platform/upstream/btrfs-progs.git
7 years agobtrfs-progs: move message helpers out of utils
David Sterba [Wed, 1 Feb 2017 15:14:06 +0000 (16:14 +0100)]
btrfs-progs: move message helpers out of utils

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move utils code out of header
David Sterba [Wed, 1 Feb 2017 15:09:47 +0000 (16:09 +0100)]
btrfs-progs: move utils code out of header

There are not performance critical static inlines, we can do the normal
declaration/definition split for various helpers.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: improve assert in make_convert_btrfs
David Sterba [Wed, 1 Feb 2017 14:41:06 +0000 (15:41 +0100)]
btrfs-progs: convert: improve assert in make_convert_btrfs

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: rename members that clash with other functions
David Sterba [Wed, 1 Feb 2017 14:06:04 +0000 (15:06 +0100)]
btrfs-progs: convert: rename members that clash with other functions

Rename 'free' and also 'used' for consistency.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs: use const char for label
David Sterba [Wed, 1 Feb 2017 13:27:24 +0000 (14:27 +0100)]
btrfs-progs: mkfs: use const char for label

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs: make make_cfg::blocks an internal member
David Sterba [Wed, 1 Feb 2017 13:25:18 +0000 (14:25 +0100)]
btrfs-progs: mkfs: make make_cfg::blocks an internal member

The caller of make_btrfs does not need to set it, it's for internal use
and can be read after mkfs ends.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs: describe fields of btrfs_mkfs_config
David Sterba [Wed, 1 Feb 2017 13:17:12 +0000 (14:17 +0100)]
btrfs-progs: mkfs: describe fields of btrfs_mkfs_config

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs: clear whole mkfs_cfg at once
David Sterba [Wed, 1 Feb 2017 13:15:08 +0000 (14:15 +0100)]
btrfs-progs: mkfs: clear whole mkfs_cfg at once

Zero out the structure and set only the requested fields.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: remove unused variables from docs makefile
David Sterba [Wed, 1 Feb 2017 12:37:42 +0000 (13:37 +0100)]
btrfs-progs: build: remove unused variables from docs makefile

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: use MAKEOPTS where missing
David Sterba [Wed, 1 Feb 2017 12:23:04 +0000 (13:23 +0100)]
btrfs-progs: build: use MAKEOPTS where missing

$(MAKE) should also use MAKEOPTS.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: drop deprecated utility from test dependencies
David Sterba [Wed, 1 Feb 2017 10:53:21 +0000 (11:53 +0100)]
btrfs-progs: build: drop deprecated utility from test dependencies

Standalone btrfs-debug-tree has been superseded by
'btrfs inspect-internal dump-tree'.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: move internal bg size definition
David Sterba [Wed, 1 Feb 2017 10:42:01 +0000 (11:42 +0100)]
btrfs-progs: convert: move internal bg size definition

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: rename ext2 function to create a symlink
David Sterba [Tue, 31 Jan 2017 10:08:09 +0000 (11:08 +0100)]
btrfs-progs: convert: rename ext2 function to create a symlink

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move mkfs helper implementation out of utils
David Sterba [Tue, 31 Jan 2017 22:13:54 +0000 (23:13 +0100)]
btrfs-progs: move mkfs helper implementation out of utils

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move more mkfs declarations to the common header
David Sterba [Tue, 31 Jan 2017 22:13:54 +0000 (23:13 +0100)]
btrfs-progs: move more mkfs declarations to the common header

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: kerncompat: print trace from ASSERT, if enabled
David Sterba [Tue, 31 Jan 2017 19:34:27 +0000 (20:34 +0100)]
btrfs-progs: kerncompat: print trace from ASSERT, if enabled

Commit bc2d4ccc46e2807964c74ce805fb6405ca2bc842 "btrfs-progs:
kerncompat: disconnect assert and warning messages" forgot to add the
print_trace call to assert_trace.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: build library by default
David Sterba [Mon, 30 Jan 2017 17:15:31 +0000 (18:15 +0100)]
btrfs-progs: build: build library by default

