clone.2: Document raw syscall interfaces on various other architectures

Reported-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

clone.2: Change types for 'ptid' and 'ctid' in syscall prototypes

These types changed from 'void *' to 'int *' back in Linux 3.8.
The new types are closer to reality, so just update the page
without discussing the history.

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

clone.2: Adjust syscall prototype and expand CLONE_SETTLS description

The prototype for the system call was added in kernel commit
81f10dad, but looking at the kernel's fork.c, I believe the
relevant definition is

SYSCALL_DEFINE5(clone, unsigned long, clone_flags,
                unsigned long, newsp,
                int __user *, parent_tidptr,
                int __user *, child_tidptr,
                unsigned long, tls)

so the last argument is the tls argument, not a pt_regs argument.
I stumbled upon this while trying to understand CLONE_SETTLS, so
I expanded that description a little to cover other architectures.

Reviewed-by: Josh Triplett <josh@joshtriplett.org>

pipe.7: Document FIONREAD

Reported-by: Patrick McLean <patrickm@gaikai.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

namespaces.7: tfix

Reported-by: Nikola Forró <nforro@redhat.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

execve.2: Note that real UID, real GID, and supplementary GIDs are unchanged

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

fork.2: ffix

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

fork.2: PID of new process also does not match any existing session ID

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

quotactl.2: wfix

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

quotactl.2: ffix

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

quotactl.2: Tweaks to Eugene Syromyatnikov's patches

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

quotactl.2: wfix

quotactl.2: Updated information regarding XFS-specific quotactl subcommands

Added information regarding structure definitions used for
XFS-specific subcommands, updated flag constants, added
information regarding ignored syscall arguments, added notes on
usage of kernel UAPI header.

quotactl.2: Updated information regarding disk quota flags

Added information regarding DQF_SYS_FILE flag; updated definition
of V1_DQF_RSQUASH, which has been defined privately and defined
publicly as DQF_ROOT_SQUASH.

quotactl.2: Additions regarding project quotas

Added information regarding presence of project quotas.

dlopen.3: dlmopen() is still broken in glibc 2.24

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

prctl.2: Simplify list of cases where "dumpable" attribute is reset

Simplify list of cases that cause dumpable attribute to
reset to suid_dumpable: there were two lists that together
in effect had many duplicates.

Also some minor rewordings.

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

prctl.2: Refer to proc(5) for effects of dumpability on ownership of /proc/PID/*

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

prctl.2: Minor fix: add a reference to ptrace(2) for PR_SET_DUMPABLE

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

proc.5: srcfix

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

proc.5: Note that 'suid_dumpable' mode 1 is insecure

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

proc.5: Add reference to core(5) in discussion of 'suid_dumpable'

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

proc.5: Refer to ptrace(2) for info on effect of suid_dumpable on ptraceability

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

proc.5: Explain rules determining ownership of /proc/PID/* files

Describe the effect of the "dumpable" attribute on ownership
of /proc/PID files.

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

proc.5: wfix

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

proc.5: Note effect of 'suid_dumpable' on ownership of /proc/PID files

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

capabilities.7: SEE ALSO: add proc(5)

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

epoll_wait.2: Clarify that the timeout is measured against CLOCK_MONOTONIC

The existing page left the choice of clock as ambiguous. My
reading of the kernel implementation is that CLOCK_MONOTONIC is
always used since ep_poll() calls schedule_hrtimeout_range(),
which calls schedule_hrtimeout_range_clock() passing
CLOCK_MONOTONIC.

References:
 http://lxr.free-electrons.com/source/fs/eventpoll.c?v=4.7#L1614
 http://lxr.free-electrons.com/source/kernel/time/hrtimer.c?v=4.7#L1785

The "measured against" terminology was borrowed from nanosleep.2.

I've checked all the way back to Linux v2.6.12 where the
calculation was done using jiffies so I think that this has
effectively always been true.

Signed-off-by: Mike Crowe <mac@mcrowe.com>

futex.2: srcfix

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

bpf.2: srcfix

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

clone.2: tfix

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

lirc.4, proc.5, netlink.7: tfix

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

proc.5: tfix

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

proc.5: tfix

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

execve.2: tfix

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

prctl.2: ERRORS: Add EACCES error for PR_SET_SECCOMP-SECCOMP_MODE_FILTER

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

clone.2: EINVAL is generated by glibc wrapper for NULL 'fn' or 'child_stack'

Clarify that this error is produced by the wrapper function, not
the underlying system call. In particular, the point is that the
raw system call can accommodate a NULL pointer for 'child_stack'.

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

ld.so.8: Clarify text describing whether secure-mode programs preload libraries

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

proc.5: tfix

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

proc.5: Minor tweaks to Namhyung Kim's patch

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

proc.5: Add description of CLEAR_REFS_MM_HIWATER_RSS

The Linux kernel commit 695f05593693 ("fs/proc/task_mmu.c: add
user-space support for resetting mm->hiwater_rss (peak RSS)") added a
way to reset peak RSS of a process but missed to update manpage.

