1 ==================== Changes in man-pages-4.12 ====================
3 Released: 2017-07-13, London
9 The following people contributed patches/fixes or (noted in brackets
10 in the changelog below) reports, notes, and ideas that have been
11 incorporated in changes in this release:
13 Alex Henrie <alexhenrie24@gmail.com>
14 Andi Kleen <andi-bz.firstfloor.org>
15 Arjun Shankar <arjun.is@lostca.se>
16 Brad Bendily <bendily@gmail.com>
17 Cameron Wright <wright18@uw.edu>
18 Carlos O'Donell <carlos@redhat.com>
19 Darrick J. Wong <darrick.wong@oracle.com>
20 David Lewis <davlewi3@cisco.com>
21 DJ Delorie <dj@delorie.com>
22 Douglas Caetano dos Santos <douglascs@taghos.com.br>
23 Dr. Tobias Quathamer <toddy@debian.org>
24 Eric Biggers <ebiggers@google.com>
25 Ferdinand Thiessen <f.thiessen@gmx.de>
26 G. Branden Robinson <g.branden.robinson@gmail.com>
27 Heinrich Schuchardt <xypron.glpk@gmx.de>
28 Henry Bent <henry.r.bent@gmail.com>
29 Jakub Wilk <jwilk@jwilk.net>
30 Janne Snabb <snabb@epipe.com>
31 Joe Brown <joe.brown.89@outlook.com>
32 Jorge Nerin <jnerin@gmail.com>
33 Kirill Tkhai <ktkhai@virtuozzo.com>
34 lilydjwg <lilydjwg@gmail.com>
35 Long Wang <w@laoqinren.net>
36 Michael Kerrisk <mtk.manpages@gmail.com>
37 Mike Frysinger <vapier@chromium.org>
38 Nadav Har'El <nyh@math.technion.ac.il>
39 NeilBrown <neilb@suse.com>
40 Pavel Tikhomirov <ptikhomirov@virtuozzo.com>
41 Quentin Rameau <quinq@fifth.space>
42 Ruben Kerkhof <ruben@rubenkerkhof.com>
43 sulit <sulitsrc@foxmail.com>
44 石井大貴 <hiroki.ishii@miraclelinux.com>
46 Apologies if I missed anyone!
49 Newly documented interfaces in existing pages
50 ---------------------------------------------
53 Kirill Tkhai [Michael Kerrisk]
54 Document the /proc/[pid]/ns/pid_for_children file
57 Changes to individual pages
58 ---------------------------
62 'objdump -p prog | grep NEEDED' doesn't give quite same info as 'ldd'
66 Put fchmod() feature test macro requirements in a more readable format
68 Note glibc 2.24 feature test macro requirements changes for fchmod()
72 When file owner or group is changed, file capabilities are cleared
74 Changes to file owner by root also clear set-UID and set-GID bits
78 Update BUGS to reflect fact that PID caching was removed in glibc 2.25
82 Clarify semantics of returned 'data' field
83 The returned 'data' is the 'data' most recently set via
87 Michael Kerrisk [Nadav Har'El, Andi Kleen]
88 SYNOPSIS: fix return type of get_mempolicy()
89 See https://bugzilla.kernel.org/show_bug.cgi?id=97051
92 Carlos O'Donell, Michael Kerrisk
93 Note that PID caching is removed as of glibc 2.25
94 Since glibc 2.25 the PID cache is removed.
96 Rationale given in the release notes:
97 https://sourceware.org/glibc/wiki/Release/2.25#pid_cache_removal
101 SEE ALSO: add ioctl_getfsmap(2)
105 Fix ordering of sections
111 Document the GETFSMAP ioctl
112 Document the new GETFSMAP ioctl that returns the physical layout of a
113 (disk-based) filesystem.
115 ERRORS: order alphabetically
119 Remove bogus text re POSIX_MADV_NOREUSE
120 There is a POSIX_FADV_NOREUSE for posix_fadvise(),
121 but no POSIX_MADV_NOREUSE for any API in POSIX.
