1 ==================== Changes in man-pages-3.41 ====================
3 Released: 2011-05-11, Christchurch
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 Akihiro MOTOKI <amotoki@gmail.com>
14 Andries E. Brouwer <Andries.Brouwer@cwi.nl>
15 Angelo Borsotti <angelo.borsotti@gmail.com >
16 Bjarni Ingi Gislason <bjarniig@rhi.hi.is>
17 Brian M. Carlson <sandals@crustytoothpaste.ath.cx>
19 David Prévot <taffit@debian.org>
20 D. Barbier <bouzim@gmail.com>
21 Eric Blake <eblake@redhat.com>
22 Hugh Dickins <hughd@google.com>
23 Ivana Varekova <varekova@redhat.com>
24 Jakub Jelinek <jakub@redhat.com>
25 Jan Kara <jack@suse.cz>
26 Jason Baron <jbaron@redhat.com>
27 Jean-Michel Vourgère <jmv_deb@nirgal.com>
28 Jeff Moyer <jmoyer@redhat.com>
29 Josh Triplett <josh@joshtriplett.org>
30 Kasper Dupont <kasperd@gmail.com>
31 KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
32 Lauri Kasanen <curaga@operamail.com>
33 Mel Gorman <mgorman@suse.de>
34 Michael Kerrisk <mtk.manpages@gmail.com>
35 Mike Frysinger <vapier@gentoo.org>
36 Nick Piggin <npiggin@gmail.com>
37 Paul Pluzhnikov <ppluzhnikov@google.com>
38 Petr Baudis <pasky@ucw.cz>
39 Ralph Corderoy <ralph@inputplus.co.uk >
40 Rich Felker <bugdal@aerifal.cx>
41 Simone Piccardi <piccardi@truelite.it>
42 Simon Paillard <spaillard@debian.org>
43 Stefan Puiu <stefan.puiu@gmail.com>
44 Stephen Hemminger <shemminger@vyatta.com>
45 Vincent Lefevre <vincent@vinc17.net>
46 Yuri Kozlov <yuray@komyakino.ru>
48 Apologies if I missed anyone!
51 New and rewritten pages
52 -----------------------
55 Ivana Varekova [Michael Kerrisk]
56 New page documenting get_robust_list(2) and set_robust_list(2)
59 Michael Kerrisk [KOSAKI Motohiro, Paul Pluzhnikov]
60 New page for mallinfo(3)
63 Michael Kerrisk [Jakub Jelinek]
64 New page for malloc_info(3)
67 Michael Kerrisk [KOSAKI Motohiro]
68 New man page for malloc_stats(3)
71 Newly documented interfaces in existing pages
72 ---------------------------------------------
76 Document MADV_DONTDUMP and MADV_DODUMP
84 New link to new get_robust_list.2 page
102 CIRCLEQ_INSERT_AFTER.3
103 CIRCLEQ_INSERT_BEFORE.3
104 CIRCLEQ_INSERT_HEAD.3
105 CIRCLEQ_INSERT_TAIL.3
109 The queue(3) page documents these macros, so it makes sense to
110 have links for the names.
114 New link to des_crypt.3
115 The des_crypt(3) page documents this macro, so it makes sense
116 to have a link for the name.
121 Overlooked to add this link in 3.38, when documentation of
122 qsort_r() was added to the qsort.3 page.
147 Michael Kerrisk [Lauri Kasanen]
148 Global fix: note glibc version that added library support
163 Michael Kerrisk [Andries E. Brouwer]
164 Clarify that these functions operate on bytes, not (wide) characters
165 Change 'character(s)' to 'byte(s)' to make clear that these
166 functions operate on bytes, not wide / UTF8 characters.
167 (POSIX uses 'byte(s)' similarly, to make this point.)
177 Remove names of constants from NAME line
178 Some of the sockets/network protocol pages included names of
179 the corresponding address family constants in the NAME line,
180 but this wasn't done consistently across all pages, and probably
181 it adds little value in those pages that did do this. So, remove
182 these constants from those pages that have them in the NAME
186 Changes to individual pages
187 ---------------------------
190 Michael Kerrisk [Josh Triplett]
191 Expand description of CLOCK_REALTIME
192 Make it clear that this clock may be discontinuous, and is
193 affected my incremental NTP and clock-adjtime(2) adjustments.
195 See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=540872
199 Clarify that 'timeout' is a *minimum* interval
200 Make it clear that 'timeout' is a minimum interval; the actual
201 interval will be rounded up to the system clock granularity,
202 and may overrun because of kernel scheduling delays.
