This is obviously wrong and makes ./check -r skip over tests on ext4
with "ext4 on $DEV not configured with metadata journaling".
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
As posix standard, if the file offset is at or past the end of file,
no bytes are read, and read() returns zero. There was a bug, when
DIO read offset is just past the EOF a little, but in the same block
with EOF, read returns different negative values.
Kernel commit 74cedf9b6c60 ("direct-io: Fix negative return from dio
read beyond eof") and commit 2d4594acbf6d ("fix the regression from
"direct-io: Fix negative return from dio read beyond eof"") fixed
the bug.
This case reads from range within EOF, past EOF and at EOF, to make
sure the return value as expected, especially read from past/at EOF
returns 0.
[eguan: update commit log and comments about information of the
specific bug, adjust read_test param order (offset, count, ret) and
test description]
Signed-off-by: Zorro Lang <zlang@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Some tests and common helpers don't properly clean up tmp files and
leave them behind in /tmp dir, and these tmp files are accumulating
over time.
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Some tests don't have x permission set in mode, git complains about
file mode change after running tests in fstests git repo dir. So
change all such tests to 755 mode. Performed by:
find tests ! -perm /111 -name [0-9][0-9][0-9] -exec chmod 755 {} \;
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
generic/224 is doing concurrent direct and buffered I/O to the same
set of files, and this triggers some expected warnings on XFS. So
filter out these warnings just like what we did in generic/095 and
generic/247.
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Setting acls on an xfs filesystem will succeed even after running
out of space for user attributes. Use trusted attributes instead.
Also speed up the test by setting large values for the attributes.
[eguan: use perl to generate attr value, and add comments on trusted
namespace]
Signed-off-by: Ernesto A. Fernández <ernesto.mnd.fernandez@gmail.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
The $dread_pid refers to the while-true-do loop, wait for $dread_pid
doesn't mean the xfs_io direct read process is already dead,
sometimes xfs_io process is still running and blocking
_scratch_unmount.
Fix it by making the direct read does a fixed number of loop and
break out the second mmap-fpunch loop if the first loop exits. At
this point we're sure that there's no unfinished background process
blocking the umount.
Reviewed-by: Xiao Yang <yangx.jy@cn.fujitsu.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
1) This pattern is repeated in several seek_data/hole tests
(e.g. generic/285, generic/436, generic/445 generic/448)
and generic/009. A common _ext4_disable_extent_zeroout()
helper could be added and applied by generic/009 and
_require_seek_data_hole().
2) On some old kernels(e.g. v3.1-v3.6), when vfs recognizes
SEEK_DATA/HOLE flag && ext4 has no extent zeroout tunable
in sysfs, these cases may trigger "sysfs entry not found"
issue. We can add check if extent_max_zeroout_kb exists
on ext4 filesystem.
The extent_max_zeroout_kb is introduced by:
'67a5da564f97 ("ext4: make the zero-out chunk size tunable")'
3) Declare several vars as local in _require_seek_data_hole().
Signed-off-by: Xiao Yang <yangx.jy@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Some tests had separate output files for IRIX and Linux. Now that the
IRIX ones have been removed, rename the Linux ones and remove the
now-unneeded .cfg files and calls to _link_out_file.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Many tests claimed (via _supported_os) to work on both Linux and IRIX.
Since IRIX is no longer supported by xfstests, update these to claim
Linux support only. Then remove any obvious IRIX-specific logic in the
tests, and any IRIX-specific golden output files.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
The ACL test shared/051 was very similar to generic/099 which was not
being run and was just removed; most likely the script was copy+pasted
at some point. Since shared/051 has been getting maintained+run and is
not really XFS and UDF-specific, move it to generic, reusing the old
number of 099. One change was required for it to work on other
filesystems: the output of 'find' must be sorted.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
This IRIX-specific ACL test was nearly identical to shared/051, which
has been better maintained and will be made a generic test in the next
commit. Therefore, remove the existing generic/099.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
This IRIX-specific test did some basic testing of extended attributes.
