On Mon, 2019-07-29 at 14:46 +0100, Daniel P. Berrangé wrote:
On Mon, Jul 29, 2019 at 10:33:08AM +0200, Andrea Bolognani wrote:
> On Tue, 2019-07-23 at 17:03 +0100, Daniel P. Berrangé wrote:
> > + VIR_DEBUG("Pointing secondary drivers to primary");
> > + priv->interfaceConn = virObjectRef(priv->conn);
> > + priv->networkConn = virObjectRef(priv->conn);
> > + priv->nodedevConn = virObjectRef(priv->conn);
> > + priv->nwfilterConn = virObjectRef(priv->conn);
> > + priv->secretConn = virObjectRef(priv->conn);
> > + priv->storageConn = virObjectRef(priv->conn);
>
> Do we even need this code for the non-libvirtd case? We have listed
> all drivers, primary and secondary, above, so I can't think of any
> valid reason we'd end up here unless there's a bug, and in that case
> we'd just be masking it, no?
It is handling the remote driver case for virtproxyd, but we could
make that more explicit.
Yeah, that'd be great! I didn't realize that was the case at all.
--
Andrea Bolognani / Red Hat / Virtualization