.\"	$NetBSD: libnpf.3,v 1.4 2016/12/27 21:25:12 wiz Exp $
.\"
.\" Copyright (c) 2011-2015 The NetBSD Foundation, Inc.
.\" All rights reserved.
.\"
.\" This material is based upon work partially supported by The
.\" NetBSD Foundation under a contract with Mindaugas Rasiukevicius.
.\"
.\" 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 THE NETBSD FOUNDATION, INC. 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 THE FOUNDATION 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.
.\"
.Dd April 19, 2015
.Dt LIBNPF 3
.Os
.Sh NAME
.Nm libnpf
.Nd NPF packet filter library
.Sh LIBRARY
.Lb libnpf
.Sh SYNOPSIS
.In npf.h
.\" ---
.Ft nl_config_t *
.Fn npf_config_create "void"
.Ft int
.Fn npf_config_submit "nl_config_t *ncf" "int fd" "nl_error_t *errinfo"
.Ft nl_config_t *
.Fn npf_config_retrieve "int fd" "bool *active" "bool *loaded"
.Ft int
.Fn npf_config_flush "int fd"
.Ft void
.Fn npf_config_export "const nl_config_t *ncf" "size_t *len"
.Ft nl_config_t *
.Fn npf_config_import "const void *blob" "size_t len"
.Ft bool
.Fn npf_config_active_p "nl_config_t *ncf"
.Ft void
.Fn npf_config_destroy "nl_config_t *ncf"
.\" ---
.Ft nl_rule_t *
.Fn npf_rule_create "char *name" "uint32_t attr" "const char *ifname"
.Ft int
.Fn npf_rule_setcode "nl_rule_t *rl" "int type" "const void *code" "size_t len"
.Ft int
.Fn npf_rule_setkey "nl_rule_t *rl" "int type" "const void *code" "size_t len"
.Ft int
.Fn npf_rule_setinfo "nl_rule_t *rl" "const void *info" "size_t len"
.Ft bool
.Fn npf_rule_exists_p "nl_config_t *ncf" "const char *name"
.Ft int
.Fn npf_rule_setprio "nl_rule_t *rl" "int pri"
.Ft int
.Fn npf_rule_setproc "nl_config_t *ncf" "nl_rule_t *rl" "const char *name"
.Ft int
.Fn npf_rule_insert "nl_config_t *ncf" "nl_rule_t *parent" "nl_rule_t *rl"
.Ft void *
.Fn npf_rule_export "nl_rule_t *rl" "size_t *length"
.Ft void
.Fn npf_rule_destroy "nl_rule_t *rl"
.\" ---
.Ft nl_rproc_t *
.Fn npf_rproc_create "char *name"
.Ft bool
.Fn npf_rproc_exists_p "nl_config_t *ncf" "const char *name"
.Ft int
.Fn npf_rproc_insert "nl_config_t *ncf" "nl_rproc_t *rp"
.\" ---
.Ft nl_nat_t *
.Fn npf_nat_create "int type" "u_int flags" "const char *ifname" \
"npf_addr_t *addr" "int af" "in_port_t port"
.Ft int
.Fn npf_nat_insert "nl_config_t *ncf" "nl_nat_t *nt" "pri_t pri"
.\" ---
.Ft nl_table_t *
.Fn npf_table_create "const char *name" "u_int id" "int type"
.Ft int
.Fn npf_table_add_entry "nl_table_t *tl" "int af" \
"in_addr_t addr" "in_addr_t mask"
.Ft int
.Fn npf_table_insert "nl_config_t *ncf" "nl_table_t *tl"
.Ft void
.Fn npf_table_destroy "nl_table_t *tl"
.\" -----
.Sh DESCRIPTION
The
.Nm
library provides an interface to create an NPF configuration having rules,
tables, procedures, or translation policies.
The configuration can be submitted to the kernel.
.\" -----
.Sh FUNCTIONS
.Ss Configuration
.Bl -tag -width 4n
.It Fn npf_config_create
Create a configuration.
.It Fn npf_config_submit "ncf" "fd" "errinfo"
Submit configuration
.Fa ncf
to the kernel.