206 Rewording to deemphasize libc5 details
211 Can occur on, for example, non-MMU hardware.
215 Add RETURN VALUE and ERRORS sections
217 Refer reader to NOTES for more info about 'tcache'
219 DESCRIPTION: reword a sentence to be clearer
227 Rewrite to focus on system call API
228 Rewrite to focus on the system call interface, adding
229 some notes on the libaio wrapper differences.
230 See the following mail:
231 2012-05-07 "aio manuals", linux-man@vger.kernel.org
232 http://thread.gmane.org/gmane.linux.man/1935/focus=2910
234 Other minor rewrites.
238 Comment out an old Linux libc detail
241 Nick Piggin [KOSAKI Motohiro, Jan Kara, Hugh Dickins]
242 Describe race of direct I/O and fork()
243 Rework 04cd7f64, which didn't capture the details correctly.
244 See the April/May 2012 linux-man@ mail thread "[PATCH]
245 Describe race of direct read and fork for unaligned buffers"
246 http://thread.gmane.org/gmane.linux.kernel.mm/77571
250 Clarify that 'timeout' is a *minimum* interval
251 Make it clear that 'timeout' is a minimum interval; the actual
252 interval will be rounded up to the system clock granularity,
253 and may overrun because of kernel scheduling delays.
255 Clarify discussion of wrapper function emulation
256 Clarify that glibc (as well as old libc) provides emulation
257 using select(2) on older kernels that don't have a poll()
260 Make the meaning of a zero timeout explicit
261 Clarify that timeout==0 causes an immediate return, even if
262 no file descriptors are ready.
265 Michael Kerrisk [Kasper Dupont]
266 BUGS: Note O_APPEND + pwrite() does the wrong thing
267 See https://bugzilla.kernel.org/show_bug.cgi?id=43178
271 Clarify that 'timeout' is a *minimum* interval
272 Make it clear that 'timeout' interval will be rounded up to the
273 system clock granularity, and may overrun because of kernel
278 Clarify that 'timeout' is a *minimum* interval
279 Make it clear that 'timeout' is a minimum interval; the actual
280 interval will be rounded up to the system clock granularity,
281 and may overrun because of kernel scheduling delays.
283 Expand description of the self-pipe trick
285 Add further details on pselect6() system call that underlies pselect()
289 Clarify that 'timeout' of semtimedop() is a *minimum* interval
290 Make it clear that 'timeout' interval will be rounded up to the
291 system clock granularity, and may overrun because of kernel
296 Note that 'sig_t' requires _BSD_SOURCE
297 Also remove some old Linux libc details
301 Clarify that 'timeout' of sigtimedwait() is a *minimum* interval
302 Make it clear that 'timeout' is a minimum interval; the actual
303 interval will be rounded up to the system clock granularity,
304 and may overrun because of kernel scheduling delays.
309 From "groff -ww" (or "man --warnings=w ..."):
311 warning: around line 442: table wider than line width
313 GNU man uses line length of 78.
315 Use text blocks. Two spaces between sentences or better: start
316 each sentence in a new line.
321 From "groff -ww ..." (or "man --warnings=w ..."):
323 warning: around line 157: table wider than line width
325 Have to use text blocks. Move some text to its correct column.
326 Split text to two columns to avoid hyphenation.
330 Remove reference to obsolete libc5
334 Remove some details about obsolete Linux libc
344 ERRORS: Add/update ENOSYS error
348 Clarify what happens when a request isn't successfully canceled
350 Add pointers to aio(7) and sigevent(7)
354 SYNOPSIS: Add header file <fcntl.h>
355 Upstreamed from Debian, and consistent with FreeBSD
360 Note details of POSIX.1-2008 specification of 'b' in 'mode'
361 Michael Kerrisk [Rich Felker]
362 BUGS: fmemopen() doesn't correctly set file position in some cases
363 If 'mode' is append, but 'size' does not cover a null byte
364 in 'buf', then fmemopen() incorrectly sets the initial file
365 position to -1, rather than the next byte after the end of
368 See http://sourceware.org/bugzilla/show_bug.cgi?id=13151
370 BUGS: fmemopen() incorrectly handles size==0 case
371 If size is zero, fmemopen() fails, This is surprising behavior,
372 and not specified in POSIX.1-2008.
374 See http://sourceware.org/bugzilla/show_bug.cgi?id=11216
376 Reported-by; Alex Shinn <alexshinn@gmail.com>
378 BUGS: Note silent ABI change for fmemopen() in glibc 2.9
379 Michael Kerrisk [Rich Felker]
380 BUGS: Append mode does not force writes to append
381 Append mode correctly sets the initial offset but does
382 not force subsequent writes to append at end of stream.
