/* * Syn-proxy step 2 logic * Receive client's 3-handshakes Ack packet, do cookie check * and then send syn to rs after creating a session. * */ int ip_vs_synproxy_ack_rcv(int af, struct sk_buff *skb, struct tcphdr *th, struct ip_vs_protocol *pp, struct ip_vs_conn **cpp, struct ip_vs_iphdr *iph, int *verdict) { struct ip_vs_synproxy_opt opt; struct ip_vs_service *svc; int res_cookie_check; /* * Don't check svc syn-proxy flag, as it may * be changed after syn-proxy step 1. */ if (!th->syn && th->ack && !th->rst && !th->fin && (svc = ip_vs_service_get(af, skb->mark, iph->protocol, &iph->daddr, th->dest))) { if (ip_vs_todrop()) { /* * It seems that we are very loaded. * We have to drop this packet :( */ ip_vs_service_put(svc); *verdict = NF_DROP; return 0; } if (sysctl_ip_vs_synproxy_defer && !syn_proxy_ack_has_data(skb, iph, th)) { /* update statistics */ IP_VS_INC_ESTATS(ip_vs_esmib, SYNPROXY_NULL_ACK); /* * When expecting ack packet with payload, * we get a pure ack, so have to drop it. */ ip_vs_service_put(svc); *verdict = NF_DROP; return 0; } /* * Import: set tcp hdr before cookie check, as it * will be used in cookie_check funcs. */ skb_set_transport_header(skb, iph->len); #ifdef CONFIG_IP_VS_IPV6 if (af == AF_INET6) { res_cookie_check = ip_vs_synproxy_v6_cookie_check(skb, ntohl (th-> ack_seq) - 1, &opt); } else #endif { res_cookie_check = ip_vs_synproxy_v4_cookie_check(skb, ntohl (th-> ack_seq) - 1, &opt); } if (!res_cookie_check) { /* update statistics */ IP_VS_INC_ESTATS(ip_vs_esmib, SYNPROXY_BAD_ACK); /* * Cookie check fail, drop it. */ IP_VS_DBG(6, "syn_cookie check failed seq=%u\n", ntohl(th->ack_seq) - 1); ip_vs_service_put(svc); *verdict = NF_DROP; return 0; } /* update statistics */ IP_VS_INC_ESTATS(ip_vs_esmib, SYNPROXY_OK_ACK); /* * Let the virtual server select a real server for the * incoming connection, and create a connection entry. */ *cpp = ip_vs_schedule(svc, skb, 1); if (!*cpp) { IP_VS_DBG(6, "ip_vs_schedule failed\n"); *verdict = ip_vs_leave(svc, skb, pp); return 0; } /* * Release service, we don't need it any more. */ ip_vs_service_put(svc); /* * Do anything but print a error msg when fail. * Because session will be correctly freed in ip_vs_conn_expire. */ if (!syn_proxy_send_rs_syn(af, th, *cpp, skb, pp, &opt)) { IP_VS_ERR_RL("syn_proxy_send_rs_syn failed!\n"); } /* count in the ack packet (STOLEN by synproxy) */ ip_vs_in_stats(*cpp, skb); /* * Active sesion timer, and dec refcnt. * Also stole the skb, and let caller return immediately. */ ip_vs_conn_put(*cpp); *verdict = NF_STOLEN; return 0; } return 1; }
/* * Check if it's for virtual services, look it up, * and send it on its way... */ static unsigned int ip_vs_in(unsigned int hooknum, struct sk_buff **pskb, const struct net_device *in, const struct net_device *out, int (*okfn)(struct sk_buff *)) { struct sk_buff *skb = *pskb; struct iphdr *iph; struct ip_vs_protocol *pp; struct ip_vs_conn *cp; int ret, restart; int ihl; /* * Big tappo: only PACKET_HOST (neither loopback nor mcasts) * ... don't know why 1st test DOES NOT include 2nd (?) */ if (unlikely(skb->pkt_type != PACKET_HOST || skb->dev == &loopback_dev || skb->sk)) { IP_VS_DBG(12, "packet type=%d proto=%d daddr=%d.%d.%d.%d ignored\n", skb->pkt_type, ip_hdr(skb)->protocol, NIPQUAD(ip_hdr(skb)->daddr)); return NF_ACCEPT; } iph = ip_hdr(skb); if (unlikely(iph->protocol == IPPROTO_ICMP)) { int related, verdict = ip_vs_in_icmp(pskb, &related, hooknum); if (related) return verdict; skb = *pskb; iph = ip_hdr(skb); } /* Protocol supported? */ pp = ip_vs_proto_get(iph->protocol); if (unlikely(!pp)) return NF_ACCEPT; ihl = iph->ihl << 2; /* * Check if the packet belongs to an existing connection entry */ cp = pp->conn_in_get(skb, pp, iph, ihl, 0); if (unlikely(!cp)) { int v; if (!pp->conn_schedule(skb, pp, &v, &cp)) return v; } if (unlikely(!cp)) { /* sorry, all this trouble for a no-hit :) */ IP_VS_DBG_PKT(12, pp, skb, 0, "packet continues traversal as normal"); return NF_ACCEPT; } IP_VS_DBG_PKT(11, pp, skb, 0, "Incoming packet"); /* Check the server status */ if (cp->dest && !(cp->dest->flags & IP_VS_DEST_F_AVAILABLE)) { /* the destination server is not available */ if (sysctl_ip_vs_expire_nodest_conn) { /* try to expire the connection immediately */ ip_vs_conn_expire_now(cp); } /* don't restart its timer, and silently drop the packet. */ __ip_vs_conn_put(cp); return NF_DROP; } ip_vs_in_stats(cp, skb); restart = ip_vs_set_state(cp, IP_VS_DIR_INPUT, skb, pp); if (cp->packet_xmit) ret = cp->packet_xmit(skb, cp, pp); /* do not touch skb anymore */ else { IP_VS_DBG_RL("warning: packet_xmit is null"); ret = NF_ACCEPT; } /* increase its packet counter and check if it is needed to be synchronized */ atomic_inc(&cp->in_pkts); if ((ip_vs_sync_state & IP_VS_STATE_MASTER) && (cp->protocol != IPPROTO_TCP || cp->state == IP_VS_TCP_S_ESTABLISHED) && (atomic_read(&cp->in_pkts) % sysctl_ip_vs_sync_threshold[1] == sysctl_ip_vs_sync_threshold[0])) ip_vs_sync_conn(cp); ip_vs_conn_put(cp); return ret; }
