platform/upstream/btrfs-progs.git
6 years agobtrfs-progs: mkfs: delete un-used parameter fd
Gu Jinxiang [Fri, 18 Aug 2017 08:32:46 +0000 (01:32 -0700)]
btrfs-progs: mkfs: delete un-used parameter fd

Parameter fd is not used in function make_image and traverse_directory
of mkfs.  Delete it.

Signed-off-by: Gu Jinxiang <gujx@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Allow inspect dump-tree to show specified tree block even some tree...
Qu Wenruo [Fri, 18 Aug 2017 09:36:30 +0000 (18:36 +0900)]
btrfs-progs: Allow inspect dump-tree to show specified tree block even some tree roots are corrupted

For btrfs inspect-internal dump-tree, if we use "-b" parameter to show
specified tree block, then we don't really need extra tree roots.

Only chunk root is needed to build up the whole chunk mapping so we can
read tree blocks.

This patch will add __OPEN_CTREE_RETURN_CHUNK_ROOT flag when show
speicifed tree block.
So even root tree is corrupted, we can still use inspect-internal
dump-tree to do some debugging.

Reported-by: Zirconium Hacker <jared.e.vb@gmail.com>
Signed-off-by: Qu Wenruo <quwenruo.btrfs@gmx.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Doc: Fix asciidoc grammar of btrfs-rescue
Qu Wenruo [Mon, 21 Aug 2017 06:57:04 +0000 (15:57 +0900)]
btrfs-progs: Doc: Fix asciidoc grammar of btrfs-rescue

Code block of kernel backtrace lacks leading change line, causing the
following man page result:
------
           One can determine whether zero-log is needed according to the
           kernel backtrace:

           ? replay_one_dir_item+0xb5/0xb5 [btrfs]
           ? walk_log_tree+0x9c/0x19d [btrfs]
           ? btrfs_read_fs_root_no_radix+0x169/0x1a1 [btrfs]
           ? btrfs_recover_log_trees+0x195/0x29c [btrfs]
           ? replay_one_dir_item+0xb5/0xb5 [btrfs]
           ? btree_read_extent_buffer_pages+0x76/0xbc [btrfs]
           ? open_ctree+0xff6/0x132c [btrfs]

       + If the errors are like above, then zero-log should be used to clear
       the log and the filesystem may be mounted normally again. The keywords
------

Not only "+" is rendered as is, but also wrong indent.

Fix it by adding change line before code block.

Signed-off-by: Qu Wenruo <quwenruo.btrfs@gmx.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Specify C standard to gnu90 explicitly
Qu Wenruo [Tue, 15 Aug 2017 00:17:12 +0000 (09:17 +0900)]
btrfs-progs: Specify C standard to gnu90 explicitly

Different C compilers have different default language standard.
This sometimes causes problem on different system.

For distribution like CentOS/RHEL7, its gcc is still 4.8 and will report
error for c90 style declaration, while most developers are using newer
gcc which will just ignore it.
This makes us hard to detect such language standard problem.

This patch will specify standard to gnu90 explicitly to avoid such problem.
Gnu90 is a good mix of c90 while provide a lot of useful gnu extension,
and is supported by all modern gcc and clang.

Reported-by: Marco Lorenzo Crociani <marcoc@prismatelecomtesting.com>
Signed-off-by: Qu Wenruo <quwenruo.btrfs@gmx.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: kernel-lib: add generated tables.c
David Sterba [Mon, 14 Aug 2017 13:14:16 +0000 (15:14 +0200)]
btrfs-progs: kernel-lib: add generated tables.c

The contents of tables.c hasn't changed for more than 15 years, we don't
expect any changes to current contents. New tables might be still added,
in that case the file should be regenerated using the included mktables
tool and updated.

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: build: do not regenerate kernel-lib/tables.c
David Sterba [Mon, 14 Aug 2017 13:09:45 +0000 (15:09 +0200)]
btrfs-progs: build: do not regenerate kernel-lib/tables.c

As reported in
https://www.mail-archive.com/linux-btrfs@vger.kernel.org/msg66765.html
cross-compilation does not work very well with current tables.c. We can
simpy keep the pergenerated file in the repository and update if needed.

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agoBtrfs progs v4.12
David Sterba [Fri, 28 Jul 2017 13:29:46 +0000 (15:29 +0200)]
Btrfs progs v4.12

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: update CHANGES for v4.12
David Sterba [Fri, 31 Mar 2017 12:20:21 +0000 (14:20 +0200)]
btrfs-progs: update CHANGES for v4.12

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: docs: update wording for compression mount options
David Sterba [Mon, 24 Jul 2017 14:45:40 +0000 (16:45 +0200)]
btrfs-progs: docs: update wording for compression mount options

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: tests: Use '-t btrfs' mount option in tests
Adam Buchbinder [Wed, 12 Jul 2017 20:05:22 +0000 (13:05 -0700)]
btrfs-progs: tests: Use '-t btrfs' mount option in tests

Without it, mount (at least from util-linux 2.20.1) tries (and fails) to
mount some filesystems as NTFS.

