Re: [PATCH] xfstests: add _require_freeze and minor cleanups

[Eric Sandeen <sandeen@redhat.com>]

On 9/21/12 11:38 AM, Ben Myers wrote:
> Hey Eric,
> 
> On Wed, Sep 19, 2012 at 05:53:56PM -0500, Eric Sandeen wrote:
>> More filesystems have grown freeze capability, so rather than
>> hardcoding several in _supported_fs, make tests 068 and 280
>> generic and then add a new _require_freeze() which checks whether
>> the fs under test can be frozen before beginning the test.
>>
>> Minor other cleanups to 280:
>> - remove extra _supported_fs line
>> - clear $seq.full before beginning
>>
>> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
>> ---
>>
>> diff --git a/068 b/068
>> index b595d1d..617420c 100755
>> --- a/068
>> +++ b/068
>> @@ -51,10 +51,11 @@ trap "_cleanup" 0 1 2 3 15
>>  . ./common.filter
>>  
>>  # real QA test starts here
>> -_supported_fs btrfs ext3 ext4 xfs
>> +_supported_fs generic
>>  _supported_os Linux IRIX
>>  
>>  _require_scratch
>> +_require_freeze
>>  
>>  echo "*** init FS"
>>  
>> diff --git a/280 b/280
>> index 55849ed..5e26173 100755
>> --- a/280
>> +++ b/280
>> @@ -45,13 +45,15 @@ _cleanup()
>>  
>>  _require_scratch
>>  _require_quota
>> +_require_freeze
>>  
>>  # real QA test starts here
>>  
>>  # Modify as appropriate.
>> -_supported_fs generic
>>  _supported_os Linux
>> -_supported_fs ext3 ext4 xfs
>> +_supported_fs generic
>> +
>> +rm -f $seq.full
>>  
>>  umount $SCRATCH_DEV 2>/dev/null
>>  _scratch_mkfs >> $seq.full 2>&1
>> diff --git a/common.rc b/common.rc
>> index 602513a..0e8a306 100644
>> --- a/common.rc
>> +++ b/common.rc
>> @@ -1758,6 +1758,15 @@ _require_btrfs()
>>  	[ $? -eq 0 ] || _notrun "$BTRFS_UTIL_PROG too old (must support $cmd)"
>>  }
>>  
>> +# Does freeze work on this fs?
>> +_require_freeze()
>> +{
>> +	xfs_freeze -f "$TEST_DIR" >/dev/null 2>&1
>> +	result=$? 
>> +	xfs_freeze -u "$TEST_DIR" >/dev/null 2>&1
>> +	[ $result -eq 0 ] || _notrun "$FSTYP does not support freezing"
>> +}
>> +
> 
> Pretty good idea to generalize _require_freeze.  It looks like xfs_freeze is a
> script that uses xfs_io which uses xfsctl XFS_IOC_FREEZE.  So isn't what you
> have here xfs specific?  It wouldn't work for the other filesystems that
> implement s_op.freeze_fs:

It got elevated to a generic ioctl:

fs/xfs/xfs_fs.h:
/*      XFS_IOC_FREEZE            -- FIFREEZE   119      */
/*      XFS_IOC_THAW              -- FITHAW     120      */

to:

include/linux/fs.h:
#define FIFREEZE        _IOWR('X', 119, int)    /* Freeze */
#define FITHAW          _IOWR('X', 120, int)    /* Thaw */

-Eric

>    1   1502  btrfs/super.c <<GLOBAL>>
>              .freeze_fs = btrfs_freeze,
>    2    804  ext3/super.c <<GLOBAL>>
>              .freeze_fs = ext3_freeze,
>    3   1238  ext4/super.c <<GLOBAL>>
>              .freeze_fs = ext4_freeze,
>    4   1578  gfs2/super.c <<GLOBAL>>
>              .freeze_fs = gfs2_freeze,
>    5    760  jfs/super.c <<GLOBAL>>
>              .freeze_fs = jfs_freeze,
>    6    688  nilfs2/super.c <<GLOBAL>>
>              .freeze_fs = nilfs_freeze,
>    7    620  reiserfs/super.c <<GLOBAL>>
>              .freeze_fs = reiserfs_freeze,
>    8   1536  xfs/linux-2.6/xfs_super.c <<GLOBAL>>
>              .freeze_fs = xfs_fs_freeze,
> 
> Maybe it would be better if we had some kind of interface to test whether
> .freeze_fs is defined (if there isn't one already) rather than freeze and thaw
> to find out.  Oddly freeze_super seems to just return 0 when .freeze_fs is not
> defined.
> 
> Regards,
> Ben
> 

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs