Luis Chamberlain [Fri, 20 Aug 2021 01:04:02 +0000 (18:04 -0700)]
common/scsi_debug: use the patient module remover
If you try to run tests such as generic/108 in a loop
you'll eventually see a failure, but the failure can
be a false positive and the test was just unable to remove
the scsi_debug module.
We need to give some time for the refcnt to become 0. For
instance for the test generic/108 the refcnt lingers between
2 and 1. It should be 0 when we're done but a bit of time
seems to be required. The chance of us trying to run rmmod
when the refcnt is 2 or 1 is low, about 1/30 times if you
run the test in a loop on linux-next today.
Likewise, even when its 0 we just need a tiny breather before
we can remove the module (sleep 10 suffices) but this is
only required on older kernels. Otherwise removing the module
will just fail.
Some of these races are documented on the korg#212337, and
Doug Gilbert has posted at least one patch attempt to try
to help with this [1]. The patch does not resolve all the
issues though, it helps though.
This let's us remove the cheesy try loop. We keep the
udevadm settle call as it can help salvage buggy tests
which forgot to call it.
We also special-case where MODPROBE_PATIENT_RM_TIMEOUT_SECONDS
is set to "forever" and the initial module check finds its
in use, for that case we just try removing the module once
since fstests would not be the one leaving modues lingering
around, and waiting forever could mean you won't discover
the issue for a while.
[0] https://bugzilla.kernel.org/show_bug.cgi?id=212337
[1] https://lkml.kernel.org/r/
20210508230745.27923-1-dgilbert@interlog.com
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Luis Chamberlain [Fri, 20 Aug 2021 01:04:01 +0000 (18:04 -0700)]
common/module: add patient module rmmod support
When we call rmmod it will fail if the refcnt is greater than 0.
This is expected, however, if using test modules such as scsi_debug,
userspace tests may expect that once userspace is done issuing out
commands it can safely remove the module, and the module will be
removed.
This is not true for few reasons. First, a module might take a while
to quiesce after its used. This varies module by module. For example,
at least for scsi_debug there is one patch to help with this but
that is not sufficient to address all the removal issues, it just helps
quiesce the module faster. If something like LVM pvremove is used, as in
the case of generic/108, it may take time before the module's refcnt goes
to 0 even if DM_DEFERRED_REMOVE is *not* used and even if udevadm settle
is used. Even *after* all this... the module refcnt is still very
fickle. For example, any blkdev_open() against a block device will bump
a module refcnt up and we have little control over stopping these
sporadic userspace calls after a test. A failure on module removal then
just becomes an inconvenience on false positives.
This was first observed on scsi_debug [0]. Doug worked on a patch to
help the driver quiesce [1]. Later the issue has been determined to be
generic [2]. The only way to properly resolve these issues is with a
patient module remover. The kernel used to support a wait for the
delete_module() system call, however this was later deprecated into
kmod with a 10 second userspace sleep. That 10 second sleep is long gone
from kmod now though. I've posted patches now for a kmod patient module
remover then [3], in light of the fact that this issue is generic and
the only way to then properly deal with this is implementing a userspace
patient module remover.
Use the kmod patient module remover when supported, otherwise we open
code our own solution inside fstests. We default to a timeout of 100
seconds. Each test can override the timeout by setting the variable
MODPROBE_PATIENT_RM_TIMEOUT_SECONDS or setting it to "forever" if they
wish for the patience to be infinite.
This uses kmod's patient module remover if you have that feature,
otherwise we open code a solution in fstests which is a simplified
version of what has been proposed for kmod.
[0] https://bugzilla.kernel.org/show_bug.cgi?id=212337
[1] https://lore.kernel.org/linux-scsi/
20210508230745.27923-1-dgilbert@interlog.com/
[2] https://bugzilla.kernel.org/show_bug.cgi?id=214015
[3] https://lkml.kernel.org/r/
20210810051602.
3067384-1-mcgrof@kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Kai-Wen Hu [Fri, 20 Aug 2021 12:10:31 +0000 (20:10 +0800)]
generic/457: post-check fs on thin-vol device instead of scratch device
Recently, we found that this test running with btrfs didn't
pass the post-check fsck on scratch device. This issue occured
due to the commit
5faafc77f ("dm space maps: don't reset space map
allocation cursor when committing") and the use of SSD as scratch
device.
In this test, it does not directly use the filesystem on scratch
device. It runs test on dm-thin volume tying two linear-mapping
devices, and those two linear devices are created from the scratch
device. When the test runs, it will write data on dm-thin volume,
and dirty the underlying scratch device.
Since the commit (
5faafc77f) changes the dm space allocation policy,
it may cause more dirty blocks on underlying scratch device when we
write data on thin volume. So it makes more chance to overwrite the
position of btrfs-meta on the scratch device. Besides, our scratch
device is SSD, so there is only one copy of btrfs meta on the scratch
device. Once the btrfs-meta is overwritten, the btrfs volume will
crash, and cannot pass the fsck check.
Because, this test won't directly use the fs on scratch device. We
just disable the post-check fsck on scratch device, instead, we run
post-check fsck on thin-vol device.
Reviewed-by: Chung-Chiang Cheng <cccheng@synology.com>
Reviewed-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Kai-Wen Hu <kevinhu@synology.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Tue, 17 Aug 2021 23:53:05 +0000 (16:53 -0700)]
xfs/176: fix the group name
Filesystem shrink tests for xfs are supposed to be in the 'shrinkfs'
group, not 'shrink'.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Dave Chinner [Thu, 19 Aug 2021 06:00:44 +0000 (16:00 +1000)]
xfs/187: remove the test
Upstream changes mean that "-o noattr2" no longer removes the
on-disk ATTR2 feature bit from the superblock and so the failure
case this test exercises is no longer guaranteed to be a failure
case. Hence the test is now useless as a regression test, so remove
it.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Boyang Xue [Thu, 19 Aug 2021 08:07:51 +0000 (16:07 +0800)]
ext4: regression test for "tune2fs -l" after ext4 shutdown
Regression test for e2fsprogs commit:
e905fbe3fd0f ext4: Fix tune2fs checksum failure for mounted filesystem
This test runs "tune2fs -l" after ext4 shutdown. tune2fs reads
superblock checksum from the buffer cache. On unfixed kernels, the
checksum is incorrect until the writeout happens, so tune2fs fails
with "superblock checksum does not match" in this case.