Add a convenience rule and build the library by default, not just at the
installation phase.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: add stub makefile to convert
David Sterba [Mon, 30 Jan 2017 17:00:39 +0000 (18:00 +0100)]
btrfs-progs: build: add stub makefile to convert

A shortcut to compile btrfs-convert from inside the directory.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: use bit field for convert flags
David Sterba [Mon, 30 Jan 2017 16:52:15 +0000 (17:52 +0100)]
btrfs-progs: convert: use bit field for convert flags

Use one flag field instead of several variables. The change cascades
down to the callchain and modifies several functions.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: use helper for special inode number check
David Sterba [Mon, 30 Jan 2017 16:14:03 +0000 (17:14 +0100)]
btrfs-progs: convert: use helper for special inode number check

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: better error handling in ext2_read_used_space
David Sterba [Mon, 30 Jan 2017 15:58:57 +0000 (16:58 +0100)]
btrfs-progs: convert: better error handling in ext2_read_used_space

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: remove unused includes
David Sterba [Mon, 30 Jan 2017 15:52:07 +0000 (16:52 +0100)]
btrfs-progs: convert: remove unused includes

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: use on-stack buffer for subvol name dir
David Sterba [Mon, 30 Jan 2017 13:20:30 +0000 (14:20 +0100)]
btrfs-progs: convert: use on-stack buffer for subvol name dir

Get rid of dynamic allocation.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: use fixed lenght array for source fs name
David Sterba [Mon, 30 Jan 2017 13:11:19 +0000 (14:11 +0100)]
btrfs-progs: convert: use fixed lenght array for source fs name

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: move struct initialization to the init function
David Sterba [Mon, 30 Jan 2017 12:41:36 +0000 (13:41 +0100)]
btrfs-progs: convert: move struct initialization to the init function

The context is zeroed in convert_open_fs after and overwrites the rbtree
initialization, which accidentally is the same (NULL).

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: add rule for building cscope index
David Sterba [Mon, 30 Jan 2017 12:29:22 +0000 (13:29 +0100)]
btrfs-progs: build: add rule for building cscope index

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: split makefile to generated and stable parts
David Sterba [Mon, 30 Jan 2017 10:04:16 +0000 (11:04 +0100)]
btrfs-progs: build: split makefile to generated and stable parts

It's not really necessary to configure and regenerate Makefiles in cases
like adding a new source file. The build environment and optional
features are not affected by that.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: add rule for ctags
David Sterba [Mon, 30 Jan 2017 09:49:55 +0000 (10:49 +0100)]
btrfs-progs: build: add rule for ctags

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: update some forward declarations
David Sterba [Mon, 30 Jan 2017 09:36:16 +0000 (10:36 +0100)]
btrfs-progs: convert: update some forward declarations

Add missing and reorder so they come next to the structures using them.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: use wider types types for inode counts for progress reports
David Sterba [Mon, 30 Jan 2017 09:33:01 +0000 (10:33 +0100)]
btrfs-progs: convert: use wider types types for inode counts for progress reports

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs: make list of source fs more visible
David Sterba [Mon, 30 Jan 2017 09:11:22 +0000 (10:11 +0100)]
btrfs-progs: mkfs: make list of source fs more visible

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: remove directory-specific include paths
David Sterba [Fri, 27 Jan 2017 18:12:08 +0000 (19:12 +0100)]
btrfs-progs: build: remove directory-specific include paths

The convert and image utilities used include, but we specify the headers
by path relative to the toplevel directory for clarity. The compiler
option is not needed.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: use target deps on commandline via automatic variable
David Sterba [Fri, 27 Jan 2017 18:09:05 +0000 (19:09 +0100)]
btrfs-progs: build: use target deps on commandline via automatic variable