Signed-off-by: Adam Buchbinder <abuchbinder@google.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: tests: use separate helper for mounting convert filesystems
David Sterba [Thu, 20 Jul 2017 16:44:49 +0000 (18:44 +0200)]
btrfs-progs: tests: use separate helper for mounting convert filesystems

We'll enforce the btrfs type for mount everwhere so we must provide a
way to mount converted filesystems. Add a new helper that will try to
mount the given type.

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: tests: enhance API to request type of the converted filesystem
David Sterba [Thu, 20 Jul 2017 15:53:56 +0000 (17:53 +0200)]
btrfs-progs: tests: enhance API to request type of the converted filesystem

We'll use mount -t $fstype later on, extend the API so we don't have to
parse the type from other parameters.

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: docs: adjust wording for subvol delete
David Sterba [Wed, 19 Jul 2017 19:01:30 +0000 (21:01 +0200)]
btrfs-progs: docs: adjust wording for subvol delete

Slightly update the text about deletion after the discussion on IRC.

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: docs: enhance documentation of 'btrfs device ready'
David Sterba [Mon, 17 Jul 2017 15:58:50 +0000 (17:58 +0200)]
btrfs-progs: docs: enhance documentation of 'btrfs device ready'

There were questions raised about the purpose of the 'dev ready'
command.

Link: https://marc.info/?l=linux-btrfs&m=149944574207325
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: docs: move deprecated mount option to own section
David Sterba [Mon, 17 Jul 2017 14:44:02 +0000 (16:44 +0200)]
btrfs-progs: docs: move deprecated mount option to own section

- alloc_start
- recovery
- subvolrootid

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: subvol show: add support to search subvolume by rootid or uuid
Anand Jain [Wed, 12 Jul 2017 22:47:11 +0000 (06:47 +0800)]
btrfs-progs: subvol show: add support to search subvolume by rootid or uuid

Unless the top level is mounted there is no way to know the
details of all the subvolume.  For example:

mount -o subvol=sv1/newsv1 /dev/sdb /btrfs

btrfs su list /btrfs
ID 257 gen 12 top level 5 path sv1
ID 258 gen 9 top level 257 path sv1/snap
ID 259 gen 11 top level 257 path sv1/newsv1

You can't subvol show for sv1 and sv1/snap as its paths aren't
accessible to the user unless the its top level is mounted.

This patch adds two new options to the existing btrfs subvol show
cli. They are --rootid/-r or --uuid/-u, with this now the user will
be able to look for a subvolume using the rootid OR the uuid.

./btrfs su show -r 257 /btrfs
sv1
Name:  sv1
UUID:  30129358-c69d-3e4a-a662-29509cc69c95
Parent UUID:  -
Received UUID:  -
Creation time:  2017-07-11 20:32:57 +0800
Subvolume ID:  257
Generation:  12
Gen at creation:  7
Parent ID:  5
Top level ID:  5
Flags:  -
Snapshot(s):
sv1/snap

Signed-off-by: Anand Jain <anand.jain@oracle.com>
[ minor adjustments in the help text ]
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: build: Enable ThreadSanitizer, using D=tsan
Adam Buchbinder [Wed, 12 Jul 2017 20:04:49 +0000 (13:04 -0700)]
btrfs-progs: build: Enable ThreadSanitizer, using D=tsan

Tested with clang-3.9. We have to enable PIE,
(https://clang.llvm.org/docs/ThreadSanitizer.html)

Signed-off-by: Adam Buchbinder <abuchbinder@google.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Fix an infinite loop in btrfs_next_bg
Justin Maggard [Fri, 9 Jun 2017 18:09:35 +0000 (11:09 -0700)]
btrfs-progs: Fix an infinite loop in btrfs_next_bg

I've run into a couple filesystems where btrfs-find-root would spin
indefinitely.

If the first cache extent start location is 0, we end up in an infinite
loop in btrfs_next_bg().  Fix it by checking for that situation, and
jumping to the next bg if necessary.

Fixes: e2e0dae9 (btrfs-progs: volume: Fix a bug causing btrfs-find-root to skip first chunk)
Signed-off-by: Justin Maggard <jmaggard@netgear.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agoBtrfs-progs: fix infinite loop in find_free_extent
Liu Bo [Mon, 26 Jun 2017 17:34:41 +0000 (11:34 -0600)]
Btrfs-progs: fix infinite loop in find_free_extent

If the found %ins is crossing a stripe len, ie. BTRFS_STRIPE_LEN, we'd
search again with a stripe-aligned %search_start.  The current code
calculates %search_start by adding a wrong offset, in order to fix it, the
start position of the block group should be taken, otherwise, it'll end up
with looking at the same block group forever.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Tighten integer types in print-tree
Adam Buchbinder [Mon, 10 Jul 2017 21:29:09 +0000 (14:29 -0700)]
btrfs-progs: Tighten integer types in print-tree

There are likely more places where the wrong size types are used, but
these tripped Clang's warnings because they eventually get passed to
printf.

Signed-off-by: Adam Buchbinder <abuchbinder@google.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: tests: Fix missing internal deps in check and misc tests
Adam Buchbinder [Mon, 10 Jul 2017 21:29:08 +0000 (14:29 -0700)]
btrfs-progs: tests: Fix missing internal deps in check and misc tests

Doing a straight 'make test' would fail because some misc and fsck
tests require particular tools to already be built. Add dependencies
at the Makefile and shell-script level.

