==================== Changes in man-pages-3.69 ==================== Released: 2014-06-14, Munich Contributors ------------ The following people contributed patches/fixes or (noted in brackets in the changelog below) reports, notes, and ideas that have been incorporated in changes in this release: Cyril Hrubis Jan Chaloupka Jeff Layton Kirill A. Shutemov KOSAKI Motohiro Marko Myllynen Michael Kerrisk NeilBrown Peng Haitao Petr Gajdos Qian Lei Rasmus Villemoes Vasiliy Kulikov Walter Harms Yuri Kozlov Apologies if I missed anyone! New and rewritten pages ----------------------- iconv.1 Marko Myllynen [Michael Kerrisk] New page for the iconv(1) command iconvconfig.8 Marko Myllynen New page for iconvconfig(8) Newly documented interfaces in existing pages --------------------------------------------- fcntl.2 Jeff Layton, Michael Kerrisk Document open file description locks As provided by the fcntl() operations F_OFD_SETLK, F_OFD_SETLKW, and F_OFD_GETLK Changes to individual pages --------------------------- locale.1 Marko Myllynen Add FILES section, add charmap(5) reference Marko Myllynen Align with recent charmap(5) / repertoiremap(5) changes execve.2 Michael Kerrisk [NeilBrown] Before kernel 2.6.0, RLIMIT_NPROC had no effect for set*uid() Michael Kerrisk [Vasiliy Kulikov] RLIMIT_NPROC is checked only if preceded by set*uid() Michael Kerrisk [Vasiliy Kulikov, NeilBrown, KOSAKI Motohiro] Document EAGAIN error See also https://bugzilla.kernel.org/show_bug.cgi?id=42704 fcntl.2 Michael Kerrisk Detail the limitations of traditional (process-associated) locks Michael Kerrisk [Jeff Layton] Describe how to check whether the kernel supports a particular command Michael Kerrisk ERRORS: add EINVAL for invalid 'cmd' Michael Kerrisk Add para introducing advisory locks and noting existence of OFD locks Michael Kerrisk [Jeff Layton] Add notes on F_SETLKW deadlock detection and its limitations Michael Kerrisk Add an explicit note that mandatory locking is not in POSIX Michael Kerrisk Rewrite introductory paragraphs on mandatory locking Make the structure more logical, and also explicitly mention OFD locks. Michael Kerrisk [Jeff Layton] Reword discussion of mandatory lock bug a little Jeff Layton confirmed that the bug remains even in modern kernels. Michael Kerrisk Explain POSIX background to EACCES/EAGAIN error for F_GETLK Michael Kerrisk Add NOTES subhead for record locking and NFS Michael Kerrisk [NeilBrown] Note treatment of locks when an NFS client loses contact with the server Michael Kerrisk [NeilBrown] Refine discussion of locks when NFSv4 client loses contact with server Michael Kerrisk [Jeff Layton] nfsv4leasetime controls the "contact lost" interval for NFSv4 Jeff Layton: flock.2 Michael Kerrisk In some modern BSDs, fcntl() and flock() locks do interact So, reword and extend the discussion of this topic in NOTES. Michael Kerrisk Move NOTES text describing implementation of flock() Michael Kerrisk [NeilBrown] Add more details on NFS, including Linux 2.6.37 changes Also: move NOTES text describing interaction of fcntl() and flock() locks. fork.2 Michael Kerrisk Add notes on inheritance of flock() and OFD locks across fork() lseek.2 Michael Kerrisk Add reference to open(2) for discussion of file descriptors and OFDs open.2 Michael Kerrisk Rework and extend the discussion of open file descriptions open_by_handle_at.2 Rasmus Villemoes Add reference to feature_test_macros(7) recvmmsg.2 Rasmus Villemoes Add reference to feature_test_macros(7) remap_file_pages.2 Michael Kerrisk [Kirill A. Shutemov] Note that remap_file_pages() is deprecated sendmmsg.2 Rasmus Villemoes Add reference to feature_test_macros(7) seteuid.2 Michael Kerrisk seteuid() and setegid() are implemented as library functions Michael Kerrisk Error checking should always be performed, even when caller is UID 0 setresuid.2 Michael Kerrisk Document EAGAIN error that can occur after kernel alloc_uid() failure Michael Kerrisk Since Linux 3.1, the EAGAIN case for RLIMIT_NPROC no longer occurs Michael Kerrisk Correct the description of the EAGAIN error Michael Kerrisk Error checking should always be performed, even when caller is UID 0 setreuid.2 Michael Kerrisk Document EAGAIN error that can occur after kernel alloc_uid() failure Michael Kerrisk Error checking should always be performed, even when caller is UID 0 Michael Kerrisk Add EAGAIN error for hitting RLIMIT_NPROC limit Michael Kerrisk Since Linux 3.1, the EAGAIN case for RLIMIT_NPROC no longer occurs setuid.2 Michael Kerrisk Document EAGAIN error that can occur after kernel alloc_uid() failure Michael Kerrisk Correct the description of the EAGAIN error Michael Kerrisk Error checking should always be performed, even when caller is UID 0 Michael Kerrisk Since Linux 3.1, the EAGAIN case for RLIMIT_NPROC no longer occurs statfs.2 Cyril Hrubis Update MAGIC constants Most of the updates are taken from /usr/include/linux/magic.h, some were found by grepping the Linux source code. Cyril Hrubis [Michael Kerrisk] fstatfs(2) was broken on file descriptors from pipe(2) syscalls.2 Michael Kerrisk Note that remap_file_pages() is deprecated basename.3 Peng Haitao ATTRIBUTES: Note functions that are thread-safe The functions basename() and dirname() are thread safe. catgets.