libfdisk: undefined or garbage value returned to caller [clang analyze]

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

fdisk: value is never read [clang analyze]

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

fdisk: value is never read [clang analyze]

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

lsipc: value is never read [clang analyze]

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

fdisk: value is never read [clang analyze]

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

po: update fr.po (from translationproject.org)

po: update da.po (from translationproject.org)

unshare: remove duplicate [make checkincludes]

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

lslogins: remove duplicate [make checkincludes]

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

libmount: sort pseudo filesystems by LC_COLLATE=C

Addresses: https://github.com/karelzak/util-linux/issues/212
Signed-off-by: Karel Zak <kzak@redhat.com>

tasts: add logger octet counting test

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

logger: Add support to logger for RFC6587 octet counting

This patch adds support to logger for RFC6587 octet counting.
RFC6587 provides support for two sorts of framing:

1. Octet counting (at RFC6587 s3.4.1)

   In essence each frame is preceded by a decimal length and a
   space.

2. Non-transparent framing (at RFC6587 s3.4.2), also called
   'octet stuffing'

   In essence each frame is terminated by a `\n`

Prior to this patch, logger used option 2 (non-transparent framing)
on TCP, and used no framing on UDP. After this patch, the default
behaviour is unchanged, but if the '--octet-count' option is supplied,
option 1 is used for both TCP and UDP. Arguably octet count framing
makes little sense on UDP, but some servers provide it and this
allows testing of those servers.

Signed-off-by: Alex Bligh <alex@alex.org.uk>

sfdisk: add --list-free

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

fdisk: add 'F' command to list free unpartitioned space

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

Merge branch 'test-fixes' of https://github.com/rudimeier/util-linux

Revert "tests: start script/race test only when --force specified"

This reverts commit 8ba3f35e07f736a0165669ac787b016b4311eb29.

This test is not too slow anymore and BTW since last script refactoring
it does work now even on slow/heavy-loaded systems.

Signed-off-by: Ruediger Meier <ruediger.meier@ga-group.nl>

script: improve SIG{TERM,QUIT,KILL} sensitivity

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

cfdisk: don't print obsolete CHS addresses

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

cfdisk: make sure that output fits to terminal width

Reported-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Karel Zak <kzak@redhat.com>

libsmartcols: reduce relative columns more aggressively

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

fdisk: init libsmartcols debug

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

cfdisk: remove extra line between menu and info line

Reported-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Karel Zak <kzak@redhat.com>

cfdisk: spell "label" in lower case

To be less shouty.  It is not an abbreviation like "UUID".

Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

cfdisk: allow an uppercase X to toggle the extra info

To make true the help text line that says that all commands
can be entered with either upper or lower case.

Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

build-sys: add --disable-cal

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

build-sys: fix test_colors build

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

tests: update build-sys tests

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

libsmartcols: don't link with tinfo

Let's move color names to sequence translation to separate file to
make it usable without all the stuff in lib/colors.c.

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

tests: add udevadm settle

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

libfdisk: add comment to the header file

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

lsipc: make --id usable with all output formats and with -o

This patch allows to use things like:

 # lsipc -m --id 47611910 -o COMMAND,SIZE,KEY --json

or

 # lsipc -m --id 47611910 -o SIZE --bytes --noheadings --raw

to get just one value for the resource.

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

lsipc: the default --global is mutually exclusive with -c --id and -t

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

lsipc: remove --print0 and -colon-separate

These options have been originally designed for lslogins for
compatibility another passwd-like tools. It does not make sense for
IPC.

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

lsipc: remove duplicate "Usage:"

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

cfdisk: make sure that extra info is refreshed/removed

For example when you move from a Partition to the FreeSapce then we
need to remove old extra info and draw nothing.

