DragonFly On-Line Manual Pages

Search: Section:  

NG_CISCO(4)	      DragonFly Kernel Interfaces Manual	   NG_CISCO(4)


ng_cisco -- Cisco HDLC protocol netgraph node type


#include <netinet/in.h> #include <netgraph/cisco/ng_cisco.h>


The cisco node type performs encapsulation and de-encapsulation of pack- ets using the Cisco HDLC protocol. This is a fairly simple protocol for the transmission of packets across high speed synchronous lines. Each packet is prepended with an Ethertype, indicating the protocol. There is also a ``keep alive'' and an ``inquire'' capability. The downstream hook should connect to the synchronous line. On the other side of the node are the inet, inet6, atalk, and ipx hooks, which trans- mit and receive raw IP, IPv6, AppleTalk, and IPX packets, respectively. Typically these hooks would connect to the corresponding hooks on an ng_iface(4) type node. IP Configuration In order to function properly for IP traffic, the node must be informed of the local IP address and netmask setting. This is because the proto- col includes an ``inquire'' packet which we must be prepared to answer. There are two ways to accomplish this, manually and automatically. Whenever such an inquire packet is received, the node sends a NGM_CISCO_GET_IPADDR control message to the peer node connected to the inet hook (if any). If the peer responds, then that response is used. This is the automatic method. If the peer does not respond, the node falls back on its cached value for the IP address and netmask. This cached value can be set at any time with a NGM_CISCO_SET_IPADDR message, and this is the manual method. If the inet hook is connected to the inet hook of an ng_iface(4) node, as is usually the case, then configuration is automatic as the ng_iface(4) understands the NGM_CISCO_GET_IPADDR message.


This node type supports the following hooks: downstream The connection to the synchronous line. inet IP hook. inet6 IPv6 hook. atalk AppleTalk hook. ipx IPX hook


This node type supports the generic control messages, plus the following: NGM_CISCO_SET_IPADDR This command takes an array of two struct in_addr arguments. The first is the IP address of the corresponding interface and the sec- ond is the netmask. NGM_CISCO_GET_IPADDR This command returns the IP configuration in the same format used by NGM_CISCO_SET_IPADDR. This command is also sent by this node type to the inet peer whenever an IP address inquiry packet is received. NGM_CISCO_GET_STATUS Returns a struct ngciscostat: struct ngciscostat { u_int32_t seq_retries; /* # unack'd retries */ u_int32_t keepalive_period; /* in seconds */ };


This node shuts down upon receipt of a NGM_SHUTDOWN control message, or when all hooks have been disconnected.


netgraph(4), ng_iface(4), ngctl(8) D. Perkins, Requirements for an Internet Standard Point-to-Point Protocol, RFC 1547.


Cisco is a trademark of Cisco Systems, Inc.


The ng_cisco node type was implemented in FreeBSD 4.0.


Julian Elischer <julian@FreeBSD.org> Archie Cobbs <archie@FreeBSD.org>


Not all of the functionality has been implemented. For example, the node does not support querying the remote end for its IP address and netmask. DragonFly 5.1 January 19, 1999 DragonFly 5.1

Search: Section: