The Linux man-pages maintainer proudly announces:
man-pages-4.10 - man pages for Linux
This release resulted from patches, bug reports, reviews, and comments
from over 40 contributors. The release sees a large number of changes:
over 600 commits changing around 160 pages. The changes include the
addition of 11 pages, significant rewrites of 3 other pages, and
enhancements to many other pages.
Tarball download:
http://www.kernel.org/doc/man-pages/download.html
Git repository:
https://git.kernel.org/cgit/docs/man-pages/man-pages.git/
Online changelog:
http://man7.org/linux/man-pages/changelog.html#release_4.10
A short summary of the release is blogged at:
http://linux-man-pages.blogspot.com/2017/03/man-pages-410-is-released.html
The current version of the pages is browsable at:
http://man7.org/linux/man-pages/
A selection of changes in this release that may be of interest
to readers on LKML is shown below.
Cheers,
Michael
==================== Changes in man-pages-4.10 ====================
New and rewritten pages
-----------------------
add_key.2
Michael Kerrisk [Eugene Syromyatnikov, David Howells]
Major improvements and additions
The page has doubled in length.
ioctl_iflags.2
Michael Kerrisk
New page describing inode flags and ioctl() operations
ioctl_ns.2
Michael Kerrisk
New page created by splitting ioctl(2) operations out of namespaces(7)
keyctl.2
Michael Kerrisk, Eugene Syromyatnikov [David Howells, Mat Martineau]
A vast number of additions and improvements
The page has gone from somewhat over 100 lines to well over
1000 lines and now more or less documents the complete interface
provided by this system call.
getentropy.3
Michael Kerrisk
New page documenting getentropy(3)
getentropy(3) is added to glibc in version 2.25.
keyrings.7
David Howells
New page (written by David Howells) adopted from keyutils
Michael Kerrisk [Eugene Syromyatnikov, David Howells]
Very many additions and improvements
persistent-keyring.7
Michael Kerrisk
New page (written by David Howells) adopted from keyutils
Since this page documents kernel-user-space interfaces,
it makes sense to have it as part of man-pages, rather
than the keyutils package.
Michael Kerrisk
Various clean-ups and additions
process-keyring.7
Michael Kerrisk
New page (written by David Howells) adopted from keyutils
Since this page documents kernel-user-space interfaces,
it makes sense to have it as part of man-pages, rather
than the keyutils package.
Michael Kerrisk
Various additions and improvements
request_key.2
Michael Kerrisk, Eugene Syromyatnikov [David Howells]
Very many additions and improvements
The page is now three times its former length.
session-keyring.7
Michael Kerrisk
New page (written by David Howells) adopted from keyutils
Michael Kerrisk
Various reworking and additions
signal-safety.7
Michael Kerrisk
New page created by migrating the signal-safety discussion from
signal(7). Along the way some more details got added.
Michael Kerrisk [KASAKI Motohiro]
Note async-signal-safety problems caused by pthread_atfork()
See https://bugzilla.kernel.org/show_bug.cgi?id=25292
Michael Kerrisk [KASAKI Motohiro]
Note glibc deviations from POSIX requirements
See https://bugzilla.kernel.org/show_bug.cgi?id=25292
thread-keyring.7
Michael Kerrisk
New page (written by David Howells) adopted from keyutils
Michael Kerrisk
Various rewordings and additions
user-keyring.7
Michael Kerrisk
New page (written by David Howells) adopted from keyutils
Michael Kerrisk
Various reworking and improvements
user-session-keyring.7
Michael Kerrisk
New page (written by David Howells) adopted from keyutils
Michael Kerrisk
Various rewordings and additions
Newly documented interfaces in existing pages
---------------------------------------------
bzero.3
Michael Kerrisk
Document explicit_bzero() (new in glibc 2.25)
Also, reword the description of bzero somewhat.
proc.5
Michael Kerrisk
Document /proc/sys/vm/user_reserve_kbytes
Michael Kerrisk
Document /proc/sys/vm/admin_reserve_kbytes
Michael Kerrisk
Document /proc/sys/fs/mount-max
Michael Kerrisk
Document /proc/PID/status 'NoNewPrivs' field
Changes to individual pages
---------------------------
clone.2
Michael Kerrisk
clone() does not execute fork handlers
execve.2
Michael Kerrisk
File capabilities can be ignored for the same reasons as set-UID/set-GID
Michael Kerrisk
The 'no_new_privs' bit inhibits transformations of the effective IDs
fork.2
Michael Kerrisk
cgroup PIDs controller may also be trigger for EAGAIN error
getrlimit.2
Michael Kerrisk
Document role of RLIMIT_NOFILE for FD passing over UNIX sockets
link.2
Michael Kerrisk
When using linkat() AT_EMPTY_PATH, 'olddirfd' must not be a directory
Michael Kerrisk
ERRORS: add EPERM for immutable/append-only files
Michael Kerrisk
Note limits where EMLINK is encountered on ext4 and Btrfs
mbind.2
Krzysztof Kulakowski [Michael Kerrisk]
Update MPOL_BIND description
The behavior of MPOL_BIND changed in Linux 2.6.26.
nanosleep.2
Michael Kerrisk
Describe "creeping sleep" problem
nanosleep() has a problem if used in a program that catches
signals and those signals are delivered at a very high rate.
