libblkid: minor changes to dm topology code

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

libblkid: cleanup topology fallback

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

libblkid: announce Joliet extension

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

fallocate: new command

The fallocate(1) utility is used to preallocate blocks to a file.

This can be useful for virtual images, database files, testing, etc.
Normally we'd hope that various tools will start using preallocation
internally, but until then such a utility may be useful, and could be
scripted as well.

The original Eric's version is available at:
http://thread.gmane.org/gmane.linux.utilities.util-linux-ng/2490

This version:
 - checks for fallocate glibc function and kernel syscall
 - does not provide a fallback and does not call posix_fallocate()
 - adds long options
 - uses err.h for errro messages
 - adds NLS support
 - cleanups man page

Co-Author: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

mount: document changed semantics of tmpfs size option in mount.8

Linux kernel commit 818db35992c249dc32c1d86daf7d533fb0952f5d fixed a
problem where invoking mount -t tmpfs with a size argument smaler than the
size of one page caused the limits on the size of a tmpfs to not operate.
The commit also caused the number supplied by the size option to be
rounded up to the  nearest page size instead of down.  This alters the man
page to match the behavior of the kernel in this instance.

Signed-off-by: Kevin Granade <kevin.granade@gmail.com>

libblkid: prefer ISO9660 PVD Label to Joliet Label

The old good libvolume_id had the same behavior.

Reported-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>

libblkid: trim tailing whitespace from unicode LABELs

old version:
$ ./blkid -o udev -p iso-joliet.img
ID_FS_LABEL=ThisIsLabel
ID_FS_LABEL_ENC=ThisIsLabel\x20\x20\x20\x20\x20
ID_FS_VERSION=Joliet\x20Extension
ID_FS_TYPE=iso9660
ID_FS_USAGE=filesystem

new version:
$ ./blkid -o udev -p iso-joliet.img
ID_FS_LABEL=ThisIsLabel
ID_FS_LABEL_ENC=ThisIsLabel
ID_FS_VERSION=Joliet\x20Extension
ID_FS_TYPE=iso9660
ID_FS_USAGE=filesystem

Reported-by: Maxim Levitsky <maximlevitsky@gmail.com>
Addresses-Ubuntu-Bug: #432215
Signed-off-by: Karel Zak <kzak@redhat.com>

libblkid: fix blkid_probe_set_utf8label() call for Joliet

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

libblkid: minor changes to samples

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

libblkid: add EVMS topology support (for old kernels)

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

libblkid: add DM topology support (for old kernels)

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

libblkid: add MD topology support (for old kernels)

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

libblkid: cosmetic change in topology sample

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

libblkid: fix probing for binary interface

The probing for binary interface has to always start from scratch.

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

libblkid: add private blkid_topology_set_*() functions

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

libblkid: add blkid_driver_has_major()

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

libblkid: fix blkid_devno_to_wholedisk()

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

libblkid: allows more probing methods for topology chain

This patch moves the current topology code to the separate sysfs.c
file.

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

libblkid: minor fix in topology sample

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

libblkid: fix topology information values

Looks like the topology values returned are being
inflated by x 512, but the kernel reports them in bytes;
no need for the multiplication, just return the value as-is.

With the patch it all matches:

# mdadm -D /dev/md0  | grep Chunk
     Chunk Size : 64K
# cat /sys/devices/virtual/block/md0/queue/minimum_io_size
65536
# blkid/samples/topology /dev/md0
----- NAME=value interface (values: 0):
MINIMUM_IO_SIZE = 65536
OPTIMAL_IO_SIZE = 131072
----- binary interface:
alignment offset : 0
minimum io size  : 65536
optimal io size  : 131072

Signed-off-by: Eric Sandeen <sandeen@redhat.com>

mount: fix mount.8, xfs attr2 is enabled by default

Address-Red-Hat-Bug: #510845
Signed-off-by: Karel Zak <kzak@redhat.com>

libblkid: add blkid_partition_get_type_string()

.. because for example Mac and GPT don't use numbers, but strings or
UUIDs as partition type identifier.