Port it to Linux; this mainly involved updating it to use the 'getfattr'
and 'setfattr' programs instead 'attr'. Note that although 'attr' is
available on Linux, it's mainly for IRIX compatibility, the man page
recommends against using it on non-XFS filesystems, and it doesn't
support listing user xattrs only. (In the last point it actually
differs from IRIX 'attr', but probably no one cares anymore.) getfattr
also sorts its output by xattr name, so its output will be the same on
all filesystems unlike 'attr -l'.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
This IRIX-specific test mainly tested whether a file's capabilities are
cleared when it is written to. Port the test to the Linux libcap tools
and update it to expect the Linux semantics which are a little simpler:
capabilities are always cleared even if the program is root (or has
CAP_FSETID). The test also tests that chmod doesn't affect open file
descriptors; this is mostly unrelated, but keep it in for now.
[eguan: add _require_test_program rule for src/writemod]
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
generic/192 would sleep 40 seconds, update a file's atime, and then
fail if the atime was not exactly 40 seconds later. This is
unreliable since things may be slow enough to cause an extra second
to elapse. "Fix" this by allowing for 2 seconds of delay. Also,
while we're at it shorten the sleep to a much more reasonable 5
seconds.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
On some buggy kernels, the 'find' command in generic/421 encounters
a "Permission denied" error when trying to search $dir, so it
doesn't find the file it's supposed to. This causes 'cat' to read
from stdin, hanging the test. Quote the argument to cat to make the
test fail right away rather than hanging.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Currently we just have this test run on a whitelist of filesystems,
but it would be best to be able to run it on all of them. The
problem is that a lot of filesystems basically shut down once they
hit metadata errors.
Allow the fsync-err testcase to operate in two different modes. One
mode just does basic testing to ensure that we get an error back on
all fd's when we fsync. The other does a more thorough test to
ensure that we get back 0 on subsequent fsyncs when there hasn't
been any write activity.
For now, we just opt-in to the more thorough testing on certain
filesystems: xfs, ext3 and ext4 on the generic test. All other
filesystems will run in simple mode.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Check that the group permission bits of a file are not altered when
setfacl fails. At the time of this patch the test fails for at least
ext2, ext4, jfs and btrfs.
The failure in setfacl is induced by filling the device and setting
as many user attributes to the file as possible. For xfs this is not
enough, and the test will not run.
[eguan replaced setfattr with $SETFATTR_PROG and added enospc group]
Signed-off-by: Ernesto A. Fernández <ernesto.mnd.fernandez@gmail.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Check that we get -ENXIO if the user calls SEEK_HOLE/SEEK_DATA with
a negative file offset.
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>
Test a regression in XFS where we blow out a transaction reservation if
we create a big file, share every other block, and delete the first
file. There's nothing particularly fs-specific about this stress test,
so put it in generic.
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>
In commit d55123c080 ("generic/247: filter out expected XFS warnings
for mixed mmap/direct I/O"), _scratch_unmount was removed from
generic/095 and added to generic/247. But actually generic/095 *should*
be unmounting SCRATCH_DEV; and generic/247 should be unmounting
TEST_DEV, not SCRATCH_DEV, since it doesn't use a scratch device (it was
failing if SCRATCH_DEV was not defined). Fix it.
[eguan: the original bug was introduced by me not Brian, as I
removed _scratch_unmount from wrong test]
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
This test reproduces a race between a direct I/O read and
a mapped write to a hole in a file. On xfs filesystem, it
will trigger a BUG_ON(), and this XFS bug has been fixed by:
04197b3 ("xfs: don't BUG() on mixed direct and mapped I/O")
[ eguan: umount before check dmesg log, and add rw group ]
Signed-off-by: Xiao Yang <yangx.jy@cn.fujitsu.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
generic/247 reproduces some of the same, expected warnings from XFS
as generic/095. These warnings occur due to mixed buffered/mapped
I/O racing with direct I/O to the same file.
generic/095 contains a custom dmesg filter to prevent test failure
in the event of such warnings. Lift the helper from generic/095 to
common/xfs and reuse it in generic/247 to implement the same
behavior.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
The cmp command was using the wrong file to perform the comparison.
Use $testdir/beyond instead of $testdir/end.
Signed-off-by: Luis Henriques <lhenriques@suse.com>
Reviewed-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
Ilya Dryomov
Zorro Lang
Ernesto A. Fernández
Eryu Guan
Xiao Yang
Eric Biggers
Jeff Layton
Darrick J. Wong
Brian Foster
Luis Henriques