The call Statement

Syntax

call <server> {
    [ statements ]
}

The call statement will run the current request through another virtual server. Once that virtual server is complete, the request will continue being processed in the original virtual server.

<server>: The virtual server which will process the request.

The called virtual server must have the same namespace value as the calling virtual server. Enforcing the same namespace prevents calls to a virtual server with an incompatible protocol during compile or run-times.

The called virtual server will execute the request as if it had been received from the network. For example, if the packet is a RADIUS Access-Request, then the called virtual server will process the recv Access-Request section, followed by the authenticate <type> section, followed by the send Access-Accept section.
[ statements ]: Once the called virtual server has finished, the [ statements] from inside of the call section are executed.

The return code of the call statement is taken from the final return code of the called virtual server.

The statments inside of a call { … } execute after the called server returns. Execution of these statements returns an rcode that becomes the call rcode, overriding the rcode of the called server.

The call keyword takes the current request and sends it through the named virtual server. This means that the called virtual server can edit the request list, reply list, session-state list, etc.

In many cases (e.g. proxying), the call keyword should be wrapped in a subrequest keyword. The subrequest can be used to create a child request that the other server can edit.

Example

Processing the contents of a DHCPv6 relay-message by using a subrequest and recursively calling the DHCPv6 virtual server.

subrequest Packet-Type Relay-Message reply.Relay-Message {
    call dhcpv6 {}
}