getdirentries.3: tfix

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

drand48.3: tfix (Oxford comma)

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

dl_iterate_phdr.3: wsppfix

Reported-by: Alejandro Colomar (man-pages) <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

perfmonctl.2, pidfd_send_signal.2: ffix: Ensure statement about no wrapper function is a new paragraph

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

copy_file_range.2: Document glibc wrapper instead of kernel syscall

Glibc uses 'off64_t' instead of 'loff_t'.

......

Glibc prototype:

$ syscall='copy_file_range';
$ ret='ssize_t';
$ find glibc/ -type f -name '*.h' \
  |xargs pcregrep -Mn "(?s)^[\w\s]*${ret}\s*${syscall}\s*\(.*?;";
glibc/posix/unistd.h:1121:
ssize_t copy_file_range (int __infd, __off64_t *__pinoff,
 int __outfd, __off64_t *__poutoff,
 size_t __length, unsigned int __flags);

......

Testing example:

$ man ./man2/copy_file_range.2 \
  |sed -n '/^EXAMPLES/,/^SEE ALSO/p' \
  |head -n -1 \
  |tail -n +2 \
  >copy_file_range.c
$ gcc -Wall -Wextra -Werror -pedantic
copy_file_range.c -o copy_file_range
$ ./copy_file_range
Usage: ./copy_file_range <source> <destination>
$ tee a >/dev/null
asdf
$ tee b >/dev/null
qwerty
zxcvbn
$ ./copy_file_range a b
$ cat a
asdf
$ cat b
asdf

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

wait4.2, endian.3, isalpha.3, unlocked_stdio.3: Minor consistency fix for Feature Test Macro info

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

strsignal.3: ffix

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

Various pages: s/Glibc versions </Glibc </ in Feature Test Macro discussions

Use a more compact notation that will render less wide
on the terminal.

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

pread.2, rename.2, syscall.2, utimensat.2, gethostid.3, rcmd.3, seekdir.3, setnetgrent.3, strerror.3, strsignal.3: ffix: Various tweaks for greater consistency in Feature Test Macros

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

chown.2, ffs.3, fpclassify.3, gamma.3, getgrent.3, ilogb.3, strdup.3: ffix

Bring a bit more consistency to Feature Test Macro information
(mainly .PP between differnt FTM lists).

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

Changes.old: tfix

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

Various pages: Formatting fix

Use .PP (which gives a bit of vertical white space) rather than
.br to separate functions in Feature Test Macro requirement lists.

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

Various pages: srcfix: remove redundant .br

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

recv.2: srcfix

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

get_phys_pages.3: glibc gets the info from sysinfo(2) since 2.23

See glibc's commit: 0ce657c576bf1b24
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=0ce657c576bf1b24

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

keyctl.2: SYNOPSIS: Fix prototype parameter types

The Linux kernel uses 'unsigned long'.
There's no reason to use the typedef '__kernel_ulong_t'.

......

$ syscall='keyctl';
$ find linux/ -type f -name '*.c' \
  |xargs pcregrep -Mn "(?s)^[\w_]*SYSCALL_DEFINE.\(${syscall},.*?\)";
linux/security/keys/keyctl.c:1869:
SYSCALL_DEFINE5(keyctl, int, option, unsigned long, arg2, unsigned long, arg3,
unsigned long, arg4, unsigned long, arg5)
linux/security/keys/compat.c:17:
COMPAT_SYSCALL_DEFINE5(keyctl, u32, option,
       u32, arg2, u32, arg3, u32, arg4, u32, arg5)

Cc: Eugene Syromyatnikov <evgsyr@gmail.com>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

delete_module.2: SYNOPSIS: Fix prototype parameter types

The Linux kernel uses 'unsigned int' instead of 'int' for the
'flags' parameter.  As glibc provides no wrapper, use the same
type the kernel uses.

......

$ syscall='delete_module';
$ find linux/ -type f -name '*.c' \
  |xargs pcregrep -Mn "(?s)^[\w_]*SYSCALL_DEFINE.\(${syscall},.*?\)";
linux/kernel/module.c:977:
SYSCALL_DEFINE2(delete_module, const char __user *, name_user,
unsigned int, flags)

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

kexec_load.2: wfix

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

Various pages: Normalize NOTES about nonexistent glibc wrappers

See previous commit.

This commit normalizes texts under sections other than SYNOPSIS
(most of them in NOTES).

Signed-off-by: Ganimedes Colomar <gacoan.linux@gmail.com>
Cowritten-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Various pages: Normalize SYNOPSIS notes about nonexistent glibc wrappers

To easily distinguish documentation about glibc wrappers from
documentation about kernel syscalls, let's have a normalized
'Note' in the SYNOPSIS, and a further explanation in the page body
(NOTES in most of them), as already happened in many (but not all)
of the manual pages for syscalls without a wrapper.  Furthermore,
let's normalize the messages, following membarrier.2 (because it's
already quite extended), so that it's easy to use grep to find
those pages.