On error, the the description is written into the structure specified by
.Fa errinfo .
.It Fn npf_config_export "fd" "len"
Serialize the given configuration and return binary object and its
length in
.Fa len
parameter.
The binary object is dynamically allocated and should be destroyed using
.Xr free 3 .
.It Fn npf_config_import "blob" "len"
Read the configuration from a binary object of the specified length,
unserialize, construct and return the configuration object.
.It Fn npf_config_flush "fd"
Flush the current configuration.
.It Fn npf_config_retrieve "fd" "active" "loaded"
Retrieve and return the loaded configuration from the kernel.
.It Fn npf_config_active_p "ncf"
Indicate whether the retrieved configuration is active (true if yes
and false otherwise).
.It Fn npf_config_destroy "ncf"
Destroy the configuration
.Fa ncf .
.El
.\" ---
.Ss Rule interface
.Bl -tag -width 4n
.It Fn npf_rule_create "name" "attr" "ifname"
Create a rule with a given name, attribute and priorty.
Name can be
.Dv NULL ,
in which case rule has no unique identifier.
Otherwise, rules shall not have duplicate names.
The following attributes, which can be ORed, are available:
.Bl -tag -width indent
.It Dv NPF_RULE_PASS
Decision of this rule is "pass".
If this attribute is not
specified, then packet "block" (drop) is the default.
.It Dv NPF_RULE_IN
Match incoming packets.
.It Dv NPF_RULE_OUT
Match outgoing packets.
.It Dv NPF_RULE_FINAL
Indicates that on rule match, further processing of the
ruleset should be stopped and this rule applied instantly.
.It Dv NPF_RULE_STATEFUL
Create a state (session) on match, track the connection and
therefore pass the backwards stream without inspection.
The state is uniquely identified by a 5-tuple (source and destination
IP addresses, port numbers and an interface identifier).
.It Dv NPF_RULE_MULTIENDS
Exclude the interface from the state identifier.
.It Dv NPF_RULE_RETRST
Return TCP RST packet in a case of packet block.
.It Dv NPF_RULE_RETICMP
Return ICMP destination unreachable in a case of packet block.
.It Dv NPF_RULE_GROUP
Allow this rule to have sub-rules.
If used with
.Dv NPF_RULE_DYNAMIC
flag set, the can be added dynamically.
Otherwise rules must be added statically i.e. created with the configuration.
.It Dv NPF_RULE_DYNAMIC
Indicate that the rule is dynamic.
.El
.Pp
Interface is specified by
.Fa ifname ,
which is a string.
.Dv NULL
indicates any interface.
.\" ---
.It Fn npf_rule_setcode "rl" "type" "code" "len"
Assign compiled code for the rule specified by
.Fa rl ,
used for filter criteria.
Pointer to the binary code is specified by
.Fa code ,
and size of the memory area by
.Fa len .
Type of the code is specified by
.Fa type .
Currently, only BPF byte-code is supported and
.Dv NPF_CODE_BPF
should be passed.
.\" ---
.It Fn npf_rule_setkey "rl" "type" "key" "len"
Assign a key for the rule specified by
.Fa rl .
Binary key is specified by
.Fa key ,
and its size by
.Fa len .
The size shall not exceed
.Dv NPF_RULE_MAXKEYLEN .
The kernel does not validate the key is unique, it is the responsibility
of the caller.
.\" ---
.It Fn npf_rule_setinfo "rl" "info" "len"
Associate arbitrary information blob specified by
.Fa info ,
and its size by
.Fa len .
This may be used for such purposes as byte-code annotation.
.\" ---
.It Fn npf_rule_setprio "rl" "pri"
Set priority to the rule.
Negative priorities are invalid.
.Pp
Priority is the order of the rule in the ruleset.
Lower value means first to process, higher value - last to process.
If multiple rules are inserted with the same priority,
the order is unspecified.
.Pp
The special constants
.Dv NPF_PRI_FIRST
and
.Dv NPF_PRI_LAST
can be passed to indicate that the rule should be inserted into the
beginning or the end of the priority level 0 in the ruleset.