384 See http://sourceware.org/bugzilla/show_bug.cgi?id=13152
385 Michael Kerrisk [Eric Blake]
386 BUGS: Note inconsistent treatment of 'b' in 'mode'
387 fopen() permits, for example, both "w+b" and "wb+",
388 but only the latter is meaningful to fmemopen().
390 See http://sourceware.org/bugzilla/show_bug.cgi?id=12836
393 Michael Kerrisk [Petr Baudis]
394 Correct description of return for user-supplied 'write' function
395 See http://sourceware.org/bugzilla/show_bug.cgi?id=2074
399 Note that AI_ADDRCONFIG is not affected by loopback addresses
400 See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=660479
404 Upstream useful NOTE from Debian
405 Warn the reader that the pointer arguments can't be
406 interpreted as C style strings. Also, note possible
407 alignment requirements for the referenced bytes sequences,
409 Write a better paragraph introducing iconv() and its arguments
412 Michael Kerrisk [Vincent Lefevre]
413 Clarify that the arguments to these macros must be real-floating
414 See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=609033
418 Clarify that async notification occurs when *all* I/Os complete
422 SYNOPSIS: Correct return types of major() and minor()
423 See https://bugzilla.redhat.com/show_bug.cgi?id=754188
425 Reported-by; Zdenek Kabelac <zkabelac@redhat.com>
429 SEE ALSO: Add malloc_info(3)
433 Fix wordos in function names in NAME line
438 The example code was a version that was not consistent with
439 the shell output shown on the page.
441 Reported-bY: Simon Paillard <spaillard@debian.org>
443 Restore accidentally omitted line in shell session
445 SEE ALSO: Add malloc_stats(3)
449 Change target of link to mmap.2 (was mmap2.2)
450 Upstreamed from Red Hat / Fedora
453 Michael Kerrisk [Casper.Dik@oracle.com]
454 Remove note about Solaris possibly returning a relative path
457 Michael Kerrisk [Ralph Corderoy]
458 Document behavior when 'ident' argument to openlog() is NULL
459 See https://bugs.launchpad.net/ubuntu/+source/manpages/+bug/382096
461 Update CONFORMING TO for POSIX.1-2008
462 POSIX.1-2008 doesn't change any details, but make
467 Remove some functions that have been documented
471 Remove reference to nonexistent scsi(4) page
472 Upstreamed from RedHat / Fedora
475 Michael Kerrisk [Stephen Hemminger]
476 Note that this driver was removed in 2.6.28
477 See https://bugs.launchpad.net/ubuntu/+source/manpages/+bug/528020
481 Upstream pieces from Red Hat/Fedora
482 Note mention of empty password field.
483 Add description of "*NP*" in password field.
485 Various minor fixes and improvements
489 Note that CAP_SYS_ADMIN processes can override file-max
490 Upstreamed from red Hat / Fedora
492 Document /proc/[pid]/cgroup
493 Upstreamed from Red Hat / Fedora
497 Take a Debian improvement into upstream
501 Mention timezone directories in DESCRIPTION
502 Note that timezone files are usually in /usr/lib/zoneinfo
503 or /usr/share/zoneinfo.
506 The SYNOPSIS doesn't correspond to a user-visible file.
508 SEE ALSO: Add pointer to glibc source file timezone/tzfile.h
510 SEE ALSO: add tzset(3) and tzselect(8)
514 Indent for "troff" makes table too wide
515 Fix following from "groff -t -ww ...":
517 warning: around line 53: table wider than line width
519 Extra indent for "troff" makes the table look misplaced
520 (default "ps" output).
525 From "nroff -ww -t ...":
527 warning: around line 44: table wider than line width
529 Columns are made narrower (column gutter decreased).
533 Add ENODEV error for bind() to link-local IPv6 address
536 Michael Kerrisk [Simone Piccardi]
537 Clarify that SIGLOST is unused
539 Comment out crufty BUGS text on SIGLOST
540 It must be a very long time since the statement there
541 about SIGLOST was true. (The text seems to date back to
544 Update architectures for tables of signal numbers
549 This patch clarifies that 0xc0 and 0xc1 are not valid in any UTF-8
550 encoding[0], and it also references RFC 3629 instead of RFC 2279.
552 [0] In order to have 0xc0, you'd have to have a two-byte encoding
553 with all the data bits zero in the first byte (and thus only six
554 bits of data), which would be an ASCII character encoded in the
555 non-shortest form. Similarly with 0xc1.
557 See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=538641
562 Remove path prefix from NAME line
563 Command names shown in NAME are normally just the basename,
564 not the full pathname of the command.