This reverts commit a0c92a5871 (disable
space cache), as option nospace_cache was newly introduced in linux-3.2,
so we'll fail to mount btrfs in older kernels.
As an alternative fix, we just list subvolme names, don't assume
what ID numbers those subvolumes will have.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
I can't pass 254, and below is the output:
254 3s ... - output mismatch (see 254.out.bad)
...
ID 256 top level 5 path snap
-ID 257 top level 5 path subvol
+ID 258 top level 5 path subvol
When space cache is enabled (and now mkfs.btrfs always enables it),
there will be some space cache inodes in the root tree, and they
consume some IDs, and that's why "subvol" has the ID 258 but not 257.
Just disable space cache for this test case.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The usage is 'btrfs subvolume set-default <id> <path>', not
'<path> <path>'.
The code happens to work because strotoull(path) returns 0, but it
will fail if in the future we check the argument more strict in
btrfs-progs.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Debian unstable has move the location of the gdbm libraries as part
of it's multi-arch conversion. The libraries are now only found by
the configured library search path in /etc/ld.conf.d, so change the
autoconf macros to test and find the libraries by link tests rather
than hard coding library paths when a header is found.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This test is for write-posix test. If writing a file when the disk is almost
full, the posix wants the call to write as much as possible but not none.
quote the POSIX:
If a write() requests that more bytes be written than there is room for
(for example, [XSI] [Option Start] the process' file size limit or
[Option End] the physical end of a medium), only as many bytes as there
is room for shall be written. For example, suppose there is space for 20
bytes more in a file before reaching a limit. A write of 512 bytes will
return 20. The next write of a non-zero number of bytes would give a
failure return (except as noted below).
Signed-off-by: Wu Bo <Wu.Bo@cn.fujitsu.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This test is for preallocation test. If the disk is full, just with a prealloc
file has some free space that prealloc early. We need to check whether the write
to the free space is success or not.
Signed-off-by: Wu Bo <Wu.Bo@cn.fujitsu.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The first argument of _scratch_mkfs_sized() should be file system size
in bytes, so '100m' is not a valid number, and the 'expr' expression in
_scratch_mkfs_sized() complains 'non-numeric argument' about it.
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Make the logging of operations in test 205 to the 205.full file work
correctly. This prevents the overwrite of mkfs status, and leaves
the 205.full file around for later analysis even on successful test
runs.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This test is a stress test. It creates a set of threads for coping small files
into disk. I use a 2G disk for test, the ENOSPC arises usually but the disk is
not full under kenerl 3.0 with intel64.
Signed-off-by: Wu Bo <Wu.Bo@cn.fujitsu.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The recently added fsstress operations result in a different
number of directories and files being created for xfsdump test
022, which causes the test to fail since the expected xfsrestore
output records these numbers. Filter out the numbers to prevent
this from breaking in the future.
Also increase the number of fsstress operations used to populate
the filesystem so that we get roughly the same number of
directories and files as we used to for this test.
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Bill Kendall <wkendall@sgi.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Implement the multi-stream placeholder tests 059 and 060. The comments
imply that these tests existed on IRIX, but I was unable to find them.
Test 059 does a simple 4-way dump and restore. Test 060 does a 4-way
dump, then restores each dump file individually (in a cumulative
fashion).
The tests are skipped if the installed xfsdump does not support
multi-stream output.
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Bill Kendall <wkendall@sgi.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tests that are not executable are made executable when they are
run, causing git to report them as modified. Fix this by
changing non-executable tests to have permissions of 0755.
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Bill Kendall <wkendall@sgi.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
_require_scratch unmounts the SCRATCH_DEV. So it is not necessery
to do it explicitly at the first time.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
dd should not have count argument because it suposed to work untill ENOSPC.
Also it is reasonable to log dd's output, do that for both 269'th and 270'th
tests
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
It wants to grant CAP_CHOWN to fsstress, without this capability
chown(2) will fail due to EPERM. Since chown(2) is one of the most
error prone places of quota accounting this makes 270'th
test almost useless, so it is better to simply skip it.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
- Same as 264 but with quota enabled.
- IO performed from $qa_user user
- fsstress granted with CAP_CHOWN capability.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
During stress testing we want to cover as much code paths as possible
fsstress is very good for this purpose. But it has expandable nature
(disk usage almost continually grow). So once it goes in no ENOSPC
condition it will be where till the end. But by running 'dd' writers
in parallel we can regularly trigger ENOSPC but only for a limited
periods of time because each time it opens the same file with O_TRUNC.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Add two new operations:
- getattr: ioctl(fd, FS_IOC_GETFLAGS, &fl)
- setattr: ioctl(fd, FS_IOC_SETFLAGS, &random_flags)
Attribute mask may be passed via -M opt, by default is (~0).
By default FS_IOC_SETFLAGS has zero probability because
it may produce inodes with APPEND or IMMUTABLE flags which
are not deletable by default. Let's assumes that one who
enable it knows how to delete such inodes.
For example like follows:
find $TEST_PATH -exec chattr -i -a {} \;
rm -rf $TEST_PATH
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Add tests for fallocate(2) syscall
- fallocate: reserve the disk space
- punch: de-allocates the disk space
Since FALLOC_FL_PUNCH_HOLE is relatively new it's value defined
explicitly if not yet defined. Later we may clear that define.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This is regression tescase for explicit combination of seed+fs_opts.
Let's hardcode all options to prevent interactions with fsstress
changes in future.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Once some combination of seed+fs_ops result in regression it is
reasonable to document that combination. It is usefull to dump
that configuration in command line style. Later this line may be
simply hardcoded in to regression test.
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Li Zefan
Dave Chinner
WuBo
Eryu Guan
Bill Kendall
Dmitry Monakhov