All rules inserted using these constants will have the priority 0
assigned and will share this level in the ordered way.
.\" ---
.It Fn npf_rule_setproc "ncf" "rl" "name"
Set a procedure for the specified rule.
.\" ---
.It Fn npf_rule_insert "ncf" "parent" "rl"
Insert the rule into the set of parent rule specified by
.Fa parent .
If value of
.Fa parent
is
.Dv NULL ,
then insert into the main ruleset.
.\" ---
.It Fn npf_rule_export "rl" "length"
Serialize the rule (including the byte-code), return a binary object
and set its
.Fa length .
The binary object is dynamically allocated and should be destroyed using
.Xr free 3 .
.\" ---
.It Fn npf_rule_destroy "rl"
Destroy the given rule.
.El
.\" -----
.Ss Rule procedure interface
.Bl -tag -width 4n
.It Fn npf_rproc_create "name"
Create a rule procedure with a given
.Fa name .
Name must be unique for each procedure.
.It Fn npf_rproc_insert "ncf" "rp"
Insert rule procedure into the specified configuration.
.El
.\" -----
.Ss Translation interface
.Bl -tag -width 4n
.It Fn npf_nat_create "type" "flags" "ifname" "addr" "af" "port"
Create a NAT translation policy of a specified type.
There are two types:
.Bl -tag -width "NPF_NAT_PORTMAP "
.It Dv NPF_NATIN
Inbound NAT policy (rewrite destination).
.It Dv NPF_NATOUT
Outbound NAT policy (rewrite source).
.El
.Pp
A bi-directional NAT is obtained by combining two policies.
The following
.Fa flags
are supported:
.Bl -tag -width "NPF_NAT_PORTMAP "
.It Dv NPF_NAT_STATIC
Perform static (stateless) NAT rather than dynamic (stateful).
.It Dv NPF_NAT_PORTS
Indicates to perform port translation.
Otherwise, port translation is not performed and
.Fa port
is ignored.
.It Dv NPF_NAT_PORTMAP
Effective only if
.Dv NPF_NAT_PORTS
flag is set.
Indicates to create a port map and select a random port for translation.
Otherwise, port is translated to the value specified by
.Fa port
is used.
.El
.Pp
Translation address is specified by
.Fa addr ,
and its family by
.Fa af .
Family must be either
.Dv AF_INET
for IPv4 or
.Dv AF_INET6
for IPv6 address.
.\" ---
.It Fn npf_nat_setalgo "nt" "algo"
Set a particular NAT algorithm.
Currently, only
.Dv NPF_ALGO_NPT66
algorithm is supported for NPTv6 (RFC 6296).
.\" ---
.It Fn npf_nat_insert "ncf" "nt" "pri"
Insert NAT policy, its rule, into the specified configuration.
.El
.\" -----
.Ss Table interface
.Bl -tag -width 4n
.It Fn npf_table_create "name" "index" "type"
Create NPF table of specified type.
The following types are supported:
.Bl -tag -width "NPF_TABLE_TREE "
.It Dv NPF_TABLE_HASH
Indicates to use hash table for storage.
.It Dv NPF_TABLE_TREE
Indicates to use red-black tree for storage.
Table is identified by the
.Fa name
and
.Fa index ,
which should be in the range between 1 and
.Dv NPF_MAX_TABLE_ID .
.El
.\" ---
.It Fn npf_table_add_entry "tl" "af" "addr" "mask"
Add an entry of IP address and mask, specified by
.Fa addr
and
.Fa mask ,
to the table specified by
.Fa tl .
Family, specified by
.Fa af ,
must be either
.Dv AF_INET
for IPv4 or
.Dv AF_INET6
for IPv6 address.
.It Fn npf_table_insert "ncf" "tl"
Insert table into set of configuration.
Routine performs a check for duplicate table ID.
.\" ---
.It Fn npf_table_destroy "tl"
Destroy the specified table.
.El
.\" -----
.Sh SEE ALSO
.Xr bpf 4 ,
.Xr npf 7 ,
.Xr npfctl 8
.Sh HISTORY
The NPF library first appeared in
.Nx 6.0 .