Merge branch 'mount_fuzz' of https://github.com/evverx/util-linux

* 'mount_fuzz' of https://github.com/evverx/util-linux:
  mount_fuzz: reject giant files early

Merge branch 'topic/hardlink'

tests: update hardlink --maximum-size

Signed-off-by: Karel Zak <kzak@redhat.com>

hardlink: add new option: -S/--maximum-size

Support limiting comparison by maximum file size.

[kzak@redhat.com: - resolve conflicts]

Signed-off-by: Daniele Pizzolli <dan+dev@toel.it>
Signed-off-by: Karel Zak <kzak@redhat.com>

hardlink: add --cache-size

Signed-off-by: Karel Zak <kzak@redhat.com>

lib/fileeq: fix for small memsiz

hardlink: rename --buffer-size to --io-size

Addresses: https://github.com/karelzak/util-linux/pull/1467
Signed-off-by: Karel Zak <kzak@redhat.com>

mount_fuzz: reject giant files early

It should help to address https://github.com/google/oss-fuzz/issues/6703

Signed-off-by: Evgeny Vereshchagin <evvers@ya.ru>

hardlink: rewrite files content comparison

Example:

 # hardlink --ignore-time --dry-run /var/lib

Old version, based on memcmp(), nothing cached:

 Mode:           dry-run
 Files:          93453
 Linked:         36172 files
 Compared:       0 xattrs
 Compared:       25755615 files
 Saved:          80.11 MiB
 Duration:       204.557185 seconds

New version, --method=memcmp, 32 bytes cached:

 Mode:           dry-run
 Method:         memcmp
 Files:          93453
 Linked:         36172 files
 Compared:       0 xattrs
 Compared:       25755615 files
 Saved:          80.11 MiB
 Duration:       5.248426 seconds

New version, zero-copy (Linux crypto API), sha1, crc32 and sha256:

 Mode:           dry-run
 Method:         sha1
 Files:          93453
 Linked:         36172 files
 Compared:       0 xattrs
 Compared:       25755615 files
 Saved:          80.11 MiB
 Duration:       2.870973 seconds

 Mode:           dry-run
 Method:         crc32
 Files:          93453
 Linked:         36172 files
 Compared:       0 xattrs
 Compared:       25755615 files
 Saved:          80.11 MiB
 Duration:       2.582651 seconds

 Mode:           dry-run
 Method:         sha256
 Files:          93453
 Linked:         36172 files
 Compared:       0 xattrs
 Compared:       25755615 files
 Saved:          80.11 MiB
 Duration:       3.152825 seconds

The default is sha256 to be robust..

Signed-off-by: Karel Zak <kzak@redhat.com>

include/fileeq: add functions to compare files content

Signed-off-by: Karel Zak <kzak@redhat.com>

chsh: fflush stdout before reading from stdin

Same problem as described in https://github.com/karelzak/util-linux/pull/1481

Signed-off-by: Lorenzo Beretta <vc.net.loreb@gmail.com>

chfn: flush stdout before reading stdin and fix uninitialized variable

Same problem as described in https://github.com/karelzak/util-linux/pull/1481

Signed-off-by: Lorenzo Beretta <vc.net.loreb@gmail.com>

Merge branch 'refactor-tests' of https://github.com/Ajchler/util-linux

* 'refactor-tests' of https://github.com/Ajchler/util-linux:
  tests: split additional tests into subtests
  tests: split several tests into subtests
  tests: split test into subtest
  tests: split cal/colorw test into subtests
  tests: split cal/color test into subtests

build-sys: disable libmount when missing mntent.h

Addresses: https://github.com/karelzak/util-linux/issues/1432
Signed-off-by: Karel Zak <kzak@redhat.com>

build-sys: include xlocale.h for locale_t on MacOS

Addresses: https://github.com/karelzak/util-linux/issues/1432
Signed-off-by: Karel Zak <kzak@redhat.com>

build-sys: (hardlink) check for llistxattr and lgetxattr

It seems check for sys/xattr.h is not enough. The header file exists
on MacOS, but without these functions.

Addresses: https://github.com/karelzak/util-linux/issues/1432
Signed-off-by: Karel Zak <kzak@redhat.com>

libbuid: use _UL_LIBUUID_UUID_H to cover uuid.h

It seems _UUID_UUID_H is too generic.

Addresses: https://github.com/karelzak/util-linux/issues/1432
Signed-off-by: Karel Zak <kzak@redhat.com>

docs: add hint for non-public reports

Fixes: https://github.com/karelzak/util-linux/issues/1482
Signed-off-by: Karel Zak <kzak@redhat.com>

Merge branch 'flush' of https://github.com/ericonr/util-linux

* 'flush' of https://github.com/ericonr/util-linux:
  vipw: flush stdout before getting answer.

agetty: use getttynam() if available

/etc/ttys seems to be a rather archaic concept that is not meant to
exist on Linux. Nevertheless it does. glibc has getttynam() which
correctly parses /etc/ttys. So let's give it a try before falling back
to the built in defaults. One can set the terminal type for a
specific tty using e.g.:

echo 'ttyS0 "" xterm' > /etc/ttys

[kzak@redhat.com: - improve configure.ac part
                  - log error on failed strdup()]

Signed-off-by: Karel Zak <kzak@redhat.com>

Merge branch 'uniformize-title-markups' of https://github.com/rffontenelle/util-linux

* 'uniformize-title-markups' of https://github.com/rffontenelle/util-linux:
  docs: Uniformize references to section titles

vipw: flush stdout before getting answer.

Otherwise the question is displayed only after the user presses Return,
and the program looks like it's hanging.

This happens at least on musl libc.

Reported by @loreb.

Signed-off-by: Érico Nogueira <erico.erc@gmail.com>

logger: fix --size use for stdin

The stdin version counts log header into the message size, but
for example when it reads message from argv[] it counts only message
itself.

 $ logger --stderr  --size 3 "abcd"
 <13>Oct 21 18:48:29 kzak: abc

 $ echo "abcd" | logger --stderr  --size 3
 logger: cannot allocate 18446744073709551597 bytes: Cannot allocate memory

Addresses: https://bugzilla.redhat.com/show_bug.cgi?id=2011602
Signed-off-by: Karel Zak <kzak@redhat.com>

Revert "logger: verify --size of large enough for message header"

This reverts commit 2fdea5a26480d46ec9445b0ea17fc984f84eecc9.

build-sys: use set +e before patch --try in ./autogen.sh

Addresses: https://github.com/karelzak/util-linux/issues/1476
Signed-off-by: Karel Zak <kzak@redhat.com>

logger: realloc buffer when header size changed

This is probably paranoid optimization, but when we generate a new
header we need to be sure that buffer is not smaller than calculated
maximal size of user's data.

Signed-off-by: Karel Zak <kzak@redhat.com>

logger: verify --size of large enough for message header

Addresses: https://bugzilla.redhat.com/show_bug.cgi?id=2011602
Signed-off-by: Karel Zak <kzak@redhat.com>

wdctl: update man page

Addresses: https://github.com/karelzak/util-linux/issues/1475
Signed-off-by: Karel Zak <kzak@redhat.com>

wdctl: add --setpregovernor

Addresses: https://github.com/karelzak/util-linux/issues/1475
Signed-off-by: Karel Zak <kzak@redhat.com>

wdctl: set_watchdog() refactoring

* keep wanted values in struct wd_control
* make set_watchdog() more readable

Signed-off-by: Karel Zak <kzak@redhat.com>

wdctl: print the current and available governors

 # wdctl
 ...
 Pre-timeout:    0 seconds
 Pre-timeout governor: noop
 Available pre-timeout governors: userspace noop panic
 ...

Addresses: https://github.com/karelzak/util-linux/issues/1475
Signed-off-by: Karel Zak <kzak@redhat.com>

docs: add links to adjtime_config manpage

Fixes: https://github.com/karelzak/util-linux/issues/1478
Signed-off-by: Karel Zak <kzak@redhat.com>

tests: split additional tests into subtests

build-sys: remove bashism

Addresses: https://github.com/karelzak/util-linux/issues/1476
Signed-off-by: Karel Zak <kzak@redhat.com>

tests: Fix test/misc/swaplabel failure due to change in mkswap behaviour.

mkswap now warns if the image file has holes. If fallocate is used to create the
file, use POSIX semantics to ensure the file has no holes.