The dependencies are duplicated in many targets, but we can use the $^
variable to simplify the command line.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: replace target names with automatic variable
David Sterba [Fri, 27 Jan 2017 18:00:10 +0000 (19:00 +0100)]
btrfs-progs: build: replace target names with automatic variable

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: reorder target dependencies
David Sterba [Fri, 27 Jan 2017 17:50:56 +0000 (18:50 +0100)]
btrfs-progs: build: reorder target dependencies

Reorder target dependencies:

1) target-specific objects
2) common objects
3) library objects
4) static libraries

1-3 are for faster detection of build failures, the specific objects
reflect changes to the source, but otherwise is only a cosmetic change.
The 4 is for correct order during linking, if we'd use the $^ variable.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: split LIBS
David Sterba [Fri, 27 Jan 2017 17:34:51 +0000 (18:34 +0100)]
btrfs-progs: build: split LIBS

Not all the build targets use all libs listed in LIBS, separate the
common and compression ones.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: list mkfs.btrfs build objects in a variable
David Sterba [Fri, 27 Jan 2017 16:55:25 +0000 (17:55 +0100)]
btrfs-progs: build: list mkfs.btrfs build objects in a variable

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: build: list convert build objects in a variable
David Sterba [Fri, 27 Jan 2017 16:55:25 +0000 (17:55 +0100)]
btrfs-progs: build: list convert build objects in a variable

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: move implementation for interal conversion API to own file
David Sterba [Fri, 27 Jan 2017 14:47:17 +0000 (15:47 +0100)]
btrfs-progs: convert: move implementation for interal conversion API to own file

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: move ext2 conversion out of main.c
David Sterba [Fri, 27 Jan 2017 15:56:37 +0000 (16:56 +0100)]
btrfs-progs: convert: move ext2 conversion out of main.c

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: move ext2 definitions out of main
David Sterba [Fri, 27 Jan 2017 15:51:03 +0000 (16:51 +0100)]
btrfs-progs: convert: move ext2 definitions out of main

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: move definitions for interal conversion API to own file
David Sterba [Fri, 27 Jan 2017 14:47:17 +0000 (15:47 +0100)]
btrfs-progs: convert: move definitions for interal conversion API to own file

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move fs features implementation to own file
David Sterba [Thu, 26 Jan 2017 12:27:58 +0000 (13:27 +0100)]
btrfs-progs: move fs features implementation to own file

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move fs features declarations to own header from utils
David Sterba [Thu, 26 Jan 2017 12:13:36 +0000 (13:13 +0100)]
btrfs-progs: move fs features declarations to own header from utils

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: move common api implementation to own file
David Sterba [Wed, 25 Jan 2017 16:21:02 +0000 (17:21 +0100)]
btrfs-progs: convert: move common api implementation to own file

Lots of moved code but no actual changes.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs: move common api implementation to own file
David Sterba [Wed, 25 Jan 2017 16:21:02 +0000 (17:21 +0100)]
btrfs-progs: mkfs: move common api implementation to own file

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move convert definitions to own header
David Sterba [Wed, 25 Jan 2017 16:09:10 +0000 (17:09 +0100)]
btrfs-progs: move convert definitions to own header

Create a header for filesystem conversion API, the config and the main
entry function.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move mkfs definitions to own header
David Sterba [Wed, 25 Jan 2017 15:54:35 +0000 (16:54 +0100)]
btrfs-progs: move mkfs definitions to own header

Create a header for filesystem creation API, the config and the main
entry function.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move some common definitions to own header
David Sterba [Wed, 25 Jan 2017 15:54:35 +0000 (16:54 +0100)]
btrfs-progs: move some common definitions to own header

Start a header for various defines that are supposed to be used in any
context or subpart.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move help implemetnation to own file
David Sterba [Wed, 25 Jan 2017 15:35:00 +0000 (16:35 +0100)]
btrfs-progs: move help implemetnation to own file