Signed-off-by: Adam Buchbinder <abuchbinder@google.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: subvol show: fix the path use full_path as provided by the root info
Anand Jain [Wed, 12 Jul 2017 07:20:27 +0000 (15:20 +0800)]
btrfs-progs: subvol show: fix the path use full_path as provided by the root info

This is a kind of preparatory patch for the patch which will add
--rootid and --uuid options for the btrfs subvol show command.

As of now btrfs subvol show is using the external user provided subvol
path to show in the output. Which is kind of confusing.

btrfs su show /btrfs
/btrfs <--
Name:  <FS_TREE>

It will be even more confusing when proposed --uuid or --rootid
options are used.

btrfs su show --rootid 258 /btrfs
/btrfs <--
Name:  snap <--
UUID:  9630a45f-e647-4242-bd19-97590b4e20b2
Parent UUID:  30129358-c69d-3e4a-a662-29509cc69c95
Received UUID:  -
Creation time:  2017-07-12 12:43:28 +0800
Subvolume ID:  258
Generation:  9
Gen at creation:  9
Parent ID:  257
Top level ID:  257
Flags:  -
Snapshot(s):

Now with this patch, it will only show what is provided by the root_info.

btrfs su show --rootid 258 /btrfs
sv1/snap <--
Name:  snap
UUID:  9630a45f-e647-4242-bd19-97590b4e20b2
Parent UUID:  30129358-c69d-3e4a-a662-29509cc69c95
Received UUID:  -
Creation time:  2017-07-12 12:43:28 +0800
Subvolume ID:  258
Generation:  9
Gen at creation:  9
Parent ID:  257
Top level ID:  257
Flags:  -
Snapshot(s):

Signed-off-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: image: fix typos in messages
Adam Buchbinder [Wed, 12 Jul 2017 22:10:13 +0000 (00:10 +0200)]
btrfs-progs: image: fix typos in messages

Signed-off-by: Adam Buchbinder <abuchbinder@google.com>
[ separated from the previous patch ]
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: image: Fix data races when reporting progress
Adam Buchbinder [Wed, 12 Jul 2017 20:05:10 +0000 (13:05 -0700)]
btrfs-progs: image: Fix data races when reporting progress

Making the code data-race safe requires that reads *and* writes
happen under a mutex lock, if any of the access are writes. See
Dmitri Vyukov, "Benign data races: what could possibly go wrong?"
for more details.

The fix here was to put most of the main loop of restore_worker
under a mutex lock.

This race was detected using fsck-tests/012-leaf-corruption.

==================
WARNING: ThreadSanitizer: data race
  Write of size 4 by main thread:
    #0 add_cluster btrfs-progs/image/main.c:1931
    #1 restore_metadump btrfs-progs/image/main.c:2566
    #2 main btrfs-progs/image/main.c:2859

  Previous read of size 4 by thread T6:
    #0 restore_worker btrfs-progs/image/main.c:1720

  Location is stack of main thread.

  Thread T6 (running) created by main thread at:
    #0 pthread_create <null>
    #1 mdrestore_init btrfs-progs/image/main.c:1868
    #2 restore_metadump btrfs-progs/image/main.c:2534
    #3 main btrfs-progs/image/main.c:2859

SUMMARY: ThreadSanitizer: data race btrfs-progs/image/main.c:1931 in
add_cluster

Signed-off-by: Adam Buchbinder <abuchbinder@google.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: convert: Fix data race when reporting progress
Adam Buchbinder [Wed, 12 Jul 2017 20:05:38 +0000 (13:05 -0700)]
btrfs-progs: convert: Fix data race when reporting progress

The status display was reading the state while the task was updating
it. Use a mutex to prevent the race.

This race was detected using ThreadSanitizer and
misc-tests/005-convert-progress-thread-crash.

==================
WARNING: ThreadSanitizer: data race
  Write of size 8 by main thread:
    #0 ext2_copy_inodes btrfs-progs/convert/source-ext2.c:853
    #1 copy_inodes btrfs-progs/convert/main.c:145
    #2 do_convert btrfs-progs/convert/main.c:1297
    #3 main btrfs-progs/convert/main.c:1924

  Previous read of size 8 by thread T1:
    #0 print_copied_inodes btrfs-progs/convert/main.c:124

  Location is stack of main thread.

  Thread T1 (running) created by main thread at:
    #0 pthread_create <null>
    #1 task_start btrfs-progs/task-utils.c:50
    #2 do_convert btrfs-progs/convert/main.c:1295
    #3 main btrfs-progs/convert/main.c:1924

SUMMARY: ThreadSanitizer: data race
btrfs-progs/convert/source-ext2.c:853 in ext2_copy_inodes

Signed-off-by: Adam Buchbinder <abuchbinder@google.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor chunk creation functions to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:35 +0000 (17:19 +0800)]
btrfs-progs: Refactor chunk creation functions to use btrfs_fs_info

4 functions are involved in this refactor: btrfs_make_block_group()
btrfs_make_block_groups(), btrfs_alloc_chunk, btrfs_alloc_data_chunk().

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Remove btrfs_read_super_device as there is no implementation
Qu Wenruo [Tue, 13 Jun 2017 09:19:34 +0000 (17:19 +0800)]
btrfs-progs: Remove btrfs_read_super_device as there is no implementation

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_add_device() to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:33 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_add_device() to use btrfs_fs_info

BTW, there is a duplicated definition of btrfs_add_device() in
volumes.h, also remove it.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_chunk_readonly to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:32 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_chunk_readonly to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_add_system_chunk to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:31 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_add_system_chunk to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_read_sys_array/chunk_tree to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:30 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_read_sys_array/chunk_tree to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_find_device to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:29 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_find_device to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_check_chunk_valid to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:28 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_check_chunk_valid to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor write_all_supers and its callers to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:27 +0000 (17:19 +0800)]
btrfs-progs: Refactor write_all_supers and its callers to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_readahead_tree_block to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:26 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_readahead_tree_block to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_find_tree_block to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:25 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_find_tree_block to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor read_extent_data to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:24 +0000 (17:19 +0800)]
btrfs-progs: Refactor read_extent_data to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: btrfstune: Refactor change_devices_uuid to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:23 +0000 (17:19 +0800)]
btrfs-progs: btrfstune: Refactor change_devices_uuid to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor write_tree_block to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:22 +0000 (17:19 +0800)]
btrfs-progs: Refactor write_tree_block to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor write_and_map_eb to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:21 +0000 (17:19 +0800)]
btrfs-progs: Refactor write_and_map_eb to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor csum_tree_block to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:20 +0000 (17:19 +0800)]
btrfs-progs: Refactor csum_tree_block to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_next_bg and its callers to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:19 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_next_bg and its callers to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_num_copies to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:18 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_num_copies to use btrfs_fs_info

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor btrfs_map_block and its variants to use btrfs_fs_info
Qu Wenruo [Tue, 13 Jun 2017 09:19:17 +0000 (17:19 +0800)]
btrfs-progs: Refactor btrfs_map_block and its variants to use btrfs_fs_info

Just to keep the 1st paramter the same as kernel.

We can also save a few lines since the parameter is shorter now.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: lowmem check: Fix false alert on missing chunk or dev extent
Qu Wenruo [Wed, 31 May 2017 05:56:09 +0000 (13:56 +0800)]
btrfs-progs: lowmem check: Fix false alert on missing chunk or dev extent

When checking chunk or dev extent, lowmem mode uses chunk length as dev
extent length, and if they mismatch, report missing chunk or dev extent
like:
------
ERROR: chunk[256 4324327424) stripe 0 did not find the related dev extent
ERROR: chunk[256 4324327424) stripe 1 did not find the related dev extent
ERROR: chunk[256 4324327424) stripe 2 did not find the related dev extent
------

However, only for Single/DUP/RAID1 profiles chunk length is the same as
dev extent length.
For other profiles, this will cause tons of false alert.

Fix it by using correct stripe length when checking chunk and dev extent
items.

This fixes the mkfs test failure when using lowmem mode check.

Reported-by: Kai Krakow <hurikhan77@gmail.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Introduce function to get correct stripe length
Qu Wenruo [Wed, 31 May 2017 05:56:08 +0000 (13:56 +0800)]
btrfs-progs: Introduce function to get correct stripe length

Introduce a new function, btrfs_get_chunk_stripe_len() to get correct
stripe length.
This is very handy for lowmem mode, which checks the mapping between
device extent and chunk item.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: check: Reuse btrfs_check_chunk_valid in lowmem mode
Qu Wenruo [Wed, 31 May 2017 05:56:07 +0000 (13:56 +0800)]
btrfs-progs: check: Reuse btrfs_check_chunk_valid in lowmem mode

Before this patch, btrfs check lowmem mode manually checks found chunk
item, even we already have the generic chunk validation checker,
btrfs_check_chunk_valid().

This patch will use btrfs_check_chunk_valid() to replace open-coded
chunk validation checker in check_chunk_item().

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Enhance chunk item validation check
Qu Wenruo [Wed, 31 May 2017 05:56:06 +0000 (13:56 +0800)]
btrfs-progs: Enhance chunk item validation check

btrfs_check_chunk_valid() doesn't check if
1) chunk flag has conflicting flags
   For example chunk type DATA|METADATA|RAID1|RAID10 is completely
   invalid, while current check_chunk_valid() can't detect it.
2) num_stripes is invalid for RAID10
   Num_stripes 5 is not valid for RAID10.

This patch will enhance btrfs_check_chunk_valid() to handle above cases.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Introduce wrapper to recover raid56 data
Qu Wenruo [Thu, 25 May 2017 06:21:50 +0000 (14:21 +0800)]
btrfs-progs: Introduce wrapper to recover raid56 data

Introduce a wrapper to recover raid56 data.

The logical is the same with kernel one, but with different interfaces,
since kernel ones cares the performance while in btrfs we don't care
that much.

And the interface is more caller friendly inside btrfs-progs.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: raid56: Allow raid6 to recover data and P
Qu Wenruo [Thu, 25 May 2017 06:21:49 +0000 (14:21 +0800)]
btrfs-progs: raid56: Allow raid6 to recover data and P

Copied from kernel lib/raid6/recov.c.

