platform/upstream/btrfs-progs.git
9 years agobtrfs-progs: autoconf: use paths and $*_LIBS from ./configure
Karel Zak [Tue, 9 Dec 2014 15:57:10 +0000 (16:57 +0100)]
btrfs-progs: autoconf: use paths and $*_LIBS from ./configure

Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: autoconf: check for build programs in ./configure
Karel Zak [Tue, 9 Dec 2014 15:41:10 +0000 (16:41 +0100)]
btrfs-progs: autoconf: check for build programs in ./configure

Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: autoconf: use ./configure to generate version.h
Karel Zak [Tue, 9 Dec 2014 14:58:11 +0000 (15:58 +0100)]
btrfs-progs: autoconf: use ./configure to generate version.h

The original homemade solution is unnecessary, autotools provides better
infrastructure to generate files.

Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: autoconf: use standard PACKAGE_* macros
Karel Zak [Tue, 9 Dec 2014 14:37:03 +0000 (15:37 +0100)]
btrfs-progs: autoconf: use standard PACKAGE_* macros

- use standard PACKAGE_{NAME,VERSION,STRING,URL,...} autoconf macros
  rather than homemade BTRFS_BUILD_VERSION

- don't #include version.h, now the file is necessary for library API only

Note that "btrfs version" returns "btrfs-progs <version>" instead of
the original confusing "btrfs <version>".

Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: autoconf: use config.h
Karel Zak [Tue, 9 Dec 2014 14:15:47 +0000 (15:15 +0100)]
btrfs-progs: autoconf: use config.h

- the header file is generated by ./configure, the standard autotools
  way is to use -include config.h on compiler command line rather than
  include the file directly from code

- remove _GNU_SOURCE from code, the macros is already defined in config.h
  by AC_USE_SYSTEM_EXTENSIONS autoconf macro

Signed-off-by: Karel Zak <kzak@redhat.com>
[_GNU_SOURCE changes already done]
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: autoconf: add ./configure script
Karel Zak [Thu, 4 Dec 2014 12:26:46 +0000 (13:26 +0100)]
btrfs-progs: autoconf: add ./configure script

- add ./autogen.sh script (necessary after git clean/clone)
- add ./configure.ac
- copy autotool helper scripts from automake
- modify version.sh to be usable from the configure script
- rename Makefile to Makefile.in and use basic variables from configure.ac

Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agoBtrfs progs v3.18.2
David Sterba [Tue, 27 Jan 2015 16:42:28 +0000 (17:42 +0100)]
Btrfs progs v3.18.2

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: update wording for qgroup limits
David Sterba [Tue, 27 Jan 2015 16:07:40 +0000 (17:07 +0100)]
btrfs-progs: update wording for qgroup limits

Based on a user report, 'max' in help does not clearly point to the
limits that are commonly used as a quota-related term.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: don't output operational status information to stderr
Xing Gu [Mon, 26 Jan 2015 08:26:38 +0000 (16:26 +0800)]
btrfs-progs: don't output operational status information to stderr

Signed-off-by: Xing Gu <gux.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: check: Add --readonly flag
Hugo Mills [Tue, 27 Jan 2015 15:05:53 +0000 (15:05 +0000)]
btrfs-progs: check: Add --readonly flag

Particularly during support conversations, people get confused about
which options to use with btrfs check. Adding a flag, --readonly, which
implies the default read-only behaviour and which conflicts with the
read-write operations, should help make the behaviour of the tool clear.

Signed-off-by: Hugo Mills <hugo@carfax.org.uk>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Make option parsing more robust to code modifications
Hugo Mills [Tue, 27 Jan 2015 15:05:52 +0000 (15:05 +0000)]
btrfs-progs: Make option parsing more robust to code modifications

The current approach to option parsing, where long-only options are
selected on the basis of their position in the long_options array is
fragile and painful to modify if options are to be inserted into the
list, rather than appended.

Instead, use the last field of struct option to return a value which
cannot be a char (and hence a short option), and simply switch on those
within the case statement.

Signed-off-by: Hugo Mills <hugo@carfax.org.uk>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: qgroup: make large size aligned
Fan Chengniang [Fri, 23 Jan 2015 01:50:49 +0000 (09:50 +0800)]
btrfs-progs: qgroup: make large size aligned

Problem: when the size is too big, the output format will be unaligned.
The __update__columns_max_len function has been updated to fix this
problem.

