config root man

.\"
.\" Copyright (c) 2004-2005
.\"	Hartmut Brandt
.\"	All rights reserved.
.\" Copyright (c) 2001-2003
.\"	Fraunhofer Institute for Open Communication Systems (FhG Fokus).
.\"	All rights reserved.
.\"
.\" Author: Harti Brandt <harti@FreeBSD.org>
.\" 
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\" 
.\" THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" $Begemot: bsnmp/snmp_mibII/snmp_mibII.3,v 1.10 2005/10/04 08:46:52 brandt_h Exp $
.\"
.Dd October 4, 2005
.Dt SNMP_MIBII 3
.Os
.Sh NAME
.Nm mibII ,
.Nm mibif_notify_f ,
.Nm mib_netsock ,
.Nm mib_if_set_dyn ,
.Nm mib_refresh_iflist ,
.Nm mib_find_if ,
.Nm mib_find_if_sys ,
.Nm mib_find_if_name ,
.Nm mib_first_if ,
.Nm mib_next_if ,
.Nm mib_register_newif ,
.Nm mib_unregister_newif ,
.Nm mib_fetch_ifmib ,
.Nm mib_if_admin ,
.Nm mib_find_ifa ,
.Nm mib_first_ififa ,
.Nm mib_next_ififa ,
.Nm mib_ifstack_create ,
.Nm mib_ifstack_delete ,
.Nm mib_find_rcvaddr ,
.Nm mib_rcvaddr_create ,
.Nm mib_rcvaddr_delete ,
.Nm mibif_notify ,
.Nm mibif_unnotify
.Nd "mib-2 module for bsnmpd."
.Sh LIBRARY
.Pq begemotSnmpdModulePath."mibII" = "@MODPATH@snmp_mibII.so"
.Sh SYNOPSIS
.In bsnmp/snmpmod.h
.In bsnmp/snmp_mibII.h
.Ft typedef void
.Fn (*mibif_notify_f) "struct mibif *ifp" "enum mibif_notify event" "void *uarg"
.Vt extern int mib_netsock ;
.Ft void
.Fn mib_if_set_dyn "const char *ifname"
.Ft void
.Fn mib_refresh_iflist "void"
.Ft struct mibif *
.Fn mib_find_if "u_int ifindex"
.Ft struct mibif *
.Fn mib_find_if_sys "u_int sysindex"
.Ft struct mibif *
.Fn mib_find_if_name "const char *ifname"
.Ft struct mibif *
.Fn mib_first_if "void"
.Ft struct mibif *
.Fn mib_next_if "const struct mibif *ifp"
.Ft int
.Fn mib_register_newif "int (*func)(struct mibif *)" "const struct lmodule *mod"
.Ft void
.Fn mib_unregister_newif "const struct lmodule *mod"
.Ft int
.Fn mib_fetch_ifmib "struct mibif *ifp"
.Ft int
.Fn mib_if_admin "struct mibif *ifp" "int up"
.Ft struct mibifa *
.Fn mib_find_ifa "struct in_addr ipa"
.Ft struct mibifa *
.Fn mib_first_ififa "const struct mibif *ifp"
.Ft struct mibifa *
.Fn mib_next_ififa "struct mibifa *ifa"
.Ft int
.Fn mib_ifstack_create "const struct mibif *lower" "const struct mibif *upper"
.Ft void
.Fn mib_ifstack_delete "const struct mibif *lower" "const struct mibif *upper"
.Ft struct mibrcvaddr *
.Fn mib_find_rcvaddr "u_int ifindex" "const u_char *addr" "size_t addrlen"
.Ft struct mibrcvaddr *
.Fn mib_rcvaddr_create "struct mibif *ifp" "const u_char *addr" "size_t addrlen"
.Ft void
.Fn mib_rcvaddr_delete "struct mibrcvaddr *addr"
.Ft void *
.Fn mibif_notify "struct mibif *ifp" "const struct lmodule *mod" "mibif_notify_f func" "void *uarg"
.Ft void
.Fn mibif_unnotify "void *reg"
.Sh DESCRIPTION
The
.Nm snmp_mibII
module implements parts of the internet standard MIB-2.
