Brian Foster [Tue, 17 Jun 2014 23:29:22 +0000 (09:29 +1000)]
xfs/013: allow non-write fsstress operations in background workload
It has been reported that test xfs/013 probably uses more space than
necessary, exhausting space if run against a several GB sized ramdisk.
xfs/013 primarily creates, links and removes inodes. Most of the space
consumption occurs via the background fsstress workload.
Remove the fsstress -w option that suppresses non-write operations. This
slightly reduces the storage footprint while still providing a
background workload for the test.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Regression test for the btrfs ioctl clone operation when the source range
contains hole(s) and the FS has the NO_HOLES feature enabled (file holes
don't need file extent items in the btree to represent them).
This issue is fixed by the following linux kernel btrfs patch:
Btrfs: fix clone to deal with holes when NO_HOLES feature is enabled
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
btrfs: test for btrfs send when nested subvols/snapshots exist
Regression test for a btrfs incremental send issue where the difference
between the snapshots used by the incremental send consists of one of
these cases:
1) First snapshot has a directory with name X and in the second snapshot
that directory doesn't exist anymore but a subvolume/snapshot with
the same name (X) exists;
2) First snapshot has a subvolume/snapshot with name X and in the second
snapshot that subvolume/snapshot doesn't exist anymore (might have been
replaced by a directory with the same name or not).
This issue is fixed by the following linux kernel btrfs patches:
Btrfs: send, don't error in the presence of subvols/snapshots
Btrfs: set dead flag on the right root when destroying snapshot
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Verify that btrfs send is able to replicate xattrs larger than
PATH_MAX. This is possible if the b+tree leaf size is larger
than 4Kb (mkfs.btrfs's default is max(16Kb, PAGE_SIZE) as of
btrfs-progs v3.12, and max(4Kb, PAGE_SIZE in older versions).
This issue is fixed by the following linux kernel btrfs patch:
Btrfs: send, use the right limits for xattr names and values
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
This is a test to verify that the btrfs ioctl clone operation is
able to clone extents of a file to different positions of the file,
that is, the source and target files are the same. Existing tests
only cover the case where the source and target files are different.
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: David Disseldorp <ddiss@suse.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
common: add helper require function _require_btrfs_cloner
So that the same check (btrfs cloner program presence) can be reused
by other tests.
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: David Disseldorp <ddiss@suse.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Regression test for btrfs send where long paths (exceeding 230 characters)
made send produce paths with random characters from a memory buffer returned
by kmalloc, as send forgot to populate the new buffer with the path string.
This issue is fixed by the following linux kernel btrfs patch:
Btrfs: send, fix corrupted path strings for long paths
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Brian Foster [Thu, 15 May 2014 01:37:55 +0000 (11:37 +1000)]
config: make sure tests check for loop device support
Several tests happen to make use of loop device support without the
requisite pre-test checks. This results in spurious failures for systems
that might not have loop device support. Add _require_loop checks to
shared/298, xfs/206 and xfs/259.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Regression for btrfs send when an inode only has extended references
associated to it (no regular references present). This used to cause
incorrect access to a b+tree leaf, where an extended reference item
was accessed as if it were a regular reference item, causing unexpected
and unpredictable behaviour such as producing a random/weird path string
or a crash.
This issue is fixed by the following linux kernel btrfs patch:
Btrfs: send, fix incorrect ref access when using extrefs
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Brian Foster [Thu, 15 May 2014 01:37:49 +0000 (11:37 +1000)]
xfs/013: use killall -9 and wait in _cleanup()
Kill any lingering fsstress processes and wait properly should we abort
the test. This prevents the workload from inadvertently affecting
subsequent tests.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Jeff Layton [Thu, 15 May 2014 01:34:56 +0000 (11:34 +1000)]
loggen: fix build vs. newer xfsprogs headers
v2: just include xfs/libxfs.h for the conftest prerequisites
commit cc085d770adb in xfsprogs removed some header files that are
included by loggen.c. Add an autoconf test that checks to see whether
xfs_log_format.h is present. If it is, include that instead of the
deprecated headers and add an alias for XFS_TRANS_MAGIC since that
value no longer exists.
With this patch, xfstests builds on both f20 and rawhide (f21) boxes.