Minor modifications includes:
- Rename from raid6_datap_recov_intx() to raid5_recov_datap()
- Rename parameter from faila to dest1

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: raid56: Allow raid6 to recover 2 data stripes
Qu Wenruo [Thu, 25 May 2017 06:21:48 +0000 (14:21 +0800)]
btrfs-progs: raid56: Allow raid6 to recover 2 data stripes

Copied from kernel lib/raid6/recov.c raid6_2data_recov_intx1() function.
With the following modification:
- Rename to raid6_recov_data2() for shorter name
- s/kfree/free/g modification

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: raid56: Introduce tables for RAID6 recovery
Qu Wenruo [Thu, 25 May 2017 06:21:47 +0000 (14:21 +0800)]
btrfs-progs: raid56: Introduce tables for RAID6 recovery

Use kernel RAID6 galois tables for later RAID6 recovery.

Galois tables file, kernel-lib/tables.c is generated by user space
program, mktable.

Galois field tables declaration, in kernel-lib/raid56.h, is completely
copied from kernel.

The mktables.c is copied from kernel with minor header/macro
modification, to ensure the generated tables.c works well in
btrfs-progs.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: raid56: Introduce raid56 header for later recovery usage
Qu Wenruo [Thu, 25 May 2017 06:21:46 +0000 (14:21 +0800)]
btrfs-progs: raid56: Introduce raid56 header for later recovery usage

Introduce a new header, kernel-lib/raid56.h, for later raid56 works.

It contains 2 functions, from original btrfs-progs code:
void raid6_gen_syndrome(int disks, size_t bytes, void **ptrs);
int raid5_gen_result(int nr_devs, size_t stripe_len, int dest, void **data);

Will be expanded later and some part of it(RAID6 recover part) may keep
sync with kernel later.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ unify gpl header, rename header macro ]
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Fix slot >= nritems
Philipp Hahn [Mon, 15 May 2017 17:00:23 +0000 (19:00 +0200)]
btrfs-progs: Fix slot >= nritems

Running "btrfsck --repair /dev/sdd2" crashed as it can happen in
(corrupted) file systems, that slot > nritems:
> (gdb) bt full
> #0  0x00007ffff7020e71 in __memmove_sse2_unaligned_erms () from /lib/x86_64-linux-gnu/libc.so.6
> #1  0x0000000000438764 in btrfs_del_ptr (trans=<optimized out>, root=0x6e4fe0, path=0x1d17880, level=0, slot=7)
>     at ctree.c:2611
>         parent = 0xcd96980
>         nritems = <optimized out>
>         __func__ = "btrfs_del_ptr"
> #2  0x0000000000421b15 in repair_btree (corrupt_blocks=<optimized out>, root=<optimized out>) at cmds-check.c:3539
>         key = {objectid = 77990592512, type = 168 '\250', offset = 16384}
>         trans = 0x8f48c0
>         path = 0x1d17880
>         level = 0
> #3  check_fs_root (wc=<optimized out>, root_cache=<optimized out>, root=<optimized out>) at cmds-check.c:3703
>         corrupt = 0x1d17880
>         corrupt_blocks = {root = {rb_node = 0x6e80c60}}
>         path = {nodes = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, slots = {0, 0, 0, 0, 0, 0, 0, 0}, locks = {0, 0,
>             0, 0, 0, 0, 0, 0}, reada = 0, lowest_level = 0, search_for_split = 0, skip_check_block = 0}
>         nrefs = {bytenr = {271663104271646720560021504, 0, 0, 0, 0, 0}, refs = {1, 1, 1, 0, 0, 0, 0, 0}}
>         wret = 215575372
>         root_node = {cache = {rb_node = {__rb_parent_color = 0, rb_right = 0x0, rb_left = 0x0}, objectid = 0,
>             start = 0, size = 0}, root_cache = {root = {rb_node = 0x0}}, inode_cache = {root = {
>               rb_node = 0x781c80}}, current = 0x819530, refs = 0}
>         status = 215575372
>         rec = 0x1
> #4  check_fs_roots (root_cache=0xcd96b6d, root=<optimized out>) at cmds-check.c:3809
>         path = {nodes = {0x6eed90, 0x6a2f40, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, slots = {18, 2, 0, 0, 0, 0, 0, 0},
>           locks = {0, 0, 0, 0, 0, 0, 0, 0}, reada = 0, lowest_level = 0, search_for_split = 0,
>           skip_check_block = 0}
>         key = {objectid = 323, type = 132 '\204', offset = 18446744073709551615}
>         wc = {shared = {root = {rb_node = 0x0}}, nodes = {0x0, 0x0, 0x7fffffffe428, 0x0, 0x0, 0x0, 0x0, 0x0},
>           active_node = 2, root_level = 2}
>         leaf = 0x6e4fe0
>         tmp_root = 0x6e4fe0
> #5  0x00000000004287c3 in cmd_check (argc=215575372, argv=0x1d17880) at cmds-check.c:11521
>         root_cache = {root = {rb_node = 0x98c2940}}
>         info = 0x6927b0
>         bytenr = 6891440
>         tree_root_bytenr = 0
>         uuidbuf = "f65ff1a1-76ef-456e-beb5-c6c3841e7534"
>         num = 215575372
>         readonly = 218080104
>         qgroups_repaired = 0
> #6  0x000000000040a41f in main (argc=3, argv=0x7fffffffebe8) at btrfs.c:243
>         cmd = 0x689868
>         bname = <optimized out>
>         ret = <optimized out>

in that case the count of remaining items (nritems - slot - 1) gets
negative. That is then casted to (unsigned long len), which leads to the
observed crash.