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

cfdisk: properly toggle extra info window, make it more robust

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

lsipc: tiny change about --global to the man page

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

ipcmk: support <size>{K,M,G,...}

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

lsipc: default to --global

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

lsipc: add USE% to --global

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

lsipc: use sub-table for semphore elements

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

lsipc: improve JSON output

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

lsipc: move outmode to coltrol struct

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

lsipc: fix coding style

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

lsipc: differentiate between UID and USER names

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

lsipc: fix ID parsing, use full time for pretty print

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

lsipc: cleanup -o usage

* support -o+AAA,BBB,CCC
* keep default resource columns in getopt switch()

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

lsipc: improve --global

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

include/xalloc: add err_oom()

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

lsipc: cleanup columns indexes

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

lsipc: add COMMAND for shmem

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

lib/procutils: add proc_get_cmdline()

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

include/all-io: read_all() don't retry on EOF

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

lsipc: add CTIME to shmem default

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

lsipc: new command to list IPC facilities

Co-Author: Karel Zak <kzak@redhat.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Ondrej Oprala <ooprala@redhat.com>

agetty: Reprompt once the network addresses change if address displayed

Several of the /etc/issue escape codes such as \4 and \6 depend on
the current addresses of the system that can change after the agetty
prompt is displayed.  This can cause stale data to be displayed
when a user looks at a VT, especially in cases of DHCP racing with
system start up.

Similar to the --reload mechanism, if we're displaying an address
in the issue output, and the user hasn't typed anything yet: then
redisplay the prompt with the new address.

We use netlink to watch for address changes. We only open the netlink
socket if we display an address in the issue file.

logger: do not exit when socket errors are not enforced

The libc openlog(3) does not have error detection whether unix socket
could be opened.  As a side effect that made it possible to use logger
even if syslogd was not running.  Of course user message in these cases
were lost.  This change makes the logger do behave similar way again, so
that sysvinit scripts can successfully pipe messages to logger when ever.

Addresses: https://bugs.debian.org/787864
Addresses: https://bugs.debian.org/790875
Reported-by: Andreas Beckmann <anbe@debian.org>
Reported-by: Andreas Henriksson <andreas@fatal.se>
Tested-by: Robie Basak <robie.basak@ubuntu.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

hwclock: add -D to the man page

Addresses: https://bugzilla.redhat.com/show_bug.cgi?id=1244609
Signed-off-by: Karel Zak <kzak@redhat.com>

Merge branch 'kill-man' of https://github.com/ronin13/util-linux

lslogins: merge read_utmp() code

The code is used only in lslogins, so it does not make sense to
maintain it in libcommon.

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

lslogins: fix --user-accs and --system-accs docs

Reported-by: Mantas Mikulėnas <grawity@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

bash-completion: cal: add --twelve and --color options

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

ldattach: tiny coding style changes

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

ldattach: order line disciplines canonically

List line disciplines in numerical order and put official names
before convenience aliases.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>

ldattach: always set line speed and format for N_GIGASET_M101

The Gigaset M101 exhibits particularly awkward and confusing behaviour
when the serial line speed isn't set to what the device expects:
Configuration mode detects the used speed and works anyway, but data
mode silently fails. Troubleshooting this is enormously time-consuming.

Spare the users this trouble by setting the line speed and format to
these devices' factory default values as long as the command line
doesn't say otherwise.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>

ldattach: adapt print_table column width to widest entry

The hardcoded column width of 10 in print_table() is not enough, as
the currently longest table entry to be printed has 12 characters.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>

libfdisk: fix fdisk_label_parse_parttype() for unknown types

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

libfdisk: (gpt) add missing GUIDs

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

libfdisk: add GPT GUIDs for PReP

Power Firmware (OFW) scans chosen devices for PReP partition IDs.