125 Add ENOSYS error for 'nohz_full' CPU setting
129 Revise description of MS_REMOUNT | MS_BIND
130 MS_REMOUNT|MS_BIND affects all per-mount-point
131 flag. MS_RDONLY is only special because it,
132 uniquely, is both a per-mount-point flag *and* a
135 So the sections of per-mount-point flags and
136 MS_REMOUNT can usefully be clarified.
140 Note some further advantages of the *at() APIs
144 SEE ALSO: add tee(2) and vmsplice(2)
148 glibc 2.26 adds library support for preadv2() and pwritev2()
152 Mention cpuset cgroups as a cause of EINVAL error
156 Expand SECCOMP_RET_KILL documentation
160 Note feature test macro requirements for 'si_code' constants
162 Add a subheading for the description of 'si_code'
164 TRAP_BRANCH and TRAP_HWBKPT are present only on IA64
168 Note that specifying SS_ONSTACK in ss.ss_flags decreases portability
169 In the Illumos source (which presumably mirrors its Solaris
170 ancestry), there is this check in the sigaltstack()
173 if (ss.ss_flags & ~SS_DISABLE)
174 return (set_errno(EINVAL));
176 And in the FreeBSD source we find similar:
178 if ((ss->ss_flags & ~SS_DISABLE) != 0)
181 Note buggy addition of ss.ss_flags==SS_ONSTACK
182 Note buggy addition of ss.ss_flags==SS_ONSTACK as a synonym
183 for ss_flags==0. No other implemenation does this, AFAIK.
184 And it was not needed :-(.
186 Specifying 'ss' returns the current settings without changing them
188 Give 'oss' argument a more meaningful name: 'old_ss'
190 Some minor reworking of the text
192 ERRORS: update description of EINVAL error
198 SEE ALSO: add pipe(7)
202 ERRORS: split EINVAL error cases
204 ERRORS: add EINVAL for case where both descriptors refer to same pipe
208 Document the CONFIG_POSIX_TIMERS option added in Linux 4.10
212 Note glibc 2.26 changes to feature test macro requirements for waitid()
218 Remove C89 designation.
219 See https://bugzilla.kernel.org/show_bug.cgi?id=196319
221 [mtk: confirmed from review of draft of C90 standard]
225 Note feature test macro requirements changes for glibc 2.26
229 dl_iterate_phdr() shows the order in which objects were loaded
230 dl_iterate_phdr() tells us not just which objects are
231 loaded, but also the order in which they are loaded
232 (the "link-map order"). Since the order is relevant for
233 understanding symbol resolution, give the reader this clue.
235 Expand the code example, and show sample output
237 List values for the 'p_type' field
241 _GNU_SOURCE is needed to get RTLD_DEFAULT and RTLD_NEXT definitions
245 Note glibc 2.24 feature test macro requirement changes
249 Rework RETURN VALUE description to add more detail
251 Add an errors section
253 Largely rewrite the description of _PC_CHOWN_RESTRICTED
255 Rewrite description of _PC_PIPE_BUF
256 The existing description was not accurate, and lacked details.
260 BUGS: document a probable glibc regression in FTW_SLN case
261 See https://bugzilla.redhat.com/show_bug.cgi?id=1422736
262 and http://austingroupbugs.net/view.php?id=1121.
266 Fix _POSIX_C_SOURCE value for getaddrinfo()
267 The correct _POSIX_C_SOURCE value is 200112L, not 201112L in features.h.
271 Exemplar structure should use 'ucontext_t'.
272 In glibc we just finished a round of purging 'struct ucontext'
273 which is not in the POSIX reserved namespace of *_t tags. This has
274 some consequences to applications using the non-standard struct
276 https://sourceware.org/glibc/wiki/Release/2.26#Removal_of_.27struct_ucontext.27
277 but it also fixes a namespace conformance issue which is always a
278 longterm pain for large portable programs.
280 It was noted by Peter Maydell
281 (https://sourceware.org/bugzilla/show_bug.cgi?id=21457) that the
282 linux man pages still had references to 'struct ucontext' but only
283 in the form of an exemplar structure.
285 The patch fixes the exemplar to match what is in glibc
286 and therefore it won't ever suggest users can use 'struct
291 Note glibc 2.22 changes for feature test macro requirements
296 Ferdinand Thiessen [Michael Kerrisk]
297 Update feature test macro-requirements for glibc 2.24
300 Douglas Caetano dos Santos
301 Add ENODEV error for if_nametoindex()
302 Checked against glibc and kernel code: when given name isn't
303 a valid interface, kernel returns ENODEV and glibc just
308 Document the reallocarray() added in glibc 2.26
312 Note feature test macro requirements for nl_langinfo_l()
316 Remove paragraph about POSIX_FADV_NOREUSE
317 POSIX_FADV_NOREUSE is documented for posix_fadvise, and a
318 corresponding POSIX_MADV_NOREUSE flag is not specified by POSIX.