Change the tests before the move to handle only the non-corrupted case,
were slow < nritems.

This does not fix the corruption, but allows btrfsck to finish without
crashing.

Signed-off-by: Philipp Hahn <hahn@univention.de>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Refactor read_node_slot function to get rid of btrfs_root
Qu Wenruo [Thu, 18 May 2017 03:11:55 +0000 (11:11 +0800)]
btrfs-progs: Refactor read_node_slot function to get rid of btrfs_root
parameter

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor read_tree_block to get rid of btrfs_root
Qu Wenruo [Thu, 18 May 2017 02:51:08 +0000 (10:51 +0800)]
btrfs-progs: Refactor read_tree_block to get rid of btrfs_root

The only reasom read_tree_block() needs a btrfs_root parameter is to get
its node/sector size.

And long ago, I have already introduced a compactible interface,
read_tree_block_fs_info() to pass btrfs_fs_info instead of btrfs_root.

Since we have cleaned up all root->sector/node/stripesize users, we
should be OK to refactor read_tree_block() function.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor btrfs_root paramters in btrfs-corrupt-block.c
Qu Wenruo [Thu, 18 May 2017 02:36:19 +0000 (10:36 +0800)]
btrfs-progs: Refactor btrfs_root paramters in btrfs-corrupt-block.c

Refactor the following functions and its callers to get rid of
incorrectly passed btrfs_root parameters:

1) corrupt_keys()
2) corrupt_metadata_block()

The only reason passing btrfs_root for them is to get block sizes.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Remove block size members in btrfs_root
Qu Wenruo [Thu, 18 May 2017 02:22:22 +0000 (10:22 +0800)]
btrfs-progs: Remove block size members in btrfs_root

Finally, we can get rid of per tree block size members now.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in cmds-inspect-dump-tree.c
Qu Wenruo [Thu, 18 May 2017 02:21:34 +0000 (10:21 +0800)]
btrfs-progs: Refactor sectorsize users in cmds-inspect-dump-tree.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in convert/source-ext2.c
Qu Wenruo [Thu, 18 May 2017 02:14:19 +0000 (10:14 +0800)]
btrfs-progs: Refactor sectorsize users in convert/source-ext2.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in convert/main.c
Qu Wenruo [Thu, 18 May 2017 02:13:27 +0000 (10:13 +0800)]
btrfs-progs: Refactor sectorsize users in convert/main.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize in convert/source-fs.c
Qu Wenruo [Thu, 18 May 2017 02:11:46 +0000 (10:11 +0800)]
btrfs-progs: Refactor sectorsize in convert/source-fs.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in free-space-tree.c
Qu Wenruo [Thu, 18 May 2017 02:09:53 +0000 (10:09 +0800)]
btrfs-progs: Refactor sectorsize users in free-space-tree.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in volumes.c
Qu Wenruo [Thu, 18 May 2017 02:09:10 +0000 (10:09 +0800)]
btrfs-progs: Refactor sectorsize users in volumes.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in file.c
Qu Wenruo [Thu, 18 May 2017 02:06:40 +0000 (10:06 +0800)]
btrfs-progs: Refactor sectorsize users in file.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in free-space-cache.c
Qu Wenruo [Thu, 18 May 2017 02:05:22 +0000 (10:05 +0800)]
btrfs-progs: Refactor sectorsize users in free-space-cache.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsizes users in file-item.c
Qu Wenruo [Thu, 18 May 2017 02:02:24 +0000 (10:02 +0800)]
btrfs-progs: Refactor sectorsizes users in file-item.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor sectorsize users in mkfs/main.c
Qu Wenruo [Thu, 18 May 2017 02:00:01 +0000 (10:00 +0800)]
btrfs-progs: Refactor sectorsize users in mkfs/main.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor nodesize users in cmds-inspect-tree-stats.c
Qu Wenruo [Thu, 18 May 2017 01:54:49 +0000 (09:54 +0800)]
btrfs-progs: Refactor nodesize users in cmds-inspect-tree-stats.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor nodesize users in qgroup-verify.c
Qu Wenruo [Thu, 18 May 2017 01:51:55 +0000 (09:51 +0800)]
btrfs-progs: Refactor nodesize users in qgroup-verify.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor nodesize user in print-tree.c
Qu Wenruo [Thu, 18 May 2017 01:50:51 +0000 (09:50 +0800)]
btrfs-progs: Refactor nodesize user in print-tree.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in extent-tree.c
Qu Wenruo [Thu, 18 May 2017 01:48:52 +0000 (09:48 +0800)]
btrfs-progs: Refactor block sizes users in extent-tree.c

And in the refactoring, also refactors a open-coded round_up().

