Re: [nfsv4] No WG meeting for IETF 91

"nfsv4" <nfsv4-bounces@ietf.org> wrote on 10/04/2014 12:03:39 PM:

> From: "J. Bruce Fields" <bfields@fieldses.org>
> To: Chuck Lever <chuck.lever@oracle.com>
> Cc: Christoph Hellwig <hch@lst.de>, "nfsv4@ietf.org" <nfsv4@ietf.org>
> Date: 10/04/2014 12:03 PM
> Subject: Re: [nfsv4] No WG meeting for IETF 91
> Sent by: "nfsv4" <nfsv4-bounces@ietf.org>
>
> On Sat, Oct 04, 2014 at 12:04:54PM -0400, Chuck Lever wrote:
> >
> > On Oct 4, 2014, at 3:11 AM, Christoph Hellwig <hch@lst.de> wrote:
> >
> > > The current version of draft-naik-nfsv4-xattrs-01 totally ignores
> > > the issue of access control to different namespaces.  To be actually
> > > interoperable between Linux, FreeBSD, MacOS and Window (and more
> theoretically
> > > OS/2 and IRIX) it should be defined to operate on only what Linux
calls
> > > "user." xattrs, FreeBSD calls EXTATTR_NAMESPACE_USER extattrs, IRIX
calls
> > > user namespace attrs, and the the only one available to the rest.
> >
> > I thought we had already agreed that only the “user.” namespace would
> > be exposed via NFS. If that’s not in the draft specification, it really
> > needs to be added.
>
> Section 3 does say that the "user" namespace is the only one for which
> xattrs operations "can be considered interoperable across platforms and
> vendor implementations", so it seems strange if they're trying to keep
> the others.

Definitely not trying to keep the others. But, based on what I thought was
a request, leaving open the option of allowing xattrs from other namespaces
to be added in the future if they could be made interoperable.

>
> I think they were trying to keep NFS xattr names the same as local
> attribute names instead of requiring implementations to strip the
> "user.", and also maybe hoping to leave open the possibility of future
> standardization of behavior in other namespaces without requiring new
> xdr.  But I think all the feedback has been that that's a quagmire.

Yes.

>
> So I'm not sure what the appropriate language is.  Maybe something
> roughly like:
>
>    Some operating systems define multiple "namespaces" for xattrs,
>    with the system enforcing different policies for different
>    xattrs, and giving some xattrs special meaning to the system.
>    (For example, some systems use xattr operations to set and query
>    ACLs.)
>
>    NFS extended attributes are not organized into namespaces and do
>    not have special meaning to the server.  Several operating
>    systems that define multiple namespaces have a "user" namespace
>    with these semantics.  Such systems will want to keep NFS
>    extended attributes under the "user" namespace by translating
>    any NFS xattr name "X" to a local extended attribute name
>    "user.X".

I can go with this.

Manoj.

>
> --b.
>
> _______________________________________________
> nfsv4 mailing list
> nfsv4@ietf.org
> https://www.ietf.org/mailman/listinfo/nfsv4