==================== Changes in man-pages-4.05 ==================== Released: ????-??-??, 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: Adhemerval Zanella Akihiro Suda Alan Aversa Alan Cox Alec Leamas Alex Henrie Alexander Miller Andreas Gruenbacher Andreas Schwab Anna Schumaker Askar Safin Bill O. Gallmeister Carlos O'Donell Chris Pick Christoph Hellwig Craig Gallek Darrick J. Wong Davidlohr Bueso Dmitry V. Levin Dr. Tobias Quathamer Eric Blake Eric Dumazet Florian Weimer Gabriel Corona Heinrich Schuchardt Ivan Shapovalov Jakub Wilk Jason Baron Jason Vas Dias Jérémie Galarneau Jeremy Harris Joachim Wuttke Joe Stein John Stultz Josh Triplett Kondo, Naoya Krzysztof Adamski Manfred Spraul Marianne CHEVROT Marko Myllynen Mark Post Martin Gebert Mats Wichmann Matt Zimmerman Michael Kerrisk ` Mike Frysinger Minchan Kim Naoya Kondo Naresh Kamboju Nikola Forró Nikos Mavrogiannopoulos Orion Poplawski Pakin Yury Patrick Donnelly Paul Eggert Paul Pluzhnikov Peter Hurley Peter Wu Petr Gajdos Philip Semanchuk Rasmus Villemoes Rich Felker Simon Que Stephan Bergmann Stéphane Aulery Stephen Hurd Vincent Bernat William Preston Yuri Kozlov Zefram Apologies if I missed anyone! New and rewritten pages ----------------------- copy_file_range.2 Anna Schumaker [Darrick J. Wong, Christoph Hellwig, Michael Kerrisk] New page documenting copy_file_range() copy_file_range() is a new system call for copying ranges of data completely in the kernel. This gives filesystems an opportunity to implement some kind of "copy acceleration", such as reflinks or server-side-copy (in the case of NFS). personality.2 Michael Kerrisk This page has been greatly expanded, to add descriptions of personality domains. fmemopen.3 Michael Kerrisk [Adhemerval Zanella] Significant reworking of this page: * Rework discussion of the (obsolete) binary mode * Split open_memstream(3) description into a separate page. * Note various fmemopen() bugs that were fixed in glibc 2.22 * Greatly expand description of 'mode' argument * Rework description of 'buf' and 'len' arguments * Expand discussion of "current position" for fmemopen() stream ntp_gettime.3 Michael Kerrisk New page describing ntp_gettime(3) and ntp_gettimex(3) open_memstream.3 Michael Kerrisk New page created by split of fmemopen(3). At the same time, add and rework a few details in the text. posix_spawn.3 Bill O. Gallmeister, Michael Kerrisk New man page documenting posix_spawn(3) and posix_spawnp(3) readdir.3 Michael Kerrisk [Florian Weimer] Split readdir_r() content into separate page As suggested by Florian Weimer: It may make sense to move this documentation to a separate manual page, specific to readdir_r. This will keep the readdir() documentation nice and crisp. Most programmers will never have to consult all these details. Michael Kerrisk Near complete restructuring of the page and add some further details Michael Kerrisk [Florian Weimer, Rich Felker, Paul Eggert] Add a lot more detail on portable use of the 'd_name' field readdir_r.3 Michael Kerrisk [Florian Weimer] New page created after split of readdir(3). Michael Kerrisk [Florian Weimer] Explain why readdir_r() is deprecated and readdir() is preferred Michael Kerrisk [Florian Weimer] Remove misleading code example using pathconf() lirc.4 Alec Leamas New page documenting lirc device driver Newly documented interfaces in existing pages --------------------------------------------- adjtimex.2 Michael Kerrisk Document ntp_adjtime(3) epoll_ctl.2 Michael Kerrisk [Jason Baron] Document EPOLLEXCLUSIVE madvise.2 Minchan Kim [Michael Kerrisk] Document MADV_FREE Document the MADV_FREE flag added to madvise() in Linux 4.5. proc.5 Michael Kerrisk Document CmaTotal and CmaFree fields of /proc/meminfo Michael Kerrisk Document additional /proc/meminfo fields Document DirectMap4k, DirectMap4M, DirectMap2M, DirectMap1G Michael Kerrisk Document MemAvailable /proc/meminfo field Michael Kerrisk Document inotify /proc/PID/fdinfo entries Michael Kerrisk Document fanotify /proc/PID/fdinfo entries Michael Kerrisk Add some kernel version numbers for /proc/PID/fdinfo entries Michael Kerrisk [Patrick Donnelly] /proc/PID/fdinfo displays the setting of the close-on-exec flag Note also the pre-3.1 bug in the display of this info. socket.7 Craig Gallek [Michael Kerrisk, Vincent Bernat] Document some BPF-related socket options Document the behavior and the first kernel version for each of the following socket options: SO_ATTACH_FILTER SO_ATTACH_BPF SO_ATTACH_REUSEPORT_CBPF SO_ATTACH_REUSEPORT_EBPF SO_DETACH_FILTER SO_DETACH_BPF SO_LOCK_FILTER New and changed links --------------------- isalpha_l.3 Michael Kerrisk New link to isalpha.3 longjmp.3 Michael Kerrisk Replace page with link to setjmp(3), which now incorporates longjmp() ntp_adjtime.3 Michael Kerrisk New link to adjtimex(2) ntp_gettimex.3 Michael Kerrisk New link to ntp_gettime.3 open_wmemstream.3 Michael Kerrisk Update link to point to new open_memstream(2) page posix_spawnp.3 Michael Kerrisk New link to new posix_spawn.3 page siglongjmp.3 Michael Kerrisk Rewire link to point to setjmp(3) strerror_l.3 Michael Kerrisk New link to strerror.3 Fix missing link Global changes -------------- Various pages Michael Kerrisk Update FTM requirements (_DEFAULT_SOURCE) Michael Kerrisk Update feature test macro requirements Update to use _DEFAULT_SOURCE, and also changes brought by glibc commit 266865c0e7b79d4196e2cc393693463f03c90bd8. Various pages Michael Kerrisk Simplify FTM requirements Looking at (or feature_test_macros(7)), one can see that when _XOPEN_SOURCE is defined with the value 700 (or greater), then _POSIX_C_SOURCE is defined with the value 200809L (or greater). Therefore, terms in the man pages such as _XOPEN_SOURCE\ >=\ 700 || _POSIX_C_SOURCE\ >=\ 200809L can be simplified to: _POSIX_C_SOURCE\ >=\ 200809L Various pages Michael Kerrisk Simplify FTM requirements Looking at (or feature_test_macros(7)), one can see that when _XOPEN_SOURCE is defined with the value 600 (or greater), then _POSIX_C_SOURCE is defined with the value 200112L (or greater). Therefore, terms in the man pages such as _XOPEN_SOURCE\ >=\ 600 || _POSIX_C_SOURCE\ >=\ 200112L can be simplified to: _POSIX_C_SOURCE\ >=\ 200112L Various pages Michael Kerrisk Simplify FTM requirements _XOPEN_SOURCE implies _POSIX_C_SOURCE >=2, so simplify FTM requirements in various pages. Various pages Michael Kerrisk Remove "or 'cc -std=c99'" from SYNOPSIS Under the FTM requirements all of these pages document the requirement for _ISOC99_SOURCE. And feature_test_macros(7) now documents that "cc -std=c99" produces the same effect as defining _ISOC99_SOURCE. So, all of these pages don't additionally need to specify "or 'cc -std=c99'" under the FTM requirements in the SYNOPSIS. Removing that redundant text also simplifies the SYNOPSIS a little. Various pages Michael Kerrisk Simplify FTM requirements Looking at (or feature_test_macros(7)), one can see that when _XOPEN_SOURCE is defined with the value 600 (or greater), then _POSIX_C_SOURCE is defined with the value 200112L (or greater). Therefore, terms in the man pages such as _XOPEN_SOURCE\ >=\ 600 || _POSIX_C_SOURCE\ >=\ 200112L can be simplified to: _POSIX_C_SOURCE\ >=\ 200112L Various pages Michael Kerrisk Remove references to _XOPEN_SOURCE_EXTENDED in SYNOPSIS _XOPEN_SOURCE_EXTENDED is obsolete (it existed in SUSv1, but not subsequent standards). _XOPEN_SOURCE >= 500 produces the same effects as (_XOPEN_SOURCE && _XOPEN_SOURCE_EXTENDED). Modifying the SYNOPSIS of various ages that contain: _XOPEN_SOURCE\ >=\ 500 || _XOPEN_SOURCE\ &&\ _XOPEN_SOURCE_EXTENDED to just: _XOPEN_SOURCE\ >=\ 500 This has the following benefits: a) Simplifying the SYNOPSIS by removing ancient historical information. b) Preventing users from being misled into using _XOPEN_SOURCE_EXTENDED in new source code. Various pages Michael Kerrisk Remove mention of the obsolete _POSIX_SOURCE macro from SYNOPSIS _POSIX_SOURCE was a POSIX.1-1990 creation that was soon made obsolete bu _POSIX_C_SOURCE. Retaining mention of it in the feature test macro requirements section of the SYNOPSIS doesn't contain important information, and may mislead readers into actually trying to use this macro. A few mentions of it are maintained in a some pages where defining _POSIX_SOURCE inhibits some behavior. Various sockets-related pages Michael Kerrisk [Carlos O'Donell] Use consistent argument/variable names for socket addresses and lengths As noted by Carlos, there's quite a bit of inconsistency across pages. Use 'addr' and 'addrlen' consistently in variables and function arguments. Various pages Michael Kerrisk Wording fix: "current file offset" ==> "file offset" "File offset" is the preferred POSIX terminology. Various pages Michael Kerrisk Word "descriptor" more precisely Use either "file descriptor" or message queue descriptor". Various pages Michael Kerrisk ERRORS: add reference to signal(7) in description of EINTR Changes to individual pages --------------------------- locale.1 Marko Myllynen Add "locale -c charmap" as an example Addresses https://bugzilla.kernel.org/show_bug.cgi?id=104511. localedef.1 Marko Myllynen Add hint on purpose of --no-archive Indicate why using --no-archive might be a good idea. The issue is that if you create a custom locale with localedef(1) and put it to the locale archive then during the next glibc upgrade the locale archive is updated as well and your custom locale is gone.) accept.2 Michael Kerrisk ERRORS: improve description for EBADF adjtimex.2 Michael Kerrisk [John Stultz] Various improvements after feedback from John Stultz Michael Kerrisk Remove FTM requirements It seems that adjtimex() never needed _BSD_SOURCE (and my earlier commit 5918743bc8b02b was simply a blunder). Michael Kerrisk Split EINVAL error cases Michael Kerrisk Note treatment of out-of-range buf.offset Michael Kerrisk Don't refer reader to adjtime(3) Probably, it's not wise to suggest adjtime(3) as the more portable API. Rather, ntp_adjtime(3) should be used. Michael Kerrisk [Naresh Kamboju] Update details of buf.offset EINVAL error Michael Kerrisk SEE ALSO: add ntp_gettime(3) Michael Kerrisk Improve description of some PPS timex fields Michael Kerrisk Add ATTRIBUTES section William Preston [Petr Gajdos] Update a detail in adjtimex return value description Michael Kerrisk Note range constraints and clamping for ADJ_FREQUENCY bdflush.2 Michael Kerrisk Note that glibc support for this system call went away in version 2.23 bind.2 Michael Kerrisk Improve description of ENOENT error bpf.2 Michael Kerrisk Document close-on-exec semantics The close-on-exec file descriptor flag is automatically enabled for FDs returned by bpf(). chmod.2 Michael Kerrisk Clarify terminology (file mode versus file permission bits) chown.2 Michael Kerrisk ERRORS: improve EBADF description clone.2 unshare.2 Michael Kerrisk Remove mention of _BSD_SOURCE and _SVID_SOURCE The right way to expose declarations for these Linux-specific system calls was always _GNU_SOURCE. Mentioning the historical use of _BSD_SOURCE and _SVID_SOURCE just clouds the issue. connect.2 Michael Kerrisk ERRORS: improve EBADF description create_module.2 Michael Kerrisk Glibc 2.23 removed last vestiges of support for this system call delete_module.2 Michael Kerrisk Glibc 2.23 removed last