This fixes the test failure

misc: swaplabel                      ... FAILED (misc/swaplabel)
========= script: /build/util-linux-2.37.2/tests/ts/misc/swaplabel =================
================= OUTPUT =====================
     1  Setting up swapspace version 1, size = 9 pages (9xPGSZ bytes)
     2  LABEL=1234567890abcde, UUID=12345678-abcd-abcd-abcd-1234567890ab
     3  LABEL: 1234567890abcde
     4  UUID:  12345678-abcd-abcd-abcd-1234567890ab
================= EXPECTED ===================
     1  Setting up swapspace version 1, size = 9 pages (9xPGSZ bytes)
     2  LABEL=1234567890abcde, UUID=12345678-abcd-abcd-abcd-1234567890ab
     3  LABEL: 1234567890abcde
     4  UUID:  12345678-abcd-abcd-abcd-1234567890ab
================= O/E diff ===================
==============================================

The additional error appears in swaplabel.err:

 mkswap: <swapfile> contains holes or other unsupported extents.
         This swap file can be rejected by kernel on swap activation!
         Use --verbose for more details.

[zeha@debian.org: this appears to be important/required when building
inside cowbuilder, on an ext3 filesystem.]

Signed-off-by: Mark Hindley <mark@hindley.org.uk>

build-sys: make libtool patching more robust

Fixes: https://github.com/karelzak/util-linux/issues/1476
Signed-off-by: Karel Zak <kzak@redhat.com>

wdctl: sysfs open refactoring

Let's make the sysfs handler usable in more functions.

Signed-off-by: Karel Zak <kzak@redhat.com>

wdctl: add --setpretimeout

Addresses: https://github.com/karelzak/util-linux/issues/1475
Signed-off-by: Karel Zak <kzak@redhat.com>

blockdev: remove accidental non-breaking spaces

commit 9147d2ad8abb73cea5799323fc73ccdaf675826f introduced these
C2A0 spaces, but without saying anything. Likely this was an
accident.

Bug-Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=996751
Signed-off-by: Chris Hofstaedtler <zeha@debian.org>

mcookie: fix infinite-loop when use -f

Signed-off-by: Karel Zak <kzak@redhat.com>

tests: split several tests into subtests

build-sys: patch libtool.m4 for darwin

Fixes: https://github.com/karelzak/util-linux/issues/1468
Signed-off-by: Karel Zak <kzak@redhat.com>

tests: (logger) check for socat

Addresses: https://github.com/karelzak/util-linux/issues/1466
Signed-off-by: Karel Zak <kzak@redhat.com>

su: reset RLIMIT_AS too

Fixes: https://github.com/karelzak/util-linux/issues/1465
Signed-off-by: Karel Zak <kzak@redhat.com>

docs: add note about GitHub PR

Signed-off-by: Karel Zak <kzak@redhat.com>

findmnt: (adoc) Added section stating exit code semantics

per issue #1464:

    https://github.com/karelzak/util-linux/issues/1464

[kzak@redhat.com: - move the text from DESCRIPTION to EXIT STATUS section]

Signed-off-by: Glenn Golden <gdg@zplane.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

hardlink: improve verbose messages

Signed-off-by: Karel Zak <kzak@redhat.com>

more: fix -e in non-interactive mode

Signed-off-by: Karel Zak <kzak@redhat.com>

tests: split test into subtest

su: reset also RLIMIT_FSIZE and RLIMIT_NOFILE

Addresses: https://github.com/linux-pam/linux-pam/issues/85
Signed-off-by: Karel Zak <kzak@redhat.com>

mount.8.adoc: note that mandatory locking is fully deprecated in Linux 5.15

Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Jeff Layton <jlayton@kernel.org>

fstrim: fix typo

Addresses: https://github.com/karelzak/util-linux/issues/1463
Signed-off-by: Karel Zak <kzak@redhat.com>

fstrim: don't trigger autofs

- ignore read-only entries
- ignore autofs entries (for example from /proc/self/mountinfo)
- ignore autofs mountpoints where automounter has not been triggered yet

