This test is motivated by an issue found in btrfs.
It tests that after syncing the filesystem, adding a hard link to a file,
syncing the filesystem again, doing a write to the file that increases
its size and then doing a fsync against that file, durably persists the
data written to the file. That is, after log/journal replay, the data
is available.
The btrfs issue is fixed by the commit titled:
"Btrfs: fix fsync data loss after append write"
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
These changes make it possible to run more of the tests on busybox.
Signed-off-by: Ari Sundholm <ari@tuxera.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Add a regression test for a problem where attempting to delete the
default subvolume would fail (as expected), but not until after all
submounts under the subvolume were unmounted.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
XFS dynamic inode allocation has a fundamental limitation in that an
inode chunk requires a contiguous extent of a minimum size. Depending on
the level of free space fragmentation, inode allocation can fail with
ENOSPC where the filesystem might not be near 100% usage.
The sparse inodes feature was implemented to provide an inode allocation
strategy that maximizes the ability to allocate inodes under free space
fragmentation. This test fragments free space and verifies that
filesystems that support sparse inode allocation can allocate a minimum
percentage of inodes on the fs.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
When we enable quota, btrfs will rescan quota numbers. We need
to wait the rescan finished before any more operations on btrfs
qgroups. Otherwith, the new btrfs-progs would WARN out:
WARNING: Rescan is running, qgroup data may be incorrect.
It would make btrfs/022 failed.
Signed-off-by: Dongsheng Yang <yangds.fnst@cn.fujitsu.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Commit bbe051c841d5 ("xfs: disallow ro->rw remount on norecovery mount")
disabled rw remount on norecovery ro mount, this test makes sure the
behavior is correct.
Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Call 'udevadm settle' or 'udevsettle' or 'sleep 1' to make sure new lv
is ready for use before making filesystem on it, depends on which
command is available on the system.
Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
The testcase tests 2 corner cases:
Length is zero
Length is smaller than block size
Correct the beginning description by changing "of" to "or".
Signed-off-by: Wang Sheng-Hui <shhuiw@foxmail.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
In certain cases, the extent size hints can cause maximum extent
size overflows resulting in extent tree corruptions. This test
exercises the original reproducer, and another corner case
demonstrated to expose problems on 1k block size filesystems.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
--
Version 2:
- TESTDIR->TEST_DIR
- append output to $seqres.full
This option causes fsstress to delete the test directory between each
run. This is the opposite from the ltp version of fsstress, which
deletes the test directory by default, and the -c option caused ltp's
fsstress _not_ to delete the test directory.
It can be useful to be able to have the same test behavior as ltp
version, and although reversing the sense of the option is
unfortunate, it also reserves the -c option, which makes it a bit
easier if we want to eventually have one version which is a superset
of the xfstest and ltp version of fsstress.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
This feature is in the ltp version of fsstress; port it into
xfstests.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
generic/019 was failing with:
./tests/generic/019: line 65: /sys/block/pmem0p2/make-it-fail: No such file or directory
When using a partition, the file needed is located at
/sys/block/pmem0/pmem0p2/make-it-fail.
Rather than attempt to deduce whether a block device is a partition
or not, use the symlinks located in /sys/dev/block/ to find the right
location for the make-it-fail file.
Also change btrfs/088 to use the new _sysfs_dev function.
Signed-off-by: Matthew Wilcox <willy@linux.intel.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
There was some confused about what the fs was supposed to do when you truncate
at i_size with preallocated space past i_size. We decided on the following
things
1) truncate(i_size) will trim all blocks past i_size.
2) truncate(x) where x > i_size will not trim all blocks past i_size.
This test is to make sure we're all acting sanely. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
xfs/111 is failing today, primarily because it obliterates the
root inode chunk and the verifiers fail the mount - i.e. the test
fails to properly test the thing it's meant to test.
Change the test so that the induced corruption is further into the
filesystem, but still hitting the inodes which have been created
for the test, so that the fs can mount and continue.
This requires that the helper binary (itrash) take an offset, which
we will figure out by using xfs_db.