Cc: Petr Cermak <petrcermak@chromium.org>
Acked-by: Petr Cermak <petrcermak@chromium.org>
Signed-off-by: Namhyung Kim <namhyung@gmail.com>

raw.7: Clarify user namespace requirements for CAP_NET_RAW

Also remove mention of UID 0 as a method or creating
a raw socket. As far as I can tell from reading the
kernel source (net/ipv4/af_inet.c), this is not true.

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

packet.7: Clarify user namespace requirements for CAP_NET_RAW

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

open.2: Clarify user namespace capability requirements for O_NOATIME

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

proc.5: Clarify user namespace requirements for /proc/sys/fs/protected_hardlinks

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

bindresvport.3, rcmd.3, ip.7: Note user namespace requirements for CAP_NET_BIND_SERVICE

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

ip.7: wfix

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

msgctl.2, msgget.2, msgop.2, semctl.2, semget.2, semop.2, shmctl.2, shmget.2, shmop.2: Note the user namespace requirements for CAP_IPC_OWNER

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

chroot.2: Note user namespace requirements for CAP_SYS_CHROOT

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

setuid.2: Note user namespace requirements for CAP_SETUID

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

setreuid.2: Note user namespace requirements for CAP_SETUID and CAP_SETGID

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

setresuid.2: Note user namespace requirements for CAP_SETUID

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

setgid.2: Note user namespace requirements for CAP_SETGID

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

setgid.2: ffix

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

seteuid.2: wfix

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

seteuid.2: Note user namespace requirements for CAP_SETUID and CAP_SETGID

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

setuid.2: wfix

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

setresuid.2: wfix

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

setgid.2: wfix

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

kill.2: Minor wording fix

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

kill.2: wfix

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

seccomp.2: CAP_SYS_ADMIN is required only in caller's user namespace

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

getrlimit.2: Note user namespace semantics for CAP_SYS_RESOURCE

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

getgroups.2: Note user namespace requirements for CAP_SETGID

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

getdomainname.2: Note user namespace requirements for CAP_SYS_ADMIN

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

gethostname.2: Note user namespace requirements for CAP_SYS_ADMIN

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

prctl.2: Note user namespace requirements for PR_CAPBSET_DROP CAP_SETPCAP

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

reboot.2: Note user namespace requirements around CAP_SYS_BOOT

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

sched_setaffinity.2: Note user namespace requirements for CAP_SYS_NICE

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

kill.2: tfix

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

kill.2: srcfix

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

kill.2: tfix

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

kill.2: Note the user namespace requirement for CAP_KILL

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

killpg.2: Refer reader to kill(2) for signal permission rules

Rather than repeating details here, refer the reader to kill(2)
(so that the rules are in a canonical location, and need only
be edited in one place for future changes--see next commit).

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

fcntl.2: Note an important detail of F_SETOWN permission rules for signals

F_SETOWN records the caller's credentials at the time of
the fcntl() call, and it is these saved credentials that
are used for subsequent permission checks.

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

socket.7: SIOCSPGRP: refer to fcntl(2) F_SETOWN for correct permission rules

The permission rules described for SIOCCPGRP are wrong. Rather
than repeat the rules here, just refer the reader to fcntl(2),
where the rules are described for F_SETOWN.

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

close.2: Add mention of the close-on-exec flag

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

umask.2: tfix

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

open.2: Clarify the rules about how the group ID of a new file is determined

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

open.2: wfix

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

open.2: ffix

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

getsid.2: Rework description to be somewhat clearer

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

setsid.2: Refer to credentials(7) for details for details on controlling terminal

Refer to credentials(7) for details of how a session obtains
a controlling terminal.

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

getsid.2: wfix

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

getsid.2: Correct the definition of "session ID"

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

setpgid.2: Minor wording fix

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

setsid.2: Minor wording fixes

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

setpgid.2, setsid.2: Relocate some text on sessions and sessions leaders

Some text that was in setpgid(2) is better placed in setsid(2).

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

setpgid.2: Add a reference to credentials(7)

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

credentials.7: SEE ALSO: add setsid(2) and setpgid(2)

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

execve.2: Mention use of 'environ' to access environment list

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

execve.2: Minor clarification

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

execve.2: wfix

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

execve.2: Minor clarification

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

execve.2: ffix

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

execve.2: Minor clarification

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

fcntl.2: Make the description of the effect of close-on-exec a little clearer

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

close.2: Clarify discussion noting that close() does not flush buffer cache

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

fcntl.2: Clarify that F_GETFD and F_GETFL return flags via the function result

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

ld.so.8: Remove mention of "ELF only"

Drawing a distinction between ELF-only features versus a,out
ceased to be relevant long ago, so cluttering the page
with "ELF-only" serves no purpose.

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