h' | |
4771 | [19534] libc: execle, execlp may use malloc | |
4772 | [19568] localedata: *_CH: Swiss locales have inconsistent start of week | |
4773 | [19573] network: res_nclose and __res_maybe_init disagree about name | |
4774 | server initialization, breaking Hesiod | |
4775 | [19575] localedata: Status of GB18030 tables | |
4776 | [19581] localedata: sr_* date_fmt string contains additional newline | |
4777 | [19583] string: SSSE3_Fast_Copy_Backward flag needs to be enabled for AMD | |
4778 | Excavator core | |
4779 | [19592] math: [ldbl-128ibm] ceill incorrect in non-default rounding modes | |
4780 | [19593] math: [ldbl-128ibm] truncl incorrect in non-default rounding modes | |
4781 | [19594] math: [ldbl-128ibm] roundl incorrect in non-default rounding modes | |
4782 | [19595] math: [ldbl-128ibm] fmodl incorrect for results in subnormal | |
4783 | double range | |
4784 | [19602] math: [ldbl-128ibm] fmodl handling of equal arguments with low | |
4785 | part zero incorrect | |
4786 | [19603] math: [ldbl-128ibm] remainderl, remquol incorrect sign handling in | |
4787 | equality tests | |
4788 | [19610] dynamic-link: ldconfig -X removes stale symbolic links | |
4789 | [19613] libc: s390x (64 bit) macro expansion WCOREDUMP and others | |
4790 | [19633] locale: strfmon_l applies global locale to number formatting | |
4791 | [19642] network: Memory leak in getnameinfo | |
4792 | [19648] libc: test-skeleton.c: Do not set RLIMIT_DATA | |
4793 | [19653] libc: Potential for NULL pointer dereference (CWE-476) in | |
4794 | glibc-2.22 | |
4795 | [19654] math: [x86_64] Need testcase for BZ #19590 fix | |
4796 | [19671] localedata: Missing Sanity Check for malloc() in 'tst-fmon.c' & | |
4797 | 'tst-numeric.c' | |
4798 | [19674] math: [ldbl-128ibm] powl incorrect overflow handling | |
4799 | [19677] math: [ldbl-128ibm] remainderl equality test incorrect for zero | |
4800 | low part | |
4801 | [19678] math: [ldbl-128ibm] nextafterl, nexttowardl incorrect sign of zero | |
4802 | result | |
4803 | [19679] dynamic-link: gcc-4.9.3 C++ exception handling broken due to | |
4804 | unaligned stack | |
4805 | [19726] locale: Converting UCS4LE to INTERNAL with iconv() does not update | |
4806 | pointers and lengths in error-case. | |
4807 | [19727] locale: Converting from/to UTF-xx with iconv() does not always | |
4808 | report errors on UTF-16 surrogates values. | |
4809 | [19755] nscd: nscd assertion failure in gc | |
4810 | [19758] dynamic-link: Typo in EXTRA_LD_ENVVARS for x86-64 | |
4811 | [19759] libc: mempcpy shouldn't be inlined | |
4812 | [19762] dynamic-link: HAS_CPU_FEATURE/HAS_ARCH_FEATURE are easy to misuse | |
4813 | [19765] libc: s390 needs an optimized mempcpy | |
4814 | [19779] glob: glob: buffer overflow with GLOB_ALTDIRFUNC due to incorrect | |
4815 | NAME_MAX limit assumption (CVE-2016-1234) | |
4816 | [19783] build: benchtests don't support --enable-hardcoded-path-in-tests | |
4817 | [19787] network: Missing and incorrect truncation checks in getnameinfo | |
4818 | [19790] math: [ldbl-128ibm] nearbyintl incorrect in non-default rounding | |
4819 | modes | |
4820 | [19791] network: Assertion failure in res_query.c with un-connectable name | |
4821 | server addresses | |
4822 | [19792] libc: MIPS: backtrace yields infinite backtrace with makecontext | |
4823 | [19822] math: libm.so install clobbers old version | |
4824 | [19825] network: resolv: send_vc can return uninitialized data in second | |
4825 | response to getaddrinfo | |
4826 | [19830] network: nss_dns: should check RDATA length against buffer length | |
4827 | [19831] network: nss_dns: getaddrinfo returns uninitialized data when | |
4828 | confronted with A/AAAA records of invalid size | |
4829 | [19837] nss: nss_db: No retries for some long lines with a larger buffer | |
4830 | [19848] math: powl(10,n) for n=-4,-5,-6,-7 is off by more than 1 ULP | |
4831 | [19853] stdio: Printing IBM long double in decimal with high precision is | |
4832 | sometimes incorrect | |
4833 | [19860] build: x86_64: compile errors for tst-audit10 and tst-auditmod10b | |
4834 | [19861] nptl: libpthread IFUNC resolver for fork can lead to crash | |
4835 | [19862] network: resolv, nss_dns: Remove remaining logging of unexpected | |
4836 | record types | |
4837 | [19865] network: Assertion failure or memory leak in | |
4838 | _nss_dns_getcanonname_r | |
4839 | [19868] network: nss_dns: netent code does not skip over non-PTR records | |
4840 | [19879] network: nss_dns: Stack overflow in getnetbyname implementation | |
4841 | (CVE-2016-3075) | |
4842 | [19881] string: Improve x86-64 memset | |
4843 | [19907] string: Incorrect memcpy tests | |
4844 | [19916] dynamic-link: S390: fprs/vrs are not saved/restored while | |
4845 | resolving symbols | |
4846 | [19925] libc: termios.h XCASE namespace | |
4847 | [19928] string: memmove-vec-unaligned-erms.S is slow with large data size | |
4848 | [19929] libc: limits.h NL_NMAX namespace | |
4849 | [19931] stdio: Memory leak in vfprintf | |
4850 | [19957] libc: clone(CLONE_VM) access invalid parent memory | |
4851 | [19963] localedata: en_IL: New locale | |
4852 | [19989] stdio: stdio.h cuserid namespace | |
4853 | [19994] network: getaddrinfo does not restore RES_USE_INET6 flag in | |
4854 | gethosts | |
4855 | [19996] locale: langinfo.h nl_langinfo_l namespace | |
4856 | [20005] stdio: fflush on a file opened with fmemopen resets position to 0 | |
4857 | [20010] network: getaddrinfo: Stack overflow in hostent translation | |
4858 | (CVE-2016-3706) | |
4859 | [20012] stdio: libio: fmemopen append mode failure | |
4860 | [20014] stdio: stdio.h namespace for pre-threads POSIX | |
4861 | [20017] network: resolv: Use gmtime_r instead of gmtime in p_secstodate | |
4862 | [20023] libc: fcntl.h timespec namespace | |
4863 | [20024] math: [x86_64] vectorized sincos trashes the stack | |
4864 | [20031] network: nss_hesiod: Heap overflow in get_txt_records | |
4865 | [20041] time: sys/time.h timespec namespace | |
4866 | [20043] libc: unistd.h missing cuserid for UNIX98 and before | |
4867 | [20044] libc: unistd.h missing pthread_atfork for UNIX98 | |
4868 | [20051] libc: ttyslot in wrong header under wrong conditions | |
4869 | [20054] libc: gethostname not declared for XPG4 | |
4870 | [20055] libc: termios.h missing tcgetsid for XPG4 | |
4871 | [20072] dynamic-link: x86 init_cpu_features is called twice in static | |
4872 | executable | |
4873 | [20073] libc: sys/stat.h fchmod namespace | |
4874 | [20074] libc: stdlib.h rand_r namespace | |
4875 | [20076] libc: sys/stat.h missing S_IFSOCK, S_ISSOCK for XPG4 | |
4876 | [20094] libc: stdlib.h should not declare grantpt, ptsname, unlockpt for | |
4877 | XPG3 | |
4878 | [20111] libc: struct sockaddr_storage cannot be aggregate-copied | |
4879 | [20112] network: sunrpc: stack (frame) overflow in Sun RPC clntudp_call | |
4880 | (CVE-2016-4429) | |
4881 | [20115] string: Extra alignment in memset-vec-unaligned-erms.S | |
4882 | [20119] libc: Wrong mask for processors level type from CPUID | |
4883 | [20139] dynamic-link: Upper part of zmm is zeroed if Glibc is built with | |
4884 | AS not supporting AVX512 | |
4885 | [20151] math: [ldbl-128/ldbl-128ibm] j0l, j1l, y0l, y1l return sNaN for | |
4886 | sNaN argument | |
4887 | [20153] math: [ldbl-128ibm] sqrtl (sNaN) returns sNaN | |
4888 | [20156] math: [ldbl-128ibm] ceill, rintl etc. return sNaN for sNaN | |
4889 | argument | |
4890 | [20157] math: [powerpc] fabsl (sNaN) wrongly raises "invalid" | |
4891 | [20160] math: [powerpc] ceil, rint etc. return sNaN for sNaN input | |
4892 | [20178] libc: posix_spawn{p} should not call exit | |
4893 | [20191] stdio: libio: vtables hardening | |
4894 | [20195] string: FMA4 detection requires CPUID execution with register | |
4895 | eax=0x80000001 | |
4896 | [20198] libc: quick_exit incorrectly destroys C++11 thread objects. | |
4897 | [20205] math: [i386/x86_64] nextafterl incorrect incrementing negative | |
4898 | subnormals | |
4899 | [20212] math: acos (sNaN) returns sNaN | |
4900 | [20213] math: asin (sNaN) returns sNaN | |
4901 | [20214] network: Linux header sync with linux/in6.h and ipv6.h again. | |
4902 | [20218] math: [i386] asinhl (sNaN) returns sNaN | |
4903 | [20219] math: [i386] atanhl (sNaN) returns sNaN | |
4904 | [20222] stdio: fopencookie: Mangle function pointers | |
4905 | [20224] math: [i386] cbrtl (sNaN) returns sNaN | |
4906 | [20225] math: ldexp, scalbn, scalbln return sNaN for sNaN input | |
4907 | [20226] math: [i386/x86_64] expl, exp10l, expm1l return sNaN for sNaN | |
4908 | input | |
4909 | [20227] math: [i386/x86_64] logl (sNaN) returns sNaN | |
4910 | [20228] math: [i386/x86_64] log10l (sNaN) returns sNaN | |
4911 | [20229] math: [i386/x86_64] log1pl (sNaN) returns sNaN | |
4912 | [20232] math: [ldbl-128] expm1l (sNaN) returns sNaN | |
4913 | [20233] math: [ldbl-128ibm] expm1l (sNaN) returns sNaN | |
4914 | [20234] math: [ldbl-128ibm] log1pl (sNaN) returns sNaN | |
4915 | [20235] math: [i386/x86_64] log2l (sNaN) returns sNaN | |
4916 | [20237] nss: nss_db: get*ent segfaults without preceding set*ent | |
4917 | [20240] math: modf (sNaN) returns sNaN | |
4918 | [20248] libc: debug/tst-longjump_chk2 calls printf from a signal handler | |
4919 | [20250] math: frexp (sNaN) returns sNaN | |
4920 | [20252] math: atan2 (sNaN, qNaN) fails to raise "invalid" | |
4921 | [20255] math: [i386] fdim, fdimf return with excess range and precision / | |
4922 | double rounding | |
4923 | [20256] math: [i386/x86_64] fdiml returns sNaN for sNaN input | |
4924 | [20260] string: ../sysdeps/x86/bits/string.h:1092:3: error: array | |
4925 | subscript is below array bounds [-Werror=array-bounds] | |
4926 | [20262] nis: _nss_nis_initgroups_dyn always returns NSS_STATUS_NOTFOUND | |
4927 | [20263] nptl: robust mutex deadlocks if other thread requests timedlock | |
4928 | (Only arm/linux) | |
4929 | [20277] libc: $dp is not initialized correctly in sysdeps/hppa/start.S | |
4930 | [20284] malloc: malloc: Corrupt arena avoidance causes unnecessary mmap | |
4931 | fallbacks | |
4932 | [20296] math: [i386/x86_64] scalbl returns sNaN for sNaN input, missing | |
4933 | "invalid" exceptions | |
4934 | [20314] nptl: make[4]: *** [/usr/include/stdlib.h] Error 1 | |
4935 | [20316] localedata: id_ID: Februari instead of Pebruari | |
4936 | [20327] string: POWER8 strcasecmp returns incorrect result | |
4937 | [20347] math: Failure: Test: j0_downward (0xap+0) | |
4938 | [20348] libc: FAIL: misc/tst-preadvwritev64 | |
4939 | [20349] libc: 64-bit value is passed differently in p{readv,writev}{64} | |
4940 | [20350] libc: There is no test for p{read,write}64 | |
4941 | [20357] math: Incorrect cos result for 1.5174239687223976 | |
4942 | [20384] build: Don't run libmvec-sincos-avx* tests on non avx machines | |
4943 | \f | |
4944 | Version 2.23 | |
4945 | ||
4946 | * Unicode 8.0.0 Support: Character encoding, character type info, and | |
4947 | transliteration tables are all updated to Unicode 8.0.0, using new | |
4948 | and/or improved generator scripts contributed by Mike FABIAN (Red Hat). | |
4949 | These updates cause user visible changes, such as the fixes for bugs | |
4950 | 89, 16061, and 18568. | |
4951 | ||
4952 | * sched_setaffinity, pthread_setaffinity_np no longer attempt to guess the | |
4953 | kernel-internal CPU set size. This means that requests that change the | |
4954 | CPU affinity which failed before (for example, an all-ones CPU mask) will | |
4955 | now succeed. Applications that need to determine the effective CPU | |
4956 | affinities need to call sched_getaffinity or pthread_getaffinity_np after | |
4957 | setting it because the kernel can adjust it (and the previous size check | |
4958 | would not detect this in the majority of cases). | |
4959 | ||
4960 | * The fts.h header can now be used with -D_FILE_OFFSET_BITS=64. With LFS | |
4961 | the following new symbols are used: fts64_children, fts64_close, | |
4962 | fts64_open, fts64_read and fts64_set. | |
4963 | ||
4964 | * getaddrinfo now detects certain invalid responses on an internal netlink | |
4965 | socket. If such responses are received, an affected process will | |
4966 | terminate with an error message of "Unexpected error <number> on netlink | |
4967 | descriptor <number>" or "Unexpected netlink response of size <number> on | |
4968 | descriptor <number>". The most likely cause for these errors is a | |
4969 | multi-threaded application which erroneously closes and reuses the netlink | |
4970 | file descriptor while it is used by getaddrinfo. | |
4971 | ||
4972 | * A defect in the malloc implementation, present since glibc 2.15 (2012) or | |
4973 | glibc 2.10 via --enable-experimental-malloc (2009), could result in the | |
4974 | unnecessary serialization of memory allocation requests across threads. | |
4975 | The defect is now corrected. Users should see a substantial increase in | |
4976 | the concurrent throughput of allocation requests for applications which | |
4977 | trigger this bug. Affected applications typically create and | |
4978 | destroy threads frequently. (Bug 19048 was reported and analyzed by | |
4979 | Ericsson.) | |
4980 | ||
4981 | * There is now a --disable-timezone-tools configure option for disabling the | |
4982 | building and installing of the timezone related utilities (zic, zdump, and | |
4983 | tzselect). This is useful for people who build the timezone data and code | |
4984 | independent of the GNU C Library. | |
4985 | ||
4986 | * The obsolete header <regexp.h> has been removed. Programs that require | |
4987 | this header must be updated to use <regex.h> instead. | |
4988 | ||
4989 | * The obsolete functions bdflush, create_module, get_kernel_syms, | |
4990 | query_module and uselib are no longer available to newly linked binaries; | |
4991 | the header <sys/kdaemon.h> has been removed. These functions and header | |
4992 | were specific to systems using the Linux kernel and could not usefully be | |
4993 | used with the GNU C Library on systems with version 2.6 or later of the | |
4994 | Linux kernel. | |
4995 | ||
4996 | * Optimized string, wcsmbs and memory functions for IBM z13. | |
4997 | Implemented by Stefan Liebler. | |
4998 | ||
4999 | * Newly linked programs that define a variable called signgam will no longer | |
5000 | have it set by the lgamma, lgammaf and lgammal functions. Programs that | |
5001 | require signgam to be set by those functions must ensure that they use the | |
5002 | variable provided by the GNU C Library and declared in <math.h>, without | |
5003 | defining their own copy. | |
5004 | ||
5005 | * The minimum GCC version that can be used to build this version of the GNU | |
5006 | C Library is GCC 4.7. Older GCC versions, and non-GNU compilers, can | |
5007 | still be used to compile programs using the GNU C Library. | |
5008 | ||
5009 | Security related changes: | |
5010 | ||
5011 | * An out-of-bounds value in a broken-out struct tm argument to strftime no | |
5012 | longer causes a crash. Reported by Adam Nielsen. (CVE-2015-8776) | |
5013 | ||
5014 | * The LD_POINTER_GUARD environment variable can no longer be used to disable | |
5015 | the pointer guard feature. It is always enabled. Previously, | |
5016 | LD_POINTER_GUARD could be used to disable security hardening in binaries | |
5017 | running in privileged AT_SECURE mode. Reported by Hector Marco-Gisbert. | |
5018 | (CVE-2015-8777) | |
5019 | ||
5020 | * An integer overflow in hcreate and hcreate_r could lead to an | |
5021 | out-of-bounds memory access. Reported by Szabolcs Nagy. (CVE-2015-8778) | |
5022 | ||
5023 | * The catopen function no longer has unbounded stack usage. Reported by | |
5024 | Max. (CVE-2015-8779) | |
5025 | ||
5026 | * The nan, nanf and nanl functions no longer have unbounded stack usage | |
5027 | depending on the length of the string passed as an argument to the | |
5028 | functions. Reported by Joseph Myers. (CVE-2014-9761) | |
5029 | ||
5030 | * A stack-based buffer overflow was found in libresolv when invoked from | |
5031 | libnss_dns, allowing specially crafted DNS responses to seize control | |
5032 | of execution flow in the DNS client. The buffer overflow occurs in | |
5033 | the functions send_dg (send datagram) and send_vc (send TCP) for the | |
5034 | NSS module libnss_dns.so.2 when calling getaddrinfo with AF_UNSPEC | |
5035 | family. The use of AF_UNSPEC triggers the low-level resolver code to | |
5036 | send out two parallel queries for A and AAAA. A mismanagement of the | |
5037 | buffers used for those queries could result in the response of a query | |
5038 | writing beyond the alloca allocated buffer created by | |
5039 | _nss_dns_gethostbyname4_r. Buffer management is simplified to remove | |
5040 | the overflow. Thanks to the Google Security Team and Red Hat for | |
5041 | reporting the security impact of this issue, and Robert Holiday of | |
5042 | Ciena for reporting the related bug 18665. (CVE-2015-7547) | |
5043 | ||
5044 | The following bugs are resolved with this release: | |
5045 | ||
5046 | [89] localedata: Locales nb_NO and nn_NO should transliterate æøå | |
5047 | [887] math: Math library function "logb" and "nextafter" inconsistent | |
5048 | [2542] math: Incorrect return from float gamma (-0X1.FA471547C2FE5P+1) | |
5049 | [2543] math: Incorrect return from float gamma (-0X1.9260DCP+1) | |
5050 | [2558] math: Incorrect return from double gamma (-0X1.FA471547C2FE5P+1) | |
5051 | [2898] libc: [improve] warning: the use of `mktemp' is dangerous, better | |
5052 | use `mkstemp' | |
5053 | [4404] localedata: German translation of "Alarm clock" is misleading | |
5054 | [6799] math: nextafter() and nexttoward() doesn't set errno on | |
5055 | overflow/underflow errors | |
5056 | [6803] math: scalb(), scalbln(), scalbn() do not set errno on | |
5057 | overflow/underflow | |
5058 | [10432] nis: _nss_nis_setnetgrent assertion failure | |
5059 | [11460] libc: fts has no LFS support | |
5060 | [12926] network: getaddrinfo()/make_request() may spin forever | |
5061 | [13065] nptl: Race condition in pthread barriers | |
5062 | [13690] nptl: pthread_mutex_unlock potentially cause invalid access | |
5063 | [14341] dynamic-link: Dynamic linker crash when DT_JMPREL and DT_REL{,A} | |
5064 | are not contiguous | |
5065 | [14551] math: [ldbl-128ibm] strtold overflow handling for IBM long double | |
5066 | [14912] libc: Rename non-installed bits/*.h headers | |
5067 | [15002] libc: Avoid undefined behavior in posix_fallocate overflow check | |
5068 | [15367] math: Let gcc use __builtin_isinf | |
5069 | [15384] math: One constant fewer in ieee754/dbl-64/wordsize-64/s_finite.c | |
5070 | [15421] math: lgamma wrongly sets signgam for ISO C | |
5071 | [15470] math: [arm] On ARM llrintl() and llroundl() do not raise | |
5072 | FE_INVALID with argument out of range | |
5073 | [15491] math: [i386/x86_64] x86 nearbyint implementations wrongly clear | |
5074 | all exceptions | |
5075 | [15786] dynamic-link: ifunc resolver functions can smash function | |
5076 | arguments | |
5077 | [15918] math: Unnecessary check for equality in hypotf() | |
5078 | [16061] localedata: Review / update transliteration data | |
5079 | [16068] math: [i386/x86_64] x86 and x86_64 fesetenv exclude state they | |
5080 | should include | |
5081 | [16141] time: strptime %z offset restriction | |
5082 | [16171] math: drem should be alias of remainder | |
5083 | [16296] math: fegetround is pure? | |
5084 | [16347] math: [ldbl-128ibm] ldbl-128/e_lgammal_r.c may not be suitable. | |
5085 | [16364] libc: sleep may leave SIGCHLD blocked on sync cancellation on | |
5086 | GNU/Linux | |
5087 | [16399] math: [mips] lrint / llrint / lround / llround missing exceptions | |
5088 | [16415] math: Clean up ldbl-128 / ldbl-128ibm expm1l for large positive | |
5089 | arguments | |
5090 | [16422] math: [powerpc] math-float, math-double failing llrint tests with | |
5091 | "Exception "Inexact" set" on ppc32 | |
5092 | [16495] localedata: nl_NL: date_fmt: shuffle year/month around | |
5093 | [16517] math: Missing underflow exception from tanf/tan/tanl | |
5094 | [16519] math: Missing underflow exception from sinhf | |
5095 | [16520] math: Missing underflow exception from tanhf | |
5096 | [16521] math: Missing underflow exception from exp2 | |
5097 | [16620] math: [ldbl-128ibm] exp10l spurious overflows / bad directed | |
5098 | rounding results | |
5099 | [16734] stdio: fopen calls mmap to allocate its buffer | |
5100 | [16961] math: nan function incorrect handling of bad sequences | |
5101 | [16962] math: nan function unbounded stack allocation (CVE-2014-9761) | |
5102 | [16973] localedata: Fix lang_lib/lang_term as per ISO 639-2 | |
5103 | [16985] locale: localedef: confusing error message when opening output | |
5104 | fails | |
5105 | [17118] math: ctanh(INFINITY + 2 * I) returns incorrect value | |
5106 | [17197] locale: Redundant shift character in iconv conversion output at | |
5107 | block boundary | |
5108 | [17243] libc: trunk/posix/execl.c:53: va_args problem ? | |
5109 | [17244] libc: trunk/sysdeps/unix/sysv/linux/semctl.c:116: va_args muxup ? | |
5110 | [17250] dynamic-link: static linking breaks nss loading | |
