On Tue, Jun 19, 2018 at 03:06:06PM -0500, Eric Blake wrote:
> On 06/15/2018 10:50 AM, Daniel P. Berrangé wrote:
> > From: "Daniel P. Berrange" <berra...@redhat.com>
> >
> > Currently any client which can complete the TLS handshake is able to use
> > the NBD server. The server admin can turn on the 'verify-peer' option
> > for the x509 creds to require the client to provide a x509 certificate.
> > This means the client will have to acquire a certificate from the CA
> > before they are permitted to use the NBD server. This is still a fairly
> > low bar to cross.
> >
> > This adds a '--tls-authz OBJECT-ID' option to the qemu-nbd command which
> > takes the ID of a previously added 'QAuthZ' object instance. This will
> > be used to validate the client's x509 distinguished name. Clients
> > failing the authorization check will not be permitted to use the NBD
> > server.
> >
> > For example to setup authorization that only allows connection from a client
> > whose x509 certificate distinguished name contains 'CN=fred', you would
> > use:
> >
> > qemu-nbd -object tls-creds-x509,id=tls0,dir=/home/berrange/qemutls,\
> > endpoint=server,verify-peer=yes \
> > -object authz-simple,id=authz0,policy=deny,\
> > rules.0.match=*CN=fred,rules.0.policy=allow \
>
> s/-object/--object/g
>
> > -tls-creds tls0 \
> > -tls-authz authz0
>
> s/-tls/--tls/g
>
> (qemu-nbd requires double-dash long-opts, -o means --offset except that
> 'bject' is not an offset; similarly for -t meaning --persistent)
Sigh, yes, just another reason for us to standardize on -- everywhere
> > +++ b/qemu-nbd.c
>
> > @@ -533,6 +535,7 @@ int main(int argc, char **argv)
> > { "image-opts", no_argument, NULL, QEMU_NBD_OPT_IMAGE_OPTS },
> > { "trace", required_argument, NULL, 'T' },
> > { "fork", no_argument, NULL, QEMU_NBD_OPT_FORK },
> > + { "tls-authz", no_argument, NULL, QEMU_NBD_OPT_TLSAUTHZ },
>
> Not your fault, but worth sorting these alphabetically?
>
> Bummer that pre-patch, you could use '--tls' as an unambiguous abbreviation
> for --tls-creds; now it is an ambiguous prefix (you have to type --tls-c or
> --tls-a to get to the point of no ambiguity). If we really cared, we could
> add:
>
> { "t", required_argument, NULL, QEMU_NBD_OPT_TLSCREDS },
> { "tl", required_argument, NULL, QEMU_NBD_OPT_TLSCREDS },
> { "tls", required_argument, NULL, QEMU_NBD_OPT_TLSCREDS },
> { "tls-", required_argument, NULL, QEMU_NBD_OPT_TLSCREDS },
>
> since getopt_long() no longer reports ambiguity if there is an exact match
> to what is otherwise the common prefix of two ambiguous options. But I don't
> think backwards-compatibility on this front is worth worrying about
> (generally, scripts don't rely on getopt_long()'s unambiguous prefix
> handling).
Personally I think this is not worth worrying about. We've never documented
ability to abbreviate nor ever promised they are stable. Abbreviations are
inherantly unstable as you illustrate, so if anything we should just document
that you should never abbreviate args.
>
> > +++ b/qemu-nbd.texi
> > @@ -91,6 +91,10 @@ of the TLS credentials object previously created with
> > the --object
> > option.
> > @item --fork
> > Fork off the server process and exit the parent once the server is
> > running.
> > +@item --tls-authz=ID
> > +Specify the ID of a qauthz object previously created with the
>
> s/qauthz/authz-simple/ ?
No, qauthz is a QOM interface, which is implemented by many subclasses,
of which authz-simple is just one example.
> > +--object option. This will be used to authorize users who
> > +connect against their x509 distinguished name.
>
> Sounds like someone is "connecting against their name", rather than
> "authorizing against their name". Better might be:
>
> This will be used to authorize connecting users against their x509
> distinguished name.
Ok
Regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