vestiges of support for this system call epoll_ctl.2 Michael Kerrisk Document ELOOP error for circular monitoring loops eventfd.2 Michael Kerrisk Note that eventfd info is available in /proc/PID/fdinfo execve.2 Michael Kerrisk [Krzysztof Adamski] Add EPERM error for capabilities check of capability-dumb binaries Michael Kerrisk Add reference to ld-linux.so(8) Michael Kerrisk SEE ALSO: add system(3) fanotify_init.2 Michael Kerrisk Note kernel version that allowed O_CLOEXEC for event_f_flags fcntl.2 flock.2 Michael Kerrisk SEE ALSO: add lslocks(8) fcntl.2 Michael Kerrisk [Jason Vas Dias] Rework description of F_SETOWN As suggested by Jason, make it clearer that I/O signalling requires the use of both F_SETOWN and O_ASYNC. While we're at, make a few other cleanups to the text. Michael Kerrisk Remove mention of _BSD_SOURCE to get definition of F_SETOWN/F_GETOWN This usage went away in glibc 2.20, and the simplest remedy is just to omit mention of it. futex.2 Michael Kerrisk FUTEX_CLOCK_REALTIME can now be used with FUTEX_WAIT get_kernel_syms.2 Michael Kerrisk Note that glibc des not support this system call init_module.2 Michael Kerrisk Glibc 2.23 removed last vestiges of support for this system call ioctl_list.2 Heinrich Schuchardt Include uapi/linux/wireless.h Add the list of wireless IOCTLs. Heinrich Schuchardt Path to sockios.h sockios.h is now in include/uapi Heinrich Schuchardt Add reference to netdevice.7 netdevice.7 describes most of the IOCTLs of sockios.h Heinrich Schuchardt Transfer structure (wireless.h IOCTLs) The sole parameter to be passed to the wireless.h IOCTLs is of type struct iwreq *. ioperm.2 Michael Kerrisk [Alex Henrie] ioperm.2: Permissions are inherited across fork(2) See https://bugzilla.kernel.org/show_bug.cgi?id=99911 iopl.2 Michael Kerrisk [Alex Henrie] Permissions are not inherited across fork(2) or preserved on execve(2) See https://bugzilla.kernel.org/show_bug.cgi?id=99901 lseek.2 Michael Kerrisk FUSE now supports SEEK_HOLE and SEEK_DATA Michael Kerrisk NFS supports SEEK_HOLE and SEEK_DATA Michael Kerrisk SEE ALSO: add open(2) madvise.2 Michael Kerrisk Clarify MADV_HWPOISON wording to say that it applies to a page range mknod.2 Michael Kerrisk SEE ALSO: add mknod(1) mount.2 Michael Kerrisk SEE ALSO: add findmnt(8) open.2 Michael Kerrisk NOTES: mention existence of proc/PID/fd and /proc/PID/fdinfo Mark Post [Petr Gajdos] O_TMPFILE support is now provided bt Btrfs pipe.2 Michael Kerrisk [Eric Blake] Note treatment of 'pipefd' on error poll.2 Michael Kerrisk [Josh Triplett] Document spurious EAGAIN error that can occur on other systems Light reworking of text proposed by Josh Triplett. readlink.2 Michael Kerrisk Clarify EINVAL error description recv.2 Heinrich Schuchardt Equivalence to read() Describe the recv(2)-read(2) and the recvfrom(2)-recv(2) equivalences for zero-valued arguments. Michael Kerrisk MSG_WAITALL has no effect for datagram sockets recv.2 cmsg.3 Nikola Forró Fix type of cmsg_len member of cmsghdr structure The type shown for cmsg_len member of cmsghdr structure is socklen_t, but the actual type used by glibc and the kernel is size_t. The information was obtained from glibc source code: http://bit.ly/21m1RMp Michael Kerrisk Note that cmsg_len is typed as socklen_t in POSIX.1 sched_setaffinity.2 Michael Kerrisk [Florian Weimer, Florian Weimer] Warn that CPU_ALLOC() may allocate a slightly CPU set than requested Michael Kerrisk [Florian Weimer] Add reference to CPU_ALLOC(3) sched_setattr.2 Michael Kerrisk [Akihiro Suda] EPERM depends on affinity mask of target thread, not calling thread select.2 Michael Kerrisk [Josh Triplett] Document spurious EAGAIN error that can occur on other systems Light reworking of text proposed by Josh Triplett. Nikos Mavrogiannopoulos