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

mount: fix reference to samba-client in mount.8

Address-Red-Hat-Bug: #521367
Signed-off-by: Karel Zak <kzak@redhat.com>

po: update POTFILES.in

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

tests: add partitions probing test

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

tests: rename blkid/images to blkid/images-fs

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

libblkid: gtkdocize (API docs generated by gtk-docs)

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

libblkid: add missing comments

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

libblkid: add mkfs sample

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

libblkid: add partitions sample

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

libblkid: add EFI GPT partitions support

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

lib: add a generic crc32()

This is public domain implementation, based on static crc32_table[].
This implementation is used on may places (libparted, FreeBSD kernel,
PostgreSQL, ...). The Linux kernel uses on-the-fly generated and
allocated tables. That's useless in shared libraries.

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

libblkid: add MINIX partitions support

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

libblkid: add MS-DOS partitions support

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

libblkid: add MAC partitions support

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

libblkid: add SGI partitions support

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

libblkid: add SUN partitions support

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

libblkid: add SOLARIS-X86 partitions support

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

libblkid: add UNIXWARE partitions support

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

libblkid: add BSD partitions support

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

libblkid: add AIX partitions support

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

libblkid: add partitions filter routines

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

libblkid: add partitions parsing support

This patch is large, but not invasive. There is a new code only.

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

libblkid: add samples/topology.c

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

libblkid: add topology support

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

libblkid: add blkid_devno_to_wholedisk()

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

libblkid: use the new API in whole u-l-ng

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

libblkid: refresh blkid.{h,sym}

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

libblkid: add generic function for binary data

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

libblkid: allow to read in sectors

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

libblkid: allow to change dimension of probing area

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

libblkid: add blkid_probe_get_{size,sectorsize,devno}

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

libblkid: add missing comments to probe.c

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

libblkid: use chains in blkid_do_{safe,full,}_probe()

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

libblkid: use chains in prober (de)initialization

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

libblkid: use private {lookup,get}_value functions

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

libblkid: remove superblock functions from probe.c

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

libblkid: use superblock filter functions

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

libblkid: add superblocks filter functions

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

libblkid: move blkid_known_fstype() to superblocks.c

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

libblkid: add superblocks chain

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

libblkid: add BLKID_SUBLKS_* flags

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

libblkid: add generic filter functions

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

libblkid: add functions for chain tags

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

libblkid: add chain structs

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

libblkid: move filter macros to header file

 * move blkid_bmp_* macros to blkidP.h
 * add blkid_bmp_nbytes() -- returns size of bitmap in bytes
 * add blkid_bmp_nwords() -- returns size of bitmap in words
                             (rename from blkid_bmp_size())

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

libblkid: create a generic blkid_encode_to_utf8()

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

libblkid: create a generic blkid_unparse_uuid()

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

libblkid: use superblocks.h

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

libblkid: move FS/raid stuff to superblocks directory

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

libblkid: fix typo (swsupend -> swsuspend)

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

mount: more explicitly explain 'strictatime' in mount.8

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

sfdisk: dump has to be $LANG insensitive

This stupid bug has been introduced by:

commit add5133f4ad5136aac3ce7627e615d14893d0aeb
Author: Pedro Ribeiro <p.m42.ribeiro@gmail.com>
Date:   Fri Oct 3 08:52:35 2008 +0200
fdisk: several strings without gettext calls

... so it shows that we need to improve our review process... :-(

Address-Red-Hat-Bug: #522718
Signed-off-by: Karel Zak <kzak@redhat.com>

libuuid: remove .UE macro from libuuid man pages.

On some system it produces insertion of "()".
There is missing .UR anyway.

[kzak@redhat.com: - this is u-l-ng bug, we don't use UR+UE macros like the
                    original man pages in e2fsprogs]

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

docs: README width and language correction

Signed-off-by: Jan Sarenik <jsarenik@redhat.com>

cfdisk: more key alternatives

Signed-off-by: Jan Sarenik <jsarenik@redhat.com>

rtcwake: ignore the tm_isdst field returned from the RTC

Signed-off-by: Paul Fox <pgf@laptop.org>

partx: work properly with 512 sectors (dos PT)

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

rtcwake: add S5 support

Based on patch from
   Piergiorgio Sartor <piergiorgio.sartor@nexgo.de>
at
   https://bugzilla.redhat.com/show_bug.cgi?id=449115

Piergiorgio's note about S5:
> According to the ACPI specifications, chapter 4.7.2.4 "Real Time
> Clock Alarm", the wakeup from RTC, when supported, should work from
> *sleep* state S1-S3 and, optionally, from S4.
>
> Note 3 (same chapter) says that S5 is *not* a sleep state and should
> not be supported. Actually it also says that: "The OS will disable
> the RTC_EN bit prior to entering the G2/S5 or G3 states regardless."
>
> Nevertheless, on all PC supporting the RTC wakeup I tested, all were
> able to wake from S5.

