logindefs: fix coding style

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

islocal: fix coding style

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

chsh: fix coding style

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

chfn: fix coding style

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

chfn, setpwnam: let config.h be in control of extensions in use

The autotools will set _GNU_SOURCE which removes necessity to have
any _*_SOURCE definition in *.c files.

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

chfn, chsh: new file pamfail.h for error printing

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

chsh: build bug: do not override config.h definition

The ONLY_LISTED_SHELLS is defined by --disable-chsh-only-listed
configuration option.

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

chfn, chsh, setpwnam: get true/false from stdbool.h

Requires C99, which should not be a problem.

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

login: add version printing option

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

newgrp: add version and help options

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

vipw: stop printing non-sense version string

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

vipw: use libc error printing facilities

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

chsh: use libc error printing facilities

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

chsh: align with howto-usage-function.txt

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

chfn: use libc error printing facilities

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

chfn: align with howto-usage-function.txt

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

fallocate: clean up fallbacks for FALLOC_FL_* flags

Reported-by: Voelker, Bernhard <bernhard.voelker@siemens-enterprise.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

tests: add mkfs.carmfs test for small FS with duplicate files

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

small bug in mkfs.cramfs + patch

mkfs.cramfs on version v2.17.2 has a small bug when dealing with very small
filesystems.  look at upstream code on git, and the issue is still here.
ex:

         $ cd /tmp
         $ mkdir content
         $ echo hello > content/a
         $ echo hello > content/b
         $ mkfs.cramfs content content.cramfs
         not enough space allocated for ROM image (4090 allocated, 4096 used)

Here is a patch for fixing (simply rounding to page size AFTER eleminating
doubles).

Signed-off-by: Arnaud Mouiche <arnaud.mouiche@invoxia.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

fallocate: add FALLOC_FL_PUNCH_HOLE support

Recent Linux kernel supports FALLOC_FL_PUNCH_HOLE in fallocate(2).
This patch adds FALLOC_FL_PUNCH_HOLE support to fallocate utility,
by introducing a new option -p|--punch-hole.

[kzak@redhat.com: - fix merge conflict in fallocate.1]

Cc: Karel Zak <kzak@redhat.com>
Cc: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

prlimit: s/amount/number/

prlimit: add support for executing a command

prlimit.c: Alternatively to applying the limits to an existing process via the
--pid option, allow a command to be executed. Adapted usage() accordingly.

prlimit.1: mention new syntax.

Signed-off-by: Bernhard Voelker <mail@bernhard-voelker.de>

libblkid: fix mac partition detection

This fixes the buffer length passed to blkid_probe_get_buffer() calls,
and the block size byte order (on little-endian systems).

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

mount: enable fstab.d

This patch makes mount command read fstab.d/*.fstab if the command
is built with --enable-libmount-mount.

[kzak@redhat.com: - add some info to fstab.5 and mount.8 man pages
                  - use error callback for tab parser]

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

flock: use O_RDWR as fallback if O_RDONLY returns EIO

The commit 75aaee08f06b92d119ed827c53d1af5474eb16ff introduces
regression:

     $ echo '#!/bin/sh' > test.sh
     $ chmod a+rx test.sh
     $ flock -eon ./test.sh ./test.sh
     flock: ./test.sh: Text file busy

The lock file cannot be opened in read-write mode by default, because
then we cannot use flock(1) to lock executable files.

The read-write mode for lock files is necessary on NFSv4 where
flock(2) is emulated by by fcntl() -- this situation is possible to
detect by flock(2) EIO error.

This patch reverts the default to O_RDONLY and use O_RDWR only if EIO
error is detected.

Reported-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Karel Zak <kzak@redhat.com>

flock: timer code refactoring

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

login: fix compiler warning [-Wreturn-type]

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

build-sys: add missing BUILD_PRLIMIT

Reported-by: Thorsten Kukuk <kukuk@suse.de>
Signed-off-by: Karel Zak <kzak@redhat.com>

login: improve pam_setcred() usage

Reported-by: Thorsten Kukuk <kukuk@suse.de>
Signed-off-by: Karel Zak <kzak@redhat.com>

column: avoid memory overrun and/or use of uninitialized buffer

* text-utils/column.c (maketbl): Use the right starting point
and the right length when zeroing new memory after xrealloc.

prlimit: don't share pointer for old and new in prlimit(2)

This patch makes the code more robust. We should not share the same
pointer for old and new arguments for prlimit(2) syscall.

Reported-by: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Karel Zak <kzak@redhat.com>

prlimit: avoid segfault due to array-out-of-bounds error

Example:
$ prlimit -l -l -l -l -l -l -l -l -l -l -l -l -l -l -l -l -l -l -l
RESOURCE DESCRIPTION                         SOFT   HARD UNITS
MEMLOCK  max locked-in-memory address space 65536 262144 bytes
MEMLOCK  max locked-in-memory address space 65536 262144 bytes
...
Segmentation fault

Reported-by: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Karel Zak <kzak@redhat.com>

prlimit: fix typo in man page, option --verbose

Signed-off-by: Bernhard Voelker <mail@bernhard-voelker.de>

prlimit: replace lims[i] with pointer

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

prlimit: fix case when PID is given later

 # prlimit --nofile=:4000 --pid $$

When the PID is given later than a partially given limit, then prlimit
used the current PID for getting the missing part (hard, soft) of the
limit.

