Previously, our XFS fuzzing efforts were limited to using the xfs_db
blocktrash command to scribble garbage all over a block. This is
pretty easy to discover; it would be far more interesting if we could
fuzz individual fields looking for unhandled corner cases. Since we
now have an online scrub tool, use it to check for our targeted
corruptions prior to the usual steps of writing to the FS, taking it
offline, repairing, and re-checking.
These tests use the new xfs_db 'fuzz' command to test corner case
handling of every field. The 'print' command tells us which fields
are available, and the fuzz command can write zeroes or ones to the
field; set the high, middle, or low bit; add or subtract numbers; or
randomize the field. We loop through all fields and all fuzz verbs to
see if we can trip up the kernel.
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>
Move some fuzzing helper functions into a new common/fuzzy file.
We'll add a lot more fuzzing helpers in subsequent patches.
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>
Back when I created common/populate, I thought it was sufficient to
_require the tools that the populate functions need in the main
file. This turned out to be a bit sloppy, so move them into a
helper function and call it from everything that uses populate.
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>
Since commit 9538d1b69399 (ext4: avoid split extents for DAX
writes), currently sitting in ext4 tree, the block layout of files
has changed when the filesystem is mounted with -o dax and now
matches 'nodelalloc' variant instead of 'nozero'. Fixup the test
configuration.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Upstream xfs_io has been converted to always use LFS compliant
(i.e. 64 bit) pwrite() rather than pwrite64(). Similar changes have
been made for multiple syscalls that have "*64" variants. hence the
error output of all these commands has changed, such as "pwrite64:
..." to "pwrite: ....".
Make a filter to catch the *64 variants and strip it, and
convert all the golden output to use the non-*64 variant. This will
make all golden output matching work correctly regardless of what
version of xfs_io is in use.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Enhance _exclude_scratch_mount_option() function to normalize mount
options. Now it can understand and extract real mount option from
string like "-o opt1,opt2 -oopt3".
And now we do word grep to handle mount options like noinode_cache
and inode_cache.
Finally, allow it to accept multiple options at the same time. No
need for multiple _exclude_scratch_mount_option lines now
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Define test groups for those tests which have _require_xfs_io_command
for punch, collapse, insert, and zero. This makes it easier to
exclude tests that use one of these fallocate commands. Or if you
want to specifically test for those fallocate commands you can do
this.
This obviates an out-of-tree xfstests patch I maintain which used an
XFS_IO_AVOID environment variable to suppress running tests that use
punch, collapse, insert, etc. This was rejected because of the
claim that it could be done using groups. So this commit is in
response to those upstream comments.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
ext4 supports dynamic expansion of inode size via inode's
i_extra_size field. Test the code performing the expansion in the
kernel.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
xfstests supports extended test names like 314-foo-bar, but
installation of these tests was skipped (not matching a regexp). So
this patch fixes the makefiles in tests/*/
The include/buildrules change was written by Dave Chinner.
Signed-off-by: Jan Tulak <jtulak@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
An ext4 file system can be created without a journal, but ext4/021
presumes it will contain one. Make that requirement explicit to
avoid unnecessary failures when testing "nojournal" file systems.
Signed-off-by: Eric Whitney <enwlinux@gmail.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
ext4/271 runs in no journal mode (-onoload), so running test with
journal related mount options makes no sense, and test fails after
kernel commit 1e381f60dad9 ("ext4: do not allow journal_opts for fs
w/o journal"), journal related mount options are not allowed in no
journal mode. So _notrun if there're journal related mount options.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Commit 688f869 fixed this bug:
ext4: Initialize fsync transaction ids in ext4_new_inode()
We manually modify jbd2 journal_superblock_s.s_sequence to be a very
large number, which will greatly reduce the time taken to trigger
this bug, though it seems some too hacked.
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>
There're many tests don't remove $seqres.full before writing to it, and
accumulating logs there, then the logs are always growing over time.
Let's fix them once.
generic/16[1-8] generic/170 and generic/33[34] truncate $seqres.full in
the middle of the test, which results in partial logs. Fix them as well.
xfs/227 has duplicated lines to remove $seqres.full, remove the extra
line.
Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
There were a number of tests that use Direct I/O that weren't testing
to make sure O_DIRECT is actually supported. This will be important
for avoiding false positives when testing ext4 encryption (which does
not support DIO for obvious reasons).
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
We have a wrapper around chattr to make sure people don't do the wrong thing on
their boxes, so we need to be able to specify CHATTR_PROG and have it actually
work, so replace all chattr calls with $CHATTR_PROG.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
The check script requires that it be run as root, so adding
individualized checks for this in each teat is not needed.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
If the delayed allocation is disabled, we need a slightly different
output for the delayed allocation portion of the tests.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: Jan Kara <jack@suse.cz>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Calling EXT4_IOC_MOVE_EXTENT on file not aligned with block size and
block size is smaller than page size would cause integrity issue on the
partial-blocksize part when copying data between orign file and donor
file.
This ext4 kernel patch would fix it, titled
"ext4: don't read blocks from disk after extents being swapped in
move_extent_per_page())"
Though this bug only happens in the blocksize smaller than pagesize
case, there's no harm to test on various block size fs, so no block size
is specified in the test, it depends on the test configurations.
Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Replace every explicit mount/umount of scratch or test devices with
helper functions. This allows the next patch to add in hooks to these
functions in order to set up & tear down overlayfs on every mount/umount
(also adds _test_unmount(), which didn't exist prior)
[Eryu Guan rebased the patch agains latest master and replaced more
mount/umount with helpers]
Signed-off-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eric Sandeen <esandeen@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Depending on mount options, we handle unwritten extents somewhat
differently. So sometimes we end up zeroing out unwritten extent and
converting it to written one and sometimes we just split it. Choose
expected output based on mount options.
Signed-off-by: Jan Kara <jack@suse.com>
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Dave Chinner <david@fromorbit.com>
After patch "ext4: Fix races of writeback with punch hole and zero
range" we don't flush range that's going to be zeroed out which results
in different final extent layout because some extents will be zeroed-out
instead of being split. Update the output file accordingly.
Signed-off-by: Jan Kara <jack@suse.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
The ext4/301, ext4/302, ext4/303, and ext4/304 tests are not crashing
on 3.10.89, 3.14.53, 3.18.21, 4.1.8, and 4.3-rc2. So promote these
tests from the dangerous to the auto group.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
ext4/305 sleeps 3 minutes and does mount/umount loop in background,
which produces lots of logs in dmesg and 3 minutes is not necessary.
Ted pointed out that 30 mount/umount cycles is enough to crash a buggy
kernel, so just limit the mount/umount loop to reduce the runtime. And
now the runtime is about 2s.
Reported-by: Theodore Ts'o <tytso@mit.edu>
Suggested-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Targeted fuzzing tests which destroy various pieces of file,
directory, and symlink metadata; the tests look for (a) kernel
detection of corruption, (b) e2fsck repair of said corruption, and (c)
post-repair fs usability.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Darrick J. Wong
Jan Kara
Dave Chinner
Qu Wenruo
Xiao Yang
Theodore Ts'o
Jan Tulak
Eric Whitney
Eryu Guan
Wang Xiaoguang
Josef Bacik
Eric Sandeen
Jan Kara