320 Thanks to Marc Lehmann <debian-reportbug@plan9.de>
321 See https://bugs.debian.org/865699
324 Michael Kerrisk [Arjun Shankar]
325 Since glibc 2.26, ptsname_r() no longer gives EINVAL for buf==NULL
329 Note glibc 2.24 feature test macro requirement changes for rand_r()
333 Add basic notes on 'op' argument of res_nmkquery() and res_mkquery()
337 Note glibc 2.26 changes to feature test macro requirements
341 Note glibc 2.26 feature test macro changes
346 The prefix 0x may be capitalized as 0X.
348 See ISO/IEC 9899:1999.
351 Michael Kerrisk [Pavel Tikhomirov]
352 Rework RETURN VALUE description to add more detail
353 Make the discussion clearer, and add a few details.
354 Also, fix the problem report from Pavel Tikhomirov
355 who noted that the man page falsely said that errno
356 is not changed on a successful return.
358 Addresses https://bugzilla.kernel.org/show_bug.cgi?id=195955
366 Fix error in feature test macro requirements
368 Note feature test macro requirements changes in glibc 2.24
370 Clarify details of use of <stdlib.h> file
374 Note glibc 2.24 feature test macro requirement changes
378 SEE ALSO: add dl_iterate_phdr(3)
382 Clarify group merge rules
383 This minor patch clarifies when merging is not done,
384 and how duplicate entries are merged.
386 Found while extending glibc's NSS test suite, discussed with glibc
387 team and Stephen Gallagher.
391 Document that 'iowait' field of /proc/stat is unreliable
392 Text taken from Chao Fan's kernel commit 9c240d757658a3ae996.
395 Michael Kerrisk [Jorge Nerin]
396 SEE ALSO: add some references to relevant kernel source files
400 SEE ALSO: add memfd_create(2), mmap(2), shm_open(3)
404 Clarify the effect on process capabilities when UID 0 does execve(2)
406 Note effect on capabilities when a process with UID != 0 does execve(2)
407 Michael Kerrisk [David Lewis]
408 Fix reversed descriptions of CAP_MAC_OVERRIDE and CAP_MAC_ADMIN
410 SEE ALSO: add filecap(8), netcap(8), pscap(8)
414 Add some further explanation of the example shell session
416 Fix a bug in shell session example
418 Some minor wording fixes and clarifications
422 Note glibc 2.24 feature test macro changes for S_IFSOCK and S_ISSOCK()
426 Undocument "URL" macro in man(7) in favor .UR+.UE
430 Mention /proc/[pid]/ns/pid_for_children
434 SEE ALSO: add tee(2) and vmsplice(2)
438 Mention signal.h header file
441 Michael Kerrisk [lilydjwg]
442 Since Linux 3.8, read(2) on an inotify FD is restartable with SA_RESTART
443 See https://bugzilla.kernel.org/show_bug.cgi?id=195711
445 read() from an inotify FD is no longer interrupted by a stop signal
446 (Change was in Linux 3.8.)
450 Document value '2' for tcp_timestamps
451 Since Linux 4.10, the value '2' is meaningful for tcp_timestamps
453 Change default value of tcp_frto
454 The default changed in c96fd3d461fa495400df24be3b3b66f0e0b152f9
459 Greatly expand the explanation of LD_DYNAMIC_WEAK
461 Expand DT_RUNPATH details.
462 ld.so.8: Expand DT_RUNPATH details.
464 Every 3 years we get asked why DT_RUNPATH doesn't work like DT_RPATH.
465 The most recent question was here:
466 https://www.sourceware.org/ml/libc-help/2017-06/msg00013.html
468 We need to expand the description of DT_RUNPATH to cover this
469 situation and explain that the DT_RUNPATH entries apply only to the
470 immediate DT_NEEDED, not that of another, say dlopen'd child object.
472 Since glibc 2.2.5, LD_PROFILE is ignored in secure-execution mode
474 Make notes on secure-execute mode more prominent
475 Place each note on secure-execution mode in a separate
476 paragraph, to make it more obvious.
478 Note that libraries in standard directories are not normally set-UID
479 In secure mode, LD_PRELOAD loads only libraries from standard
480 directories that are marked set-UID. Note that it is unusual for
481 a library to be marked in this way.
485 Note version where secure-execution started ignoring LD_USE_LOAD_BIAS
487 Correct glibc version that ignores LD_SHOW_AUXV in secure-execution mode
488 Ignored since 2.3.4 (not 2.3.5).
490 Rewrite LD_DEBUG_OUTPUT description and note that .PID is appended