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

mkswap: fix memory leaks, cleanup check_blocks()

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

mkswap: unbreak -c ("check") option.

In the check_blocks() loop, current_page is not incremented. Because of this
bug, the loop does not end when the end of the device is reached.  Instead it
tries to continue reading, which of course fails, and eventually mkswap aborts
("too many bad pages").

Note that this can only be tested with swap partitions, mkswap ignores the -c
option for swap files.

Signed-off-by: Peter De Wachter <pdewacht@gmail.com>

libblkid: fix buffer overflow in blkid_encode_string()

[kzak@redhat.com: - this is patch is originally from udev repository
                    commit 8cfcf9980a3a7037a12a3052c38e4981cb0f0190]

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

libblkid: add *.ko.gz support to modules.dep parser

Address-Red-Hat-Bug: #518572
Signed-off-by: Karel Zak <kzak@redhat.com>

replace usleep() for systems that don't have them

This function is marked obsolete in POSIX.1-2001 and removed in
POSIX.1-2008.

Conditionally replaced with nanosleep().

Signed-off-by: Daniel Mierswa <impulze@impulze.org>

hwclocks: use time limit for KDGHWCLK busy wait

Currently the busy wait in synchronize_to_clock_tick_kd() is
restricted by number of loops. It's better to use time limit
(1.5s). We already use this method for RTC.

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

mount: mention mtab for single mount point mount in mount.8

Karel Zak wrote: "This is very old (10+years) feature. The mount(8) is
trying to found mountpoint/device in fstab and then in mtab.

I guess the original purpose for this feature is user-friendly 'mount
/node -o remount'. Unfortunately, the code does not check for
MS_REMOUNT flag and consults mtab in all cases. (The code is
consulting fstab/mtab before mount options parsing.)"

This commit documents this feature.

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

mount: clean up mount.8

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

Revert "mount: clean up mount.8"

This reverts commit 98c7944b52c54c86fb5b9d73fbead9442bfa0e30.

Unfortunately, the patch has been committed by "git commit -a" and
includes unwanted changes in configure.ac and sys-utils/Makefile.am...

Sorry.

mount: clean up mount.8

 * use "filesystem" everywhere (currently, the mount.8 man page is inconsistent
   and uses "file system" and "filesystem")

 * fix "The extN" to "The extN filesystem" (reported by Theodore Tso)

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

mount: add info about ext{3,4} barriers to mount.8

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

tests: test for basic functionality of sun labels

Tests the following:
* Creation and deletion of first two partitions
* setting sysid to one (randomly chosen) value
* setting readonly/mountable flags

Signed-off-by: Zdenek Behan <rain@matfyz.cz>

tests: fdisk doslabel test also checks setting partition active

* Yet another test, this time for setting active partition

Signed-off-by: Zdenek Behan <rain@matfyz.cz>

tests: fdisk doslabel test also checks changing partition type

* Changed apostrophes around test commands to quotes, this makes vim
syntax highlighting more happy, and makes the newlines stand out.

* Added check for changing partition type. After the change is done and
checked, the partition is deleted and re-created to not mess with the
other md5's already in place...

Signed-off-by: Zdenek Behan <rain@matfyz.cz>

fsck.8: formatting

Signed-off-by: Peter Breitenlohner <peb@mppmu.mpg.de>

ionice.1: formatting

Signed-off-by: Peter Breitenlohner <peb@mppmu.mpg.de>

hwclock.8: formatting

Signed-off-by: Peter Breitenlohner <peb@mppmu.mpg.de>

getopt.1: formatting

Signed-off-by: Peter Breitenlohner <peb@mppmu.mpg.de>

swapon.8: formatting

Signed-off-by: Peter Breitenlohner <peb@mppmu.mpg.de>

mount.8: formatting

Signed-off-by: Peter Breitenlohner <peb@mppmu.mpg.de>

losetup.8: formatting

Signed-off-by: Peter Breitenlohner <peb@mppmu.mpg.de>