Changes

   1 ==================== Changes in man-pages-3.70 ====================
   2
   3 Released: ????-??-??, Munich
   4
   5
   6 Contributors
   7 ------------
   8
   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:
  12
  13 Carlos O'Donell <carlos@redhat.com>
  14 Elie De Brauwer <eliedebrauwer@gmail.com>
  15 Florian Weimer <fweimer@redhat.com>
  16 Heinrich Schuchardt <xypron.glpk@gmx.de>
  17 Marko Myllynen <myllynen@redhat.com>
  18 Michael Kerrisk <mtk.manpages@gmail.com>
  19 Nadav Har'El <nyh@math.technion.ac.il>
  20 NeilBrown <neilb@suse.de>
  21 Rich Felker <dalias@libc.org>
  22 Robert P. J. Day <rpjday@crashcourse.ca>
  23 Simon Paillard <spaillard@debian.org>
  24 Tomi Salminen <tlsalmin@gmail.com>
  25 Walter Harms <wharms@bfs.de>
  26 Yuri Kozlov <yuray@komyakino.ru>
  27 Кирилл <kirillnow@gmail.com>
  28
  29 Apologies if I missed anyone!
  30
  31
  32 New and rewritten pages
  33 -----------------------
  34
  35 sprof.1
  36     Michael Kerrisk [Marko Myllynen]
  37         New page for the glibc sprof(1) command
  38
  39
  40 Newly documented interfaces in existing pages
  41 ---------------------------------------------
  42
  43 epoll_ctl.2
  44     NeilBrown
  45         Document EPOLLWAKEUP
  46
  47 epoll.7
  48     NeilBrown
  49         Document EPOLLWAKEUP
  50
  51
  52 Changes to individual pages
  53 ---------------------------
  54
  55 iconv.1
  56 iconvconfig.8
  57     Marko Myllynen
  58         Clarify gconv file terminology a bit
  59
  60 ldd.1
  61     Michael Kerrisk
  62         SEE ALSO: add sprof(1)
  63
  64 connect.2
  65     Michael Kerrisk
  66         ERRORS: add EPROTOTYPE
  67             See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=708394
  68
  69 dup.2
  70     Michael Kerrisk  [Rich Felker]
  71         Fix erroneous discussion regarding closing 'newfd' before calling dup2()
  72             And propose a workaround if the caller cares about catching
  73             close() time errors.
  74
  75             See http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2#comment36888604_23444465
  76             and http://austingroupbugs.net/view.php?id=411
  77     Michael Kerrisk
  78         Rework and enhance discussion of dup2()
  79             In particular, note that dup2() performs the steps of closing
  80             and reusing 'newfd' atomically.
  81     Michael Kerrisk
  82         Add subhead for dup3()
  83
  84 epoll_ctl.2
  85     Michael Kerrisk
  86         BUGS: EPOLLWAKEUP is silently ignored without CAP_BLOCK_SUSPEND
  87             If the caller does not have CAP_BLOCK_SUSPEND, then EPOLLWAKEUP
  88             is silently ignored.
  89
  90 fcntl.2
  91     Michael Kerrisk  [Tomi Salminen]
  92         The return value for F_SETPIPE_SZ is the pipe capacity
  93     Michael Kerrisk
  94         ERRORS: Document ENOTDIR error for F_NOTIFY
  95     Michael Kerrisk
  96         Use proper page cross-references in F_NOTIFY discussion
  97     Michael Kerrisk
  98         Suggest the use of real-time signals with F_NOTIFY
  99
 100 getitimer.2
 101     Michael Kerrisk
 102         Rewrite a few pieces to clarify some details
 103
 104 inotify_add_watch.2
 105     Michael Kerrisk
 106         Clarify that the target of a watch is an i-node
 107             The target of a watch is an i-node, not a pathname. Clarify
 108             the text to prevent the reader possibly misunderstanding
 109             that establishing watches by two different links to the same
 110             file might create different watch descriptors.
 111
 112 open.2
 113     Michael Kerrisk  [Кирилл]
 114         O_CLOEXEC is also one the flags not ignored when O_PATH is specified
 115
 116 pipe.2
 117     Elie De Brauwer
 118         PIPE_BUF is defined in limits.h
 119             To make use of PIPE_BUF in an application one should include
 120             <limits.h>.  Adding a reference to this inclusion.
