On Tue, Oct 23, 2012 at 08:30:44AM -0400, Christoph Hellwig wrote:
> On Thu, Oct 18, 2012 at 04:11:17PM +1100, Dave Chinner wrote:
> > So, I was bored a few days ago, and I was sick of having to run
> > xfs_db incorrectly report free space extents when the filesytem is
> > mounted, so I decided to extend fiemap to export freespace mappings
> > to userspace so I could get the information coherently through the
> > mounted filesystem.
> > Yes, this could probably be considered interface abuse but, well, it
> > was simple to do because extent mapping is exactly what fiemap is
> > designed to do. Hence I didn't have to write new walkers/formatters
> > and I was using code I knew worked correctly.
> I think the right way to handle this is to introduce a new ioctl which
> uses the same structures. That way we have a reasonable interface,
> without issue like which file does it need to be called on because the
> VFS glue can turn it into a superblock op.
A VFS level ioctl or an XFS ioctl?
I thought about a new ioctl, but then what's the point of having an
extensible fiemap interface if we create new ioctls with an
identical interface for doing something that the existing ioctl is
perfectly capable of doing? I'd still need special flags to control
the ioctl behaviour even though it uses struct fiemap and plumbing,
so it seemed pointless to introduce a new ioctl....
As it is, the only reason fiemap doesn't work on directory ioctls
for XFS is that it hasn't been hooked up to directories. I can't see
anything in the fiemap VFS layers that prevents us from mapping
directories and we know the mapping code in XFS works on
directories. So that would the "what file" problem go away - any
file would do as long as the user has the permissions to run the
free space mapping command....