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: move help defines to own header
David Sterba [Wed, 25 Jan 2017 15:35:00 +0000 (16:35 +0100)]
btrfs-progs: move help defines to own header

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: find-root: rename usage helper
David Sterba [Wed, 25 Jan 2017 15:23:24 +0000 (16:23 +0100)]
btrfs-progs: find-root: rename usage helper

This would clash with the generic usage from commands.h.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: Introduce macro to calculate backup superblock offset
Qu Wenruo [Tue, 24 Jan 2017 03:03:06 +0000 (11:03 +0800)]
btrfs-progs: Introduce macro to calculate backup superblock offset

Introduce a new macro, BTRFS_SB_OFFSET() to calculate backup superblock
offset, this is handy if one wants to initialize static array at
declaration time.

Suggested-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: Introduce kernel sizes to cleanup large intermediate number
Qu Wenruo [Tue, 24 Jan 2017 03:03:05 +0000 (11:03 +0800)]
btrfs-progs: Introduce kernel sizes to cleanup large intermediate number

Large numbers like (1024 * 1024 * 1024) may cost reader/reviewer to
waste one second to convert to 1G.

Introduce kernel include/linux/sizes.h to replace any intermediate
number larger than 4096 (not including 4096) to SZ_*.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: introduce global config
David Sterba [Mon, 21 Nov 2016 12:52:01 +0000 (13:52 +0100)]
btrfs-progs: introduce global config

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agoBtrfs progs v4.9.1
David Sterba [Fri, 27 Jan 2017 14:15:59 +0000 (15:15 +0100)]
Btrfs progs v4.9.1

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: update CHANGES for v4.9.1
David Sterba [Wed, 25 Jan 2017 09:14:32 +0000 (10:14 +0100)]
btrfs-progs: update CHANGES for v4.9.1

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs/convert: separate the convert part from make_btrfs
David Sterba [Tue, 24 Jan 2017 18:36:15 +0000 (19:36 +0100)]
btrfs-progs: mkfs/convert: separate the convert part from make_btrfs

The regulare mkfs_btrfs does not anything special about convert and just
passes the arguments. Make that two functions.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: tests: use built binaries for 004-send-parent-multi-subvol
David Sterba [Tue, 24 Jan 2017 17:19:27 +0000 (18:19 +0100)]
btrfs-progs: tests: use built binaries for 004-send-parent-multi-subvol

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: tests: add variable quotation to cli-tests
David Sterba [Tue, 24 Jan 2017 16:59:47 +0000 (17:59 +0100)]
btrfs-progs: tests: add variable quotation to cli-tests

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: tests: add test for --sync option of qgroup show
Tsutomu Itoh [Thu, 15 Dec 2016 04:33:05 +0000 (13:33 +0900)]
btrfs-progs: tests: add test for --sync option of qgroup show

Simple test script for the following patch.

   btrfs-progs: qgroup: add sync option to 'qgroup show'

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: qgroup show: do not error if sync fails
David Sterba [Tue, 24 Jan 2017 16:46:10 +0000 (17:46 +0100)]
btrfs-progs: qgroup show: do not error if sync fails

The sync call is unlikely to fail so we optimistically turn it to a
warning if that happens so we return inaccurate numbers instead of a
hard exit.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: qgroup: change the value of sort option
Tsutomu Itoh [Thu, 15 Dec 2016 04:30:55 +0000 (13:30 +0900)]
btrfs-progs: qgroup: change the value of sort option

The value of sort option ('S') is not used for option letter.
Therefore, I'll change the single letter to non-character.

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: qgroup: add sync option to 'qgroup show'
Tsutomu Itoh [Thu, 15 Dec 2016 04:29:28 +0000 (13:29 +0900)]
btrfs-progs: qgroup: add sync option to 'qgroup show'

The 'qgroup show' command does not synchronize filesystem.
Therefore, 'qgroup show' may not display the correct value unless
synchronized with 'filesystem sync' command etc.

So add the '--sync' option so that we can choose whether or not
to synchronize when executing the command.

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: kerncompat: simplify warning_trace
David Sterba [Tue, 24 Jan 2017 15:05:24 +0000 (16:05 +0100)]
btrfs-progs: kerncompat: simplify warning_trace

The value of assertion is always set.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: kerncompat: disconnect assert and warning messages
David Sterba [Tue, 24 Jan 2017 15:02:44 +0000 (16:02 +0100)]
btrfs-progs: kerncompat: disconnect assert and warning messages

The assertion and warn/bug have reversed condition checks, using the
same helpers drops the exact value, so we'll print the message directly
without the helpers.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: kerncompat: pass exact condition value from ASSERT
David Sterba [Tue, 24 Jan 2017 12:17:25 +0000 (13:17 +0100)]
btrfs-progs: kerncompat: pass exact condition value from ASSERT

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: kerncompat: add separate trace print for BUG_ON
David Sterba [Tue, 24 Jan 2017 12:15:41 +0000 (13:15 +0100)]
btrfs-progs: kerncompat: add separate trace print for BUG_ON

We want to pass unmodified condition down to the handlers so we can't
use assert_trace for that purpose.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: Fix disable backtrace assert error
Qu Wenruo [Thu, 19 Jan 2017 04:00:03 +0000 (12:00 +0800)]
btrfs-progs: Fix disable backtrace assert error

Due to commit 00e769d04c2c83029d6c71(btrfs-progs: Correct value printed
by assertions/BUG_ON/WARN_ON), which changed the assert_trace()
parameter, the condition passed to assert/WARN_ON/BUG_ON are logical
notted for backtrace enabled and disabled case.

Such behavior makes us easier to pass value wrong, and in fact it did
cause us to pass wrong condition for ASSERT().

Instead of passing different conditions for ASSERT/WARN_ON/BUG_ON()
manually, this patch will use ASSERT() to implement the resting
ASSERT/WARN_ON/BUG() for disable backtrace case, and use assert_trace()
to implement ASSERT() and BUG_ON(), to allow them to print correct
value.

Also, move WARN_ON() out of the ifdef branch, as it's completely the
same for both branches.

Cc: Goldwyn Rodrigues <rgoldwyn@suse.de>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: quota: fix printing during wait mode
Jeff Mahoney [Wed, 18 Jan 2017 01:56:39 +0000 (20:56 -0500)]
btrfs-progs: quota: fix printing during wait mode

If we call "btrfs quota rescan -w", it will attempt to start the rescan
operation, wait for it, and then print the "quota rescan started" message.
The wait could last an arbitrary amount of time, so printing it after
the wait isn't very helpful.

This patch reworks how we print the rescan started message as well as the
printing of the messages, including adding an error message for status
query failures (which could be EPERM/EFAULT/ENOMEM, not just no rescan
in progress) and wait failures.

Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: sanitize - Use correct source for memcpy
Goldwyn Rodrigues [Fri, 20 Jan 2017 19:03:33 +0000 (13:03 -0600)]
btrfs-progs: sanitize - Use correct source for memcpy

While performing a memcpy, we are copying from uninitialized dst
as opposed to src->data. Though using eb->len is correct, I used
src->len to make it more readable.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: check: fix false alert on dropped leaf in lowmem mode
Qu Wenruo [Wed, 18 Jan 2017 05:21:07 +0000 (13:21 +0800)]
btrfs-progs: check: fix false alert on dropped leaf in lowmem mode

For btrfs-progs test case 021-partially-dropped-snapshot-case, if the
first leaf is already dropped, btrfs check low-memory mode will report
false alert:

checking fs roots
checksum verify failed on 29917184 found E4E3BDB6 wanted 00000000
checksum verify failed on 29917184 found E4E3BDB6 wanted 00000000
checksum verify failed on 29917184 found E4E3BDB6 wanted 00000000
checksum verify failed on 29917184 found E4E3BDB6 wanted 00000000

This is caused by we are calling check_fs_first_inode() function,
unlike the rest part of check_fs_root_v2(), it doesn't have enough check
on dropping progress, and caused the false alert.

Fix it by checking dropping progress before searching slot.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: tests: 005-qgroup-show
David Sterba [Fri, 20 Jan 2017 16:54:04 +0000 (17:54 +0100)]
btrfs-progs: tests: 005-qgroup-show

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: qgroup show: refine error messages
David Sterba [Fri, 20 Jan 2017 16:52:12 +0000 (17:52 +0100)]
btrfs-progs: qgroup show: refine error messages

When qgroup show is called on a filesystem that does not have quotas
enabled, the error message is very unclear:

  ERROR: can't perform the search - No such file or directory
  ERROR: can't list qgroups: No such file or director

Remove the error from low level helper and let the command handler
decide what to print.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: qgroups show: clean up errno passing
David Sterba [Fri, 20 Jan 2017 16:41:55 +0000 (17:41 +0100)]
btrfs-progs: qgroups show: clean up errno passing

We should use return value for errors, not errno.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: docs: fix typo in btrfs-subvolume
Esteve Fernandez [Mon, 9 Jan 2017 16:59:02 +0000 (17:59 +0100)]
btrfs-progs: docs: fix typo in btrfs-subvolume

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: corrupt-block: Include more inode fields
Lakshmipathi.G [Thu, 5 Jan 2017 17:03:58 +0000 (18:03 +0100)]
btrfs-progs: corrupt-block: Include more inode fields

Signed-off-by: Lakshmipathi.G <Lakshmipathi.G@giis.co.in>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: corrupt-block: Include inode nlink field
Lakshmipathi.G [Thu, 5 Jan 2017 10:08:32 +0000 (11:08 +0100)]
btrfs-progs: corrupt-block: Include inode nlink field

Signed-off-by: Lakshmipathi.G <Lakshmipathi.G@giis.co.in>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: defrag: remove v1 ioctl support
David Sterba [Thu, 12 Jan 2017 14:12:09 +0000 (15:12 +0100)]
btrfs-progs: defrag: remove v1 ioctl support

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: defrag: force using v2 defrag ioctl and make default 32M threshold actua...
David Sterba [Thu, 12 Jan 2017 14:00:47 +0000 (15:00 +0100)]
btrfs-progs: defrag: force using v2 defrag ioctl and make default 32M threshold actually work

A user reported on IRC that the new 32M default for target extent size
does not work. This happens because if there are no commandline options,
the v1 ioctl is used that does not do any fine grained defrag. As the v2
ioctl has been introduced 6 years ago (2010, kernel 2.6.33) we won't
keep backward compatibility anymore.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: Fix spelling/typos in user-facing strings
Nicholas D Steeves [Fri, 6 Jan 2017 03:40:47 +0000 (22:40 -0500)]
btrfs-progs: Fix spelling/typos in user-facing strings

Signed-off-by: Nicholas D Steeves <nsteeves@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: mkfs: print device name while trimming
David Sterba [Fri, 6 Jan 2017 12:47:07 +0000 (13:47 +0100)]
btrfs-progs: mkfs: print device name while trimming

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: enable negative numbers for unallocated device space
David Sterba [Tue, 3 Jan 2017 17:26:43 +0000 (18:26 +0100)]
btrfs-progs: enable negative numbers for unallocated device space

The unallocated space is not calculated correctly when a device
deltion/replace is in progress. This appears as huge number, EiB-sized.
It's really a negative number, because we don't have entire information
available and mixing numbers that do and don't take the deleted device
size into account.