/* * Create syn packet and send it to rs. * ATTENTION: we also store syn skb in cp if syn retransimition * is tured on. */ static int syn_proxy_send_rs_syn(int af, const struct tcphdr *th, struct ip_vs_conn *cp, struct sk_buff *skb, struct ip_vs_protocol *pp, struct ip_vs_synproxy_opt *opt) { struct sk_buff *syn_skb; int tcp_hdr_size; __u8 tcp_flags = TCPCB_FLAG_SYN; unsigned int tcphoff; struct tcphdr *new_th; if (!cp->packet_xmit) { IP_VS_ERR_RL("warning: packet_xmit is null"); return 0; } syn_skb = alloc_skb(MAX_TCP_HEADER + 15, GFP_ATOMIC); if (unlikely(syn_skb == NULL)) { IP_VS_ERR_RL("alloc skb failed when send rs syn packet\n"); return 0; } /* Reserve space for headers */ skb_reserve(syn_skb, MAX_TCP_HEADER); tcp_hdr_size = (sizeof(struct tcphdr) + TCPOLEN_MSS + (opt->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0) + (opt->wscale_ok ? TCPOLEN_WSCALE_ALIGNED : 0) + /* SACK_PERM is in the place of NOP NOP of TS */ ((opt->sack_ok && !opt->tstamp_ok) ? TCPOLEN_SACKPERM_ALIGNED : 0)); new_th = (struct tcphdr *)skb_push(syn_skb, tcp_hdr_size); /* Compose tcp header */ skb_reset_transport_header(syn_skb); syn_skb->csum = 0; /* Set tcp hdr */ new_th->source = th->source; new_th->dest = th->dest; new_th->seq = htonl(ntohl(th->seq) - 1); new_th->ack_seq = 0; *(((__u16 *) new_th) + 6) = htons(((tcp_hdr_size >> 2) << 12) | tcp_flags); /* FIX_ME: what window should we use */ new_th->window = htons(5000); new_th->check = 0; new_th->urg_ptr = 0; new_th->urg = 0; new_th->ece = 0; new_th->cwr = 0; syn_proxy_syn_build_options((__be32 *) (new_th + 1), opt); /* * Set ip hdr * Attention: set source and dest addr to ack skb's. * we rely on packet_xmit func to do NATs thing. */ #ifdef CONFIG_IP_VS_IPV6 if (af == AF_INET6) { struct ipv6hdr *ack_iph = ipv6_hdr(skb); struct ipv6hdr *iph = (struct ipv6hdr *)skb_push(syn_skb, sizeof(struct ipv6hdr)); tcphoff = sizeof(struct ipv6hdr); skb_reset_network_header(syn_skb); memcpy(&iph->saddr, &ack_iph->saddr, sizeof(struct in6_addr)); memcpy(&iph->daddr, &ack_iph->daddr, sizeof(struct in6_addr)); iph->version = 6; iph->nexthdr = NEXTHDR_TCP; iph->payload_len = htons(tcp_hdr_size); iph->hop_limit = IPV6_DEFAULT_HOPLIMIT; new_th->check = 0; syn_skb->csum = skb_checksum(syn_skb, tcphoff, syn_skb->len - tcphoff, 0); new_th->check = csum_ipv6_magic(&iph->saddr, &iph->daddr, syn_skb->len - tcphoff, IPPROTO_TCP, syn_skb->csum); } else #endif { struct iphdr *ack_iph = ip_hdr(skb); u32 rtos = RT_TOS(ack_iph->tos); struct iphdr *iph = (struct iphdr *)skb_push(syn_skb, sizeof(struct iphdr)); tcphoff = sizeof(struct iphdr); skb_reset_network_header(syn_skb); *((__u16 *) iph) = htons((4 << 12) | (5 << 8) | (rtos & 0xff)); iph->tot_len = htons(syn_skb->len); iph->frag_off = htons(IP_DF); /* FIX_ME: what ttl shoule we use */ iph->ttl = IPDEFTTL; iph->protocol = IPPROTO_TCP; iph->saddr = ack_iph->saddr; iph->daddr = ack_iph->daddr; ip_send_check(iph); new_th->check = 0; syn_skb->csum = skb_checksum(syn_skb, tcphoff, syn_skb->len - tcphoff, 0); new_th->check = csum_tcpudp_magic(iph->saddr, iph->daddr, syn_skb->len - tcphoff, IPPROTO_TCP, syn_skb->csum); } /* Save syn_skb if syn retransmission is on */ if (sysctl_ip_vs_synproxy_syn_retry > 0) { cp->syn_skb = skb_copy(syn_skb, GFP_ATOMIC); atomic_set(&cp->syn_retry_max, sysctl_ip_vs_synproxy_syn_retry); } /* Save info for fast_response_xmit */ if(sysctl_ip_vs_fast_xmit && skb->dev && likely(skb->dev->type == ARPHRD_ETHER) && skb_mac_header_was_set(skb)) { struct ethhdr *eth = (struct ethhdr *)skb_mac_header(skb); if(likely(cp->indev == NULL)) { cp->indev = skb->dev; dev_hold(cp->indev); } if (unlikely(cp->indev != skb->dev)) { dev_put(cp->indev); cp->indev = skb->dev; dev_hold(cp->indev); } memcpy(cp->src_hwaddr, eth->h_source, ETH_ALEN); memcpy(cp->dst_hwaddr, eth->h_dest, ETH_ALEN); IP_VS_INC_ESTATS(ip_vs_esmib, FAST_XMIT_SYNPROXY_SAVE); IP_VS_DBG_RL("syn_proxy_send_rs_syn netdevice:%s\n", netdev_name(skb->dev)); } /* count in the syn packet */ ip_vs_in_stats(cp, skb); /* If xmit failed, syn_skb will be freed correctly. */ cp->packet_xmit(syn_skb, cp, pp); return 1; }
