Alex Elder [Wed, 19 May 2010 22:44:14 +0000 (22:44 +0000)]
xfstests: 226: have xfs_io use bigger buffers
By default xfs_io uses a buffer size of 4096 bytes. On test 226,
the result is that the test runs much slower (at least an order
of magnitude) than it needs to.
Add a flag to the "pwrite" command sent to xfs_io so it uses
larger buffers, thereby speeding things up considerably.
Signed-off-by: Alex Elder <aelder@sgi.com> Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Jan Kara [Mon, 24 May 2010 09:46:44 +0000 (11:46 +0200)]
Fix fallocate() test
The test for fallocate was broken because it used $TEST_DIR/$tmp.io. Because
$tmp is usually something like /tmp/1234 or /mnt/1234 the file cannot be
created and xfs_io fails regardless of existance of fallocate support. Moreover
the subsequent message parsing decides that fallocate is actually supported
because it does not expect this message.
Jan Kara [Wed, 19 May 2010 09:25:52 +0000 (11:25 +0200)]
Fix test whether kernel supports quotas
For all 2.6 kernels presence of quota support in kernel can be detected by
checking /proc/sys/fs/quota. This is actually more reliable than trying to
mount a filesystem with quota options (for example because SCRATCH_DEV does
not have to contain a filesystem type we are going to test).
Alex Elder [Thu, 6 May 2010 17:26:07 +0000 (17:26 +0000)]
xfstests: honor comments in the test group file
There are some spots in the "group" file where test numbers have
groups listed after a '#' character, clearly intending for those
groups to be commented out. But the way the group list gets
generated that commenting doesn't work, and in fact these tests
explicitly *are* included in such commented-out groups.
This patch fixes that, stripping out all comments (which start
with a '#' character and end with a newline) from the file before
building the set of test numbers for a group.
Signed-off-by: Alex Elder <aelder@sgi.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Amit Arora [Wed, 5 May 2010 19:15:08 +0000 (14:15 -0500)]
xfstests 228: New testcase to check if fallocate respects RLIMIT_FSIZE
Add a new testcase to the xfstests suite to check if fallocate respects
the limit imposed by RLIMIT_FSIZE (can be set by "ulimit -f XXX") or
not, on a particular filesystem.
Signed-off-by: Amit Arora <aarora@in.ibm.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Tao Ma [Tue, 4 May 2010 06:01:00 +0000 (16:01 +1000)]
xfstests: Add query fiemap count test.
According to Documentation/filesystems/fiemap.txt, If fm_extent_count
is zero, then the fm_extents[] array is ignored (no extents will be
returned), and the fm_mapped_extents count will hold the number of
extents needed.
Dave Chinner [Tue, 4 May 2010 06:01:26 +0000 (16:01 +1000)]
xfsqa: build dbtest on debian platforms
Set up autoconf to find the correct headers and compat libraries for
debian squeeze, and massage the includes to ensure the right headers get
included. Also fix a compile warning that was emitted now that it is being
compiled.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 4 May 2010 06:01:36 +0000 (16:01 +1000)]
xfsqa: keep xfs_fsr output around in test 222
To confirm that xfs_fsr is doing the right thing, make it output debug
and verbose messages and store them in the 222.full file so that it
can be checked after the fact for correct operation.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 4 May 2010 06:01:47 +0000 (16:01 +1000)]
xfsqa: clean up 030 repair output
With the new checks in xfs_repair, it outputs more information
about errors found than previously. This new output can be ignored
for the purposeѕ of this test, so filter it all out. This will
allow the test to run on new and old reapir binaries.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Wed, 28 Apr 2010 05:34:06 +0000 (15:34 +1000)]
xfsqa: new fsr defragmentation test
This test aims to recreate the conditions that caused xfs_fsr to
corrupt inode forks. The problem was that the data forks between the
two inodes were in different formats due to different inode fork
offsets, so when they swapped the data forks the formats were
invalid.
This test generates a filesystem with a known fragmented freespace pattern and
then abuses known "behaviours" of the allocator to generate files
with a known number of extents. It creates attributes to generate a
known inode fork offset, then uses a debug feature of xfs_fsr to
attempt to defrag the inode to a known number of extents.