Signed-off-by: Scott Moser <smoser@ubuntu.com>

libfdisk: (gpt) fix "MidnightBSD UFS" UUID

Reported-by: Андрей Гаврилин <gaal.dev@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

script: evaluate errno only if read() sets it

[kzak@redhat.com: - be careful with errno and DBG
                  - add EINTR check
                  (both suggested by Rudi]

Signed-off-by: Ruediger Meier <ruediger.meier@ga-group.nl>
Signed-off-by: Karel Zak <kzak@redhat.com>

kill: Add verbose to man page.

kill adds a verbose option to print the pid(s) and the signal. It is
added to man page here.

script: make sure errno is zero

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

swapon: fix restart waitpid with errno == EINTR

Reported-by: Maks Naumov (https://github.com/karelzak/util-linux/pull/205)
Signed-off-by: Karel Zak <kzak@redhat.com>

lib/strutils: make strmode() more generic

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

Merge branch 'sulogin'

fstrim: close dir before exit [coverity scan]

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

rtcwake: cosmetic changes

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

rtcwake: fix ARRAY_SIZE() usage

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

rtcwake: don't use usage() on error

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

rtcwake: fix valgrind warning

==3445== Syscall param ioctl(generic) points to uninitialised byte(s)
==3445==    at 0x4F16477: ioctl (in /usr/lib/libc-2.21.so)
==3445==    by 0x402785: setup_alarm (rtcwake.c:231)
==3445==    by 0x403302: main (rtcwake.c:520)

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: read accepted mode strings from /sys/power/state

The rtcwake has various mode names that are internal to the command,
listed in rtcwake_mode_string array, while some modes are dynamicaly
listed in /sys/power/state.  Together these form the allowed modes.

[kzak@redhat.com: - cleanup strv usage]

Reviewed-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Signed-off-by: Karel Zak <kzak@redhat.com>

rtcwake: make some command line options mutually exclusive

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: improve coding style

Reviewed-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: add --list-modes

Commit ece44f19f423408f576f348fed2845c876d72c6e missed freeze mode, which
to a simple patch adding the missing argument but Karel pointed out it
would be better to make the rtcwake to tell what arguments are supported
so that possible changes end up to be automatically correct in bash
completion.

Proposed-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: do not overwrite device name

This improves error messaging, and removes a memory leak.

Reviewed-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: clean up struct tm initializations

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: fix preprocessor redefinition

The RTC_AF is expected to be part of linux/rtc.h file.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: add human readable --date timestamp format

Reviewed-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: improve read_clock_mode()

Make skipping two lines more robust, and add message about unexpected
adjfile contents when running with --verbose.

Reviewed-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: remove RTC_ALM_READ and RTC_ALM_SET compatibility

The RTC_WKALM_RD and RTC_WKALM_SET have been available since 2.6.17, and
preferred way since 2007.  Keeping the fallbacks to the old interface is
no longer needed.

Reference: https://github.com/torvalds/linux/commit/e824290e5dcfaf2120da587b16d10dfdff8d5d3e
Reference: https://github.com/torvalds/linux/commit/f8245c26886c912627ebc49f714e4491261224c4
Reviewed-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: replace long if else statement with switch case

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: enumerate constant mode strings

Reviewed-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

rtcwake: add rtcwake_control and remove global variables

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

reset: deprecate the script, and remove reset.033c

Most distributions probably ship reset from ncurses, which means
util-linux reset(1) is unlikely to be in use anywhere.  Assuming the
previous being true it is almost unimaginable anyone one would be using
the alternative script, so remove it and mark the script deprecated.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

getopt: use $(...) command substitution in example script [shellcheck]

Reference: https://github.com/koalaman/shellcheck/wiki/SC2006
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

docs: unify mkswap(8) with swapon(8) about holes warning

Reviewed-by: Dave Rutherford <dave@evilpettingzoo.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>

zramctl: add link to kernel documentation

The manual page does lacks explanation how to create zram with modprobe.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>

last-deprecated: remove from source tree

The last/lastb(1) from sysvinit has been around for about two years,
and the better implementation is already part of releases 2.24 to 2.26.
It should be safe to remove the unused last code from the source tree.

Reference: ce60272039ea11952b15fefb653892dd0da02217
Signed-off-by: Sami Kerola <sami.kerola@lastminute.com>

script: fix variable initialization warning

term-utils/script.c:402:19: warning: obsolete use of designated
initializer without '=' [-Wpedantic]

Signed-off-by: Sami Kerola <kerolasa@iki.fi>