This test requires XFS_SB_VERSION_MOREBITSBIT to be zero. ftype (which
is now enabled by default) causes this to be set, so detect it in mkfs
and disable it.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Use the xfs set/get metadata field helpers to detect the correct sfdir
field name prefix on v4-v5 filesystems. This enables us to test inode
link count corrections on a (deliberately) disconnected directory.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
From the kernel patch that this test examines ("xfs: detect agfl
count corruption and reset agfl"):
"The struct xfs_agfl v5 header was originally introduced with
unexpected padding that caused the AGFL to operate with one less
slot than intended. The header has since been packed, but the fix
left an incompatibility for users who upgrade from an old kernel
with the unpacked header to a newer kernel with the packed header
while the AGFL happens to wrap around the end. The newer kernel
recognizes one extra slot at the physical end of the AGFL that the
previous kernel did not. The new kernel will eventually attempt to
allocate a block from that slot, which contains invalid data, and
cause a crash.
"This condition can be detected by comparing the active range of the
AGFL to the count. While this detects a padding mismatch, it can
also trigger false positives for unrelated flcount corruption. Since
we cannot distinguish a size mismatch due to padding from unrelated
corruption, we can't trust the AGFL enough to simply repopulate the
empty slot.
"Instead, avoid unnecessarily complex detection logic and and use a
solution that can handle any form of flcount corruption that slips
through read verifiers: distrust the entire AGFL and reset it to an
empty state. Any valid blocks within the AGFL are intentionally
leaked. This requires xfs_repair to rectify (which was already
necessary based on the state the AGFL was found in). The reset
mitigates the side effect of the padding mismatch problem from a
filesystem crash to a free space accounting inconsistency."
This test exercises the reset code by mutating a fresh filesystem to
contain an agfl with various list configurations of correctly wrapped,
incorrectly wrapped, not wrapped, and actually corrupt free lists; then
checks the success of the reset operation by fragmenting the free space
btrees to exercise the agfl. Kernels without this reset fix will shut
down the filesystem with corruption errors.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Previously _scratch_mount didn't check the mount status and most
tests continue to run even if the mount failed (unless test checks
for the mount status explicitly). This would result in running tests
on the underlying filesystem (usually rootfs) and implicit test
failures, and such failures can be annoying and are usually hard to
debug.
Now _fail test by default if _scratch_mount failed and introduce
_try_scratch_mount for tests that need to check mount results
themselves.
Suggested-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Kernel commit f59cf5c29919 ("xfs: remove "no-allocation"
reservations for file creations") is known to cause earlier ENOSPC
conditions, and xfs/015 is affected in the way that no new file/dir
can be created in a newly created 16M XFS with both reflink and
rmapbt enabled, thus xfs/015 fails due to the missing test dirs.
So enlarge the initial fs size to 32M that we're able to create new
dir/file before growing the filesystem size.
Note that we doubled the fs initial size, we need to double the
required space too.
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
The XFS rmapbt extent swap mechanism performs an extent by extent
swap to ensure the rmapbt is rectified with the appropriate extent
owner information after the operation. This implementation suffers
from a corner case that requires extra reservation if the swap
operation results in bouncing one of the associated inodes between
extent and btree formats. When this corner case occurs, it results
in a transaction block reservation overrun and possible corruption
of the free space accounting.
This regression test provides coverage for this corner case. It
creates two files with a large enough extent count to require btree
format, regardless of inode size, and performs a sequence of extent
swaps between them with a decreasing extent count until all extents
are removed from the file(s). This ensures that one of the swaps
covers the btree <-> extent fork format boundary case.
This test reproduces fs corruption on rmapbt enabled filesystems
running on kernels without the associated extent swap fix.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Add three tests to look for quota bugs in xfs reflink. The first
test looks for problems when we have speculative cow reservations in
memory, we chown the file, but the reservations don't move to the
new owner. The second test checks that we remembered to dqattach
the inodes before performing reflink operations. The third is a
stress test for reflink quota handling near enospc and helped us to
find a directio cow write corruption bug when free space is
fragmented.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
A directory corrupted into a symlink is caught by the local format
ifork verifiers, and previously this test failed to catch such
kernel bug, so fix the test.
Note that the local format ifork verifiers were introduced by commit
0795e004fd4f ("xfs: create structure verifier function for short
form symlinks"), and didn't get wired up until commit 9cfb9b47479e
("xfs: provide a centralized method for verifying inline fork
data").