Signed-off-by: Boyang Xue <bxue@redhat.com>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Thu, 19 Aug 2021 01:43:26 +0000 (18:43 -0700)]
generic: add swapfile maxpages regression test
Add regression test for "mm/swap: consider max pages in
iomap_swapfile_add_extent".
Cc: Gang Deng <gavin.dg@linux.alibaba.com>
Cc: Xu Yu <xuyu@linux.alibaba.com>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Mon, 16 Aug 2021 11:35:10 +0000 (20:35 +0900)]
generic: add checks for zoned block device
Modify generic tests to require non-zoned block device
generic/108 is disabled on zoned block device because the LVM device
not always aligned to the zone boundary.
generic/471 is disabled because we cannot enable NoCoW on zoned
btrfs.
generic/570 is disabled because swap file which require nocow is not
usable on zoned btrfs.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Mon, 16 Aug 2021 11:35:09 +0000 (20:35 +0900)]
btrfs: add checks for zoned block device
Modify btrfs tests to require non-zoned block device or limit some
part of tests not to be run on zone block devices.
Modified tests by the reasons:
* Mixed BG
- btrfs/011
* Non-single profile
- btrfs/003
- btrfs/011
- btrfs/023
- btrfs/124
- btrfs/195
- btrfs/197
- btrfs/198
- and these are restricted indirectly by "_require_btrfs_fs_feature raid56"
- btrfs/125
- btrfs/148
- btrfs/157
- btrfs/158
* Convert from ext4
- btrfs/012
- btrfs/136
* nodatacow
- btrfs/236
* inode cache
- btrfs/049
* space cache (v1)
- btrfs/131
* write outside of FS code
- btrfs/116
- btrfs/140
- btrfs/215
* verbose output
- btrfs/194
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Mon, 16 Aug 2021 11:35:08 +0000 (20:35 +0900)]
common: add zoned block device checks
dm-error and dm-snapshot does not have DM_TARGET_ZONED_HM nor
DM_TARGET_MIXED_ZONED_MODEL feature and does not implement
.report_zones(). So, it cannot pass the zone information from the down
layer (zoned device) to the upper layer.
Loop device also cannot pass the zone information.
This patch requires non-zoned block device for the tests using these
ones.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 28 Jul 2021 00:10:35 +0000 (17:10 -0700)]
xfs: test regression in shrink when the new EOFS splits a sparse inode cluster
This is a targeted regression test for commit
da062d16a897 ("xfs:
check for sparse inode clusters that cross new EOAG when shrinking"),
which was found by running the random-loopy shrink stresser xfs/168.
The original shrink implementation assumed that if we could allocate
the last free extent in the filesystem, it was ok to proceed with
the fs shrink. Unfortunately, this isn't quite the case -- if
there's a sparse inode cluster such that the blocks at the end of
the cluster are free, it is not ok to shrink the fs to the point
that part of the cluster hangs off the end of the filesystem. Doing
so results in repair and scrub marking the filesystem corrupt, so we
must not.
(EOFS == "end of filesystem"; EOAG == "end of allocation group")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 28 Jul 2021 00:10:24 +0000 (17:10 -0700)]
generic: test xattr operations only
Exercise extended attribute operations.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Anand Jain [Fri, 13 Aug 2021 01:59:35 +0000 (09:59 +0800)]
btrfs/220: clear_cache fix for older kernel
mount option -o clear_cache shown in /proc/self/mounts isn't
supported in the newer kernel, make this test case older kernel
compatible by checking if clear_cache is shown in the
/proc/self/mounts.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Anand Jain [Fri, 13 Aug 2021 01:59:34 +0000 (09:59 +0800)]
btrfs/220: nologreplay support older kernel
mount option -o rescue=nologreplay isn't supported on the older
kernel, make this test case older kernel compatible by checking if
the mount option is supported.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Anand Jain [Fri, 13 Aug 2021 01:59:33 +0000 (09:59 +0800)]
btrfs/220: discard=sync support older kernel
mount option -o discard=sync isn't supported on the older kernel,
make this test case older kernel compatible by checking if the mount
option is supported.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Anand Jain [Fri, 13 Aug 2021 01:59:32 +0000 (09:59 +0800)]
common/rc: debug add _scratch_mount_options to the _scratch_mount
When the _scratch_mount fails, there isn't any clue on why it failed
as of now, change this and add the $(_scratch_mount_options $*) to
the _fail.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Boris Burkov [Tue, 12 Jan 2021 21:17:47 +0000 (13:17 -0800)]
btrfs/220: fix clear_cache and inode_cache option tests
I recently changed clear_cache to not appear in mount options, as it
has one shot semantics, which was breaking this test. Test
explicitly that it _doesn't_ appear, which properly fails on old
filesystems and passes on misc-next.
The patch that changed this behavior was:
8b228324a8ce btrfs: clear free space tree on ro->rw remount
Separately, inode_cache is deprecated and will never appear in mount
options; remove it entirely.
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Wed, 11 Aug 2021 15:12:30 +0000 (00:12 +0900)]
shared/032: add check for zoned block device
Mkfs on zoned block device won't work on most filesystem. Let's
disable the test.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Wed, 11 Aug 2021 15:12:28 +0000 (00:12 +0900)]
btrfs: add minimal file system size check
Some btrfs tests call _scratch_pool_mkfs or _scratch_mkfs by themselves to
specify file system size limit. It slips through the check in
_scratch_mkfs_sized(). Let's add size check call for each of them.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Wed, 11 Aug 2021 15:12:27 +0000 (00:12 +0900)]
btrfs/057: use _scratch_mkfs_sized to set filesystem size
Btrfs/057 is using _scratch_mkfs directly to set filesystem size. This
can be _scratch_mkfs_sized instead, to go through several
checks (e.g., minimal filesystem size check).
