ADC Home > Reference Library > Reference > Darwin > KPI Reference
|
kpi_ipfilter.h |
Includes: | <sys/kernel_types.h> |
This header defines an API to attach IP filters. IP filters may be attached to intercept either IPv4 or IPv6 packets. The filters can intercept all IP packets in to and out of the host regardless of interface.
ipf_addv4 |
errno_t ipf_addv4( const struct ipf_filter*filter, ipfilter_t *filter_ref);
filter
filter_ref
0 on success otherwise the errno error.
Attaches an IPv4 ip filter.
ipf_addv6 |
errno_t ipf_addv6( const struct ipf_filter*filter, ipfilter_t *filter_ref);
filter
filter_ref
0 on success otherwise the errno error.
Attaches an IPv6 ip filter.
ipf_inject_input |
errno_t ipf_inject_input( mbuf_t data, ipfilter_t filter_ref);
data
filter_ref
0 on success otherwise the errno error.
Inject an IP packet as though it had just been
reassembled in ip_input. When re-injecting a packet intercepted
by the filter's ipf_input function, an IP filter can pass its
reference to avoid processing the packet twice. This also
prevents ip filters installed before this filter from
getting a chance to process the packet. If the filter modified
the packet, it should not specify the filter ref to give other
filters a chance to process the new packet.
Caller is responsible for freeing mbuf chain in the event that
ipf_inject_input returns an error.
ipf_inject_output |
errno_t ipf_inject_output( mbuf_t data, ipfilter_t filter_ref, ipf_pktopts_t options);
data
filter_ref
options
0 on success otherwise the errno error. ipf_inject_output will always free the mbuf.
Inject an IP packet as though it had just been sent to ip_output. When re-injecting a packet intercepted by the filter's ipf_output function, an IP filter can pass its reference to avoid processing the packet twice. This also prevents ip filters installed before this filter from getting a chance to process the packet. If the filter modified the packet, it should not specify the filter ref to give other filters a chance to process the new packet.
ipf_remove |
errno_t ipf_remove( ipfilter_t filter_ref);
filter_ref
0 on success otherwise the errno error.
Detaches an IPv4 or IPv6 filter.
ipf_detach_func |
typedef void (*ipf_detach_func)( void*cookie);
cookie
- The cookie specified when your filter was attached.
ipf_detach_func is called to notify your filter that it has been detached.
ipf_filter |
- ipf_filter
- opaque_ipfilter
- ipfilter_t
typedef struct opaque_ipfilter* ipfilter_t;
cookie
- A kext defined cookie that will be passed to all filter functions.
name
- A filter name used for debugging purposes.
ipf_input
- The filter function to handle inbound packets.
ipf_output
- The filter function to handle outbound packets.
ipf_detach
- The filter function to notify of a detach.
This structure is used to define an IP filter for use with the ipf_addv4 or ipf_addv6 function.
ipf_input_func |
typedef errno_t (*ipf_input_func)( void*cookie, mbuf_t *data, int offset, u_int8_t protocol);
cookie
- The cookie specified when your filter was attached.
data
- The reassembled ip packet, data will start at the ip header.
offset
- An offset to the next header (udp/tcp/icmp/esp/etc...).
protocol
- The protocol type (udp/tcp/icmp/etc...) of the IP packet
Return: 0 - The caller will continue with normal processing of the packet. EJUSTRETURN - The caller will stop processing the packet, the packet will not be freed. Anything Else - The caller will free the packet and stop processing.
ipf_input_func is used to filter incoming ip packets. The IP filter is called for packets from all interfaces. The filter is called between when the general IP processing is handled and when the packet is passed up to the next layer protocol such as udp or tcp. In the case of encapsulation, such as UDP in ESP (IPSec), your filter will be called once for ESP and then again for UDP. This will give your filter an opportunity to process the ESP header as well as the decrypted packet. Offset and protocol are used to determine where in the packet processing is currently occuring. If you're only interested in TCP or UDP packets, just return 0 if protocol doesn't match TCP or UDP.
ipf_output_func |
typedef errno_t (*ipf_output_func)( void*cookie, mbuf_t *data, ipf_pktopts_t options);
cookie
- The cookie specified when your filter was attached.
data
- The ip packet, will contain an IP header followed by the rest of the IP packet.
Return: 0 - The caller will continue with normal processing of the packet. EJUSTRETURN - The caller will stop processing the packet, the packet will not be freed. Anything Else - The caller will free the packet and stop processing.
ipf_output_func is used to filter outbound ip packets. The IP filter is called for packets to all interfaces. The filter is called before fragmentation and IPSec processing. If you need to change the destination IP address, call ipf_inject_output and return EJUSTRETURN.
ipfilter_t |
- ipf_filter
- opaque_ipfilter
- ipf_filter
typedef struct opaque_ipfilter* ipfilter_t;
cookie
- A kext defined cookie that will be passed to all filter functions.
name
- A filter name used for debugging purposes.
ipf_input
- The filter function to handle inbound packets.
ipf_output
- The filter function to handle outbound packets.
ipf_detach
- The filter function to notify of a detach.
This structure is used to define an IP filter for use with the ipf_addv4 or ipf_addv6 function.
ipf_filter |
- opaque_ipfilter
- ipf_filter
- ipfilter_t
struct ipf_filter { void*cookie; const char*name; ipf_input_func ipf_input; ipf_output_func ipf_output; ipf_detach_func ipf_detach; };
This structure is used to define an IP filter for use with the ipf_addv4 or ipf_addv6 function.
opaque_ipfilter |
- ipf_filter
- ipf_filter
- ipfilter_t
struct opaque_ipfilter;
This structure is used to define an IP filter for use with the ipf_addv4 or ipf_addv6 function.
|