Most of the relevant MIBs are implemented.
Some of the tables are restricted to be read-only instead of read-write.
The exact current implementation can be found in
.Pa @DEFPATH@mibII_tree.def .
The module also exports a number of functions and global variables for use
by other modules, that need to handle network interfaces.
This man page describes these functions.
.Ss DIRECT NETWORK ACCESS
The
.Nm
module opens a socket that is used to execute all network related
.Xr ioctl 2
functions.
This socket is globally available under the name
.Va mib_netsock .
.Ss NETWORK INTERFACES
The
.Nm
module handles a list of all currently existing network interfaces.
It allows
other modules to handle their own interface lists with special information
by providing a mechanism to register to events that change the interface list
(see below).
The basic data structure is the interface structure:
.Bd -literal -offset indent
struct mibif {
	TAILQ_ENTRY(mibif) link;
	u_int		flags;
	u_int		index;	/* logical ifindex */
	u_int		sysindex;
	char		name[IFNAMSIZ];
	char		descr[256];
	struct ifmibdata mib;
	uint64_t	mibtick;
	void		*specmib;
	size_t		specmiblen;
	u_char		*physaddr;
	u_int		physaddrlen;
	int		has_connector;
	int		trap_enable;
	uint64_t	counter_disc;
	mibif_notify_f	xnotify;
	void		*xnotify_data;
	const struct lmodule *xnotify_mod;
	struct asn_oid	spec_oid;
};
.Ed
.Pp
The
.Nm
module tries to implement the semantic if
.Va ifIndex
as described in RFC-2863.
This RFC states, that an interface indexes may not be reused.
That means, for example, if
.Pa tun
is a synthetic interface type and the system creates the interface
.Pa tun0 ,
destroys this interfaces and again creates a
.Pa tun 0 ,
then these interfaces must have different interface indexes, because in fact
they are different interfaces.
If, on the other hand, there is a hardware interface
.Pa xl0
and this interface disappears, because its driver is unloaded and appears
again, because the driver is loaded again, the interface index must stay
the same.
.Nm
implements this by differentiating between real and synthetic (dynamic)
interfaces.
An interface type can be declared dynamic by calling the function
.Fn mib_if_set_dyn
with the name if the interface type (for example
.Qq tun ).
For real interfaces, the module keeps the mapping between the interface name
and its
.Va ifIndex
in a special list, if the interface is unloaded.
For dynamic interfaces
a new
.Va ifIndex
is generated each time the interface comes into existence.
This means, that the interface index as seen by SNMP is not the same index
as used by the system.
The SNMP
.Va ifIndex
is held in field
.Va index ,
the system's interface index is
.Va sysindex .
.Pp
A call to
.Nm mib_refresh_iflist
causes the entire interface list to be re-created.
.Pp
The interface list can be traversed with the functions
.Fn mib_first_if
and
.Fn mib_next_if .
Be sure not to change the interface list while traversing the list with
these two calls.
.Pp
There are three functions to find an interface by name or index.
.Fn mib_find_if
finds an interface by searching for an SNMP
.Va ifIndex ,
.Fn mib_find_if_sys
finds an interface by searching for a system interface index and
.Fn mib_find_if_name
finds an interface by looking for an interface name.
Each of the function returns
.Li NULL
if the interface cannot be found.
.Pp
The function
.Fn mib_fetch_ifmib
causes the interface MIB to be refreshed from the kernel.
.Pp
The function
.Fn mib_if_admin
can be used to change the interface administrative state to up
(argument is 1) or down (argument is 0).
.Ss INTERFACE EVENTS
A module can register itself to receive a notification when a new entry is
created in the interface list.