5111 | (getaddrinfo/getpwnam/etc...) | |
5112 | [17404] libc: atomic_exchange_rel lacking a barrier on MIPS16, GCC before | |
5113 | 4.7? | |
5114 | [17441] math: isnan() should use __builtin_isnan() in GCC | |
5115 | [17514] nptl: Assert failure unlocking ERRORCHECK mutex after timedlock | |
5116 | (related to lock elision) | |
5117 | [17787] manual: Exponent on page 324 of the PDF ends prematurely | |
5118 | [17886] time: strptime should be able to parse "Z" as a timezone with %z | |
5119 | [17887] time: strptime should be able to parse "+01:00" style timezones | |
5120 | [17905] libc: catopen() Multiple unbounded stack allocations | |
5121 | (CVE-2015-8779) | |
5122 | [18084] libc: backtrace (..., 0) dumps core on x86 | |
5123 | [18086] libc: nice() sets errno to 0 on success | |
5124 | [18240] libc: hcreate, hcreate_r should fail with ENOMEM if element count | |
5125 | is too large (CVE-2015-8778) | |
5126 | [18251] dynamic-link: SONAME missing when audit modules provides path | |
5127 | [18265] libc: add attributes for wchar string and memory functions | |
5128 | [18370] math: csqrt missing underflows | |
5129 | [18421] libc: [hppa] read-only segment has dynamic relocations | |
5130 | [18472] libc: Obsolete syscall wrappers should be compat symbols | |
5131 | [18480] libc: hppa glibc miscompilation in sched_setaffinity() | |
5132 | [18491] localedata: Update tr_TR LC_CTYPE as part of Unicode updates | |
5133 | [18525] localedata: Remove locale timezone information | |
5134 | [18560] libc: [powerpc] spurious bits/ipc.h definitions | |
5135 | [18568] localedata: Update locale data to Unicode 8.0 | |
5136 | [18589] locale: sort-test.sh fails at random | |
5137 | [18595] math: ctan, ctanh missing underflows | |
5138 | [18604] libc: assert macro-expands its argument | |
5139 | [18610] math: S390: fetestexcept() reports any exception if DXC-code | |
5140 | contains a vector instruction exception. | |
5141 | [18611] math: j1, jn missing errno setting on underflow | |
5142 | [18618] localedata: sync Chechen locale definitions with other *_RU | |
5143 | locales | |
5144 | [18647] math: powf(-0x1.000002p0, 0x1p30) returns 0 instead of +inf | |
5145 | [18661] libc: Some x86-64 assembly codes don't align stack to 16 bytes | |
5146 | [18665] network: In send_dg, the recvfrom function is NOT always using the | |
5147 | buffer size of a newly created buffer (CVE-2015-7547) | |
5148 | [18674] libc: [i386] trunk/sysdeps/i386/tst-auditmod3b.c:84: possible | |
5149 | missing break ? | |
5150 | [18675] libc: fpathconf(_PC_NAME_MAX) fails against large filesystems for | |
5151 | 32bit processes | |
5152 | [18681] libc: regexp.h is obsolete and buggy, and should be desupported | |
5153 | [18699] math: tilegx cproj() for various complex infinities does not yield | |
5154 | infinity | |
5155 | [18724] libc: Harden put*ent functions against data injection | |
5156 | [18743] nptl: PowerPC: findutils testcase fails with --enable-lock-elision | |
5157 | [18755] build: build errors with -DNDEBUG | |
5158 | [18757] stdio: fmemopen fails to set errno on failure | |
5159 | [18778] dynamic-link: ld.so crashes if failed dlopen causes libpthread to | |
5160 | be forced unloaded | |
5161 | [18781] libc: openat64 lacks O_LARGEFILE | |
5162 | [18787] libc: [hppa] sysdeps/unix/sysv/linux/hppa/bits/atomic.h:71:6: | |
5163 | error: can’t find a register in class ‘R1_REGS’ while reloading ‘asm’ | |
5164 | [18789] math: [ldbl-128ibm] sinhl inaccurate near 0 | |
5165 | [18790] math: [ldbl-128ibm] tanhl inaccurate | |
5166 | [18795] libc: stpncpy fortification misses buffer lengths that are | |
5167 | statically too large | |
5168 | [18796] build: build fails for --disable-mathvec | |
5169 | [18803] math: hypot missing underflows | |
5170 | [18820] stdio: fmemopen may leak memory on failure | |
5171 | [18823] math: csqrt spurious underflows | |
5172 | [18824] math: fma spurious underflows | |
5173 | [18825] math: pow missing underflows | |
5174 | [18857] math: [ldbl-128ibm] nearbyintl wrongly uses signaling comparisons | |
5175 | [18868] nptl: pthread_barrier_init typo has in-theory-undefined behavior | |
5176 | [18870] build: sem_open.c fails to compile with missing symbol | |
5177 | FUTEX_SHARED | |
5178 | [18872] stdio: Fix memory leak in printf_positional | |
5179 | [18873] libc: posix_fallocate overflow check ineffective | |
5180 | [18875] math: Excess precision leads incorrect libm | |
5181 | [18877] libc: arm: mmap offset regression | |
5182 | [18887] libc: memory corruption when using getmntent on blank lines | |
5183 | [18918] localedata: hu_HU: change time to HH:MM:SS format | |
5184 | [18921] libc: Regression: extraneous stat() and fstat() performed by | |
5185 | opendir() | |
5186 | [18928] dynamic-link: LD_POINTER_GUARD is not ignored for privileged | |
5187 | binaries (CVE-2015-8777) | |
5188 | [18951] math: tgamma missing underflows | |
5189 | [18952] math: [ldbl-128/ldbl-128ibm] lgammal spurious "invalid", incorrect | |
5190 | signgam | |
5191 | [18953] localedata: lt_LT: change currency symbol to the euro | |
5192 | [18956] math: powf inaccuracy | |
5193 | [18961] math: [i386] exp missing underflows | |
5194 | [18966] math: [i386] exp10 missing underflows | |
5195 | [18967] math: math.h XSI POSIX namespace (gamma, isnan, scalb) | |
5196 | [18969] build: multiple string test failures due to missing locale | |
5197 | dependencies | |
5198 | [18970] libc: Reference of pthread_setcancelstate in libc.a | |
5199 | [18977] math: float / long double Bessel functions not in XSI POSIX | |
5200 | [18980] math: i386 libm functions return with excess range and precision | |
5201 | [18981] math: i386 scalb*, ldexp return with excess range and precision | |
5202 | [18982] stdio: va_list and vprintf | |
5203 | [18985] time: Passing out of range data to strftime() causes a segfault | |
5204 | (CVE-2015-8776) | |
5205 | [19003] math: [x86_64] fma4 version of pow inappropriate contraction | |
5206 | [19007] libc: FAIL: elf/check-localplt with -z now and binutils 2.26 | |
5207 | [19012] locale: iconv_open leaks memory on error path | |
5208 | [19016] math: clog, clog10 inaccuracy | |
5209 | [19018] nptl: Mangle function pointers in tls_dtor_list | |
5210 | [19032] math: [i386] acosh (-qNaN) spurious "invalid" exception | |
5211 | [19046] math: ldbl-128 / ldbl-128ibm lgamma bad overflow handling | |
5212 | [19048] malloc: malloc: arena free list can become cyclic, increasing | |
5213 | contention | |
5214 | [19049] math: [powerpc] erfc incorrect zero sign | |
5215 | [19050] math: [powerpc] log* incorrect zero sign | |
5216 | [19058] math: [x86_64] Link fail with -fopenmp and -flto | |
5217 | [19059] math: nexttoward overflow incorrect in non-default rounding modes | |
5218 | [19071] math: ldbl-96 lroundl incorrect just below powers of 2 | |
5219 | [19074] network: Data race in _res_hconf_reorder_addrs | |
5220 | [19076] math: [ldbl-128ibm] log1pl (-1) wrong sign of infinity | |
5221 | [19077] math: [ldbl-128ibm] logl (1) incorrect sign of zero result | |
5222 | [19078] math: [ldbl-128ibm] expl overflow incorrect in non-default | |
5223 | rounding modes | |
5224 | [19079] math: dbl-64/wordsize-64 lround based on llround incorrect for | |
5225 | ILP32 | |
5226 | [19085] math: ldbl-128 lrintl, lroundl missing exceptions for 32-bit long | |
5227 | [19086] manual: posix_fallocate64 documented argument order is wrong. | |
5228 | [19088] math: lround, llround missing exceptions close to overflow | |
5229 | threshold | |
5230 | [19094] math: lrint, llrint missing exceptions close to overflow threshold | |
5231 | [19095] math: dbl-64 lrint incorrect for 64-bit long | |
5232 | [19122] dynamic-link: Unnecessary PLT relocations in librtld.os | |
5233 | [19124] dynamic-link: ld.so failed to build with older assembler | |
5234 | [19125] math: [powerpc32] llroundf, llround incorrect exceptions | |
5235 | [19129] dynamic-link: [arm] Concurrent lazy TLSDESC resolution can crash | |
5236 | [19134] math: [powerpc32] lround, lroundf spurious exceptions | |
5237 | [19137] libc: i386/epoll_pwait.S doesn't support cancellation | |
5238 | [19143] nptl: Remove CPU set size checking from sched_setaffinity, | |
5239 | pthread_setaffinity_np | |
5240 | [19156] math: [ldbl-128] j0l spurious underflows | |
5241 | [19164] nptl: tst-getcpu fails with many possible CPUs | |
5242 | [19168] math: math/test-ildoubl and math/test-ldouble failure | |
5243 | [19174] nptl: PowerPC: TLE enabled pthread mutex performs poorly. | |
5244 | [19178] dynamic-link: ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA confuses | |
5245 | prelink | |
5246 | [19181] math: [i386/x86_64] fesetenv (FE_DFL_ENV), fesetenv | |
5247 | (FE_NOMASK_ENV) do not clear SSE exceptions | |
5248 | [19182] malloc: malloc deadlock between ptmalloc_lock_all and | |
5249 | _int_new_arena/reused_arena | |
5250 | [19189] math: [ldbl-128] log1pl (-qNaN) spurious "invalid" exception | |
5251 | [19201] math: dbl-64 remainder incorrect sign of zero result | |
5252 | [19205] math: bits/math-finite.h conditions do not match math.h and | |
5253 | bits/mathcalls.h | |
5254 | [19209] math: bits/math-finite.h wrongly maps ldexp to scalbn | |
5255 | [19211] math: lgamma functions do not set signgam for -ffinite-math-only | |
5256 | for C99-based standards | |
5257 | [19212] libc: features.h not -Wundef clean | |
5258 | [19213] math: [i386/x86_64] log* (1) incorrect zero sign for -ffinite- | |
5259 | math-only | |
5260 | [19214] libc: Family and model identification for AMD CPU's are incorrect. | |
5261 | [19219] libc: GLIBC build fails for ia64 with missing __nearbyintl | |
5262 | [19228] math: [powerpc] nearbyint wrongly clears "inexact", leaves traps | |
5263 | disabled | |
5264 | [19235] math: [powerpc64] lround, lroundf, llround, llroundf spurious | |
5265 | "inexact" exceptions | |
5266 | [19238] math: [powerpc] round, roundf spurious "inexact" for integer | |
5267 | arguments | |
5268 | [19242] libc: strtol incorrect in Turkish locales | |
5269 | [19243] malloc: reused_arena can pick an arena on the free list, leading | |
5270 | to an assertion failure and reference count corruption | |
5271 | [19253] time: tzset() ineffective when temporary TZ did not include DST | |
5272 | rules | |
5273 | [19266] math: strtod ("NAN(I)") incorrect in Turkish locales | |
5274 | [19270] math: [hppa] Shared libm missing __isnanl | |
5275 | [19285] libc: [hppa] sysdeps/unix/sysv/linux/hppa/bits/mman.h: missing | |
5276 | MAP_HUGETLB and MAP_STACK defines | |
5277 | [19313] nptl: Wrong __cpu_mask for x32 | |
5278 | [19347] libc: grantpt: try to force a specific gid even without pt_chown | |
5279 | [19349] math: [ldbl-128ibm] tanhl inaccurate for small arguments | |
5280 | [19350] math: [ldbl-128ibm] sinhl spurious overflows | |
5281 | [19351] math: [ldbl-128ibm] logl inaccurate near 1 | |
5282 | [19363] time: x32: times() return value wrongly truncates/sign extends | |
5283 | from 32bit | |
5284 | [19367] dynamic-link: Improve branch prediction on Silvermont | |
5285 | [19369] network: Default domain name not reset by res_ninit when "search" | |
5286 | / "domain" entry is removed from resolv.conf | |
5287 | [19375] math: powerpc: incorrect results for POWER7 logb with negative | |
5288 | subnormals | |
5289 | [19385] localedata: bg_BG: time separator should be colon, not comma | |
5290 | [19408] libc: linux personality syscall wrapper may erroneously return an | |
5291 | error on 32-bit architectures | |
5292 | [19415] libc: dladdr returns wrong names on hppa | |
5293 | [19432] libc: iconv rejects redundant escape sequences in IBM900, IBM903, | |
5294 | IBM905, IBM907, and IBM909 | |
5295 | [19439] math: Unix98 isinf and isnan functions conflict with C++11 | |
5296 | [19443] build: build failures with -DDEBUG | |
5297 | [19451] build: Make check fails on test-double-vlen2 | |
5298 | [19462] libc: Glibc failed to build with -Os | |
5299 | [19465] math: Wrong code with -Os | |
5300 | [19466] time: time/tst-mktime2.c is compiled into an infinite loop with | |
5301 | -Os | |
5302 | [19467] string: Fast_Unaligned_Load needs to be enabled for Excavator core | |
5303 | CPU's. | |
5304 | [19475] libc: Glibc 2.22 doesn't build on sparc [PATCH] | |
5305 | [19486] math: S390: Math tests fail with "Exception Inexact set". | |
5306 | [19529] libc: [ARM]: FAIL: stdlib/tst-makecontext | |
5307 | [19550] libc: [mips] mmap negative offset handling inconsistent with other | |
5308 | architectures | |
5309 | [19590] math: Fail to build shared objects that use libmvec.so functions. | |
5310 | \f | |
5311 | Version 2.22 | |
5312 | ||
5313 | * The following bugs are resolved with this release: | |
5314 | ||
5315 | 438, 2981, 4719, 6544, 6792, 11216, 12836, 13028, 13064, 13151, 13152, | |
5316 | 14094, 14113, 14292, 14841, 14906, 14958, 15319, 15467, 15790, 15969, | |
5317 | 16159, 16339, 16350, 16351, 16352, 16353, 16361, 16512, 16526, 16538, | |
5318 | 16559, 16560, 16704, 16783, 16850, 17053, 17090, 17195, 17269, 17293, | |
5319 | 17322, 17403, 17475, 17523, 17542, 17569, 17581, 17588, 17596, 17620, | |
5320 | 17621, 17628, 17631, 17692, 17711, 17715, 17776, 17779, 17792, 17833, | |
5321 | 17836, 17841, 17912, 17916, 17930, 17932, 17944, 17949, 17964, 17965, | |
5322 | 17967, 17969, 17977, 17978, 17987, 17991, 17996, 17998, 17999, 18007, | |
5323 | 18019, 18020, 18029, 18030, 18032, 18034, 18036, 18038, 18039, 18042, | |
5324 | 18043, 18046, 18047, 18049, 18068, 18078, 18080, 18093, 18100, 18104, | |
5325 | 18110, 18111, 18116, 18125, 18128, 18134, 18138, 18185, 18196, 18197, | |
5326 | 18206, 18210, 18211, 18217, 18219, 18220, 18221, 18234, 18244, 18245, | |
5327 | 18247, 18287, 18319, 18324, 18333, 18346, 18371, 18383, 18397, 18400, | |
5328 | 18409, 18410, 18412, 18418, 18422, 18434, 18444, 18457, 18468, 18469, | |
5329 | 18470, 18479, 18483, 18495, 18496, 18497, 18498, 18502, 18507, 18508, | |
5330 | 18512, 18513, 18519, 18520, 18522, 18527, 18528, 18529, 18530, 18532, | |
5331 | 18533, 18534, 18536, 18539, 18540, 18542, 18544, 18545, 18546, 18547, | |
5332 | 18549, 18553, 18557, 18558, 18569, 18583, 18585, 18586, 18592, 18593, | |
5333 | 18594, 18602, 18612, 18613, 18619, 18633, 18635, 18641, 18643, 18648, | |
5334 | 18657, 18676, 18694, 18696, 18887. | |
5335 | ||
5336 | * Cache information can be queried via sysconf() function on s390 e.g. with | |
5337 | _SC_LEVEL1_ICACHE_SIZE as argument. | |
5338 | ||
5339 | * A buffer overflow in gethostbyname_r and related functions performing DNS | |
5340 | requests has been fixed. If the NSS functions were called with a | |
5341 | misaligned buffer, the buffer length change due to pointer alignment was | |
5342 | not taken into account. This could result in application crashes or, | |
5343 | potentially arbitrary code execution, using crafted, but syntactically | |
5344 | valid DNS responses. (CVE-2015-1781) | |
5345 | ||
5346 | * The time zone file parser has been made more robust against crafted time | |
5347 | zone files, avoiding heap buffer overflows related to the processing of | |
5348 | the tzh_ttisstdcnt and tzh_ttisgmtcnt fields, and a stack overflow due to | |
5349 | large time zone data files. Overly long time zone specifiers in the TZ | |
5350 | variable no longer result in stack overflows and crashes. | |
5351 | ||
5352 | * A powerpc and powerpc64 optimization for TLS, similar to TLS descriptors | |
5353 | for LD and GD on x86 and x86-64, has been implemented. You will need | |
5354 | binutils-2.24 or later to enable this optimization. | |
5355 | ||
5356 | * Character encoding and ctype tables were updated to Unicode 7.0.0, using | |
5357 | new generator scripts contributed by Pravin Satpute and Mike FABIAN (Red | |
5358 | Hat). These updates cause user visible changes, such as the fix for bug | |
5359 | 17998. | |
5360 | ||
5361 | * CVE-2014-8121 The NSS backends shared internal state between the getXXent | |
5362 | and getXXbyYY NSS calls for the same database, causing a denial-of-service | |
5363 | condition in some applications. | |
5364 | ||
5365 | * Added vector math library named libmvec with the following vectorized x86_64 | |
5366 | implementations: cos, cosf, sin, sinf, sincos, sincosf, log, logf, exp, expf, | |
5367 | pow, powf. | |
5368 | The library can be disabled with --disable-mathvec. Use of the functions is | |
5369 | enabled with -fopenmp -ffast-math starting from -O1 for GCC version >= 4.9.0. | |
5370 | Shared library libmvec.so is linked in as needed when using -lm (no need to | |
5371 | specify -lmvec explicitly for not static builds). | |
5372 | Visit <https://sourceware.org/glibc/wiki/libmvec> for detailed information. | |
5373 | ||
5374 | * A new fmemopen implementation has been added with the goal of POSIX | |
5375 | compliance. The new implementation fixes the following long-standing | |
5376 | issues: BZ#6544, BZ#11216, BZ#12836, BZ#13151, BZ#13152, and BZ#14292. The | |
5377 | old implementation is still present for use be by existing binaries. | |
5378 | ||
5379 | * The 32-bit sparc sigaction ABI was inadvertently broken in the 2.20 and 2.21 | |
5380 | releases. It has been fixed to match 2.19 and older, but binaries built | |
5381 | against 2.20 and 2.21 might need to be recompiled. See BZ#18694. | |
5382 | ||
5383 | * Port to Native Client running on ARMv7-A (--host=arm-nacl). | |
5384 | Contributed by Roland McGrath (Google). | |
5385 | ||
5386 | * The header <regexp.h> is deprecated, and will be removed in a future | |
5387 | release. Use of this header will trigger a deprecation warning. | |
5388 | Application developers should update their code to use <regex.h> instead. | |
5389 | ||
5390 | This header was formerly part of SUS, but was deprecated in 1994 and | |
5391 | removed from the standard in 2001. Also, the glibc implementation | |
5392 | leaks memory. See BZ#18681 for more details. | |
5393 | \f | |
5394 | Version 2.21 | |
5395 | ||
5396 | * The following bugs are resolved with this release: | |
5397 | ||
5398 | 6652, 10672, 12674, 12847, 12926, 13862, 14132, 14138, 14171, 14498, | |
5399 | 15215, 15378, 15884, 16009, 16418, 16191, 16469, 16576, 16617, 16618, | |
5400 | 16619, 16657, 16740, 16857, 17192, 17266, 17273, 17344, 17363, 17370, | |
5401 | 17371, 17411, 17460, 17475, 17485, 17501, 17506, 17508, 17522, 17555, | |
5402 | 17570, 17571, 17572, 17573, 17574, 17582, 17583, 17584, 17585, 17589, | |
5403 | 17594, 17601, 17608, 17616, 17625, 17630, 17633, 17634, 17635, 17647, | |
5404 | 17653, 17657, 17658, 17664, 17665, 17668, 17682, 17702, 17717, 17719, | |
5405 | 17722, 17723, 17724, 17725, 17732, 17733, 17744, 17745, 17746, 17747, | |
5406 | 17748, 17775, 17777, 17780, 17781, 17782, 17791, 17793, 17796, 17797, | |
5407 | 17801, 17803, 17806, 17834, 17844, 17848, 17868, 17869, 17870, 17885, | |
5408 | 17892. | |
5409 | ||
5410 | * CVE-2015-1472 CVE-2015-1473 Under certain conditions wscanf can allocate | |
5411 | too little memory for the to-be-scanned arguments and overflow the | |
5412 | allocated buffer. The implementation now correctly computes the required | |
5413 | buffer size when using malloc, and switches to malloc from alloca as | |
5414 | intended. | |
5415 | ||
5416 | * A new semaphore algorithm has been implemented in generic C code for all | |
5417 | machines. Previous custom assembly implementations of semaphore were | |
5418 | difficult to reason about or ensure that they were safe. The new version | |
5419 | of semaphore supports machines with 64-bit or 32-bit atomic operations. | |
5420 | The new semaphore algorithm is used by sem_init, sem_open, sem_post, | |
5421 | sem_wait, sem_timedwait, sem_trywait, and sem_getvalue. | |
5422 | ||
5423 | * Port to Altera Nios II has been contributed by Mentor Graphics. | |
5424 | ||
5425 | * Optimized strcpy, stpcpy, strncpy, stpncpy, strcmp, and strncmp | |
5426 | implementations for powerpc64/powerpc64le. | |
5427 | Implemented by Adhemerval Zanella (IBM). | |
5428 | ||
5429 | * Added support for TSX lock elision of pthread mutexes on powerpc32, powerpc64 | |
5430 | and powerpc64le. This may improve lock scaling of existing programs on | |
5431 | HTM capable systems. The lock elision code is only enabled with | |
5432 | --enable-lock-elision=yes. Also, the TSX lock elision implementation for | |
5433 | powerpc will issue a transaction abort on every syscall to avoid side | |
5434 | effects being visible outside transactions. | |
5435 | ||
5436 | * Optimized strcpy, stpcpy, strchrnul and strrchr implementations for | |
5437 | AArch64. Contributed by ARM Ltd. | |
5438 | ||
5439 | * i386 memcpy functions optimized with SSE2 unaligned load/store. | |
5440 | ||
5441 | * CVE-2104-7817 The wordexp function could ignore the WRDE_NOCMD flag | |
5442 | under certain input conditions resulting in the execution of a shell for | |
5443 | command substitution when the application did not request it. The | |
5444 | implementation now checks WRDE_NOCMD immediately before executing the | |
5445 | shell and returns the error WRDE_CMDSUB as expected. | |
5446 | ||
5447 | * CVE-2012-3406 printf-style functions could run into a stack overflow when | |
5448 | processing format strings with a large number of format specifiers. | |