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor nodesize users in utils.c
Qu Wenruo [Thu, 18 May 2017 01:40:25 +0000 (09:40 +0800)]
btrfs-progs: Refactor nodesize users in utils.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor nodesize user in btrfstune.c
Qu Wenruo [Thu, 18 May 2017 01:39:03 +0000 (09:39 +0800)]
btrfs-progs: Refactor nodesize user in btrfstune.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in cmds-check.c
Qu Wenruo [Thu, 18 May 2017 01:38:03 +0000 (09:38 +0800)]
btrfs-progs: Refactor block sizes users in cmds-check.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor nodesize users in image/main.c
Qu Wenruo [Thu, 18 May 2017 01:29:54 +0000 (09:29 +0800)]
btrfs-progs: Refactor nodesize users in image/main.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor nodesize user in extent_io.c
Qu Wenruo [Thu, 18 May 2017 01:27:32 +0000 (09:27 +0800)]
btrfs-progs: Refactor nodesize user in extent_io.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in cmds-restore.c
Qu Wenruo [Thu, 18 May 2017 01:26:48 +0000 (09:26 +0800)]
btrfs-progs: Refactor block sizes users in cmds-restore.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in backref.c
Qu Wenruo [Thu, 18 May 2017 01:24:55 +0000 (09:24 +0800)]
btrfs-progs: Refactor block sizes users in backref.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in chunk-recover.c
Qu Wenruo [Thu, 18 May 2017 01:23:59 +0000 (09:23 +0800)]
btrfs-progs: Refactor block sizes users in chunk-recover.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in btrfs-map-logical.c
Qu Wenruo [Thu, 18 May 2017 01:21:05 +0000 (09:21 +0800)]
btrfs-progs: Refactor block sizes users in btrfs-map-logical.c

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in ctree.c and ctree.h
Qu Wenruo [Thu, 18 May 2017 01:17:41 +0000 (09:17 +0800)]
btrfs-progs: Refactor block sizes users in ctree.c and ctree.h

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in btrfs-corrupt-block.c
Qu Wenruo [Thu, 18 May 2017 01:11:20 +0000 (09:11 +0800)]
btrfs-progs: Refactor block sizes users in btrfs-corrupt-block.c

Convert root->sectorsize/nodesize users in btrfs-corrupt-block.

This provides the basis to further refactor incorrect btrfs_root
parameter to btrfs_fs_info parameter.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Refactor block sizes users in disk-io.c
Qu Wenruo [Wed, 17 May 2017 09:17:56 +0000 (17:17 +0800)]
btrfs-progs: Refactor block sizes users in disk-io.c

Since we have cached block sizes in fs_info, there is no need to specify
these sizes in btrfs_setup_root() function.

And refactor all root->sector/node/stripesize users in disk-io.c.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Introduce sectorsize nodesize and stripesize members for
Qu Wenruo [Wed, 17 May 2017 08:57:35 +0000 (16:57 +0800)]
btrfs-progs: Introduce sectorsize nodesize and stripesize members for
btrfs_fs_info

Just like what we do in kernel, since we will not support different
leaf/node/stripe size per tree, there is no need to store these block
sizes in btrfs_root.

This patch will introduce these block size members into btrfs_fs_info
structure, allowing us to convert such usage in later patches.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Remove deprecated leafsize usage
Qu Wenruo [Wed, 17 May 2017 08:42:50 +0000 (16:42 +0800)]
btrfs-progs: Remove deprecated leafsize usage

Leafsize is deprecated for a long time, and kernel has already updated
ctree.h to rename sb->leafsize to sb->__unused_leafsize.

This patch will remove normal users of leafsize:
1) Remove leafsize member from btrfs_root structure
   Now only root->nodesize and root->sectorisze.
   No longer root->leafsize.

2) Remove @leafsize parameter from btrfs_setup_root() function
   Since no root->leafsize, no need for @leafsize parameter.

The remaining user of leafsize will be:
1) btrfs inspect-internal dump-super
   Reformat the "leafsize" output to "leafsize (deprecated)" and
   use le32_to_cpu() to do the cast manually.

2) mkfs
   We still need to set sb->__unused_leafsize to nodesize.
   Do the manual cast too.

3) convert
   Same as mkfs, these two superblock setup should be merged later

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
6 years agobtrfs-progs: Cleanup open-coded btrfs_chunk_item_size
Qu Wenruo [Mon, 15 May 2017 08:27:38 +0000 (16:27 +0800)]
btrfs-progs: Cleanup open-coded btrfs_chunk_item_size

In btrfs_check_chunk_valid() we calculate chunk item using open code,
use an existing helper btrfs_chunk_item_size() instead.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: docs: document conventions
David Sterba [Thu, 18 May 2017 16:49:00 +0000 (18:49 +0200)]
btrfs-progs: docs: document conventions

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agoBtrfs progs v4.11.1
David Sterba [Fri, 30 Jun 2017 16:34:24 +0000 (18:34 +0200)]
Btrfs progs v4.11.1

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: update CHANGES for v4.11.1
David Sterba [Fri, 31 Mar 2017 12:20:21 +0000 (14:20 +0200)]
btrfs-progs: update CHANGES for v4.11.1

Signed-off-by: David Sterba <dsterba@suse.com>
6 years agoBtrfs-progs: convert: do not clear header rev
Liu Bo [Mon, 26 Jun 2017 23:55:02 +0000 (17:55 -0600)]
Btrfs-progs: convert: do not clear header rev