Fixes: https://github.com/karelzak/util-linux/issues/1463
Signed-off-by: Karel Zak <kzak@redhat.com>

tests: split cal/colorw test into subtests

tests: split cal/color test into subtests

su: reset RLIMIT_{NICE,RTPRIO} to zero

Addresses: https://github.com/linux-pam/linux-pam/issues/85
Reported-by: Lennart Poettering <lennart@poettering.net>
Signed-off-by: Karel Zak <kzak@redhat.com>

prlimit: improve --help output

Signed-off-by: Karel Zak <kzak@redhat.com>

more: POSIX compliance patch preventing exit on EOF without -e

[kzak@redhat.com: - merge all patches to one commit,
                  - improve prompt for non-tty stdin]

Signed-off-by: Karel Zak <kzak@redhat.com>

fallocate: add verbose messages

Addresses: https://bugzilla.redhat.com/show_bug.cgi?id=2003927
Signed-off-by: Karel Zak <kzak@redhat.com>

hardlink: use more passive wording in hardlink.1

And fix "same (basename)".

Signed-off-by: Eduard Bloch <blade@debian.org>

Let user choose larger buffers for IO reading

Simultaneous reading of multiple files through a small one-page buffer
is slow with classic HDDs. Let the user improve it by chosing buffers of
several mebibytes if needed.

[kzak@redhat.com: - tiny changes to coding style
                  - mark buffers static
                  - use xalloc.h]

Signed-off-by: Eduard Bloch <blade@debian.org>
Signed-off-by: Karel Zak <kzak@redhat.com>

setterm: (man) improve dosc about optional arguments

Fixes: https://github.com/karelzak/util-linux/issues/1457
Signed-off-by: Karel Zak <kzak@redhat.com>

meson: fix typo

Addresses: https://github.com/karelzak/util-linux/issues/1442
Signed-off-by: Karel Zak <kzak@redhat.com>

meson: make raw(7) optional

Fixes: https://github.com/karelzak/util-linux/issues/1442
Signed-off-by: Karel Zak <kzak@redhat.com>

test_mount_optstr: use xstrdup()

Fixes: https://github.com/karelzak/util-linux/issues/1434
Signed-off-by: Karel Zak <kzak@redhat.com>

lib/env: don't ignore failed malloc

Fixes: https://github.com/karelzak/util-linux/issues/1434
Signed-off-by: Karel Zak <kzak@redhat.com>

mkswap: support -U {clear,random,time,uuid}

Let's follow -U from mkfs.ext4 and make it easy to fully control
UUIDs for the swap area.

Fixes: https://github.com/karelzak/util-linux/issues/1453
Signed-off-by: Karel Zak <kzak@redhat.com>

Merge branch 'int/alex/hifive_uuid' of https://github.com/AlexGhiti/util-linux

* 'int/alex/hifive_uuid' of https://github.com/AlexGhiti/util-linux:
  include: Rename HiFive partition UUIDs

Merge branch 'master' of https://github.com/mariobl/util-linux

* 'master' of https://github.com/mariobl/util-linux:
  Man pages: Fix end extend formatting

mount: add hint about dmesg(8) to error messages

Fixes: https://github.com/karelzak/util-linux/pull/1444
Signed-off-by: Karel Zak <kzak@redhat.com>

include: Rename HiFive partition UUIDs

Both HiFive Unleashed and HiFive Unmatched bootloaders seek for the same
UUIDs to load the next stage bootloader: the current name makes partitions
on Unmatched board appear as 'Unleashed'.

Fix that by removing the 'Unleashed' part of the current naming so it
fits both.

Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>

Man pages: Fix end extend formatting

mount.8.adoc: fix misformatting

Signed-off-by: Karel Zak <kzak@redhat.com>

unshare.1.adoc: Improve wording re creation of bind mounts

Use consistent wording for the options that create
bind mounts, and improve the wording.

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

unshare.1.adoc: Improve wording re namespace creation

The terminonoly "unshare the namespace" is not very helpful.
Instead, use wording that explains what these options actually
fo: "create a new namespace".

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

github: add meson build target

Signed-off-by: Karel Zak <kzak@redhat.com>

Merge branch 'meson-fixes' of https://github.com/Hello71/util-linux

* 'meson-fixes' of https://github.com/Hello71/util-linux:
  meson: make asciidoc optional
  meson: fix building libsmartcols
  meson: fix building logger

dmesg: translate ctime strings

Let's make the date-time strings accessible for translators.

