There is a bug in trim code which leads to fstrim accessing beyond
device boundary.
The test case will check if fstrim, then shrink, then fstrim, all of
them works without problem.
The fix is titled "btrfs: trim: fix underflow in trim length to prevent
access beyond device boundary".
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
In xfs/030, we zero the entire AGFL which makes it corrupt because the
AG superblock block is /never/ free. In commit dcd6c2e1 ("xfs_repair:
convert to libxfs_verify_agbno") we enhanced repair to complain about
zeroed entries in the AGFL, so add that to the golden output.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
XFS filesystems require V5 format to support all three quota types at
the same time, so skip the test if we're testing XFS older than that.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
btrfs is clever enough to figure out which devices constitute the
sprout fs even without specifying them explicitly with -o device.
Additionally, explicitly settings the devices via -o device reduces
coverage of the test since it didn't detect breakage a local change
introduced. Without -o device instead this breakage was detected.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Currently generic/501 fails on machines with a page size different from 4K
(like ppc64le), because the clone operation fails with -EINVAL due to the
fact we pass it an offset that is 4K aligned but not aligned to the page
size of the machine.
The test doesn't actually need offsets and lengths to be 4K aligned, so
just update the test to use offsets and lenghts that work for page size.
Also add a comment mentioning that a file size of at least 16Mb was a
necessary condition to trigger the btrfs bug.
The test is a regression test for a btrfs issue fixed by kernel commit
bd3599a0e142cd ("Btrfs: fix file data corruption after cloning a range
and fsync"), which landed in kernel 4.18.
Since I couldn't compile a 4.17 kernel on debian testing, I tried this
with a 4.18 kernel with that commit reverted, and it fails as expected
on a x86_64 box:
$ ./check generic/501
FSTYP -- btrfs
PLATFORM -- Linux/x86_64 debian9 4.18.0-btrfs-next-64 #1 SMP (...)
MKFS_OPTIONS -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1
generic/501 1s ... - output mismatch (see .../xfstests/results//generic/501.out.bad)
--- tests/generic/501.out 2020-07-22 14:50:12.585674202 +0100
+++ /home/fdmanana/git/hub/xfstests/results//generic/501.out.bad ...
@@ -2,4 +2,4 @@
File bar digest before power failure:
69319d0343ab8f5ea564167da445addc SCRATCH_MNT/bar
File bar digest after power failure:
-69319d0343ab8f5ea564167da445addc SCRATCH_MNT/bar
+21de7d7325fe4dae1f3311d5a76f819f SCRATCH_MNT/bar
...
(Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/generic/501.out ...
Ran: generic/501
Failures: generic/501
Failed 1 of 1 tests
Without the commit reverted it passes as expected.
Reported-by: Murphy Zhou <jencce.kernel@gmail.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Command "quota enable" triggers a quota rescan, but it can finish
quick in some machines leading to the next command "quota rescan" to
be able to start scanning again, and then printing "quota rescan
started" making the test fail.
In some machines this don't happen because the first rescan
initiated by "quota enable" is still running when "quota rescan" is
executed, returning -EINPROGRESS from ioctl
BTRFS_IOC_QUOTA_RESCAN_STATUS and not printing the message.
Ignoring any output from "quota rescan" solves the issue in both
cases, and this is already being done by others tests as well.
Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Some generic tests explicitly run 'mknod' or 'mkfifo' command, however
the correspondent runtime check is missing, this results in a test
failure instead of selecting not to execute the test.
The change adds _require_mknod check to the next generic tests:
* generic/157
* generic/158
* generic/294
* generic/423
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
In runtime the modified tests requires and utilizes fpunch command
of xfs_io, thus a tested filesystem shall support the feature and
running a filesystem check by "punch" group selection should include
these tests:
* generic/176
* generic/415
* generic/447
* generic/469
* generic/562
In addition and due to the same reasons add generic/469 to zero test
group, the test already contains _require_xfs_io_command "fzero" check.
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Test if dirtying many inodes (which can delay umount) then
unmounting and quickly mounting again causes the mount to fail.
A race, which breaks the test in btrfs, is fixed by kernel commit
48cfa61b58a1 ("btrfs: fix mount failure caused by race with umount")
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
These tests use an overlay mount as lower layer for a nested overlay.
That provides test coverage for a lower layer with null uuid and
lower layer with real inode numbers that use the high xino bits.
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The tests with multi lower layers rotate the upper layer into a lower
layer on specific operations such as mkdir, rename and link.
That provides test coverage for redirect_dir and index features.