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Wed, 11 Aug 2021 15:12:26 +0000 (00:12 +0900)]
common/rc: fix blocksize detection for btrfs
Mkfs.btrfs's "-b" does not specify the block size but the file system size.
Use "-s" to detect it instead.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Wed, 11 Aug 2021 15:12:25 +0000 (00:12 +0900)]
common/rc: introduce minimal fs size check
_scratch_mkfs_sized() create a file system with specified size
limit. It can, however, too small for certain kind of devices. For
example, zoned btrfs requires at least 5 zones to make a file system.
This commit introduces MIN_FSSIZE, which specify the minimum size of the
possible file system. We can set this variable e.g. $ZONE_SIZE *
$MIN_ZONE_COUNT.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Luis Chamberlain [Wed, 11 Aug 2021 15:45:10 +0000 (08:45 -0700)]
fstests: use udevadm settle after pvremove
As with creation, we also need to use udevadm settle
when removing a pv, otherwise we can trip on races with
module removals for the block devices in use.
This reduces the amount of time in which a block device
module refcnt for test modules such as scsi_debug spends
outside of 0.
Races with the refcnt being greater than 0 means module
removal can fail causing false positives. This helps
ensure that the pv is really long gone. These issues
are tracked for scsi_debug [0] and later found to be a
generic issue regardless of filesystem with pvremove [1].
Using udevadm settle *helps*, it does not address all
possible races with the refcnt as noted in the generic
bug entry [1].
[0] https://bugzilla.kernel.org/show_bug.cgi?id=212337
[1] https://bugzilla.kernel.org/show_bug.cgi?id=214015
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Marcos Paulo de Souza [Thu, 12 Aug 2021 11:30:48 +0000 (08:30 -0300)]
btrfs/177: Handle the different fi resize output formats
Commit
78aa1d95dd99 ("btrfs-progs: fi resize: make output more
readable") added the device id of the resized fs along with a pretty
printed size. Create a new function to simplify the output message
using size in bytes.
Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Tue, 20 Jul 2021 01:09:13 +0000 (18:09 -0700)]
dmerror: export configuration so that subprograms don't have to reinit
Export the dmerror configuration variables so that subprograms don't
have to reinitialize the configuration in their own subprograms before
calling the helpers. In the next patch (where we allow dmerror for log
and rt devices) it will become important to avoid these
reinitializations.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Tue, 20 Jul 2021 01:09:08 +0000 (18:09 -0700)]
dmflakey: support external log and realtime devices
Upgrade the dmflakey code to coordinate making external scratch log and
scratch realtime devices flakey along with the scratch device. Note
that unlike SCRATCH_DEV, we save the old rt/log devices as separate
variables and replace SCRATCH_{RT,LOG}DEV so that helper functions
continue to work without modification.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Amir Goldstein [Tue, 3 Aug 2021 07:07:58 +0000 (10:07 +0300)]
overlay/078: Fix cleanup with unpatched kernel
Darrick wrote: "...
I noticed a massive regression with this week's fstests,
probably because something didn't get cleaned up properly:
...
+rm: cannot remove '/opt/ovl-upper/testfile': Operation not permitted
...
and then the tests after it (e.g. generic/030) fail with:
+mount: /opt/ovl-mnt: mount(2) system call failed: Stale file handle.
..."
Link: https://lore.kernel.org/fstests/20210802230727.GC3601425@magnolia/
Reported-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Zorro Lang [Fri, 6 Aug 2021 05:10:12 +0000 (13:10 +0800)]
xfs: quota state command should report upg grace times separately
Regression test for xfsprogs
d8a94546 ("xfs_quota: state command
should report ugp grace times"). When give "-ugp" or "-a" options to
xfs_quota state command, it should report grace times for all three
types separately.
Signed-off-by: Zorro Lang <zlang@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Qu Wenruo [Fri, 6 Aug 2021 11:33:33 +0000 (19:33 +0800)]
btrfs: verify the behavior of deleting non-existing device
There is a kernel regression for btrfs, that when passing
non-existing devid to "btrfs device remove" command, kernel will
crash due to NULL pointer dereference.
The test case is for such regression, it will:
- Create and mount an empty single-device btrfs
- Try to remove devid 3, which doesn't exist for above fs
- Make sure the command exits properly with expected error message
The kernel fix is titled "btrfs: fix NULL pointer dereference when
deleting device by invalid id".
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Jeffle Xu [Thu, 29 Jul 2021 04:47:58 +0000 (12:47 +0800)]
common/rc: only force nfs4.2 non-default SEEK_HOLE behaviour
Only NFSv4.2 supports non-defautl SEEK_HOLE behaviour. Thus default
SEEK_HOLE behaviour shall be allowed for NFSv4.0/4.1, or it will fail
generic/285, generic/448, generic/490 on NFSv4.0/4.1, complaining they
should support non-default SEEK_HOLE behaviour.
The *.full log is like:
File system supports the default behavior.
Default behavior is not allowed. Aborting.
Signed-off-by: Jeffle Xu <jefflexu@linux.alibaba.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Ritesh Harjani [Wed, 21 Jul 2021 05:28:02 +0000 (10:58 +0530)]
common/attr: Reduce MAX_ATTRS to leave some overhead for 64K blocksize
Test generic/020 fails for ext4 with 64K blocksize.
This adds changes in common/attr for MAX_ATTRS calculations for
ext2|ext3|ext4 along with comments explaining the calculations.
Suggested-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Ritesh Harjani [Wed, 21 Jul 2021 05:28:01 +0000 (10:58 +0530)]
common/attr: Cleanup end of line whitespaces issues
This patch clears the end of line whitespace issues in this file.