/* * Pass or drop the packet. * Called by ip_vs_in, when the virtual service is available but * no destination is available for a new connection. */ int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb, struct ip_vs_protocol *pp) { __be16 _ports[2], *pptr; struct iphdr *iph = ip_hdr(skb); pptr = skb_header_pointer(skb, iph->ihl*4, sizeof(_ports), _ports); if (pptr == NULL) { ip_vs_service_put(svc); return NF_DROP; } /* if it is fwmark-based service, the cache_bypass sysctl is up and the destination is RTN_UNICAST (and not local), then create a cache_bypass connection entry */ if (sysctl_ip_vs_cache_bypass && svc->fwmark && (inet_addr_type(iph->daddr) == RTN_UNICAST)) { int ret, cs; struct ip_vs_conn *cp; ip_vs_service_put(svc); /* create a new connection entry */ IP_VS_DBG(6, "ip_vs_leave: create a cache_bypass entry\n"); cp = ip_vs_conn_new(iph->protocol, iph->saddr, pptr[0], iph->daddr, pptr[1], 0, 0, IP_VS_CONN_F_BYPASS, NULL); if (cp == NULL) return NF_DROP; /* statistics */ ip_vs_in_stats(cp, skb); /* set state */ cs = ip_vs_set_state(cp, IP_VS_DIR_INPUT, skb, pp); /* transmit the first SYN packet */ ret = cp->packet_xmit(skb, cp, pp); /* do not touch skb anymore */ atomic_inc(&cp->in_pkts); ip_vs_conn_put(cp); return ret; } /* * When the virtual ftp service is presented, packets destined * for other services on the VIP may get here (except services * listed in the ipvs table), pass the packets, because it is * not ipvs job to decide to drop the packets. */ if ((svc->port == FTPPORT) && (pptr[1] != FTPPORT)) { ip_vs_service_put(svc); return NF_ACCEPT; } ip_vs_service_put(svc); /* * Notify the client that the destination is unreachable, and * release the socket buffer. * Since it is in IP layer, the TCP socket is not actually * created, the TCP RST packet cannot be sent, instead that * ICMP_PORT_UNREACH is sent here no matter it is TCP/UDP. --WZ */ icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PORT_UNREACH, 0); return NF_DROP; }
/* * Handle ICMP messages in the outside-to-inside direction (incoming). * Find any that might be relevant, check against existing connections, * forward to the right destination host if relevant. * Currently handles error types - unreachable, quench, ttl exceeded. */ static int ip_vs_in_icmp(struct sk_buff **pskb, int *related, unsigned int hooknum) { struct sk_buff *skb = *pskb; struct iphdr *iph; struct icmphdr _icmph, *ic; struct iphdr _ciph, *cih; /* The ip header contained within the ICMP */ struct ip_vs_conn *cp; struct ip_vs_protocol *pp; unsigned int offset, ihl, verdict; *related = 1; /* reassemble IP fragments */ if (ip_hdr(skb)->frag_off & htons(IP_MF | IP_OFFSET)) { skb = ip_vs_gather_frags(skb, hooknum == NF_IP_LOCAL_IN ? IP_DEFRAG_VS_IN : IP_DEFRAG_VS_FWD); if (!skb) return NF_STOLEN; *pskb = skb; } iph = ip_hdr(skb); offset = ihl = iph->ihl * 4; ic = skb_header_pointer(skb, offset, sizeof(_icmph), &_icmph); if (ic == NULL) return NF_DROP; IP_VS_DBG(12, "Incoming ICMP (%d,%d) %u.%u.%u.%u->%u.%u.%u.%u\n", ic->type, ntohs(icmp_id(ic)), NIPQUAD(iph->saddr), NIPQUAD(iph->daddr)); /* * Work through seeing if this is for us. * These checks are supposed to be in an order that means easy * things are checked first to speed up processing.... however * this means that some packets will manage to get a long way * down this stack and then be rejected, but that's life. */ if ((ic->type != ICMP_DEST_UNREACH) && (ic->type != ICMP_SOURCE_QUENCH) && (ic->type != ICMP_TIME_EXCEEDED)) { *related = 0; return NF_ACCEPT; } /* Now find the contained IP header */ offset += sizeof(_icmph); cih = skb_header_pointer(skb, offset, sizeof(_ciph), &_ciph); if (cih == NULL) return NF_ACCEPT; /* The packet looks wrong, ignore */ pp = ip_vs_proto_get(cih->protocol); if (!pp) return NF_ACCEPT; /* Is the embedded protocol header present? */ if (unlikely(cih->frag_off & htons(IP_OFFSET) && pp->dont_defrag)) return NF_ACCEPT; IP_VS_DBG_PKT(11, pp, skb, offset, "Checking incoming ICMP for"); offset += cih->ihl * 4; /* The embedded headers contain source and dest in reverse order */ cp = pp->conn_in_get(skb, pp, cih, offset, 1); if (!cp) return NF_ACCEPT; verdict = NF_DROP; /* Ensure the checksum is correct */ if (!skb_csum_unnecessary(skb) && ip_vs_checksum_complete(skb, ihl)) { /* Failed checksum! */ IP_VS_DBG(1, "Incoming ICMP: failed checksum from %d.%d.%d.%d!\n", NIPQUAD(iph->saddr)); goto out; } /* do the statistics and put it back */ ip_vs_in_stats(cp, skb); if (IPPROTO_TCP == cih->protocol || IPPROTO_UDP == cih->protocol) offset += 2 * sizeof(__u16); verdict = ip_vs_icmp_xmit(skb, cp, pp, offset); /* do not touch skb anymore */ out: __ip_vs_conn_put(cp); return verdict; }