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Remove the -y parameter from scrub runs since we're removing that
option from xfs_scrub.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
XFS headers are supposed to be included in a certain order so that
inline functions actually compile correctly. For the most part the
shell feeds us the files in an order that works, but with the
addition of the xfs_dir2_dirblock_bytes function this doesn't always
work now. Therefore, explicitly #include the headers in the
required order.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
If log stripe unit isn't a multiple of the fs blocksize and
mounting, the invalid sb_logsunit leads to crash as soon as we try
to write to the log.
Signed-off-by: xiao yang <yangx.jy@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
1) Introduce _require_no_xfs_bug_on_assert helper to check if XFS is
built with CONFIG_XFS_ASSERT_FATAL, and call _require_no_xfs_debug
if bug_on_assert is not available.
2) Apply _require_no_xfs_bug_on_assert in xfs/098 and xfs/115.
3) Move filter_xfs_dmesg from xfs/098 to common/filter, and rename
it as _filter_assert_dmesg.
[eguan: update comment and _notrun message a bit]
Signed-off-by: xiao yang <yangx.jy@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Fixes direct invocations of xfs_repair to add in -r option if required.
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Richard Wareing <rwareing@fb.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
To better exercise the data path code of realtime subvolumes, we
will set rtinherit=1 during mkfs calls. For tests which this is not
desired we introduce a _require_no_rtinherit function to opt out of
this behavior.
Signed-off-by: Richard Wareing <rwareing@fb.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Some tests do not play well with realtime devices, in an effort to
produce a stable set of test which exercise the realtime code paths
we introduce a _require_no_realtime function to allow tests to opt
out of realtime subvolume test runs.
And to make tests generic/409-411 work well with rt device, teach
_get_mount now honors $SCRATCH_OPTIONS.
Signed-off-by: Richard Wareing <rwareing@fb.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
When the first writeback and the retried writeback of dquota buffer
get the same IO error, XFS will let xfsaild to restart the writeback
and xfs_qm_dqflush_done() will not be invoked. xfsaild will try to
re-push the quota log item in AIL, the push will return early
everytime after checking xfs_dqflock_nowait(), and xfsaild will try
to push it again.
IOWs, AIL will never be empty, and the umount process will wait for
the drain of AIL, so the umount process hangs.
Signed-off-by: Hou Tao <houtao1@huawei.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
It's common to get and set the values of fields in XFS super block,
so factor them out as scratch_xfs_[get|set]_sb_field, reimplement
them based on _scratch_xfs_[get|set]_metadata_field, and update the
related test cases accordingly.
Also move _scratch_xfs_[get|set]_metadata_field from common/fuzzy to
common/xfs.
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Getting xfs/040 to "pass" takes a bit of effort, however the effort
to require updating xfsprogs is purely an xfsprogs maintainer task
only. There no functional gain by users of xfs or a QA team to get
this test to pass. This is not trivial from the current description
so document this.
Signed-off-by: Luis R. Rodriguez <mcgrof@kernel.org>
Reviewed-by: Bill O'Donnell <billodo@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
We're removing from XFS the ability to perform no-allocation file
creation. This was added years ago because some customer of SGI
demanded that we still be able to create (empty?) files with zero
free blocks remaining so long as there were free inodes and space in
existing directory blocks. This came at an unacceptable risk of
ENOSPC'ing midway through a transaction and shutting down the fs, so
we're removing it for the create case having changed our minds 20
years later.
However, some tests fail as a result, so fix them to be more
flexible about not failing when a dir/file creation fails due to
ENOSPC.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
util-linux commit ea848180dd34 ("libmount: add
mnt_context_get_excode()") since v2.30 changed the error message on
EUCLEAN and ESTALE again (and maybe other errno too):
- mount: <device> on <mountpoint> failed: Structure needs cleaning
+ mount: <mountpoint>: mount(2) system call failed: Structure needs cleaning.
and it causes xfs/005, overlay/037 to fail (and probably xfs/333 too,
but it's always _notrun for now).
And what's more, the mentioned tests would also fail when testing
with util-linux prior to v2.21, no one complained just because the
tests are usually _notrun on such old distributions that ship
util-linux < v2.21.
So let's filter out the changing parts and keep the error message
simple.
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Add a couple of tests to check that we don't leak inodes or dquots
if CoW recovery fails and therefore the mount fails.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
xfs_io 4.14 will gain the ability to print error messages when
pwrite+fsync fail. Certain tests use the error injector to cause
failures, so the errors are expected. Since we test for a shut down
filesystem after the error injection, we can push the error messages to
the log.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Darrick J. Wong
Eryu Guan
Brian Foster
xiao yang
Richard Wareing
Hou Tao
Luis R. Rodriguez