ODL Release - sulphur via Light.io 16.1

Issue:
Given a NETCONF device setup for call home, "odl-netconf-callhome-server" yang is used to add the device to an allowed list mapping a name to a host key. The device connects to OpenDayLight (good).

The yang models from the device are all placed in the "cache/schema" directory of OpenDayLight - is this a bug? Given the call home device has a unique name, shouldn't there be a cache/<deviceName> for those learned schemas?

Also, we have devices that must take advantage of "user defined" schema list when they are added using "netconf-topology" APIs. OpenDaylight allows us to specify user defined schemas as well as specify the schema directory using a non "callhome" API
PUT /restconf/data/network-topology:network-topology/topology=

/node=

But when using call home, this does not seem possible. Devices are added that call home, their schema ends up in the "cache/schemas" directory. There is no way to make a call home device point to a specific schema directory and/or implement user defined schema capabilities.

Not sure if this is a bug or a feature request here.. or are we just missing something obvious?

To recreate, turn on the call home feature in ODL, and add a sysrepo2 simulator which is provisioned to use the ODL instance. Using the call home yang, add an allowed device to set the unique name, host key, etc.. OR configure the global provisioning to accept any device.

The device will connect to ODL and show up in the "operational/non-config" GET network-topology list. This list will show the device connected and its capabilities.
The network-topology will allow the user to "config" a device with the same name as the call home device, but there doesn't seem to be any cohesion between the two - call home and configured netconf-topology devices seem to be a separate entity.