Mention the 'fd_set' size limitation early and refer to poll(2) Change this because of the serious limitation of select() imposing a limit on the range of file descriptors that can be monitored. This is currently mentioned too late in the documentation (in the NOTES section). The man page should warn early and refer to poll(2) as soon as possible. Michael Kerrisk Add details on the glibc fixed-size fd_set limitation No modern application should use select() on Linux. select_tut.2 Michael Kerrisk Some readability fixes to example program Michael Kerrisk Better variable names in example program Michael Kerrisk Simplify 'if' logic in in example program Michael Kerrisk Use correct type (socklen_t) for addrlen semctl.2 Michael Kerrisk [Davidlohr Bueso, Manfred Spraul, Philip Semanchuk] NOTES: note when 'sempid' is set on various implementations See https://bugzilla.kernel.org/show_bug.cgi?id=112271 and http://thread.gmane.org/gmane.linux.kernel/2162754/ Subject: [PATCH] Don't set sempid in semctl syscall. Date: 2016-02-26 12:21:38 GMT semop.2 Michael Kerrisk Tweak comment describing 'sempid' sendfile.2 Askar Safin Fix incorrect description in text referring to splice(2) Michael Kerrisk SEE ALSO: add copy_file_range(2) setpgid.2 Michael Kerrisk Correct/simplify FTM requirements for BSD setpgrp() and getpgrp() signalfd.2 Michael Kerrisk Note that signalfd info is available in /proc/PID/fdinfo sigprocmask.2 Michael Kerrisk [Mike Frysinger] Explicitly refer the reader to sigsetops(3) This man page did not make it obvious which functions should be used for manipulating signals sets, nor where those functions were documented. socketpair.2 Michael Kerrisk [Eric Blake] Note treatment of 'sv' on error splice.2 Askar Safin Improve description of 0 return value. See https://bugzilla.kernel.org/show_bug.cgi?id=90911 statfs.2 Michael Kerrisk [Jakub Wilk] Use consistent case for hex constants sync.2 Christoph Hellwig Clarify description and document the Linux data integrity guarantees syscall.2 Mike Frysinger Add more architectures and improve error documentation Move the error register documentation into the main table rather than listing them in sentences after the fact. Add sparc error return details. Add details for alpha/arc/m68k/microblaze/nios2/powerpc/superh/ tile/xtensa. syscalls.2 Michael Kerrisk Add copy_file_range(2) times.2 Kondo, Naoya Fix an incorrect description in NOTES The text has an incorrect description in NOTES, it says that (2^32/HZ) - 300 is about 429 million. It is correct only if HZ=10 which does not look common today. So just removing "(i.e., about 429 million)" is good enough. truncate.2 Michael Kerrisk SEE ALSO: add truncate(1) uselib.2 Michael Kerrisk Mention CONFIG_USELIB Michael Kerrisk Note that glibc does not support this (obsolete) system call wait.2 wait4.2 Michael Kerrisk Rename the "status" argument to "wstatus" The fact that exit(3)/_exit(2) has an argument called "status" and the same name is used in the arguments to the wait*() calls can a little too easily lead the user into thinking that the two arguments hold the same information, when of course they don't. So, use a different name for the argument of the wait*() functions, to reduce the chances of such confusion. backtrace.3 Michael Kerrisk [Martin Gebert] Small fixes to example program clearenv.3 Michael Kerrisk [Matt Zimmerman] Clarify the use and effect of clearenv() See http://bugs.debian.org/679323 Michael Kerrisk Variables can be added to the environment after calling clearenv() clog10.3 Michael Kerrisk Show an alternative equivalence for clog10() Michael Kerrisk Update CONFORMING TO Fix grammar error and add C11. dl_iterate_phdr.3 Michael Kerrisk [Paul Pluzhnikov] Describe 'struct dl_phdr_info' fields added