Signed-off-by: Fan Chengniang <fancn.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: drop feature defines from C files, in favour of CFLAGS defines
Dimitri John Ledkov [Sun, 18 Jan 2015 01:52:58 +0000 (01:52 +0000)]
btrfs-progs: drop feature defines from C files, in favour of CFLAGS defines

glibc 2.10+ (5+ years old) enables all the desired features:
_XOPEN_SOURCE 700, __XOPEN2K8, POSIX_C_SOURCE, DEFAULT_SOURCE; with a
single _GNU_SOURCE define in the makefile alone. For portability to
other libc implementations (e.g. dietlibc) _XOPEN_SOURCE=700 is also
defined.

This also resolves Debian bug report filed by Michael Tautschnig -
"Inconsistent use of _XOPEN_SOURCE results in conflicting
declarations". Whilst I was not able to reproduce the results, the
reported fact is that _XOPEN_SOURCE set to 500 in one set of files
(e.g. cmds-filesystem.c) generates/defines different struct stat from
other files (cmds-replace.c).

This patch thus cleans up all feature defines, and sets them at a
consistent level.

Bug-Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=747969
Signed-off-by: Dimitri John Ledkov <dimitri.j.ledkov@intel.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: unify getopt table terminators
David Sterba [Wed, 21 Jan 2015 17:53:02 +0000 (18:53 +0100)]
btrfs-progs: unify getopt table terminators

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: add --human-readable option where applicable
David Sterba [Wed, 21 Jan 2015 17:47:24 +0000 (18:47 +0100)]
btrfs-progs: add --human-readable option where applicable

Add an alias to -h to 'filesystem usage', 'filesystem df' and
'device usage' commands, same as the traditional 'df'.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: use predefined getopt values for unit suffixes
David Sterba [Wed, 21 Jan 2015 17:09:32 +0000 (18:09 +0100)]
btrfs-progs: use predefined getopt values for unit suffixes

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: define common geopt values for unit suffixes
David Sterba [Wed, 21 Jan 2015 17:02:42 +0000 (18:02 +0100)]
btrfs-progs: define common geopt values for unit suffixes

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: make btrfs qgroups show human readable sizes
Fan Chengniang [Mon, 19 Jan 2015 07:18:18 +0000 (15:18 +0800)]
btrfs-progs: make btrfs qgroups show human readable sizes

add --raw, --si, --iec, --kbytes, --mbytes, --gbytes, --tbytes options
make columns which show sizes align to right. Others aligned to left.

example:
qgroupid         rfer         excl     max_rfer     max_excl parent  child
--------         ----         ----     --------     -------- ------  -----
0/5         299.58MiB    299.58MiB    300.00MiB    300.00MiB 1/1     ---
0/265       299.58MiB     16.00KiB    400.00MiB        0.00B 1/1     ---
0/266       299.58MiB     16.00KiB    350.00MiB        0.00B ---     ---
1/1         599.16MiB    299.59MiB    800.00MiB        0.00B ---     0/5,0/265

Signed-off-by: Fan Chengniang <fancn.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: unify header file inclusion protections
David Sterba [Wed, 21 Jan 2015 16:49:26 +0000 (17:49 +0100)]
btrfs-progs: unify header file inclusion protections

There are missing ifdefs or defines with very generic names.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Doc: Add explain on 'source' and 'target' for btrfs-image.
Qu Wenruo [Mon, 24 Nov 2014 07:27:44 +0000 (15:27 +0800)]
btrfs-progs: Doc: Add explain on 'source' and 'target' for btrfs-image.

Add explain on 'source' and 'target', which is somewhat confusing for
users who want to restore dumped image.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[reworded texts]
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Cleanup, use bitshift instead of immediate number in btrfs_open_ctree_flags.
Qu Wenruo [Mon, 19 Jan 2015 06:45:03 +0000 (14:45 +0800)]
btrfs-progs: Cleanup, use bitshift instead of immediate number in btrfs_open_ctree_flags.

Change the immediate number in btrfs_open_ctree_flags to bit shift.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: btrfstune: force to set seeding flags
Fan Chengniang [Mon, 19 Jan 2015 07:27:30 +0000 (15:27 +0800)]
btrfs-progs: btrfstune: force to set seeding flags

Now we can use -f with -S option when setting seeding flags or clearing
seeding flags

Reported-by: Chen Hanxiao <chenhanxiao@cn.fujitsu.com>
Signed-off-by: Fan Chengniang <fancn.fnst@cn.fujitsu.com>
[updated docs]
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: tests, use non-interactive sudo helper
David Sterba [Mon, 19 Jan 2015 18:09:28 +0000 (19:09 +0100)]
btrfs-progs: tests, use non-interactive sudo helper

Sudo may not be configured to run without user prompt, in that case the
tests would be stuck.

Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: fix check of running scrub
David Sterba [Mon, 19 Jan 2015 17:47:20 +0000 (18:47 +0100)]
btrfs-progs: fix check of running scrub

Scrub on multiple devices may report wrong status if scrub finishes
early on one of them.

Reported-by: Sandy McArthur Jr <sandymac@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: make getopt tables static const
David Sterba [Mon, 19 Jan 2015 12:44:49 +0000 (13:44 +0100)]
btrfs-progs: make getopt tables static const

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: cleanup, move getop long options close to their use
David Sterba [Mon, 19 Jan 2015 12:30:06 +0000 (13:30 +0100)]
btrfs-progs: cleanup, move getop long options close to their use

Move long_option defintions just before getopt_long everywhere.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: tests, adjust alignment of the pretty command name
David Sterba [Wed, 14 Jan 2015 17:32:17 +0000 (18:32 +0100)]
btrfs-progs: tests, adjust alignment of the pretty command name

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: tests, use the root helper in 012
David Sterba [Wed, 14 Jan 2015 17:13:21 +0000 (18:13 +0100)]
btrfs-progs: tests, use the root helper in 012

We need it to mount/umount, to traverse lost+found and stat the results.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: tests, add support for running commands under root
David Sterba [Wed, 14 Jan 2015 17:07:43 +0000 (18:07 +0100)]
btrfs-progs: tests, add support for running commands under root

Most of the checks run fine without root, but some of them may need to
do a mount test or access the data. Add the support to selectively run
commands under root, hardcoded to sudo for now.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Move extent tree rebuild test to its dir
Qu Wenruo [Thu, 25 Dec 2014 01:32:15 +0000 (09:32 +0800)]
btrfs-progs: Move extent tree rebuild test to its dir

Move extent tree rebuild teset to its dir.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Move leaf-corruption no extent data case and add verification script
Qu Wenruo [Thu, 25 Dec 2014 01:32:14 +0000 (09:32 +0800)]
btrfs-progs: Move leaf-corruption no extent data case and add verification script

Move leaf-corruption type no extent data case to its dir, and add
verification script in test.sh

The verification script is based on manual btrfs-debug-tree check.

The image can also be reused to other leaf-corruption type, like
corrupted leaf contains regular file extent data case.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Move bad root items test cases to its corresponding dir
Qu Wenruo [Thu, 25 Dec 2014 01:32:13 +0000 (09:32 +0800)]
btrfs-progs: Move bad root items test cases to its corresponding dir

Now 006-bad-root-items has two cases, one for default case, and one for
skinny metadata case.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Move btrfs-image dump to corresponding dir
Qu Wenruo [Thu, 25 Dec 2014 01:32:12 +0000 (09:32 +0800)]
btrfs-progs: Move btrfs-image dump to corresponding dir

Move these obvious btrfs-image to its corresponding dir to use the new
infrastructure.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: New btrfsck test infrastructure
Qu Wenruo [Thu, 25 Dec 2014 01:32:11 +0000 (09:32 +0800)]
btrfs-progs: New btrfsck test infrastructure

Change the old btrfsck test infrastructure (btrfs-image dump or xz raw
dump) to the new test infrastructure.

1) Test case layout
The new infrastructure is dir based, each dir is one test type, and can
contain multiple images/scripts for different corner cases.
So layout will be the following:
btrfs-progs
|-tests
    |-fsck-tests
|-001-SOME-CORRUPT-TYPE
|-IMAGE-FOR-CASE1
|-IMAGE-FOR-CASE2

2) Test case image types
Only 2 types for test case images.
a) btrfs-image dump
   This one is the simplest case, one only needs to add the image to
   corresponding dir.

b) custom script
   This one is for all the resting cases which can't fit btrfs-image,
   like csum error or script can generate the image (this reduces the
   size obviously and good for review)
   The old binary dump also belongs to this type, so need to add script
   to extract them.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: remove uneccessary subvol name check for subvol deletion
Gui Hecheng [Thu, 18 Dec 2014 03:32:35 +0000 (11:32 +0800)]
btrfs-progs: remove uneccessary subvol name check for subvol deletion

When we want to delete a subvol, we first check to see whether it is
a subvolume or not. After the check, we are sure that it is a valid
subvol, don't have to check its name.

Signed-off-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
[removed unused variable 'len']
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: btrfstune: fix multiple options error
Fan Chengniang [Fri, 9 Jan 2015 08:11:42 +0000 (16:11 +0800)]
btrfs-progs: btrfstune: fix multiple options error

when we use multiple options, error return status will be override by the
last option status.

example: btrfstune -S 1 -r /dev/loop0

when -S option fails and -r option succeeds, return value is 0, rather than
1, where 1 is the right return status.