Signed-off-by: Jeff Layton <jlayton@poochiereds.net> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Jeff Layton [Tue, 13 May 2014 05:30:23 +0000 (15:30 +1000)]
locktest: consolidate do_lock and do_unlock, and add ability to F_GETLK
Eliminate do_unlock as it's just cut-and-paste of do_lock (down to the
bad stderr fprintf). Fix the debug messages to print the cmd and type,
and add the ability to do a WRTEST and RDTEST.
Signed-off-by: Jeff Layton <jlayton@poochiereds.net> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Jeff Layton [Tue, 13 May 2014 05:30:18 +0000 (15:30 +1000)]
locktest: don't assume that F_OPEN should use O_RDWR
In a later patch, I'll be adding a test for F_GETLK that will require
that we open the file with a different f_mode. Change do_open not to
assume that the it needs to open the file O_RDWR.
While we're at it, fix a bug in do_open. Just because we pass in '0' for
the flags, doesn't mean that this is the initial open. Move the exit(1)
for that case to the caller.
Signed-off-by: Jeff Layton <jlayton@poochiereds.net> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Josef Bacik [Tue, 13 May 2014 05:30:15 +0000 (15:30 +1000)]
config: fix selinux context handling
With the new config stuff we lost the selinux options being set for systems with
selinux turned on. We want the selinux context set all the time, wether we
provide a MOUNT_OPTIONS value or not, so take this logic out of _mount_opts()
and just put it in the body of common/config
Signed-off-by: Josef Bacik <jbacik@fb.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Josef Bacik [Tue, 13 May 2014 05:30:03 +0000 (15:30 +1000)]
common: fix flink check
I don't have flink support in my xfsprogs, but it doesn't fail with "command not
found" or whatever, it fails because I don't have the -T option, whereas Eric
gets an error about $TEST_DIR being a directory because his xfs_io tries to open
the directory first before it parses the options. So fix this by checking flink
with these two cases and don't run if we hit either of them. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Wang Shilong [Tue, 13 May 2014 05:29:37 +0000 (15:29 +1000)]
btrfs: add regression test for inode cache vs tree log
This patch adds a regression test to verify btrfs can not
reuse inode id until we have committed transaction. Which was
addressed by the following kernel patch:
Btrfs: fix inode cache vs tree log
Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
All commands run as $qa_user should use a relative path so that
missing access permissions on $HOME for root don't prevent running
it. This fixes common/233 for me.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Brian Foster [Tue, 13 May 2014 05:28:35 +0000 (15:28 +1000)]
xfs/013: stress the free inode btree
Create a stress test for the free inode btree. Allocate a set of inodes
sequentually and run a hard link clone and random replacement algorithm
across the set. Background removal of the oldest directories creates a
sparse set of free inodes over time. Run an fsstress workload
concurrently to exercise the fs.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Brian Foster [Tue, 13 May 2014 05:27:47 +0000 (15:27 +1000)]
xfs/010: test repair for finobt corruption
The finobt creates a duplicate subset of inode allocation metadata from
the inobt. xfs_repair should detect and repair inconsistencies in the
finobt that could be caused by bugs or corruption. This test uses xfs_db
to cause targeted corruptions in the finobt and verify repair detects
and corrects the filesystem.
In particular, the test corrupts individual finobt records to cause
inconsistency between the inode allocation count fields as well as
causing the finobt to contain a record with no free inodes.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Brian Foster [Tue, 13 May 2014 05:27:47 +0000 (15:27 +1000)]
repair: filter agno/ino repair output for finobt
finobt enabled filesystems can generate new repair output. Update
_filter_repair() to ensure specific AG and inode numbers are filtered
from test output.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Brian Foster [Tue, 13 May 2014 05:26:59 +0000 (15:26 +1000)]
xfs/030: filter out extra repair noise for finobt enabled fs'
xfs/030 nukes various on-disk data structures to test for repair. This
can result in extra output when testing finobt enabled filesystems. For
example, xfs_repair detects an invalid free inode btree root block when
the agi is zeroed.
Filter this output directly in xfs/030 such that the test passes for
finobt and non-finobt filesystems.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Namjae Jeon [Tue, 13 May 2014 05:26:59 +0000 (15:26 +1000)]
fsstress: fix incorrect if condition check for collapse range mode
There is if condition to be block aligned for off and len of Collapse range.