/* * Handle ICMP messages in the outside-to-inside direction (incoming) * and sometimes in outgoing direction from ip_vs_forward_icmp. * Find any that might be relevant, check against existing connections, * forward to the right destination host if relevant. * Currently handles error types - unreachable, quench, ttl exceeded. */ static int ip_vs_in_icmp(struct sk_buff **skb_p) { struct sk_buff *skb = *skb_p; struct iphdr *iph; struct icmphdr *icmph; struct iphdr *ciph; /* The ip header contained within the ICMP */ __u16 *pptr; /* port numbers from TCP/UDP contained header */ unsigned short len; unsigned short clen, csize; struct ip_vs_conn *cp; struct rtable *rt; /* Route to the other host */ int mtu; if (skb_is_nonlinear(skb)) { if (skb_linearize(skb, GFP_ATOMIC) != 0) return NF_DROP; } iph = skb->nh.iph; ip_send_check(iph); icmph = (struct icmphdr *)((char *)iph + (iph->ihl << 2)); len = ntohs(iph->tot_len) - (iph->ihl<<2); if (len < sizeof(struct icmphdr)) return NF_DROP; IP_VS_DBG(12, "icmp in (%d,%d) %u.%u.%u.%u -> %u.%u.%u.%u\n", icmph->type, ntohs(icmp_id(icmph)), NIPQUAD(iph->saddr), NIPQUAD(iph->daddr)); if ((icmph->type != ICMP_DEST_UNREACH) && (icmph->type != ICMP_SOURCE_QUENCH) && (icmph->type != ICMP_TIME_EXCEEDED)) return NF_ACCEPT; /* * If we get here we have an ICMP error of one of the above 3 types * Now find the contained IP header */ clen = len - sizeof(struct icmphdr); if (clen < sizeof(struct iphdr)) return NF_DROP; ciph = (struct iphdr *) (icmph + 1); csize = ciph->ihl << 2; if (clen < csize) return NF_DROP; /* We are only interested ICMPs generated from TCP or UDP packets */ if (ciph->protocol != IPPROTO_UDP && ciph->protocol != IPPROTO_TCP) return NF_ACCEPT; /* Skip non-first embedded TCP/UDP fragments */ if (ciph->frag_off & __constant_htons(IP_OFFSET)) return NF_ACCEPT; /* We need at least TCP/UDP ports here */ if (clen < csize + sizeof(struct udphdr)) return NF_DROP; /* Ensure the checksum is correct */ if (ip_compute_csum((unsigned char *) icmph, len)) { /* Failed checksum! */ IP_VS_ERR_RL("incoming ICMP: failed checksum from " "%d.%d.%d.%d!\n", NIPQUAD(iph->saddr)); return NF_DROP; } pptr = (__u16 *)&(((char *)ciph)[csize]); IP_VS_DBG(11, "Handling incoming ICMP for " "%u.%u.%u.%u:%d -> %u.%u.%u.%u:%d\n", NIPQUAD(ciph->saddr), ntohs(pptr[0]), NIPQUAD(ciph->daddr), ntohs(pptr[1])); /* This is pretty much what ip_vs_conn_in_get() does, except parameters are in the reverse order */ cp = ip_vs_conn_in_get(ciph->protocol, ciph->daddr, pptr[1], ciph->saddr, pptr[0]); if (cp == NULL) return NF_ACCEPT; ip_vs_in_stats(cp, skb); /* The ICMP packet for VS/TUN, VS/DR and LOCALNODE will be forwarded directly here, because there is no need to translate address/port back */ if (IP_VS_FWD_METHOD(cp) != IP_VS_CONN_F_MASQ) { int ret; if (cp->packet_xmit) ret = cp->packet_xmit(skb, cp); else ret = NF_ACCEPT; atomic_inc(&cp->in_pkts); ip_vs_conn_put(cp); return ret; } /* * mangle and send the packet here */ if (!(rt = __ip_vs_get_out_rt(cp, RT_TOS(iph->tos)))) goto tx_error_icmp; /* MTU checking */ mtu = rt->u.dst.pmtu; if ((skb->len > mtu) && (iph->frag_off&__constant_htons(IP_DF))) { ip_rt_put(rt); icmp_send(skb, ICMP_DEST_UNREACH,ICMP_FRAG_NEEDED, htonl(mtu)); IP_VS_DBG_RL("ip_vs_in_icmp(): frag needed\n"); goto tx_error; } /* drop old route */ dst_release(skb->dst); skb->dst = &rt->u.dst; /* copy-on-write the packet before mangling it */ if (ip_vs_skb_cow(skb, rt->u.dst.dev->hard_header_len, &iph, (unsigned char**)&icmph)) { ip_vs_conn_put(cp); return NF_DROP; } ciph = (struct iphdr *) (icmph + 1); pptr = (__u16 *)&(((char *)ciph)[csize]); /* The ICMP packet for VS/NAT must be written to correct addresses before being forwarded to the right server */ /* First change the dest IP address, and recalc checksum */ iph->daddr = cp->daddr; ip_send_check(iph); /* Now change the *source* address in the contained IP */ ciph->saddr = cp->daddr; ip_send_check(ciph); /* the TCP/UDP source port - cannot redo check */ pptr[0] = cp->dport; /* And finally the ICMP checksum */ icmph->checksum = 0; icmph->checksum = ip_compute_csum((unsigned char *) icmph, len); skb->ip_summed = CHECKSUM_UNNECESSARY; IP_VS_DBG(11, "Forwarding incoming ICMP to " "%u.%u.%u.%u:%d -> %u.%u.%u.%u:%d\n", NIPQUAD(ciph->saddr), ntohs(pptr[0]), NIPQUAD(ciph->daddr), ntohs(pptr[1])); #ifdef CONFIG_NETFILTER_DEBUG skb->nf_debug = 1 << NF_IP_LOCAL_OUT; #endif /* CONFIG_NETFILTER_DEBUG */ ip_send(skb); ip_vs_conn_put(cp); return NF_STOLEN; tx_error_icmp: dst_link_failure(skb); tx_error: dev_kfree_skb(skb); ip_vs_conn_put(cp); return NF_STOLEN; }