5449 | ||
5450 | * CVE-2014-9402 The nss_dns implementation of getnetbyname could run into an | |
5451 | infinite loop if the DNS response contained a PTR record of an unexpected | |
5452 | format. | |
5453 | ||
5454 | * The minimum GCC version that can be used to build this version of the GNU | |
5455 | C Library is GCC 4.6. Older GCC versions, and non-GNU compilers, can | |
5456 | still be used to compile programs using the GNU C Library. | |
5457 | ||
5458 | * The GNU C Library is now built with -Werror by default. This can be | |
5459 | disabled by configuring with --disable-werror. | |
5460 | ||
5461 | * New locales: tu_IN, bh_IN, raj_IN, ce_RU. | |
5462 | ||
5463 | * The obsolete sigvec function has been removed. This was the original | |
5464 | 4.2BSD interface that inspired the POSIX.1 sigaction interface, which | |
5465 | programs have been using instead for about 25 years. Of course, ABI | |
5466 | compatibility for old binaries using sigvec remains intact. | |
5467 | ||
5468 | * Merged gettext 0.19.3 into the intl subdirectory. This fixes building | |
5469 | with newer versions of bison. | |
5470 | ||
5471 | * Support for MIPS o32 FPXX, FP64A and FP64 ABI Extensions. | |
5472 | The original MIPS o32 hard-float ABI requires an FPU where double-precision | |
5473 | registers overlay two consecutive single-precision registers. MIPS32R2 | |
5474 | introduced a new FPU mode (FR=1) where double-precision registers extend the | |
5475 | corresponding single-precision registers which is incompatible with the | |
5476 | o32 hard-float ABI. The MIPS SIMD ASE and the MIPSR6 architecture both | |
5477 | require the use of FR=1 making a transition necessary. New o32 ABI | |
5478 | extensions enable users to migrate over time from the original o32 ABI | |
5479 | through to the updated o32 FP64 ABI. To achieve this the dynamic linker now | |
5480 | tracks the ABI of any loaded object and verifies that new objects are | |
5481 | compatible. Mode transitions will also be requested as required and | |
5482 | unsupportable objects will be rejected. The ABI checks include both soft and | |
5483 | hard float ABIs for o32, n32 and n64. | |
5484 | ||
5485 | GCC 5 with GNU binutils 2.25 onwards: | |
5486 | It is strongly recommended that all o32 system libraries are built using the | |
5487 | new o32 FPXX ABI (-mfpxx) to facilitate the transition as this is compatible | |
5488 | with the original and all new o32 ABI extensions. Configure a MIPS GCC | |
5489 | compiler using --with-fp-32=xx to set this by default. | |
5490 | \f | |
5491 | Version 2.20 | |
5492 | ||
5493 | * The following bugs are resolved with this release: | |
5494 | ||
5495 | 6804, 9894, 11505, 12994, 13347, 13651, 14308, 14770, 15119, 15132, 15347, | |
5496 | 15514, 15698, 15804, 15894, 15946, 16002, 16064, 16095, 16194, 16198, | |
5497 | 16275, 16284, 16287, 16315, 16348, 16349, 16354, 16357, 16362, 16447, | |
5498 | 16516, 16532, 16539, 16545, 16561, 16562, 16564, 16574, 16599, 16600, | |
5499 | 16609, 16610, 16611, 16613, 16619, 16623, 16629, 16632, 16634, 16639, | |
5500 | 16642, 16648, 16649, 16670, 16674, 16677, 16680, 16681, 16683, 16689, | |
5501 | 16695, 16701, 16706, 16707, 16712, 16713, 16714, 16724, 16731, 16739, | |
5502 | 16740, 16743, 16754, 16758, 16759, 16760, 16770, 16786, 16789, 16791, | |
5503 | 16796, 16799, 16800, 16815, 16823, 16824, 16831, 16838, 16839, 16849, | |
5504 | 16854, 16876, 16877, 16878, 16882, 16885, 16888, 16890, 16892, 16912, | |
5505 | 16915, 16916, 16917, 16918, 16922, 16927, 16928, 16932, 16943, 16958, | |
5506 | 16965, 16966, 16967, 16977, 16978, 16984, 16990, 16996, 17009, 17022, | |
5507 | 17031, 17042, 17048, 17050, 17058, 17061, 17062, 17069, 17075, 17078, | |
5508 | 17079, 17084, 17086, 17088, 17092, 17097, 17125, 17135, 17137, 17150, | |
5509 | 17153, 17187, 17213, 17259, 17261, 17262, 17263, 17319, 17325, 17354. | |
5510 | ||
5511 | * Reverted change of ABI data structures for s390 and s390x: | |
5512 | On s390 and s390x the size of struct ucontext and jmp_buf was increased in | |
5513 | 2.19. This change is reverted in 2.20. The introduced 2.19 symbol versions | |
5514 | of getcontext, setjmp, _setjmp, __sigsetjmp, longjmp, _longjmp, siglongjmp | |
5515 | are preserved pointing straight to the same implementation as the old ones. | |
5516 | Given that, new callers will simply provide a too-big buffer to these | |
5517 | functions. Any applications/libraries out there that embed jmp_buf or | |
5518 | ucontext_t in an ABI-relevant data structure that have already been rebuilt | |
5519 | against 2.19 headers will have to rebuilt again. This is necessary in any | |
5520 | case to revert the breakage in their ABI caused by the glibc change. | |
5521 | ||
5522 | * Support for file description locks is added to systems running the | |
5523 | Linux kernel. The standard file locking interfaces are extended to | |
5524 | operate on file descriptions, not file descriptors, via the use of | |
5525 | F_OFD_GETLK, F_OFD_SETLK, and F_OFD_SETLKW. File description locks | |
5526 | are associated with an open file instead of a process. | |
5527 | ||
5528 | * Optimized strchr implementation for AArch64. Contributed by ARM Ltd. | |
5529 | ||
5530 | * The minimum Linux kernel version that this version of the GNU C Library | |
5531 | can be used with is 2.6.32. | |
5532 | ||
5533 | * Running the testsuite no longer terminates as soon as a test fails. | |
5534 | Instead, a file tests.sum (xtests.sum from "make xcheck") is generated, | |
5535 | with PASS or FAIL lines for individual tests. A summary of the results is | |
5536 | printed, including a list of failing lists, and "make check" exits with | |
5537 | error status if there were any unexpected failures. "make check | |
5538 | stop-on-test-failure=y" may be used to keep the old behavior. | |
5539 | ||
5540 | * The am33 port, which had not worked for several years, has been removed | |
5541 | from ports. | |
5542 | ||
5543 | * The _BSD_SOURCE and _SVID_SOURCE feature test macros are no longer | |
5544 | supported; they now act the same as _DEFAULT_SOURCE (but generate a | |
5545 | warning). Except for cases where _BSD_SOURCE enabled BSD interfaces that | |
5546 | conflicted with POSIX (support for which was removed in 2.19), the | |
5547 | interfaces those macros enabled remain available when compiling with | |
5548 | _GNU_SOURCE defined, with _DEFAULT_SOURCE defined, or without any feature | |
5549 | test macros defined. | |
5550 | ||
5551 | * Optimized strcmp implementation for ARMv7. Contributed by ARM Ltd. | |
5552 | ||
5553 | * Added support for TX lock elision of pthread mutexes on s390 and s390x. | |
5554 | This may improve lock scaling of existing programs on TX capable systems. | |
5555 | The lock elision code is only built with --enable-lock-elision=yes and | |
5556 | then requires a GCC version supporting the TX builtins. With lock elision | |
5557 | default mutexes are elided via __builtin_tbegin, if the cpu supports | |
5558 | transactions. By default lock elision is not enabled and the elision code | |
5559 | is not built. | |
5560 | ||
5561 | * CVE-2014-4043 The posix_spawn_file_actions_addopen implementation did not | |
5562 | copy the path argument. This allowed programs to cause posix_spawn to | |
5563 | deference a dangling pointer, or use an unexpected pathname argument if | |
5564 | the string was modified after the posix_spawn_file_actions_addopen | |
5565 | invocation. | |
5566 | ||
5567 | * All supported architectures now use the main glibc sysdeps directory | |
5568 | instead of some being in a separate "ports" directory (which was | |
5569 | distributed separately before glibc 2.17). | |
5570 | ||
5571 | * The NPTL implementation of POSIX pthreads is no longer an "add-on". | |
5572 | On configurations that support it (all Linux configurations), it's now | |
5573 | used regardless of the --enable-add-ons switch to configure. It is no | |
5574 | longer possible to build such configurations without pthreads support. | |
5575 | ||
5576 | * Locale names, including those obtained from environment variables (LANG | |
5577 | and the LC_* variables), are more tightly checked for proper syntax. | |
5578 | setlocale will now fail (with EINVAL) for locale names that are overly | |
5579 | long, contain slashes without starting with a slash, or contain ".." path | |
5580 | components. (CVE-2014-0475) Previously, some valid locale names were | |
5581 | silently replaced with the "C" locale when running in AT_SECURE mode | |
5582 | (e.g., in a SUID program). This is no longer necessary because of the | |
5583 | additional checks. | |
5584 | ||
5585 | * On x86-64, the dynamic linker's lazy-binding support is now compatible | |
5586 | with application code using Intel MPX instructions. (With all previous | |
5587 | versions, the MPX register state could be clobbered when making calls | |
5588 | into or out of a shared library.) Note that while the new dynamic | |
5589 | linker is compatible with all known x86 hardware whether or not it | |
5590 | supports Intel MPX, some x86 instruction-set emulators might fail to | |
5591 | handle the new instruction encodings. This is known to affect Valgrind | |
5592 | versions up through 3.9 (but will be fixed in the forthcoming 3.10 | |
5593 | release), and might affect other tools that do instruction emulation. | |
5594 | ||
5595 | * Support for loadable gconv transliteration modules has been removed. | |
5596 | The support for transliteration modules has been non-functional for | |
5597 | over a decade, and the removal is prompted by security defects. The | |
5598 | normal gconv conversion modules are still supported. Transliteration | |
5599 | with //TRANSLIT is still possible, and the //IGNORE specifier | |
5600 | continues to be supported. (CVE-2014-5119) | |
5601 | ||
5602 | * Decoding a crafted input sequence in the character sets IBM933, IBM935, | |
5603 | IBM937, IBM939, IBM1364 could result in an out-of-bounds array read, | |
5604 | resulting a denial-of-service security vulnerability in applications which | |
5605 | use functions related to iconv. (CVE-2014-6040) | |
5606 | \f | |
5607 | Version 2.19 | |
5608 | ||
5609 | * The following bugs are resolved with this release: | |
5610 | ||
5611 | 156, 387, 431, 762, 832, 926, 2801, 4772, 6786, 6787, 6807, 6810, 6981, | |
5612 | 7003, 9721, 9954, 10253, 10278, 11087, 11157, 11214, 12100, 12486, 12751, | |
5613 | 12986, 13028, 13982, 13985, 14029, 14032, 14120, 14143, 14155, 14286, | |
5614 | 14547, 14699, 14752, 14782, 14876, 14910, 15004, 15048, 15073, 15089, | |
5615 | 15128, 15218, 15268, 15277, 15308, 15362, 15374, 15400, 15425, 15427, | |
5616 | 15483, 15522, 15531, 15532, 15593, 15601, 15608, 15609, 15610, 15632, | |
5617 | 15640, 15670, 15672, 15680, 15681, 15723, 15734, 15735, 15736, 15748, | |
5618 | 15749, 15754, 15760, 15763, 15764, 15797, 15799, 15825, 15843, 15844, | |
5619 | 15846, 15847, 15849, 15850, 15855, 15856, 15857, 15859, 15867, 15886, | |
5620 | 15887, 15890, 15892, 15893, 15895, 15897, 15901, 15905, 15909, 15915, | |
5621 | 15917, 15919, 15921, 15923, 15939, 15941, 15948, 15963, 15966, 15968, | |
5622 | 15985, 15988, 15997, 16032, 16034, 16036, 16037, 16038, 16041, 16046, | |
5623 | 16055, 16071, 16072, 16074, 16077, 16078, 16103, 16112, 16143, 16144, | |
5624 | 16146, 16150, 16151, 16153, 16167, 16169, 16172, 16195, 16214, 16245, | |
5625 | 16271, 16274, 16283, 16289, 16293, 16314, 16316, 16330, 16337, 16338, | |
5626 | 16356, 16365, 16366, 16369, 16372, 16375, 16379, 16381, 16384, 16385, | |
5627 | 16386, 16387, 16390, 16394, 16398, 16400, 16407, 16408, 16414, 16430, | |
5628 | 16431, 16453, 16474, 16506, 16510, 16529 | |
5629 | ||
5630 | * Slovenian translations for glibc messages have been contributed by the | |
5631 | Translation Project's Slovenian team of translators. | |
5632 | ||
5633 | * The public headers no longer use __unused nor __block. This change is to | |
5634 | support compiling programs that are derived from BSD sources and use | |
5635 | __unused internally, and to support compiling with Clang's -fblock | |
5636 | extension which uses __block. | |
5637 | ||
5638 | * CVE-2012-4412 The strcoll implementation caches indices and rules for | |
5639 | large collation sequences to optimize multiple passes. This cache | |
5640 | computation may overflow for large collation sequences and may cause a | |
5641 | stack or buffer overflow. This is now fixed to use a slower algorithm | |
5642 | which does not use a cache if there is an integer overflow. | |
5643 | ||
5644 | * CVE-2012-4424 The strcoll implementation uses malloc to cache indices and | |
5645 | rules for large collation sequences to optimize multiple passes and falls | |
5646 | back to alloca if malloc fails, resulting in a possible stack overflow. | |
5647 | The implementation now falls back to an uncached collation sequence lookup | |
5648 | if malloc fails. | |
5649 | ||
5650 | * CVE-2013-4788 The pointer guard used for pointer mangling was not | |
5651 | initialized for static applications resulting in the security feature | |
5652 | being disabled. The pointer guard is now correctly initialized to a | |
5653 | random value for static applications. Existing static applications need | |
5654 | to be recompiled to take advantage of the fix (bug 15754). | |
5655 | ||
5656 | * CVE-2013-4237 The readdir_r function could write more than NAME_MAX bytes | |
5657 | to the d_name member of struct dirent, or omit the terminating NUL | |
5658 | character. (Bugzilla #14699). | |
5659 | ||
5660 | * CVE-2013-4332 The pvalloc, valloc, memalign, posix_memalign and | |
5661 | aligned_alloc functions could allocate too few bytes or corrupt the | |
5662 | heap when passed very large allocation size values (Bugzilla #15855, | |
5663 | #15856, #15857). | |
5664 | ||
5665 | * CVE-2013-4458 Stack overflow in getaddrinfo with large number of results | |
5666 | for AF_INET6 has been fixed (Bugzilla #16072). | |
5667 | ||
5668 | * New locales: ak_GH, anp_IN, ar_SS, cmn_TW, hak_TW, lzh_TW, nan_TW, pap_AW, | |
5669 | pap_CW, quz_PE, the_NP. | |
5670 | ||
5671 | * Substantially revised locales: gd_GB, ht_HT | |
5672 | ||
5673 | * The LC_ADDRESS field was updated to support country_car for almost all | |
5674 | supported locales. | |
5675 | ||
5676 | * ISO 1427 definitions were updated. | |
5677 | ||
5678 | * ISO 3166 definitions were updated. | |
5679 | ||
5680 | * The localedef utility now supports --big-endian and --little-endian | |
5681 | command-line options to generate locales for a different system from that | |
5682 | for which the C library was built. | |
5683 | ||
5684 | * Binary locale files now only depend on the endianness of the system for | |
5685 | which they are generated and not on other properties of that system. As a | |
5686 | consequence, binary files generated with new localedef may be incompatible | |
5687 | with old versions of the GNU C Library, and binary files generated with | |
5688 | old localedef may be incompatible with this version of the GNU C Library, | |
5689 | in the following circumstances: | |
5690 | ||
5691 | + Locale files may be incompatible on m68k systems. | |
5692 | ||
5693 | + Locale archive files (but not separate files for individual locales) may | |
5694 | be incompatible on systems where plain "char" is signed. | |
5695 | ||
5696 | * The configure option --disable-versioning has been removed. Builds with | |
5697 | --disable-versioning had not worked for several years. | |
5698 | ||
5699 | * ISO 639 definitions were updated for Chiga (cgg) and Chinese (gan, hak, czh, | |
5700 | cjy, lzh, cmn, mnp, cdo, czo, cpx, wuu, hsn, yue). | |
5701 | ||
5702 | * SystemTap probes for malloc have been introduced. | |
5703 | ||
5704 | * SystemTap probes for slow multiple precision fallback paths of | |
5705 | transcendental functions have been introduced. | |
5706 | ||
5707 | * Support for powerpc64le has been added. | |
5708 | ||
5709 | * The soft-float powerpc port now supports e500 processors. | |
5710 | ||
5711 | * Support for STT_GNU_IFUNC symbols added for ppc32/power4+ and ppc64. | |
5712 | ||
5713 | * A new feature test macro _DEFAULT_SOURCE is available to enable the same | |
5714 | set of header declarations that are enabled by default, even when other | |
5715 | feature test macros or compiler options such as -std=c99 would otherwise | |
5716 | disable some of those declarations. | |
5717 | ||
5718 | * The _BSD_SOURCE feature test macro no longer enables BSD interfaces that | |
5719 | conflict with POSIX. The libbsd-compat library (which was a dummy library | |
5720 | that did nothing) has also been removed. | |
5721 | ||
5722 | * Preliminary documentation about Multi-Thread, Async-Signal and | |
5723 | Async-Cancel Safety has been added. | |
5724 | ||
5725 | * Change of ABI data structures for s390 and s390x: | |
5726 | On s390 and s390x the size of struct ucontext and jmp_buf was increased to | |
5727 | allow for future hardware extensions. All library functions that accept or | |
5728 | return these structures were versioned in order to provide backwards | |
5729 | compatibility with existing code. However, not all cases can be handled | |
5730 | with interface versioning. If an application embeds either structure into | |
5731 | another structure and that structure is passed to another compilation unit | |
5732 | which expects the newer larger-sized object then any access to the new | |
5733 | fields will result in undefined behaviour. Similarly any access to fields | |
5734 | that were moved by the enlarging of the structures will cause undefined | |
5735 | behaviour. To fix the undefined behaviour all objects that are part of the | |
5736 | final in-memory image must agree on the size of structures, and this may | |
5737 | require recompilation. | |
5738 | \f | |
5739 | Version 2.18 | |
5740 | ||
5741 | * The following bugs are resolved with this release: | |
5742 | ||
5743 | 2546, 2560, 5159, 6809, 7006, 10060, 10062, 10283, 10357, 10686, 11120, | |
5744 | 11561, 12310, 12387, 12492, 12515, 12723, 13550, 13889, 13951, 13988, | |
5745 | 14142, 14176, 14200, 14256, 14280, 14293, 14317, 14327, 14478, 14496, | |
5746 | 14582, 14686, 14812, 14888, 14894, 14907, 14908, 14909, 14920, 14941, | |
5747 | 14952, 14964, 14981, 14982, 14985, 14991, 14994, 14996, 15000, 15003, | |
5748 | 15006, 15007, 15014, 15020, 15022, 15023, 15036, 15054, 15055, 15062, | |
5749 | 15078, 15084, 15085, 15086, 15100, 15160, 15214, 15221, 15232, 15234, | |
5750 | 15283, 15285, 15287, 15304, 15305, 15307, 15309, 15327, 15330, 15335, | |
5751 | 15336, 15337, 15339, 15342, 15346, 15359, 15361, 15366, 15380, 15381, | |
5752 | 15394, 15395, 15405, 15406, 15409, 15416, 15418, 15419, 15423, 15424, | |
5753 | 15426, 15429, 15431, 15432, 15441, 15442, 15448, 15465, 15480, 15485, | |
5754 | 15488, 15490, 15492, 15493, 15497, 15506, 15529, 15536, 15553, 15577, | |
5755 | 15583, 15618, 15627, 15631, 15654, 15655, 15666, 15667, 15674, 15711, | |
5756 | 15755, 15759. | |
5757 | ||
5758 | * CVE-2013-2207 Incorrectly granting access to another user's pseudo-terminal | |
5759 | has been fixed by disabling the use of pt_chown (Bugzilla #15755). | |
5760 | Distributions can re-enable building and using pt_chown via the new configure | |
5761 | option `--enable-pt_chown'. Enabling the use of pt_chown carries with it | |
5762 | considerable security risks and should only be used if the distribution | |
5763 | understands and accepts the risks. | |
5764 | ||
5765 | * CVE-2013-0242 Buffer overrun in regexp matcher has been fixed (Bugzilla | |
5766 | #15078). | |
5767 | ||
5768 | * CVE-2013-1914 Stack overflow in getaddrinfo with many results has been | |
5769 | fixed (Bugzilla #15330). | |
5770 | ||
5771 | * Add support for calling C++11 thread_local object destructors on thread | |
5772 | and program exit. This needs compiler support for offloading C++11 | |
5773 | destructor calls to glibc. | |
5774 | ||
5775 | * Improved worst case performance of libm functions with double inputs and | |
5776 | output. | |
5777 | ||
5778 | * Support for priority inherited mutexes in pthread condition variables on | |
5779 | non-x86 architectures. | |
5780 | ||
5781 | * Port to Xilinx MicroBlaze contributed by David Holsgrove. | |
5782 | ||
5783 | * Optimized string functions for AArch64. Implemented by Marcus Shawcroft. | |
5784 | ||
5785 | * Optimized string functions for ARM. Implemented by Will Newton and | |
5786 | Richard Henderson. | |
5787 | ||
5788 | * Optimized libm functions for SPARC. Implemented by David S. Miller. | |
5789 | ||
5790 | * Support for building more of ARM glibc as Thumb-2 code. Implemented by | |
5791 | Richard Henderson. | |
5792 | ||
5793 | * Support for building most of MIPS glibc as MIPS16 code. Implemented by | |
5794 | Chung-Lin Tang, Maciej W. Rozycki and Maxim Kuvyrkov. | |
5795 | ||
5796 | * Added a benchmark framework to track performance of functions in glibc. | |
5797 | ||
5798 | * New <math.h> macro named issignaling to check for a signaling NaN (sNaN). | |
5799 | It is based on draft TS 18661 and currently enabled as a GNU extension. | |