This changes the locations of the inodes we hit; we're not really
going to be able to predict that terribly well, so remove the
output which shows inode offsets, and just keep track of whether
we managed to obliterate any at all.
The test still fails, because the fs is corrupted; this was done
intentionally, so run xfs_repair before the test exits to fix
things up.
(This test doesn't run often; it's not in the auto group, and
all the failures are extremely noisy and time consuming...)
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
This test checks block usage on quota inodes based on the inode number
values stored in the superblock. Version 5 superblocks (crc=1) have an
independent project quota inode field to support concurrent group and
project quotas. Older superblocks do not have the pquotino field. The
gquotino field is reused for project quotas.
The test currently unconditionally uses the pquotino field to determine
the project quota inode. This causes failures on pre-v5 superblocks as
the test queries the block usage of an incorrect inode number. Update
the test to use gquotino as the project quota inode on such filesystems.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
fixes 9435b92 common: _require_command needs to handle parameters
Also quoted $_command because _require_command may be called with an
empty $1 parameter, e.g.:
_require_command "$MY_UTIL_PROG" my_util # but $MY_UTIL_PROG is empty
[ -x ] returns true.
[ -x "" ] returns false, as required here.
Signed-off-by: Omer Zilberberg <omzg@plexistor.com>
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Use generic quota tools with gfs2.
Fixes "xfs_quota: cannot setup path for mount /mnt/scratch: No such
device or address"
Signed-off-by: Andrew Price <anprice@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
From the xfs_repair manpage:
xfs_repair run without the -n option will always return
a status code of 0.
So we must use "-n" to detect corruption in this test.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
xfs_repair compares attr names in the root namespace to
two special/reserved names, "SGI_ACL_FILE" and "SGI_ACL_DEFAULT"
and if the value in them aren't valid acls, flags this as
an inconsistency.
However, due to various bugs, xfs_repair may only compare
a smaller portion of the on-disk value; hence either
substrings or superstrings may match, and false-positive
corruption will be detected. This test checks for those
false positives; i.e. the ACL names created in this test
may cause xfs_repair to "fix" them, but it should not.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Check if setting the file access and modification times to the current time
and to a specific timestamp is allowed when expected.
In generic/126, remove a left-over temporary file.
Signed-off-by: Andreas Gruenbacher <andreas.gruenbacher@gmail.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
With the change to CRCs by default, some tests are updated to call mkfs
with "-m crc=0" option directly, and this breaks testings on older
distros where mkfs.xfs doesn't have crc support.
Introduce a new variable to tell if mkfs.xfs supports v5 xfs and do
tweaks in _scratch_mkfs_xfs_opts() based on it.
Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
When a large IO is done as a single buffer, there is no guarantee
that it will partially succeed when close to ENOSPC. The test
assumes that the kernel is going to break the write down into
smaller chunks (i.e. buffered IO breaking it down into PAGE_SIZE
allocations), but certain configurations will not do this. e.g.
extent size hints are set or DAX is being used) and hence the large
write fails completely as there is not space for the entire
allocation to be made.
Hence break the final write in the test up into multiple small
writes, thereby acheiving the same effect - ensuring that we can
write more data after removing some space....
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
On certain configurations (e.g. MOUNT_OPTIONS="-o dax") we get
different allocation patterns due to the writes being done in
multiple pwrite() calls. e.g. the write is 8k, but the buffer size
is 4k, and so the filesystem sees 4k writes. If the filesytem is not
using delayed allocation, then the allocation context is a 4k write
rather than an 8k write and so they don't get appropriately aligned.
Fix this by making the write buffer the same size and the writes
being done.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Filipe Manana
Ari Sundholm
Omar Sandoval
Brian Foster
Dongsheng Yang
Eryu Guan
Wang Sheng-Hui
Dave Chinner
Theodore Ts'o
Matthew Wilcox
Josef Bacik
Eric Sandeen
Omer Zilberberg
Andrew Price
Eric Sandeen
Andreas Gruenbacher