/* * Pass or drop the packet. * Called by ip_vs_in, when the virtual service is available but * no destination is available for a new connection. */ static int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb) { struct iphdr *iph = skb->nh.iph; __u16 *portp = (__u16 *)&(((char *)iph)[iph->ihl*4]); /* if it is fwmark-based service, the cache_bypass sysctl is up and the destination is RTN_UNICAST (and not local), then create a cache_bypass connection entry */ if (sysctl_ip_vs_cache_bypass && svc->fwmark && (inet_addr_type(iph->daddr) == RTN_UNICAST)) { int ret; struct ip_vs_conn *cp; ip_vs_service_put(svc); /* create a new connection entry */ IP_VS_DBG(6, "ip_vs_leave: create a cache_bypass entry\n"); cp = ip_vs_conn_new(iph->protocol, iph->saddr, portp[0], iph->daddr, portp[1], 0, 0, IP_VS_CONN_F_BYPASS, NULL); if (cp == NULL) { kfree_skb(skb); return NF_STOLEN; } /* statistics */ ip_vs_in_stats(cp, skb); /* set state */ ip_vs_set_state(cp, VS_STATE_INPUT, iph, portp); /* transmit the first SYN packet */ ret = cp->packet_xmit(skb, cp); atomic_inc(&cp->in_pkts); ip_vs_conn_put(cp); return ret; } /* * When the virtual ftp service is presented, packets destined * for other services on the VIP may get here (except services * listed in the ipvs table), pass the packets, because it is * not ipvs job to decide to drop the packets. */ if ((svc->port == FTPPORT) && (portp[1] != FTPPORT)) { ip_vs_service_put(svc); return NF_ACCEPT; } ip_vs_service_put(svc); /* * Notify the client that the destination is unreachable, and * release the socket buffer. * Since it is in IP layer, the TCP socket is not actually * created, the TCP RST packet cannot be sent, instead that * ICMP_PORT_UNREACH is sent here no matter it is TCP/UDP. --WZ */ icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PORT_UNREACH, 0); kfree_skb(skb); return NF_STOLEN; }
/* * Check if it's for virtual services, look it up, * and send it on its way... */ static unsigned int ip_vs_in(unsigned int hooknum, struct sk_buff **skb_p, const struct net_device *in, const struct net_device *out, int (*okfn)(struct sk_buff *)) { struct sk_buff *skb = *skb_p; struct iphdr *iph = skb->nh.iph; union ip_vs_tphdr h; struct ip_vs_conn *cp; struct ip_vs_service *svc; int ihl; int ret; /* * Big tappo: only PACKET_HOST (nor loopback neither mcasts) * ... don't know why 1st test DOES NOT include 2nd (?) */ if (skb->pkt_type != PACKET_HOST || skb->dev == &loopback_dev) { IP_VS_DBG(12, "packet type=%d proto=%d daddr=%d.%d.%d.%d ignored\n", skb->pkt_type, iph->protocol, NIPQUAD(iph->daddr)); return NF_ACCEPT; } if (iph->protocol == IPPROTO_ICMP) return ip_vs_in_icmp(skb_p); /* let it go if other IP protocols */ if (iph->protocol != IPPROTO_TCP && iph->protocol != IPPROTO_UDP) return NF_ACCEPT; /* make sure that protocol header available in skb data area, note that skb data area may be reallocated. */ ihl = iph->ihl << 2; if (ip_vs_header_check(skb, iph->protocol, ihl) == -1) return NF_DROP; iph = skb->nh.iph; h.raw = (char*) iph + ihl; /* * Check if the packet belongs to an existing connection entry */ cp = ip_vs_conn_in_get(iph->protocol, iph->saddr, h.portp[0], iph->daddr, h.portp[1]); if (!cp && (h.th->syn || (iph->protocol!=IPPROTO_TCP)) && (svc = ip_vs_service_get(skb->nfmark, iph->protocol, iph->daddr, h.portp[1]))) { if (ip_vs_todrop()) { /* * It seems that we are very loaded. * We have to drop this packet :( */ ip_vs_service_put(svc); return NF_DROP; } /* * Let the virtual server select a real server for the * incoming connection, and create a connection entry. */ cp = ip_vs_schedule(svc, iph); if (!cp) return ip_vs_leave(svc, skb); ip_vs_conn_stats(cp, svc); ip_vs_service_put(svc); } if (!cp) { /* sorry, all this trouble for a no-hit :) */ IP_VS_DBG(12, "packet for %s %d.%d.%d.%d:%d continue " "traversal as normal.\n", ip_vs_proto_name(iph->protocol), NIPQUAD(iph->daddr), ntohs(h.portp[1])); return NF_ACCEPT; } IP_VS_DBG(11, "Incoming %s %u.%u.%u.%u:%d->%u.%u.%u.%u:%d\n", ip_vs_proto_name(iph->protocol), NIPQUAD(iph->saddr), ntohs(h.portp[0]), NIPQUAD(iph->daddr), ntohs(h.portp[1])); /* Check the server status */ if (cp->dest && !