By using these features, we can pretty much cover the entire matrix of inode
fork configurations, hence reproducing the conditions that lead to corruptions.
This test has already uncovered one bug in the current kernel code, and the
current fsr (with it's naive attribute fork handling) is aborted on a couple of
hundred of the files created by this test.
The -s option to repquota used to be a no-op, but actually changes
output to different units in quota tools 4.0. Remove it from the
repquota invocation in test 219.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Eric Sandeen <sandeen@sandeen.net> Tested-by: "Theodore Ts'o" <tytso@mit.edu>
Dave Chinner [Sat, 6 Mar 2010 00:25:10 +0000 (11:25 +1100)]
xfsqa: define resblks for tests near ENOSPC
Several tests assume a certain amount of disk space free after the
reserve block pool is filled. Changing the default size of the
reserve block pool breaks these tests because there is less space
available that first thought.
Change these tests to specify a known reserve block pool size of
1024 blocks to ensure that they continue to work correctly even if
the default size changes.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Eric Sandeen [Sat, 6 Mar 2010 00:24:54 +0000 (11:24 +1100)]
xfsqa: Add fiemap exerciser
Preliminary fiemap testing support based on a test util written by
Josef Bacik.
For now it's only run with preallocation disabled, because xfs has a
tendency to fill in holes with data blocks (EOF prealloc stuff I
think) and similar for explicit preallocation, so this is breaking
the preallocation tests for now, when it finds a "data" block where
it expects a preallocated block.
Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Sat, 6 Mar 2010 00:24:38 +0000 (11:24 +1100)]
xfsqa: fix size specification for scratch mkfs
When making a specifically sized scratch filesystem, we need to
calculate the size rather than assigning the caclculation expression
to a variable and passing that into the _scratch_mkfs_sized
function. Currently sized filesystems are failing to be made and the
code is falling back to full sized filesystems so we are not testing
enospc issues correctly.
Dave Chinner [Sat, 6 Mar 2010 00:22:58 +0000 (11:22 +1100)]
xfsqa: add delayed allocation @ ENOSPC exerciser
When we deplete the reserve block pool or receive an error during
delayed allocation, we currently toss the page away. If that page
has delayed allocation extents on it, we can fail to remove them and
leave stale delayed llocation extents lying around on the inode.
This can, in certain circumstances, trigger errors later on when the
stale delalloc extent it found again, including tripping a BUG().
Exercise this failure path so that we get code coverage of the fix
that prevents stale delalloc mappings from being left on the inode
when pages are tossed.
This is based on a test case supplied by Lachlan McIlroy.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Recent fixes to 073 added umount -d, but if we happen
to be using a scratch device on loopback, subsequent
tests get very, very unhappy when their loopback block
device goes away!
Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
blkid without options usually gets no probe results at all just after
creating a filesystem. This problem is due to the cache that blkid
uses in it's default mode, and is unlikely to get fixed. Use the -p
option to bypass the cache layer.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <david@fromorbit.com>
Alex Elder [Wed, 3 Feb 2010 19:54:04 +0000 (19:54 +0000)]
xfstests: 219: fix awk filter for duplicate users
The filter I added for removing duplicate users from the
output of repquota didn't do the job very well. This
fixes that, making it so the first time a user is seen
its line is printed, not thereafter.
Signed-off-by: Alex Elder <aelder@sgi.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Eric Sandeen [Tue, 9 Feb 2010 17:59:23 +0000 (11:59 -0600)]
xfstests: fix up fs_perms test used by 126
Test 126 was failing intermittently for Ted & I; it seems that
this is because we were passing an unterminated string to
fopen for the mode; I'm not certain why this made it fail,
but it's pretty clearly not a good thing to do, and fixing
it fixes the test.
Rather than passing around characters, do things string-wise,
since that is what is ultimately used in fopen().