Factored out the retrieval of the unknown limit from parse_prlim() to
new get_unknown_hardsoft() which is to be called by do_prlimit() based
on the struct prlimit member 'modify' set by add_prlim().

[kzak@redhat.com: - use prlimit->mofify as mask
                  - add soft vs. hard limit check from another
                    Bernhard's patch]

Signed-off-by: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Karel Zak <kzak@redhat.com>

prlimit: fix error message

 - differentiate between set and get operations
 - add limit name to the error message
 - remove PID from the message (whole util works always with only one
   PID, so this info is superfluous)

Reported-by: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Karel Zak <kzak@redhat.com>

prlimit: show all limits if called without options

Fix option parsing: prlimit ran into usage() if called
without arguments which should only be done for surplus
arguments. Instead, it should display all limits.

Signed-off-by: Bernhard Voelker <mail@bernhard-voelker.de>

wipefs: add -t <list> option

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

lib,match: split match_fstype() from libmount

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

libblkid: export PTMAGIC also if only PTTYPE requested

Reported-by: David Zeuthen <davidz@redhat.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

fdisk: split delete_partition() off from dos_delete_partition()

This accidentally fixes a mistake printing the "Partition n is deleted"
message as the 'i' variable get decremented or incremented in case of
logical partitions.

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

fdisk: avoid segfault validating a sgi label (boot/swap not set)

swap_part or boot_part can be set to -1 when they don't exist.

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

fdisk: use ngettext() in str_units() for plural forms

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

switch_root: umount mount points we cannot move with MNT_DETACH

If a mount point cannot be moved to the new root, umount it with
MNT_DETACH, so that it is lazy umounted and does not show up in
/proc/mounts anymore.

[kzak@redhat.com: - add MNT_DETACH macro fallback]

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

fdisk: enable expert menu with SGI label

This allows the creation of a new label if the disk already has an existing
SGI disklabel without working around this problem such as creating a DOS
disklabel or overwriting with zero the partition table (problem reported in
'Gentoo Linux/MIPS Handbook').

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

fdisk: rewrite print menu system

DOS and SGI label menus are unchanged. BSD label command
descriptions change slightly to use a common form.

This also removes an useless menu entry in SUN label menu to
edit bsd disklabel.

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

fdisk: move command prompt code to a separate function

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

fdisk: rename bselect, xselect functions

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

lib,tt: fix compiler warning [-Wreturn-type]

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

readprifile: remove unnecessary optstring array

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

Merge branch 'sys-utils-3rd' of https://github.com/kerolasa/lelux-utiliteetit

* 'sys-utils-3rd' of https://github.com/kerolasa/lelux-utiliteetit:
  ldattach: add default case to switches
  ldattach: convert debug macro to function
  ldattach: check numeric user inputs with strtol_or_err
  ldattach: align with howto-usage-function.txt
  ldattach: fix coding style
  docs: add long options to readprofile.8
  readprofile: fix coding style
  readprofile: use libc error printing facilities
  readprofile: add long options
  docs: align tunelp.8 with howto-man-page.txt
  docs: add long options to tunelp.8
  tunelp: add long options and align with howto-usage-function.txt
  tunelp: use libc error printing facilities
  tunelp: use symbolic exit values
  tunelp: fix coding style
  switch_root: add nls support
  switch_root: align with howto-usage-function.txt
  unshare: align with howto-usage-function.txt
  rtcwake: align with howto-usage-function.txt
  setsid: add long options and fix coding style

libblkid: export info about PT magic strings

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

libblkid: add BLKID_PARTS_MAGIC

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

mkswap: wipe all old signatures

mkswap(8) zaps the begin of the device (1024 bytes) only. This is not
enough, because for example ReiserFS superblock is outside this area.

This patch add blkid_do_wipe() call to mkswap(8) and all block device
is scanned and all signature are removed.

Reported-by: Martin Pitt <martin.pitt@ubuntu.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

docs: update TODO file

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

wipefs: improve -a, use blkid_do_wipe()

Some filesystems (e.g. FAT) is possible to detect in many ways -- for
example if a magic string is stored on more places on the device. The
current implementation always removes from the device only the firstly
detected possibility.

