This is a regression test for kernel commit 65cd913ec9d9
("ovl: invalidate readdir cache on changes to dir with origin")
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Kernel commit 708fa01597fa ("ovl: allow upperdir inside lowerdir")
changes the rules w.r.t allowed overlayfs configurations, so the
upperdir/lowerdir test that expects an error fails.
Adapt the test to check the configuration that is still not allowed
(lowerdir inside upperdir), which had no test coverage.
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Combine these two helpers into a single generic function so that we can
use it in the next patch to test a regression when running overlayfs
atop xfs.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Overlayfs added the ability to set inode flags (e.g. chattr +i) in
kernel 5.10 by commit 61536bed2149 ("ovl: support [S|G]ETFLAGS and
FS[S|G]ETXATTR ioctls for directories").
Icenowy Zheng reported [1] a regression in that commit that causes
a deadlock when setting inode flags on lower dir.
The regression was fixed by commit b854cc659dcb ("ovl: avoid deadlock
on directory ioctl") and applied to kernel 5.10.15.
[1] https://lore.kernel.org/linux-unionfs/20210101201230.768653-1-icenowy@aosc.io/
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Chengguang Xu reported [1] that append-only flag is lost on copy-up.
I had noticed that for directories, immutable flag can also be lost
on copy up (when parent is copied up). That's an old overlayfs bug.
Fixing this requires some VFS API changes that Miklos has proposed[2]
[1] https://lore.kernel.org/linux-unionfs/20201226104618.239739-1-cgxu519@mykernel.net/
[2] https: //lore.kernel.org/linux-fsdevel/20210203124112.1182614-1-mszeredi@redhat.com/
[Eryu: referencing the thread of the pending fix]
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
commit 61536bed2149 ("ovl: support [S|G]ETFLAGS and FS[S|G]ETXATTR
ioctls for directories") makes the comment in test header inaccurate.
Fix the comment to include this information.
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
When nfs_export is enabled, the link count of upper dir
objects are more then the expected number in this testcase.
Because extra index entries are linked to upper inodes.
Signed-off-by: Murphy Zhou <jencce.kernel@gmail.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Starting with version 2.41 of libcap, the output of the getcap program
changed and therefore some existing tests fail when the installed version
of libcap is >= 2.41 (the latest version available at the moment is 2.44).
The change was made by the following commit of libcap:
commit 177cd418031b1acfcf73fe3b1af9f3279828681c
Author: Andrew G. Morgan <morgan@kernel.org>
Date: Tue Jul 21 22:58:05 2020 -0700
A more compact form for the text representation of capabilities.
While this does not change anything about the supported range of
equivalent text specifications for capabilities, as accepted by
cap_from_text(), this does alter the preferred output format of
cap_to_text() to be two characters shorter in most cases. That is,
what used to be summarized as:
"= cap_foo+..."
is now converted to the equivalent text:
"cap_foo=..."
which is also more intuitive.
So add a filter to change the old format to the new one, an helper that
calls getcap with that filter, make existing tests use the new helper and
update their golden output to match the new output format of getcap.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
1) If underlying filesystem is ext4, the lower overlay inodes do not
have the MSB set (e.g. file ino 8590721028 = 0x2000C0004) because
ext4 has a known inode number limit of 32bit (see ovl_can_decode_fh).
2) With nested xino feature, the nested overlay inodes have the MSB set
(e.g. file ino 9223372045445496836 = 0x80000002000C0004).
It is expected different inode numbers which casue the failure of
overlay/071, so fix this failure by removing the unneeded comparision.
Suggested-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Xiao Yang <yangx.jy@cn.fujitsu.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
fstests only supports Linux, so get rid of this unnecessary predicate.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Change OVL_BASE_TEST_MNT -> OVL_BASE_TEST_DIR, since the former is not
actually defined anywhere.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Explicitly specify the xattr namespace required for tests.
This allows tests to be skipped correctly for filesystems
that don't support all xattr namespaces.
This changes all tests that require anything other than
the "user" xattr namespace. When called without arguments
as before, _require_attrs() still defaults to the "user"
namespace, so those tests do not need to be changed.
Signed-off-by: Frank van der Linden <fllinden@amazon.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.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>
Test two overlayfs file handle bugs:
1. Failure to query file handle size
Fixed by kernel commit 144da23beab8:
ovl: return required buffer size for file handles
2. Kernel OOPS on open by hand crafted malformed file handle
Fixed by kernel commit 9aafc1b01873:
ovl: potential crash in ovl_fid_to_fh()
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Kernel commit c21c839b8448 "ovl: whiteout inode sharing" results in
a temp whiteout file resident inside work dir.
Test overlay/031 is a regression test for two user visible bugs:
1. Exposed whiteouts in overlay
2. Failure to remove directory
It also has a sanity tests for a harmless by-product of the bug -
a residue file in work dir.
The new temp whiteout file looks like a residue and causes the test
to fail.
Drop this sanity test, because it is not vital to the regression test.
We could also check if the residue is a single whiteout, but that is
not really needed, so best not poke into overlay internal work dir.
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
This is a test for whiteout inode sharing feature.
[Amir] added check for whiteout sharing support
and whiteout of lower dir.
Signed-off-by: Chengguang Xu <cgxu519@mykernel.net>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
This is a variant on test overlay/034.
This variant is mangling upper hardlinks instead of lower hardlinks
and does not require the inodes index feature.
This is a regression test for kernel commit 83552eacdfc0
("ovl: fix WARN_ON nlink drop to zero")
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
When overlayfs nfs_export feature is enabled by default in either kernel
config or module parameters, this test fails:
mount: /tmp/8751/mnt: mount(2) system call failed: Stale file handle.
cat: /tmp/8751/mnt/bar: No such file or directory
The reason is that nfs_export depends on index feature and with index
feature enabled, an upper/work dirs cannot be reused for mounting with
a different lower layer.
To reproduce the failure do:
echo Y > /sys/module/overlay/parameters/index
echo Y > /sys/module/overlay/parameters/nfs_export
before running the test.
Fix the failure by explicitly re-creating upper/work dirs.
Reported-by: Chengguang Xu <cgxu519@mykernel.net>
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 were checking that nfs_export feature was successfully enabled
by greping for nfs_export=on option in /proc/mounts.
This check was incorrect if the module default was nfs_export enabled
and caused test to not run with the message:
cannot enable nfs_export feature on nested overlay
Use a helper that checks this condition correctly.
Reported-by: Chengguang Xu <cgxu519@mykernel.net>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
If the arbitrary char/block devices 1:1 do not exist in the system,
the tests fail.
Use /dev/zero and loop device instead of made up device numbers.
Reposted-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Also test that d_ino of readdir entries and i_ino from /proc/locks are
consistent with st_ino and that inode numbers persist after rename to
new parent, drop caches and mount cycle.
With nested xino configuration, directory st_ino is not persistent and
its st_ino/d_ino/i_ino values are not consistent, so test only non-dir
in this test.
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Amir Goldstein
Darrick J. Wong
Murphy Zhou
Filipe Manana
Xiao Yang
Darrick J. Wong
Frank van der Linden
Chengguang Xu