5800 | ||
5801 | * On Linux, the clock function now uses the clock_gettime system call | |
5802 | for improved precision, rather than old times system call. | |
5803 | ||
5804 | * Added support for version-3 tz format files. This is needed when using | |
5805 | the tz database release 2013e or later, and affects a few unusual cases -- | |
5806 | currently only TZ='America/Godthab' for time stamps after 2037. | |
5807 | ||
5808 | * Added new API functions pthread_getattr_default_np and | |
5809 | pthread_setattr_default_np to get and set the default pthread | |
5810 | attributes of a process. | |
5811 | ||
5812 | * Added support for TSX lock elision for pthread mutexes on i386 and x86-64. | |
5813 | This may improve lock scaling of existing programs on TSX capable systems. | |
5814 | When the --enable-lock-elision=yes parameter is specified at configure | |
5815 | time lock elision will be enabled by default for all PTHREAD_MUTEX_DEFAULT | |
5816 | mutexes. | |
5817 | ||
5818 | * Support added for AT_HWCAP2 (to coincide with Linux kernel 3.10 | |
5819 | availability). Implemented by Ryan S. Arnold. | |
5820 | ||
5821 | * Support added for POWER8 platform. Implemented by Ryan S. Arnold. | |
5822 | \f | |
5823 | Version 2.17 | |
5824 | ||
5825 | * The following bugs are resolved with this release: | |
5826 | ||
5827 | 1349, 3439, 3479, 3665, 5044, 5246, 5298, 5400, 6530, 6677, 6778, 6808, | |
5828 | 9685, 9914, 10014, 10038, 10114, 10191, 10631, 10873, 11438, 11607, 11638, | |
5829 | 11741, 12140, 13013, 13412, 13542, 13601, 13603, 13604, 13629, 13679, | |
5830 | 13696, 13698, 13717, 13741, 13759, 13761, 13763, 13881, 13939, 13950, | |
5831 | 13952, 13966, 14042, 14047, 14090, 14150, 14151, 14152, 14154, 14157, | |
5832 | 14166, 14173, 14195, 14197, 14237, 14246, 14251, 14252, 14283, 14298, | |
5833 | 14303, 14307, 14328, 14331, 14336, 14337, 14347, 14349, 14368, 14376, | |
5834 | 14417, 14447, 14459, 14476, 14477, 14501, 14505, 14510, 14516, 14518, | |
5835 | 14519, 14530, 14532, 14538, 14543, 14544, 14545, 14557, 14562, 14568, | |
5836 | 14576, 14579, 14583, 14587, 14595, 14602, 14610, 14621, 14638, 14645, | |
5837 | 14648, 14652, 14660, 14661, 14669, 14672, 14683, 14694, 14716, 14719, | |
5838 | 14743, 14767, 14783, 14784, 14785, 14793, 14796, 14797, 14801, 14803, | |
5839 | 14805, 14807, 14811, 14815, 14821, 14822, 14824, 14828, 14831, 14833, | |
5840 | 14835, 14838, 14856, 14863, 14865, 14866, 14868, 14869, 14871, 14872, | |
5841 | 14879, 14889, 14893, 14898, 14914. | |
5842 | ||
5843 | * Optimization of memcpy for MIPS. | |
5844 | ||
5845 | * CVE-2011-4609 svc_run() produces high cpu usage when accept fails with | |
5846 | EMFILE has been fixed (Bugzilla #14889). | |
5847 | ||
5848 | * The add-on ports collection is now distributed in the "ports" subdirectory | |
5849 | of the main GNU C Library distribution, rather than separately. | |
5850 | ||
5851 | * Port to ARM AArch64 contributed by Linaro. | |
5852 | ||
5853 | * Support for STT_GNU_IFUNC symbols added for s390 and s390x. | |
5854 | Optimized versions of memcpy, memset, and memcmp added for System z10 and | |
5855 | zEnterprise z196. | |
5856 | Implemented by Andreas Krebbel. | |
5857 | ||
5858 | * The new function secure_getenv allows secure access to the environment, | |
5859 | returning NULL if running in a SUID/SGID process. This function replaces | |
5860 | the internal function __secure_getenv. | |
5861 | ||
5862 | * SystemTap static probes have been added into the dynamic linker. | |
5863 | Implemented by Gary Benson. | |
5864 | ||
5865 | * Optimizations of string functions strstr, strcasestr and memmem. | |
5866 | Implemented by Maxim Kuvyrkov. | |
5867 | ||
5868 | * The minimum Linux kernel version that this version of the GNU C Library | |
5869 | can be used with is 2.6.16. | |
5870 | ||
5871 | * Optimizations of string functions memchr, wcschr, wcscpy, and wcsrchr for | |
5872 | powerpc POWER7. Implemented by Will Schmidt. | |
5873 | ||
5874 | * New configure option --disable-nscd builds the C library such that it | |
5875 | never attempts to contact the Name Service Caching Daemon (nscd). | |
5876 | New configure option --disable-build-nscd avoids building nscd itself; | |
5877 | this is the default if --disable-nscd is used. | |
5878 | ||
5879 | * Improved support for cross-compilation, including cross-testing and | |
5880 | bootstrap builds without a previously built glibc. | |
5881 | ||
5882 | * Several testsuite tests are now able to test multiple IFUNC variants of an | |
5883 | interface, rather than just testing the one that would be chooen by | |
5884 | default. | |
5885 | ||
5886 | * New configure options --with-bugurl and --with-pkgversion, for | |
5887 | distributors to use to embed their bug-reporting and package version | |
5888 | information in --help and --version output. | |
5889 | ||
5890 | * The ttyname and ttyname_r functions on Linux now fall back to searching for | |
5891 | the tty file descriptor in /dev/pts or /dev if /proc is not available. This | |
5892 | allows creation of chroots without the procfs mounted on /proc. | |
5893 | ||
5894 | * The `crypt' function now fails if passed salt bytes that violate the | |
5895 | specification for those values. On Linux, the `crypt' function will | |
5896 | consult /proc/sys/crypto/fips_enabled to determine if "FIPS mode" is | |
5897 | enabled, and fail on encrypted strings using the MD5 or DES algorithm | |
5898 | when the mode is enabled. | |
5899 | ||
5900 | * The `clock_*' suite of functions (declared in <time.h>) is now available | |
5901 | directly in the main C library. Previously it was necessary to link with | |
5902 | -lrt to use these functions. This change has the effect that a | |
5903 | single-threaded program that uses a function such as `clock_gettime' (and | |
5904 | is not linked with -lrt) will no longer implicitly load the pthreads | |
5905 | library at runtime and so will not suffer the overheads associated with | |
5906 | multi-thread support in other code such as the C++ runtime library. | |
5907 | ||
5908 | * New locales: ayc_PE, doi_IN, ia_FR, mni_IN, nhn_MX, niu_NU, niu_NZ, | |
5909 | sat_IN, and szl_PL. | |
5910 | \f | |
5911 | Version 2.16 | |
5912 | ||
5913 | * The following bugs are resolved with this release: | |
5914 | ||
5915 | 174, 208, 350, 369, 411, 706, 766, 2074, 2541, 2547, 2548, 2550, 2551, | |
5916 | 2552, 2553, 2554, 2562, 2563, 2565, 2566, 2570, 2576, 2636, 2678, 3335, | |
5917 | 3440, 3748, 3768, 3866, 3868, 3906, 3976, 3992, 4026, 4108, 4596, 4822, | |
5918 | 5077, 5461, 5805, 5993, 6471, 6486, 6578, 6649, 6730, 6770, 6794, 6884, | |
5919 | 6890, 6894, 6895, 6907, 6911, 6959, 7064, 9739, 9902, 10110, 10135, 10140, | |
5920 | 10153, 10210, 10254, 10346, 10375, 10545, 10716, 10846, 10882, 11174, | |
5921 | 11261, 11322, 11365, 11451, 11494, 11521, 11677, 11837, 11959, 12047, | |
5922 | 12097, 12193, 12194, 12297, 12298, 12301, 12340, 12354, 12416, 12495, | |
5923 | 13058, 13223, 13361, 13525, 13526, 13527, 13528, 13529, 13530, 13531, | |
5924 | 13532, 13533, 13547, 13551, 13552, 13553, 13555, 13556, 13559, 13563, | |
5925 | 13566, 13576, 13579, 13583, 13592, 13594, 13613, 13618, 13637, 13656, | |
5926 | 13658, 13673, 13691, 13695, 13704, 13705, 13706, 13718, 13726, 13738, | |
5927 | 13739, 13743, 13750, 13758, 13760, 13761, 13775, 13786, 13787, 13792, | |
5928 | 13806, 13824, 13840, 13841, 13844, 13846, 13848, 13851, 13852, 13854, | |
5929 | 13871, 13872, 13873, 13879, 13882, 13883, 13884, 13885, 13886, 13892, | |
5930 | 13895, 13908, 13910, 13911, 13912, 13913, 13914, 13915, 13916, 13917, | |
5931 | 13918, 13919, 13920, 13921, 13922, 13923, 13924, 13926, 13927, 13928, | |
5932 | 13938, 13941, 13942, 13954, 13955, 13956, 13963, 13967, 13968, 13970, | |
5933 | 13973, 13979, 13983, 13986, 13996, 14012, 14027, 14033, 14034, 14036, | |
5934 | 14040, 14043, 14044, 14048, 14049, 14050, 14053, 14055, 14059, 14064, | |
5935 | 14075, 14080, 14083, 14103, 14104, 14109, 14112, 14117, 14122, 14123, | |
5936 | 14134, 14153, 14183, 14188, 14199, 14210, 14218, 14229, 14241, 14273, | |
5937 | 14277, 14278. | |
5938 | ||
5939 | * Support for the x32 ABI on x86-64 added. The x32 target is selected by | |
5940 | configuring glibc with: | |
5941 | BUILD_CC='gcc' CC='gcc -mx32' CXX='g++ -mx32' | |
5942 | Visit <https://sites.google.com/site/x32abi/> for more x32 ABI info. | |
5943 | Implemented by H.J. Lu. | |
5944 | ||
5945 | * ISO C11 support: | |
5946 | ||
5947 | + define static_assert | |
5948 | ||
5949 | + do not declare gets | |
5950 | ||
5951 | + declare at_quick_exit and quick_exit also for ISO C11 | |
5952 | ||
5953 | + aligned_alloc. NB: The code is deliberately allows the size parameter | |
5954 | to not be a multiple of the alignment. This is a moronic requirement | |
5955 | in the standard but it is only a requirement on the caller, not the | |
5956 | implementation. | |
5957 | ||
5958 | + timespec_get added | |
5959 | ||
5960 | + uchar.h support added | |
5961 | ||
5962 | + CMPLX, CMPLXF, CMPLXL added | |
5963 | ||
5964 | Implemented by Ulrich Drepper. | |
5965 | ||
5966 | * Support for the IA-64 has been moved to ports. | |
5967 | ||
5968 | * Remove support for anything but ELF binary format | |
5969 | ||
5970 | * Checking versions of poll, ppoll added. | |
5971 | Implemented by Ulrich Drepper. | |
5972 | ||
5973 | * More generic and 64-bit performance optimizations to math functions. | |
5974 | Implemented by Ulrich Drepper. | |
5975 | ||
5976 | * New configure option --enable-obsolete-rpc makes the deprecated RPC | |
5977 | headers and functions available at compile time as they were before | |
5978 | version 2.14. This option will be removed at some time in the future | |
5979 | after the TI-RPC library becomes fully sufficient for the needs of | |
5980 | existing applications. | |
5981 | ||
5982 | * Compatibility code for Linux kernel versions before 2.4 has been removed. | |
5983 | Note that glibc is not expected to work with any Linux kernel version | |
5984 | before 2.6. | |
5985 | ||
5986 | * New header <sys/auxv.h> and function getauxval allowing easy access to | |
5987 | the AT_* key-value pairs passed from the Linux kernel. The header also | |
5988 | defines the HWCAP_* bits associated with the AT_HWCAP key. | |
5989 | ||
5990 | * New locales: mag_IN | |
5991 | ||
5992 | * New configure option --enable-systemtap builds SystemTap static probes | |
5993 | into libc for setjmp and longjmp and into libpthread for various operations. | |
5994 | So far the setjmp/longjmp probes and some of the libpthread probes are | |
5995 | provided only for i*86 and x86_64. | |
5996 | Implemented by Roland McGrath and Rayson Ho. | |
5997 | ||
5998 | * Optimized expf for x86-32 and x86-64. Implemented by Liubov Dmitrieva. | |
5999 | ||
6000 | * More optimized functions for PowerPC. Implemented by Adhemerval Zanella | |
6001 | and Will Schmidt. | |
6002 | ||
6003 | * More optimized functions for SPARC. Implemented by David S. Miller. | |
6004 | ||
6005 | * Improved support for cross-compilation, especially bootstrap builds | |
6006 | without a previously built glibc. | |
6007 | ||
6008 | * Ports for the TILE-Gx and TILEPro families of processors. Contributed by | |
6009 | Chris Metcalf from Tilera. | |
6010 | ||
6011 | * Support for the old ARM ABI has been removed from ports. Only the EABI is | |
6012 | now supported for ARM processors. | |
6013 | ||
6014 | * The hard-float variant of the ARM EABI now uses /lib/ld-linux-armhf.so.3 | |
6015 | as the name of the dynamic linker, to distinguish it from the | |
6016 | /lib/ld-linux.so.3 dynamic linker used for the base EABI. | |
6017 | ||
6018 | * Support for CRIS has been removed from ports. | |
6019 | ||
6020 | * A new class of installed header has been documented for low-level | |
6021 | platform-specific functionality. PowerPC added the first instance with a | |
6022 | function to provide time base register access. Contributed by Tulio | |
6023 | Magno Quites Machado Filho. | |
6024 | ||
6025 | * ix86 configurations now install header files that are consistent with | |
6026 | what x86-64 configurations install. These same header files can be used | |
6027 | for -m32, -m64, or -mx32 builds. | |
6028 | Contributed by H.J. Lu. | |
6029 | ||
6030 | * Math library bug fixes. A thorough audit of all open math library bugs was | |
6031 | conducted by Joseph Myers. Significant progress was made on many math | |
6032 | library bugs resulting in more accurate exceptions and function results. | |
6033 | Many thanks to all those that contributed including Andreas Jaeger for his | |
6034 | patch review and work on the x87 trigonometric instruction issues. | |
6035 | ||
6036 | * Timezone data is no longer installed. Timezone-related binaries and scripts | |
6037 | will continue to be installed. Users should obtain their timezone data from | |
6038 | their distribution provider or from the tzdata package at | |
6039 | <ftp://munnari.oz.au/pub/>. | |
6040 | \f | |
6041 | Version 2.15 | |
6042 | ||
6043 | * The following bugs are resolved with this release: | |
6044 | ||
6045 | 6779, 6783, 9696, 10103, 10709, 11589, 11929, 12403, 12786, 12840, 12847, | |
6046 | 12868, 12852, 12874, 12885, 12892, 12906, 12907, 12922, 12935, 12962, | |
6047 | 13007, 13021, 13061, 13062, 13067, 13068, 13085, 13088, 13090, 13092, | |
6048 | 13096, 13114, 13118, 13123, 13134, 13138, 13147, 13150, 13166, 13179, | |
6049 | 13185, 13189, 13192, 13268, 13276, 13282, 13291, 13305, 13328, 13335, | |
6050 | 13337, 13344, 13358, 13367, 13413, 13416, 13423, 13439, 13446, 13472, | |
6051 | 13484, 13506, 13515, 13523, 13524, 13538, 13540 | |
6052 | ||
6053 | * New program pldd to list loaded object of a process | |
6054 | Implemented by Ulrich Drepper. | |
6055 | ||
6056 | * Add nss_db support back to glibc. No more dependency on Berkeley db | |
6057 | and support for initgroups lookups. | |
6058 | Implemented by Ulrich Drepper. | |
6059 | ||
6060 | * Optimized strcpy, strncpy, stpcpy, stpncpy for SSE2 and SSSE3 on x86-32. | |
6061 | Contributed by HJ Lu. | |
6062 | ||
6063 | * Improved strcpy, strncpy, stpcpy, stpncpy for SSE2 and SSSE3 on x86-64. | |
6064 | Contributed by HJ Lu. | |
6065 | ||
6066 | * Optimized strcat, strncat on x86-64 and optimized wcscmp, wcslen, strnlen | |
6067 | on x86-32 and x86-64. | |
6068 | Contributed by Liubov Dmitrieva. | |
6069 | ||
6070 | * Optimized strchr and strrchr for SSE on x86-32. | |
6071 | Contributed by Liubov Dmitrieva. | |
6072 | ||
6073 | * Optimized memchr, memrchr, rawmemchr, memcmp, wmemcmp, wcschr, wcscpy | |
6074 | for x86-64 and x86-32. | |
6075 | Contributed by Liubov Dmitrieva. | |
6076 | ||
6077 | * New interfaces: scandirat, scandirat64 | |
6078 | Implemented by Ulrich Drepper. | |
6079 | ||
6080 | * Checking versions of FD_SET, FD_CLR, and FD_ISSET added. | |
6081 | Implemented by Ulrich Drepper. | |
6082 | ||
6083 | * nscd now also caches the netgroup database. | |
6084 | Implemented by Ulrich Drepper. | |
6085 | ||
6086 | * Integrate libm with gcc's -ffinite-math-only option. | |
6087 | Implemented by Ulrich Drepper. | |
6088 | ||
6089 | * Lots of generic, 64-bit, and x86-64-specific performance optimizations | |
6090 | to math functions. Implemented by Ulrich Drepper. | |
6091 | ||
6092 | * Optimized strcasecmp and strncasecmp for AVX on x86-64. | |
6093 | Implemented by Ulrich Drepper. | |
6094 | ||
6095 | * New Linux interfaces: process_vm_readv, process_vm_writev | |
6096 | ||
6097 | * Optimized strcasecmp and strncasecmp for SSSE3 and SSE4.2 on x86-32. | |
6098 | Implemented by Ulrich Drepper. | |
6099 | ||
6100 | * Optimized nearbyint and strcasecmp for PPC. | |
6101 | Implemented by Adhemerval Zanella. | |
6102 | ||
6103 | * New locales: bho_IN, unm_US, es_CU, ta_LK | |
6104 | \f | |
6105 | Version 2.14 | |
6106 | ||
6107 | * The following bugs are resolved with this release: | |
6108 | ||
6109 | 386, 6420, 7101, 9730, 9732, 9809, 10138, 10149, 10157, 11099, 11257, | |
6110 | 11258, 11487, 11532, 11558, 11578, 11634, 11653, 11668, 11697, 11724, | |
6111 | 11781, 11799, 11820, 11837, 11857, 11884, 11892, 11895, 11901, 11945, | |
6112 | 11947, 11952, 11987, 12052, 12083, 12158, 12178, 12200, 12346, 12350, | |
6113 | 12393, 12420, 12432, 12445, 12449, 12453, 12454, 12460, 12469, 12489, | |
6114 | 12509, 12510, 12511, 12518, 12527, 12541, 12545, 12551, 12582, 12583, | |
6115 | 12587, 12597, 12601, 12611, 12625, 12626, 12631, 12650, 12653, 12655, | |
6116 | 12660, 12671, 12681, 12684, 12685, 12711, 12713, 12714, 12717, 12723, | |
6117 | 12724, 12734, 12738, 12746, 12766, 12775, 12777, 12782, 12788, 12792, | |
6118 | 12795, 12811, 12813, 12814, 12841 | |
6119 | ||
6120 | * The RPC implementation in libc is obsoleted. Old programs keep working | |
6121 | but new programs cannot be linked with the routines in libc anymore. | |
6122 | Programs in need of RPC functionality must be linked against TI-RPC. | |
6123 | The TI-RPC implementation is IPv6 enabled and there are other benefits. | |
6124 | ||
6125 | Visible changes of this change include (obviously) the inability to link | |
6126 | programs using RPC functions without referencing the TI-RPC library and the | |
6127 | removal of the RPC headers from the glibc headers. | |
6128 | Implemented by Ulrich Drepper. | |
6129 | ||
6130 | * New Linux interfaces: clock_adjtime, name_to_handle_at, open_by_handle_at, | |
6131 | syncfs, setns, sendmmsg | |
6132 | ||
6133 | * New locales: os_RU, bem_ZA, en_ZA, ff_SN, sw_KE, sw_TZ, lb_LU, wae_CH, | |
6134 | yue_HK, lij_IT, mhr_RU | |
6135 | ||
6136 | * New encodings: CP770, CP771, CP772, CP773, CP774 | |
6137 | ||
6138 | * New program sotruss to trace calls through PLTs | |
6139 | Implemented by Ulrich Drepper. | |
6140 | ||
6141 | * The malloc hook implementation is marked deprecated and will be removed | |
6142 | from the default implementation in the next version. The design never | |
6143 | worked ever since the introduction of threads. Even programs which do | |
6144 | not create threads themselves can use multiple threads created internally. | |
6145 | \f | |
6146 | Version 2.13 | |
6147 | ||
6148 | * The following bugs are resolved with this release: | |
6149 | ||
6150 | 3268, 6812, 7066, 10085, 10484, 10563, 10851, 11149, 11155, 11577, 11611, | |
6151 | 11640, 11655, 11701, 11840, 11856, 11883, 11903, 11904, 11968, 11979, | |
6152 | 12005, 12037, 12067, 12077, 12078, 12092, 12093, 12107, 12108, 12113, | |
6153 | 12140, 12159, 12167, 12191, 12194, 12201, 12204, 12205, 12207, 12348, | |
6154 | 12378, 12394, 12397 | |
6155 | ||
6156 | * New Linux interfaces: prlimit, prlimit64, fanotify_init, fanotify_mark | |
6157 | ||
6158 | * POWER7 optimizations: memset, memcmp, strncmp | |
6159 | ||
6160 | * New optimized string functions for x86-64: strnlen (SSE2), | |
6161 | strcasecmp (SSE2, SSSE3, SSE4.2), strncasecmp (SSE2, SSSE3, SSE4.2) | |
6162 | Implemented by Ulrich Drepper. | |
6163 | \f | |
6164 | Version 2.12 | |
6165 | ||
6166 | * The following bugs are resolved with this release: | |
6167 | ||
6168 | 3662, 4457, 5553, 10162, 10401, 10414, 10554, 10824, 10864, 10915, | |
6169 | 10918, 10936, 10939, 10958, 10968, 10969, 10972, 10992, 11000, 11001, | |
6170 | 11007, 11010, 11027, 11039, 11040, 11041, 11043, 11046, 11056, 11070, | |
6171 | 11093, 11115, 11125, 11126, 11127, 11134, 11141, 11149, 11183, 11184, | |
6172 | 11185, 11186, 11187, 11188, 11189, 11190, 11191, 11192, 11193, 11194, | |
6173 | 11200, 11230, 11235, 11242, 11254, 11258, 11271, 11272, 11276, 11279, | |
6174 | 11287, 11292, 11319, 11332, 11333, 11387, 11389, 11390, 11394, 11397, | |
6175 | 11410, 11438, 11449, 11470, 11471, 11520, 11537, 11538, 11571 | |
6176 | ||
6177 | * New interfaces: pthread_getname_np, pthread_setname_np | |
6178 | ||
6179 | * New Linux interface: recvmmsg | |
6180 | ||
6181 | * STT_GNU_IFUNC implemented for Sparc by David Miller. | |
6182 | ||
6183 | * The dynamic linker now recognizes supported ABI versions from the | |
6184 | EI_ABIVERSION field in the ELF header. | |
6185 | Implemented by Ulrich Drepper. | |
6186 | ||
6187 | * New NIS mode selector ADJUNCT_AS_SHADOW. The passwd.adjunct.byname table | |
6188 | will not be used to fill in password fields in the passwd.byname replies. | |
6189 | Instead it is used to synthesize the shadow.byname table, should it be | |
6190 | missing. This is a useful mode in some installations involving Solaris. | |
6191 | Implemented by Ulrich Drepper. | |
6192 | ||
6193 | * New locales: kok_IN, sq_MK, cv_RU | |
6194 | \f | |
6195 | Version 2.11 | |
6196 | ||
6197 | * The following bugs are resolved with this release: | |
6198 | ||
6199 | 3270, 7094, 9924, 9986, 10011, 10085, 10107, 10159, 10162, 10166, 10183, | |
6200 | 10195, 10196, 10200, 10203, 10207, 10211, 10213, 10217, 10221, 10229, | |
6201 | 10262, 10286, 10312, 10315, 10319, 10349, 10360, 10391, 10402, 10416, | |
6202 | 10418, 10422, 10425, 10446, 10448, 10452, 10540, 10553, 10560, 10564, | |
6203 | 10609, 10643, 10692, 10713, 10717, 10730, 10731, 10742, 10780, 10784, | |
6204 | 10789, 10817, 10823, 10840, 10847 | |
6205 | ||
6206 | * New interfaces: execvpe, pthread_sigqueue, mkstemps, mkstemps64, mkostemps, | |
