In btrfs-progs v4.10 we had a behaviour change where starting a balance
operation without any filters results in a delay of 10 seconds and a
warning is printed to stdout that warns that a full balance is about to
be made and that it can be a slow operation. The new flag '--full-balance'
was added in that release to avoid the 10 seconds delay and the warning
message.
Our existing helper _run_btrfs_balance_start() uses that new balance flag
if we are running a btrfs-progs version that has it, to avoid that 10
seconds wait.
Make all existing btrfs tests that trigger balance operations use the
_run_btrfs_balance_start() helper, so that we avoid wasting time and
speed up some of the tests. In particular test btrfs/014 is now about 10x
faster and tests btrfs/060 to btrfs/064 3 to 5 times faster (depending
on the fsstress random load).
Besides speeding up many tests that do balance operations it also fixes
functional problems:
1) Since btrfs-progs v4.10 the test case btrfs/014 got broken, because
its purpose is to run balance and snapshot creation in parallel,
and that wasn't happening anymore because all snapshots were being
created during the 10 seconds delay of the first balance operation,
so balance and snapshot creation was being serialized instead of
running in parallel.
Fixing this test to avoid the 10 seconds delay immediately
exposes a regression that went into kernel 5.7-rc1 which is fixed
by the following commit
aec7db3b13a0 ("btrfs: fix setting last_trans for reloc roots")
2) Test cases btrfs/060 to btrfs/064 now spend much more time running
fsstress, balance and other operations in parallel, there's no
longer intervals of 10 seconds where balance is not running
concurrently with those other operations, making the tests a lot
more useful again.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The use of run_check() immediately stops a test because it calls the
_fail() function when execution of its argument fails. This is
generally not encouraged in fstests as it prevents a test from
detecting further problems after that failure.
Since the next patch in this series updates other tests to use
_run_btrfs_balance_start() for which a failure to run balance can be
expected (btrfs/187 for example), remove the use of run_check() from
_run_btrfs_balance_start(). Existing tests that use
_run_btrfs_balance_start() now redirect standard output to the
test's .full file for debugging purposes. In case balance fails the
tests will fail due to unexpected output from the standard error.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Some tests exercise balance but are not included in the 'balance' group,
so just add them to that group.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The rationale behind the proposed changes is to correct and improve
granularity of some generic tests, so it becomes possible to control
lists of run tests better, if there is given additional information
about a test environment and tested filesystems.
* generic/459 test is added to freeze group
* generic/482 test is added to thin group
* generic/529 test is added to acl group
* generic/530 test is added to shutdown group
Signed-off-by: Vladimir Zapolskiy <vladimir@tuxera.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
We have a bug in the current kernel merge window (for 5.7) that results
in data loss when cloning an inline extent into a new file (or an empty
file. This change adds a test for such case into the existing test case
btrfs/205, because it's the test case that tests all the supported
scenarios for cloning inline extents in btrfs.
Linux kernel commit 4fdb688c7071 ("btrfs: fix lost i_size update
after cloning inline extent") fixes the regression.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.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>
Add a regression test for the bug fixed by commit 10a98cb16d80 ("xfs:
clear PF_MEMALLOC before exiting xfsaild thread").
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Add a regression test for a bug in the FS_IOC_REMOVE_ENCRYPTION_KEY
ioctl fixed by commit 2b4eae95c736 ("fscrypt: don't evict dirty inodes
after removing key").
This ioctl is also tested by generic/580 and generic/581, but they
didn't cover the case where this bug occurs.
This test detects the bug on ext4, f2fs, and ubifs. The multi-threaded
part of the test actually still fails on ubifs even with the fix, due to
another kernel bug which I'm working on fixing.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
If $TEST_DIR is a substring of $SCRATCH_MNT (like /mnt and /mnt2),
this test always fail. Change to use _filter_testdir_and_scratch
which will filter the longer string first if the other string is a
substring of the longer one.
Signed-off-by: Yongcheng Yang <yongcheng.yang@gmail.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Add an option to the locktest program to check for fcntl setlease
support, and skip the generic/571 test on filesystems where the
check returns EINVAL.
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Reviewed-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
When I run xfs/191 with upstream xfsprogs, I get the following
errors because mkfs.xfs binary has changed a lot(use loop device
to avoid stripe alignment affect).
-------------------------
pass -n size=2b /dev/loop0
pass -d agsize=8192b /dev/loop0
pass -d agsize=65536s /dev/loop0
pass -d su=0,sw=64 /dev/loop0
pass -d su=4096s,sw=64 /dev/loop0
pass -d su=4096b,sw=64 /dev/loop0
pass -l su=10b /dev/loop0
fail -n log=15 /dev/loop0
fail -r rtdev=/mnt/xfstests/test/191-input-validation.img /dev/loop0
fail -r size=65536,rtdev=/mnt/xfstests/test/191-input-validation.img /dev/loop0
fail -i log=10 /dev/loop
--------------------------
"pass -d su=0,sw=64 /dev/loop0", expect fail, this behavior has been
fixed by commit 16adcb88(mkfs: more sunit/swidth sanity checking).