Reported-by: Chen Hanxiao <chenhanxiao@cn.fujitsu.com>
Signed-off-by: Fan Chengniang <fancn.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: add missing options to qgroup limit
Fan Chengniang [Tue, 13 Jan 2015 03:53:16 +0000 (11:53 +0800)]
btrfs-progs: add missing options to qgroup limit

btrfs qgroup limit has two options -c and -e,. They were forgotten to add
to manpage.

Signed-off-by: Fan Chengniang <fancn.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Fix wrong return value when executing 'fi show' on umounted device.
Qu Wenruo [Wed, 14 Jan 2015 06:50:40 +0000 (14:50 +0800)]
btrfs-progs: Fix wrong return value when executing 'fi show' on umounted device.

When executing 'btrfs fi show' on unmounted device, even no problem
happens, the return value is still 1 not 0.

The problem lies in search_umounted_fs_uuids(), where when it finds the
given uuid, it should return 1, but later uuid copy overwrites the return
value, causing it always return 0 under that case.

Fix it by pass found as pointer, and return value only indicates
whether anything wrong happens, whether found or not is stored in the
new parameter.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: restore backtrace functionality on glibc
Merlijn Wajer [Wed, 14 Jan 2015 12:13:15 +0000 (13:13 +0100)]
btrfs-progs: restore backtrace functionality on glibc

Originally broke in commit c2691f807ddd2c6b261c5707f6838a45d9275390

__glibc__ should have been __GLIBC__

We also include features.h ; although most includes (at least stdlib.h)
typically already include it -- at least on glibc, where it matters.

Signed-off-by: Merlijn Wajer <merlijn@wizzup.org>
Reported-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Remove redundant short description from cmd_snapshot_usage
Martin Volf [Fri, 9 Jan 2015 21:27:21 +0000 (22:27 +0100)]
btrfs-progs: Remove redundant short description from cmd_snapshot_usage

cmd_snapshot_usage in cmds-subvolume.c contains the short description
twice. Remove the first one.

Signed-off-by: Martin Volf <martin.volf.42@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agoBtrfs progs v3.18.1
David Sterba [Fri, 9 Jan 2015 18:22:29 +0000 (19:22 +0100)]
Btrfs progs v3.18.1

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Don't increase error count if the inode can be repaired.
Qu Wenruo [Mon, 22 Dec 2014 02:53:36 +0000 (10:53 +0800)]
btrfs-progs: Don't increase error count if the inode can be repaired.

The original check_inode_recs() will return -1 if found any error in a
inode_record. This is OK for original design since there is almost
nothing can repair at that time.

However more and more error from nlink mismatch to missing inode item
can be repaired in try_repair_inode(), check_inode_recs() should not
increase the error count if the inode can be repair.

With this patch, repair function for leaf-corruption will not return
error if all corruption inode can be recovered.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: fix some format errors in doc
Fan Chengniang [Fri, 9 Jan 2015 07:08:56 +0000 (15:08 +0800)]
btrfs-progs: fix some format errors in doc

Signed-off-by: Fan Chengniang <fancn.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc: fix format of btrfs-replace
Tsutomu Itoh [Thu, 8 Jan 2015 02:40:53 +0000 (11:40 +0900)]
btrfs-progs: doc: fix format of btrfs-replace

Current 'man btrfs-replace' is as follows:

========================================================================
...
...
           -f
               force using and overwriting <targetdev> even if it looks like
               containing a valid btrfs filesystem.

               A valid filesystem is assumed if a btrfs superblock is found
               which contains a correct checksum. Devices which are currently
               mounted are never allowed to be used as the <targetdev>. -B::::
               no background replace.
...
...
========================================================================

The format of 'B' option is wrong. So, fix it.

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Fix a copy-n-paste bug in btrfs_read_fs_root().
Qu Wenruo [Wed, 7 Jan 2015 09:23:43 +0000 (17:23 +0800)]
btrfs-progs: Fix a copy-n-paste bug in btrfs_read_fs_root().

Introduced in commit 96ec888aad41969 ("btrfs-progs: add quota group
verify code").

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Fix a buffer overflow causing segfault in fstests/btrfs/069
Qu Wenruo [Tue, 6 Jan 2015 09:30:44 +0000 (17:30 +0800)]
btrfs-progs: Fix a buffer overflow causing segfault in fstests/btrfs/069

The newly introduced search_chunk_tree_for_fs_info() won't count devid 0
in fi_arg->num_devices, which will cause buffer overflow since later
get_device_info() will fill di_args with devid.

