==================== Changes in man-pages-3.45 ==================== Released: 2012-12-21, Christchurch 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: Andi Kleen Cyril Hrubis David Prévot Elie De Brauwer Eric Dumazet Felipe Pena Florian Weimer Gao Feng Jan Glauber Jim Paris Jon Grant Julien Cristau Michael Kerrisk Mike Frysinger Rens van der Heijden Simon Paillard Thierry Vignaud Trevor Woerner YOSHIFUJI Hideaki Apologies if I missed anyone! New and rewritten pages ----------------------- s390_runtime_instr.2 Jan Glauber New page for s390-specific s390_runtime_instr(2) if_nameindex.3 YOSHIFUJI Hideaki Document if_nameindex(3) and if_freenameindex(3) Michael Kerrisk Edits, improvements and corrections to Hideaki's page Michael Kerrisk Add an example program if_nametoindex.3 YOSHIFUJI Hideaki New page documenting if_nametoindex(3) and if_indextoname(3) Newly documented interfaces in existing pages --------------------------------------------- New and changed links --------------------- if_freenameindex.3 Michael Kerrisk New link to if_nameindex.3 if_indextoname.3 Michael Kerrisk New link to if_nametoindex.3 Global changes -------------- sysconf.3 cciss.4 Michael Kerrisk Global fix: s/runtime/run time/ Changes to individual pages --------------------------- clone.2 Michael Kerrisk Since 2.6.30, CLONE_NEWIPC also supports POSIX message queues delete_module.2 Michael Kerrisk Small rewording of description of effect of O_TRUNC getrlimit.2 Michael Kerrisk [Trevor Woerner] Document Linux's nonstandard treatment or RLIMIT_CPU soft limit Upon encountering the RLIMIT_CPU soft limit when a SIGXCPU handler has been installed, Linux invokes the signal handler *and* raises the soft limit by one second. This behavior repeats until the limit is encountered. No other implementation that I tested (Solaris 10, FreeBSD 9.0, OpenBSD 5.0) does this, and it seems unlikely to be POSIX-conformant. The (Linux-specific) RLIMIT_RTTIME soft limit exhibits similar behavior. Michael Kerrisk Point reader at discussion of /proc/[pid]/limits in proc(5) io_getevents.2 Michael Kerrisk io_getevents() may cause segfault when called with invalid ctx_id For reference see: http://marc.info/?l=linux-aio&m=130089887002435&w=2 recv.2 Michael Kerrisk [Eric Dumazet] UNIX domain sockets support MSG_TRUNC since 3.4 sendmmsg.2 Elie De Brauwer Add example program for sendmmsg() stat.2 Simon Paillard Clarify description of EOVERFLOW error The EOVERFLOW error is not only for st_size, but also inode and block size fields. See glibc source file sysdeps/unix/sysv/linux/xstatconv.c and kernel source file fs/stat.c. Also, fix bit/byte confusion See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=604928 syscalls.2 Michael Kerrisk Update various references to "i386" to "x86" Michael Kerrisk Add s390_runtime_instr(2) sysctl.2 Michael Kerrisk Mention CONFIG_SYSCTL_SYSCALL Michael Kerrisk Calls to sysctl() log warnings to the kernel log since 2.6.24 syslog.2 Michael Kerrisk Substantially reorganize discussion of commands Make the layout of the discussion of the commands more readable. Michael Kerrisk Add kernel symbolic 'type' names Michael Kerrisk Clarify SYSLOG_ACTION_SIZE_UNREAD semantics SYSLOG_ACTION_SIZE_UNREAD returns the number of bytes available for reading via SYSLOG_ACTION_READ. Michael Kerrisk Clarify where SYSLOG_ACTION_READ_ALL places data it reads Michael Kerrisk Clarify semantics of SYSLOG_ACTION_CLEAR The SYSLOG_ACTION_CLEAR command (5) does not really clear the ring buffer; rather it affects the semantics of what is returned by commands 3 (SYSLOG_ACTION_READ_ALL) and 4 (SYSLOG_ACTION_READ_CLEAR). Michael Kerrisk Clarify discussion of privileges for commands 3 and 10 Michael Kerrisk Add mention of CONFIG_LOG_BUF_SHIFT wait.2 Michael Kerrisk BUGS: Document odd waitid() behavior when 'infop' is NULL getifaddrs.3 Michael Kerrisk [Julien Cristau] Update description of ifa_data to Linux 2.6+ reality See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=526778 memcmp.3 Michael Kerrisk [Jon Grant] Enhance RETURN VALUE text and remove redundant text from DESCRIPTION Note that sign of result equals sign of difference between first two bytes that differ (treated as "unsigned char")." mkstemp.3 Michael Kerrisk [Florian Weimer] Deemphasize discussion of mode 066 in glibc 2.0.6 Glibc 2.0.6 is now so ld that the discussion of details of that version can be deemphasized placing just under NOTES. See https://bugzilla.kernel.org/show_bug.cgi?id=51811 strcmp.3 Michael Kerrisk [Jon Grant] Enhance RETURN VALUE text and remove redundant text from DESCRIPTION Note that sign of result equals sign of difference between first two bytes that differ (treated as "unsigned char")." ttyname.3 Michael Kerrisk Fix confused text in ERRORS The existing text suggested that the ERRORS applied only for ttyname_r(). However, 2 of the 3 errors can occur for ttyname(). undocumented.3 Michael Kerrisk Remove some now documented functions proc.5 Michael Kerrisk [Jim Paris] Correct description of SwapFree in /proc/meminfo Michael Kerrisk Note change of /proc/[pid]/limits file permissions in 2.6.36 resolv.conf.5 Simon Paillard Document IPv6 format for nameserver See: http://bugs.debian.org/610036 capabilities.7 Michael Kerrisk [Rens van der Heijden] Correct URL for POSIX.1e draft ipv6.7 Gao Feng Add description of getsockopt() for IPV6_MTU In IPv4,IP_MTU is only supported by getsockopt. In IPv6, we can use IPV6_MTU to set socket's MTU, but the return value of getsockopt() is the path MTU. rtnetlink.7 Michael Kerrisk [Julien Cristau] Update description of IFLA_STATS to Linux 2.6+ reality See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=526778 socket.7 Michael Kerrisk [YOSHIFUJI Hideaki] Document 'sockaddr' and 'sockaddr_storage' types Andi Kleen Explain effect of SO_SNDTIMEO for connect() When SO_SNDTIMEO is set before connect(), then connect() may return EWOULDBLOCK when the timeout fires.