Calling _mount doesn't work when we want to add mount options such
as realtime devices. Fix it by calling _scratch_options first then
mount with $SCRATCH_OPTIONS.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
This test uses the scratch device, so cycle that, not the test dev.
This is also a xfs_fsr test, so put it in the fsr group.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Run xfs_repair twice more at the end of each test -- once to rebuild
the btree indices, and again with -n to check the rebuild work.
This is in addition to the regular dry-run spot check.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Use the helper functions for scratch devices. This fixes a problem
where xfs/234 fails when there's a realtime device.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Use the helper functions for scratch devices. This fixes a problem
where xfs/179 fails when there's a realtime device.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Use the standard _scratch_mount to mount the filesystem from the
restored image, instead of trying to call mount directly. This is
needed in case we had custom mount options (like rtdev).
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
In _require_scratch_dedupe, test the scratch device, not the
testdev.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
List the new log redo items. These should have stable sizes.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Fix the calculation of the dmhuge size. The previous calculation
tried to calculate the size correctly, but got it wrong for 1k block
sizes. Therefore, clean the whole mess up.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Don't open-code _notrun checks for the rmapbt, just use the helper.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
These two tests were accidentally double-added as xfs/30[78], but
the newer versions have fixed up helper usage and fewer whitespace
problems, so nuke the old tests.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Eryu Guan <eguan@redhat.com>
The recent direct I/O path refactor changed the function name in the
racing read case. Update the test case to handle the new name in
addition to the old one.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Sometimes it's useful to keep track of why we're excluding a particular
test. Technically, we can include whatever explanation we want in the
exclude file since we just grep for the test names, but properly
supporting comments is a little more robust. This patch makes it so that
the rest of a line after a '#' is ignored.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Test truncate running executable binaries from lower and upper dirs.
truncate(2) should return ETXTBSY, not other errno nor segfault
Commit 03bea6040932 ("ovl: get_write_access() in truncate") fixed
this issue.
Reviewed-by: Xiong Zhou <xzhou@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
The upper dentry may become stale before we remove it from overlay,
e.g. file was removed from upperdir directly, which triggered
warning/oops in ovl_remove_and_whiteout().
Commit cfc9fde0b07c ("ovl: verify upper dentry in
ovl_remove_and_whiteout()") fixed this issue.
Reviewed-by: Xiong Zhou <xzhou@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Test dump overlay private xattr on overlayfs with multiple lower
dirs. All the lower layers, except the lowest one, should honor
overlay private xattr "trusted.overlay", which means it should not
be visible to user.
Commit b581755b1c56 ("ovl: xattr filter fix") fixed this issue.
Reviewed-by: Xiong Zhou <xzhou@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Currently in btrfs, there is something wrong with fallocate(2)'s
data space reservation, it'll temporarily occupy more data space
thant it really needs, which in turn will impact other operations'
data request.
In this test case, it runs write(2) and fallocate(2) in parallel and
the total needed data space for these two operations don't exceed
whole fs free data space, to see whether we will get any unexpected
ENOSPC error.
Signed-off-by: Wang Xiaoguang <wangxg.fnst@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
xfs/106 is testing basic functions of xfs_quota command, but
there're three problems prevent people from running it as a reliable
regression test (i.e. it's not in 'auto' group):
- It tests unavailable commands in current xfs_quota, e.g. "warn"
- Lack of test coverage of important commands, e.g. "disable,
enable, off and remove" commands
- Test output doesn't match golden image
So rewrite this case to test as many xfs_quota sub-commands and
options as possible and add it to 'auto' group, except:
- some "-v" options
- all "-a" options, because it's hard to get deterministic output
- default quota, this should be already covered by other tests
- project command options, I will write another case to test it
- report command -l option, which is a new option and will cause
failure when testing with xfsprogs prior to 3.1 version.
- quot command -c option, which may output different histogram in
different systems or hosts.
- warn command, which is not currently implemented.
- print, df, free, help, quit and path commands.
Signed-off-by: Zorro Lang <zlang@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Different quota format may print additional information in
repquota(8) output after standard quota information is printed. If
format does support this additional printouts, repquota(8) will
separate possible output by two empty lines even if format handler
doesn't actually printout anything (which is currently always the
case for queries test generic/235 does). If format doesn't support
additional printouts, these two empty lines are not present in the
output. This inconsistency causes false failures for some quota
formats.
Fix the problem by filtering out empty lines out of repquota(8)
output.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Update the following quick/auto tag based on their execution time
btrfs/007
btrfs/050
btrfs/100
btrfs/101
Two systems are used to determine their execution time. One is
backed by an SATA spinning rust, whose maximum R/W speed is about
100MB/s, modern desktop performance. (VM1)
Another one is a VM inside a openstack pool, with stronger CPU and
memory performance along with high latency storage. Maximum R/W
speed is around 150MB/s, latency is much higher than normal HDD
though. (VM2)
The 'quick' standard is a little more restrict, only when both
systems pass the test within 30s(+/- 10%), while 'auto' is less
restrict, any system can pass within 5min(+/- 10%) will still stay
in 'auto' group.
Other test cases don't fit both standards on both systems will not
be modified.
Execution time result: (Unit: seconds)
------------------------------------------------------
Test case No. | VM1 | VM2 | Modification |
------------------------------------------------------
btrfs/007 | 4 | 2 | +quick |
btrfs/050 | 4 | 13 | +quick |
btrfs/100 | 57 | 151 | -quick |
btrfs/101 | 45 | 59 | -quick |
------------------------------------------------------
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
On my test setup xfs_io reports 'nan' in bytes/s and ops/s fields
when the operation takes zero time. Account for that in
_filter_xfs_io.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Really old versions of coreutils (mine are 8.12) quote a filename in the
output with a backquote in the beginning and normal quote in the end.
Improve _filter_mknod to handle that.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
When fsx fails we try to copy failure state to the results/
dir, but in some cases we are using $seqres instead of
$seq or $seq instead of $here/$seq; fix this up so the
failure state is accurately captured in the results/ dir.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
For fully deduped file, whose file extents are all pointing to the
same extent, btrfs backref walk can be very time consuming, long
enough to trigger softlock.
Unfortunately, btrfs send is one of the caller of such backref walk
under an O(n) loop, making the total time complexity to O(n^3) or
more.
And even worse, btrfs send will allocate memory in such loop, to
trigger OOM on system with small memory(<4G).
This test case will check if btrfs send will cause these problems.
Reporeted-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
xfs/083 will corrupt the fs intentionally, there will be WARNINGs
in dmesg as expected, so here disable dmesg check.
Signed-off-by: Xiao Yang <yangx.jy@cn.fujitsu.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Darrick J. Wong
Christoph Hellwig
Omar Sandoval
Eryu Guan
Wang Xiaoguang
Zorro Lang
Jan Kara
Qu Wenruo
Eric Sandeen
Xiao Yang