 121
 122 poll.2
 123     Michael Kerrisk  [Nadav Har'El]
 124         The negate-fd-to-ignore technique does not work for file descriptor 0
 125             See https://bugzilla.kernel.org/show_bug.cgi?id=79411
 126
 127 set_tid_address.2
 128     Michael Kerrisk  [Rich Felker]
 129         Use "thread" rather than "process" in DESCRIPTION
 130     Michael Kerrisk
 131         SEE ALSO: add gettid(2)
 132
 133 shmop.2
 134     Michael Kerrisk
 135         Explain SHMLBA in much more detail
 136
 137 splice.2
 138     Michael Kerrisk
 139         Document EAGAIN error
 140             See https://bugzilla.kernel.org/show_bug.cgi?id=48641
 141
 142 syscalls.2
 143     Carlos O'Donell
 144         Add prlimit64(2)
 145             While trying to reconcile the new features in glibc with the
 146             documented entries in the linux kernel man pages I noticed that
 147             glibc exports prlimit64 for use by 32-bit applications (as does
 148             the linux kernel), but that prlimit64 was not defined in the
 149             syscalls list or in the prlimit-related page.
 150
 151             This is not the complete fix for this, but I don't have the time
 152             to explain why and when prlimit64 should be used (or how it should
 153             be used safely). Therefore I'm just patching the syscalls.2 list
 154             to show that prlimit64 exists and was added in 2.6.36 (verified
 155             with git by checking out the tags before and after).
 156
 157             Unless you've purposely excluded prlimit64 to avoid telling users
 158             about it because it's complicated, please apply.
 159
 160 syslog.2
 161     Michael Kerrisk
 162         Rework introductory paragraph
 163     Michael Kerrisk  [Robert P. J. Day]
 164         Rework text describing loglevels
 165             The kernel header file mentioned in the discussion of the KERN_*
 166             constants has morphed and is no longer exported inside glibc.
 167             And the definitions of the constants themselves changed subtly
 168             with kernel commit 04d2c8c83d0e3ac5f78aeede51babb3236200112.
 169             So, rewrite the description of the constants to be a bit more
 170             abstract.
 171     Michael Kerrisk
 172         Rewrite parts of the page, and import /proc/sys/kernel/printk
 173             * Move /proc/sys/kernel/printk from proc(5) to this page,
 174               and correct various details in the discussion of that file.
 175             * Rewrite and correct various other details on the page.
 176             * Clean out some crufty text.
 177             * Miscellaneous minor fixes.
 178     Michael Kerrisk
 179         Update SYSLOG_ACTION_CONSOLE_OFF + SYSLOG_ACTION_CONSOLE_ON description
 180             Details changed in Linux 2.6.32
 181
 182 tee.2
 183     Michael Kerrisk
 184         Document EAGAIN error
 185             See https://bugzilla.kernel.org/show_bug.cgi?id=48641
 186
 187 vmsplice.2
 188     Michael Kerrisk
 189         Document EAGAIN error
 190             See https://bugzilla.kernel.org/show_bug.cgi?id=48641
 191
 192 ether_aton.3
 193     Michael Kerrisk
 194         Make description of ether_line() bug a little more informative
 195
 196 mallopt.3
 197     Michael Kerrisk  [Florian Weimer]
 198         MALLOC_MMAP_THRESHOLD_ and MALLOC_MMAP_MAX_ *do* work in setgid programs
 199             My testing on this point was bogus, overlooking details of
 200             strace(1)'s behavior with setuid programs.
 201
 202             See https://sourceware.org/bugzilla/show_bug.cgi?id=12155
 203
 204 printf.3
 205     Michael Kerrisk  [Rich Felker]
 206         Remove libc4 and libc5 details
 207             Rich Felker noted that "scare text" in the man page warned about
 208             the use of snprintf() on libc, and that some people had cited
 209             this as a reason not to use snprintf().  Linux libc is now
 210             ancient history, so there is no real need to keep that text.
 211             But, while we're at it, we may as well clear out all of the
 212             other ancient libc4 and libc5 pieces in the page. They are
 213             nowadays more clutter than help.