This can be trigger by fstests/btrfs/069 and any operations needs to
iterate over all the devices like 'fi show' or 'dev stat' while
replacing.

The fix is do an extra probe specifically for devid 0 after
search_chunk_tree_for_fs_info() and change num_devices if needed.

Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Documentation: fix broken conversion
Peter Wu [Mon, 5 Jan 2015 14:10:09 +0000 (15:10 +0100)]
btrfs-progs: Documentation: fix broken conversion

Fixes various formatting issues:

Fixes quoting issues in various manpages. Before/After excerpts for
mkfs.btrfs(8), btrfs-qgroup(8):

    mkfs.btrfs [-A|--alloc-start '<alloc-start>']
    mkfs.btrfs [-A|--alloc-start <alloc-start>]
    id 0/<subvolume id>''
    id 0/<subvolume id>
    If multiple '<attr>'s is given, use comma to separate.
    If multiple <attr>s is given, use comma to separate.

The one that is most apparent is the removal of text in
btrfs-subvolume(8) and others. Before/After:

    list [options] [-G <value>] [-C <value>] [--sort=rootid,gen,ogen,path] <path>
    list [options] [-G [+|-]<value>] [-C [+|-]<value>] [--sort=rootid,gen,ogen,path] <path>
    create [-i <qgroupid>] <name>
    create [-i <qgroupid>] [<dest>]<name>

This happens because `<foo>` was replaced by `'<foo>'` in the sed
expression, which is then treated as a constrained quote. As a result,
the `[...]` before the string gets interpreted as "quoted text
attributes".

In this patch, the sed expression is dropped and asciidoc is configured
to recognize `<...>` as a unconstrained quoted string (such that
`<attr>s` is correctly emphasized) and to avoid adding quotes in
contexts where it is not needed.

A remaining problem is that some texts (the one between brackets) do not
add the emphasis for `<...>`. While this could be solved by replacing
`[` and `]` with `&#91;` and `&#93;` using sed, it introduces formatting
problems in btrfs-zero-log(8) because the context is ignored.

Signed-off-by: Peter Wu <peter@lekensteyn.nl>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Fix a extent buffer leak in count_csum_range().
Qu Wenruo [Mon, 5 Jan 2015 07:56:15 +0000 (15:56 +0800)]
btrfs-progs: Fix a extent buffer leak in count_csum_range().

The commit f495a2ac6611 ("btrfs-progs: fsck: remove unfriendly BUG_ON()
for searching tree failure") is causing tons of extent buffer leak if some
csum mismatches in btrfsck.

This is caused by a misplaced btrfs_release_path(), fix it.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Allow debug-tree to be executed on regular file.
Qu Wenruo [Mon, 5 Jan 2015 06:24:27 +0000 (14:24 +0800)]
btrfs-progs: Allow debug-tree to be executed on regular file.

The commit 1bad43fbe002 ("btrfs-progs: refine btrfs-debug-tree error
prompt when a mount point given")
add judgement on btrfs-debug-tree to restrict only block device to be
executed on, but the command can also be used on regular file, so add
regular file support for the judgement.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: convert, handle pwrite error in do_rollback
David Sterba [Tue, 6 Jan 2015 17:54:48 +0000 (18:54 +0100)]
btrfs-progs: convert, handle pwrite error in do_rollback

Resolves-coverity-id: 1242984
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: handle orphan creation error in btrfs_unlink
David Sterba [Tue, 6 Jan 2015 17:46:03 +0000 (18:46 +0100)]
btrfs-progs: handle orphan creation error in btrfs_unlink

The error handling is simple and just returns an error upward, no
cleanup.

Resolves-coverity-id: 1260251
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: fix leak of path in reset_balance
David Sterba [Sat, 3 Jan 2015 02:33:31 +0000 (03:33 +0100)]
btrfs-progs: fix leak of path in reset_balance

A path may leak after an unlikely error condition.

Resolves-coverity-id: 1199474
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: safely copy path in btrfs_scan_lblkid
David Sterba [Sat, 3 Jan 2015 02:24:10 +0000 (03:24 +0100)]
btrfs-progs: safely copy path in btrfs_scan_lblkid

We should terminate the string, let's use the wrapper.

Resolves-coverity-id: 1127097
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: cleanup whitespace in is_same_blk_file
David Sterba [Sat, 3 Jan 2015 02:13:11 +0000 (03:13 +0100)]
btrfs-progs: cleanup whitespace in is_same_blk_file

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: use strncpy in is_same_blk_file
David Sterba [Sat, 3 Jan 2015 02:10:55 +0000 (03:10 +0100)]
btrfs-progs: use strncpy in is_same_blk_file

Although we're getting strings from the caller, let's be safe.