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function catgets() is thread safe. getdate.3 Rasmus Villemoes Use blank definition of _GNU_SOURCE in example program getdirentries.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function getdirentries() is thread safe. getdtablesize.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function getdtablesize() is thread safe. iconv.3 Qian Lei [Peng Haitao] ATTRIBUTES: Note function that is thread-safe The function iconv() is thread safe. Michael Kerrisk SEE ALSO: add iconvconfig(8) lockf.3 Qian Lei [Peng Haitao] ATTRIBUTES: Note function that is thread-safe The function lockf() is thread safe. malloc_get_state.3 Rasmus Villemoes SYNOPSIS: use correct header The nonstandard functions malloc_set_state() and malloc_get_state() are provided by not . malloc_usable_size.3 Qian Lei ATTRIBUTES: Note function that is thread-safe The function malloc_usable_size() is thread safe. matherr.3 Qian Lei [Peng Haitao] ATTRIBUTES: Note function that is thread-safe The function matherr() is thread safe. mkdtemp.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function mkdtemp() is thread safe. mkstemp.3 Peng Haitao ATTRIBUTES: Note functions that are thread-safe The functions mkstemp(), mkostemp(), mkstemps() and mkostemps() are thread safe. mq_close.3 Qian Lei ATTRIBUTES: Note function that is thread-safe The function mq_close() is thread safe. mq_getattr.3 Qian Lei ATTRIBUTES: Note function that is thread-safe The functions mq_getattr() and mq_setattr() are thread safe. mq_open.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function mq_open() is thread safe. mq_receive.3 Peng Haitao ATTRIBUTES: Note functions that are thread-safe The functions mq_receive() and mq_timedreceive() are thread safe. mq_send.3 Peng Haitao ATTRIBUTES: Note functions that are thread-safe The functions mq_send() and mq_timedsend() are thread safe. mq_unlink.3 Qian Lei ATTRIBUTES: Note function that is thread-safe The function mq_unlink() is thread safe. posix_fallocate.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function posix_fallocate() is thread safe. posix_openpt.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function posix_openpt() is thread safe. siginterrupt.3 Peng Haitao ATTRIBUTES: Note function that is not thread-safe The function siginterrupt() is not thread safe. system.3 Peng Haitao ATTRIBUTES: Note function that is thread-safe The function system() is thread safe. charmap.5 Marko Myllynen Update to match current glibc charmap(5) was outdated, bring it to closer to reality by fixing syntax descriptions to match current glibc code and practices, adding missing options, removing obsolete comments and references, and removing now incorrect examples. locale.5 Marko Myllynen Clarify AM/PM settings a bit localedef(1) complains if really undefined, should be empty instead. Also: add some SEE ALSO references. Marko Myllynen Document glibc conventions regarding days and week Based on existing practice and glibc community wiki page at https://sourceware.org/glibc/wiki/Locales proc.5 Michael Kerrisk [Jan Chaloupka, Walter Harms] Add a brief description of /proc/fs repertoiremap.5 Marko Myllynen New page for repertoiremap(5) Rather obsolete feature but localedef(1) refers to repertoiremaps. bootparam.7 Petr Gajdos Describe 'rootflags' and 'rootdelay' kernel parameters Patch based on text from Documentation/kernel-parameters.txt charsets.7 Marko Myllynen Update to reflect past developments Rewrite the introduction to make Unicode's prominence more obvious. Reformulate parts of the text to reflect current Unicode world. Minor clarification for ASCII/ISO sections, some other minor fixes. Marko Myllynen List CJK encodings in the order of C, J, K environ.7 Michael Kerrisk SEE ALSO: add env(1), printenv(1), ld.so(8) locale.7 Marko Myllynen Add some SEE ALSO references man-pages.7 Michael Kerrisk Note that .TH 'date' field is nowadays automatically updated by scripts signal.7 Michael Kerrisk Describe EINTR semantics for recvmmsg(2) Michael Kerrisk Clarify text describing EINTR semantics for socket interfaces unicode.7 Marko Myllynen Update to reflect past developments The unicode(7) page will look more modern with few small changes: - drop old BUGS section, editors cope with UTF-8 ok these days, and perhaps the state-of-the-art is better described elsewhere anyway than in a man page - drop old suggestion about avoiding combined characters - refer to LANANA for Linux zone, add registry file reference - drop a reference to an inactive/dead mailing list - update some reference URLs utf-8.7 Marko Myllynen Drop an old comment about UTF-8 replacing ISO 8859 And add locale(1) under SEE ALSO.