 214     Michael Kerrisk
 215         SUSv3 and later agree with C99 for the snprintf() return value
 216             Determined by inspection of the SUSv3 and SUSv4 specifications.
 217     Michael Kerrisk
 218         Remove some old text about glibc 2.0 changes
 219             We probably don't now need such ancient info.
 220     Michael Kerrisk
 221         Update references to standards for C and S conversion specifiers
 222
 223 profil.3
 224     Michael Kerrisk
 225         SEE ALSO: add sprof(1)
 226
 227 charmap.5
 228     Marko Myllynen
 229         Various minor updates and improvements
 230             - more precise title
 231             - extend description a bit
 232             - document previously omitted WIDTH_DEFAULT
 233     Marko Myllynen
 234         Remove accidental ISO C compliance reference
 235             glibc refers in locale/programs/charmap.c to ISO C 99 section
 236             7.17.(2) and ISO C 99 section 5.2.1.(3) that if a character map
 237             is not ASCII compatible then the locale using it is not ISO C
 238             compliant. This does not state anything about the character set
 239             itself.
 240
 241 proc.5
 242     Michael Kerrisk
 243         Replace /proc/sys/kernel/printk discussion with reference to syslog(2)
 244             It makes more sense to have the /proc/sys/kernel/printk with
 245             the related material in syslog(2).
 246     Michael Kerrisk
 247         Rewrite /proc/sys/kernel/printk description
 248
 249 inotify.7
 250     Michael Kerrisk
 251         Clarify with events are generated for watched directories
 252             Really, with respect to watched directories, events fall into
 253             three classes (not two, as was shown):
 254
 255             * Events that can be generated only for the watched directory.
 256             * Events that can be generated only for objects that are inside
 257               the watched directory.
 258             * Events that can be generated both for the watched directory
 259               and for objects inside the directory.
 260
 261             So, mark these three classes more clearly in the list of inotify
 262             events.
 263     Heinrich Schuchardt  [Michael Kerrisk]
 264         BUGS: Note possible bug triggered by watch descriptor reuse
 265             Watch descriptor IDs are returned by inotify_add_watch().
 266             When calling inotify_rm_watch() an IN_IGNORE is placed on the
 267             inotify queue pointing to the ID of the removed watch.
 268
 269             inotify_add_watchi() should not return a watch descriptor ID for
 270             which events are still on the queue but should return an
 271             unused ID.
 272
 273             Unfortunately, the existing Kernel code does not provide such a
 274             guarantee.
 275
 276             Actually, in rare cases watch descriptor IDs are returned by
 277             inotify_add_watch() for which events are still on the inotify
 278             queue.
 279
 280             See https://bugzilla.kernel.org/show_bug.cgi?id=77111
 281     Michael Kerrisk
 282         Add further detail to the watch descriptor reuse bug
 283             As well as inotify_rm_watch(), file deletion and unmounting a
 284             filesystem can also cause a watch descriptor to be deleted.
 285     Michael Kerrisk
 286         The watch descriptor reuse bug may be hard to hit in practice
 287             Explain the circumstances in detail, indicating that the
 288             bug may be very unlikely to occur in practice.
 289     Michael Kerrisk
 290         Clarify description of IN_EXCL_UNLINK
 291             Clarify the text a little, in particular making it clearer
 292             that the target of a watch is an i-node (not a pathname).
 293     Michael Kerrisk
 294         Clarify IN_ONESHOT explanation
 295             Make it clearer that the target of monitoring is an i-node,
 296             not a pathname.
 297     Michael Kerrisk
 298         Make comment on 'mask' field more accurate
 299
 300 libc.7
 301     Michael Kerrisk
 302         Clarify man-pages policy on documenting C libraries other tahn glibc
 303     Michael Kerrisk
 304         Use absolute dates in discussion of libc vs glibc
 305
 306 pipe.7
 307     Elie De Brauwer
 308         Add reference that the pipe capacity can be changed
 309             In fcntl(2) F_SETPIPE_SZ, F_GETPIPE_SZ and
 310             /proc/sys/fs/pipe-max-size are defined, however
 311             pipe(7) still defines the pipe capacity as being
 312             a static entity. Adding a reference to fcntl(2).
 313     Michael Kerrisk  [Walter Harms]
 314         Clarify that since 2.6.35, 65535 bytes is the default pipe capacity
 315
 316 ld.so.8
 317     Michael Kerrisk
 318         Clarify that LD_PROFILE can specify just a single shared object
 319     Michael Kerrisk
 320         Clarify that LD_PROFILE output is appended to target file
 321             The LD_PROFILE output is appended to any existing
 322             contents of the target file.
 323     Michael Kerrisk
 324         SEE ALSO: add sprof(1)