Describe the problem, and note that clock_nanosleep(2)
provides a solution.
posix_fadvise.2
Michael Kerrisk
Mention /proc/sys/vm/drop_caches
It may be helpful for the reader of this page to know about
/proc/sys/vm/drop_caches.
Michael Kerrisk
One can use open(2) + mmap(2) + mincore(2) as a 'fincore'
Note that open(2) + mmap(2) + mincore(2) can be used to get a view
of which pages of a file are currently cached.
Michael Kerrisk [Maik Zumstrull]
Note that POSIX_FADV_DONTNEED *may* try to write back dirty pages
read.2
Michael Kerrisk [Steven Luo]
Remove crufty text about EINTR and partial read
Remove bogus text saying that POSIX permits partial read
to return -1/EINTR on interrupt by a signal handler.
That statement already ceased to be true in SUSv1 (1995)!
See https://bugzilla.kernel.org/show_bug.cgi?id=193111
remap_file_pages.2
Michael Kerrisk
remap_file_pages() has been replaced by a slower in-kernel emulation
setxattr.2
Michael Kerrisk
ERRORS: add EPERM for immutable/append-only files
signalfd.2
Michael Kerrisk
signalfd() doesn't play well with helper programs spawned by libraries
See https://lwn.net/Articles/415684/.
Michael Kerrisk
signalfd can't be used to receive synchronously generated signals
Signals such as the SIGSEGV that results from an invalid
memory access can be caught only with a handler.
syscall.2
Mike Frysinger
Add endian details with 64-bit splitting
Architectures that split 64-bit values across register pairs
usually do so according to their C ABI calling convention (which
means endianness). Add some notes to that effect, and change the
readahead example to show a little endian example (since that is
way more common than big endian).
Also start a new list of syscalls that this issue does not apply
to.
Mike Frysinger
Note parisc handling of aligned register pairs
While parisc would normally have the same behavior as ARM/PowerPC,
they decide to write shim syscall stubs to unpack/realign rather
than expose the padding to userspace.
vfork.2
Michael Kerrisk
Note some caveats re the use of vfork()
Inspired by Rich Felker's post at http://ewontfix.com/7/.
See also https://sourceware.org/bugzilla/show_bug.cgi?id=14749 and
See also https://sourceware.org/bugzilla/show_bug.cgi?id=14750.
write.2
Michael Kerrisk [Kai NODA]
Alert the reader that there is a limit on 'count'
See https://bugzilla.kernel.org/show_bug.cgi?id=86061
exec.3
Michael Kerrisk
execl() and execle() were not async-signal-safe before glibc 2.24
sem_wait.3
Michael Kerrisk [Fabjan Sukalia]
Remove statement that SA_RESTART does not cause restarting
This has not been true since Linux 2.6.22. The description
of EINTR maintains a reference to signal(7), which explains
the historical details.
See https://bugzilla.kernel.org/show_bug.cgi?id=192071
core.5
Michael Kerrisk
Clarify that dumping program's initial CWD is root directory
Michael Kerrisk
The target of core dump piping can also be a script
proc.5
Mike Frysinger [Michael Kerrisk]
Clarify /proc/pid/environ behavior
/proc/pid/environ reflects process environment at
*start* of program execution; it is set at time of execve(2)
capabilities.7
Michael Kerrisk [Casey Schaufler]
Add subsection with notes to kernel developers
Provide some notes to kernel developers considering how to choose
which capability should govern a new kernel feature.
Michael Kerrisk
Further enhance the recommendation against new uses of CAP_SYS_ADMIN
inotify.7
Michael Kerrisk
Point out that inotify monitoring is inode based
mount_namespaces.7
Michael Kerrisk [Rob Landley]
Rework the discussion of defaults for mount propagation types
Add rather more detail. In particular, note the cases where the
default propagation type is MS_PRIVATE vs MS_SHARED.
netlink.7
Michael Kerrisk
NETLINK_IP6_FW went away in Linux 3.5
Michael Kerrisk
NETLINK_W1 went away in Linux 2.6.18
Michael Kerrisk
NETLINK_FIREWALL was removed in Linux 3.5
Michael Kerrisk
NETLINK_NFLOG was removed in Linux 3.17
pid_namespaces.7
Keno Fischer
CLONE_SIGHAND|CLONE_VM|CLONE_NEWPID is no longer disallowed
pipe.7
Michael Kerrisk
Since Linux 4.9, pipe-max-size is ceiling for the default pipe capacity
Michael Kerrisk
Clarify that default pipe capacity is 16 pages
The statement that the default pipe capacity is 65536 bytes
is accurate only on systems where the page size is 4096B.
See the use of PIPE_DEF_BUFFERS in the kernel source.
unix.7
Michael Kerrisk [Willy Tarreau]
Document ETOOMANYREFS for SCM_RIGHTS send exceeding RLIMIT_NOFILE limit
ld.so.8
Michael Kerrisk
LD_BIND_NOT has effect only for function symbols
Michael Kerrisk
Describe use of LD_DEBUG with LD_BIND_NOT
Michael Kerrisk
In secure mode, LD_AUDIT restricts the libraries that it will load
Michael Kerrisk
LD_AUDIT understands $ORIGIN, $LIB, and $PLATFORM
--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/