Fixes: https://github.com/karelzak/util-linux/issues/1451
Signed-off-by: Karel Zak <kzak@redhat.com>

logger: fix --prio-prefix doesn't use --priority default

The commit b9ef27f have added priority check, but it introduced
regression as the default priority (as specified by --priority) is
ignored.

This patch fixes this problem, but it also removes extra check for
"kern facility", it's unnecessary and inconsistent with the rest of
logger.

Fixes: https://github.com/karelzak/util-linux/issues/1450
Signed-off-by: Karel Zak <kzak@redhat.com>

logger: dealloc login name

Signed-off-by: Karel Zak <kzak@redhat.com>

Merge branch 'losetup-direct-dio' of https://github.com/Hello71/util-linux

* 'losetup-direct-dio' of https://github.com/Hello71/util-linux:
  losetup: directly set dio instead of afterwards

Merge branch 'more-shell-execution' of https://github.com/ritikrajdev/util-linux

* 'more-shell-execution' of https://github.com/ritikrajdev/util-linux:
  more - Bug Resolve - new line separated command execution

Merge branch 'refactor-tests' of https://github.com/Ajchler/util-linux

* 'refactor-tests' of https://github.com/Ajchler/util-linux:
  tests: use subtests
  tests: make use of subtests

more - Bug Resolve - new line separated command execution

tests: use subtests

losetup: directly set dio instead of afterwards

This avoids an extra syscall, and allows the kernel to automatically set
block size [0], avoiding unnecessary failure with 4096-byte devices.

This changes the observable behavior of losetup --direct-io in the case
where DIO is not supported to fully fail, instead of creating a
non-direct-io device. If the user explicitly specifies --direct-io, then
they should get either a DIO loopdev or no loopdev, not a non-DIO
loopdev and a misleading error.

Additionally, loopcxt_setup_device now uses O_CLOEXEC in the read-only
fallback.

[0] https://github.com/torvalds/linux/commit/85560117d00f5d528e928918b8f61cadcefff98b

tests: make use of subtests

meson: make asciidoc optional

it should be under an option like autoconf, but cba to do that now, and
this is an improvement over the status quo.

meson: fix building libsmartcols

meson: fix building logger

Merge branch 'master' of https://github.com/yugart/util-linux

* 'master' of https://github.com/yugart/util-linux:
  Fix memory leaks in the chcpu

eject: fix typo in docs

Fixes: https://github.com/karelzak/util-linux/issues/1436
Signed-off-by: Karel Zak <kzak@redhat.com>

Fix memory leaks in the chcpu

Fix misspellings

docs: Uniformize references to section titles

Wrap section title references with asterisks, and prioritize
'see the *TITLE* section' over 'see section TITLE'.

tests: Skip lsns/ioctl_ns test if unshare fails

Some parts of the Debian build infrastructure uses unshare to run the
package build, and that appears to cause a "nested" unshare in the
lsns/ioctl_ns test to fail. Unfortunately the tests then hang at this
point.

Try running unshare before the actual test, and skip the test if unshare
already fails.

[kzak@redhat.com: - add --fork to the test
                  - don't write to stdout/err]

Signed-off-by: Chris Hofstaedtler <zeha@debian.org>
Signed-off-by: Karel Zak <kzak@redhat.com>

su: (bash-completion) offer usernames rather than files

Fixes: https://github.com/karelzak/util-linux/issues/1424
Signed-off-by: Karel Zak <kzak@redhat.com>

column: segmentation fault on invalid unicode input passed to -s option

The function mbs_to_wcs() returns NULL on invalid UTF.

Fixes: https://github.com/karelzak/util-linux/issues/1425
Signed-off-by: Karel Zak <kzak@redhat.com>

Merge branch 'bug/1416' of https://github.com/V-Smeets/util-linux

* 'bug/1416' of https://github.com/V-Smeets/util-linux:
  Erase line before writing the filename

Merge branch 'patch-2' of https://github.com/rffontenelle/util-linux

* 'patch-2' of https://github.com/rffontenelle/util-linux:
  asciidoc: unconstrained formatting pair in fdisk