The following variations are covered:
1. All layers all on scratch fs (--ov=10 --samefs)
2. All layers on scratch fs expect for lowermost on test fs (--ov=10)
3. Lowermost layer on test fs, one layer on tmpfs and the rest
on sratch fs (--ov=10 --maxfs=1)
4. Lowermost layer on test fs, one layer on scratch fs and the rest
are unique tmpfs instances (--ov=10 --maxfs=10)
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Add support for running unionmount-testsuite from xfstests.
This requires that unionmount-testsuite is installed under src dir or
that UNIONMOUNT_TESTSUITE variable points to the location of the
testsuite. It also requires a recent version of unionmount-testsuite
that supports setting basedir path via UNIONMOUNT_* environment variables.
Add tests for three basic configurations:
1. overlay with upper/lower on same fs
2. overlay with upper/lower not on same fs without xino
3. overlay with upper/lower not on same fs with xino
The samefs test uses scratch partition for lower/upper layers.
The non samefs tests use the scratch partition for upper layer and the
test partition for lower layer.
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The test checks regressions on quota accounting and therefore it relies
on quota support by a tested filesystem, this makes it reasonable to
include the test into 'quota' test group for better managing of test
groups.
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The change makes one-in-one correspondence between generic tests in
'insert' test group and tests with $(_require_xfs_io_command "finsert")
runtime check:
* generic/348 -- insert group association is removed,
* generic/458 -- insert group association is added.
The intention of the change is to get better summarized association
between tests and test groups to be able to operate with test groups
in a better way.
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
clear WARN_ONCE state before each test to allow a potential problem
to be reported for each test
[Eryu: replace "/sys/kernel/debug" with $DEBUGFS_MNT ]
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The test runs fstrim utility to stress a filesystem, however it implies
that the filesystem supports ioctl(FITRIM), this is checked in runtime by
_require_batched_discard() in the test itself, however the assignment of
the test to trim group is missing.
The intention of the change is to get better association between tests
and test groups.
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The test does not belong to acl group of tests, and expectedly the test
is supposed to pass on a filesystem which supports file attributes but
not ACLs.
In addition the test should be removed from attr group of tests, because
the attr group specifies tests which operate over extended file attributes,
this test does not fall into this category.
The change is intended to improve the assignment of tests to groups
to get better testing coverage, if a list of run tests is selected by
test group attributes.
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
A few generic tests have a _require_attrs runtime check for support
of extended file attributes by a tested filesystem, review shows that
the tests listed below are definitely using extended file attributes,
thus this implies that the tested filesystem should provide support of
the feature in general.
Adding the tests explicitly to attr group of tests allows to get a better
granularity of selecting or deselecting tests by groups in accordance to
extended file attributes feature given by a filesystem, the change affects
the next list of tests:
* generic/037
* generic/066
* generic/103
* generic/337
* generic/449
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The change fixes false positives from generic/554 and generic/565 tests,
when supplementary output of an mkfs utility to stderr apprears to be mixed
into the test output, which results into non-empty diff with the expected
output found in *.out files.
The change unifies _scratch_mkfs redirections among all generic tests, these
two modified tests are the only ones with the swapped order of output
redirections.
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The change fixes false positives reported by a few generic tests under
circumstances, when an mkfs utility prints its output to stderr and then it
appears in produced *.out files. The change is intended to fix such issues
in the tests, which utilize _require_scratch_swapfile().
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Test generic/530 doesn't make sence without metadata journalling as in
that case, there's no way to recover shutdown fs besides fsck. ext4
can be configured without a journal and it supports shutdown ioctl even
in that mode which makes this test fail for that configuration. Add
requirement for metadata journalling to this test so that it's properly
skipped.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Frequently when trying to reproduce a problem I want to run a set of
specific tests in a loop, over and over again. I run fstests from a
set of run scripts that have non-trivial overhead (e.g. patterning
block devices before the runs start), so if all I want to do is run
the same test 100x, using a shell loop over the entire run
scripts reduces the iteration rate substantially.
Hence add an option to check to allow fstests to loop a number of
times over the configured test set without stopping.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Add a regression test to see if kernel hangs in order to
look after CVE-2020-12655 and check if the corresponding
fix is applied as well.
Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Qu Wenruo
Darrick J. Wong
Nikolay Borisov
Filipe Manana
Marcos Paulo de Souza
Vladimir Zapolskiy
Boris Burkov
Anand Jain
Amir Goldstein
Lukas Czerner
Jan Kara
Dave Chinner
Gao Xiang