(cp->dest->flags & IP_VS_DEST_F_AVAILABLE)) { /* the destination server is not available */ if (sysctl_ip_vs_expire_nodest_conn) { /* try to expire the connection immediately */ ip_vs_conn_expire_now(cp); } else { /* don't restart its timer, and silently drop the packet. */ __ip_vs_conn_put(cp); } return NF_DROP; } ip_vs_in_stats(cp, skb); ip_vs_set_state(cp, VS_STATE_INPUT, iph, h.portp); if (cp->packet_xmit) ret = cp->packet_xmit(skb, cp); else { IP_VS_DBG_RL("warning: packet_xmit is null"); ret = NF_ACCEPT; } /* increase its packet counter and check if it is needed to be synchronized */ atomic_inc(&cp->in_pkts); if (ip_vs_sync_state & IP_VS_STATE_MASTER && (cp->protocol != IPPROTO_TCP || cp->state == IP_VS_S_ESTABLISHED) && (atomic_read(&cp->in_pkts) % 50 == sysctl_ip_vs_sync_threshold)) ip_vs_sync_conn(cp); ip_vs_conn_put(cp); return ret; }
static struct ip_vs_conn * ip_vs_sched_persist(struct ip_vs_service *svc, struct sk_buff *skb, __be16 src_port, __be16 dst_port, int *ignored) { struct ip_vs_conn *cp = NULL; struct ip_vs_iphdr iph; struct ip_vs_dest *dest; struct ip_vs_conn *ct; __be16 dport = 0; unsigned int flags; struct ip_vs_conn_param param; const union nf_inet_addr fwmark = { .ip = htonl(svc->fwmark) }; union nf_inet_addr snet; ip_vs_fill_iphdr(svc->af, skb_network_header(skb), &iph); #ifdef CONFIG_IP_VS_IPV6 if (svc->af == AF_INET6) ipv6_addr_prefix(&snet.in6, &iph.saddr.in6, svc->netmask); else #endif snet.ip = iph.saddr.ip & svc->netmask; IP_VS_DBG_BUF(6, "p-schedule: src %s:%u dest %s:%u " "mnet %s\n", IP_VS_DBG_ADDR(svc->af, &iph.saddr), ntohs(src_port), IP_VS_DBG_ADDR(svc->af, &iph.daddr), ntohs(dst_port), IP_VS_DBG_ADDR(svc->af, &snet)); { int protocol = iph.protocol; const union nf_inet_addr *vaddr = &iph.daddr; __be16 vport = 0; if (dst_port == svc->port) { if (svc->port != FTPPORT) vport = dst_port; } else { if (svc->fwmark) { protocol = IPPROTO_IP; vaddr = &fwmark; } } if (ip_vs_conn_fill_param_persist(svc, skb, protocol, &snet, 0, vaddr, vport, ¶m) < 0) { *ignored = -1; return NULL; } } ct = ip_vs_ct_in_get(¶m); if (!ct || !ip_vs_check_template(ct)) { dest = svc->scheduler->schedule(svc, skb); if (!dest) { IP_VS_DBG(1, "p-schedule: no dest found.\n"); kfree(param.pe_data); *ignored = 0; return NULL; } if (dst_port == svc->port && svc->port != FTPPORT) dport = dest->port; ct = ip_vs_conn_new(¶m, &dest->addr, dport, IP_VS_CONN_F_TEMPLATE, dest, skb->mark); if (ct == NULL) { kfree(param.pe_data); *ignored = -1; return NULL; } ct->timeout = svc->timeout; } else { dest = ct->dest; kfree(param.pe_data); } dport = dst_port; if (dport == svc->port && dest->port) dport = dest->port; flags = (svc->flags & IP_VS_SVC_F_ONEPACKET && iph.protocol == IPPROTO_UDP)? IP_VS_CONN_F_ONE_PACKET : 0; ip_vs_conn_fill_param(svc->net, svc->af, iph.protocol, &iph.saddr, src_port, &iph.daddr, dst_port, ¶m); cp = ip_vs_conn_new(¶m, &dest->addr, dport, flags, dest, skb->mark); if (cp == NULL) { ip_vs_conn_put(ct); *ignored = -1; return NULL; } ip_vs_control_add(cp, ct); ip_vs_conn_put(ct); ip_vs_conn_stats(cp, svc); return cp; } struct ip_vs_conn * ip_vs_schedule(struct ip_vs_service *svc, struct sk_buff *skb, struct ip_vs_proto_data *pd, int *ignored) { struct ip_vs_protocol *pp = pd->pp; struct ip_vs_conn *cp = NULL; struct ip_vs_iphdr iph; struct ip_vs_dest *dest; __be16 _ports[2], *pptr; unsigned int flags; *ignored = 1; ip_vs_fill_iphdr(svc->af, skb_network_header(skb), &iph); pptr = skb_header_pointer(skb, iph.len, sizeof(_ports), _ports); if (pptr == NULL) return NULL; if (pptr[0] == FTPDATA) { IP_VS_DBG_PKT(12, svc->af, pp, skb, 0, "Not scheduling FTPDATA"); return NULL; } if ((!skb->dev || skb->dev->flags & IFF_LOOPBACK) && (cp = pp->conn_in_get(svc->af, skb, &iph, iph.len, 1))) { IP_VS_DBG_PKT(12, svc->af, pp, skb, 0, "Not scheduling reply for existing connection"); __ip_vs_conn_put(cp); return NULL; } if (svc->flags & IP_VS_SVC_F_PERSISTENT) return ip_vs_sched_persist(svc, skb, pptr[0], pptr[1], ignored); *ignored = 0; if (!svc->fwmark && pptr[1] != svc->port) { if (!svc->port) pr_err("Schedule: port zero only supported " "in persistent services, " "check your ipvs configuration\n"); return NULL; } dest = svc->scheduler->schedule(svc, skb); if (dest == NULL) { IP_VS_DBG(1, "Schedule: no dest found.