Re: [PATCH] btrfs: fix file truncation if FALLOC_FL_KEEP_SIZE is specified

[Jeff Liu <jeff.liu@oracle.com>]

On 06/29/2013 10:22 AM, Dave Chinner wrote:

> On Fri, Jun 28, 2013 at 08:41:00AM -0400, Josef Bacik wrote:
>> On Fri, Jun 28, 2013 at 01:15:52PM +0800, Jeff Liu wrote:
>>> From: Jie Liu <jeff.liu@oracle.com>
>>>
>>> Create a small file and fallocate it to a big size with
>>> FALLOC_FL_KEEP_SIZE option, then truncate it back to the
>>> small size again, the disk free space is not changed back
>>> in this case. i.e,
>>>
>>> # dd if=/dev/zero of=/mnt/test bs=512 count=1
>>> # ls -l /mnt
>>> total 4
>>> -rw-r--r-- 1 root root 512 Jun 28 11:35 test
>>>
>>> # df -h
>>> Filesystem      Size  Used Avail Use% Mounted on
>>> ....
>>> /dev/sdb1       8.0G   56K  7.2G   1% /mnt
>>>
>>> # xfs_io -c 'falloc -k 512 5G' /mnt/test 
>>> # ls -l /mnt/test
>>> -rw-r--r-- 1 root root 512 Jun 28 11:35 /mnt/test
>>>
>>> # sync; df -h
>>> Filesystem      Size  Used Avail Use% Mounted on
>>> ....
>>> /dev/sdb1       8.0G  5.1G  2.2G  70% /mnt
>>>
>>> # xfs_io -c 'truncate 512' /mnt/test
>>> # sync; df -h
>>> Filesystem      Size  Used Avail Use% Mounted on
>>> ....
>>> /dev/sdb1       8.0G  5.1G  2.2G  70% /mnt
>>>
>>> With this fix, the truncated up space is back as:
>>> # sync; df -h
>>> Filesystem      Size  Used Avail Use% Mounted on
>>> ....
>>> /dev/sdb1       8.0G   56K  7.2G   1% /mnt
>>>
>>> Signed-off-by: Jie Liu <jeff.liu@oracle.com>
>>> ---
>>>  fs/btrfs/inode.c |    3 ---
>>>  1 file changed, 3 deletions(-)
>>>
>>> diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
>>> index 4f9d16b..7e1a5ff 100644
>>> --- a/fs/btrfs/inode.c
>>> +++ b/fs/btrfs/inode.c
>>> @@ -4509,9 +4509,6 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr)
>>>  	int mask = attr->ia_valid;
>>>  	int ret;
>>>  
>>> -	if (newsize == oldsize)
>>> -		return 0;
>>> -
>>>  	/*
>>>  	 * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
>>>  	 * special case where we need to update the times despite not having
>>
>> Cc'ing a few people on this since I'd like their opinion.  Looking at other fs's
>> it looks like ext4 does the same thing we do and would leave the prealloc'ed
>> space, but it appears that xfs will truncate it.  What do we think is the
>> correct behavior? 
> 
> XFS has had this truncate behaviour since at least the start of the
> git tree history (2005). Given that these fallocate()
> prealloc-blocks-beyond-EOF behaviours are modelled on what XFS has
> historically provided, I think y'all can see what i think should be
> done...
> 
>> I'm inclined to take this patch, but I'd like to have an
>> xfstest made for it so other file systems can be made to be consistent, and I'd
>> like to make sure we all agree what is the correct behavior before we wander
>> down that road.  Thanks,
> 
> I couldn't have said it better myself. Jeff, can you take care of
> this, please?

Yes, I'll take care of this.

Thanks,
-Jeff