To find these pages, we used:
$ grep -rn wrapper man? | sort -V
and
$ grep -rni support.*glibc | sort -V

delete_module.2, init_module.2:  glibc 2.23 is no longer
  maintained, so we changed the notes about wrappers, to say that
  there are no glibc wrappers for these system calls; see NOTES.

We didn't fix some obsolete pages such as create_module.2.

Signed-off-by: Ganimedes Colomar <gacoan.linux@gmail.com>
Cowritten-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

syscalls.2: Add epoll_pwait2()

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

readlink.2: Emphasize that the returned buffer is not null-terminated

Reported-by: Jonny Grant <jg@jguk.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

filesystems.5: tfix

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

ftime.3: tfix

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

cgroups.7: tfix

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

signal.7: tfix

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

pkeys.7: srcfix: remove redundant .PP

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

user_namespaces.7: srcfix: remove redundant .PP after .SS

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

userfaultfd.2, posix_spawn.3: srcfix: remove redundant .PP

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

pthread_mutexattr_setrobust.3: wfix: remove redundant comments

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

bind.2: wfix: remove redundant comment

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

spu_run.2: Minor fix-ups for comments

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

fanotify.7, inotify.7: wfix: Tweak location of tag comments

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

ioctl_ns.2, keyctl.2, memfd_create.2, pidfd_send_signal.2, poll.2, seccomp.2, signalfd.2, timerfd_create.2, unshare.2, dlinfo.3, duplocale.3, fopencookie.3, getgrouplist.3, getifaddrs.3, getprotoent_r.3, getservent_r.3, insque.3, printf.3, pthread_attr_init.3, pthread_cancel.3, pthread_create.3, pthread_mutexattr_setrobust.3, pthread_setaffinity_np.3, qsort.3, sem_wait.3, core.5, elf.5, aio.7, epoll.7, pkeys.7, system_data_types.7, unix.7: Use periods more consistently inside code comments

In general, complete sentences in free-standing comments
should be terminated by periods.

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

bind.2, clone.2, dup.2, futex.2, mprotect.2, open_by_handle_at.2, pivot_root.2, select_tut.2, shmop.2, timer_create.2, userfaultfd.2, getaddrinfo.3, getaddrinfo_a.3, inet_net_pton.3, malloc_info.3, mbstowcs.3, newlocale.3, posix_spawn.3, pthread_sigmask.3, shm_open.3, strtol.3, inotify.7, user_namespaces.7: Add periods at end of sentences in comments

Try to bring a little more consistency on the use of periods
in comments in example code.

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

man-pages.7: Add some notes on comments in example code

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

slist.3: tfix

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

stailq.3: ffix

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

list.3, slist.3, stailq.3, tailq.3: tfix

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

seccomp.2, circleq.3, list.3, malloc_hook.3, slist.3, stailq.3, tailq.3, ip.7, unix.7: tfix

Remove "." at the end of sentence fragments/short single sentences
in comments.

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

system_data_types.7: tfix

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

Various pages: Remove explicit blank lines, replacing them with .PP when appropriate

Rationale:
$ man 7 man-pages 2>/dev/null | sed -n /Paragraphs/,/^$/p
       Paragraphs should be separated by suitable markers (usually
       either .PP or .IP).  Do not separate paragraphs using blank
       lines,  as  this  results  in poor rendering in some output
       formats (such as PostScript and PDF).

Fix:
$ sed -i -e '1,/^\.EX/s/^$/.PP/' -e '/^\.EE/,/^\.EX/s/^$/.PP/' man?/*

And then some manual adjustments.

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

syscall.2: Update superh syscall convention

I noticed this while working on some silly "hello, world"
programs, see https://git.sr.ht/~phf/hello-again if you're
curious.  Disassembling sh4 code showed trap #31 all over the
place but the syscall(2) man page talked about trap #0x17 and
friends. Checking the kernel sources I got lucky in
arch/sh/kernel/entry-common.S where in commit 3623d138213ae Rich
Felker clarifies the situation.

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

mallinfo2.3: New link to mallinfo(3)

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

mallinfo.3: Update example program to use mallinfo2()

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

mallinfo.3: Document mallinfo2() and note that mallinfo() is deprecated

Document the mallinfo2() function added in glibc 2.33.

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

ld.so.8: srcfix

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

ld.so.8: Document the --argv0 option added in glibc 2.33

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

ftime.3: glibc 2.33 has removed ftime()

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

getrusage.2: Starting in 2.33, glibc no longer provides vtimes()

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

pthread_attr_setsigmask_np.3: tfix

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

nl_langinfo.3: tfix

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

isfdtype.3: tfix

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

aio_suspend.3: tfix

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

setns.2: tfix

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

perf_event_open.2: wfix

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

statfs.2: tfix (Section number)

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

Changes.old: Minor fixes to 5.10 changelog

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

Start of man-pages-5.11: updating Changes and Changes.old

Start of man-pages-5.11: updating .Announce and .lsm files

Start of man-pages-5.11: renaming .Announce and .lsm files

Ready for 5.10

getent.1, access.2, cacheflush.2, chroot.2, clock_getres.2, fcntl.2, getrusage.2, io_cancel.2, io_destroy.2, io_getevents.2, io_setup.2, io_submit.2, link.2, llseek.2, mmap.2, mount.2, readv.2, restart_syscall.2, semctl.2, set_mempolicy.2, set_tid_address.2, shmctl.2, sigaction.2, sigaltstack.2, spu_create.2, statfs.2, subpage_prot.2, syscalls.2, timer_getoverrun.2, uselib.2, INFINITY.3, __ppc_set_ppr_med.3, bstring.3, btree.3, ctime.3, fgetc.3, fopen.3, getcontext.3, gethostbyname.3, getnetent.3, getprotoent.3, getservent.3, inet.3, j0.3, list.3, makecontext.3, nextafter.3, posix_memalign.3, profil.3, pthread_tryjoin_np.3, puts.3, rcmd.3, resolver.3, rtime.3, sigsetops.3, strnlen.3, tailq.3, elf.5, filesystems.5, nscd.conf.5, proc.5, utmp.5, attributes.7, ipv6.7, packet.7, pthreads.7, signal-safety.7, signal.7, spufs.7, system_data_types.7, tcp.7: tstamp

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

Changes: Ready for 5.10

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

rcmd.3: SEE ALSO: remove intro(2)

There's no reason to link to intro(2)...

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

cacheflush.2: ffix

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

cacheflush.2: Document Architecture-specific variants

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

cacheflush.2: Document __builtin___clear_cache() as a more portable alternative

Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Cc: Martin Sebor <msebor@redhat.com>
Cc: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

packet.7: Update references to kernel documentation

Files moved from .txt to .rst.

Also, drop / prefix from kernel source tree references.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

filesystems.5: wfix: ntfs: remove FAT comparison

The main point I was driving at with this patch was to fix
"Microsoft Window's FAT filesystems" (i.e., FAT filesystems which
belong to Microsoft Windows, which is decidedly wrong).

FAT32 first shipped with MS-DOS 7.1, as part of Windows 95 OSR2,
but it's a (relatively) simple logical extension of the previous
FATx filesystems (16 and 12 as we know and love them today, I
don't think the PC ever saw 8), hence the "VFAT" driver name ‒
calling FAT-anything a Windows filesystem would be a flat-out lie,
calling it a Microsoft filesystem would be, uh, facetious.

NTFS (as part of Windows NT), on the other hand, is wholly
different WRT the scope and feature-set (it does borrow some
layouting from FAT, but reading NTFS as FAT doesn't get you very
far, or much).

The replacing bit is also questionable, especially in a.d. 2020:
while it is true that you cannot install NT on FAT (after a
certain point? my memory ain't what it used to be), and must
therefore replace your existing FAT partitions with NTFS during
upgrades; Windows NT 4.0, the last product to be NT-branded came
out in 1996, i.e. you could not install Windows on FAT (and,
therefore, upgrade it to NTFS, replacing it) during my entire
lifetime.

Indeed, in $(date +%Y) we live in a post-NTFS world ‒ putting NTFS
in the same class as FAT beyond "is a filesystem" is a joke.

Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

filesystems.5: Note ncpfs removal from kernel

Relevant Linux commits:
 * moved to staging in 1bb8155080c652c4853e6228f8f0d262b3049699
   (describe: v4.15-rc1-129-g1bb8155080c6) in Nov 2017,
   described as "broken" and "obsolete"
 * purged in bd32895c750bcd2b511bf93917bf7ae723e3d0b6
   (describe: v4.17-rc3-1010-gbd32895c750b) in Jun 2018,
   "since no one has complained or even noticed it was gone"

Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

filesystems.5: Fix link to user space tooling for ncpfs

Corresponds to Linux commit 1b83df308f69a5a3cc59be03bd7fb23e4bcebd8e

Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

pthreads.7: ffix

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

pthreads.7: Rephrase function list in terms of POSIX rather than SUS

The list was using an inconsistent mixture of "POSIX" and "SUS".