"fail -n log=15 /dev/sda11" "fail -i log=10 /dev/sda11", expect pass,
this option has been removed since commit 2cf637c(mkfs: remove
logarithm based CLI option).
"fail -r size=65536,rtdev=$fsimg /dev/sda11" "fail -r rtdev=$fsimg
/dev/sda11" works well if we disable reflink, fail if we enable
reflink. It fails because reflink was not supported in realtime
devices since commit bfa66ec(mkfs: don't create realtime filesystems
with reflink enabled).
Since xfsprogs v4.15.0-rc1(commit 68344ba0f mkfs: introduce default
configuration structure), we have deault sectorsize and blocksize.
So some cases without 's' or 'b' suffix trun into pass.
Signed-off-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Set different grace time, make sure each of quota (user, group and
project) timers can be set (by setquota) and get (by repquota)
correctly.
Signed-off-by: Zorro Lang <zlang@redhat.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Some filesystems do not support negative c/m/atime timestamps. And
for these filesystems, generic/258 will fail.
Since there's currently no way to check sb->s_time_min from
userspace, this patch adds a new common/rc function
_require_negative_timestamps() that will _notrun depending on the
$FSTYP variable.
Signed-off-by: Luis Henriques <lhenriques@suse.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
It turns out that repquota (which reports in units of 1k blocks) reports
rounded up numbers when the fs blocksize is 512 bytes. However, xfs_io
stat always reports block counts in units of 512 bytes. If the number
of (512b) file blocks is not an even number, the "$3 / 2" expression
will round down, causing the test to fail. Round up to the nearest 1k
to match repquota's behavior.
Reported-by: zlang@redhat.com
Fixes: 6b04ed0545 ("generic: test unwritten extent conversion extent mapping quota accounting")
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>
The 'grep -c "error"' check was causing false positive results for
cifs.ko when running it with 'echo 1 > /proc/fs/cifs/cifsFYI'.
That is, the test would fail when cifs.ko prints out a non-error
message like this:
[ 2320.903987] fs/cifs/smb2maperror.c: Mapping SMB2 status code 0x80000006 to POSIX err -61
Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Suggested-by: Zorro Lang <zlang@redhat.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>
Test a scenario were we fsync a range of a file and have a power
failure. We want to check that after a power failure and mounting
the filesystem, we do not end up with a missing file extent
representing a hole. This applies only when not using the NO_HOLES
feature.
This currently fails on btrfs but it is fixed by a patch for the kernel
that has the following subject:
"Btrfs: fix missing file extent item for hole after ranged fsync"
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Now btrfs can delete subvolumes based in ther subvolume id. This
makes easy for the user willing to delete a subvolume that cannot be
accessed by the mount point, since btrfs allows to mount a specific
subvolume and hiding the other from the mount point.
Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
In the blockgroup type single with nodesize 64K, the relocation of
the bg containing the swapfile is not being attempted during the
resize. So due to this the resize is successful and does not
generate the required 'Text file busy' error message as in the
golden output and so the testcase fails.
Fix this by replacing the mkfs created chunk with the bigger kernel
created chunk using balance, and then fill it up to the full. Upsize
to 3x of fssize once instead of first to 2G and then to 3G. Also
drop the unnecessary downsize to 2G step.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
This test case needs at least 3g scratch device space, check for it
before starting the test case.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
The minimum accepted allocsize mount option value is page size, which
causes the particular test to fail in architectures where page size >
block size. Fix it by basing the value on the platform page size rather
than the block size as obtained from mkfs. In addition add a filter so
that different values can be used without breaking the golden output.
Signed-off-by: Anthony Iliopoulos <ailiopoulos@suse.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Create a helper to run xfs_admin on the scratch device, then
refactor all tests to use it.
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>
If xfs_io's utimes command cannot interpret the arguments that are
given to it, it will print out "Bad value for [am]time". Detect
when this happens and drop the file out of the test entirely.
This is particularly noticeable on 32-bit platforms and the largest
timestamp seconds supported by the filesystem is INT_MAX. In this
case, the maximum value we can cram into tv_sec is INT_MAX, and
there is no way to actually test setting a timestamp of INT_MAX + 1
to test the clamping.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Apparently we don't have any tests which exercise the code path in
Btrfs that has to split up direct I/Os for RAID stripes. Add one to
catch the bug fixed by "btrfs: fix RAID direct I/O reads with
alternate csums".
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Filipe Manana
Vladimir Zapolskiy
Amir Goldstein
Eric Biggers
Yongcheng Yang
Scott Mayhew
Yang Xu
Zorro Lang
Luis Henriques
Darrick J. Wong
Paulo Alcantara (SUSE)
Marcos Paulo de Souza
Anand Jain
Anthony Iliopoulos
Omar Sandoval