Ready for 5.04

Removed trailing white space at end of lines

bpf-helpers.7: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2, fallocate.2, ioctl_iflags.2, ioctl_list.2, pidfd_open.2, pivot_root.2, quotactl.2, seccomp.2, select.2, wait.2, proc.5, cgroups.7, netdevice.7, uts_namespaces.7: tstamp

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Changes: Ready for 5.04

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

bpf-helpers.7: Fix .TH line

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

bpf-helpers.7: srcfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

bpf-helpers.7: Refresh against kernel v5.4-rc7

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Use pid_t for clone3() {child,parent}_tid

Advertise to userspace that they should use proper pid_t types
for arguments returning a pid.

The kernel-internal struct kernel_clone_args currently uses int
as type and since POSIX mandates that pid_t is a signed integer
type and glibc and friends use int this is not an issue. After
the merge window for v5.5 closes we can switch struct
kernel_clone_args over to using pid_t as well without any danger
in regressing current userspace.

Also note, that the new set tid feature which will be merged for
v5.5 uses pid_t types as well.

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Check for MAP_FAILED not NULL on mmap()

If mmap() fails it will return MAP_FAILED which according to the manpage
is (void *)-1 not NULL.

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: tfix

Fix two spelling mistakes in manpage describing the clone{2,3}()
syscalls/syscall wrappers.

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mmap.2: tfix

Reported-by: Marko Myllynen <myllynen@redhat.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mmap.2: Some rewording of the description of MAP_STACK

Reword a little to allow for the fact that there are now
*two* reasons to consider using this flag.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mmap.2: Note that MAP_STACK exists on some other systems

As noted in man-pages commit 99c3a000279919cc4875c9dfa9c,
MAP_STACK exists on at least OpenBSD and FreeBSD.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pivot_root.2: EXAMPLE: allocate stack using mmap() MAP_STACK rather than malloc()

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Allocate child's stack using mmap(2) rather than malloc(3)

Christian Brauner suggested mmap(MAP_STACKED), rather than
malloc(), as the canonical way of allocating a stack for the
child of clone(), and Jann Horn noted some reasons why:

    Not on Linux, but on OpenBSD, they do use MAP_STACK now
    AFAIK; this was announced here:
    <http://openbsd-archive.7691.n7.nabble.com/stack-register-checking-td338238.html>.
    Basically they periodically check whether the userspace
    stack pointer points into a MAP_STACK region, and if not,
    they kill the process. So even if it's a no-op on Linux, it
    might make sense to advise people to use the flag to improve
    portability? I'm not sure if that's something that belongs
    in Linux manpages.

    Another reason against malloc() is that when setting up
    thread stacks in proper, reliable software, you'll probably
    want to place a guard page (in other words, a 4K PROT_NONE
    VMA) at the bottom of the stack to reliably catch stack
    overflows; and you probably don't want to do that with
    malloc, in particular with non-page-aligned allocations.

And the OpenBSD 6.5 manual pages says:

    MAP_STACK
        Indicate that the mapping is used as a stack. This
        flag must be used in combination with MAP_ANON and
        MAP_PRIVATE.

And I then noticed that MAP_STACK seems already to be on
FreeBSD for a long time:

    MAP_STACK
        Map the area as a stack.  MAP_ANON is implied.
        Offset should be 0, fd must be -1, and prot should
        include at least PROT_READ and PROT_WRITE.  This
        option creates a memory region that grows to at
        most len bytes in size, starting from the stack
        top and growing down.  The stack top is the start‐
        ing address returned by the call, plus len bytes.
        The bottom of the stack at maximum growth is the
        starting address returned by the call.

        The entire area is reserved from the point of view
        of other mmap() calls, even if not faulted in yet.

Reported-by: Jann Horn <jannh@google.com>
Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

strftime.3: wfix

Fix comma splice.

Signed-off-by: Jakub Wilk <jwilk@jwilk.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Minor tweaks after feedback from Christian Brauner

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: tfix

Remove duplicated word.

Signed-off-by: Jakub Wilk <jwilk@jwilk.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Tidy up the description of CLONE_DETACHED

The obsolete CLONE_DETACHED flag has never been properly
documented, but now the discussion CLONE_PIDFD also requires
mention of CLONE_DETACHED. So, properly document CLONE_DETACHED,
and mention its interactions with CLONE_PIDFD.

