Update the existing stress tests to ensure that we can handle
reflinking the same block a million times, and that we can handle
reflinking million different extents. Add a couple of tests to ensure
that we can ^C and SIGKILL our way out of long-running reflinks.
v2: Don't run the signal tests on NFS, as we cannot interrupt NFS
clone operations.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
[hch@lst.de: don't run on NFS]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Ensure that CoW operations against shared blocks in the source file
work correctly.
v2: remove filefrag dependencies
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Test various scenarios (with dm-flakey) where we simulate write
failures during CoW, to see if the FS can get through it without
blowing up or corrupting data. Plumb in a FS-generic method to
sort out repairing filesystems after they get hit by IO errors.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Ensure that we correctly handle a CoW operation immediately followed
by a truncate, falloc, fpunch, fzero, fcollapse, and finsert operation
in the middle of the CoW'd region before any flush can occur.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Check that we don't expose old disk contents when a directio write to
an unwritten extent fails due to IO errors. This primarily affects
XFS and ext4.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
The new Q_GETNEXTQUOTA quotactl (not yet merged) is designed
to take an ID as input ala Q_GETQUOTA, and return the quota
for the next active ID >= the input ID. This lets us quickly
iterate over all existing quotas by leveraging the kernel's
knowledge of which quotas are allocated and active.
The test contains a new helper binary, test-nextquota, which
tests both the "vfs" and "xfs" versions of the quotactl.
It accepts an ID, and outputs the returned ID, ihard, and
isoft values for that quota. It doesn't return block information
simply because that can vary depending on fs, block size, etc,
and we want something very consistent as output, for verifiation.
The test harness sets quotas for 100 random IDs, remounts,
and uses these quotactls to iterate over all the IDs we set,
using the test binary, making sure we get back what we expect.
Not the prettiest thing, but it works!
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Test what happens when we send largeish buffers to CoW all at once.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Test CoW operations when blocksize < pagesize and the only reflink
block is in the middle of the page.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
These tests examine the behavior of advanced and tricky copy on write
situations.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Add more tests for unaligned copy-on-write things, and explicitly
test the ability to pass "len == 0" to mean reflink/dedupe all
the way to the end of the file".
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Currently we're developing a new in-band deduplication feature for btrfs,
when enabing this feature, it will take much long time to hit the enospc
condition which 275 tries to create. I think 275 is also certain enospc
test and we should add it to 'enospc' group, then we can skip it easily by
excluding 'enospc' group when running fstests.
Signed-off-by: Xiaoguang Wang <wangxg.fnst@cn.fujitsu.com>
This test fails 100% of the time for me on xfs and current git head, and
is not run for ext4 since ext4 does not support shutdown. After talking
with bfoster, it isn't expected to succeed right now. Since the auto
group is for tests that *are* expected to succeed, let's move this one
out.
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Test that a file fsync works after punching a hole for the same file
range multiple times, and that after log/journal replay the file's
content and layout are correct.
This test is motivated by a bug found in btrfs, which is fixed by
the following linux kernel patch:
"Btrfs: fix hole punching when using the no-holes feature"
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Add a few horrible opt-in stress tests to see what happens if we try
to reflink the same block billions of times, and what happens if we
run out of space while reflinking a file.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Ensure that copy-on-writing a reflinked file when there's no free disk
space reflects the desired ENOSPC back to userspace during the write
call. Tests the buffered IO, direct IO, and mmap write paths.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Make sure that running reflink ops while other IO is ongoing doesn't
break the filesystem.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Check that we can feed bad inputs to reflink/dedupe and it'll reject
them.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Check that the free block counts seem to be handled correctly in
the reflink operation and subsequent attempts to rewrite reflinked
copies.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Check that the variants of fallocate (allocate, punch, zero range,
collapse range, insert range) do the right thing when they're run
against a range of reflinked blocks.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Ensure that CoW happens correctly with buffered, directio, and mmap writes.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Test the operation of the btrfs (and now xfs) reflink and dedupe
ioctls at various file offsets and with matching and nonmatching
files.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Move the cp --reflink tests from btrfs/ to generic/ since xfs now
supports that ioctl.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
This tests Brian Foster's fix for xfs:
xfs: always drain dio before extending aio write submission
It launches four adjacent 1k IOs past EOF, then reads back
to see if we have 4k worth of the data we wrote, or something else -
possibly zeros from sub-block zeroing and eof racing.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Calls like fsync() should report failure on partial I/O failure, e.g. a
single failed disk in a raid 0 stripe.
This test is motivated by an XFS bug, and this commit fixed the issue
xfs: return errors from partial I/O failures to files
This case is written by David Jeffery <djeffery@redhat.com> originally.
Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Darrick J. Wong
Eric Sandeen
Xiaoguang Wang
Jeff Moyer
Filipe Manana
Eryu Guan