Reported-by: Theodore Tso <tytso@mit.edu> Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Christoph Hellwig <hch@lst.de>
Eric Sandeen [Wed, 3 Feb 2010 16:43:53 +0000 (10:43 -0600)]
xfstests: 223 - test file alignment on stripe geometry
A first-cut test to ensure that files are well-aligned on filesystems
with stripe geometry.
Several sizes of stripe units are mkfs'd, and then files are written
and fallocated in various multiples of those stripe sizes.
Each file is checked to ensure that the first block is
stripe-aligned.
(Ideally, for any fragmented files, we should ensure that each
fragment start is well-aligned, but this does not do that yet)
(slightly unrelated: don't send scratch mkfs output to /dev/null,
we'd like to see mkfs output and direct it to $seq.full - this
more or less matches _scratch_mkfs_xfs behavior and doesn't
break any tests that I can see)
Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 26 Jan 2010 03:22:35 +0000 (14:22 +1100)]
xfstests: Automatic build dependency calculations
Currently the xfstest builds do not have any automatic dependency
calculations. It relies on a separate make depend run to build or
update dependency information. It also relies on an external
makedepend binary. If that binary does not exist, the dependencies
do not get calculated.
To remove the dependency on makedepend, gcc can be used instead as
it has a command to generate dependency information. This patch
changes the dependency rule building to use gcc.
In case anyone uses an old (several years) gcc compiler or a
compiler that doesn't support gcc compatible dependency generation,
a new configure check is added to turn off dependency checking so
builds can still be done.
To use the dependencies automatically, we need to use a special
include makefile directive to include the build dependencies into
the current makefile. Essentially once the dependencies are
calculated, they can be included into the makefile and make will
recalculate the build dependencies automatically based on that
information.
Hence we get a build that automatically calculates and keeps
dependencies up to date without dependence on any external tools.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 26 Jan 2010 03:21:47 +0000 (14:21 +1100)]
xfstests: rebuild aclocal.m4
The local m4 configuration has not been built in a long time.
Trying to rebuild it results in failures with dmapi library
detection and libxfs.h detection macros. Fix the errors and rebuild
it.
Note: the new format for the aclocal.m4 file uses m4_include
directives which means it is much smaller than before and won't
need updating unless we include new m4 macro files.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 26 Jan 2010 03:26:42 +0000 (14:26 +1100)]
xfstests: deprecate busted log printing tests
Tests 018, 081 and 082 read the contents of the log and assume that the
contents will always be the same. They are trying to ensure that the
contents of the log don't change for a given fixed load.
This has several problems - high level changes to the filesystem and
VFS code can change the order and contents of the log. Changes to
the way we sync the filesystem will change the contents of the log.
background writeback occurring in the middle of the test will change
the contents of the log by allowing the tail to move. Even changes
to the default mkfs parameters can break them!
The tests also assume that unmount leaves a dirty log behind. We've
fixed lots of problems in sync and the unmount paths over recent
times, so now a clean unmount leaves a clean log behind. That is,
there is nothing left in the log print output for these tests to
check. IOWs, major surgery is required for these tests to be
returned to their former break-when-something-changes behaviour.
However, these tests are a maintenance nightmare. They spend more
time broken and failing than they do passing, and then it's not long
before they get broken again. They have to cover all sorts of
different permutations of log configurations and that will continue
to grow and increase the complexity of making these tests continue
to work. And to top it all off, I can't remember a bug actually ever
being found by these tests. Hence I think we should just stop using
them altogether.
So this patch deprecates 018, 081 and 082 rather than fixes them.
It introduces a "deprecated" test group and puts them in it. That
means the tests can still be run on older systems where they may
have some use, but will not be run automatically any more, nor
will any attempt be made to keep them up to date or working.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Wed, 20 Jan 2010 21:54:38 +0000 (08:54 +1100)]
xfstests: kill in 131 needs to be quiet
If the kill fails because the processes have already terminates,
the output from kill (no such process) will cause the test to
fail. Make sure kill doesn't output anything.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Wed, 20 Jan 2010 21:53:53 +0000 (08:53 +1100)]
xfstests: Don't use tempnam in growfiles.c
tempnam() generates a compiler warning as a dangerous function.