Mostly since many kernel developers also keep this editor config to clear
any end of line whitespaces on file save.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Ritesh Harjani [Wed, 21 Jul 2021 05:27:57 +0000 (10:57 +0530)]
ext4/022: exclude this test for dax config on 64KB pagesize platform
This test case assumes blocksize to be 4KB and hence it fails
to mount with "-o dax" option on a 64kb pagesize platform (e.g. PPC64).
This leads to test case reported as failed with dax config on PPC64.
This patch exclude this test when pagesize is 64KB and for dax config.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Ritesh Harjani [Wed, 21 Jul 2021 05:27:56 +0000 (10:57 +0530)]
ext4/306: Add -b blocksize parameter too to avoid failure with DAX config
mkfs.ext4 by default uses 4K blocksize. On DAX config with a 64K
pagesize platform (PPC64), this will fail to mount since DAX requires bs
== ps.
Hence add the -b blocksize paramter in ext4/306.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Ritesh Harjani [Wed, 21 Jul 2021 05:27:55 +0000 (10:57 +0530)]
ext4/027: Correct the right code of block and inode bitmap
Observed occasional failure of this test sometimes say with 64k config
and small device size. Reason is we were grepping for wrong values for
inode and block bitmap.
Correct those values according to [1] to fix this test.
[1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/fs/ext4/fsmap.h#n53
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Ritesh Harjani [Wed, 21 Jul 2021 05:27:54 +0000 (10:57 +0530)]
ext4/003: Fix this test on 64K platform for dax config
mkfs.ext4 by default uses 4K blocksize which doesn't mount when
testing with dax config and the test fails. This patch fixes it.
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Luis Chamberlain [Tue, 27 Jul 2021 20:10:43 +0000 (13:10 -0700)]
common/scsi_debug: use udevadm settle instead of sleeping
The variable UDEV_SETTLE_PROG is already defined and used for
lvm to either use `udevadm settle` in case it is available, and
if not, use 'sleep 1' otherwise (ancient distros or CONFIG_NET
is missing).
Use it on scsi_debug to replace the sleep calls sprinkled in
place after module removal. The correct thing to do is to
just use udevadm settle when available, and only fall back to
calling sleep when udevadm is not available or CONFIG_NET
is disabled.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Luis Chamberlain [Tue, 27 Jul 2021 20:10:42 +0000 (13:10 -0700)]
common/config: disable udevadm settle if CONFIG_NET is disabled
If CONFIG_NET is disabled kobject_uevent_net_broadcast() will be a no-op
and so no uevent are sent and so 'udevadm settle' won't really do
anything for you.
We check for /proc/net to see if CONFIG_NET was enabled.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 28 Jul 2021 20:00:45 +0000 (13:00 -0700)]
xfs/007: fix regressions on V4 filesystems
Following commit
eae40404, I noticed the following regression when
running a V4 fstests run on an 5.13 kernel:
--- /tmp/fstests/tests/xfs/007.out 2021-05-13 11:47:55.
793859995 -0700
+++ /var/tmp/fstests/xfs/007.out.bad 2021-07-28 09:23:42.
856000000 -0700
@@ -16,4 +16,4 @@
*** umount
*** Usage after quotarm ***
core.nblocks = 0
-core.nblocks = 0
+core.nblocks = 1
The underlying cause of this problem is the fact that we now remount the
filesystem with no quota options because that will (soon) become the
only means to turn off quota accounting on XFS. Because V4 filesystems
don't support simultaneous project and group quotas and play weird
remapping games with the incore superblock field, we actually have to
issue a remove command for the group quota file if we're trying to
truncate the project quota file on a V4 filesystem.
Due to stupid limitations in xfs_quota we actually have to issue a
separate 'remove' command.
Fixes: eae40404 ("xfs/007: unmount after disabling quota")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 28 Jul 2021 20:00:20 +0000 (13:00 -0700)]
xfs/530: skip test if user MKFS_OPTIONS screw up formatting
Skip this test if the user's MKFS_OPTIONS are not compatible with the
realtime parameters that we're injecting in order to test growfs bugs.
Because this test is trying to trigger a specific kernel bug, we need
mkfs to format a filesystem with very specific geometry parameters.
The first problem stems from the fact that the test performs a default
mkfs, computes a suitable realtime geometry from that filesystem, and
then formats a second time with an explicit blocksize option to mkfs.
If the original MKFS_OPTS contained a blocksize directive, the mkfs will
fail because the option was respecified. The two blocksize options will
be the same, so we drop the explicit blocksize option.
However, this exposes a second problem: MKFS_OPTIONS might contain
options that are not compatible with any realtime filesystem. If that
happens, _scratch_do_mkfs will "helpfully" drop MKFS_OPTIONS and try
again with only the options specified by the test. This gets us a
filesystem with the given rt geometry, but it could be missing critical
parameters from MKFS_OPTIONS (like blocksize). The test will then fail
to exercise the growfs bugfix, so the second part of the fix is to check
that the filesystem we're going to test actually has the geometry
parameters that we require.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 28 Jul 2021 00:09:59 +0000 (17:09 -0700)]
check: back off the OOM score adjustment to -500
Dave Chinner complained that fstests really shouldn't be running at
-1000 oom score adjustment because that makes it more "important" than
certain system daemons (e.g. journald, udev). That's true, so increase
it to -500.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 28 Jul 2021 00:09:48 +0000 (17:09 -0700)]
check: don't leave the scratch filesystem mounted after _notrun
Unmount the scratch filesystem if a test decides to _notrun itself
because _try_wipe_scratch_devs will not be able to wipe the scratch
device prior to the next test run. We don't want to let scratch state
from one test leak into subsequent tests if we can help it.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 28 Jul 2021 00:09:43 +0000 (17:09 -0700)]
xfs/106: fix golden output regression in quota off test
In commit
4c5df338, we reshuffled this test a bit in preparation to
disable quotaoff by rearranging the test to testing disabling of quota
by remounting the filesystem. Unfortunately, extra blank lines were
added to the golden output, leading to test regressions.
The "extra" blank lines are a result of the "echo ; test_off";
test_off() itself doesn't print anything. Make it print /something/ so
that we know what the test was trying to do when a particular line of
golden output appears, then fix the blank lines.
