We have this check in the kernel but not in userspace, which makes fsck
fail when we wouldn't have a problem in the kernel. This was meant to
catch this case because it really isn't good, unfortunately it will
require a design change to fix in the kernel so in the meantime add this
check so we can be sure our tests only catch real problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
PPC64 arch use such following IOC values "
\#define _IOC_NONE 1U
\#define _IOC_READ 2U
\#define _IOC_WRITE 4U
" comparing to the default IOC values "
\#define _IOC_NONE 0U
\#define _IOC_READ 2U
\#define _IOC_WRITE 1U"
This means the value "_IOW*" will be negative when we store it in the int
variables. Such as the "BTRFS_IOC_QGROUP_CREATE", it will be "0x4010942e" on
X86_64, but "0x8010942e" on PPC64.
Notice that the IOC values are the "unsigned long" type, so we use the
"unsigned long" to store it, and this can insure the comparison between the
variable and BTRFS_IOC_* valid.
Signed-off-by: George Wang <xuw2015@gmail.com>
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
[manpage states that it's int, kernel uses unsigned int, glibc wants
unsigned long int]
Tested-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Copy the functionality of standalone btrfs-zero-log to the main tool.
Delete man page for btrfs-zero-log and copy the relevant parts into
btrfs-rescue(8). The standalone utility will be removed later.
Signed-off-by: David Sterba <dsterba@suse.cz>
Resize of a filesystem image does not work as expected. This has been
confusing and can have bad consequences as people have reported,
resizing the wrong filesystem.
Signed-off-by: David Sterba <dsterba@suse.cz>
A few minor benefits:
* editors set highliting according to the extensions
* web access to the git repository (github) renders the .asciidoc
files:
* we can link to them from the wiki
* the files are editable via browser and such editations can be
submitted for merge easily
Signed-off-by: David Sterba <dsterba@suse.cz>
my troubleshooting experience says have unique error string per module.
In the below eg, its one additional step to know error line,
cat -n cmds-device.c | egrep "error removing the device"
185 "ERROR: error removing the device '%s' - %s\n",
190 "ERROR: error removing the device '%s' - %s\n",
which is completely avoidable.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
[merged the two messages into one]
Signed-off-by: David Sterba <dsterba@suse.cz>
simple compile time warning fixes.
btrfs-calc-size.c: In function ‘print_seek_histogram’:
btrfs-calc-size.c:221: warning: ‘group_start’ may be used uninitialized in this function
btrfs-calc-size.c:223: warning: ‘group_end’ may be used uninitialized in this function
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Currently this test uses the system btrfs-image. If there isn't a
btrfs-image on $PATH, the test fails. The test should be using the
locally compiled btrfs-image, not the system one.
Signed-off-by: WorMzy Tykashi <wormzy.tykashi@gmail.com>
Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
We're not using it anywhere. The best practice is to add enums with
values > 255 for the long options, option index counting is error prone.
Signed-off-by: David Sterba <dsterba@suse.cz>
mkfs and convert will not support the same features, -O will print only
the list according to the given mask.
Signed-off-by: David Sterba <dsterba@suse.cz>
We've passed blocksize to prepare_system_chunk and used it to read and
write superblock. While this does not cause a bug (SUPER_INFO is
blocksize ie. page size on most arches), we should really use the
correct size.
Signed-off-by: David Sterba <dsterba@suse.cz>
We'd like to run each class of tests separately. There were some missing
prerequisities that should be/are verified by the tests, makefile rules
have been synced.
Signed-off-by: David Sterba <dsterba@suse.cz>
If the test image is eg. on NFS it's not writable for root, so chmod
0777 fixes that but then we must not delete the file.
Signed-off-by: David Sterba <dsterba@suse.cz>
Although btrfs-convert can rollback converted btrfs, it still has some
limitation to ensure rollback.
Add a warning on the limitations.
Also add a note for users who decides to go on with btrfs and don't need
the rollback ability.
Reported-by: Vytautas D <vytdau@gmail.com>
Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Also change the test() to convert_test(), to avoid conflict name with
bash test function.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Add fallback TEST_DEV for test case 013.
Fallback to $TOP/tests/test.img.
Now all test cases of btrfs-progs need no extra setting except sudo.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Update 013-extent-tree-rebuild to use more framework infrastructure,
including:
1) Use run_check other than open-coded redirect
2) Add root privillege
3) Add dependency on 'btrfs-debug-tree' command
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Since we have already had TEST_MNT fallback setup to $TOP/tests/mnt,
just remove duplicated setting in 012-leaf-corruption/test.sh
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Use upper case variant name for the following variants:
1) top -> TOP
2) script_dir -> SCRIPT_DIR
And change the following variant name:
1) RESULT -> RESULTS
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Before this patch, ext*_image is always set NODATACSUM inode flag.
However btrfs-convert will set normal file with DATACUSM flag by
default, and generate checksum for regular file extent.
Now, a regular file extent is shared by a btrfs file inode with DATACSUM
and ext*_image with NODATACSUM, and it has checksum in csum tree.
This will cause btrfsck complain about odd checksum, since ext*_image is
set NODATACSUM but has checksum generated from regular file extent.
This patch makes convert completely obey datacsum setting, meaning
btrfs-convert will generate csum for every file extent by default.
Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
We provided format <path>|<device> in command line.
But btrfs device stats doesn't work if device is not mounted.
Also fix some tailing whitespace.
