[PATCH] Introduce SEEK_DATA/SEEK_HOLE support v9

[Mark Tinguely]

On 05/10/12 15:08, Ben Myers wrote:
> On Thu, May 10, 2012 at 02:51:15PM -0500, Mark Tinguely wrote:
>> On 05/10/12 08:29, Jeff Liu wrote:
>>> Hello,
>>>
>>> This is the revised patch according to Mark's comments.
>>> Is it possible to get it to accept if there are no other issues,
>>> so that I can continue to work on the unwritten part based on that.
>>>
>>> Thanks,
>>> -Jeff
>>>
>>> Signed-off-by: Jie Liu<jeff.liu at oracle.com>
>>>
>>> ---
>>
>>> +
>>> +STATIC loff_t
>>> +xfs_seek_hole(
>>> +	struct file		*file,
>>> +	loff_t			start,
>>> +	u32			type)
>>> +{
>>> +	struct inode		*inode = file->f_mapping->host;
>>> +	struct xfs_inode	*ip = XFS_I(inode);
>>> +	struct xfs_mount	*mp = ip->i_mount;
>>> +	loff_t			uninitialized_var(offset);
>>> +	loff_t			holeoff;
>>> +	xfs_fsize_t		isize;
>>> +	xfs_fileoff_t		fsbno;
>>> +	uint			lock;
>>> +	int			error;
>>> +
>>> +	if (XFS_FORCED_SHUTDOWN(mp))
>>> +		return -XFS_ERROR(EIO);
>>
>> I think xfs_bmapi_read() returns just a XFS_ERROR(EIO), not negative.
>> Ben can fix up if you agree.
>
> Looks like error is negated at the bottom of the function and is not a problem.
>

Yes, but xfs_seek_data() also negates at the error at the end of the 
routine. They should be consistent.

--Mark.