6207 | mkostemps64 | |
6208 | Implemented by Ulrich Drepper. | |
6209 | ||
6210 | * Checking version of longjmp added that fails if an uninitialized stack | |
6211 | frame would be created. Implemented by Ulrich Drepper. | |
6212 | ||
6213 | * STT_GNU_IFUNC is now supported in static executables. | |
6214 | Implemented by H.J. Lu. | |
6215 | ||
6216 | * STT_GNU_IFUNC implemented for PPC by Alan Modra. | |
6217 | ||
6218 | * New optimized string functions for x86-64: strstr, strcasestr, memcmp, | |
6219 | strcspn, strpbrk, strspn, strcpy, stpcpy, strncpy, strcmp (SSE2, SSE4.2), | |
6220 | strncmp (SSE2, SSE4.2), strchr (SSE4.2), strrchr (SSE4.2). | |
6221 | Contributed by H.J. Lu. | |
6222 | ||
6223 | strlen, rawmemchr, strcmp (SSSE3), strncmp (SSSE3). | |
6224 | Implemented by Ulrich Drepper. | |
6225 | ||
6226 | * New optimized string functions for x86: strlen, strcspn, strspn, strpbrk, | |
6227 | strstr, strcasestr. | |
6228 | Contributed by H.J. Lu. | |
6229 | ||
6230 | * Support for fma instruction in AVX on x86-64. | |
6231 | Implemented by H.J. Lu and Ulrich Drepper. | |
6232 | ||
6233 | * AVX support in x86-64 auditing support in ld.so. | |
6234 | Implemented by H.J. Lu. | |
6235 | ||
6236 | * STB_GNU_UNIQUE support added. | |
6237 | Implemented by Ulrich Drepper. | |
6238 | ||
6239 | * Implement second fallback mode for optimized DNS lookup for even more | |
6240 | broken environments. If two requests from the same port are not | |
6241 | handled correctly close the socket and open a new one before sending | |
6242 | the second request. The 'single-request-reopen' option in /etc/resolv.conf | |
6243 | can be used to select this mode right away, instead of rediscovering the | |
6244 | necessity is every process again. | |
6245 | Implemented by Ulrich Drepper. | |
6246 | ||
6247 | * New resolver flag RES_USE_DNSSEC to enable use of verified lookup. | |
6248 | Implemented by Adam Tkac. | |
6249 | ||
6250 | * Optimized iconv conversions for S390x. | |
6251 | Implemented by Andreas Krebbel. | |
6252 | ||
6253 | * Using condvars with PI mutexes is now more efficient due to kernel | |
6254 | support for requeueing to PI futexes. NPTL support added for x86-64. | |
6255 | Implemented by Ulrich Drepper. | |
6256 | ||
6257 | * New locales: ps_AF, my_MM | |
6258 | \f | |
6259 | Version 2.10 | |
6260 | ||
6261 | * The following bugs are resolved with this release: | |
6262 | ||
6263 | 697, 5381, 5807, 6411, 6545, 7040, 7058, 7067, 7080, 7083, 7095, 7098, | |
6264 | 9677, 9694, 9697, 9701, 9704, 9705, 9706, 9720, 9726, 9730, 9733, 9736, | |
6265 | 9741, 9750, 9753, 9759, 9781, 9791, 9793, 9823, 9833, 9844, 9880, 9881, | |
6266 | 9891, 9893, 9895, 9913, 9916, 9920, 9948, 9955, 9957, 9963, 9985, 10052, | |
6267 | 10069, 10086, 10087, 10090, 10092, 10093, 10100, 10118, 10128 | |
6268 | ||
6269 | * New interfaces: preadv, preadv64, pwritev, pwritev64, malloc_info | |
6270 | Implemented by Ulrich Drepper. | |
6271 | ||
6272 | * New Linux interfaces: accept4, fallocate, fallocate64. | |
6273 | Implemented by Ulrich Drepper. | |
6274 | ||
6275 | * Correct declarations of string function when used in C++ code. This | |
6276 | could lead to compile errors for invalid C++ code. | |
6277 | ||
6278 | * XPG7/POSIX 2008 compilation environment. Many GNU-specific functions are | |
6279 | now in POSIX. | |
6280 | ||
6281 | * New POSIX 2008 interface: psiginfo | |
6282 | Implemented by Ulrich Drepper. | |
6283 | ||
6284 | * New ISO C++1x interfaces: quick_exit, at_quick_exit | |
6285 | Implemented by Ulrich Drepper. | |
6286 | ||
6287 | * Support for selecting between multiple function definitions at runtime | |
6288 | using STT_GNU_IFUNC symbols. Implemented by Ulrich Drepper. | |
6289 | ||
6290 | * The libcrypt library can now use the hash function implementations in | |
6291 | NSS. Implemented by Ulrich Drepper. | |
6292 | ||
6293 | * The malloc implementation can be compiled to be less memory efficient | |
6294 | but higher performing in multi-threaded programs. | |
6295 | Implemented by Ulrich Drepper. | |
6296 | ||
6297 | * New locales: nan_TW@latin, ks_IN | |
6298 | ||
6299 | * Faster strlen, strchr, strchrnul, strrchr, memchr, and rawmemchr for x86-64. | |
6300 | Implemented by Ulrich Drepper. | |
6301 | ||
6302 | * Extended printf hook support. It is possible to use user-defined types | |
6303 | and extend existing format specifiers. | |
6304 | Implemented by Ulrich Drepper. | |
6305 | ||
6306 | * Handling for group shadow files has been added. | |
6307 | Implemented by Ulrich Drepper. | |
6308 | ||
6309 | * DNS IPv4-IPv6 parallel lookup now deals better with broken DNS | |
6310 | servers (the case, e.g., for some people using the built-in DNS | |
6311 | server in ADSL modems/routers). There is a once-per-process timeout | |
6312 | in case of a broken server. To avoid it, users can run nscd or put | |
6313 | 'options single-request' in /etc/resolv.conf. | |
6314 | Implemented by Ulrich Drepper. | |
6315 | \f | |
6316 | Version 2.9 | |
6317 | ||
6318 | * The following bugs are resolved with this release: | |
6319 | ||
6320 | 2447, 3406, 5209, 5210, 5381, 5794, 5814, 5911, 6428, 6438, 6442, 6447, | |
6321 | 6461, 6472, 6505, 6544, 6589, 6612, 6634, 6653, 6654, 6657, 6698, 6712, | |
6322 | 6713, 6719, 6723, 6724, 6740, 6763, 6771, 6790, 6791, 6817, 6824, 6839, | |
6323 | 6843, 6845, 6856, 6857, 6860, 6867, 6875, 6919, 6920, 6942, 6947, 6955, | |
6324 | 6966, 6968, 6974, 6980, 6995, 7008, 7009, 7029 | |
6325 | ||
6326 | * Unified lookup for getaddrinfo: IPv4 and IPv6 addresses are now looked | |
6327 | up at the same time. Implemented by Ulrich Drepper. | |
6328 | ||
6329 | * TLS descriptors for LD and GD on x86 and x86-64. | |
6330 | Implemented by Alexandre Oliva. | |
6331 | ||
6332 | * getaddrinfo now handles DCCP and UDPlite. | |
6333 | Implemented by Ulrich Drepper. | |
6334 | ||
6335 | * New fixed-size conversion macros: htobe16, htole16, be16toh, le16toh, | |
6336 | htobe32, htole32, be32toh, le32toh, htobe64, htole64, be64toh, le64toh. | |
6337 | Implemented by Ulrich Drepper. | |
6338 | ||
6339 | * New implementation of memmem, strstr, and strcasestr which is O(n). | |
6340 | Implemented by Eric Blake. | |
6341 | ||
6342 | * New Linux interfaces: inotify_init1, dup3, epoll_create1, pipe2 | |
6343 | ||
6344 | * Implement "e" option for popen to open file descriptor with the | |
6345 | close-on-exec flag set. Implemented by Ulrich Drepper. | |
6346 | ||
6347 | * Implement "b" mode for fmemopen. In this mode writes writes don't | |
6348 | implicitly add a NUL byte and seeks from the end of the buffer really | |
6349 | use the buffer end, not the string length as the basis. | |
6350 | Implemented by Ulrich Drepper. | |
6351 | ||
6352 | * Many functions, exported and internal, now atomically set the close-on-exec | |
6353 | flag when run on a sufficiently new kernel. Implemented by Ulrich Drepper. | |
6354 | ||
6355 | * Sorting rules for some Indian languages (Kannada, Gurumukhi, Telugu, | |
6356 | Sinhala) | |
6357 | Implemented by Pravin Satpute. | |
6358 | ||
6359 | * New locales: sd_IN, sd_IN@devanagari, ks_IN@devanagari, ht_HT, en_AG, nl_AW. | |
6360 | \f | |
6361 | Version 2.8 | |
6362 | ||
6363 | * The following bugs are resolved with this release: | |
6364 | ||
6365 | 2549, 3112, 4314, 4407, 4997, 5012, 5040, 5112, 5204, 5208, 5209, 5220, | |
6366 | 5222, 5224, 5225, 5237, 5238, 5240, 5245, 5277, 5346, 5375, 5378, 5382, | |
6367 | 5424, 5427, 5428, 5435, 5436, 5439, 5441, 5442, 5443, 5451, 5452, 5454, | |
6368 | 5455, 5463, 5464, 5465, 5475, 5477, 5478, 5520, 5541, 5545, 5553, 5600, | |
6369 | 5602, 5607, 5614, 5627, 5628, 5736, 5737, 5741, 5753, 5760, 5762, 5768, | |
6370 | 5774, 5777, 5778, 5779, 5786, 5790, 5818, 5854, 5857, 5903, 5939, 5979, | |
6371 | 5983, 5995, 5998, 6004, 6007, 6020, 6021, 6024, 6040, 6042 | |
6372 | ||
6373 | * New locales: bo_CN, bo_IN, shs_CA. | |
6374 | ||
6375 | * New encoding: HP-ROMAN9, HP-GREEK8, HP-THAI8, HP-TURKISH8. | |
6376 | ||
6377 | * Sorting rules for some Indian languages (Devanagari and Gujarati). | |
6378 | Implemented by Pravin Satpute. | |
6379 | ||
6380 | * IPV6 addresses in /etc/resolv.conf can now have a scope ID | |
6381 | ||
6382 | * nscd caches now all timeouts for DNS entries | |
6383 | Implemented by Ulrich Drepper. | |
6384 | ||
6385 | * nscd is more efficient and wakes up less often. | |
6386 | Implemented by Ulrich Drepper. | |
6387 | ||
6388 | * More checking functions: asprintf, dprintf, obstack_printf, vasprintf, | |
6389 | vdprintf, and obstack_vprintf. | |
6390 | Implemented by Jakub Jelinek. | |
6391 | ||
6392 | * Faster memset for x86-64. | |
6393 | Implemented by Harsha Jagasia and H.J. Lu. | |
6394 | ||
6395 | * Faster memcpy on x86. | |
6396 | Implemented by Ulrich Drepper. | |
6397 | ||
6398 | * ARG_MAX is not anymore constant on Linux. Use sysconf(_SC_ARG_MAX). | |
6399 | Implemented by Ulrich Drepper. | |
6400 | ||
6401 | * Faster sqrt and sqrtf implementation for some PPC variants. | |
6402 | Implemented by Steven Munroe. | |
6403 | \f | |
6404 | Version 2.7 | |
6405 | ||
6406 | * The following bugs are resolved with this release: | |
6407 | ||
6408 | 4125, 4126, 4178, 4359, 4407, 4512, 4514, 4525, 4554, 4556, 4557, 4566, | |
6409 | 4582, 4586, 4588, 4599, 4610, 4647, 4702, 4705, 4726, 4745, 4772, 4773, | |
6410 | 4775, 4776, 4792, 4813, 4814, 4816, 4833, 4858, 4860, 4896, 4905, 4925, | |
6411 | 4936, 4937, 4938, 4941, 4946, 4963, 4972, 5010, 5028, 5043, 5058, 5063, | |
6412 | 5071, 5103, 5104, 5112, 5113, 5184, 5186 | |
6413 | ||
6414 | * More checking functions: fread, fread_unlocked, open*, mq_open. | |
6415 | Implemented by Jakub Jelinek and Ulrich Drepper. | |
6416 | ||
6417 | * Extend fortification to C++. Implemented by Jakub Jelinek. | |
6418 | ||
6419 | * Implement 'm' modifier for scanf. Add stricter C99/SUS compliance | |
6420 | by not recognizing 'a' as a modifier when those specs are requested. | |
6421 | Implemented by Jakub Jelinek. | |
6422 | ||
6423 | * PPC optimizations to math and string functions. | |
6424 | Implemented by Steven Munroe. | |
6425 | ||
6426 | * New interfaces: mkostemp, mkostemp64. Like mkstemp* but allow additional | |
6427 | options to be passed. Implemented by Ulrich Drepper. | |
6428 | ||
6429 | * More CPU set manipulation functions. Implemented by Ulrich Drepper. | |
6430 | ||
6431 | * New Linux interfaces: signalfd, eventfd, eventfd_read, and eventfd_write. | |
6432 | Implemented by Ulrich Drepper. | |
6433 | ||
6434 | * Handle private futexes in the NPTL implementation. | |
6435 | Implemented by Jakub Jelinek and Ulrich Drepper. | |
6436 | ||
6437 | * Add support for O_CLOEXEC. Implement in Hurd. Use throughout libc. | |
6438 | Implemented by Roland McGrath and Ulrich Drepper. | |
6439 | ||
6440 | * Linux/x86-64 vDSO support. Implemented by Ulrich Drepper. | |
6441 | ||
6442 | * SHA-256 and SHA-512 based password encryption. | |
6443 | Implemented by Ulrich Drepper. | |
6444 | ||
6445 | * New locales: ber_DZ, ber_MA, en_NG, fil_PH, fur_IT, fy_DE, ha_NG, ig_NG, | |
6446 | ik_CA, iu_CA, li_BE, li_NL, nds_DE, nds_NL, pap_AN, sc_IT, tk_TM, ug_CN, | |
6447 | yo_NG. | |
6448 | ||
6449 | + New iconv modules: MAC-CENTRALEUROPE, ISO-8859-9E, KOI8-RU. | |
6450 | Implemented by Ulrich Drepper. | |
6451 | \f | |
6452 | Version 2.6 | |
6453 | ||
6454 | * The following bugs are resolved with this release: | |
6455 | ||
6456 | 3156, 3213, 3285, 3291, 3306, 3313, 3320, 3322, 3325, 3326, 3334, 3348, | |
6457 | 3352, 3363, 3369, 3426, 3427, 3429, 3451, 3458, 3465, 3480, 3483, 3493, | |
6458 | 3514, 3515, 3559, 3632, 3664, 3673, 3674, 3745, 3747, 3818, 3840, 3842, | |
6459 | 3851, 3855, 3884, 3885, 3902, 3905, 3919, 3944, 3954, 3955, 3957, 3991, | |
6460 | 3995, 3996, 4024, 4040, 4069, 4070, 4074, 4076, 4096, 4101, 4102, 4114, | |
6461 | 4130, 4131, 4181, 4276, 4306, 4342, 4344, 4349, 4364, 4368, 4381, 4392, | |
6462 | 4403, 4405, 4406, 4411, 4438, 4439, 4455, 4456, 4465, 4512, 4514, 4586, | |
6463 | 4702, 4858 | |
6464 | ||
6465 | * New Linux interfaces: epoll_pwait, sched_getcpu. | |
6466 | ||
6467 | * New generic interfaces: strerror_l. | |
6468 | ||
6469 | * nscd can now cache the services database. Implemented by Ulrich Drepper. | |
6470 | \f | |
6471 | Version 2.5 | |
6472 | ||
6473 | * The following bugs are resolved with this release: | |
6474 | ||
6475 | 39, 68, 192, 352, 388, 531, 935, 1006, 1201, 1203, 1386, 1782, 1783, 1784, | |
6476 | 1785, 1786, 1787, 1951, 1974, 1996, 2072, 2098, 2125, 2134, 2158, 2182, | |
6477 | 2349, 2376, 2386, 2389, 2415, 2418, 2420, 2423, 2450, 2451, 2466, 2472, | |
6478 | 2473, 2477, 2498, 2499, 2501, 2502, 2505, 2507, 2508, 2509, 2517, 2526, | |
6479 | 2569, 2571, 2592, 2611, 2625, 2632, 2644, 2662, 2680, 2683, 2684, 2693, | |
6480 | 2695, 2703, 2734, 2751, 2764, 2766, 2775, 2792, 2821, 2823, 2832, 2841, | |
6481 | 2843, 2883, 2892, 2908, 2914, 2926, 2961, 2978, 2980, 2987, 2997, 2998, | |
6482 | 3013, 3018, 3034, 3040, 3044, 3123, 3124, 3137, 3138, 3143, 3155, 3189, | |
6483 | 3225, 3251, 3252, 3253, 3273 | |
6484 | ||
6485 | * For Linux, the sorting of addresses returned by getaddrinfo now also | |
6486 | handles rules 3, 4, and 7 from RFC 3484. I.e., all rules are handled. | |
6487 | Implemented by Ulrich Drepper. | |
6488 | ||
6489 | * Allow system admin to configure getaddrinfo with the /etc/gai.conf file. | |
6490 | Implemented by Ulrich Drepper. | |
6491 | ||
6492 | * New Linux interfaces: splice, tee, sync_file_range, vmsplice. | |
6493 | ||
6494 | * New iconv module for MIK. Contributed by Alexander Shopov. | |
6495 | ||
6496 | * For sites with broken group and/or passwd database, the auto-propagate | |
6497 | option of nscd can prevent creating ID lookup entries from the results | |
6498 | of a name lookup and vice versa. This usually is no problem but some | |
6499 | site might have problems with the default behavior. | |
6500 | Implemented by Ulrich Drepper. | |
6501 | ||
6502 | * Iterating over entire database in NIS can be slow. With the | |
6503 | SETENT_BATCH_READ option in /etc/default/nss a system admin can decide | |
6504 | to trade time for memory. The entire database will be read at once. | |
6505 | Implemented by Ulrich Drepper. | |
6506 | ||
6507 | * The interfaces introduced in RFC 3542 have been implemented by | |
6508 | Ulrich Drepper. | |
6509 | ||
6510 | * Support for the new ELF hash table format was added by Ulrich Drepper. | |
6511 | ||
6512 | * Support for priority inheritance mutexes added by Jakub Jelinek and | |
6513 | Ulrich Drepper. | |
6514 | ||
6515 | * Support for priority protected mutexes added by Jakub Jelinek. | |
6516 | ||
6517 | * New locales: nr_ZA, pa_PK, ca_AD, ca_FR, ca_IT, el_CY, tr_CY, as_IN, | |
6518 | or_IN, csb_PL, fy_NL, sr_ME. | |
6519 | \f | |
6520 | Version 2.4 | |
6521 | ||
6522 | * More overflow detection functions. | |
6523 | ||
6524 | * New iconv converters for IBM1025, IBM1122, IBM1137, IBM1153, | |
6525 | IBM1154, IBM1155, IBM1156, IBM1157, and IBM1158 contributed by Jiro Sekiba. | |
6526 | ||
6527 | More converters for IBM803, IBM901, IBM902, IBM921, IBM1008, IBM1097, | |
6528 | IBM1112, IBM1123, IBM1130, IBM1140, IBM1141, IBM1142, IBM1143, IBM1144, | |
6529 | IBM1145, IBM1146, IBM1147, IBM1148, IBM1149, IBM1166, IBM1167, IBM4517, | |
6530 | IBM4899, IBM4909, IBM4971, IBM5347, IBM9030, IBM9066, IBM9448, IBM12712, | |
6531 | IBM16804, IBM1364, IBM1371, IBM1388, IBM1390, and IBM1399 contributed | |
6532 | by Masahide Washizawa. | |
6533 | ||
6534 | * It is now possible to install audit libraries for the dynamic linker using | |
6535 | LD_AUDIT. Implemented by Ulrich Drepper. | |
6536 | ||
6537 | * The LinuxThreads add-on, providing pthreads on Linux 2.4 kernels, is no | |
6538 | longer supported. The new NPTL implementation requires Linux 2.6 kernels. | |
6539 | For a libc and libpthread that works well on Linux 2.4 kernels, we | |
6540 | recommend using the stable 2.3 branch. | |
6541 | ||
6542 | * The new function open_wmemstream acts like open_memstream, | |
6543 | but uses a wchar_t wide character buffer. | |
6544 | ||
6545 | * The new function ppoll is an improved version of poll, similar to pselect. | |
6546 | ||
6547 | * New interfaces: fdopendir, openat, fstatat, fchmodat, fchownat, | |
6548 | futimesat, faccessat, mkdirat, mkfifoat, mknodat, | |
6549 | renameat, unlinkat, linkat, symlinkat, readlinkat. | |
6550 | ||
6551 | * New Linux kernel system calls: unshare, | |
6552 | inotify_init, inotify_add_watch, inotify_rm_watch. | |
6553 | ||
6554 | * The euidaccess function is now also known by the alias eaccess, | |
6555 | for compatibility with some other systems. | |
6556 | ||
6557 | * Timezone data updated to 2006b version. | |
6558 | \f | |
6559 | Version 2.3.6 | |
6560 | ||
6561 | * The following bugs are resolved with this release: | |
6562 | ||
6563 | 38, 253, 549, 622, 653, 721, 758, 851, 877, 915, 934, 955, 961, | |
6564 | 1016, 1037, 1076, 1079, 1080, 1081, 1082, 1083, 1084, 1085, 1086, | |
6565 | 1087, 1088, 1090, 1091, 1092, 1093, 1094, 1095, 1096, 1097, 1098, | |
6566 | 1099, 1100, 1101, 1102, 1103, 1104, 1105, 1106, 1107, 1108, 1109, | |
6567 | 1110, 1111, 1112, 1113, 1125, 1137, 1138, 1249, 1250, 1251, 1252, | |
6568 | 1253, 1254, 1350, 1358, 1394, 1438, 1498, 1534 | |
6569 | ||
6570 | Visit <https://sourceware.org/bugzilla/> for the details of each bug. | |
6571 | ||
6572 | * As of this release, GCC 4 can be used to compile the C Library. | |
6573 | ||
6574 | * Timezone data updated to 2005m version. | |
6575 | \f | |
6576 | Version 2.3.5 | |
6577 | ||
6578 | * The following bugs are resolved with this release: | |
6579 | ||
6580 | 284, 592, 602, 626, 633, 640, 650, 661, 671, 681, 693, 700, 710, 719, | |
6581 | 722, 723, 725, 726, 727, 728, 729, 730, 731, 732, 733, 734, 735, 736, | |
6582 | 737, 738, 739, 740, 741, 742, 743, 744, 745, 765, 767, 768, 769, 776, | |
6583 | 777, 787, 821, 822, 823, 825 | |
6584 | ||
6585 | Visit <https://sourceware.org/bugzilla/> for the details of each bug. | |
6586 | \f | |
6587 | Version 2.3.4 | |
6588 | ||
6589 | * Support for RFC 3678. Real implementations exist only for Linux so far. | |
6590 | Implemented by Ulrich Drepper. | |
6591 | ||
6592 | * nscd can now cache entries persistently. Expiring entries are reloaded. | |
6593 | For speedups the cache can be shared in memory with client processes. | |
6594 | Implemented by Ulrich Drepper. | |
6595 | ||
6596 | * nscd can now perform SELinux checks. | |
6597 | Implemented by Matthew Rickard <mjricka@epoch.ncsc.mil>. | |
6598 | ||
6599 | * getaddrinfo queries are now cached. Canonical name lookup is performed | |
6600 | efficiently. | |
6601 | Implemented by Ulrich Drepper. | |
6602 | ||
6603 | * The nothrow function attribute is used when headers are used by gcc when | |
6604 | compiling C code. This can avoid significant amounts of exception | |
6605 | handling data. | |
6606 | ||
6607 | * The malloc functions perform more error checking and are stricter when | |
6608 | it comes to reacting on errors. The default action is to terminate | |
6609 | the process after showing an error message. Implemented by Ulrich Drepper. | |
6610 | ||
6611 | * Reverse lookups of IPv6 addresses does not use bit string or .ip6.int | |
6612 | lookups anymore unless explicitly requested. Implemented by Ulrich Drepper. | |
6613 | ||
6614 | * Namespaces in ld.so are implemented. DSOs can be loaded in separate | |
6615 | namespaces using the new function dlmopen(). This feature is of course, | |
6616 | like most other dynamic loading functionality, not available in statically | |
6617 | linked applications. Implemented by Ulrich Drepper. | |
6618 | ||
6619 | * Low-overhead boundary checking variants of string and some stdio functions | |
6620 | were added. These are to be used in conjunction with a gcc patch by | |
6621 | Jakub Jelinek which adds calls to these functions if possible. | |
6622 | Implemented by Jakub Jelinek and Ulrich Drepper. | |
6623 | ||
6624 | * Old code for several operating systems and machine architectures that | |
6625 | have not been in working condition in a long time have been removed from | |
6626 | the main source tree maintained by the GNU C Library's maintainers. | |
6627 | These files are now reside in the separate `ports' source module | |
6628 | that is usable as an add-on when building the library. | |
6629 | \f | |
6630 | Version 2.3.3 | |
6631 | ||
6632 | * New functions `dladdr1' and `dlinfo' in <dlfcn.h> provide more ways to | |
6633 | interrogate the dynamic linker, compatible with the Solaris interface. | |
6634 | ||
6635 | * ELF thread-local storage support (TLS) now works on PowerPC and PowerPC64; | |
6636 | implemented by Paul Mackerras, Steven Munroe, and Roland McGrath. | |
6637 | ||
6638 | * getifaddrs now uses the netlink interface on Linux to get its information. | |
6639 | Implemented by Thorsten Kukuk. | |
6640 | ||
6641 | * getaddrinfo now implements AI_V4MAPPED, AI_ALL, and AI_ADDRCONF. | |
6642 | Implemented by Ulrich Drepper. | |
6643 | ||
6644 | * support for non-executable stacks on x86 has been added. Changes mostly | |
6645 | by Roland McGrath. | |
6646 | ||
6647 | * regex is now much faster for multibyte locales. Changes by Jakub Jelinek | |
6648 | and Ulrich Drepper. | |
6649 | ||
6650 | * getaddrinfo now performs destination address selection according to | |
6651 | RFC 3484. | |
6652 | \f | |
6653 | Version 2.3.2 | |
6654 | ||
6655 | * Thread-safe interfaces for many functions that access locale data | |
6656 | were added in version 2.3, but these features were omitted from NEWS. | |
6657 | Many functions have variants with an `_l' suffix that take a `locale_t' | |
6658 | object as a parameter rather than consulting the current locale. | |
6659 | The new functions `newlocale', `duplocale', and `freelocale' in <locale.h> | |
6660 | create and maintain `locale_t' objects. Additionally, the new function | |
6661 | `uselocale' sets "the current locale" (as used by functions not so | |
6662 | parameterized) set for an individual thread. These features were added | |
6663 | in version 2.3, implemented by Ulrich Drepper and Roland McGrath. | |
6664 | ||
6665 | * The functions getresuid, getresgid, setresuid, and setresgid, which | |
6666 | have long been available on Linux, are now declared in <unistd.h> | |
6667 | and are now also available on the Hurd. | |
6668 | ||
6669 | * ELF thread-local storage support (TLS) now works on x86-64. | |
6670 | ||
6671 | * The new dynamic string token $LIB is expanded in shared library names. | |
6672 | This normally expands to lib, but on some 64-bit platforms to lib64 instead. | |
6673 | ||
6674 | * Aldy Hernandez contributed complete software floating point support for | |
6675 | PowerPC machines with no FPU. | |
6676 | ||
6677 | * fexecve is implemented on Linux. | |
6678 | ||
6679 | * The `btowc' function should work at least twice as fast due to | |
6680 | specialized callbacks in the iconv modules. Implemented by Bruno Haible. | |
6681 | ||
6682 | * With appropriate thread add-ons cancelable functions are now implemented | |
6683 | in libc.so as well. No need to call the function in libpthread. This | |
6684 | change allowed to finally disable the incorrect and expensive handling | |
6685 | of weak definition in ld.so. | |
6686 | ||
6687 | * Yet more PLT entries in libc.so have been removed. We finally arrived | |
6688 | at the bare minimum. Startup times improved appropriately. | |
6689 | ||
6690 | * Support for the new Linux/x86 system call interface was added. The | |
6691 | AT_SYSINFO auxiliary vector entry is recognized and handled. | |
6692 | \f | |
6693 | Version 2.3 | |
6694 | ||
6695 | * Masahide Washizawa contributed iconv modules for IBM1163 and IBM1164 | |
6696 | charsets. | |
6697 | ||
6698 | * iconv (the program and the interface) now accepts empty names (excluding | |
6699 | options like //TRANSLIT) to mean "use charset of current locale". | |
6700 | ||
6701 | * localedef can now transliterate characters in strings which are not in | |
6702 | the provided charmap. The information from the input locale is used. | |
6703 | ||
6704 | * Prelinking support was added for ELF targets. This requires additional | |
6705 | tools and recent versions of the GNU binutils. Contributed by Jakub Jelinek. | |
6706 | ||
6707 | * Read-only stdio streams now use mmap to speed up operation by eliminating | |
6708 | copying and buffer underflows. To use add 'm' to the mode string of | |
6709 | the fopen/fdopen/freopen call. Implemented by Ulrich Drepper. | |
6710 | ||
6711 | * The malloc functions were completely rewritten by Wolfram Gloger based | |
6712 | on Doug Lea's malloc-2.7.0.c. | |
6713 | ||
6714 | * Isamu Hasegawa contributed a completely new and POSIX-conformant | |
6715 | implementation of regex. | |
6716 | ||
6717 | * Bruno Haible upgraded the iconv and locale implementation to support | |
6718 | Unicode 3.2. | |
6719 | ||
6720 | * Contents of the LC_* and LANG environment variables in the CEN style are | |
6721 | not recognized anymore. It never was used. Change by Ulrich Drepper. | |
6722 | ||
6723 | * The runtime (ld.so, libc, libpthread for Linux) now can handle the ELF | |
6724 | thread-local storage (TLS) ABI on some platforms. | |
6725 | Changes by Ulrich Drepper. SH support by Kaz Kojima. | |
6726 | ||
6727 | * Bruno Haible contributed iconv converters for ISO-2022-JP-3, SHIFT JIS-X0213, | |
6728 | EUC-JISX0213, and TSCII. | |
6729 | ||
6730 | * New header <ifaddrs.h> with functions `getifaddrs' and `freeifaddrs': | |
6731 | BSD-compatible interface for getting all network interface addresses. | |
6732 | Implementation for IPv4 by Roland McGrath. | |
6733 | ||
6734 | * Loading of locale data is faster due to the introduction of a locale | |
6735 | archive. Implemented by Roland McGrath and Ulrich Drepper. | |
6736 | ||
6737 | * Startup times are significantly reduced by not using exported functions | |
6738 | inside the library itself. Changes by Jakub Jelinek, Roland McGrath, | |
6739 | and Ulrich Drepper. | |
6740 | ||
6741 | * Steven Munroe contributed a port to PowerPC64/Linux. | |
6742 | \f | |
6743 | Version 2.2.6 | |
6744 | ||
6745 | * The Hurd now uses the GNU libio implementation of stdio rather than the | |
6746 | old GNU stdio implementation, and uses a new ABI (libc.so.0.3). | |
6747 | ||
6748 | * The Hurd on x86 now has the `ioperm' function and <sys/io.h> header file | |
6749 | with the same behavior as the Linux system call of the same name. | |
6750 | \f | |
6751 | Version 2.2.5 | |
6752 | ||
6753 | * Stephen Moshier implemented log2, log10, powl and cbrtl for the | |
6754 | 128-bit long double format. | |
6755 | ||
6756 | * Masahide Washizawa contributed iconv modules for IBM1132, IBM1133, IBM1160, | |
6757 | IBM1161, and IBM1162 charsets. | |
6758 | ||
6759 | * Andreas Jaeger contributed a port to x86-64/Linux. | |
6760 | ||
6761 | * Peter Bruin contributed a port to PowerPC/Hurd. | |
6762 | ||
6763 | * libc functions using I/O streams now can handle wide-oriented streams | |
6764 | as well. | |
6765 | ||
6766 | * optimizations in the dynamic linker. Binaries created by recent binutils | |
6767 | versions start up quicker due to reduced time spent on relocations. | |
6768 | ||
6769 | * Support for use of gcc3 added by Jakub Jelinek and HJ Lu. | |
6770 | \f | |
6771 | Version 2.2.4 | |
6772 | ||
6773 | * Stephen Moshier implemented cosh, expm1, log1p, acos, sinh, tanh, | |
6774 | asinh, atanh, j0 for the 128-bit long double format. | |
6775 | ||
6776 | * Bruno Haible updated all the code handling Unicode in some form to | |
6777 | support Unicode 3.1. | |
6778 | ||
6779 | * Speed of regex for single-byte locales is back to previous levels. | |
6780 | Patch by Isamu Hasegawa. | |
6781 | ||
6782 | * Alpha, SPARC, and IA-64 now also using floating stacks. | |
6783 | ||
6784 | * Startup time of internationalized applications greatly improved through | |
6785 | iconv cache. Use iconvconfig to generate the cache file. | |
6786 | Contributed by Ulrich Drepper. | |
6787 | ||
6788 | * The IA-64 specific part of ld.so was rewritten to eliminate some pretty | |
6789 | severe performance problems. Patch by David Mosberger. | |
6790 | ||
6791 | * The Hurd port got a lot more functionality like AIO, various stdio | |
6792 | extensions, etc. Mainly done by Roland McGrath. | |
6793 | ||
6794 | * mtrace can now lookup symbols in shared libraries. | |
6795 | \f | |
6796 | Version 2.2.3 | |
6797 | ||
6798 | * Intel's IA-64 math library is largely integrated. It provides fast and | |
6799 | accurate implementatations for most basic and standard math functions | |
6800 | in float, double, and long double format. | |
6801 | ||
6802 | * Stephen Moshier implemented j0, j1, jn, y0, y1, yn, lgamma, erf, erfc, | |
6803 | and asin for the 96-bit long double format and asin, log, tan for the | |
6804 | 128-bit long double format. | |
6805 | ||
6806 | * The beginning of a last-bit accurate math library by IBM Haifa were added. | |
6807 | The basic double functions exist today. Contributed by Abraham Ziv | |
6808 | <ziv@il.ibm.com>, Moshe Olshansky <olshansk@il.ibm.com>, Ealan Henis | |
6809 | <ealan@il.ibm.com>, and Anna Reitman <reitman@il.ibm.com>. | |
6810 | ||
6811 | * An asynchronous name lookup library was added. The interface is designed | |
6812 | after POSIX AIO. The proposal was circulated beforehand to get comments. | |
6813 | No negative ones came in. Implemented by Ulrich Drepper. | |
6814 | ||
6815 | * Port to S390/64bit contributed by Martin Schwidefsky | |
6816 | <schwidefsky@de.ibm.com>. | |
6817 | ||
6818 | * David Mosberger <davidm@hpl.hp.com> implemented the setcontext family | |
6819 | of functions for Linux/IA-64. | |
6820 | ||
6821 | * The RPC code is now thread safe. Threads can now use the same service | |
6822 | of different services at the same time. Patch by Eric Norum | |
6823 | <eric.norum@usask.ca> with some help by Ulrich Drepper. | |
6824 | ||
6825 | * Martin Schwidefsky <schwidefsky@de.ibm.com> implemented the setcontext | |
6826 | family of functions for Linux/S390. | |
6827 | ||
6828 | * Ulrich Drepper <drepper@redhat.com> implemented the setcontext family | |
6829 | of functions for Linux/x86. | |
6830 | ||
6831 | * Port to Linux/CRIS contributed by Axis Communications. | |
6832 | \f | |
6833 | Version 2.2.2 | |
6834 | ||
6835 | * Lots of headers were cleaned up. Using the tool in the conform/ subdir | |
6836 | we can now check for namespace violations and missing declarations. The | |
6837 | result is that almost all headers are now Unix-compliant (as defined in | |
6838 | the upcoming XPG6). The negative side is that some programs might need | |
6839 | corrections, too, if they depend on the incorrect form of the headers in | |
6840 | previous versions which defined too many symbols and included too many | |
6841 | other headers. | |
6842 | ||
6843 | * regex now handles multibyte character sets correctly. | |
6844 | Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. | |
6845 | ||
6846 | * iconv (the program) does now conform to the upcoming XPG6 and handles | |
6847 | charmaps. Instead of the charset names the path of charmaps can be | |
6848 | provided and the conversion happens based on this data. | |
6849 | Contributed by Ulrich Drepper. | |
6850 | ||
6851 | * The locale program now provides detailed information about the installed | |
6852 | locales. While | |
6853 | ||
6854 | locale -a | |
6855 | ||
6856 | only lists the names of the supported locales | |
6857 | ||
6858 | locale -a --verbose | |
6859 | ||
6860 | provides details such as country, language, and codeset name. | |
6861 | Contributed by Ulrich Drepper. | |
6862 | \f | |
6863 | Version 2.2.1 | |
6864 | ||
6865 | * The gencat program now parses the input file according to the charset | |
6866 | selected by the LC_CTYPE category. This is important for stateful | |
6867 | character sets. To make generating catalogs easier there is a way | |
6868 | to overwrite the charset selected by the locale: before the first | |
6869 | message or $ quote line the catalog can contain a line like | |
6870 | ||
6871 | $ codeset=ISO-8859-2 | |
6872 | ||
6873 | to select the charset (ISO-8859-2 in this case). | |
6874 | ||
6875 | Implemented by Shinya Hanataka and Ulrich Drepper. | |
6876 | ||
6877 | * New codeset conversion modules: IBM-922 (Estonia PC codeset), | |
6878 | IBM-1124 (Ukraine PC codeset), IBM-1129 (Vietnamese PC codeset). | |
6879 | Contributed by Masahide Washizawa <washi@jp.ibm.com>. | |
6880 | ||
6881 | * Optimized string functions for Alpha ev6 and ev67 by Richard Henderson | |
6882 | <rth@redhat.com> and Rick Gorton <rick.gorton@alpha-processor.com>. | |
6883 | ||
6884 | * The LANGUAGE environment variable is now ignored unless the locale is | |
6885 | changed from the default "C" locale. | |
6886 | ||
6887 | * The usual bug fixes. | |
6888 | \f | |
6889 | Version 2.2 | |
6890 | ||
6891 | * Greg McGary added runtime support for bounds checking using gcc's | |
6892 | new -fbounded-pointers option. ix86 target is complete. PowerPC | |
6893 | is in progress. | |
6894 | ||
6895 | * Thorsten Kukuk added secure mode support to nscd. | |
6896 | ||
6897 | * The Berkeley db libraries have been removed. | |
6898 | ||
6899 | Related, the nss_db module is now in a separate package since it | |
6900 | obviously requires a database library being available. | |
6901 | ||
6902 | * Wide character I/O streams implemented by Ulrich Drepper. | |
6903 | ||
6904 | * Functions from the extended socket API added by Ulrich Drepper. | |
6905 | ||
6906 | * Functions feenableexcept and fedisableexcept to control the | |
6907 | behaviour of individual exceptions have been added by Andreas Jaeger. | |
6908 | ||
6909 | * ldconfig program added by Andreas Jaeger and Jakub Jelinek. | |
6910 | ||
6911 | * The resolver code has been updated from bind 8.2.3-T5B which supports | |
6912 | threads. The integration was done by Andreas Jaeger, Adam D. Bradley, | |
6913 | and Mark Kettenis. | |
6914 | ||
6915 | This change could in some situations effect backward compatibility. Since | |
6916 | now `_res' is a thread-local instead of a global variable, modifying it | |
6917 | in one thread does not have any effect in other threads. | |
6918 | ||
6919 | The resolver library was also extended to allow IPv6 as the transport | |
6920 | protocol for the requests. This work was done by Stig Venaas. | |
6921 | ||
6922 | * Compatibility code for K&R C compilers has been removed from the | |
6923 | header files. A ISO C compiler is needed to use the library | |
6924 | (conforming to either C89 or C99 standard). | |
6925 | ||
6926 | * Complete rewrite of the localedef program to support multibyte character | |
6927 | sets. Implement handling of ISO 14651 and ISO 14652. Rewrite strcoll, | |
6928 | strxfrm, wcscoll, and wcsxfrm functions. Make isw*() functions work. | |
6929 | Implemented by Ulrich Drepper. | |
6930 | ||
6931 | Bruno Haible significantly improved the generation and use of the data | |
6932 | structures for the wide character tables. | |
6933 | ||
6934 | * Plural handling in gettext implemented by Ulrich Drepper. | |
6935 | ||
6936 | * The utmp daemon has been removed. | |
6937 | ||
6938 | * The port to MIPS-Linux has been finished by Andreas Jaeger. | |
6939 | ||
6940 | * A port to Hitachi SH3 and SH4 has been contributed by Kazumoto Kojima | |
6941 | and Yutaka Niibe. | |
6942 | ||
6943 | * POSIX clocks and timers implemented by Kaz Kylheku and Ulrich Drepper. | |
6944 | ||
6945 | * POSIX barriers implemented by Kaz Kylheku. | |
6946 | ||
6947 | * POSIX spawn function family implemented by Ulrich Drepper. | |
6948 | ||
6949 | * POSIX spinlocks are now available. | |
6950 | ||
6951 | * Timed wait functions for mutex, rwlock, and semaphores are implemented. | |
6952 | ||
6953 | * the configure option --enable-kernel=X.Y.Z allows to strip out | |
6954 | compatibility for kernel versions before X.Y.Z. This is currently only | |
6955 | implemented for Linux. | |
6956 | ||
6957 | * the sockaddr_in6 structure changed. The IPv6 working group added a new | |
6958 | field sin6_scope_id. This means that all programs using IPv6 should be | |
6959 | recompiled. Don't expect binary compatibility with previous glibc | |
6960 | versions. | |
6961 | ||
6962 | * various conversion modules for IBM character sets contributed by | |
6963 | Masahide Washizawa. | |
6964 | ||
6965 | * IA-64 port by Jes Sorensen and HJ Lu. | |
6966 | \f | |
6967 | Version 2.1.3 | |
6968 | ||
6969 | * bug fixes | |
6970 | ||
6971 | \f | |
6972 | Version 2.1.2 | |
6973 | ||
6974 | * bug fixes | |
6975 | ||
6976 | \f | |
6977 | Version 2.1.1 | |
6978 | ||
6979 | * New ISO C 9x function _Exit, imaxabs, and imaxdiv are added. | |
6980 | ||
6981 | * New xdr functions are added; some rpc functions are now 64bit clean. | |
6982 | ||
6983 | * Fixed a number of bugs and memory leaks (especially in NIS+ code). | |
6984 | ||
6985 | * Fixed known incompatibilities with glibc 2.0. | |
6986 | ||
6987 | * New functions lock64, strchrnul, rawmemchr, getutmp and getutmpx. | |
6988 | ||
6989 | * Optimized a number of functions (especially the ELF dynamic loader). | |
6990 | ||
6991 | * Update timezone data files. | |
6992 | ||
6993 | * lots of charmaps corrections | |
6994 | ||
6995 | * some new locale definitions and charmaps | |
6996 | ||
6997 | \f | |
6998 | Version 2.1 | |
6999 | ||
7000 | * Richard Henderson corrected size of struct timeval on Linux/Alpha to | |
7001 | conform to POSIX member type requirements. Symbol versions have been | |
7002 | adjusted as needed within the library, and for direct use by applications, | |
7003 | but there is potential for problems if third-party libraries use | |
7004 | struct timeval as part of their interface. This does not present | |
7005 | a problem for X and other "essential" system libraries. | |
7006 | ||
7007 | * An additional locale model to support C++ Standard Library locale | |
7008 | model and probably more was implemented by Ulrich Drepper. | |
7009 | ||
7010 | * Eric Youngdale and Ulrich Drepper implemented versioning of objects on | |
7011 | symbol level. | |
7012 | ||
7013 | * Miles Bader provided the `argp' function family to support hierarchical | |
7014 | command line argument parsing, layered on top of getopt. | |
7015 | ||
7016 | * strtod accepts new hexadecimal floating-point format from ISO C 9X. | |
7017 | ||
7018 | * printf has two new specifiers %a and %A to print hexadecimal flaoting-point | |
7019 | numbers. | |
7020 | ||
7021 | * scanf recognizes the %a and %A format for scanning floating point numbers. | |
7022 | ||
7023 | * the new headers <stdint.h> and <inttypes.h> from ISO C 9X provides | |
7024 | information and interfaces for the available integer types. | |
7025 | ||
7026 | * about 130 new math functions were added to implement the ISO C9x math | |
7027 | library. | |
7028 | ||
7029 | * the new header <complex.h> contains definitions of the complex math | |
7030 | functions from ISO C 9X. | |
7031 | ||
7032 | * the new header <tgmath.h> defines generic macros to use complex or | |
7033 | real valued functions. | |
7034 | ||
7035 | * Thorsten Kukuk provided an implementation for NIS+, securelevel 0, 1 and 2. | |
7036 | ||
7037 | * Andreas Jaeger provided a test suite for the math library. | |
7038 | ||
7039 | * Mark Kettenis implemented the utmpx interface and an utmp daemon. | |
7040 | ||
7041 | * Ulrich Drepper added character set conversion functions (iconv). | |
7042 | ||
7043 | * Optimized string functions have been added. | |
7044 | ||
7045 | * The localedata addon is now part of glibc. | |
7046 | ||
7047 | * An implementation of profiling shared libraries was added by Ulrich Drepper. | |
7048 | ||
7049 | * Thorsten Kukuk and Ulrich Drepper provided an implementation for a caching | |
7050 | daemon for NSS (nscd). | |
7051 | ||
7052 | Missing a better place here are some numbers on improvements. Under | |
7053 | Linux 2.1.125 un-tar-ing the kernel sources takes | |
7054 | ||
7055 | user system wall | |
7056 | ||
7057 | using local files 12.19s 6.88s 22.91s | |
7058 | ||
7059 | using NIS 13.92s 8.91s 26.34s | |
7060 | ||
7061 | using NIS & nscd 10.37s 7.34s 25.30s | |
7062 | ||
7063 | using NIS+ 27.57s 30.37s 640.46s | |
7064 | ||
7065 | using NIS+ & nscd 10.25s 7.83s 26.51s | |
7066 | ||
7067 | using NIS & old nscd [1] 13.83s 8.32s 29.60s | |
7068 | ||
7069 | Keep in mind that non-namelookup related operations dominate above times. | |
7070 | It was just a common complain that using NIS+ unpacking the kernel is | |
7071 | horribly slow. | |
7072 | ||
7073 | [1] The old nscd implementation is not available anymore since it was | |
7074 | distributed with glibc up to version 2.0.98 and thus is now replaced. | |
7075 | ||
7076 | * Tim Waugh provided an implementation of the POSIX.2 wordexp function family. | |
7077 | ||
7078 | * Mark Kettenis provided a Hesiod NSS module. | |
7079 | ||
7080 | * The ELF dynamic loader knows how to process dynamic string tokens ($ORIGIN | |
7081 | and $PLATFORM) in RPATHs and similar strings (Ulrich Drepper). | |
7082 | ||
7083 | * rcmd can now handle netgroups (Dick Streefland). | |
7084 | ||
7085 | * A port to the ARM was contributed by Phil Blundell, Pat Beirne and Scott | |
7086 | Bambrough. | |
7087 | ||
7088 | * Support for the IPv6 protocol has been added to the socket API, as per the | |
7089 | latest draft standards. | |
7090 | ||
7091 | * Support for Linux 2.2 has been added. | |
7092 | ||
7093 | * Interface changes relative to the latest 2.0.x release: | |
7094 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
7095 | addseverity NEW: Unix98 | |
7096 | alphasort64 NEW: LFS | |
7097 | argp_err_exit_status NEW: argp, GNU ext | |
7098 | argp_error NEW: argp, GNU ext | |
7099 | argp_failure NEW: argp, GNU ext | |
7100 | argp_help NEW: argp, GNU ext | |
7101 | argp_parse NEW: argp, GNU ext | |
7102 | argp_program_bug_address NEW: argp, GNU ext | |
7103 | argp_program_version NEW: argp, GNU ext | |
7104 | argp_program_version_hook NEW: argp, GNU ext | |
7105 | argp_state_help NEW: argp, GNU ext | |
7106 | argp_usage NEW: argp, GNU ext | |
7107 | authdes_create NEW: Secure RPC | |
7108 | authdes_getucred NEW: Secure RPC | |