Fixes: 4c5df338 ("xfs/106: don't test disabling quota accounting")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Thu, 29 Jul 2021 16:53:05 +0000 (09:53 -0700)]
misc: tag all tests that examine crash recovery in a loop
Given all the recent problems that we've been finding with log recovery,
I think it would be useful to create a 'recoveryloop' group so that
developers have a convenient way to run every single test that rolls
around in a fs shutdown loop looking for subtle errors in recovery.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Zorro Lang [Mon, 26 Jul 2021 09:46:05 +0000 (17:46 +0800)]
generic: test small swapfile without page-aligned contiguous blocks
If a swapfile doesn't contain even a single page-aligned contiguous
range of blocks, it's an invalid swapfile, and might cause kernel
issue. This case covered commit
5808fecc5723 ("iomap: Fix negative
assignment to unsigned sis->pages in iomap_swapfile_activate").
Signed-off-by: Zorro Lang <zlang@redhat.com>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Filipe Manana [Tue, 27 Jul 2021 10:24:59 +0000 (11:24 +0100)]
generic: test for file loss after mix of rename, fsync and inode eviction
Test that if we fsync a directory A, evict A's inode, move one file
from directory A to directory B, fsync some other inode that is not
directory A, B or any inode inside these two directories, and then
power fail, the file that was moved is not lost.
This currently fails on btrfs and is fixed by commit
ecc64fab7d49
("btrfs: fix lost inode on log replay after mix of fsync, rename and
inode eviction")
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Dave Chinner [Wed, 21 Jul 2021 00:13:33 +0000 (10:13 +1000)]
tools: add a dm-logwrites replay tool
Trying to decipher a dm-logwrites replay failure from generic/482 is
difficult. The test destroys all the dm state and devices when the
test fails, leaving nothing behind for post-mortem analysis. This
script is intended to allow replay of the dm-logwrites device one
FUA at a time similar to generic/482, and is used in conjunction
with a modifed g/482 test that does not tear down the dm volumes
or state when the test fails.
This allows the developer to replay the logwrites up to just prior
to failure to examine just the differences between the last good
state and the first failure state, greatly reducing the complexity
of analysing failures.
Instructions for use are in the tools/dm-logwrite-replay script
itself.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Amir Goldstein [Thu, 22 Jul 2021 16:46:34 +0000 (19:46 +0300)]
overlay: add test for copy up of lower file attributes
Overlayfs copies up a subset of lower file attributes since kernel
commits:
173ff5c9ec37 ("ovl: consistent behavior for immutable/append-only inodes")
2e3f6e87c2b0 ("ovl: copy up sync/noatime fileattr flags")
This test verifies this functionality works correctly and that it
survives power failure and/or mount cycle.
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 [Thu, 22 Jul 2021 17:02:16 +0000 (20:02 +0300)]
overlay/077: check for inconsistent d_ino/st_ino
After unlink of a directory entry, that entry may still apear in getdents
results of an already open directory fd, but it should return a d_ino
value that is consistent with the already observed st_ino of that entry.
Remove redundant break condition from gendents read loop.
For testing of inconsistent d_ino/st_ino we need to unlink an entry
whose st_ino is not that of the upper inode.
In the former merge dir setup we unlink all the files in the lower
dir after copyup, so they all use st_ino of the upper inode.
Let the unlinked file f100 reside in a lower path that is not being
unlinked so it will have the st_ino of the lower inode.
This is a regression test for kernel commit
fcb7f373684d
("ovl: skip stale entries in merge dir cache iteration")
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Fri, 23 Jul 2021 05:08:59 +0000 (22:08 -0700)]
xfs/152: avoid failure when quotaoff is not supported
Switch the test that removes the quota files to just disable enforcement
and then unmount the file system as disabling quota accounting is about
to go away.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Christoph Hellwig [Thu, 22 Jul 2021 07:38:32 +0000 (09:38 +0200)]
xfs: use $XFS_QUOTA_PROG instead of hardcoding xfs_quota
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Christoph Hellwig [Thu, 22 Jul 2021 07:38:31 +0000 (09:38 +0200)]
xfs/305: don't turn quota accounting off
The test case tests just as much when just testing turning quota
enforcement off, so switch it to that. This is in preparation for
removing support to turn quota accounting off.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Christoph Hellwig [Thu, 22 Jul 2021 07:38:30 +0000 (09:38 +0200)]
xfs/304: don't turn quota accounting off
The test case tests just as much when just testing turning quota
enforcement off, so switch it to that. This is in preparation for
removing support to turn quota accounting off.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Christoph Hellwig [Thu, 22 Jul 2021 07:38:29 +0000 (09:38 +0200)]
xfs/220: avoid failure when disabling quota accounting is not supported
Doing a proper _requires for quotaoff support is rather hard, as we need
to test it on a specific file system. Instead just use sed to remove
the warning and let the test case pass. Eventually it should just be
removed entirely.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Christoph Hellwig [Thu, 22 Jul 2021 07:38:28 +0000 (09:38 +0200)]
xfs/106: don't test disabling quota accounting
Switch the test that removes the quota files to just disable
enforcement and then unmount the file system as disabling quota
accounting is about to go away.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Christoph Hellwig [Thu, 22 Jul 2021 07:38:27 +0000 (09:38 +0200)]
xfs/007: unmount after disabling quota
With the pending patches to remove support for disabling quota
accounting on a mounted file system we need to unmount the
file system first before removing the quota files.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Christoph Hellwig [Thu, 22 Jul 2021 07:38:26 +0000 (09:38 +0200)]
common/quota: allow removing quota options entirely in _qmount_option
Add support for dropping all quota related options instead of only
overriding them with new ones to _qmount_option.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Marcos Paulo de Souza [Fri, 23 Jul 2021 18:18:00 +0000 (15:18 -0300)]
btrfs: source common/module when requiring module reload
Tests 163, 219, 225 and 242 require the btrfs module to be reloaded
in the middle of the test. If btrfs is compiled builtin, these tests
should be skipped, since we cannot unload it.