\n"); return NULL; } flags = (svc->flags & IP_VS_SVC_F_ONEPACKET && iph.protocol == IPPROTO_UDP)? IP_VS_CONN_F_ONE_PACKET : 0; { struct ip_vs_conn_param p; ip_vs_conn_fill_param(svc->net, svc->af, iph.protocol, &iph.saddr, pptr[0], &iph.daddr, pptr[1], &p); cp = ip_vs_conn_new(&p, &dest->addr, dest->port ? dest->port : pptr[1], flags, dest, skb->mark); if (!cp) { *ignored = -1; return NULL; } } IP_VS_DBG_BUF(6, "Schedule fwd:%c c:%s:%u v:%s:%u " "d:%s:%u conn->flags:%X conn->refcnt:%d\n", ip_vs_fwd_tag(cp), IP_VS_DBG_ADDR(svc->af, &cp->caddr), ntohs(cp->cport), IP_VS_DBG_ADDR(svc->af, &cp->vaddr), ntohs(cp->vport), IP_VS_DBG_ADDR(svc->af, &cp->daddr), ntohs(cp->dport), cp->flags, atomic_read(&cp->refcnt)); ip_vs_conn_stats(cp, svc); return cp; } int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb, struct ip_vs_proto_data *pd) { __be16 _ports[2], *pptr; struct ip_vs_iphdr iph; #ifdef CONFIG_SYSCTL struct net *net; struct netns_ipvs *ipvs; int unicast; #endif ip_vs_fill_iphdr(svc->af, skb_network_header(skb), &iph); pptr = skb_header_pointer(skb, iph.len, sizeof(_ports), _ports); if (pptr == NULL) { ip_vs_service_put(svc); return NF_DROP; } #ifdef CONFIG_SYSCTL net = skb_net(skb); #ifdef CONFIG_IP_VS_IPV6 if (svc->af == AF_INET6) unicast = ipv6_addr_type(&iph.daddr.in6) & IPV6_ADDR_UNICAST; else #endif unicast = (inet_addr_type(net, iph.daddr.ip) == RTN_UNICAST); ipvs = net_ipvs(net); if (ipvs->sysctl_cache_bypass && svc->fwmark && unicast) { int ret; struct ip_vs_conn *cp; unsigned int flags = (svc->flags & IP_VS_SVC_F_ONEPACKET && iph.protocol == IPPROTO_UDP)? IP_VS_CONN_F_ONE_PACKET : 0; union nf_inet_addr daddr = { .all = { 0, 0, 0, 0 } }; ip_vs_service_put(svc); IP_VS_DBG(6, "%s(): create a cache_bypass entry\n", __func__); { struct ip_vs_conn_param p; ip_vs_conn_fill_param(svc->net, svc->af, iph.protocol, &iph.saddr, pptr[0], &iph.daddr, pptr[1], &p); cp = ip_vs_conn_new(&p, &daddr, 0, IP_VS_CONN_F_BYPASS | flags, NULL, skb->mark); if (!cp) return NF_DROP; } ip_vs_in_stats(cp, skb); ip_vs_set_state(cp, IP_VS_DIR_INPUT, skb, pd); ret = cp->packet_xmit(skb, cp, pd->pp); atomic_inc(&cp->in_pkts); ip_vs_conn_put(cp); return ret; }
/* * IPVS persistent scheduling function * It creates a connection entry according to its template if exists, * or selects a server and creates a connection entry plus a template. * Locking: we are svc user (svc->refcnt), so we hold all dests too * Protocols supported: TCP, UDP */ static struct ip_vs_conn * ip_vs_sched_persist(struct ip_vs_service *svc, struct sk_buff *skb, __be16 src_port, __be16 dst_port, int *ignored, struct ip_vs_iphdr *iph) { struct ip_vs_conn *cp = NULL; struct ip_vs_dest *dest; struct ip_vs_conn *ct; __be16 dport = 0; /* destination port to forward */ unsigned int flags; struct ip_vs_conn_param param; const union nf_inet_addr fwmark = { .ip = htonl(svc->fwmark) }; union nf_inet_addr snet; /* source network of the client, after masking */ /* Mask saddr with the netmask to adjust template granularity */ #ifdef CONFIG_IP_VS_IPV6 if (svc->af == AF_INET6) ipv6_addr_prefix(&snet.in6, &iph->saddr.in6, (__force __u32) svc->netmask); else #endif snet.ip = iph->saddr.ip & svc->netmask; IP_VS_DBG_BUF(6, "p-schedule: src %s:%u dest %s:%u " "mnet %s\n", IP_VS_DBG_ADDR(svc->af, &iph->saddr), ntohs(src_port), IP_VS_DBG_ADDR(svc->af, &iph->daddr), ntohs(dst_port), IP_VS_DBG_ADDR(svc->af, &snet)); /* * As far as we know, FTP is a very complicated network protocol, and * it uses control connection and data connections. For active FTP, * FTP server initialize data connection to the client, its source port * is often 20. For passive FTP, FTP server tells the clients the port * that it passively listens to, and the client issues the data * connection. In the tunneling or direct routing mode, the load * balancer is on the client-to-server half of connection, the port * number is unknown to the load balancer. So, a conn template like * <caddr, 0, vaddr, 0, daddr, 0> is created for persistent FTP * service, and a template like <caddr, 0, vaddr, vport, daddr, dport> * is created for other persistent services. */ { int protocol = iph->protocol; const union nf_inet_addr *vaddr = &iph->daddr; __be16 vport = 0; if (dst_port == svc->port) { /* non-FTP template: * <protocol, caddr, 0, vaddr, vport, daddr, dport> * FTP template: * <protocol, caddr, 0, vaddr, 0, daddr, 0> */ if (svc->port != FTPPORT) vport = dst_port; } else { /* Note: persistent fwmark-based services and * persistent port zero service are handled here. * fwmark template: * <IPPROTO_IP,caddr,0,fwmark,0,daddr,0> * port zero template: * <protocol,caddr,0,vaddr,0,daddr,0> */ if (svc->fwmark) { protocol = IPPROTO_IP; vaddr = &fwmark; } } /* return *ignored = -1 so NF_DROP can be used */ if (ip_vs_conn_fill_param_persist(svc, skb, protocol, &snet, 0, vaddr, vport, ¶m) < 0) { *ignored = -1; return NULL; } } /* Check if a template already exists */ ct = ip_vs_ct_in_get(¶m); if (!ct || !ip_vs_check_template(ct)) { struct ip_vs_scheduler *sched; /* * No template found or the dest of the connection * template is not available. * return *ignored=0 i.e. ICMP and NF_DROP */ sched = rcu_dereference(svc->scheduler); if (sched) { /* read svc->sched_data after svc->scheduler */ smp_rmb(); dest = sched->schedule(svc, skb, iph); } else { dest = NULL; } if (!dest) { IP_VS_DBG(1, "p-schedule: no dest found.