7109 | authdes_pk_create NEW: Secure RPC | |
7110 | backtrace NEW: GNU ext. | |
7111 | backtrace_symbols NEW: GNU ext. | |
7112 | backtrace_symbols_fd NEW: GNU ext. | |
7113 | cacos NEW: ISO C 9x | |
7114 | cacosf NEW: ISO C 9x | |
7115 | cacosh NEW: ISO C 9x | |
7116 | cacoshf NEW: ISO C 9x | |
7117 | cacoshl NEW: ISO C 9x | |
7118 | cacosl NEW: ISO C 9x | |
7119 | capget NEW: kernel | |
7120 | capset NEW: kernel | |
7121 | carg NEW: ISO C 9x | |
7122 | cargf NEW: ISO C 9x | |
7123 | cargl NEW: ISO C 9x | |
7124 | casin NEW: ISO C 9x | |
7125 | casinf NEW: ISO C 9x | |
7126 | casinh NEW: ISO C 9x | |
7127 | casinhf NEW: ISO C 9x | |
7128 | casinhl NEW: ISO C 9x | |
7129 | casinl NEW: ISO C 9x | |
7130 | catan NEW: ISO C 9x | |
7131 | catanf NEW: ISO C 9x | |
7132 | catanh NEW: ISO C 9x | |
7133 | catanhf NEW: ISO C 9x | |
7134 | catanhl NEW: ISO C 9x | |
7135 | catanl NEW: ISO C 9x | |
7136 | cbc_crypt NEW: Secure RPC | |
7137 | ccos NEW: ISO C 9x | |
7138 | ccosf NEW: ISO C 9x | |
7139 | ccosh NEW: ISO C 9x | |
7140 | ccoshf NEW: ISO C 9x | |
7141 | ccoshl NEW: ISO C 9x | |
7142 | ccosl NEW: ISO C 9x | |
7143 | cexp NEW: ISO C 9x | |
7144 | cexpf NEW: ISO C 9x | |
7145 | cexpl NEW: ISO C 9x | |
7146 | cimag NEW: ISO C 9x | |
7147 | cimagf NEW: ISO C 9x | |
7148 | cimagl NEW: ISO C 9x | |
7149 | clearerr_locked REMOVED | |
7150 | clntunix_create NEW: sunrpc ext | |
7151 | clog NEW: ISO C 9x | |
7152 | clog10 NEW: ISO C 9x | |
7153 | clog10f NEW: ISO C 9x | |
7154 | clog10l NEW: ISO C 9x | |
7155 | clogf NEW: ISO C 9x | |
7156 | clogl NEW: ISO C 9x | |
7157 | conj NEW: ISO C 9x | |
7158 | conjf NEW: ISO C 9x | |
7159 | conjl NEW: ISO C 9x | |
7160 | cpow NEW: ISO C 9x | |
7161 | cpowf NEW: ISO C 9x | |
7162 | cpowl NEW: ISO C 9x | |
7163 | cproj NEW: ISO C 9x | |
7164 | cprojf NEW: ISO C 9x | |
7165 | cprojl NEW: ISO C 9x | |
7166 | creal NEW: ISO C 9x | |
7167 | crealf NEW: ISO C 9x | |
7168 | creall NEW: ISO C 9x | |
7169 | creat64 NEW: LFS | |
7170 | csin NEW: ISO C 9x | |
7171 | csinf NEW: ISO C 9x | |
7172 | csinh NEW: ISO C 9x | |
7173 | csinhf NEW: ISO C 9x | |
7174 | csinhl NEW: ISO C 9x | |
7175 | csinl NEW: ISO C 9x | |
7176 | csqrt NEW: ISO C 9x | |
7177 | csqrtf NEW: ISO C 9x | |
7178 | csqrtl NEW: ISO C 9x | |
7179 | ctan NEW: ISO C 9x | |
7180 | ctanf NEW: ISO C 9x | |
7181 | ctanh NEW: ISO C 9x | |
7182 | ctanhf NEW: ISO C 9x | |
7183 | ctanhl NEW: ISO C 9x | |
7184 | ctanl NEW: ISO C 9x | |
7185 | des_setparity NEW: Secure RPC | |
7186 | ecb_crypt NEW: Secure RPC | |
7187 | endutxent NEW: Unix98 | |
7188 | exp10 NEW: ISO C 9x | |
7189 | exp10f NEW: ISO C 9x | |
7190 | exp10l NEW: ISO C 9x | |
7191 | exp2 NEW: ISO C 9x | |
7192 | exp2f NEW: ISO C 9x | |
7193 | exp2l NEW: ISO C 9x | |
7194 | fattach NEW: STREAMS | |
7195 | fdetach NEW: STREAMS | |
7196 | fdim NEW: ISO C 9x | |
7197 | fdimf NEW: ISO C 9x | |
7198 | fdiml NEW: ISO C 9x | |
7199 | feclearexcept NEW: ISO C 9x | |
7200 | fegetenv NEW: ISO C 9x | |
7201 | fegetexceptflag NEW: ISO C 9x | |
7202 | fegetround NEW: ISO C 9x | |
7203 | feholdexcept NEW: ISO C 9x | |
7204 | feof_locked REMOVED | |
7205 | feraiseexcept NEW: ISO C 9x | |
7206 | ferror_locked REMOVED | |
7207 | fesetenv NEW: ISO C 9x | |
7208 | fesetexceptflag NEW: ISO C 9x | |
7209 | fesetround NEW: ISO C 9x | |
7210 | fetestexcept NEW: ISO C 9x | |
7211 | feupdateenv NEW: ISO C 9x | |
7212 | fflush_locked REMOVED | |
7213 | ffsl NEW: GNU ext. | |
7214 | ffsll NEW: GNU ext. | |
7215 | fgetpos64 NEW: LFS | |
7216 | fgets_unlocked NEW: GNU ext. | |
7217 | fileno_locked REMOVED | |
7218 | fma NEW: ISO C 9x | |
7219 | fmaf NEW: ISO C 9x | |
7220 | fmal NEW: ISO C 9x | |
7221 | fmax NEW: ISO C 9x | |
7222 | fmaxf NEW: ISO C 9x | |
7223 | fmaxl NEW: ISO C 9x | |
7224 | fmin NEW: ISO C 9x | |
7225 | fminf NEW: ISO C 9x | |
7226 | fminl NEW: ISO C 9x | |
7227 | fmtmsg NEW: Unix98 | |
7228 | fopen64 NEW: LFS | |
7229 | fputc_locked REMOVED | |
7230 | fputs_unlocked NEW: GNU ext. | |
7231 | fread_unlocked NEW: GNU ext. | |
7232 | freopen64 NEW: LFS | |
7233 | fseeko NEW: Unix98 | |
7234 | fsetpos64 NEW: LFS | |
7235 | fstatfs64 NEW: LFS | |
7236 | fstatvfs NEW: Unix98 | |
7237 | fstatvfs64 NEW: LFS | |
7238 | ftello NEW: Unix98 | |
7239 | ftello64 NEW: LFS | |
7240 | ftruncate64 NEW: LFS | |
7241 | ftw64 NEW: LFS | |
7242 | fwrite_unlocked NEW: GNU ext. | |
7243 | gai_strerror NEW: IPv6 | |
7244 | gamma_r REMOVED | |
7245 | gammaf_r REMOVED | |
7246 | gammal_r REMOVED | |
7247 | getchar_locked REMOVED | |
7248 | getdate NEW: Unix98 | |
7249 | getdate_err NEW: Unix98 | |
7250 | getdate_r NEW: GNU ext. | |
7251 | getmsg NEW: STREAMS | |
7252 | getnameinfo NEW: IPv6 | |
7253 | getnetname NEW: Secure RPC | |
7254 | getpmsg NEW: STREAMS | |
7255 | getpt NEW: Unix98 PTY | |
7256 | getrlimit64 NEW: LFS | |
7257 | getutxent NEW: Unix98 | |
7258 | getutxid NEW: Unix98 | |
7259 | getutxline NEW: Unix98 | |
7260 | glob64 NEW: GNU ext. | |
7261 | globfree64 NEW: GNU ext. | |
7262 | gnu_get_libc_release NEW: GNU ext. | |
7263 | gnu_get_libc_version NEW: GNU ext. | |
7264 | grantpt NEW: Unix98 PTY | |
7265 | host2netname NEW: Secure RPC | |
7266 | iconv NEW: iconv | |
7267 | iconv_close NEW: iconv | |
7268 | iconv_open NEW: iconv | |
7269 | if_freenameindex NEW: IPv6 | |
7270 | if_indextoname NEW: IPv6 | |
7271 | if_nameindex NEW: IPv6 | |
7272 | if_nametoindex NEW: IPv6 | |
7273 | in6addr_any NEW: IPv6 | |
7274 | in6addr_loopback NEW: IPv6 | |
7275 | inet6_isipv4mapped NEW: IPv6 | |
7276 | isastream NEW: STREAMS | |
7277 | iswblank NEW: GNU ext. | |
7278 | key_decryptsession NEW: Secure RPC | |
7279 | key_decryptsession_pk NEW: Secure RPC | |
7280 | key_encryptsession NEW: Secure RPC | |
7281 | key_encryptsession_pk NEW: Secure RPC | |
7282 | key_gendes NEW: Secure RPC | |
7283 | key_get_conv NEW: Secure RPC | |
7284 | key_secretkey_is_set NEW: Secure RPC | |
7285 | key_setnet NEW: Secure RPC | |
7286 | key_setsecret NEW: Secure RPC | |
7287 | llrint NEW: ISO C 9x | |
7288 | llrintf NEW: ISO C 9x | |
7289 | llrintl NEW: ISO C 9x | |
7290 | llround NEW: ISO C 9x | |
7291 | llroundf NEW: ISO C 9x | |
7292 | llroundl NEW: ISO C 9x | |
7293 | log2 NEW: ISO C 9x | |
7294 | log2f NEW: ISO C 9x | |
7295 | log2l NEW: ISO C 9x | |
7296 | lrint NEW: ISO C 9x | |
7297 | lrintf NEW: ISO C 9x | |
7298 | lrintl NEW: ISO C 9x | |
7299 | lround NEW: ISO C 9x | |
7300 | lroundf NEW: ISO C 9x | |
7301 | lroundl NEW: ISO C 9x | |
7302 | lseek64 NEW: LFS | |
7303 | makecontext NEW: Unix98 | |
7304 | mempcpy NEW: GNU ext. | |
7305 | mmap64 NEW: LFS | |
7306 | moncontrol REMOVED | |
7307 | modify_ldt NEW: kernel | |
7308 | nan NEW: ISO C 9x | |
7309 | nanf NEW: ISO C 9x | |
7310 | nanl NEW: ISO C 9x | |
7311 | nearbyint NEW: ISO C 9x | |
7312 | nearbyintf NEW: ISO C 9x | |
7313 | nearbyintl NEW: ISO C 9x | |
7314 | netname2host NEW: Secure RPC | |
7315 | netname2user NEW: Secure RPC | |
7316 | nexttoward NEW: ISO C 9x | |
7317 | nexttowardf NEW: ISO C 9x | |
7318 | nexttowardl NEW: ISO C 9x | |
7319 | nftw NEW: Unix98 | |
7320 | nftw64 NEW: LFS | |
7321 | open64 NEW: LFS | |
7322 | passwd2des NEW: Secure RPC | |
7323 | pow10 NEW: GNU ext. | |
7324 | pow10f NEW: GNU ext. | |
7325 | pow10l NEW: GNU ext. | |
7326 | pread NEW: Unix98 | |
7327 | pread64 NEW: LFS | |
7328 | printf_size NEW: GNU ext. | |
7329 | printf_size_info NEW: GNU ext. | |
7330 | profil_counter REMOVED | |
7331 | pthread_mutexattr_getkind_np REPLACED | |
7332 | pthread_mutexattr_setkind_np REPLACED | |
7333 | ptsname NEW: Unix98 PTY | |
7334 | ptsname_r NEW: Unix98 PTY | |
7335 | putc_locked REMOVED | |
7336 | putchar_locked REMOVED | |
7337 | putgrent NEW: GNU ext. | |
7338 | putmsg NEW: STREAMS | |
7339 | putpmsg NEW: STREAMS | |
7340 | pututxline NEW: Unix98 | |
7341 | pwrite NEW: Unix98 | |
7342 | pwrite64 NEW: LFS | |
7343 | readdir64 NEW: LFS | |
7344 | readdir64_r NEW: LFS | |
7345 | remquo NEW: ISO C 9x | |
7346 | remquof NEW: ISO C 9x | |
7347 | remquol NEW: ISO C 9x | |
7348 | round NEW: ISO C 9x | |
7349 | roundf NEW: ISO C 9x | |
7350 | roundl NEW: ISO C 9x | |
7351 | rtime NEW: GNU ext. | |
7352 | scalbln NEW: ISO C 9x | |
7353 | scalblnf NEW: ISO C 9x | |
7354 | scalblnl NEW: ISO C 9x | |
7355 | scandir64 NEW: LFS | |
7356 | sendfile NEW: kernel | |
7357 | setcontext NEW: Unix98 | |
7358 | setrlimit64 NEW: LFS | |
7359 | setutxent NEW: Unix98 | |
7360 | sighold NEW: Unix98 | |
7361 | sigignore NEW: Unix98 | |
7362 | sigqueue NEW: POSIX.1b | |
7363 | sigrelse NEW: Unix98 | |
7364 | sigset NEW: POSIX.1b | |
7365 | sigtimedwait NEW: POSIX.1b | |
7366 | sigwaitinfo NEW: POSIX.1b | |
7367 | sincos NEW: GNU ext. | |
7368 | sincosf NEW: GNU ext. | |
7369 | sincosl NEW: GNU ext. | |
7370 | statfs64 NEW: LFS | |
7371 | statvfs NEW: Unix98 | |
7372 | statvfs64 NEW: LFS | |
7373 | strcasestr NEW: GNU ext. | |
7374 | strtoimax NEW: ISO C 9x | |
7375 | strtoumax NEW: ISO C 9x | |
7376 | strverscmp NEW: GNU ext. | |
7377 | svcauthdes_stats NEW: Secure RPC | |
7378 | svcunix_create NEW: sunrpc etx | |
7379 | svcunixfd_create NEW: sunrpc ext | |
7380 | swapcontext NEW: Unix98 | |
7381 | tcgetsid NEW: Unix98 PTY | |
7382 | tdestroy NEW: GNU ext. | |
7383 | tgamma NEW: ISO C 9x | |
7384 | tgammaf NEW: ISO C 9x | |
7385 | tgammal NEW: ISO C 9x | |
7386 | tmpfile64 NEW: LFS | |
7387 | trunc NEW: ISO C 9x | |
7388 | truncate64 NEW: LFS | |
7389 | truncf NEW: ISO C 9x | |
7390 | truncl NEW: ISO C 9x | |
7391 | umount2 NEW: kernel | |
7392 | unlockpt NEW: Unix98 PTY | |
7393 | updwtmpx NEW: Unix98 | |
7394 | user2netname NEW: Secure RPC | |
7395 | utmpxname NEW: Unix98 | |
7396 | versionsort NEW: GNU ext. | |
7397 | versionsort64 NEW: GNU ext. | |
7398 | waitid NEW: Unix98 | |
7399 | wcscasecmp NEW: GNU ext. | |
7400 | wcsncasecmp NEW: GNU ext. | |
7401 | wcsnlen NEW: GNU ext. | |
7402 | wcstoimax NEW: ISO C 9x | |
7403 | wcstoll NEW: ISO C 9x | |
7404 | wcstoull NEW: ISO C 9x | |
7405 | wcstoumax NEW: ISO C 9x | |
7406 | wcswcs NEW: Unix98 | |
7407 | wordexp NEW: POSIX.2 | |
7408 | wordfree NEW: POSIX.2 | |
7409 | write_profiling REMOVED | |
7410 | xdecrypt NEW: Secure RPC | |
7411 | xdr_authdes_cred NEW: Secure RPC | |
7412 | xdr_authdes_verf NEW: Secure RPC | |
7413 | xdr_cryptkeyarg NEW: Secure RPC | |
7414 | xdr_cryptkeyarg2 NEW: Secure RPC | |
7415 | xdr_cryptkeyres NEW: Secure RPC | |
7416 | xdr_getcredres NEW: Secure RPC | |
7417 | xdr_key_netstarg NEW: Secure RPC | |
7418 | xdr_key_netstres NEW: Secure RPC | |
7419 | xdr_keybuf NEW: Secure RPC | |
7420 | xdr_keystatus NEW: Secure RPC | |
7421 | xdr_netnamestr NEW: Secure RPC | |
7422 | xdr_sizeof NEW: Secure RPC | |
7423 | xdr_unixcred NEW: sunrpc ext | |
7424 | xencrypt NEW: Secure RPC | |
7425 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
7426 | \f | |
7427 | Version 2.0.6 | |
7428 | ||
7429 | * more bug fixes | |
7430 | ||
7431 | \f | |
7432 | Version 2.0.5 | |
7433 | ||
7434 | * more bug fixes | |
7435 | ||
7436 | * inet_ntoa is thread-safe | |
7437 | ||
7438 | * updwtmp is moved from libutil to libc | |
7439 | ||
7440 | * rewrite of cbrt function | |
7441 | ||
7442 | * update of timezone data | |
7443 | \f | |
7444 | Version 2.0.4 | |
7445 | ||
7446 | * more bug fixes | |
7447 | \f | |
7448 | Version 2.0.3 | |
7449 | ||
7450 | * more bug fixes | |
7451 | \f | |
7452 | Version 2.0.2 | |
7453 | ||
7454 | * more bug fixes | |
7455 | ||
7456 | * add atoll function | |
7457 | ||
7458 | * fix complex problems in Berkeley DB code | |
7459 | ||
7460 | * fix math functions | |
7461 | \f | |
7462 | Version 2.0.1 | |
7463 | ||
7464 | * fixed lots of header problems (especially Linux/GNU specific) | |
7465 | ||
7466 | * dynamic loader preserves all registers | |
7467 | ||
7468 | * Roland McGrath provided support for handling of auxiliary objects in | |
7469 | the ELF dynamic loader. | |
7470 | ||
7471 | * support for parallel builds is improved | |
7472 | \f | |
7473 | Version 2.0 | |
7474 | ||
7475 | * GNU extensions are no longer declared by default. To enable them you | |
7476 | must define the macro `_GNU_SOURCE' in your program or compile with | |
7477 | `-D_GNU_SOURCE'. | |
7478 | ||
7479 | * The library has changed from using GNU ld symbol aliases to using weak | |
7480 | symbols where available. The ELF object file format supports weak | |
7481 | symbols; GNU ld also supports weak symbols in the a.out format. (There | |
7482 | is also now support for other GNU ld extensions in ELF. Use the | |
7483 | `--with-elf' option to configure to indicate you have ELF, and | |
7484 | `--with-gnu-ld' if using GNU ld.) This change resulted in the deletion | |
7485 | of many files which contained only symbol aliases, reducing the size of | |
7486 | the source and the compiled library; many other files were renamed to | |
7487 | less cryptic names previously occupied by the symbol alias files. | |
7488 | There is a new header file <elf.h> for programs which operate on | |
7489 | files in the ELF format. | |
7490 | ||
7491 | * Converted to Autoconf version 2, so `configure' has more options. | |
7492 | Run `configure --help' to see the details. | |
7493 | ||
7494 | * The library can now be configured to build profiling, highly-optimized | |
7495 | (but undebuggable), and/or shared libraries (ELF with GNU ld only). The | |
7496 | `--enable-profile', `--enable-omitfp', and `--enable-shared' options to | |
7497 | `configure' enable building these extra libraries. The shared library is | |
7498 | built by default when using both ELF and GNU ld. When shared libraries | |
7499 | are enabled, the new library `-ldl' is available for arbitrary run-time | |
7500 | loading of shared objects; its interface is defined in <dlfcn.h>. The | |
7501 | new header file <link.h> gives access to the internals of the run-time | |
7502 | dynamic linker, `ld.so'. The shell script `ldd' is similar to the | |
7503 | application of same name on other systems and it provides information | |
7504 | about dynamically linked binaries. | |
7505 | ||
7506 | * The C library now provides the run-time support code for profiling | |
7507 | executables compiled with `-pg'. Programs can control the profiling code | |
7508 | through the interface in <sys/gmon.h>. The `gmon.out' files written by | |
7509 | the GNU C library can be read only by GNU `gprof' (from GNU binutils); | |
7510 | the support for this file format was contributed by David Mosberger-Tang. | |
7511 | ||
7512 | * The math code has been replaced with a math library based on fdlibm from | |
7513 | Sun, and modified by JT Conklin and Ulrich Drepper with i387 support, by | |
7514 | Ian Taylor with `float' functions and by Ulrich Drepper with `long double' | |
7515 | functions. The math functions now reside in a separate library, so | |
7516 | programs using them will need to use `-lm' their linking commands. | |
7517 | ||
7518 | * John C. Bowman contributed optimized ix87 assembler inline functions. | |
7519 | ||
7520 | * Ulrich Drepper has contributed support for an `/etc/nsswitch.conf' | |
7521 | mechanism similar to that found in Solaris 2. This is now used for the | |
7522 | group, passwd, hosts, networks, services, protocols, rpc, ethers, | |
7523 | shadow, netgroup, publickey, and alias databases. The `nsswitch.conf' | |
7524 | file controls what services are used for each individual database. This | |
7525 | works by loading shared libraries with names specified in `nsswitch.conf', | |
7526 | so service modules can be changed or added at any time without even | |
7527 | relinking any program. Currently there are the file, db, and NIS based | |
7528 | NSS services available. | |
7529 | ||
7530 | * The new functions `strtoq' and `strtouq' parse integer values from | |
7531 | strings, like `strtol' and `strtoul', but they return `long long int' and | |
7532 | `unsigned long long int' values, respectively (64-bit quantities). | |
7533 | ||
7534 | * The new functions `strtof' and `strtold' parse floating-point values from | |
7535 | strings, like `strtod', but they return `float' and `long double' values, | |
7536 | respectively (on some machines `double' and `long double' are the same). | |
7537 | ||
7538 | * Ulrich Drepper has contributed new implementations of the floating-point | |
7539 | printing and reading code used in the `printf' family of functions and | |
7540 | `strtod', `strtof', and `strtold'. These new functions are perfectly | |
7541 | accurate, and much faster than the old ones. | |
7542 | ||
7543 | * The implementation of the POSIX locale model was completely rewritten by | |
7544 | Ulrich Drepper. This includes the new programs `localedef' and `locale' | |
7545 | to compile the POSIX locale definition. | |
7546 | ||
7547 | * The former dummy implementations of the strcoll and strxfrm function are | |
7548 | now replaced by fully functional code contributed by Ulrich Drepper. The | |
7549 | collation information comes from the POSIX locale definitions. | |
7550 | ||
7551 | * The new header <langinfo.h> defines an interface for accessing | |
7552 | various locale-dependent data (using the locale chosen with `setlocale'). | |
7553 | ||
7554 | * Ulrich Drepper has contributed a new suite of functions for operation on | |
7555 | wide-character and multibyte-character strings, in <wchar.h>; | |
7556 | and classification and case conversion of wide characters, in <wctype.h>. | |
7557 | These new functions are conforming to the ISO C, Amendment 1 specification. | |
7558 | ||
7559 | * There is now a second implementation of the standard I/O library available. | |
7560 | It comes from GNU libg++ as was written by Per Bothner, heavily modified | |
7561 | by Hongjiu Lu and made thread safe by Ulrich Drepper. | |
7562 | ||
7563 | * You can now use positional parameter specifications in format strings | |
7564 | for the `printf' and `scanf' families of functions. For example, | |
7565 | `printf ("Number %2$d, Mr %1$s\n", "Jones", 6);'' prints | |
7566 | ``Number 6, Mr Jones''. This is mainly useful when providing different | |
7567 | format strings for different languages, whose grammars may dictate | |
7568 | different orderings of the values being printed. To support this | |
7569 | feature, the interface for `register_printf_handler' has changed; see | |
7570 | the header file <printf.h> for details. | |
7571 | ||
7572 | * The `printf' and `scanf' families of functions now understand a new | |
7573 | formatting flag for numeric conversions: the ' flag (e.g. %'d or %'f) says | |
7574 | to group numbers as indicated by the locale; for `scanf' and friends, this | |
7575 | says to accept as valid only a number with all the proper grouping | |
7576 | separators in the right places. In the default "C" locale, numbers are | |
7577 | not grouped; but locales for specific countries will define the usual | |
7578 | conventions (i.e. separate thousands with `,' in the US locale). | |
7579 | ||
7580 | * The pgrp functions have been regularized, slightly incompatibly but much | |
7581 | less confusingly. The core functions are now `getpgid' and `setpgid', | |
7582 | which take arguments for the PID to operate on; the POSIX.1 `getpgrp' (no | |
7583 | argument) and BSD `setpgrp' (identical to `setpgid') functions are | |
7584 | provided for compatibility. There is no longer an incompatible `getpgrp' | |
7585 | with an argument declared under _BSD_SOURCE; no BSD code uses it. | |
7586 | ||
7587 | * The new header file <fts.h> and suite of functions simplify programs that | |
7588 | operate on directory trees. This code comes from 4.4 BSD. | |
7589 | ||
7590 | * The resolver code has been updated from the BIND 4.9.5-P1 release. | |
7591 | Parts of the code were heavily modified by Ulrich Drepper to fit in the | |
7592 | NSS scheme used in glibc. | |
7593 | ||
7594 | * There is a new malloc debugging hook `__memalign_hook'. | |
7595 | ||
7596 | * There are new typedefs `ushort' for `unsigned short int' and `uint' for | |
7597 | `unsigned int' in <sys/types.h>. These are for compatibility only and | |
7598 | their use is discouraged. | |
7599 | ||
7600 | * The `-lmcheck' library to enable standard malloc debugging hooks is now | |
7601 | done differently, so that it works even without GNU ld. | |
7602 | ||
7603 | * New function `euidaccess' checks allowed access to a file like `access', | |
7604 | but using the effective IDs instead of the real IDs. | |
7605 | ||
7606 | * The time zone data files have been updated for the latest and greatest | |
7607 | local time conventions of the countries of the world. | |
7608 | ||
7609 | * The new function `dirfd' extracts the file descriptor used by a DIR stream; | |
7610 | see <dirent.h>. | |
7611 | ||
7612 | * The new functions `ecvt', `fcvt', and `gcvt' provide an obsolete interface | |
7613 | for formatting floating-point numbers. They are provided only for | |
7614 | compatibility; new programs should use `sprintf' instead. There are | |
7615 | also equivalent function for the `long double' floating-point type and | |
7616 | all functions also exist in a reentrant form. | |
7617 | ||
7618 | * The new auxiliary library `-lutil' from 4.4 BSD contains various | |
7619 | functions for maintaining the login-record files (primarily of use to | |