Today, this is what happens:
QA output created by 242
+./common/btrfs: line 405: _require_loadable_fs_module: command not found
Sourcing common/module in these tests fixes the issue by skipping the
test:
btrfs/242 [not run] btrfs: must be a module.
Ran: btrfs/242
Not run: btrfs/242
Other tests liker btrfs/124 and btrfs/125 already source the same file
for the same reason, so follow the pattern.
Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Eric Biggers [Sun, 18 Jul 2021 19:06:58 +0000 (14:06 -0500)]
common/encrypt: accept '-' character in no-key names
Add the '-' character to the regex that generic/{419,429} use to match
no-key filenames. This is needed to prevent these tests from failing
after the kernel is changed to use a more standard variant of Base64
(https://lkml.kernel.org/r/
20210718000125.59701-1-ebiggers@kernel.org).
Note that despite breaking these tests, the kernel change is not
expected to break any real users, as the fscrypt no-key name encoding
has always been considered an implementation detail. So it is
appropriate to just update these tests.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Eric Biggers [Sun, 18 Jul 2021 19:06:57 +0000 (14:06 -0500)]
common/encrypt: add helper function for filtering no-key names
Add a helper function _filter_nokey_filenames() which replaces no-key
filenames with "NOKEY_NAME". Use it in generic/419 and generic/429.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Eric Biggers [Sun, 18 Jul 2021 19:06:56 +0000 (14:06 -0500)]
generic: update encryption tests to use term "no-key names"
To avoid ambiguity, don't use the terms "ciphertext names" or "encrypted
names" when we really mean "no-key names" (the names that the filesystem
shows when userspace lists an encrypted directory without the
directory's encryption key being present). This aligns with changes
that have been made to the kernel source code and documentation
(https://lore.kernel.org/r/
20200924042624.98439-1-ebiggers@kernel.org).
No change to the actual test logic.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Eric Biggers [Sun, 18 Jul 2021 15:17:46 +0000 (10:17 -0500)]
generic/423: test stx_size of symlink
Update generic/423 to test that the reported size of a symlink is equal
to the length of the symlink target, as required by POSIX and as
documented in various man pages.
When the test_dummy_encryption mount option is enabled, this serves as a
regression test for the bug that will be fixed by the patch series
"fscrypt: report correct st_size for encrypted symlinks"
(https://lkml.kernel.org/r/
20210702065350.209646-1-ebiggers@kernel.org).
However, this test is applicable to all filesystems and mount options.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Tue, 20 Jul 2021 01:08:29 +0000 (18:08 -0700)]
generic/561: hide assertions when duperemove is killed
Use some bash redirection trickery to capture in $seqres.full all of
bash's warnings about duperemove being killed due to assertions
triggering.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Tue, 20 Jul 2021 01:08:22 +0000 (18:08 -0700)]
new: allow users to specify a new test id
Alter the ./new script so that one can set the test id explicitly.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 7 Jul 2021 00:21:50 +0000 (17:21 -0700)]
generic/019: don't dump cores when fio/fsstress hit io errors
Disable coredumps so that fstests won't mark the test failed when the
EIO injector causes an mmap write to abort with SIGBUS.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 7 Jul 2021 00:21:45 +0000 (17:21 -0700)]
generic/371: disable speculative preallocation regressions on XFS
Once in a very long while, the fallocate calls in this test will fail
due to ENOSPC conditions. While in theory this test is careful only to
allocate at most 160M of space from a 256M filesystem, there's a twist
on XFS: speculative preallocation.
The first loop in this test is an 80M appending write done in units of
4k. Once the file size hits 64k, XFS will begin speculatively
preallocating blocks past the end of the file; as the file grows larger,
so will the speculative preallocation.
Since the pwrite/rm loop races with the fallocate/rm loop, it's possible
that the fallocate loop will free that file just before the buffered
write extends the speculative preallocation out to 160MB. With fs and
log overhead, that doesn't leave enough free space to start the 80MB
fallocate request, which tries to avoid disappointing the caller by
freeing all speculative preallocations. That fails if the pwriter
thread owns the IOLOCK on $testfile1, so fallocate returns ENOSPC and
the test fails.
The simple solution here is to disable speculative preallocation by
setting an extent size hint if the fs is XFS.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 7 Jul 2021 00:21:39 +0000 (17:21 -0700)]
xfs/084: fix test program status collection and processing
On a test VM with 1.2GB memory, I noticed that the test will
sometimes fail because resvtest leaks too much memory and gets OOM
killed. It would be useful to _notrun the test when this happens so
that it doesn't appear as an intermittent regression.
The exit code processing in this test is incorrect, since "$?" will
get us the exit status of _filter_resv, not $here/src/resvtest. Fix
that as part of learning to detect a SIGKILL and skip the test.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 7 Jul 2021 00:21:34 +0000 (17:21 -0700)]
check: run _check_filesystems in an OOM-happy subshell
While running fstests one night, I observed that fstests stopped
abruptly because ./check ran _check_filesystems to run xfs_repair.
In turn, repair (which inherited oom_score_adj=-1000 from ./check)
consumed so much memory that the OOM killer ran around killing other
daemons, rendering the system nonfunctional.
This is silly -- we set an OOM score adjustment of -1000 on the
./check process so that the test framework itself wouldn't get
OOM-killed, because that aborts the entire run. Everything else is
fair game for that, including subprocesses started by
_check_filesystems.