But off and len for all fallocate opearion can be aligned by incorrect
if condition check.
Theodore Ts'o [Mon, 12 May 2014 23:04:13 +0000 (09:04 +1000)]
check: add support for an external file containing tests to exclude
Currently the -X option is intended to specify a set of expunging
files which are stored in each test/* subdirectory. As described in
the commit description for 0b1e8abd4, in order to exclude the test
generic/280, the -X option is used as follows:
However, it is sometimes useful to store the set of expunged tests in
a single file, outside of tests/* subdirectories. This commit enables
the following:
common: Use _require_xfs_io_command() instead of helpers
Number of helpers for checking xfs_io functionality is slowly
growing. But it's as easy to simply use _require_xfs_io_command()
directly and just specify the command we want to check. It will also
avoid the need to create helper every time we need to check a new
command in xfs_io.
Remove all the helpers and use _require_xfs_io_command() in the
tests.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Dave Chinner [Mon, 28 Apr 2014 00:55:12 +0000 (10:55 +1000)]
generic: introduce new large ACL test
Having just removed the largeacl test from the shared ACL test,
reintroduce the same test as an generic test so that we can
handle the different limits in supported ACL count appropriately
across different filesystems and different configurations within
filesystem types.
Filesystems have to add support to _acl_get_max to run
this test - the default behaviour right now is to throw a
notrun error like this:
generic/026 14s ... [not run] ext4 does not define maximum ACL count
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Dave Chinner [Mon, 28 Apr 2014 00:55:04 +0000 (10:55 +1000)]
shared/051: remove ACL count subtest
Different versions of XFS support different numbers of ACLs on disk.
Remove that subtest from this shared test to prevent it form causing
spurious failures on filesystems that support more than 25 ACLs.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Dave Chinner [Mon, 28 Apr 2014 00:54:53 +0000 (10:54 +1000)]
filter: xfs_repair emits new corruption messagse
xfs_repair now dumps metadata and CRC corruption information to the
output from the verifier infrastrcuture. Filter this out so that it
doesn't cause spurious test failures such as in xfs/030 and xfs/033.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Dave Chinner [Mon, 28 Apr 2014 00:54:45 +0000 (10:54 +1000)]
xfs: remove dmapi tests from the auto group
There is no mainline kernel support for DMAPI in XFS, and so every
time the xfstests auto group is run it throws a large number of
"[not run] Assuming DMAPI modules are not loaded". I've noted that
people are excluding the dmapi group to avoid this, so rather than
inflict pain on everyone, make hose few that need to do dmapi
testing include it specifically.
IOWs, remove the DMAPI tests from the auto group.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Dave Chinner [Mon, 28 Apr 2014 00:54:38 +0000 (10:54 +1000)]
generic: cleanup space after test in TESTDIR
A couple of tests leave behind large files or directory structures
when they complete, which leads to small TEST_DEVs running out of
space during other tests. Make those space hogs clean up after
themselves so that random tests don't fail with ENOSPC errors.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Dave Chinner [Mon, 28 Apr 2014 00:54:27 +0000 (10:54 +1000)]
generic/204: tweak reserve pool size
On small block size filesystems, the reserve pool size is kept
constant at 4MB. filesystems with smaller blocks use comparitively
more blocks for indexing metadata (e.g. in the inode and extent
btrees) and so having a higher indirect block usage. Hence we need
to leave the reserve pool at 1024 block and not scale it for a
constant size.
This makes the test pass on a filesystem made with MKFS_OPTIONS="-b
size=1024 -m crc=1".
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Dave Chinner [Tue, 22 Apr 2014 01:00:55 +0000 (11:00 +1000)]
xfs/167: need at least 10GB of scratch space to run
When running on a ramdisk, the fsstress background workload consumes
a GB of disk space every 5 seconds. This leads to the test failing
with ENOSPC because the test file cannot be created due otthe
background load cosuming it all. Hence don't run this test unless
the scratch device is large enough not to hit ENOSPC conditions.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Eric Whitney [Tue, 22 Apr 2014 00:47:53 +0000 (10:47 +1000)]
generic/009: fix check for zero range support
Generic/009 fails when run on a file system that does not support byte range
zeroing. For example, an EOPNOTSUPP failure occurs when the test is run
on a pre-3.15 extent-mapped file system. The code in the test intended
to prevent this contains an apparent typo that results in a check for
fallocate() rather than zero range support.
