Re: [PATCH 2/3] cifs: use CreationTime like an i_generation field

[Jeremy Allison <jra-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org>]

On Mon, Nov 22, 2010 at 03:31:25PM -0500, Jeff Layton wrote:
> On Mon, 22 Nov 2010 14:22:50 -0600
> Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> 
> > On Mon, Nov 22, 2010 at 1:55 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > 
> > > On Mon, 28 Jun 2010 07:10:12 -0400
> > > Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > >
> > >
> > > Steve, can you clarify where we are with this patch? I originally sent
> > > this back in June. You and I discussed this at the CIFS plugfest and I
> > > had thought you had decided to merge it. It's still not in your tree
> > > however. Can you give a verdict on whether you're going to merge it and
> > > let me know when I can expect to see it in your tree if so?
> > >
> > >
> > Thanks for reminding me.  I had forgotten to ask JRA about a question
> > on legacy Samba behavior.   I remembered that older Samba
> > had problems with creation time - and was trying to remember what
> > the implication of using this patch with older Samba was.
> > 
> 
> All samba versions have issues with create times (at least on Linux
> anyway). They fake them, by using the ctime (which seems like an odd
> choice to me) in most cases.

New Samba's store create time in an xattr, so they can
return a reasonably valid create time value.

Jeremy.