tests: add libmount tags tests

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

tests: add libmount mtab tests

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

tests: add libmount tab_parse.c tests

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

libmount: cleanup fstab parser, improve mnt_tab errcb usage

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

tests: add libmount optstr.c tests

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

libmount: fix mnt_optstr_apply_flags() to work with optional variables

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

tests: use {src,build}dir variables for tests

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

libmount: improve assert() and DBG() usage

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

libmount: add umount(2) support

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

libmount: add mnt_context_get_{fstab,mtab}()

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

libmount: add mnt_tab_find_pair()

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

libmount: allows to swap source and target

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

libmount: cleanup mount.h

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

libmount: split context.c

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

libmount: cleanup helpers support

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

libmount: rewrite mountinfo/fstab parsers to use sscanf()

The old solution (without scanf()) was based on old code from
mount(8). It seems that the modern libc is able to provide all
necessary functionality by sscanf() and %ms directive.

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

libmount: fix mount helpers call

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

libmount: update mtab after mount(2)

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

libmount: always store rw/ro flags to mtab

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

libmount: update mount.sym

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

libmount: extend mnt_update_* API

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

libmount: add mount(2) and /sbin/mount.<type> support

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

libmount: add mtab managment to context API

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

libmount: use attribute __format__ for debug routines

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

libmount: add FS detection

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

libmount: better permissions and optstr evaliation

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

libmount: add DBG() to optstr.c

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

libmount: minor cleanup in optstr.c

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

libmount: add mnt_optstr_fix_user

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

libmount: clean up mountflags usage

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

libmount: improve parsers return codes

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

libmount: rewrite optstr translation

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

libmount: implement mnt_context_subst_optstr()

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

libmount: add routines for uid=, git= and context= translation

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

libmount: add utils for work with uid/gid

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

libmount: add first part of high-level API

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

libmount: cleanup mount.sym

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

libmount: add mnt_context

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

libmount: add mnt_optstr_get_userspace_mountflags()

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

libmount: add MNT_FS_SWAP flag

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

libmount: cleanup debug flags

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

libmount: improve debugging, fix optstr prepend

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

libmount: add new pseudo filesystems

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

libmount: add mnt_fs_prepend_optstr()

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

libmount: add mnt_optstr_prepend_option()

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

libmount: fix optstr test

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

libmount: rewrite DBG() macro

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

libmount: add new psetudo filesystems

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

libmount: cleanup return codes (tab_parse.c)

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

libmount: cleanup return codes (tab.c)

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

libmount: cleanup return codes (optstr.c)

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

libmount: cleanup return codes (optmap.c)

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

libmount: cleanup return codes (lock.c)

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

libmount: cleanup return codes (iter.c)

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

libmount: cleanup return codes (fs.c)

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

libmount: cleanup return codes (cache.c)

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

libmount: clean up mnt_fs_set_optstr()

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

libmount: improve an rename API for mtab management

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

libmount: update tests

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

libmount: update mount.{h,sym}

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

libmount: use mnt_fs for mtab API

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

libmount: fix datatype for mountflags

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

libmount: remove unnecessary options container

It seems that we are able to be happy with option strings only. The
mnt_optls stuff was over-engineering.

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

libmount: add support for mtab managment

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

lib: fix mange() and unmangle() to handle NULL

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

libmount: correctly use "none" values

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

findmnt: add support for fs-root (subvolumes and bind mounts)

This patch modifies the default output for SOURCE column. All btrfs
subvolume mountpoints and all bind-mount (where source is not root of
FS) will be printed as:

   SOURCE               TARGET
   /dev/sda1[/aaa]      /mnt/test

where /aaa is subvolume name or fs root for bind mounts, it means:

   # mount -t btrfs /dev/sda1 /mnt/test -o subvol=aaa

or:

   # mount --bind /aaa /mnt/test

The info about fs-root is 4th column in /proc/self/mountinfo.

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

libmount: add mnt_optstr_get_mountflags()

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

libmount: improve mnt_get_fs_root() and mnt_get_mountpoint()

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

libmount: add uhelper=, improve mnt_split_optstr()

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

libmount: add unit test for mnt_copy_fs()

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

libmount: add utils to get FS root/mountpoint

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

libmount: add mnt_copy_fs() and mnt_fs_set_root()

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

libmount: add unmangle/mangle() functions to API

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

libmount: remove fs and tab printf functions

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

libmount: add {start,end}swith() functions

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

libmount: add mnt_get_writable_mtab_path()

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

libmount: clean up lock.c

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

libmount: add mnt_tab_parse_mtab()

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

libmount: add mnt_split_optstr()

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

libmount: don't return old data from optls iterator

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

libmount: clean up mount.sym file

The API is not stable yet, so it's possible to increment the current
API version rather than create a new version.

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

libmount: add mnt_fs_get_root()

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

findmnt: add /etc/fstab.d support

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

libmount: add /etc/fstab.d support

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

lib: add wrappers for "at" functions

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

libmount: add mnt_tab_parse_stream()

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

libmount: don't store filename to mnt_tab

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

libmount: improve parser return codes

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

libmount: remove mnt_tab_strerr()

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

libmount: add mnt_tab_set_parser_errcb()

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

libblkid: use PRId64 in PT parser

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

hwclock: allow setting of RTC when it contains invalid data

In some cases the date/time stored in an RTC can be corrupted, eg due to
loss of power, before its been initially set, etc.  When this occurs
the RTC_RD_TIME ioctl can fail since the Linux kernel determines that
the RTC contains invalid data.  Currently, when setting an RTC using
hwclock, hwclock performs a number of RTC_RD_TIME ioctls before setting
the RTC.  When one of these ioctls fails, hwclock bombs out and the
corrupted RTC data can't be overwritten.  Thus once an RTC is corrupted,
it can't be fixed via hwclock*.

To work around the above issue we can make hwclock not exit when a
RTC_RD_TIME failure occurs during the process of setting the RTC.  This
allows the RTC to be set even when it contains an invalid value,
although it is not synchronized to a clock tick before it is set.

* 'hwclock --utc --noadjfile --set --date="11/23/10 17:19:00' currently
works to fix a corrupted RTC, but a user couldn't determine this without
digging through the source code.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>

lscpu: avoid len = 0 as a consequence of maxcpus = 0

This problem was observed on an x86_64 Mobile AMD Sempron 3700+ where kernel_max
returned "0" as the index of the highest CPU.

As a consequence, several variables in lscpu, which relied on maxcpus >= 1 (in
particular the 'len' value) were set to 0, resulting in the following errors:

host>./lscpu
lscpu: failed to read: /sys/devices/system/cpu/online: No such file or directory
host> cat /sys/devices/system/cpu/kernel_max
0

The fix used by this patch is to interpret kernel_max as an index and maxcpus as
a count >= 1, tested to work.

Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: Karel Zak <kzak@redhat.com>

partx: fix authors section in man page

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

cfdisk: close the program when terminal isn't present

cfdisk was getting characters without checking ERR return code, causing an
endless while loop in do_curses_fdisk() or in other functions when the terminal
is lost.

[kzak@redhat.com: - fix coding style, remove CR macro]

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

wipefs: fail if more device arguments specified

Currently wipefs process only first device argument and silently ignores
remaining.  Print error instead because it can be quite confusing.

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

hexdump: remove od functionality in favor to GNU coreutils od

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

libblkid: fix endian conversion bug in NILFS

Reported-by: Andrew Nayenko <resver@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

build-sys: remove unnecessary header

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