ipsec
consists of two sub-protocols, namely
ESP
(encapsulated security payload)
and AH
(authentication header).
ESP protects IP payload from wire-tapping by encrypting it by
secret key cryptography algorithms.
AH guarantees integrity of IP packet
and protects it from intermediate alteration or impersonation,
by attaching cryptographic checksum computed by one-way hash functions.
ipsec
has two operation modes: transport mode and tunnel mode.
Transport mode is for protecting peer-to-peer communication between end nodes.
Tunnel mode includes IP-in-IP encapsulation operation
and is designed for security gateways, like VPN configurations.
The following kernel options are available:
- options IPSEC
-
Includes support for the
IPsec
protocol.
IPSEC
will enable
secret key management part,
policy management part,
AH
and
IPComp.
Kernel binary will not be subject to export control in most of countries,
even if compiled with
IPSEC.
For example, it should be okay to export it from within the United States
to the outside.
INET6
and
IPSEC
are orthogonal so you can get IPv4-only kernel with IPsec support,
IPv4/v6 dual support kernel without IPsec, and so forth.
This option requires
INET
at this moment, but it should not.
- options IPSEC_DEBUG
-
Enables debugging code in
IPsec
stack.
This option assumes
IPSEC.
- options IPSEC_ESP
-
Includes support for
IPsec
ESP
protocol.
IPSEC_ESP
will enable source code that is subject to export control in some countries
(including the United States),
and compiled kernel binary will be subject to certain restriction.
This option assumes
IPSEC.
- options IPSEC_NAT_T
-
Includes support for
IPsec
Network Address Translator traversal (NAT-T), as described in RFCs 3947
and 3948.
This feature might be patent-encumbered in some countries.
This option assumes
IPSEC
and
IPSEC_ESP.
Kernel interface
ipsec
is controlled by key management engine and policy engine,
in the operating system kernel.
Key management engine can be accessed from the userland by using
PF_KEY
sockets.
The
PF_KEY
socket API is defined in RFC2367.
Policy engine can be controlled by extended part of
PF_KEY
API,
setsockopt(2)
operations, and
sysctl(3)
interface.
The kernel implements
extended version of
PF_KEY
interface, and allows you to define IPsec policy like per-packet filters.
setsockopt(2)
interface is used to define per-socket behavior, and
sysctl(3)
interface is used to define host-wide default behavior.
The kernel code does not implement dynamic encryption key exchange protocol
like IKE
(Internet Key Exchange).
That should be implemented as userland programs
(usually as daemons),
by using the above described APIs.
Policy management
The kernel implements experimental policy management code.
You can manage the IPsec policy in two ways.
One is to configure per-socket policy using
setsockopt(2).
The other is to configure kernel packet filter-based policy using
PF_KEY
interface, via
setkey(8).
In both cases, IPsec policy must be specified with syntax described in
ipsec_set_policy(3).
With
setsockopt(2),
you can define IPsec policy in per-socket basis.
You can enforce particular IPsec policy onto packets that go through
particular socket.
With
setkey(8)
you can define IPsec policy against packets,
using sort of packet filtering rule.
Refer to
setkey(8)
on how to use it.
In the latter case,
``
default
''
policy is allowed for use with
setkey(8).
By configuring policy to
default
,
you can refer system-wide
sysctl(8)
variable for default settings.
The following variables are available.
1
means
``
use
'',
and
2
means
``
require
''
in the syntax.
Name Type Changeable
|
net.inet.ipsec.esp_trans_deflev integer yes
|
net.inet.ipsec.esp_net_deflev integer yes
|
net.inet.ipsec.ah_trans_deflev integer yes
|
net.inet.ipsec.ah_net_deflev integer yes
|
net.inet6.ipsec6.esp_trans_deflev integer yes
|
net.inet6.ipsec6.esp_net_deflev integer yes
|
net.inet6.ipsec6.ah_trans_deflev integer yes
|
net.inet6.ipsec6.ah_net_deflev integer yes
|
If kernel finds no matching policy system wide default value is applied.
System wide default is specified by the following
sysctl(8)
variables.
0
means
``
discard
''
which asks the kernel to drop the packet.
1
means
``
none
''.
Name Type Changeable
|
net.inet.ipsec.def_policy integer yes
|
net.inet6.ipsec6.def_policy integer yes
|
Miscellaneous sysctl variables
The following variables are accessible via
sysctl(8),
for tweaking kernel IPsec behavior:
Name Type Changeable
|
net.inet.ipsec.ah_cleartos integer yes
|
net.inet.ipsec.ah_offsetmask integer yes
|
net.inet.ipsec.dfbit integer yes
|
net.inet.ipsec.ecn integer yes
|
net.inet.ipsec.debug integer yes
|
net.inet6.ipsec6.ecn integer yes
|
net.inet6.ipsec6.debug integer yes
|
The variables are interpreted as follows:
ipsec.ah_cleartos
-
If set to non-zero, the kernel clears type-of-service field in the IPv4 header
during AH authentication data computation.
The variable is for tweaking AH behavior to interoperate with devices that
implement RFC1826 AH.
It should be set to non-zero
(clear the type-of-service field)
for RFC2402 conformance.
ipsec.ah_offsetmask
-
During AH authentication data computation, the kernel will include
16bit fragment offset field
(including flag bits)
in IPv4 header, after computing logical AND with the variable.
The variable is for tweaking AH behavior to interoperate with devices that
implement RFC1826 AH.
It should be set to zero
(clear the fragment offset field during computation)
for RFC2402 conformance.
ipsec.dfbit
-
The variable configures the kernel behavior on IPv4 IPsec tunnel encapsulation.
If set to 0, DF bit on the outer IPv4 header will be cleared.
1 means that the outer DF bit is set regardless from the inner DF bit.
2 means that the DF bit is copied from the inner header to the outer.
The variable is supplied to conform to RFC2401 chapter 6.1.
ipsec.ecn
-
If set to non-zero, IPv4 IPsec tunnel encapsulation/decapsulation behavior will
be friendly to ECN
(explicit congestion notification),
as documented in
draft-ietf-ipsec-ecn-02.txt
.
gif(4)
talks more about the behavior.
ipsec.debug
-
If set to non-zero, debug messages will be generated via
syslog(3).
Variables under
net.inet6.ipsec6
tree has similar meaning as the
net.inet.ipsec
counterpart.
PROTOCOLS
The
ipsec
protocol works like plug-in to
inet(4)
and
inet6(4)
protocols.
Therefore,
ipsec
supports most of the protocols defined upon those IP-layer protocols.
Some of the protocols, like
icmp(4)
or
icmp6(4),
may behave differently with
ipsec.
This is because
ipsec
can prevent
icmp(4)
or
icmp6(4)
routines from looking into IP payload.
SEE ALSO
ioctl(2),
socket(2),
ipsec_set_policy(3),
fast_ipsec(4),
icmp6(4),
intro(4),
ip6(4),
racoon(8),
setkey(8),
sysctl(8)
STANDARDS
HISTORY
The implementation described herein appeared in WIDE/KAME IPv6/IPsec stack.
BUGS
The IPsec support is subject to change as the IPsec protocols develop.
There is no single standard for policy engine API,
so the policy engine API described herein is just for KAME implementation.
AH and tunnel mode encapsulation may not work as you might expect.
If you configure inbound
``require''
policy against AH tunnel or any IPsec encapsulating policy with AH
``
esp/tunnel/A-B/use
ah/transport/A-B/require
''
(like,)
tunneled packets will be rejected.
This is because we enforce policy check on inner packet on reception,
and AH authenticates encapsulating
(outer)
packet, not the encapsulated
(inner)
packet
(so for the receiving kernel there's no sign of authenticity.)
The issue will be solved when we revamp our policy engine to keep all the
packet decapsulation history.
Under certain condition,
truncated result may be raised from the kernel
against
SADB_DUMP
and
SADB_SPDDUMP
operation on
PF_KEY
socket.
This occurs if there are too many database entries in the kernel
and socket buffer for the
PF_KEY
socket is insufficient.
If you manipulate many IPsec key/policy database entries,
increase the size of socket buffer or use
sysctl(8)
interface.