We have to query search ioctl to retrieve the missing information, also
this requires root access for the ioctl, so we still need a workaround
for non-root case. Here the negative numbers seem to be a bit more
practical than EiB numbers.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: make negative number pretty printing optional
David Sterba [Tue, 3 Jan 2017 17:03:44 +0000 (18:03 +0100)]
btrfs-progs: make negative number pretty printing optional

Add a unit mode that will interpret the input number as a signed 64bit,
optionally and not by default for all numbers.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: utils: negative numbers are more plausible than sizes over 8 EiB
Zygo Blaxell [Sat, 3 Dec 2016 20:39:54 +0000 (15:39 -0500)]
btrfs-progs: utils: negative numbers are more plausible than sizes over 8 EiB

I got tired of seeing "16.00EiB" whenever btrfs-progs encounters a
negative size value, e.g. during resize:

Unallocated:
   /dev/mapper/datamd18   16.00EiB

This version is much more useful:

Unallocated:
   /dev/mapper/datamd18  -26.29GiB

Signed-off-by: Zygo Blaxell <ce3g8jdj@umail.furryterror.org>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: check: get the highest inode for lost+found
Goldwyn Rodrigues [Tue, 20 Dec 2016 12:08:54 +0000 (06:08 -0600)]
btrfs-progs: check: get the highest inode for lost+found

root->highest_inode is not accurate at the time of creating a lost+found
and it fails because the highest_inode+1 is already present. This could be
because of fixes after highest_inode is set. Instead, search
for the highest inode in the tree and use it for lost+found.

This makes root->highest_inode unnecessary and hence deleted.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agoBtrfs progs v4.9
David Sterba [Fri, 23 Dec 2016 18:50:39 +0000 (19:50 +0100)]
Btrfs progs v4.9

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: update CHANGES for v4.9
David Sterba [Wed, 14 Dec 2016 14:22:52 +0000 (15:22 +0100)]
btrfs-progs: update CHANGES for v4.9

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert-test: trigger chunk allocation after convert
Qu Wenruo [Mon, 19 Dec 2016 06:56:42 +0000 (14:56 +0800)]
btrfs-progs: convert-test: trigger chunk allocation after convert

Populate fs after convert so we can trigger data chunk allocation.
This can expose too restrict old rollback condition

Reported-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: file-item: Fix wrong file extents inserted
Qu Wenruo [Mon, 19 Dec 2016 06:56:37 +0000 (14:56 +0800)]
btrfs-progs: file-item: Fix wrong file extents inserted

If we specify NO_HOLES incompat feature when converting, the result
image still uses hole file extents.
And further more, the hole is incorrect as its disk_num_bytes is not
zero.

The problem is at btrfs_insert_file_extent() which doesn't check if we
are going to insert hole file extent.

Modify it to skip hole file extents to allow it follow restrict NO_HOLES
flag.

And since no_holes flag can be triggered on half-way, so current fsck
won't report such error, as it consider it as old file holes.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: Fix NULL pointer when receive clone operation
Qu Wenruo [Thu, 15 Dec 2016 08:37:02 +0000 (16:37 +0800)]
btrfs-progs: Fix NULL pointer when receive clone operation

Regression introduced by a2f7af94abe4a3491ca1280a2ae1d63edc0d62ab
"btrfs-progs: subvol_uuid_search: return error encoded pointer"
IS_ERR() will only check if it's an error code, won't check if it's
NULL.  And for all the caller the commit modifies, it can return NULL
and makes cause segfault.

Fix it by introducing new IS_ERR_OR_NULL() macro, and for NULL pointer
and needs to return int case, convert NULL pointer to -ENOENT.

Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Tested-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: check: Fix false alert on generation mismatch for tree reloc tree
Qu Wenruo [Mon, 5 Dec 2016 09:07:56 +0000 (17:07 +0800)]
btrfs-progs: check: Fix false alert on generation mismatch for tree reloc tree

For fs with tree reloc tree(under balancing), lowmem mode will report
false alert like:
ERROR: extent[62914560 4096] backref generation mismatch, wanted: <=9,
have: 13

This is because lowmem mode adds a more restrict check, to ensure
generation in fs tree won't be smaller than extent tree.