Signed-off-by: Eric Whitney <enwlinux@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Eric Whitney [Tue, 22 Apr 2014 00:47:33 +0000 (10:47 +1000)]
ext4/001: fix check for zero range support
Ext4/001 fails when run on a file system that does not support byte range
zeroing. For example, an EOPNOTSUPP failure occurs when the test is run
on a pre-3.15 extent-mapped file system. The code in the test intended
to prevent this contains an apparent typo that results in a check for
fallocate() rather than zero range support.
Signed-off-by: Eric Whitney <enwlinux@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Eric Sandeen [Tue, 22 Apr 2014 00:47:28 +0000 (10:47 +1000)]
xfs: test X_QUOTARM functionality
The Q_XQUOTARM quotactl was not working properly, because
we weren't passing in proper flags. The xfs_fs_set_xstate()
ioctl handler used the same flags for Q_XQUOTAON/OFF as
well as Q_XQUOTARM, but Q_XQUOTAON/OFF look for
XFS_UQUOTA_ACCT, XFS_UQUOTA_ENFD, XFS_GQUOTA_ACCT etc,
i.e. quota type + state, while Q_XQUOTARM looks only for
the type of quota, i.e. XFS_DQ_USER, XFS_DQ_GROUP etc.
Unfortunately these flag spaces overlap a bit, so we
got semi-random results for Q_XQUOTARM; i.e. the value
for XFS_DQ_USER == XFS_UQUOTA_ACCT, etc. yeargh.
Anyway, here's a simple test that demonstrates it,
kernel patch to fix it will follow.
Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
generic/285: fix test for generic SEEK_HOLE/DATA implementations
Generic implementation of SEEK_HOLE/DATA reports whole file as data chunk with
virtual hole at the end (generic_file_llseek, used by ext2/ext3/reiserfs/nfs).
Currently test prepares file smaller than expected for "huge file" testcases
(testcases 10-12). This patch fixes test file layout, now second data chunk
ends right at the expected end of file.
Plus it fixes type of 'filesz' argument, it should be off_t.
Signed-off-by: Konstantin Khlebnikov <k.khlebnikov@samsung.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Eric Sandeen [Tue, 22 Apr 2014 00:46:49 +0000 (10:46 +1000)]
fsx: introduce FSX_AVOID env var
Just like FSSTRESS_AVOID, FSX_AVOID can be used to add
options at the end of the default fsx runs in each test.
i.e. FSX_AVOID="-H -z -C" will disable punch hole, zero range,
and collapse range calls in all tests which run fsx.
This should handle Ted's concerns about buggy ext4 fallocate
code without needing to add tunables to the kernel to reject
these operations during xfstests runs.
Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Files that consist of an inline extent, have the corresponding
data in the filesystem btree and not on a dedicated extent. For
such extents filefrag (fiemap) will report a physical location
of 0 for that extent and set the 'inline' flag.
The btrfs inspect-internal logical-resolve command will cause a
lookup in the extent tree for the extent address we give it as
an argument, which fails with errno ENOENT if it is 0.
This error didn't happen always, as the test uses fsstress to
generate a random filesystem, which needed to generate at least
one file that could be inlined (content less than 4018 bytes).
Example, taken from results/btrfs/004.full:
# filefrag -v /home/fdmanana/btrfs-tests/scratch_1/snap1/p0/de/d1b/dcb/fb1
Filesystem type is: 9123683e
File size of /home/fdmanana/btrfs-tests/scratch_1/snap1/p0/de/d1b/dcb/fb1 is 3860 (1 block of 4096 bytes)
ext: logical_offset: physical_offset: length: expected: flags:
0: 0.. 4095: 0.. 4095: 4096: not_aligned,inline,eof
1: 280.. 344: 35190.. 35254: 65: 1: eof
/home/fdmanana/btrfs-tests/scratch_1/snap1/p0/de/d1b/dcb/fb1: 2 extents found
after filter: 0#0#0 0#0#0
# stat -c %i /home/fdmanana/btrfs-tests/scratch_1/snap1/p0/de/d1b/dcb/fb1
403
# /home/fdmanana/git/hub/btrfs-progs/btrfs inspect-internal logical-resolve -P 0 /home/fdmanana/btrfs-tests/scratch_1
ioctl ret=-1, error: No such file or directory
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
btrfs/004: fix filefrag filter for files with 1 block only
If the file consists of a single block, then filefrag mentions
'1 block of ...', and the filter expected 'blocks of ...'.
Example:
$ echo qwerty > foobar
$ filefrag -v foobar
Filesystem type is: ef53
File size of foobar is 7 (1 block of 4096 bytes)
ext: logical_offset: physical_offset: length: expected: flags:
0: 0.. 0: 0.. 0: 1: unknown,delalloc,eof
foobar: 1 extent found
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Print _fs_has_crcs=0 to stderr by default, so old xfsprogs could have
this variable set too, and a latter _fs_has_crcs=1 could overwrite it
if the fs does have crc support.
Signed-off-by: Eryu Guan <eguan@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
This test case verifies the btrfs properties feature, a new feature
introduced in the linux kernel version 3.14.
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
This test verifies that after an incremental btrfs send the
replicated file has the same exact hole and data structure as in
the origin filesystem. This didn't use to be the case before the
send stream version 2 - holes were sent as write operations of 0
valued bytes instead of punching holes with the fallocate system
call, and pre-allocated extents were sent as well as write
operations of 0 valued bytes instead of intructions for the
receiver to use the fallocate system call.
It also checks that prealloc extents that lie beyond the file's
size are replicated by an incremental send.
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Eric Whitney [Mon, 14 Apr 2014 00:37:33 +0000 (10:37 +1000)]
generic/311: add fallocate() check
Generic/311 fails when run on a test filesystem that does not
support fallocate(). Its I/O load is produced by fsync-tester,
which uses fallocate() system calls to allocate blocks for some of
its test cases. This causes EOPNOTSUPP failures when the test is
run on indirect block-mapped ext4 filesystems.
Verify that the test filesystem supports fallocate() before
proceeding with the test, checking for block allocation
capabilities. Also, fix a minor error message typo.
Signed-off-by: Eric Whitney <enwlinux@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Eric Whitney [Mon, 14 Apr 2014 00:37:32 +0000 (10:37 +1000)]
generic/300: add fallocate() checks
Generic/300 fails when run on a test filesystem that does not
support fallocate(). It uses fio's falloc ioengine to generate part
of its I/O load and both allocates blocks and punches holes. This
causes EOPNOTSUPP failures when the test is run on indirect
block-mapped ext4 filesystems or pre-3.14 ext4 filesystems created
with bigalloc.
Verify that the test filesystem supports fallocate() before
proceeding with the test, checking for both block allocation and
hole punching capabilities. Also, delete any pre-existing test
output.
Signed-off-by: Eric Whitney <enwlinux@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
config: Add -s option to run only specified sections
This commit adds -s option which allows you to specify only certain
sections from the config file to be run. The '-s' argument can be
specified multiple times to allow multiple sections to be run.
The options are still carried between section, that includes the
sections which are not going to be run.
Update README.config-sections as well.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
generic/204 fails on device with Advanced Format of 4096 bytes per
physical sector and when partition starts at the 4K boundary/./In
this case filesystem sector/block size will be of 4096 bytes size
and _scratch_mkfs_sized fails because mkfs reports that 5Mb log size
is not enough to create a filesystem, for example attempt to make
filesystem on such partition:
"log size 1280 blocks too small, minimum size is 1605 blocks"
and generic/204 fails with ENOSPC before it has finished creating
the necessary files. Log size of 7MB is enough for this test to pass.
Signed-off-by: Alexander Tsvetkov <alexander.tsvetkov@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
config: fix specifying configuration value with equality sign
Currently there is a problem with parse_config_section() when the
configuration value contains equality sign like this for example:
MOUNT_OPTIONS="-o data=journal"
the result will be
export MOUNT_OPTIONS="-o data="journal"
which is not going to work. The reason is that the expression used to
parse the configuration options uses greedy matching '.*'. Fix this by
using non greedy expression to match the first equality sign '[^=]'.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Currently if the FSTYP is not set, the code to get FSTYP using blikd
would not work. This is because we're using HOSTOS environment variable
which might not be set (at least not on my system) and because it's
already late in the code path.
