NAME

ipsec_set_policy, ipsec_get_policylen, ipsec_dump_policy — manipulate IPsec policy specification structure from human-readable policy string

LIBRARY

IPsec Policy Control Library (libipsec, -lipsec)

SYNOPSIS

#include <netinet6/ipsec.h>

char *
ipsec_set_policy(char *policy, int len);

int
ipsec_get_policylen(char *buf);

char *
ipsec_dump_policy(char *buf, char *delim);

DESCRIPTION

ipsec_set_policy() generates an IPsec policy specification structure, namely struct sadb_x_policy and/or struct sadb_x_ipsecrequest from a human-readable policy specification. The policy specification must be given as a C string policy and its length len. ipsec_set_policy() will return a buffer with the corresponding IPsec policy specification structure. The buffer is dynamically allocated, and must be free(3)'d by the caller.

You can get the length of the generated buffer with ipsec_get_policylen() (i.e. for calling setsockopt(2)).

ipsec_dump_policy() converts an IPsec policy structure into human-readable form. Therefore, ipsec_dump_policy() can be regarded as the inverse function to ipsec_set_policy(). buf points to an IPsec policy structure, struct sadb_x_policy. delim is a delimiter string, which is usually a blank character. If you set delim to NULL, a single whitespace is assumed. ipsec_dump_policy() returns a pointer to a dynamically allocated string. It is the caller's responsibility to free(3) it.

policy is formatted as either of the following:

direction [priority specification] discard

direction must be in, out, or fwd. direction specifies in which direction the policy needs to be applied. The non-standard direction fwd is substituted with in on platforms which do not support forward policies.

priority specification is used to control the placement of the policy within the SPD. The policy position is determined by a signed integer where higher priorities indicate the policy is placed closer to the beginning of the list and lower priorities indicate the policy is placed closer to the end of the list. Policies with equal priorities are added at the end of the group of such policies.

Priority can only be specified when libipsec has been compiled against kernel headers that support policy priorities (Linux >= 2.6.6). It takes one of the following formats:

{priority,prio} offset: offset is an integer in the range -2147483647..214783648.
{priority,prio} base {+,-} offset: base is either low (-1073741824), def (0), or high (1073741824).
offset is an unsigned integer. It can be up to 1073741824 for positive offsets, and up to 1073741823 for negative offsets.

The interpretation of policy priority in these functions and the kernel DOES differ. The relationship between the two can be described as p(kernel) = 0x80000000 - p(func)

With discard policy, packets will be dropped if they match the policy.

direction [priority specification] entrust

entrust means to consult the SPD defined by setkey(8).

direction [priority specification] bypass

bypass means to bypass the IPsec processing. (the packet will be transmitted in clear). This is for privileged sockets.

direction [priority specification] ipsec request ...

ipsec means that the matching packets are subject to IPsec processing. ipsec can be followed by one or more request strings, which are formatted as below:

protocol / mode / src - dst [/level]

protocol is either ah, esp, or ipcomp.

mode is either transport or tunnel.

src and dst specifies the IPsec endpoint. src always means the “sending node” and dst always means the “receiving node”. Therefore, when direction is in, dst is this node and src is the other node (peer). If mode is transport, Both src and dst can be omitted.

level must be set to one of the following: default, use, require, or unique. default means that the kernel should consult the system default policy defined by sysctl(8), such as net.inet.ipsec.esp_trans_deflev. See ipsec(4) regarding the system default. use means that a relevant SA can be used when available, since the kernel may perform IPsec operation against packets when possible. In this case, packets can be transmitted in clear (when SA is not available), or encrypted (when SA is available). require means that a relevant SA is required, since the kernel must perform IPsec operation against packets. unique is the same as require, but adds the restriction that the SA for outbound traffic is used only for this policy. You may need the identifier in order to relate the policy and the SA when you define the SA by manual keying. You can put the decimal number as the identifier after unique like unique: number. number must be between 1 and 32767 . If the request string is kept unambiguous, level and slash prior to level can be omitted. However, it is encouraged to specify them explicitly to avoid unintended behavior. If level is omitted, it will be interpreted as default.

Note that there are slight differences to the specification of setkey(8). In the specification of setkey(8), both entrust and bypass are not used. Refer to setkey(8) for details.

Here are several examples (long lines are wrapped for readability):

in discard
out ipsec esp/transport//require
in ipsec ah/transport//require
out ipsec esp/tunnel/10.1.1.2-10.1.1.1/use
in ipsec ipcomp/transport//use
        esp/transport//use

RETURN VALUES

ipsec_set_policy() returns a pointer to the allocated buffer with the policy specification if successful; otherwise a NULL pointer is returned. ipsec_get_policylen() returns a positive value (meaning the buffer size) on success, and a negative value on errors. ipsec_dump_policy() returns a pointer to a dynamically allocated region on success, and NULL on errors.

HISTORY

The functions first appeared in the WIDE/KAME IPv6 protocol stack kit.