freebsd-skq/share/man/man4/ng_patch.4
Eitan Adler 50d675f7a9 Remove trailing whitespace per mdoc lint warning
Disussed with:	gavin
No objection from:	doc
Approved by:	joel
MFC after:	3 days
2012-03-29 05:02:12 +00:00

236 lines
7.3 KiB
Groff

.\" Copyright (c) 2010 Maxim Ignatenko <gelraen.ua@gmail.com>
.\" Copyright (c) 2010 Vadim Goncharov <vadimnuclight@tpu.ru>
.\" All rights reserved.
.\"
.\" 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 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 THE 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.
.\"
.\" $FreeBSD$
.\"
.Dd March 5, 2012
.Dt NG_PATCH 4
.Os
.Sh NAME
.Nm ng_patch
.Nd "trivial mbuf data modifying netgraph node type"
.Sh SYNOPSIS
.In netgraph/ng_patch.h
.Sh DESCRIPTION
The
.Nm patch
node performs data modification of packets passing through it.
Modifications are restricted to a subset of C language operations
on unsigned integers of 8, 16, 32 or 64 bit size.
These are: set to new value (=), addition (+=), subtraction (-=),
multiplication (*=), division (/=), negation (= -),
bitwise AND (&=), bitwise OR (|=), bitwise eXclusive OR (^=),
shift left (<<=), shift right (>>=).
A negation operation is the one exception: integer is treated as signed
and second operand (the
.Va value )
is not used.
There may be several modification operations, they are all applied
to a packet sequentially in order they were specified by user.
Data payload of packet is viewed as array of bytes, with zero offset
corresponding to the very first byte of packet headers, and
.Va length
bytes beginning from
.Va offset
are taken as a single integer in network byte order.
.Sh HOOKS
This node type has two hooks:
.Bl -tag -width indent
.It Va in
Packets received on this hook are modified according to rules specified
in config and then forwarded to
.Ar out
hook, if it exists and connected.
Otherwise they are reflected back to the
.Ar in
hook.
.It Va out
Packets received on this hook are forwarded to
.Ar in
hook without any changes.
.El
.Sh CONTROL MESSAGES
This node type supports the generic control messages, plus the following:
.Bl -tag -width indent
.It Dv NGM_PATCH_SETCONFIG Pq Li setconfig
This command sets the sequence of modify operations
that will be applied to incoming data on a hook.
The following
.Vt "struct ng_patch_config"
must be supplied as an argument:
.Bd -literal -offset 4n
struct ng_patch_op {
uint64_t value;
uint32_t offset;
uint16_t length; /* 1,2,4 or 8 bytes */
uint16_t mode;
};
/* Patching modes */
#define NG_PATCH_MODE_SET 1
#define NG_PATCH_MODE_ADD 2
#define NG_PATCH_MODE_SUB 3
#define NG_PATCH_MODE_MUL 4
#define NG_PATCH_MODE_DIV 5
#define NG_PATCH_MODE_NEG 6
#define NG_PATCH_MODE_AND 7
#define NG_PATCH_MODE_OR 8
#define NG_PATCH_MODE_XOR 9
#define NG_PATCH_MODE_SHL 10
#define NG_PATCH_MODE_SHR 11
struct ng_patch_config {
uint32_t count;
uint32_t csum_flags;
struct ng_patch_op ops[];
};
.Ed
.Pp
The
.Va csum_flags
can be set to any combination of CSUM_IP, CSUM_TCP, CSUM_SCTP and CSUM_UDP
(other values are ignored) for instructing the IP stack to recalculate the
corresponding checksum before transmitting packet on output interface.
The
.Nm
node does not do any checksum correction by itself.
.It Dv NGM_PATCH_GETCONFIG Pq Li getconfig
This control message obtains current set of modify operations,
returned as
.Vt "struct ng_patch_config" .
.It Dv NGM_PATCH_GET_STATS Pq Li getstats
Returns node statistics as a
.Vt "struct ng_patch_stats" .