Resolves-coverity-id: 1199475
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: check, handle error in fix_key_order
David Sterba [Sat, 3 Jan 2015 01:47:32 +0000 (02:47 +0100)]
btrfs-progs: check, handle error in fix_key_order

In case the buffer is corrupted and the for loop does not happen, we'd
return garbage. The caller retunrs -EIO in case of any corruption, use
that value in fix_key_order.

Resolves-coverity-id: 1246944
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc subvolume, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:27:41 +0000 (02:27 +0100)]
btrfs-progs: doc subvolume, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc scrub, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:22:10 +0000 (02:22 +0100)]
btrfs-progs: doc scrub, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc replace, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:20:50 +0000 (02:20 +0100)]
btrfs-progs: doc replace, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc quota, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:20:06 +0000 (02:20 +0100)]
btrfs-progs: doc quota, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc qgroup, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:19:31 +0000 (02:19 +0100)]
btrfs-progs: doc qgroup, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc property, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:18:15 +0000 (02:18 +0100)]
btrfs-progs: doc property, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc inspect-internal, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:17:28 +0000 (02:17 +0100)]
btrfs-progs: doc inspect-internal, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc device, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:15:55 +0000 (02:15 +0100)]
btrfs-progs: doc device, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc, undocument old balance syntax
David Sterba [Sat, 3 Jan 2015 01:12:56 +0000 (02:12 +0100)]
btrfs-progs: doc, undocument old balance syntax

Standalone balance subcommand has been around since 3.3.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc balance, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 01:09:08 +0000 (02:09 +0100)]
btrfs-progs: doc balance, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc filesystem, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 00:49:24 +0000 (01:49 +0100)]
btrfs-progs: doc filesystem, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc btrfs, sort subcommands alphabetically
David Sterba [Sat, 3 Jan 2015 00:45:17 +0000 (01:45 +0100)]
btrfs-progs: doc btrfs, sort subcommands alphabetically

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc restore, cleanups for wiki conversion
David Sterba [Sat, 3 Jan 2015 00:37:40 +0000 (01:37 +0100)]
btrfs-progs: doc restore, cleanups for wiki conversion

Properly format the patch regex and add a line break.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc filesystem, cleanups for wiki conversion
David Sterba [Sat, 3 Jan 2015 00:36:39 +0000 (01:36 +0100)]
btrfs-progs: doc filesystem, cleanups for wiki conversion

Remove trailing space, add line break before NOTE.

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc balance, cleanups for wiki conversion
David Sterba [Sat, 3 Jan 2015 00:34:05 +0000 (01:34 +0100)]
btrfs-progs: doc balance, cleanups for wiki conversion

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc mkfs, cleanups for wiki conversion
David Sterba [Sat, 3 Jan 2015 00:30:25 +0000 (01:30 +0100)]
btrfs-progs: doc mkfs, cleanups for wiki conversion

- remove trailing whitespace
- use +mono+ formatting for command line example

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc mkfs, mark leafsize deprecated
David Sterba [Sat, 3 Jan 2015 00:26:31 +0000 (01:26 +0100)]
btrfs-progs: doc mkfs, mark leafsize deprecated

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc: fix incorrect format of 'l' option in mkfs.btrfs
Tsutomu Itoh [Wed, 24 Dec 2014 07:10:52 +0000 (16:10 +0900)]
btrfs-progs: doc: fix incorrect format of 'l' option in mkfs.btrfs

The format of 'l' option in mkfs.btrfs.txt is wrong.
And, when the head of the character string is 65536, the following warning
is displayed.

    $ make
    Making all in Documentation
        [ASCII]  mkfs.btrfs.xml
    asciidoc: WARNING: mkfs.btrfs.xml.tmp1: line 67: list item index: expected 1 got 65536
        [XMLTO]  mkfs.btrfs.8
        [GZ]     mkfs.btrfs.8.gz
    rm mkfs.btrfs.8 mkfs.btrfs.xml

So, fix it.

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc filesystem, add unit suffixes to option descriptions
David Sterba [Fri, 2 Jan 2015 16:51:06 +0000 (17:51 +0100)]
btrfs-progs: doc filesystem, add unit suffixes to option descriptions

Reported-by: Chris Murphy <lists@colorremedies.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: doc filesystem, fix quoting of size suffixes
David Sterba [Fri, 2 Jan 2015 16:22:07 +0000 (17:22 +0100)]
btrfs-progs: doc filesystem, fix quoting of size suffixes

Escape the first quote of two to cancel effects of ordinary '...' .

Reported-by: Chris Murphy <lists@colorremedies.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agoBtrfs progs v3.18
David Sterba [Tue, 30 Dec 2014 16:23:12 +0000 (17:23 +0100)]
Btrfs progs v3.18

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: fix minor leak of dev_info in btrfs_scan_kernel
David Sterba [Tue, 30 Dec 2014 14:37:29 +0000 (15:37 +0100)]
btrfs-progs: fix minor leak of dev_info in btrfs_scan_kernel

Resolves-coverity-id: 1127098
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: fix overflow check in btrfs_insert_inode_ref
David Sterba [Tue, 30 Dec 2014 14:16:17 +0000 (15:16 +0100)]
btrfs-progs: fix overflow check in btrfs_insert_inode_ref

Resolves-coverity-id: 1260247
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: check, missing parens around compound block in find_normal_file_extent
David Sterba [Tue, 30 Dec 2014 14:05:50 +0000 (15:05 +0100)]
btrfs-progs: check, missing parens around compound block in find_normal_file_extent

Resolves-coverity-id: 1260248
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: fi show, don't leak canonical path
David Sterba [Tue, 30 Dec 2014 14:01:13 +0000 (15:01 +0100)]
btrfs-progs: fi show, don't leak canonical path

Resolves-coverity-id: 1260252
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: check, fix path leak in error branch
David Sterba [Tue, 30 Dec 2014 13:52:55 +0000 (14:52 +0100)]
btrfs-progs: check, fix path leak in error branch

Resolves-coverity-id: 1260250
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Documentation: add T/P/E description for resize cmd
Gui Hecheng [Tue, 30 Dec 2014 01:52:17 +0000 (09:52 +0800)]
btrfs-progs: Documentation: add T/P/E description for resize cmd

Signed-off-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: refine btrfs-debug-tree error prompt when a mount point given
Gui Hecheng [Thu, 25 Dec 2014 01:16:35 +0000 (09:16 +0800)]
btrfs-progs: refine btrfs-debug-tree error prompt when a mount point given

Now, if exec:
# btrfs-debug-tree <mount_point>
it echos:
: Superblock bytenr is larger than device size

But it is quite misleading, because it is a valid btrfs.
In this case, we should tell the developer to provide a block device.

After apply:
: '<mount_point>' is not a block device

Signed-off-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: move check_arg_type() to util.c
Gui Hecheng [Thu, 25 Dec 2014 01:16:34 +0000 (09:16 +0800)]
btrfs-progs: move check_arg_type() to util.c

The check_arg_type() function does quite generic thing, move it to
utils.c.

Signed-off-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Fix btrfs fi show by uuid and label
Justin Maggard [Tue, 23 Dec 2014 19:34:52 +0000 (11:34 -0800)]
btrfs-progs: Fix btrfs fi show by uuid and label

Commit 8be2fff (btrfs-progs: apply realpath for btrfs fi
show when mount point is given) changed the behavior of
btrfs fi show to return an error if the call to realpath()
failed.  This broke the ability to specify a filesystem by
uuid or label.

So let's not consider a failed call to realpath() as an
error.  If the user really specified a bad device, just
return nothing like we did before.

Signed-off-by: Justin Maggard <jmaggard10@gmail.com>
Reviewed-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Fix a bug in reset_nlink() which may deletes the recovered file
Qu Wenruo [Tue, 23 Dec 2014 01:29:44 +0000 (09:29 +0800)]
btrfs-progs: Fix a bug in reset_nlink() which may deletes the recovered file

In reset_nlink(), we believe rec->found_link as accurate number of the
valid links. However it only records the number of found DIR_ITEM, so we
can't use it as reliable value.

Before this patch, in some case, leaf corruption recovery will believe
there is a valid backref but don't add_link() since it can't find any
valid one and don't put it into the lost+found dir.
So the recovered inode will be considered as an orphan item without
orphan item and repair_inode_orphan_item() will add orphan item for it,
causing all the filename/filetype we recovered being a waste of time.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs fix wrong memory free on check_is_root
Satoru Takeuchi [Thu, 18 Dec 2014 07:39:26 +0000 (16:39 +0900)]
btrfs-progs fix wrong memory free on check_is_root

When "/" is Btrfs, "btrfs property <subcommand> /" regards it
as non-root by mistake.

check_is_root() regards @object as a file system root if
the following two conditions are satisfied.

 a) Both @object and its parent directory are Btrfs object
    (file system root, subvolume, inode, and device
    used for Btrfs).
 b) fsid of the above mentioned two objects are different.