This code doesn't care about security issues with tempnam, so
remove it and just manually build the filenames without the
randomness of tempnam.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 19 Jan 2010 23:28:24 +0000 (10:28 +1100)]
xfsqa: Fix signal usage in aio-dio test code
Using signal() to set up signal handlers doesn't always do what you
want. A recent upgrade made test 208 fail because wait() was not
getting interrupted by a SIGALRM. Tracing showed that signal() was
being converted to a sigaction(SA_RESTART) handler, which allows
syscalls that return ERESTARTSYS to immediately restart without
returning EINTR to the calling process. The kernel code returns
ERESTARTSYS to signal interruptions while in wait().
Replace the use of signal with sigaction() to ensure that the
SA_RESTART flag is not set and the EINTR is delivered to the process
sitting in wait(). This makes test 208 terminate at 200s again.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 19 Jan 2010 23:27:08 +0000 (10:27 +1100)]
xfstests: Convert all tests to use /bin/bash
While most tests use /bin/sh, they are dependent on /bin/sh being a
bash shell. Convert all the tests to execute via /bin/bash as it is
much, much simpler than trying to debug and remove all the bashisms
throughout the test code.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Dave Chinner [Tue, 19 Jan 2010 23:26:10 +0000 (10:26 +1100)]
xfstests: Use common rules to build binaries
Remove explicit target-per-file rules in the subdirectories being
built and replace them with target based rules using the buildrules
hooks for doing this. This results in the makefiles being simpler,
smaller and more consistent. It does mean that all binaries are
linked against the common set of libraries but this does not cause
any harm.
This patch does not address the dmapi subdirectory of the xfstests
build system.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 19 Jan 2010 23:26:01 +0000 (10:26 +1100)]
xfstests: Clean up build output
We don't need to see every compiler command line for every file that
is compiled. This makes it hard to see warnings and errors during
compile. For progress notification, we really only need to see the
directory/file being operated on.
Turn down the verbosity of output by suppressing various make output
and provide better overall visibility of which directory is being
operated on, what the operation is and what is being done to the
files by the build/clean process.
This patch only addresses the top level makefile and build rules; it
does not clean up the subdirectories - these will be addressed by a
followup patch. It also does not touch the dmapi section of the
xfstests build system.
The old style verbose builds can still be run via "make V=1 ..."
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Alex Elder [Tue, 19 Jan 2010 20:26:16 +0000 (14:26 -0600)]
xfstests 219: ignore duplicates reported by repquota
Arrange to ignore duplicate entries reported by the repquota command.
This can happen if an id is used more than once (such as when two user
names are assigned the same uid).
Do this here by simply dropping any reported entries whose id number
has already been seen in the output.
Signed-off-by: Alex Elder <aelder@sgi.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Alex Elder [Tue, 19 Jan 2010 20:26:20 +0000 (14:26 -0600)]
xfstests 050: ignore duplicates reported by repquota
Arrange to ignore duplicate entries reported by the "repquota"
xfs_quota command. This can happen if an id is used more than once
(such as when two user names are assigned the same uid).
Since we are already sorting the output we can accomplish this by
adding the "-u" (unique) option to the sort command. Shorten some
long lines while we're at it.
Signed-off-by: Alex Elder <aelder@sgi.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Eric Sandeen [Sun, 17 Jan 2010 16:35:23 +0000 (10:35 -0600)]
xfstests 073: unmount source image in cleanup
Test 073 is currently failing w/ xfs_repair problems, and when
it does, the whole series stops because the test dir still has
a loopback mounted filesystem on it, and it can't unmount.
Fix that.
Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 12 Jan 2010 21:51:47 +0000 (22:51 +0100)]
add simple xfs_fsr test
We have very little fsr test coverage in the test suite right now.
042 is the only test that runs it, and it only defragments a single
file.
Improve the coverage by simply running xfs_fsr of the test
filesystem. This gives xfs_fsr a wide variety of inodes to try to
defragment. A simple test like this would have tripped over the
attr2 fork offset problems in the swap extents ioctl long ago
thanks to the varied nature of inodes the test suite leaves around
in this filesystem.
