On 9/17/13 10:58 AM, Mark Tinguely wrote:
> On 09/17/13 10:41, Eric Sandeen wrote:
>> On 9/17/13 9:59 AM, Mark Tinguely wrote:
>>> This tests triggers an assert in the XFS directory unbalance code.
>>> This test originally written by Brian Foster and suggestions
>>> from Micheal Semon.
>> cool, thanks. Comments below.
>>> Signed-off-by: Mark Tinguely<tinguely@xxxxxxx>
>>> tests/generic/319 | 62
>>> tests/generic/319.out | 2 +
>>> tests/generic/group | 1
>>> 3 files changed, 65 insertions(+)
>>> Index: b/tests/generic/319
>>> --- /dev/null
>>> +++ b/tests/generic/319
>>> @@ -0,0 +1,62 @@
>>> +#! /bin/bash
>>> +# FS QA Test No. 319
>>> +# Test directory code correctly handles fsstress filling the filesystem
>>> +# Copyright (c) 2013 SGI. All Rights Reserved.
>>> +# This program is free software; you can redistribute it and/or
>>> +# modify it under the terms of the GNU General Public License as
>>> +# published by the Free Software Foundation.
>>> +# This program is distributed in the hope that it would be useful,
>>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>>> +# GNU General Public License for more details.
>>> +# You should have received a copy of the GNU General Public License
>>> +# along with this program; if not, write the Free Software Foundation,
>>> +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
>>> +seq=`basename $0`
>>> +echo "QA output created by $seq"
>>> +status=1 # failure is the default!
>>> +trap "_cleanup; exit \$status" 0 1 2 3 15
>>> + cd /
>> That seems pointless; usually it's done w/ rm -f $tmp.*
>> right after, but we have no tmpfile, so...
> Yeah, no cleanup is needed.
>>> +# get standard environment, filters and checks
>>> +. ./common/rc
>>> +. ./common/filter
>>> +# real QA test starts here
>>> +_supported_fs generic
>>> +_supported_os IRIX Linux
>>> +_scratch_unmount> /dev/null 2>&1
>> I see this done both ways - is it required to unmount scratch at the
>> of a test? I don't think so (I know it's done in many tests, though, but
>> again, C&P& cargo cult? Or not? I'm not sure :( )
>> I guess it doesn't hurt, but at some point I'd like to get it straight
>> about who's required to umount scratch, and when (if at all).
> Have to unmount for the mkfs, as noted by Eryu, it is already done. I
> would rather manually unmount it than be surprised when someone
> changes the common files.
If that happens, tons of tests will break. I'd really just remove it
for clarity, but *shrug*
>>> +_scratch_mkfs_sized 11g>> $seqres.full 2>&1
>> _scratch_mkfs_sized doesn't take units like this ('g'), so the above fails to
>> actually make an 11g fs:
>> # Create fs of certain size on scratch device
>> # _scratch_mkfs_sized<size in bytes> [optional blocksize]
>> so we get this in 319.full:
>> expr: non-numeric argument
>> ./common/rc: line 576: [: 11g: integer expression expected
>> but then it seems like mkfs carries on anyway w/ defaults. :(
>> Apparently the mkfs 11g part isn't actually critical? ;)
> it works on xfs because mkfs.xfs size can take those values, but yes it
> breaks on other filesystem. my bad.
One other nitpick in this area, please remove $seqres.full before you start so
it doesn't grow each time the test is run.
(hm maybe we should add that to ./check or something; so many tests miss this)
But - no, it doesn't work for xfs either, at least not in all cases, because
it doesn't do the device size check. Here's xfs output on a < 11G device:
expr: non-numeric argument
./common/rc: line 582: [: 11g: integer expression expected
** mkfs failed with extra mkfs options added to "-bsize=4096" by test 319 **
** attempting to mkfs using only test 319 options: -d size=11g -b size=4096 **
size 11g specified for data subvolume is too large, maximum is 1048241 blocks
/* blocksize */ [-b log=n|size=num]
and w/o error checking (2>&1 and no || _fail) the test just carries on w/o a
fresh mkfs, on whatever size it happens to be.