It doesn't work if @object is "/" because, in this case,
fsid of "/" and its parent (it's also "/"), are the same.

* Test environment

Two Btrfs file system (not subvolume) "/" and "/home/sat/mnt".

* How to reproduce

Submit "btrfs prop get" against the above mentioned file systems.

* Test Result

** Actual result (without my patch)

==========================
ro=false
label=                 # label is displayed because it's a file system root
ro=false               # label is not displayed even if it's a file system root
==========================
** Expected result (with my patch)

==========================
ro=false
label=
ro=false
label=foo            # label is displayed
===========================

Signed-off-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Reported-by: Naohiro Aota <naota@elisp.net>
Reviewed-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: let btrfs_free_path accept NULL
David Sterba [Fri, 19 Dec 2014 16:01:40 +0000 (17:01 +0100)]
btrfs-progs: let btrfs_free_path accept NULL

Same in kernel and matches semantics of free().

Resolves-Coverity-CID: 1054881
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: check allocation result in add_clone_source
David Sterba [Fri, 19 Dec 2014 15:55:32 +0000 (16:55 +0100)]
btrfs-progs: check allocation result in add_clone_source

Resolves-Coverity-CID: 1054894
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: check result of first_cache_extent
David Sterba [Fri, 19 Dec 2014 15:22:59 +0000 (16:22 +0100)]
btrfs-progs: check result of first_cache_extent

If the tree's empty, we'll get NULL and dereference it.

Resolves-Coverity-CID: 1248828
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: fragments, close output file on error
David Sterba [Fri, 19 Dec 2014 15:15:39 +0000 (16:15 +0100)]
btrfs-progs: fragments, close output file on error

Resolves-Coverity-CID: 1258794
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: corrupt block, add break after option U
David Sterba [Fri, 19 Dec 2014 15:11:59 +0000 (16:11 +0100)]
btrfs-progs: corrupt block, add break after option U

Resolves-Coverity-CID: 1258793
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: corrupt block, add missing break to option I
David Sterba [Fri, 19 Dec 2014 15:08:48 +0000 (16:08 +0100)]
btrfs-progs: corrupt block, add missing break to option I

Using -I would imply -d.

Resolves-Coverity-CID: 1258792
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: move contents of math.h to utils.h and delete
David Sterba [Fri, 19 Dec 2014 14:31:00 +0000 (15:31 +0100)]
btrfs-progs: move contents of math.h to utils.h and delete

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: cleanup: avoid to use literal for getopt val
Satoru Takeuchi [Fri, 19 Dec 2014 05:31:54 +0000 (14:31 +0900)]
btrfs-progs: cleanup: avoid to use literal for getopt val

Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Enhance the document of btrfs property
Satoru Takeuchi [Fri, 19 Dec 2014 07:22:49 +0000 (16:22 +0900)]
btrfs-progs: Enhance the document of btrfs property

Enhance the document of btrfs property as follows.

- Add the description about the candidates of <name>.
- Enrich the description of <object>.
- Fix typos and some redundancy.

Signed-off-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Reported-by: Naohiro Aota <naota@elisp.net>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Remove deprecated _BSD_SOURCE macro.
Qu Wenruo [Fri, 19 Dec 2014 06:13:12 +0000 (14:13 +0800)]
btrfs-progs: Remove deprecated _BSD_SOURCE macro.

Fix the following gcc(>4.9) and clang warning:

In file included from cmds-receive.c:24:
In file included from ./kerncompat.h:22:
In file included from /usr/include/stdio.h:27:
/usr/include/features.h:148:3: warning: "_BSD_SOURCE and _SVID_SOURCE
are deprecated, use _DEFAULT_SOURCE" [-W#warnings]
  ^
1 warning generated.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Remove a unused function offset_to_bitmap()
Qu Wenruo [Fri, 19 Dec 2014 06:13:11 +0000 (14:13 +0800)]
btrfs-progs: Remove a unused function offset_to_bitmap()

Fix the following clang warning:

free-space-cache.c:464:19: warning: unused function 'offset_to_bitmap'
[-Wunused-function]
static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl,
                  ^
1 warning generated.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
9 years agobtrfs-progs: Remove a unused function root_gtp_mask().
Qu Wenruo [Fri, 19 Dec 2014 06:13:10 +0000 (14:13 +0800)]
btrfs-progs: Remove a unused function root_gtp_mask().

Fix the following clang warning when compiling btrfs-progs:

radix-tree.c:78:21: warning: unused function 'root_gfp_mask'
[-Wunused-function]
static inline gfp_t root_gfp_mask(struct radix_tree_root *root)
                    ^
1 warning generated.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>