In fact such assumption is not right for tree reloc tree, so remove such
check.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: check: Fix lowmem false alert on tree reloc tree
Qu Wenruo [Mon, 5 Dec 2016 09:07:55 +0000 (17:07 +0800)]
btrfs-progs: check: Fix lowmem false alert on tree reloc tree

Lowmem mode will report false alert if the fs has tree reloc tree like:
ERROR: shared extent[30011392 4096] lost its parent (parent: 30011392,
level: 1)

The problem is check_shared_block_backref() can't handle tree reloc
tree's self-pointing backref.

And still try to read out the tree block then seeking for the
referencer.

The correct method for it is to check if it's tree reloc root.
In that case, we should check found the ROOT_ITEM of tree reloc tree in
root tree.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: check: Fix lowmem mode stack overflow caused by fsck/023
Qu Wenruo [Mon, 5 Dec 2016 09:07:54 +0000 (17:07 +0800)]
btrfs-progs: check: Fix lowmem mode stack overflow caused by fsck/023

Lowmem mode fsck will overflow its stack since it will do infinite
backref check for tree reloc root.
We should not check backref if it's pointing to itself for tree reloc
root.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: check: Fix assert when using lowmem on fs with tree reloc tree
Qu Wenruo [Mon, 5 Dec 2016 09:07:53 +0000 (17:07 +0800)]
btrfs-progs: check: Fix assert when using lowmem on fs with tree reloc tree

When using lowmem mode, btrfs check will report ASSERT for calling
btrfs_read_fs_root() on tree reloc tree.

Fix it by checking objectid and call btrfs_read_fs_root_no_cache() for
tree reloc tree.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: Fix migrate_super_block() to work with 64k sectorsize
Chandan Rajendra [Thu, 8 Dec 2016 13:56:20 +0000 (19:26 +0530)]
btrfs-progs: convert: Fix migrate_super_block() to work with 64k sectorsize

migrate_super_block() uses sectorsize to refer to the size of the
superblock. Hence on 64k sectorsize filesystems, it ends up computing
checksum beyond the super block length (i.e.
BTRFS_SUPER_INFO_SIZE). This commit fixes the bug by using
BTRFS_SUPER_INFO_SIZE instead of sectorsize of the underlying
filesystem.

Signed-off-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: convert: Prevent accounting blocks beyond end of device
Chandan Rajendra [Thu, 8 Dec 2016 13:56:19 +0000 (19:26 +0530)]
btrfs-progs: convert: Prevent accounting blocks beyond end of device

When looping across data block bitmap, __ext2_add_one_block() may add
blocks which do not exist on the underlying disk. This commit prevents
this from happening by checking the block index against the maximum
block count that was present in the ext4 filesystem instance that is
being converted.

Signed-off-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: defrag: warn when deframgenting directories without -r
David Sterba [Mon, 12 Dec 2016 17:38:47 +0000 (18:38 +0100)]
btrfs-progs: defrag: warn when deframgenting directories without -r

The current implementaion of defrag ioctl on directoreis does not do
what users expect. The -r needs to be specified, but we should also
print a warning to avoid confusion.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: defrag: cleanup temporary errno value
David Sterba [Mon, 12 Dec 2016 17:22:10 +0000 (18:22 +0100)]
btrfs-progs: defrag: cleanup temporary errno value

Rename a single letter variable and move the defintion to the scope of
use.

Signed-off-by: David Sterba <dsterba@suse.com>
7 years agobtrfs-progs: recursive defrag cleanup duplicate code
Anand Jain [Tue, 6 Dec 2016 04:39:38 +0000 (12:39 +0800)]
btrfs-progs: recursive defrag cleanup duplicate code

Signed-off-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>