Signed-off-by: Chen Hanxiao <chenhanxiao@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Harald suggested that we remove the message from the fake fsck.btrfs
that some distros run at boot:
https://bugzilla.redhat.com/show_bug.cgi?id=1206502
"This output does not add anything, but is a disturbing element of
booting up a system. It's the only message I get, when starting my
system, before gdm is started."
I'm inclined to agree.
This makes the tiniest change to remove the message that's output for an
auto invocation. My guess is that it was just copied from the xfs
fsck.xfs and that no one actually cares about it.
Signed-off-by: Zach Brown <zab@zabbo.net>
Signed-off-by: David Sterba <dsterba@suse.cz>
AC_ARG_ENABLE(convert) sets $enable_convert, not $enable_btrfsconvert.
Signed-off-by: Mike Gilbert <floppym@gentoo.org>
Signed-off-by: David Sterba <dsterba@suse.cz>
Before this patch, qgroup show won't check btrfs qgroup status, so even
the INCONSISTENT flags is set, user is not aware of it.
This patch will include BTRFS_QGROUP_STATUS_ITEM in the search range and
check the flag, if there is any flag meaning the inconsistence of qgroup
data, info user.
NOTE: There is several kernel bugs from INCONSISTENT flags is always set
to RUNNING flags is not cleared until umount.
So this warning will always be here if using a newer kernel fixing these
bugs.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
This provides the basis for later qgroup related changes.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
Now parse_qgroupid() can resolve subvolume path into qgroupid.
This is quite handy for handling level 0 qgroupid, and user don't need
to resolve rootid by hand now.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[constify string in __is_subvol]
Signed-off-by: David Sterba <dsterba@suse.cz>
Now btrfs-debug-tree can print qgroup status flag as ON|INCONSISTENT
instead of 0x5.
BTW, this patch helped us to find a bug that INCONSISTENT flag is never
cleared in kernel.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[renamed to qgroup_flags_to_str]
Signed-off-by: David Sterba <dsterba@suse.cz>
Ctree.h of btrfs-progs contains wrong flags for btrfs_qgroup_status.
Update it with the one in kernel.
Also, introduce the inline function btrfs_qgroup_(level/subvid) to get
the level/subvolid of qgroup, to replace the old open-coded bit
operations.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
When runing restore under lzo compression, "bad compress length"
problems are encountered.
It is because there is a page alignment problem with the @decompress_lzo,
as follows:
|------| |----|-| |------|...|------|
page ^ page page
|
3 bytes left
When lzo pages are compressed in memory, we will ensure that the 4 bytes
length header will not cross a page boundary. There is a situation that
3 (or less) bytes are left at the end of a page, and then the 4 bytes
len is stored at the start of the next page. But the @decompress_lzo
doesn't go to the start of the next page and continue to read the next 4
bytes which crosses two pages, so a random value is fetched as a "bad
compress length".
So we check page alignment every time before we are going to fetch the
next @len and after the former piece of data is decompressed. If the
current page that we reach has less than 4 bytes left, then we should
fetch the next @len at the start of next page.
Signed-off-by: Gui Hecheng <guihc.fnst@cn.fujitsu.com>
[simplifed and moved into decompress_lzo]
Signed-off-by: David Sterba <dsterba@suse.cz>
convert-tests now test both 4096 and 16384 nodesizes.
Signed-off-by: Sebastian Thorarensen <sebth@naju.se>
Signed-off-by: David Sterba <dsterba@suse.cz>
Allow btrfs-convert to use nodesizes other than 4096. It defaults to
max(16384, pagesize), like mkfs.
Signed-off-by: Sebastian Thorarensen <sebth@naju.se>
Signed-off-by: David Sterba <dsterba@suse.cz>
check_node_or_leaf_size in utils.c now prints 'nodesize (or leafsize)'
instead of 'leafsize (or nodesize)' in the error messages, in order to
be less confusing for the user, as leafsize in mkfs is deprecated.
'ERROR: ' is also prepended to be consistent with other error messages.
Signed-off-by: Sebastian Thorarensen <sebth@naju.se>
Signed-off-by: David Sterba <dsterba@suse.cz>
Move the constant DEFAULT_MKFS_LEAF_SIZE to utils.h and rename it to
BTRFS_MKFS_DEFAULT_NODE_SIZE for consistency. Move the function
check_leaf_or_node_size to utils.c and rename it to
btrfs_check_node_or_leaf_size.
Signed-off-by: Sebastian Thorarensen <sebth@naju.se>
[added btrfs_ prefix]
Signed-off-by: David Sterba <dsterba@suse.cz>
The total size of devices was summed from raw partition size which is
wrong in two ways:
- if the device is missing, the size is 0 and it mismatches the size
summed from chunks, leading to bogus numbers like
Device unallocated: 16.00EiB
Used: 1.88TiB
Free (estimated): 8.00EiB (min: 8.00EiB)
- we should really account the device size that's occupied by btrfs, not
the real partition size altough it's the same most of the time
The sum of missing devices is now printed in the summary and any missing
device path is replaced with 'missing' instead of blank:
Data,RAID1: Size:972.00GiB, Used:962.15GiB
972.00GiB
/dev/sdb1 972.00GiB
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=94911
Reported-by: <raffix@web.de>
Signed-off-by: David Sterba <dsterba@suse.cz>
Anand Jain
Josef Bacik
George Wang
David Sterba
WorMzy Tykashi
Qu Wenruo
Chen Hanxiao
Zach Brown
Mike Gilbert
Gui Hecheng
Sebastian Thorarensen