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Consistently order paragraphs for CLONE_NEW* flags

Sometimes the descriptions of these flags mentioned the
corresponding section 7 namespace manual page and then the
required capabilities, and sometimes the order was the was
the reverse. Make it consistent.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Remove various details that are already covered in namespaces pages

Remove details of UTS, IPC, and network namespaces that are
already covered in the corresponding namespaces pages in
section 7. This change is for consistency, since corresponding
details were not provided for other namespace types in clone(2)
and these details do not appear in unshare(2).

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

uts_namespaces.7: Add a little more detail on scope of UTS namespaces

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Remove wording that suggests CLONE_NEW* flags are for containers

These flags are used for implementing many other interesting
things by now.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Adjustments to clone3() text as well as some other details in the page

After feedback from Christian Brauner [1], I've adjusted a few pieces
of the clone3() text, and also adjusted some of the older text in
the page.

[1] https://lore.kernel.org/linux-man/20191107151941.dw4gtul5lrtax4se@wittgenstein/

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Give the introductory paragraph a new coat of paint

Change the text in the introductory paragraph (which was written
20 years ago) to reflect the fact that clone*() does more things
nowadays.

Cowritten-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

resolv.conf.5: srcfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

resolv.conf.5: Attempt clarify domain/search interaction

The domain directive is historic at this point; it should not
be used.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

signal-safety.7: ffix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

ioctl_iflags.2: Emphasize that FS_IOC_GETFLAGS and FS_IOC_SETFLAGS argument is 'int *'

Reported-by: Robert Edmonds <edmonds@debian.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

ioctl_list.2: Add reference to ioctl(2) SEE ALSO section

The referenced section lists various pages that document ioctls.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

netdevice.7: Small wording fix in description of SIOCGIFCONF

SIOCGIFCONF returns "network layer" addresses (not "transport
layer").

Reported-by: Silviu Popescu <silviupopescu1990@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fallocate.2: Add gfs2 to the list of punch hole-capable filesystems

Also remove a stray " from the previous item.

Signed-off-by: Andrew Price <anprice@redhat.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: ffix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2, proc.5: Adjust references to namespaces(7)

Adjust references to namespaces(7) to be references to pages
describing specific namespace types.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

quotactl.2: wfix: consistently use 'operation', rather than 'command'

A mix of the two words was being used, with 'operation' being
more common.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

quotactl.2: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

quotactl.2: Minor tweaks to Yang Xu's patch

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

quotactl.2: Add some details about Q_QUOTAON

For Q_QUOTAON, on old kernel we can use quotacheck -ug to generate
quota files. But in current kernel, we can also hide them in
system inodes and indicate them by using "quota" or project
feature.

For user or group quota, we can do as below (etc ext4):

mkfs.ext4 -F -o quota /dev/sda5
mount /dev/sda5 /mnt
quotactl(QCMD(Q_QUOTAON, USRQUOTA), /dev/sda5, QFMT_VFS_V0, NULL);

For project quota, we can do as below (etc ext4):

mkfs.ext4 -F -o quota,project /dev/sda5
mount /dev/sda5 /mnt
quotactl(QCMD(Q_QUOTAON, PRJQUOTA), /dev/sda5, QFMT_VFS_V0, NULL);

Reported-by: Jan Kara <jack@suse.cz>
Signed-off-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

copy_file_range.2: tfix

Signed-off-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Include clone3 in NAME section.

Signed-off-by: Jakub Wilk <jwilk@jwilk.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

netlink.7: tfix

Signed-off-by: Jakub Wilk <jwilk@jwilk.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

unix.7: tfix

In the given example, the second recvmsg(2) call should receive four bytes,
as the third sendmsg(2) call only sends four.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

unix.7: tfix

Signed-off-by: Torin Carey <torin@tcarey.uk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

cgroups.7: Switch to "considerate language"

Thanks-to: https://twitter.com/expensivestevie
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

seccomp.2: Switch to "considerate language"

Thanks-to: https://twitter.com/expensivestevie
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Introduce "flags mask" as a generic term for clone()/clone3()

Use "flags mask" as a generic term to refer to the clone()
'flags' argument and the clone3() 'cl_args.flags' field.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Minor improvements following clone3() additions

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: srcfix: update copyright

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone3.2: New link to clone(2)

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Document clone3()

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: wfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Minor change: move a paragraph from DESCRIPTION to NOTES

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Add some subsection headings

Again, in preparation for adding clone3() documentation.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Rename arguments for consistency with clone3()

Sometime soon, we'll have to add documentation of clone3() to this
page. As a preparatorys step, make the names of the clone()
arguments the same as the fields in the clone3() 'args' struct:

    ctid        ==> child_pid
    ptid        ==> parent_tid
    newtls      ==> tld
    child_stack ==> stack

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

abort.3: wfix

Reported-by: Mikael Magnusson <mikachu@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mount.2, pidfd_open.2, fuse.4: Minor fix: s/file system/filesystem/

Reported-by: Marko Myllynen <myllynen@redhat.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

wait.2: Clarify semantics of waitpid(0, ...)