Signed-off-by: Dave Chinner <david@fromorbit.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Tue, 12 Jan 2010 21:49:57 +0000 (22:49 +0100)]
filter log size in 206
With the log size now going past 128MB, test 206 creates a filesystem
with a much larger log than 128MB. The golden output expects at
128MB log size, so the test breaks. Filter the log size out of the
mkfs output and golden output so that the test works on both older and
newer mkfs binaries.
Signed-off-by: Dave Chinner <david@fromorbit.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
By mounting the filesystem with quotas enabled in 087 we exercise the
replay of the quota log items in this tests, which we otherwise don't
test in xfstests. OSS bugzilla #855 is an example of something caught
by this.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <david@fromorbit.com>
Eric Sandeen [Wed, 6 Jan 2010 19:00:18 +0000 (13:00 -0600)]
[PATCH] xfstests: add another fallocate test to 214
ext4 had a regression where it double-accounted used blocks
if you fallocated on top of delalloc blocks. Ted sent a
c program to exploit it (see "fsstress-induced corruption reproduced"
on linux-ext4 on 12/31/2009) and it's trivial to do the same thing
within the xfstests framework using xfs_io.
This also changes the handcrafted xfs_io tests to use the
_require_xfs_io_falloc helper, not sure how that got missed.
Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Eric Sandeen [Thu, 3 Dec 2009 16:30:11 +0000 (10:30 -0600)]
commit bd53c922 added src/aio-dio-regress to the top-level
Makefile so that it would get caught on make install, but
doing this unconditionally means that we tried to build that
subdir even if libaio headers weren't found.
Fix this up to do a properly chained make install through
the subdirs...
Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Christoph Hellwig <hch@lst.de>
Trivial test case that mounts a filesystem with user quotas, then turns
quotas off an unmounts. Based on a testcase in a bug report from
Utako Kusaka <u-kusaka@wm.jp.nec.com>.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Alex Elder <aelder@sgi.com>
Jaroslav Barton [Wed, 16 Sep 2009 12:55:59 +0000 (09:55 -0300)]
_check_generic_filesystem support for fsck parameters
_check_generic_filesystem now support fsck parameters. Fsck parameters
can be passed by FSCK_OPTIONS environmental variable. Default values
will be used if FSCK_OPTIONS is missing or empty (same mechanism as
MOUNT_OPTIONS and MKFS_OPTIONS).
_check_extN_filesystem function is no longer needed, extN filesystem are
properly handled by default values in _check_generic_filesystem.
Theodore Ts'o [Thu, 27 Aug 2009 21:17:32 +0000 (18:17 -0300)]
add ext2/3/4-specific _check_extN_filesystem function
The _check_generic_filesystem function doesn't force a full filesystem
check, so filesystem inconsistencies after a test wouldn't be noticed.
To fix this, I added an extN specific check filesystem function.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Signed-off-by: Christoph Hellwig <hch@lst.de>
Eric Sandeen [Mon, 10 Aug 2009 18:10:26 +0000 (13:10 -0500)]
Test a few very basic defragmentation scenarios.
This test creates some files, runs defrag on them,
and compares the before/after fragmentation as well
as file md5sums and timestamps.
The test currently expects to find e4defrag in
/usr/bin
It should be relatively easy to add more interestingly
fragmented files to the tests, as well as to test
that memory-mapped files aren't touched, etc -
but this gives us a framework.
V2: remount before checking file contents, and create
common.defrag.
Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Christoph Hellwig <hch@lst.de>
Christian Kujau [Mon, 20 Jul 2009 09:31:39 +0000 (11:31 +0200)]
don't fail if nslookup is not found
When nslookup is not available, execution would stop. This patch moves the
check for nslookup and the _get_fqdn function into the ./new script (since
this is the only place where it's used) and we don't fail any more but try
to find out our FQDN without nslookup.
Signed-off-by: Christian Kujau <lists@nerdbynature.de> Signed-off-by: Christoph Hellwig <hch@lst.de>
projects / xfstests-dev.git / log