npfctl
—
control NPF packet filter
npfctl |
command [arguments] |
The npfctl
command can be used to control the NPF packet
filter. For a description of NPF's configuration file, see
npf.conf(5).
The first argument, command, specifies the
action to take. Valid commands are:
start
- Enable packet inspection using the currently loaded configuration, if any.
Note that this command does not load or reload the configuration, or
affect existing connections.
stop
- Disable packet inspection. This command does not change the currently
loaded configuration, or affect existing connections.
reload
[path]
- Load or reload configuration from file. The configuration file at
/etc/npf.conf will be used unless a file is
specified by path. All connections will be preserved
during the reload, except those which will lose NAT policy due to removal.
NAT policy is determined by the translation type and address. Note that
change of filter criteria will not expire associated connections. The
reload operation (i.e., replacing the ruleset, NAT policies and tables) is
atomic.
flush
- Flush configuration. That is, remove all rules, tables and expire all
connections. This command does not disable packet inspection.
show
- Show the current state and configuration. Syntax of printed configuration
is for the user and may not match the
npf.conf(5) syntax.
validate
[path]
- Validate the configuration file and the processed form. The configuration
file at /etc/npf.conf will be used unless a file
is specified by path. The path
may also be set to "-" for reading from
stdin.
rule
name add
⟨rule-syntax⟩
- Add a rule to a dynamic ruleset specified by name.
On success, returns a unique identifier which can be used to remove the
rule with
rem-id
command. The identifier is
alphanumeric string.
rule
name rem
⟨rule-syntax⟩
- Remove a rule from a dynamic ruleset specified by
name. This method uses SHA1 hash computed on a rule
to identify it. Although very unlikely, it is subject to hash collisions.
For a fully reliable and more efficient method, it is recommended to use
rem-id
command.
rule
name rem-id
⟨id⟩
- Remove a rule specified by unique id from a dynamic
ruleset specified by name.
rule
name list
- List all rules in the dynamic ruleset specified by
name.
rule
name flush
- Remove all rules from the dynamic ruleset specified by
name.
table
name add
⟨addr/mask⟩
- In table name, add the IP address and optionally
netmask, specified by ⟨addr/mask⟩.
Only the tables of type "lpm" support masks.
table
name rem
⟨addr/mask⟩
- In table name, remove the IP address and optionally
netmask, specified by ⟨addr/mask⟩.
Only the tables of type "lpm" support masks.
table
name test
⟨addr⟩
- Query the table name for a specific IP address,
specified by addr. If no mask is specified, a single
host is assumed.
table
name list
- List all entries in the currently loaded table specified by
name. This operation is expensive and should be used
with caution.
table
name replace
[-n
newname]
[-t
type]
⟨path⟩
- Replace the existing table specified by name with a
new table built from the file specified by path.
Optionally, the new table will:
-n
newname
- be named newname, effectively renaming the
table. If not specified, the name of the table being replaced will be
used.
-t
type
- be of type type; currently supported types are
ipset
, lpm
, or
const
. If not specified, the type of the table
being replaced will be used.
save
[path]
- Save the active configuration with a snapshot of the current connections.
The data will be stored in the /var/db/npf.db file
unless a file is specified by path. Administrator
may want to stop the packet inspection before saving.
load
- Load the saved configuration file and the connections from the file. Note
that any existing connections will be destroyed. Administrator may want to
start packet inspection after the load.
stats
- Print various statistics.
debug
(-a
|
-b
binary-config |
-c
config) [-o
outfile]
- Process the active configuration (if
-a
is set),
the given binary configuration (if -b
is set) or
the given the plain configuration (if -c
is set).
Print the byte-code of each rule and the encoded configuration data. Also,
if -o
is set, write the binary configuration data
into the given file.
This is primarily for developer use.
list
[-46hNnW
] [-i
ifname]
- Display a list of tracked connections:
-4
- Display only IPv4 connections.
-6
- Display only IPv6 connections.
-h
- Don't display a header.
-N
- Try to resolve addresses.
-n
- Only show NAT connections.
-W
- Restrict the display width.
-i
ifname
- Display only connections through the named interface.
Reloading the configuration is a relatively expensive operation. Therefore,
frequent reloads should be avoided. Use of tables should be considered as an
alternative design. See
npf.conf(5) for details.
- /dev/npf
- control device
- /etc/npf.conf
- default configuration file
Starting the NPF packet filter:
# npfctl reload
# npfctl start
# npfctl show
Addition and removal of entries in the table whose ID is
"vip":
# npfctl table "vip" add 10.0.0.1
# npfctl table "vip" rem 182.168.0.0/24
Replacing the existing table which has ID "svr" with a
new const table populated from file "/tmp/npf_vps_new", and
renamed to "vps":
# npfctl table "svr" replace -n "vps" -t const "/tmp/npf_vps_new"
NPF first appeared in NetBSD 6.0.
NPF was designed and implemented by Mindaugas
Rasiukevicius.