As noted in kernel commit 821cc7b0b205c0df64cce59aacc330af251fa8f7,
threads create an ambiguity: what if the calling process's PGID
is changed by another thread while waitpid(0, ...) is blocked?
So, clarify that waitpid(0, ...) means wait for children whose
PGID matches the caller's PGID at the time of the call to
waitpid().

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

wait.2: waitid() can be used to wait on children in same process group as caller

Since Linux 5.4, idtype == P_PGID && id == 0 can be used to wait
on children in same process group as caller.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

wait.2: Add P_PIDFD for waiting on a child referred to by a PID file descriptor

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Note the waitid() use case for PID file descriptors

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Minor fix: add some structure to text on use cases

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Add a subsection header "Use cases for PID file descriptors"

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Make it a little more explicit the CLONE_PIDFD returns a PID FD

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Minor wording improvement

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Remove a redundant sentence

clone() CLONE_PIDFD is already mentioned elsewhere in NOTES.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

getauxval.3: Clarify that AT_BASE_PLATFORM and AT_EXECFN return pointers to strings

See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=942207

Reported-by: Witold Baryluk <witold.baryluk@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

select.2: POLLIN_SET/POLLOUT_SET/POLLEX_SET are now defined in terms of EPOLL*

Since kernel commit a9a08845e9acbd224e4ee466f5c1275ed50054e8, the
equivalence between select() and poll()/epoll is defined in terms
of the EPOLL* constants, rather than the POLL* constants.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mount_namespaces.7: tfix

Remove duplicated word.

Signed-off-by: Jakub Wilk <jwilk@jwilk.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

strptime.3: tfix

Escape hyphens.

Signed-off-by: Jakub Wilk <jwilk@jwilk.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Ready for 5.04

Changes.old: Fix typos in 5.03 change log

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Ready for 5.04

Start of man-pages-5.04: updating Changes and Changes.old

Start of man-pages-5.04: updating .Announce and .lsm files

Start of man-pages-5.04: renaming .Announce and .lsm files

Ready for 5.03

localedef.1, close.2, copy_file_range.2, execve.2, get_robust_list.2, getdomainname.2, gethostname.2, inotify_add_watch.2, io_submit.2, ioctl_fideduperange.2, kcmp.2, kill.2, mmap.2, move_pages.2, perf_event_open.2, ptrace.2, rt_sigqueueinfo.2, sched_setaffinity.2, sched_setparam.2, setns.2, sigaction.2, signalfd.2, statx.2, syscall.2, syscalls.2, uname.2, write.2, errno.3, fexecve.3, getauxval.3, printf.3, pthread_mutex_consistent.3, pthread_mutexattr_init.3, pthread_mutexattr_setrobust.3, pthread_setcancelstate.3, regex.3, strtok.3, strtol.3, ttyname.3, smartpqi.4, core.5, resolv.conf.5, man-pages.7, mq_overview.7, operator.7, pthreads.7, signal-safety.7, sysvipc.7: Update timestamp

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Changes: Ready for 5.03

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Further enhancements to fork() + pidfd_open() text

Christian noted that SA_NOCLDWAIT also matters in this scenario.

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: wfix

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: tfix

Reported-by: Florian Weimer <fw@deneb.enyo.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Enhance the discussion of usage of fork() + pidfd_open()

After review comments from Christian and Daniel.

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Reported-by: Daniel Colascione <dancol@google.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Explain how pidfd_open() can be used to with fork()

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_send_signal.2: Minor wording improvement

Reported-by: Daniel Colascione <dancol@google.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_send_signal.2: tfix

Reported-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_send_signal.2: Fixes after review comments from Florian Weimer

Reported-by: Florian Weimer <fw@deneb.enyo.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_send_signal.2: wfix

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Add some missing errors

Reported-by: Florian Weimer <fw@deneb.enyo.de>
Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Improve description in example

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Add a comment on system call number in example code

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: read(2) of a PID file descriptor fails with EINVAL

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: wfix

Reported-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: wfix

Reported-by: Florian Weimer <fw@deneb.enyo.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: Add <sys/types.h> to SYNOPSIS

Reported-by: Florian Weimer <fw@deneb.enyo.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_send_signal.2: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pidfd_open.2: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: SEE ALSO: add pidfd_open(2)

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>