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

regex.3: ffix

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

pthread_attr_setguardsize.3: ffix

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

pthread_attr_setstacksize.3: ffix

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

pthread_mutexattr_setrobust.3: ffix

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

pthread_attr_setstackaddr.3: ffix

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

pthread_attr_setdetachstate.3: ffix

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

setns.2: ffix

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

elf.5: ffix

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

perf_event_open.2: tfix

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

mount_namespaces.7: ffix

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

strtol.3: tfix

Signed-off-by: John A. Leuenhagen <john@zlima12.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

elf.5: srcfix

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

elf.5: SEE ALSO: add objcopy(1)

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

tcp.7: tcp_syncookies: It is now an integer [0, 2]

Since Linux kernel 3.12, tcp_syncookies can have the value 2,
which sends out cookies unconditionally.

Related kernel commits:
5ad37d5deee1ff7150a2d0602370101de158ad86
d8513df2598e5142f8a5c4724f28411936e1dfc7

Reported-by: Philip Rowlands <linux-kernel@dimebar.com>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

modf.3: tfix

See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=976312

Reported-by: Bill Allombert <ballombe@debian.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

keyctl.2: tfix

Fixes: fa76da808eff
Signed-off-by: Arusekk <arek_koz@o2.pl>
Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

strnlen.3: Fix a small inconsistency in the text

Quoting Heinrich:

    The strnlen.3 manpage has the following sentence:

    "In doing this, strnlen() looks only at the first maxlen
    characters in the string pointed to by s and never beyond
    s+maxlen."

    This sentence is self-contradictory:

    The last visited character implied by "first maxlen
    characters" is s[maxlen-1].

    Given that "beyond a" does not include "a", the last visited
    character implied by "never beyond s+maxlen" is s[maxlen].

    A consistent sentence would be

    "In doing this, strnlen() looks only at the first maxlen
    characters in the string pointed to by s and never beyond
    s+maxlen-1."

    I would prefer

    "In doing this, strnlen() looks only at the first maxlen
    characters in the string pointed to by s and never beyond
    s[maxlen-1]"

Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

subpage_prot.2: SYNOPSIS: Fix return type: s/long/int/

The Linux kernel uses 'int' instead of 'long' for the return type.
As glibc provides no wrapper, use the same type the kernel uses.

......

$ grep -n wrapper man-pages/man2/subpage_prot.2
40:There is no glibc wrapper for this system call; see NOTES.
99:Glibc does not provide a wrapper for this system call; call it using

$ grep -rn SYSCALL_DEFINE.*subpage_prot linux/;
linux/arch/powerpc/mm/book3s64/subpage_prot.c:190:
SYSCALL_DEFINE3(subpage_prot, unsigned long, addr,

$ sed -n /SYSCALL.*subpage_prot/,/^}/p \
  linux/arch/powerpc/mm/book3s64/subpage_prot.c \
  |grep return;
return -ENOENT;
return -EINVAL;
return -EINVAL;
return 0;
return -EFAULT;
return -EFAULT;
return err;

$ sed -n /SYSCALL.*subpage_prot/,/^}/p \
  linux/arch/powerpc/mm/book3s64/subpage_prot.c \
  |grep '\<err\>';
int err;
err = -ENOMEM;
err = -ENOMEM;
err = 0;
return err;

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

spu_create.2: Clarify that spu_create() now has 4 arguments but once had only 3

Reported-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

spu_create.2: Relocate paragraph on 'mode' argument

This paragraph is a little bit hidden at the end of DESCRIPTION;
make it a little more prominent.

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

spu_create.2: Add kernel version numbers for SPU_CREATE_AFFINITY_SPU/_MEM

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

spu_run.2: srcfix

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

getpriority.2: wfix

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

getpriority.2: ffix

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

link.2: ffix

Reported-by: Alejandro Colomar (man-pages) <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

link.2: ERRORS: add ENOENT when target is deleted

Linux kernel commit aae8a97d3ec30788790d1720b71d76fd8eb44b73 (part
of kernel release v2.6.39) added a check to disallow creating a
hardlink to an unlinked file.

The manual page already describes the trick of using
AT_SYMLINK_FOLLOW as an alternative to AT_EMPTY_PATH, and for
AT_EMPTY_PATH the manual page already notes that it "will
generally not work if the file has a link count of zero". However,
the precise error (ENOENT) is not mentioned, and the error case
isn't mentioned in the ERRORS section at all.

This makes it easy to overlook the fact that the AT_SYMLINK_FOLLOW
trick on /proc/self/fd/NN won't work on deleted files, as
evidenced by the follow message (which turns up when googling
"linkat deleted ENOENT"):
https://groups.google.com/g/linux.kernel/c/zZO4lqqwp64

Signed-off-by: Mathias Rav <m@git.strova.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>