On Oct 29, 2009, at 12:24 AM, Christoph Hellwig wrote:
On Wed, Oct 28, 2009 at 03:40:12PM -0400, Jeff Layton wrote:
Solaris uses netids as values for the proto= option, so that when
someone specifies "tcp6" they get traffic over TCP + IPv6. Until
recently, this has never really been an issue for Linux since it
didn't
support NFS over IPv6. The netid and the protocol name were
generally
always the same (modulo any strange configuration in /etc/netconfig).
The solaris manpage documents their proto= option as:
proto= _netid_ | rdma
I'm strictly against adding netid braindamage to Linux. Just because
Solaris made a mistake we don't have to repeat it.
As it stands, the kernel already has some support for netids, which is
required for supporting NFS and RPC on IPv6 such that it can
interoperate with the reference NFS implementation. We might be able
to address your concern if we understood specifically what was at issue.
A feature request has been made to provide a mechanism to force the
use of IPv4 or IPv6 for mounting NFS servers that have both types of
address. The use of "proto=netid" is especially preferred for
mount.nfs since Linux often shares automounter maps with other NFS
implementations. And, netids are an internet standard at this point.
We could choose a different mechanism for the kernel's mount option
parser, but I haven't seen an alternative proposal that is cleaner
overall than what we have here.
Adding a hack to mount.nfs to parse it is fine, but there's not need
to
display it again.
That was our first pass, but as Jeff says, it's just bad user
interface design for the mount options in /proc/mounts to have
different semantics than what admins use in user space. I would
rather keep mount.nfs and the kernel's mount parser as close together
as possible.
--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