So btrfs_set_header_flags() vs btrfs_set_header_flag, the difference is
sort of similar to "=" vs "|=", when creating and initialising a new
extent buffer, convert uses the former one which clears header_rev by
accident.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agoBtrfs-progs: convert: do not clear nodatasum flag in inode item
Liu Bo [Wed, 28 Jun 2017 21:29:49 +0000 (15:29 -0600)]
Btrfs-progs: convert: do not clear nodatasum flag in inode item

With the current btrfs-convert, if we convert a ext4 without data checksum,
it'd not set nodatasum flag in inode item, nor create csum item, reading
file ends up with checksum errors.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agoBtrfs-progs: man: nodesize must be power of 2 now
Liu Bo [Tue, 27 Jun 2017 21:25:45 +0000 (15:25 -0600)]
Btrfs-progs: man: nodesize must be power of 2 now

This updates mkfs.btrfs's man page with the new limitation that nodesize must
be a power of 2 as well.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: lowmem check: Fix false alert about file extent interrupt
Lu Fengqi [Thu, 22 Jun 2017 08:12:56 +0000 (16:12 +0800)]
btrfs-progs: lowmem check: Fix false alert about file extent interrupt

As Qu mentioned in this thread
(https://www.spinics.net/lists/linux-btrfs/msg64469.html), compression
can cause regular extent to co-exist with an inlined extent. This
coexistence makes things confusing. Since it is currently allowed and
can appear in a filesystem, fix btrfsck to prevent a bunch of error
reports to appear that will make user feel uneasy.

When checking a file extent, record the extent_end of the regular extent
to check if there is a gap between the regular extents. Normally there
is only one inlined extent, so the extent_end of inlined extent is
useless. However, if a regular extent can co-exist with an inlined
extent, the extent_end of the inlined extent also needs to be recorded.

Reported-by: Marc MERLIN <marc@merlins.org>
Signed-off-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: tests: Add test case to check file hole extents with NO_HOLES flag
Qu Wenruo [Fri, 16 Jun 2017 01:43:23 +0000 (09:43 +0800)]
btrfs-progs: tests: Add test case to check file hole extents with NO_HOLES flag

Add test case which we have NO_HOLES incompat flag while still have
hole file extent.

This can be created by enabling NO_HOLES feature on an existing
filesystem, which lowmem mode would cause false alert for it.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ minor adjustments ]
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: check: Fix false alert about EXTENT_DATA that shouldn't be a hole
Qu Wenruo [Fri, 16 Jun 2017 01:43:22 +0000 (09:43 +0800)]
btrfs-progs: check: Fix false alert about EXTENT_DATA that shouldn't be a hole

Since the incompat feature NO_HOLES still allows us to have an explicit
hole file extent, current check is too strict and will cause false
alerts like:

  root 5 EXTENT_DATA[257, 0] shouldn't be hole

Fix it by removing the strict file hole extent check.

Link: https://www.spinics.net/lists/linux-btrfs/msg66374.html
Reported-by: Henk Slager <eye1tm@gmail.com>
Tested-by: Henk Slager <eye1tm@gmail.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: convert: Add larger device support
Lakshmipathi.G [Sat, 3 Jun 2017 09:57:45 +0000 (15:27 +0530)]
btrfs-progs: convert: Add larger device support

With larger file system (in this case its 22TB), ext2fs_open() returns
EXT2_ET_CANT_USE_LEGACY_BITMAPS error message with
ext2fs_read_block_bitmap().

To overcome this issue,
(a) we need pass EXT2_FLAG_64BITS flag with ext2fs_open.
(b) use 64-bit functions like ext2fs_get_block_bitmap_range2,
    ext2fs_inode_data_blocks2,ext2fs_read_ext_attr2
(c) use 64bit types with btrfs_convert_context fields

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=194795
Signed-off-by: Lakshmipathi.G <lakshmipathi.g@giis.co.in>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: convert: widen int types in convert context
Lakshmipathi.G [Mon, 26 Jun 2017 12:43:31 +0000 (14:43 +0200)]
btrfs-progs: convert: widen int types in convert context

The u32 types in the convert context might not be enough for some very
large filesytems (20TB). Use 64bit types to be safe.

Signed-off-by: Lakshmipathi.G <lakshmipathi.g@giis.co.in>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: Fix 'btrfs device stats --check' cli option
Lakshmipathi.G [Thu, 22 Jun 2017 07:57:53 +0000 (13:27 +0530)]
btrfs-progs: Fix 'btrfs device stats --check' cli option

"Bug 194961 - btrfs device stats --check <folder> does not work"

The long option --check is not recognized as it's missing from the
option table.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=194961
Reported-by: Tomas Thiemel<thiemel@centrum.cz>
Signed-off-by: Lakshmipathi.G <Lakshmipathi.G@giis.co.in>
Signed-off-by: David Sterba <dsterba@suse.com>
6 years agobtrfs-progs: docs: replace <dev> with <device>
Kasijjuf [Thu, 22 Jun 2017 17:49:55 +0000 (12:49 -0500)]
btrfs-progs: docs: replace <dev> with <device>

As discussed in #50, use 'device' everywhere a path do a device is
expected.

Pull request: #52
Signed-off-by: David Sterba <dsterba@suse.com>