.It Dv NGM_PATCH_CLR_STATS Pq Li clrstats
Clear node statistics.
.It Dv NGM_PATCH_GETCLR_STATS Pq Li getclrstats
This command is identical to
.Dv NGM_PATCH_GET_STATS ,
except that the statistics are also atomically cleared.
.El
.Sh SHUTDOWN
This node shuts down upon receipt of a
.Dv NGM_SHUTDOWN
control message, or when all hooks have been disconnected.
.Sh EXAMPLES
The
.Nm
node allows to modify TTL and TOS/DSCP fields in IP packets.
Suppose you have two adjacent simplex links to remote network
(e.g.\& satellite), so that the packets expiring in between
will generate unwanted ICMP-replies which have to go forth, not back.
Thus you need to raise TTL of every packet entering link link by 2
to ensure the TTL will not reach zero there.
So you setup
.Xr ipfw 8
rule with
.Cm netgraph
action to inject packets going to other end of simplex link by the
following
.Xr ngctl 8
script:
.Bd -literal -offset 4n
/usr/sbin/ngctl -f- <<-SEQ
mkpeer ipfw: patch 200 in
name ipfw:200 ttl_add
msg ttl_add: setconfig { count=1 csum_flags=1 ops=[ \e
{ mode=2 value=3 length=1 offset=8 } ] }
SEQ
/sbin/ipfw add 150 netgraph 200 ip from any to simplex.remote.net
.Ed
.Pp
Here
.Dq Li ttl_add
node of type
.Nm
configured to add (mode
.Dv NG_PATCH_MODE_ADD )
a
.Va value
of 3 to a one-byte TTL field, which is 9th byte of IP packet header.
.Pp
Another example would be two consecutive modifications of packet TOS
field: say, you need to clear the
.Dv IPTOS_THROUGHPUT
bit and set the
.Dv IPTOS_MINCOST
bit.
So you do:
.Bd -literal -offset 4n
/usr/sbin/ngctl -f- <<-SEQ
mkpeer ipfw: patch 300 in
name ipfw:300 tos_chg
msg tos_chg: setconfig { count=2 csum_flags=1 ops=[ \e
{ mode=7 value=0xf7 length=1 offset=1 } \e
{ mode=8 value=0x02 length=1 offset=1 } ] }
SEQ
/sbin/ipfw add 160 netgraph 300 ip from any to any not dst-port 80
.Ed
.Pp
This first does
.Dv NG_PATCH_MODE_AND
clearing the fourth bit and then
.Dv NG_PATCH_MODE_OR
setting the third bit.
.Pp
In both examples the
.Va csum_flags
field indicates that IP checksum (but not TCP or UDP checksum) should be
recalculated before transmit.
.Pp
Note: one should ensure that packets are returned to ipfw after processing
inside
.Xr netgraph 4 ,
by setting appropriate
.Xr sysctl 8
variable:
.Bd -literal -offset 4n
sysctl net.inet.ip.fw.one_pass=0
.Ed
.Sh SEE ALSO
.Xr netgraph 4 ,
.Xr ng_ipfw 4 ,
.Xr ngctl 8
.Sh HISTORY
The
.Nm
node type was implemented in
.Fx 8.1 .
.Sh AUTHORS
.An "Maxim Ignatenko" Aq gelraen.ua@gmail.com .
This manual page was written by
.An "Vadim Goncharov" Aq vadimnuclight@tpu.ru .
.Sh BUGS
Node blindly tries to apply every patching operation to each packet
(except those which offset if greater than length of the packet),
so be sure that you supply only the right packets to it (e.g. changing
bytes in the ARP packets meant to be in IP header could corrupt
them and make your machine unreachable from the network).
.Pp
.Em !!! WARNING !!!
.Pp
Output path of the IP stack assumes correct fields and lengths in the
packets - changing them by mistake to incorrect values can cause
unpredictable results including kernel panics.