Hello,
I am currently using the nfqueue library to process incoming and
outoing packets.
I now want to do some additional connection tracking and are looking
at the conntrack library for this. In general, I want to pretty much
the same as this guy [1], namely listen for TCP/UDP conntrack
DESTROY-events to remove additional info out of a list which also
tracks the connections.
Since I need to process both the nfqueue events, and the conntrack
events, I wanted implement the handling of both via polling, as
suggested in the answer to [1].
But while I know how to use the filedescriptor for the nfqueue, the
examples for libnetfilter_conntrack always use libmnl to receive
conntrack events and to run the callback on these events (and read
directly from an mnl_socket pointer) . So what I want to know is:
How can I handle conntrack events without using libmnl?
I have looked at the library setup page [2] and it seems like I can
just register a callback via "nfct_callback_register(..)", but how do
I run this callback?
I.e., for nfqueue, I would call "nfq_handle_packet()" after filling my
buffer for the packet with "recv()", but I don't know how what to do
with my conntrack buffer after I filled it via "rec()".
Basically I want to do something like this:
while ((ready = poll(fds, FD_COUNT,0)) && ready > 0 )
{
int i;
for (i = 0; i < FD_COUNT; i++) {
if ( fds[i].revents == POLLIN ) {
/* 0 is the nfqueue filedescriptor */
if (i == 0) {
int rv = recv(fd_nfq, buf_nfq, sizeof(buf_nfq), 0);
nfq_handle_packet(nf_h, buf_nfq, rv);
} else {
/* What to do here? */
recv(fd_ct, buf_ct, sizeof(buf_ct), 0);
}
}
}
}
Another question, when using the
nfct_open(u_int8_t subsys_id, unsigned subscriptions)
function, what is the "subscriptions" variable? I've seen it set to 0
while poking around, what are the options that I can subscribe to?
Regards,
Stephan
Links:
[1] https://www.spinics.net/lists/netfilter-devel/msg03308.html
--
To unsubscribe from this list: send the line "unsubscribe netfilter" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