 # wipefs /dev/sdb1
 offset               type
 ----------------------------------------------------------------
 0x36                 vfat   [filesystem]
                      UUID:  D203-A7F4

 # wipefs -a /dev/sdb1
 8 bytes [46 41 54 31 36 20 20 20] erased at offset 0x36 (vfat)

 # wipefs /dev/sdb1
 offset               type
 ----------------------------------------------------------------
 0x0                  vfat   [filesystem]
                      UUID:  D203-A7F4

This patch fixes this issue:

 # wipefs -a /dev/sdb1
 8 bytes were erased at offset 0x00000036 (vfat): 46 41 54 31 36 20 20 20
 1 bytes were erased at offset 0x00000000 (vfat): eb
 2 bytes were erased at offset 0x000001fe (vfat): 55 aa

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

libblkid: add blkid_do_wipe()

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

libblkid: don't call CDROM_GET_CAPABILITY for partitions

Reported-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

fsck.minix: fix zone_map memory allocation failure check

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

lib,tt: add TT_FL_NOEXTREMES

If you mark a column with TT_FL_NOEXTREMES flag then extremely
large fields will no have effect to column width. Foe example:

 without the TT_FL_NOEXTREMES flag for the 'AAA' column:

 AAA         BBB CCC  DDD
 aa          bbb ccc  ddd
 aaaaaaaaaaa  bb ccc  ddd
 aa           bb ccc  ddd
 aa           bb ccc  ddd

 with the flags:

 AAA  BBB CCC  DDD
 aa   bbb ccc  dddddddddd
 aaaaaaaaaaa
       bb ccc  dddddddd
 aa    bb ccc  dddddd
 aa    bb ccc  ddddddddd

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

prlimit: fix noheadings option in man page

Signed-off-by: Bernhard Voelker <mail@bernhard-voelker.de>

prlimit: fix SYS_ macro

Reported-by: "Voelker, Bernhard" <bernhard.voelker@siemens-enterprise.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

libmount: fix trivial typos in debugging output

Signed-off-by: Petr Uzel <petr.uzel@suse.cz>

whereis: remove unused value warning

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

libblkid: silence a format string warning [-Wformat]

Fix the following clang warning:

 exfat.c:130:41: warning: conversion specifies type 'unsigned short' but
 the argument has type 'uint8_t' (aka 'unsigned char') [-Wformat]
         blkid_probe_sprintf_version(pr, "%hu.%hu",
                                              ~~^
                                              %c

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>

fsck.minix: remove unnecessary memset calls

The program aborts without using this unintialized allocated memory,
setting to zero doesn't look needed.

memset calls with sizeof() of pointer as argument (reported by clang).

[kzak@redhat.com: - replace malloc with calloc for {zone,inode}_count
                    to make code more robust]

Signed-off-by: Francesco Cosoleto <cosoleto@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

mount: fix build

With the new build changes when libmount is not defined the compilation breaks:

fstab.c:440: undefined reference to `streq_except_trailing_slash'
collect2: ld returned 1 exit status
make[2]: *** [mtab_lock_test] Error 1
make[2]: Leaving directory `/home/dave/projects/ut/mount'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/dave/projects/ut'
make: *** [all] Error 2

This patch adds strutils to the fstab.c rule.

Signed-off-by: Davidlohr Bueso <dave@gnu.org>

libmount: fix chdir to parent for restricted (user) umounts

Reported-by: Petr Uzel <petr.uzel@suse.cz>
Signed-off-by: Karel Zak <kzak@redhat.com>

docs: refresh TODO file

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

libmount: ignore tailing slash in netfs source paths

Addresses: https://bugzilla.novell.com/show_bug.cgi?id=728480
Signed-off-by: Petr Uzel <petr.uzel@suse.cz>
Signed-off-by: Karel Zak <kzak@redhat.com>

lib,loopdev: add LOOP_CTL_GET_FREE support

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

docs: add note about config-gen purpose

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

docs: refresh TODO file

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

login: clean FAIL_DELAY usage

... with this patch: login(1) always uses sleep(FAIL_DELAY) before
exit() if user is not authenticated or his PAM session is not fully
initialized.

Note that user has three attempts to write his password before
login(1) is terminated.

Reported-by: Francesco Cosoleto <cosoleto@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

prlimit: add --{raw,noheadings} options

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

prlimit: add a units column

Signed-off-by: Davidlohr Bueso <dave@gnu.org>
Signed-off-by: Karel Zak <kzak@redhat.com>

docs: add notes about tools/config-gen

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

tests: add build-sys regression test

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

build-sys: remove redundant flags from mount/Makefile.am

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

build-sys: add tools/config-gen and basic compilation scenarios

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

build-sys: fix UL_INIT_BUILD macro

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

tests: fix options evaluation, add support for optional tests

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

build-sys: fix libuuid  initialization

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

prlimit: add add falbacks

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

libuuid: fix includes in man pages

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

build-sys: add UL_BUILD_INIT macro

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

build-sys: cleanup BUILD_HWCLOCK

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

build-sys: cleanup BUILD_{WALL,WRITE}

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

build-sys: cleanup BUILD_SCHEDUTILS

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

build-sys: move --with-* before --enable-*

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

build-sys: cleanup BUILD_LOGIN_UTILS

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

build-sys: cleanup BUILD_{RENAME,RESET}

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

build-sys: cleanup BUILD_{LAST,LINE,MESG}

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

build-sys: cleanup BUILD_KILL, enable kill(1) by default

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

build-sys: cleanup BUILD_ELVTUNE

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

build-sys: cleanup BUILD_PIVOT_ROOT

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

build-sys: cleanup BUILD_SWITCH_ROOT

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

build-sys: cleanup BUILD_CRAMFS

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

build-sys: cleanup BUILD_{DDATE,AGETTY}

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

build-sys: cleanup BUILD_ARCH

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