This is done by calling
.Fn mib_register_newif .
A module can register only one function, a second call to
.Fn mib_register_newif
causes the registration to be overwritten.
The registration can be removed with a call to
.Fn mib_unregister_newif .
It is unregistered automatically, when the registering module is unloaded.
.Pp
A module can also register to events on a specific interface.
This is done by calling
.Fn mibif_notify .
This causes the given callback
.Fa func
to be called with the interface pointer, a notification code and
the user argument
.Fa uarg
when any of the following events occur:
.Bl -tag -width "XXXXX"
.It Li MIBIF_NOTIFY_DESTROY
The interface is destroyed.
.El
.Pp
This mechanism can be used to implement interface type specific MIB parts
in other modules.
The registration can be removed with
.Fn mib_unnotify
which the return value from
.Fa mib_notify .
Any notification registration is removed automatically when the interface
is destroyed or the registering module is unloaded.
.Em Note that only one module can register to any given interface .
.Ss INTERFACE ADDRESSES
The
.Nm
module handles a table of interface IP-addresses.
These addresses are held in a
.Bd -literal -offset indent
struct mibifa {
	TAILQ_ENTRY(mibifa) link;
	struct in_addr	inaddr;
	struct in_addr	inmask;
	struct in_addr	inbcast;
	struct asn_oid	index;
	u_int		ifindex;
	u_int		flags;
};
.Ed
.Pp
The (ordered) list of IP-addresses on a given interface can be traversed by
calling
.Fn mib_first_ififa
and
.Fn mib_next_ififa .
The list should not be considered read-only.
.Ss INTERFACE RECEIVE ADDRESSES
The internet MIB-2 contains a table of interface receive addresses.
These addresses are handled in:
.Bd -literal -offset indent
struct mibrcvaddr {
	TAILQ_ENTRY(mibrcvaddr) link;
	struct asn_oid	index;
	u_int		ifindex;
	u_char		addr[ASN_MAXOIDLEN];
	size_t		addrlen;
	u_int		flags;
};
enum {
	MIBRCVADDR_VOLATILE	= 0x00000001,
	MIBRCVADDR_BCAST	= 0x00000002,
	MIBRCVADDR_HW		= 0x00000004,
};
.Ed
.Pp
Note, that the assignment of
.Li MIBRCVADDR_BCAST
is based on a list of known interface types.
The flags should be handled
by modules implementing interface type specific MIBs.
.Pp
A receive address can be created with
.Fn mib_rcvaddr_create
and deleted with
.Fn mib_rcvaddr_delete .
This needs to be done only for addresses that are not automatically handled
by the system.
.Pp
A receive address can be found with
.Fn mib_find_rcvaddr .
.Ss INTERFACE STACK TABLE
The
.Nm
module maintains also the interface stack table.
Because for complex stacks,
there is no system supported generic way of getting this information, interface
type specific modules need to help setting up stack entries.
The
.Nm
module handles only the top and bottom entries.
.Pp
A table entry is created with
.Fn mib_ifstack_create
and deleted with
.Fn mib_ifstack_delete .
Both functions need the pointers to the interfaces.
Entries are automatically
deleted if any of the interfaces of the entry is destroyed.
The functions handle
both the stack table and the reverse stack table.
.Sh FILES
.Bl -tag -width ".It Pa @DEFPATH@mibII_tree.def" -compact
.It Pa @DEFPATH@mibII_tree.def
The description of the MIB tree implemented by
.Nm .
.It Pa /usr/local/share/snmp/mibs
.It Pa @MIBSPATH@
The various internet MIBs.
.El
.Sh SEE ALSO
.Xr gensnmptree 1 ,
.Xr snmpmod 3
.Sh STANDARDS
This implementation conforms to the applicable IETF RFCs.
.Sh AUTHORS
.An Hartmut Brandt Aq harti@FreeBSD.org