Fix this by using OSTYP environment variable as a fallback in the case
that HOSTOS does not work and move the check to common/config.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Tested-by: Filipe David Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
config: Unset SCRATCH_DEV when deduced from SCRATCH_DEV_POOL
In the case that we already have sections in the config file we
have to make sure that we unset SCRATCH_DEV if it has been deduced from
the SCRATCH_DEV_POOL so that it does not complain about SCRATCH_DEV in
this case.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Tested-by: Filipe David Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
common/config: Error: $SCRATCH_DEV should be unset when
$SCRATCH_DEV_POOL is set
This was because get_next_config() would get called twice and hence it
would complain on the second run that SCRATCH_DEV is already set. Fix
it by making sure that we call get_next_config() only once if there
are no sections in the config file.
Also make sure that we export SCRATCH_DEV in the case we're deducing it
from SCRATCH_DEV_POOL.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reported-by: Filipe David Manana <fdmanana@gmail.com> Tested-by: Filipe David Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
btrfs: add test for btrfs incremental send data corruption
Regression test for the btrfs incremental send feature, where the kernel
would incorrectly consider a range of a file as a hole and send a stream
of 0 bytes to the destination (send stream) that would overwrite the
corresponding file region.
This issue is fixed by the following linux kernel btrfs patch:
Btrfs: send, fix data corruption due to incorrect hole detection
(https://patchwork.kernel.org/patch/3910081/)
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
[dchinner: changed binary/test file name to dirhash_collide[.c]]
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Tested-by: Mark Tinguely <tinguely@sgi.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
There are couple of tests in shared directory which really should be
made generic, so move it. It is mostly collapse range tests, which
really can be generic to make super we test every file system which adds
collapse range support.
check: unmount TEST_DEV and SCRATCH_DEV after test run
Unmount TEST_DEV and SCRATCH_DEV after each test run to avoid
mounting multiple devices on the same mount point which might result
in xfstest not being able to unmount the device later down the path.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
check: Remount file system if MOUNT_OPTIONS changed
When MOUNT_OPTIONS change we should remount TEST_DEV to put the changes
in effect. This will allow us to have different MOUNT_OPTIONS in sections
in configuration file.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Add config option RECREATE_TEST_DEV to allow to recreate file system on
the TEST_DEV device. Permitted values are true and false.
If RECREATE_TEST_DEV is set to true the TEST_DEV device will be
unmounted and FSTYP file system will be created on it. Afterwards it
will be mounted to TEST_DIR again with the default, or specified mount
options.
Also recreate the file system if FSTYP differs from the previous
section.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
This patch add support for sections in the config file. Each section can
contain configuration options in the format
OPTION=value
when one section is processed xfstests will proceed to next section
until all secitons are processed, or an error occur.
The name of the section can consist of alphanumeric characters + '_',
nothing else is allowed. Name of the section is also used to create
results subdirectory for each section. After all the sections are
processed summary of all runs is printed out.
If the config file does not contain sections, or we're not using config
file at all, nothing is changed and xfstests will work the same way as
it used to.
This is very useful for testing file system with different options. Here
is an example of the config file with sections:
Note that once the variable is set it remains set across the sections, so
you do not have to specify all the options in all sections. However one
have to make sure that unwanted options are not set from previous
sections.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
This patch only adds the indentation in place so we will be able
to clearly see and review changes made in the second patch which will
add a loop (instead of always-true condition introduced in this patch)
adding support for config sections. There are no changes in the logic,
only indentation changes.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
generic: add generic test for fallocate zero range
This is based on xfs/242. This is very similar to ext4/001 however this
test has some tweaks to make it work test zero range on generic file
system. This includes turning off ext4 extents zeroout and disabling
the test for xfs on systems where PAGE_SIZE > 4096.
It is testing extent tree manipulation with fallocate zero range
operation.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Filipe Manana [Fri, 4 Apr 2014 06:17:01 +0000 (17:17 +1100)]
btrfs: add test for btrfs send directory moves/renames
Regression test for a btrfs incremental send issue where the kernel failed
to build paths strings. This resulted either in sending a wrong path string
to the send stream or entering an infinite loop when building it.
This happened in the following scenarios:
1) A directory was made a child of another directory which has a lower inode
number and has a pending move/rename operation or there's some non-direct
ancestor directory with a higher inode number that was renamed/moved too.
This made the incremental send code go into an infinite loop when building
a path string;
2) A directory was made a child of another directory which has a higher inode
number, but the new parent wasn't moved nor renamed. Instead some other
ancestor higher in the hierarchy, with an higher inode number too, was
moved/renamed too. This made the incremental send code go into an infinite
loop when building a path string;
3) An orphan directory is created and at least one of its non-immediate
descendent directories have a pending move/rename operation. This made
an incremental send issue to the send stream an invalid path string that
didn't account for the orphan ancestor directory.
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Regression test for a btrfs incremental send issue where invalid paths for
utimes, chown and chmod operations were sent to the send stream, causing
btrfs receive to fail.
If a directory had a move/rename operation delayed, and none of its parent
directories, except for the immediate one, had delayed move/rename operations,
after processing the directory's references, the incremental send code would
issue invalid paths for utimes, chown and chmod operations.
This issue is fixed by the following linux kernel btrfs patch:
Btrfs: fix send issuing outdated paths for utimes, chown and chmod
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
btrfs: add test for btrfs send issuing premature rmdir operations
Regression test for btrfs incremental send issue where a rmdir instruction
is sent against an orphan directory inode which is not empty yet, causing
btrfs receive to fail when it attempts to remove the directory.
This issue is fixed by the following linux kernel btrfs patch:
Btrfs: fix send attempting to rmdir non-empty directories
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com> Reviewed-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Jie Liu [Thu, 13 Mar 2014 04:20:18 +0000 (15:20 +1100)]
shared/051: add filter to match the golden output for large acls
On old kernel we return EINVAL if hit the limits of maximum number of
ACLs but return E2BIG on new kernel, which cause the test failes on new
kernel as the output is mismatch to the goldens. This patch fix it by
updating the golden output with the new error message and replacing the
old error message with it via a filter.
Signed-off-by: Jie Liu <jeff.liu@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Wang Shilong [Thu, 13 Mar 2014 04:20:09 +0000 (15:20 +1100)]
btrfs: add basic functional test for btrfs quota groups
Add missing test for btrfs quota groups feature,test idea is to create
a parent qgroup that groups some subvolume groups, we try to write
some data into every subvolume and then check if we exceed parent
qgroup's limit size.
Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Reviewed-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Lukas Czerner [Thu, 13 Mar 2014 04:20:00 +0000 (15:20 +1100)]
ext4/001: Add ext4 specific test for fallocate zero range
This is based on xfs/242. However it's better to make it file system
specific because the range can be zeroes either directly by writing
zeroes, or converting to unwritten extent, so the actual result might
differ from file system to file system. Also xfs results differ
depending on the page size which is not the case for ext4.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
Lukas Czerner [Thu, 13 Mar 2014 04:19:52 +0000 (15:19 +1100)]
build: Move fallocate include into global.h
Move the inclusion of falloc.h with all it's possible defines for the
fallocate mode into global.h header file so we do not have to include
and define it manually in every tool using fallocate.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Lukas Czerner [Thu, 13 Mar 2014 04:19:03 +0000 (15:19 +1100)]
generic/008: Add test for fallocate zero range at block boundary
Add test for fallocate zero range at block boundary. This is similar to
the test xfs/290 however this one is generic and we're testing different
block sizes as well - namely 1k, 2k, 4k and 64k. Note that we're not
creating file systems with given block size buy rather test all 4
options.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
Lukas Czerner [Thu, 13 Mar 2014 04:18:56 +0000 (15:18 +1100)]
common: create _test_block_boundaries in common/punch
Create new function _test_block_boundaries() which is testing content of
the blocks after the operation such as zero, or punch hole. The test is
doing the operation around block boundaries to assure correct behaviour
of the operation on block unaligned ranges.
This has been based on test xfs/290 which has been changed to use this
new function. A small change to the output file was required.
Signed-off-by: Lukas Czerner <lczerner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
projects / xfstests-dev.git / log