7620 | system programs such as `login'), and convenient functions for | |
7621 | allocating and initializing a pseudo-terminal (pty) device. | |
7622 | ||
7623 | * Ulrich Drepper has contributed new support for System V style | |
7624 | shared memory and IPC on systems that support it. | |
7625 | ||
7626 | * Ulrich Drepper has contributed several miscellaneous new functions found | |
7627 | in System V: The `hsearch' family of functions provide an effective | |
7628 | implementation of hash tables; `a64l' and `l64a' provide a very simple | |
7629 | binary to ASCII mapping; `drand48' and friends provide a 48-bit random | |
7630 | number generator. | |
7631 | ||
7632 | * Ulrich Drepper has contributed new reentrant counterparts for the | |
7633 | `random' and `hsearch' families of functions; `random_r', `hsearch_r', etc. | |
7634 | ||
7635 | * Ulrich Drepper has contributed new, highly-optimized versions of several | |
7636 | string functions for the i486/Pentium family of processors. | |
7637 | ||
7638 | * Ulrich Drepper has updated the Linux-specific code, based largely | |
7639 | on work done in Hongjiu Lu's version of GNU libc for Linux. | |
7640 | The GNU library now supports Linux versions 2.0.10 and later, | |
7641 | using the ELF object file format (i[3456]86-*-linux). | |
7642 | ||
7643 | * Andreas Schwab has ported the C library to Linux/m68k (m68k-*-linux). | |
7644 | ||
7645 | * David Mosberger-Tang and Richard Henderson have ported the C library | |
7646 | to Linux/Alpha (alpha-*-linux). Richard Henderson contributed the | |
7647 | dynamic linking support for ELF/Alpha. | |
7648 | ||
7649 | * Richard Henderson contributed several Alpha optimized assembler function | |
7650 | for arithmetic and string handling. | |
7651 | ||
7652 | * Ulrich Drepper has contributed a new set of message catalog functions to | |
7653 | support multiple languages using the <libintl.h> interface, for use with | |
7654 | his new package GNU gettext. Translation volunteers have contributed | |
7655 | catalogs of the library's messages in Spanish, German, and Korean. | |
7656 | ||
7657 | * For compatibility with XPG4, Ulrich Drepper has contributed the `gencat' | |
7658 | program and the `catgets' function for reading the catalog files it | |
7659 | creates. (The <libintl.h> interface is preferred; we include the | |
7660 | <nl_types.h> interface using `catgets' only for source compatibility with | |
7661 | programs already written to use it.) | |
7662 | ||
7663 | * New header file <values.h> gives SVID-compatible names for <limits.h> | |
7664 | constants. | |
7665 | ||
7666 | * Various new macros, declarations, and small header files for compatibility | |
7667 | with 4.4 BSD. | |
7668 | ||
7669 | * New function `group_member' is a convenient way to check if a process has | |
7670 | a given effective group ID. | |
7671 | ||
7672 | * When using GCC 2.7 and later, the socket functions are now declared in a | |
7673 | special way so that passing an argument of type `struct sockaddr_in *', | |
7674 | `struct sockaddr_ns *', or `struct sockaddr_un *' instead of the generic | |
7675 | `struct sockaddr *' type, does not generate a type-clash warning. | |
7676 | ||
7677 | * New function `error' declared in header file <error.h> is a convenient | |
7678 | function for printing error messages and optionally exiting; this is the | |
7679 | canonical function used in GNU programs. The new functions `err', `warn', | |
7680 | and friends in header file <err.h> are the canonical 4.4 BSD interface for | |
7681 | doing the same thing. | |
7682 | ||
7683 | * The <glob.h> interface has several new flags from 4.4 BSD that extend the | |
7684 | POSIX.2 `glob' function to do ~ and {...} expansion. | |
7685 | ||
7686 | * New function `unsetenv' complements `setenv' for compatibility with 4.4 BSD. | |
7687 | `clearenv' which is used in POSIX.9 is also available. | |
7688 | ||
7689 | * New function `getsid' returns session ID number on systems that support it. | |
7690 | ||
7691 | * We have incorporated the 4.4 BSD `db' library (version 1.85). New header | |
7692 | files <db.h> and <mpool.h> provide a rich set of functions for several | |
7693 | types of simple databases stored in memory and in files, and <ndbm.h> is | |
7694 | an old `ndbm'-compatible interface using the `db' functions. Link with | |
7695 | `-ldb' to get these functions. | |
7696 | ||
7697 | * New macro `strdupa' copies a string like `strdup', but uses local stack | |
7698 | space from `alloca' instead of dynamic heap space from `malloc'. | |
7699 | ||
7700 | * New function `strnlen' is like `strlen' but searches only a given maximum | |
7701 | number of characters for the null terminator. `stpncpy', `strndup' and | |
7702 | `strndupa' are similar variants for the `stpcpy', `strdup' and `strdupa' | |
7703 | function. | |
7704 | ||
7705 | * New function `statfs' in header <sys/statfs.h>. | |
7706 | ||
7707 | * The new <argz.h> and <envz.h> interfaces contributed by Miles Bader | |
7708 | provide convenient functions for operating on blocks of null-terminated | |
7709 | strings. | |
7710 | ||
7711 | * A new suite of functions in <utmp.h> handle all the details of reading | |
7712 | and writing the utmp file. | |
7713 | ||
7714 | * An implementation of the NIS/YP(tm) based NSS service was contributed by | |
7715 | Thorsten Kukuk. | |
7716 | ||
7717 | * Paul Eggert and Ulrich Drepper modified the `strftime' function to be | |
7718 | completely POSIX compliant and also implemented the extended functionality | |
7719 | to handle alternate digit representation and alternate era date formats. | |
7720 | ||
7721 | * Ulrich Drepper provided an implementation of the `strptime' function | |
7722 | defined in XPG4.2 which transforms a string into a `struct tm' value. | |
7723 | ||
7724 | * Paul Eggert provided the tzselect shell script as part of the timezone | |
7725 | code. The shell script makes it easy to select the correct timezone | |
7726 | specification. | |
7727 | ||
7728 | * The implementation of the malloc family of functions is completely replaced | |
7729 | by a new implementation by Doug Lea with many improvements by Wolfram Gloger. | |
7730 | The implementation uses the mmap function (if available) and it is | |
7731 | optimized for the use in multi threaded programs. | |
7732 | ||
7733 | * Ulrich Drepper contributed a MD5 "encryption" for the crypt family of | |
7734 | functions. This new functionality is usable by specifying a special | |
7735 | salt string and it is compatible with implementation on *BSD systems. | |
7736 | ||
7737 | * Lots of functions from the XPG4.2 standard were added by Ulrich Drepper: | |
7738 | `getsubopt' to handle second level command line options, `bsd_signal' | |
7739 | to access BSD style `signal' functionality, the obsolete `regexp' style | |
7740 | expression matcher. | |
7741 | ||
7742 | * the `lchown' function is available on system which support this | |
7743 | functionality. | |
7744 | ||
7745 | * The implementation of the shadow password handling function was contributed | |
7746 | by Ulrich Drepper. | |
7747 | ||
7748 | * David Mosberger-Tang changed the SunRPC implementation to be 64bit safe. | |
7749 | ||
7750 | * POSIX.1g support was added. The <sys/select.h> header is available, | |
7751 | `isfdtype' and `pselect' are implemented. Craig Metz contributed an | |
7752 | implementation of `getaddrinfo'. | |
7753 | \f | |
7754 | Version 1.09 | |
7755 | ||
7756 | * For cross-compilation you should now set `BUILD_CC' instead of `HOST_CC'. | |
7757 | ||
7758 | * New header file <fstab.h> and new functions `getfsspec', `getfsent' and | |
7759 | friends, for parsing /etc/fstab. This code comes from 4.4 BSD. | |
7760 | ||
7761 | * The new function `daemon' from 4.4 BSD is useful for server programs that | |
7762 | want to put themselves in the background. | |
7763 | ||
7764 | * Joel Sherrill has contributed support for several standalone boards that | |
7765 | run without an operating system. | |
7766 | ||
7767 | * `printf', `scanf' and friends now accept a `q' type modifier for long | |
7768 | long int as well as `ll'. Formats using these might be `%qu' or `%lld'. | |
7769 | ||
7770 | * All of the code taken from BSD (notably most of the math and networking | |
7771 | routines) has been updated from the BSD 4.4-Lite release. | |
7772 | ||
7773 | * The resolver code has been updated from the BIND-4.9.3-BETA9 release. | |
7774 | ||
7775 | * The new functions `getdomainname' and `setdomainname' fetch or change the | |
7776 | YP/NIS domain name. These are system calls which exist on systems which | |
7777 | have YP (aka NIS). | |
7778 | ||
7779 | * The time zone data files have been updated for the latest international | |
7780 | conventions. | |
7781 | ||
7782 | * The SunRPC programs `portmap' and `rpcinfo' are now installed in | |
7783 | $(sbindir) (usually /usr/local/sbin) instead of $(bindir). | |
7784 | \f | |
7785 | Version 1.08 | |
7786 | ||
7787 | * The C library now includes support for Sun RPC, from Sun's free | |
7788 | RPCSRC-4.0 distribution. The `portmap', `rpcinfo', and `rpcgen' programs | |
7789 | are included. (There is still no support for YP.) | |
7790 | ||
7791 | * Tom Quinn has contributed a port of the C library to SGI machines running | |
7792 | Irix 4 (mips-sgi-irix4). | |
7793 | ||
7794 | * The new `lockf' function is a simplified interface to the locking | |
7795 | facilities of `fcntl', included for compatibility. | |
7796 | ||
7797 | * New time functions `timegm', `timelocal', and `dysize' for compatibility. | |
7798 | ||
7799 | * New header file <sys/timeb.h> and new function `ftime' for compatibility. | |
7800 | ||
7801 | * New header files <poll.h> and <sys/poll.h> and new function `poll' for | |
7802 | compatibility. | |
7803 | ||
7804 | * The error message printed by `assert' for a failed assertion now includes | |
7805 | the name of the program (if using GNU ld) and the name of the calling | |
7806 | function (with versions of GCC that support this). | |
7807 | ||
7808 | * The `psignal' function is now declared in <signal.h>, not <stdio.h>. | |
7809 | ||
7810 | * The library now includes the <sys/mman.h> header file and memory | |
7811 | management functions `mmap', `munmap', `mprotect', `msync', and | |
7812 | `madvise', on systems that support those facilities. | |
7813 | ||
7814 | * The interface for `mcheck' has changed slightly: the function called to | |
7815 | abort the program when an allocation inconsistency is detected now takes | |
7816 | an argument that indicates the type of failure. The new function | |
7817 | `mprobe' lets you request a consistency check for a particular block at | |
7818 | any time (checks are normally done only when you call `free' or `realloc' | |
7819 | on a block). | |
7820 | ||
7821 | * It is now possible to easily cross-compile the C library, building on one | |
7822 | system a library to run on another machine and/or operating system. All | |
7823 | you need to do is set the variable `HOST_CC' in `configparms' to the | |
7824 | native compiler for programs to run on the machine you are building on (a | |
7825 | few generator programs are used on Unix systems); set `CC' to the | |
7826 | cross-compiler. | |
7827 | ||
7828 | * The new function `fexecve' (only implemented on the GNU system) executes | |
7829 | a program file given a file descriptor already open on the file. | |
7830 | \f | |
7831 | Version 1.07 | |
7832 | ||
7833 | * Brendan Kehoe has contributed most of a port to the DEC Alpha | |
7834 | running OSF/1 (alpha-dec-osf1). He says it is 75% complete. | |
7835 | ||
7836 | * You can set the variable `libprefix' in `configparms' to specify a prefix | |
7837 | to be prepended to installed library files; this makes it easy to install | |
7838 | the GNU C library to be linked as `-lgnuc' or whatever. | |
7839 | ||
7840 | * The new `stpncpy' is a cross between `stpcpy' and `strncpy': It | |
7841 | copies a limited number of characters from a string, and returns the | |
7842 | address of the last character written. | |
7843 | ||
7844 | * You no longer need to check for whether the installed `stddef.h' is | |
7845 | compatible with the GNU C library. configure now checks for you. | |
7846 | ||
7847 | * You can now define a per-stream `fileno' function to convert the | |
7848 | stream's cookie into an integral file descriptor. | |
7849 | ||
7850 | * ``malloc (0)'' no longer returns a null pointer. Instead, it | |
7851 | allocates zero bytes of storage, and returns a unique pointer which | |
7852 | you can pass to `realloc' or `free'. The behavior is undefined if | |
7853 | you dereference this pointer. | |
7854 | ||
7855 | * The C library now runs on Sony NEWS m68k machines running either | |
7856 | NewsOS 3 or NewsOS 4. | |
7857 | ||
7858 | * The new `syscall' function is a system-dependent primitive function | |
7859 | for invoking system calls. It has the canonical behavior on Unix | |
7860 | systems, including unreliable return values for some calls (such as | |
7861 | `pipe', `fork' and `getppid'). | |
7862 | ||
7863 | * The error code `EWOULDBLOCK' is now obsolete; it is always defined | |
7864 | to `EAGAIN', which is the preferred name. On systems whose kernels | |
7865 | use two distinct codes, the C library now translates EWOULDBLOCK to | |
7866 | EAGAIN in every system call function. | |
7867 | \f | |
7868 | Version 1.06 | |
7869 | ||
7870 | * The GNU C Library Reference Manual is now distributed with the library. | |
7871 | `make dvi' will produce a DVI file of the printed manual. | |
7872 | `make info' will produce Info files that you can read on line using C-h i | |
7873 | in Emacs or the `info' program. | |
7874 | Please send comments on the manual to bug-glibc-manual@gnu.org. | |
7875 | ||
7876 | * The library now supports SVR4 on i386s (i386-unknown-sysv4). | |
7877 | ||
7878 | * Brendan Kehoe has contributed a port to Sun SPARCs running Solaris 2. | |
7879 | ||
7880 | * Jason Merrill has contributed a port to the Sequent Symmetry running | |
7881 | Dynix version 3 (i386-sequent-dynix). | |
7882 | ||
7883 | * The library has been ported to i386s running SCO 3.2.4 (also known as SCO | |
7884 | ODT 2.0; i386-unknown-sco3.2.4) or SCO 3.2 (i386-unknown-sco3.2). | |
7885 | ||
7886 | * New function `memory_warnings' lets you arrange to get warnings when | |
7887 | malloc is running out of memory to allocate, like Emacs gives you. | |
7888 | ||
7889 | * The C library now contains the relocating allocator used in Emacs 19 for | |
7890 | its editing buffers. This allocator (ralloc) minimizes allocation | |
7891 | overhead and fragmentation by moving allocated regions around whenever it | |
7892 | needs to. You always refer to a ralloc'd region with a "handle" (a | |
7893 | pointer to a pointer--an object of type `void **'). | |
7894 | ||
7895 | * There is a new `printf' format: `%m' gives you the string corresponding | |
7896 | to the error code in `errno'. | |
7897 | ||
7898 | * In `scanf' formats, you can now use `%as' or `%a[' to do the normal `%s' | |
7899 | or `%[' conversion, but instead of filling in a fixed-sized buffer you | |
7900 | pass, the `a' modifier says to fill in a `char **' you pass with a | |
7901 | malloc'd string. | |
7902 | ||
7903 | * The `fnmatch' function supports the new flag bits `FNM_LEADING_DIR' and | |
7904 | `FNM_CASEFOLD'. `FNM_LEADING_DIR' lets a pattern like `foo*' match a | |
7905 | name like `foo/bar'. `FNM_CASEFOLD' says to ignore case in matching. | |
7906 | ||
7907 | * `mkstemp' is a traditional Unix function to atomically create and open a | |
7908 | uniquely-named temporary file. | |
7909 | \f | |
7910 | Version 1.05 | |
7911 | ||
7912 | * The standard location for the file that says what the local timezone is | |
7913 | has changed again. It is now `/usr/local/etc/localtime' (or more | |
7914 | precisely, `${prefix}/etc/localtime') rather than `/etc/localtime'. | |
7915 | ||
7916 | * The distribution no longer contains any files with names longer than 14 | |
7917 | characters. | |
7918 | ||
7919 | * `struct ttyent' has two new flag bits: TTY_TRUSTED and TTY_CONSOLE. | |
7920 | These are set by the new `trusted' and `console' keywords in `/etc/ttys'. | |
7921 | ||
7922 | * New functions `ttyslot' and `syslog' from 4.4 BSD. | |
7923 | \f | |
7924 | Version 1.04 | |
7925 | ||
7926 | * The configuration process has changed quite a bit. The `configure' | |
7927 | script is now used just like the configuration scripts for other GNU | |
7928 | packages. The `sysdeps' directory hierarchy is much rearranged. | |
7929 | The file `INSTALL' explains the new scheme in detail. | |
7930 | ||
7931 | * The header files no longer need to be processed into ANSI C and | |
7932 | traditional C versions. There is just one set of files to install, and | |
7933 | it will work with ANSI or old C compilers (including `gcc -traditional'). | |
7934 | ||
7935 | * Brendan Kehoe and Ian Lance Taylor have ported the library to the | |
7936 | MIPS DECStation running Ultrix 4. | |
7937 | ||
7938 | * The Sun 4 startup code (crt0) can now properly load SunOS 4 shared libraries. | |
7939 | Tom Quinn contributed the initial code. The GNU C library can NOT yet be | |
7940 | made itself into a shared library. | |
7941 | ||
7942 | * Yet further improved support for the i386, running 4.3 BSD-like systems | |
7943 | (such as Mach 3 with the Unix single-server), or System V. | |
7944 | ||
7945 | * New function `strncasecmp' to do case-insensitive string comparison | |
7946 | with limited length. | |
7947 | ||
7948 | * New function `strsep' is a reentrant alternative to `strtok'. | |
7949 | ||
7950 | * New functions `scandir' and `alphasort' for searching directories. | |
7951 | ||
7952 | * New function `setenv' is a better interface to `putenv'. | |
7953 | ||
7954 | * Ian Lance Taylor has contributed an implementation of the SVID `ftw' | |
7955 | function for traversing a directory tree. | |
7956 | ||
7957 | * The GNU obstack package is now also part of the C library. | |
7958 | The new function `open_obstack_stream' creates a stdio stream that | |
7959 | writes onto an obstack; `obstack_printf' and `obstack_vprintf' do | |
7960 | formatted output directly to an obstack. | |
7961 | ||
7962 | * Miscellaneous new functions: reboot, nice, sigaltstack (4.4 BSD only), | |
7963 | cfmakeraw, getusershell, getpass, swab, getttyent, seteuid, setegid. | |
7964 | ||
7965 | * `FNM_FILE_NAME' is another name for `FNM_PATHNAME', used with `fnmatch'. | |
7966 | ||
7967 | * The new functions `strfry' and `memfrob' do mysterious and wonderful | |
7968 | things to your strings. | |
7969 | ||
7970 | * There are some new test programs: test-fseek, testmb, and testrand. | |
7971 | ||
7972 | * Some work has been done to begin porting the library to 4.4 BSD and Linux. | |
7973 | These ports are not finished, but are a good starting place for really | |
7974 | supporting those systems. | |
7975 | ||
7976 | * `/etc/localtime' is now the standard location for the file that says what | |
7977 | the local timezone is, rather than `/usr/local/lib/zoneinfo/localtime'. | |
7978 | This follows the general principle that `/etc' is the place for all local | |
7979 | configuration files. | |
7980 | ||
7981 | * The C library header files now use `extern "C"' when used by the C++ | |
7982 | compiler, so the C library should now work with C++ code. | |
7983 | ||
7984 | * The header file <bstring.h> is gone. <string.h> now declares bcopy, | |
7985 | bcmp, bzero, and ffs. (Update: nowadays these functions are declared | |
7986 | in <strings.h>.) | |
7987 | ||
7988 | * Mike Haertel (of GNU e?grep and malloc fame) has written a new sorting | |
7989 | function which uses the `merge sort' algorithm, and is said to be | |
7990 | significantly faster than the old GNU `qsort' function. Merge sort is | |
7991 | now the standard `qsort' function. The new algorithm can require a lot | |
7992 | of temporary storage; so, the old sorting function is called when the | |
7993 | required storage is not available. | |
7994 | ||
7995 | * The C library now includes Michael Glad's Ultra Fast Crypt, which | |
7996 | provides the Unix `crypt' function, plus some other entry points. | |
7997 | ||
7998 | * The code and header files taken from 4.4 BSD have been updated with the | |
7999 | latest files released from Berkeley. | |
8000 | \f | |
8001 | ---------------------------------------------------------------------- | |
8002 | Copying conditions: | |
8003 | ||
8004 | Permission is granted to anyone to make or distribute verbatim copies | |
8005 | of this document as received, in any medium, provided that the | |
8006 | copyright notice and this permission notice are preserved, | |
8007 | thus giving the recipient permission to redistribute in turn. | |
8008 | ||
8009 | Permission is granted to distribute modified versions | |
8010 | of this document, or of portions of it, | |
8011 | under the above conditions, provided also that they | |
8012 | carry prominent notices stating who last changed them. | |
8013 | \f | |
8014 | Local variables: | |
8015 | version-control: never | |
8016 | fill-column: 72 | |
8017 | End: |