123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171 |
- /*
- * Copyright 2002 Damien Miller <djm@mindrot.org> 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 ``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 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.
- */
- /* $Id$ */
- #include "common.h"
- #include "log.h"
- #include "treetype.h"
- #include "softflowd.h"
- RCSID("$Id$");
- /*
- * This is the Cisco Netflow(tm) version 1 packet format
- * Based on:
- * http://www.cisco.com/univercd/cc/td/doc/product/rtrmgmt/nfc/nfc_3_0/nfc_ug/nfcform.htm
- */
- struct NF1_HEADER {
- u_int16_t version, flows;
- u_int32_t uptime_ms, time_sec, time_nanosec;
- };
- struct NF1_FLOW {
- u_int32_t src_ip, dest_ip, nexthop_ip;
- u_int16_t if_index_in, if_index_out;
- u_int32_t flow_packets, flow_octets;
- u_int32_t flow_start, flow_finish;
- u_int16_t src_port, dest_port;
- u_int16_t pad1;
- u_int8_t protocol, tos, tcp_flags;
- u_int8_t pad2, pad3, pad4;
- u_int32_t reserved1;
- #if 0
- u_int8_t reserved2; /* XXX: no longer used */
- #endif
- };
- /* Maximum of 24 flows per packet */
- #define NF1_MAXFLOWS 24
- #define NF1_MAXPACKET_SIZE (sizeof(struct NF1_HEADER) + \
- (NF1_MAXFLOWS * sizeof(struct NF1_FLOW)))
- /*
- * Given an array of expired flows, send netflow v1 report packets
- * Returns number of packets sent or -1 on error
- */
- int
- send_netflow_v1(struct FLOW **flows, int num_flows, int nfsock, u_int16_t ifidx,
- u_int64_t *flows_exported, struct timeval *system_boot_time,
- int verbose_flag)
- {
- struct timeval now;
- u_int32_t uptime_ms;
- u_int8_t packet[NF1_MAXPACKET_SIZE]; /* Maximum allowed packet size (24 flows) */
- struct NF1_HEADER *hdr = NULL;
- struct NF1_FLOW *flw = NULL;
- int i, j, offset, num_packets, err;
- socklen_t errsz;
-
- gettimeofday(&now, NULL);
- uptime_ms = timeval_sub_ms(&now, system_boot_time);
- hdr = (struct NF1_HEADER *)packet;
- for(num_packets = offset = j = i = 0; i < num_flows; i++) {
- if (j >= NF1_MAXFLOWS - 1) {
- if (verbose_flag)
- logit(LOG_DEBUG, "Sending flow packet len = %d", offset);
- hdr->flows = htons(hdr->flows);
- errsz = sizeof(err);
- getsockopt(nfsock, SOL_SOCKET, SO_ERROR,
- &err, &errsz); /* Clear ICMP errors */
- if (send(nfsock, packet, (size_t)offset, 0) == -1)
- return (-1);
- *flows_exported += j;
- j = 0;
- num_packets++;
- }
- if (j == 0) {
- memset(&packet, '\0', sizeof(packet));
- hdr->version = htons(1);
- hdr->flows = 0; /* Filled in as we go */
- hdr->uptime_ms = htonl(uptime_ms);
- hdr->time_sec = htonl(now.tv_sec);
- hdr->time_nanosec = htonl(now.tv_usec * 1000);
- offset = sizeof(*hdr);
- }
- flw = (struct NF1_FLOW *)(packet + offset);
- flw->if_index_in = flw->if_index_out = htons(ifidx);
- /* NetFlow v.1 doesn't do IPv6 */
- if (flows[i]->af != AF_INET)
- continue;
- if (flows[i]->octets[0] > 0) {
- flw->src_ip = flows[i]->addr[0].v4.s_addr;
- flw->dest_ip = flows[i]->addr[1].v4.s_addr;
- flw->src_port = flows[i]->port[0];
- flw->dest_port = flows[i]->port[1];
- flw->flow_packets = htonl(flows[i]->packets[0]);
- flw->flow_octets = htonl(flows[i]->octets[0]);
- flw->flow_start =
- htonl(timeval_sub_ms(&flows[i]->flow_start,
- system_boot_time));
- flw->flow_finish =
- htonl(timeval_sub_ms(&flows[i]->flow_last,
- system_boot_time));
- flw->protocol = flows[i]->protocol;
- flw->tcp_flags = flows[i]->tcp_flags[0];
- offset += sizeof(*flw);
- j++;
- hdr->flows++;
- }
- flw = (struct NF1_FLOW *)(packet + offset);
- flw->if_index_in = flw->if_index_out = htons(ifidx);
- if (flows[i]->octets[1] > 0) {
- flw->src_ip = flows[i]->addr[1].v4.s_addr;
- flw->dest_ip = flows[i]->addr[0].v4.s_addr;
- flw->src_port = flows[i]->port[1];
- flw->dest_port = flows[i]->port[0];
- flw->flow_packets = htonl(flows[i]->packets[1]);
- flw->flow_octets = htonl(flows[i]->octets[1]);
- flw->flow_start =
- htonl(timeval_sub_ms(&flows[i]->flow_start,
- system_boot_time));
- flw->flow_finish =
- htonl(timeval_sub_ms(&flows[i]->flow_last,
- system_boot_time));
- flw->protocol = flows[i]->protocol;
- flw->tcp_flags = flows[i]->tcp_flags[1];
- offset += sizeof(*flw);
- j++;
- hdr->flows++;
- }
- }
- /* Send any leftovers */
- if (j != 0) {
- if (verbose_flag)
- logit(LOG_DEBUG, "Sending flow packet len = %d", offset);
- hdr->flows = htons(hdr->flows);
- errsz = sizeof(err);
- getsockopt(nfsock, SOL_SOCKET, SO_ERROR,
- &err, &errsz); /* Clear ICMP errors */
- if (send(nfsock, packet, (size_t)offset, 0) == -1)
- return (-1);
- num_packets++;
- }
- *flows_exported += j;
- return (num_packets);
- }
|