Re: [RFC 12/26] ext2 white-out support

[Dave Kleikamp <shaggy@linux.vnet.ibm.com>]

On Tue, 2007-07-31 at 13:11 -0400, Josef Sipek wrote:
> On Tue, Jul 31, 2007 at 07:00:12PM +0200, Jan Blunck wrote:
> > On Tue, Jul 31, Josef Sipek wrote:
> > 
> > > On Mon, Jul 30, 2007 at 06:13:35PM +0200, Jan Blunck wrote:
> > > > Introduce white-out support to ext2.
> > > 
> > > I think storing whiteouts on the branches is wrong. It creates all sort of
> > > nasty cases when people actually try to use unioning. Imagine a (no-so
> > > unlikely) scenario where you have 2 unions, and they share a branch. If you
> > > create a whiteout in one union on that shared branch, the whiteout magically
> > > affects the other union as well! Whiteouts are a union-level construct, and
> > > therefore storing them at the branch level is wrong.
> > 
> > So you think that just because you mounted the filesystem somewhere else it
> > should look different? This is what sharing is all about. If you share a
> > filesystem you also share the removal of objects.
> 
> The removal happens at the union level, not the branch level. Say you have:
> 
> /a/
> /b/foo
> /c/foo
> 
> And you mount /u1 as a union of {a,b}, and /u2 as union of {a,c}.

Who does this?  I'm assuming that a is the "top" layer.  Aren't union
mounts typically about sharing lower layers and having a separate rw
layer for each union mount?

> $ find /u*
> /u1
> /u1/foo
> /u2
> /u2/foo
> $ rm /u1/foo # this creates whiteout for "foo" in /a
> $ find /u*
> /u1
> /u2
> 
> Is that what you'd expect as a user? I don't think so.

That's exactly what I would expect.

If I were to:
$ echo "this is new" > /u1/foo

I would expect:
$ cat /u2/foo
this is new

So why should rm behave differently?

I haven't really been tuned into union mounts, so maybe I'm missing out
on something basic here.

Thanks,
Shaggy
-- 
David Kleikamp
IBM Linux Technology Center