in glibc 2.4 See https://bugzilla.kernel.org/show_bug.cgi?id=103011 Michael Kerrisk [Simon Que] Note that first object visited by 'callback' is the main program See https://bugzilla.kernel.org/show_bug.cgi?id=94141 errno.3 Michael Kerrisk Add some explanation of ENOENT error exec.3 Michael Kerrisk SEE ALSO: add system(3) exp.3 Michael Kerrisk [Joachim Wuttke] SEE ALSO: add expm1(3) fopen.3 Michael Kerrisk SEE ALSO: add open_memstream(3) fts.3 Michael Kerrisk BUGS: glibc-2.23 now has LFS support for the fts functions gamma.3 Michael Kerrisk [Alan Cox] gamma() was documented in SVID 2 getaddrinfo.3 Michael Kerrisk [Andreas Schwab, Orion Poplawski] Update FTM requirements for glibc 2.22 Since glibc 2.22 getaddrinfo() etc. are only declared for POSIX.1-2001 or later. getcwd.3 Michael Kerrisk SEE ALSO: add pwd(1) opendir.3 Michael Kerrisk Help the reader by explicitly mentioning the use of readdir(3) perror.3 Michael Kerrisk Suggest use of strerror(3) in place of deprecated 'sys_errlist' posix_fallocate.3 Jérémie Galarneau ERRORS: add EINTR The glibc implementation of posix_fallocate(), which calls fallocate(), may be interrupted. The fallocate() emulation also makes use of pread()/pwrite(), which may also be interrupted. posix_memalign.3 Michael Kerrisk [Eric Blake] Note posix_memalign()'s treatment of 'memptr' on error pthread_setaffinity_np.3 Michael Kerrisk SEE ALSO: add CPU_SET(3) queue.3 Dr. Tobias Quathamer Remove double CONFORMING TO section rcmd.3 Nikola Forró Add missing condition concerning .rhosts file The list of conditions determining if iruserok() and ruserok() functions automatically fail is incomplete. According to glibc source code, the functions also fail if the .rhosts file is hard linked anywhere. setbuf.3 Michael Kerrisk SEE ALSO: add stdbuf(1) setjmp.3 Michael Kerrisk Rewrite and merge longjmp()/siglongjmp() discussion into this page The discussion of nonlocal gotos is much easier to read if setjmp() and longjmp() are discussed in the same page. While we're at it, rework almost the entire text and add several more details. Michael Kerrisk Note the interactions of longjmp() and non-async-signal-safe functions POSIX.1-2008 TC2 adds explicit text on this point. See http://austingroupbugs.net/view.php?id=516#c1195 Michael Kerrisk Explain why nonlocal gotos make code harder to maintain Michael Kerrisk Reword warning on longjmp() to function that has already returned Michael Kerrisk Remove reference to obsolete _XOPEN_SOURCE_EXTENDED sleep.3 Michael Kerrisk SEE ALSO: add sleep(1) strftime.3 Michael Kerrisk [Jeremy Harris] Note which 'tm' fields are used to calculate each output string See https://bugzilla.redhat.com/show_bug.cgi?id=1162218 strlen.3 Michael Kerrisk [Alan Aversa] CONFORMING TO: add C11 system.3 Michael Kerrisk SEE ALSO: add execve(2) termios.3 Dr. Tobias Quathamer Document line length in canonical mode See https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/tty/n_tty.c#n1673 See https://bugs.debian.org/797479 Michael Kerrisk SEE ALSO: add tty(1) Michael Kerrisk [Peter Hurley] Further improvements to recent tweaks of canonical mode 4096 char limit timegm.3 Michael Kerrisk [Stephen Hurd, Mats Wichmann] Remove sample implementation of timegm() Stephen and Mats both question the wisdom of showing a portable *non-thread-safe* implementation of timegm(), and I find it hard to disagree. So, remove this code. See https://bugzilla.kernel.org/show_bug.cgi?id=103701 Michael Kerrisk Expand DESCRIPTION a little st4.4 Dr. Tobias Quathamer Remove spurious copyright section tty_ioctl.4 Michael Kerrisk SEE ALSO: add ldattach(1) elf.5 Michael Kerrisk [Gabriel Corona, Mike Frysinger] Fix description of STV_PROTECTED Michael Kerrisk Improve description of STV_DEFAULT Michael Kerrisk Improve description of STV_HIDDEN Chris Pick Remove erroneous, duplicate SHN_* section Michael Kerrisk [Chris Pick] Reword discussion of range values a little gai.conf.5 Michael Kerrisk Add VERSIONS section group.5 Michael Kerrisk SEE ALSO: add groups(2) SEE ALSO: add gpasswd(1) SEE ALSO: add sg(1) SE ALSO: add gshadow(5) SEE ALSO: add chgrp(1) locale.5 Marko Myllynen [Mike Frysinger] tel + fax are deprecated nsswitch.conf.5 Nikola Forró Update NSS compatibility mode description utmp.5 Michael Kerrisk SEE ALSO: add lslogins(1) aio.7 Andreas Gruenbacher Improve example When aio_sigevent.sigev_notify is set to SIGEV_SIGNAL, signal handlers called for asynchronous I/O operations will have si->si_code set to SI_ASYNCIO. Check to make sure that si->si_value.sival_ptr is defined. capabilities.7 Michael Kerrisk Explain safety check for capability-dumb binaries Michael Kerrisk SEE ALSO: add sg(1), su(1) SEE ALSO: add id(1), group(5), passwd(5) credentials.7 Michael Kerrisk SEE ALSO: add groups(2) environ.7 Michael Kerrisk Describe the Bourne "NAME=value command" syntax Michael Kerrisk Add some details describing hos shell's environment is initialized Michael Kerrisk Note that child of fork(2) inherits copy of parent's environment Michael Kerrisk SEE ALSO: add pam_env(3) epoll.7 Michael Kerrisk Mention that epoll info is available via /proc/PID/fdinfo fanotify.7 Michael Kerrisk Refer reader to proc(5) for info on /proc/PID/fdinfo fanotify entries feature_test_macros.7 Michael Kerrisk Add a summary of some FTM key points Michael Kerrisk Give an early hint about some macros being defined by default Michael Kerrisk Clarify relation between _XOPEN_SOURCE >=500 and _XOPEN_SOURCE_EXTENDED Emphasize that defining _XOPEN_SOURCE >=500 produces same effects as defining _XOPEN_SOURCE_EXTENDED. Michael Kerrisk Note that man pages don't mention _XOPEN_SOURCE_EXTENDED As per previous commit, mention of _XOPEN_SOURCE_EXTENDED has generally been removed from the man pages. Michael Kerrisk Note effects of "cc -std=c99" and "cc -std=c11" Michael Kerrisk Clarify some _ISOC99_SOURCE / _DEFAULT_SOURCE details Michael Kerrisk Clarify that _XOPEN_SOURCE_EXTENDED is obsolete Since SUSv2, _XOPEN_SOURCE_EXTENDED is no longer specified in the standard. inotify.7 Michael Kerrisk Refer reader to proc(5) for info on /proc/PID/fdinfo inotify entries ip.7 Eric Dumazet Document IP_BIND_ADDRESS_NO_PORT socket option mq_overview.7 Michael Kerrisk Note that the close-on-exec flag is automatically set on MQ descriptors namespaces.7 Michael Kerrisk SEE ALSO: add lsns(1) lsns(1) was recently added in util-linux, probably to appear in next release (2.28?). pipe.7 Michael Kerrisk [Jason Vas Dias] Clarify that I/O signalling requires use of both F_SETOWN and O_ASYNC Michael Kerrisk SEE ALSO: add mkfifo(1) signal.7 Michael Kerrisk Note the interactions of longjmp() and non-async-signal-safe functions See http://austingroupbugs.net/view.php?id=516#c1195. socket.7 Michael Kerrisk SEE ALSO: add pcap(3) SEE ALSO: add wireshark(1) and tcpdump(8) standards.7 Michael Kerrisk Add POSIX.1-2008 TC2 (POSIX.1-2016) svipc.7 Michael Kerrisk Tweak description of 'sempid' Michael Kerrisk SEE ALSO: add lsipc(1) symlink.7 Michael Kerrisk [Zefram] Some "magic" symlinks have permissions other than 0777 See https://bugs.debian.org/743525 time.7 Michael Kerrisk SEE ALSO: add timeout(1) SEE ALSO: add ntp_adjtime(3) and ntp_gettime(3) unicode.7 Dr. Tobias Quathamer Document private use areas See https://bugs.debian.org/285444 unix.7 Heinrich Schuchardt Add example A complete example demonstrating the usage of sockets for local interprocess communication is added. Michael Kerrisk Introduce term "sequenced-packet" for SOCK_SEQPACKET Michael Kerrisk Some wording improvements