Therefore, adapt _check_filesystems (and its children) to run in a
subshell with a much higher oom score adjustment.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 7 Jul 2021 00:21:28 +0000 (17:21 -0700)]
dmthin: erase the metadata device properly before starting
Every now and then I see the following failure when running generic/347:
--- generic/347.out
+++ generic/347.out.bad
@@ -1,2 +1,2 @@
QA output created by 347
-=== completed
+failed to create dm thin pool device
Accompanied by the following dmesg spew:
device-mapper: thin metadata: sb_check failed: blocknr
7016996765293437281: wanted 0
device-mapper: block manager: superblock validator check failed for block 0
device-mapper: thin metadata: couldn't read superblock
device-mapper: table: 253:2: thin-pool: Error creating metadata object
device-mapper: ioctl: error adding target to table
7016996765293437281 is of course the magic number 0x6161616161616161,
which are stale ondisk contents left behind by previous tests that wrote
known tests patterns to files on the scratch device. This is a bit
surprising, since _dmthin_init supposedly zeroes the first 4k of the
thin pool metadata device before initializing the pool. Or does it?
dd if=/dev/zero of=$DMTHIN_META_DEV bs=4096 count=1 &>/dev/null
Herein lies the problem: the dd process writes zeroes into the page
cache and exits. Normally the block layer will flush the page cache
after the last file descriptor is closed, but once in a while the
terminating dd process won't be the only process in the system with an
open file descriptor!
That process is of course udev. The write() call from dd triggers a
kernel uevent, which starts udev. If udev is running particularly
slowly, it'll still be running an instant later when dd terminates,
thereby preventing the page cache flush. If udev is still running a
moment later when we call dmsetup to set up the thin pool, the pool
creation will issue a bio to read the ondisk superblock. This read
isn't coherent with the page cache, so it sees old disk contents and the
test fails even though we supposedly formatted the metadata device.
Fix this by explicitly flushing the page cache after writing the zeroes.
Fixes: 4b52fffb ("dm-thinp helpers in common/dmthin")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 7 Jul 2021 00:21:23 +0000 (17:21 -0700)]
shared/298: fix random deletion when filenames contain spaces
Correct the deletion loop in this test to work properly when there are
files in $here that have spaces in their name.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Wed, 7 Jul 2021 00:21:12 +0000 (17:21 -0700)]
xfs/172: disable test when file writes don't use delayed allocation
This test tries to exploit an interaction between delayed allocation and
writeback on full filesystems to see if it can trip up the filestreams
allocator. The behaviors do not present if the filesystem allocates
space at write time, so disable it under these scenarios.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Matthew Wilcox [Tue, 6 Jul 2021 20:14:46 +0000 (21:14 +0100)]
src/seek_sanity_test: test seeks from LLONG_MIN
We don't currently test seeking from LLONG_MIN, which apparently can
produce an UBSAN warning, although I've been unable to reproduce
that. In any case, it's a good corner case to test and
straightforward to add.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Dave Chinner [Fri, 9 Jul 2021 06:10:27 +0000 (16:10 +1000)]
generic/459: catch mount failure
Because if we don't, we freeze the root filesystem instead of
$SCRATCH_MNT and it all goes downhill from there.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Pavel Reichl [Thu, 8 Jul 2021 16:35:18 +0000 (18:35 +0200)]
generic/{453,454}: Don't run for FSs restricting names
Running generic/{453,454} tests is probably only useful for filesystems
that allow unrestricted byte streams for names.
Signed-off-by: Pavel Reichl <preichl@redhat.com>
Suggested-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Naohiro Aota [Wed, 7 Jul 2021 05:01:41 +0000 (14:01 +0900)]
fstests: install btrfs_crc32c_forged_name.py
btrfs/154 uses btrfs_crc32c_forged_name.py but the file is not
installed in the PKG_LIB_DIR. Also consolidate install target files
into EXTRA_EXECS.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Filipe Manana [Tue, 6 Jul 2021 14:42:17 +0000 (15:42 +0100)]
btrfs: test fsync after increasing file size with truncate
Test that if we explicitly fsync a file that was previously renamed and
its size was increased through a truncate operation, after a power failure
the file has the size set by the truncate operation. Also, in between the
truncation and the fsync, there was a rename of another file in the same
directory and that file was also fsynced before we fsynced the file that
was truncated.
This currently fails on a 5.13 kernel and on Linus' master branch. It is
fixed by a patch with the following subject:
"btrfs: fix unpersisted i_size on fsync after expanding truncate"
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Anand Jain [Mon, 5 Jul 2021 10:43:31 +0000 (18:43 +0800)]
btrfs: test case to fstrim on a degraded filesystem
Create a degraded btrfs filesystem and run fstrim on it.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Theodore Ts'o [Mon, 21 Jun 2021 16:48:51 +0000 (12:48 -0400)]
ext4/048: skip test of filename wipe if journal checkpoint is not supported
ext4/048 will fail when running on older kernels that don't support
the filename wipe feature. The journal checkpoint ioctl is a related
feature, and landed just a little bit after filename wipe feature, so
use support for the journal checkpoint ioctl as a proxy for support
for the filename wipe feature.
Without this change, this test will fail when tesing 5.10, 5.4, and
other LTS kernels.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: Leah Rumancik <leah.rumancik@gmail.com>
Tested-by: Yang Xu <xuyang2018.jy@fujitsu.com>
Reviewed-by: Leah Rumancik <leah.rumancik@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Yang Xu [Wed, 30 Jun 2021 01:52:06 +0000 (09:52 +0800)]
overlay/Makefile: Use the unified name rules
Since xfs/ext4 uses XFS_DIR/EXT4_DIR in Makefile(also for other filesystems),
I think we don't have any reason not to use OVERLAY_DIR in here.
Signed-off-by: Yang Xu <xuyang2018.jy@fujitsu.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Yang Xu [Wed, 30 Jun 2021 01:52:05 +0000 (09:52 +0800)]
Makefile: fix bug when running make install under tests/* directory
The TESTS_DIR value is defined in TOPDIR makefile, it is empty here.
So running make install on ext4 directory will get the following info:
../../install-sh -o root -g root -m 755 -d /var/lib/xfstests//ext4
We can see they aren't installed under /var/lib/xfstests/tests/ext4 directory.
Fix this by moving TESTS_DIR definition to include/builddefs.in file.
Signed-off-by: Yang Xu <xuyang2018.jy@fujitsu.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Sat, 3 Jul 2021 03:05:48 +0000 (20:05 -0700)]
tools: make convert-group ignore converted test files
Teach the tool that rewrites tests files to use _begin_fstest group
tagging to ignore tests that have already been treated. This will make
it easier for people to rebase their dev branches.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Jeff Layton [Fri, 2 Jul 2021 13:40:24 +0000 (09:40 -0400)]
generic: ensure that page is properly filled before write
We had a broken optimization in cephfs and netfs lib that could
cause part of a page to be improperly zeroed-out when writing to an
offset that was beyond the EOF but in an existing page.
Add a simple test that would have caught this.
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Anju T Sudhakar [Mon, 28 Jun 2021 08:52:57 +0000 (14:22 +0530)]
xfs/504: Add scratch_mount before checking for xfs_scrub unicode support
We may not detect the error `Inappropriate ioctl for device`, while running
`$XFS_IO_PROG -c "scrub probe" "$mountpoint"`, if scratch device is not
mounted before invoking _check_xfs_scrub_does_unicode(). So do
_scratch_mount before checking for xfs_scrub support.
Signed-off-by: Anju T Sudhakar <anju@linux.vnet.ibm.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Theodore Ts'o [Sat, 3 Jul 2021 16:02:47 +0000 (12:02 -0400)]
check: exit with exit code 1 after printing the usage message
If check is passed an invalid command line option, exit with a
non-zero exit code so that a script calling check can detect the
failure. The check script already performs an "exit 1" if a valid
option has an invalid argument, so this is consistent with existing
practice.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Sun Ke [Wed, 30 Jun 2021 11:37:36 +0000 (19:37 +0800)]
generic/260: f2fs is also special
It fail on f2fs:
[+] Default length with start set (should succeed)
[+] Length beyond the end of fs (should succeed)
[+] Length beyond the end of fs with start set (should succeed)
+After the full fs discard 0 bytes were discarded however the file system is
12882804736 bytes long.
+It seems that fs logic handling len argument overflows
The root cause is f2fs can tag a special flag TRIMMED_FLAG to
indicate the whole filesystem is trimmed, so after mkfs/fstrim(),
following fstrim() won't trim any block.
Suggested-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Sun Ke <sunke32@huawei.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Sun Ke [Wed, 30 Jun 2021 11:37:35 +0000 (19:37 +0800)]
generic/103: special left calculation for f2fs
It failed on f2fs:
QA output created by 103
+fallocate: No space left on device
Silence is golden.
...
f2fs uses index(radix) tree as mapping metadata, its space overhead
is about one thousandth of the data.
Suggested-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Sun Ke <sunke32@huawei.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Sun Ke [Wed, 30 Jun 2021 11:37:34 +0000 (19:37 +0800)]
generic/042: make a bigger img for f2fs
f2fs-utils 1.9.0 needs at least 38 MB space for f2fs image. However,
f2fs-utils 1.14.0 needs at least 52 MB. Not sure if it will change
again. So, just set it to 128M.
Suggested-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Sun Ke <sunke32@huawei.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Dave Chinner [Tue, 29 Jun 2021 00:55:00 +0000 (10:55 +1000)]
generic/019: fix test cleanup
This never worked properly and the failure was exposed by the recent
rework of the fstest setup infrastructure.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 16:12:09 +0000 (09:12 -0700)]
misc: update documentation to reflect auto-generated group files
Update the documentation to outline the new requirements for test files
so that we can generate group files during build.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 16:12:09 +0000 (09:12 -0700)]
fstests: remove test group management code
Remove all the code that manages group files, since we now generate
them at build time.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 16:12:00 +0000 (09:12 -0700)]
fstests: remove group files
From: Darrick J. Wong <djwong@kernel.org>
Now that we autogenerate group files, get rid of them in the source
tree.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:56:23 +0000 (08:56 -0700)]
check: use generated group files
Convert the ./check script to use the automatically generated group list
membership files, as the transition is now complete.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:56:00 +0000 (08:56 -0700)]
fstests: adapt the new test script to our new group tagging scheme
Now that we autogenerate group files, adapt the new test creation script
to use autogenerated group files and to set the group data in the new
test.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:55:39 +0000 (08:55 -0700)]
fstests: convert nextid to use automatic group generation
Convert the nextid script to use the automatic group file generation to
figure out the next available test id.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:55:09 +0000 (08:55 -0700)]
fstests: automatically generate group files
Now that we've moved the group membership details into the test case
files themselves, automatically generate the group files during build.
The autogenerated files are named "group.list" instead of "group" to
avoid conflicts between generated and (stale) SCM files as everyone
rebases.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:55:09 +0000 (08:55 -0700)]
fstests: move test group info to test files
Refactor every test in the entire test suite to use the new boilerplate
functions. This also migrates all the test group information into the
test files. This patch has been autogenerated via the command:
./tools/convert-group btrfs ceph cifs ext4 f2fs generic nfs ocfs2 overlay perf shared udf xfs
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:54:54 +0000 (08:54 -0700)]
fstests: clean up open-coded golden output
Fix the handful of tests that open-coded 'QA output created by XXX'.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:54:54 +0000 (08:54 -0700)]
fstests: add tool migrate group membership data to test files
Create a tool to migrate the mapping of tests <-> groups out of the
group file and into the individual test file as a _begin_fstest
call. In the next patches we'll rewrite all the test files and auto
generate the group files from the tests.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:54:54 +0000 (08:54 -0700)]
fstests: refactor test boilerplate code
Create two new helper functions to deal with boilerplate test code:
A helper function to set the seq and seqnum variables. We will expand
on this in the next patch so that fstests can autogenerate group files
from now on.
A helper function to register cleanup code that will run if the test
exits or trips over a standard range of signals.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Darrick J. Wong [Mon, 21 Jun 2021 15:54:53 +0000 (08:54 -0700)]
misc: move exit status into trap handler
Move the "exit $status" clause of the _cleanup function into the
argument to the "trap" command so that we can standardize the
registration of the atexit cleanup code in the next few patches.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>