\n"); kfree(param.pe_data); *ignored = 0; return NULL; } if (dst_port == svc->port && svc->port != FTPPORT) dport = dest->port; /* Create a template * This adds param.pe_data to the template, * and thus param.pe_data will be destroyed * when the template expires */ ct = ip_vs_conn_new(¶m, dest->af, &dest->addr, dport, IP_VS_CONN_F_TEMPLATE, dest, skb->mark); if (ct == NULL) { kfree(param.pe_data); *ignored = -1; return NULL; } ct->timeout = svc->timeout; } else { /* set destination with the found template */ dest = ct->dest; kfree(param.pe_data); } dport = dst_port; if (dport == svc->port && dest->port) dport = dest->port; flags = (svc->flags & IP_VS_SVC_F_ONEPACKET && iph->protocol == IPPROTO_UDP) ? IP_VS_CONN_F_ONE_PACKET : 0; /* * Create a new connection according to the template */ ip_vs_conn_fill_param(svc->net, svc->af, iph->protocol, &iph->saddr, src_port, &iph->daddr, dst_port, ¶m); cp = ip_vs_conn_new(¶m, dest->af, &dest->addr, dport, flags, dest, skb->mark); if (cp == NULL) { ip_vs_conn_put(ct); *ignored = -1; return NULL; } /* * Add its control */ ip_vs_control_add(cp, ct); ip_vs_conn_put(ct); ip_vs_conn_stats(cp, svc); return cp; } /* * IPVS main scheduling function * It selects a server according to the virtual service, and * creates a connection entry. * Protocols supported: TCP, UDP * * Usage of *ignored * * 1 : protocol tried to schedule (eg. on SYN), found svc but the * svc/scheduler decides that this packet should be accepted with * NF_ACCEPT because it must not be scheduled. * * 0 : scheduler can not find destination, so try bypass or * return ICMP and then NF_DROP (ip_vs_leave). * * -1 : scheduler tried to schedule but fatal error occurred, eg. * ip_vs_conn_new failure (ENOMEM) or ip_vs_sip_fill_param * failure such as missing Call-ID, ENOMEM on skb_linearize * or pe_data. In this case we should return NF_DROP without * any attempts to send ICMP with ip_vs_leave. */ struct ip_vs_conn * ip_vs_schedule(struct ip_vs_service *svc, struct sk_buff *skb, struct ip_vs_proto_data *pd, int *ignored, struct ip_vs_iphdr *iph) { struct ip_vs_protocol *pp = pd->pp; struct ip_vs_conn *cp = NULL; struct ip_vs_scheduler *sched; struct ip_vs_dest *dest; __be16 _ports[2], *pptr; unsigned int flags; *ignored = 1; /* * IPv6 frags, only the first hit here. */ pptr = frag_safe_skb_hp(skb, iph->len, sizeof(_ports), _ports, iph); if (pptr == NULL) return NULL; /* * FTPDATA needs this check when using local real server. * Never schedule Active FTPDATA connections from real server. * For LVS-NAT they must be already created. For other methods * with persistence the connection is created on SYN+ACK. */ if (pptr[0] == FTPDATA) { IP_VS_DBG_PKT(12, svc->af, pp, skb, 0, "Not scheduling FTPDATA"); return NULL; } /* * Do not schedule replies from local real server. */ if ((!skb->dev || skb->dev->flags & IFF_LOOPBACK) && (cp = pp->conn_in_get(svc->af, skb, iph, 1))) { IP_VS_DBG_PKT(12, svc->af, pp, skb, 0, "Not scheduling reply for existing connection"); __ip_vs_conn_put(cp); return NULL; } /* * Persistent service */ if (svc->flags & IP_VS_SVC_F_PERSISTENT) return ip_vs_sched_persist(svc, skb, pptr[0], pptr[1], ignored, iph); *ignored = 0; /* * Non-persistent service */ if (!svc->fwmark && pptr[1] != svc->port) { if (!svc->port) pr_err("Schedule: port zero only supported " "in persistent services, " "check your ipvs configuration\n"); return NULL; } sched = rcu_dereference(svc->scheduler); if (sched) { /* read svc->sched_data after svc->scheduler */ smp_rmb(); dest = sched->schedule(svc, skb, iph); } else { dest = NULL; } if (dest == NULL) { IP_VS_DBG(1, "Schedule: no dest found.\n"); return NULL; } flags = (svc->flags & IP_VS_SVC_F_ONEPACKET && iph->protocol == IPPROTO_UDP) ? IP_VS_CONN_F_ONE_PACKET : 0; /* * Create a connection entry. */ { struct ip_vs_conn_param p; ip_vs_conn_fill_param(svc->net, svc->af, iph->protocol, &iph->saddr, pptr[0], &iph->daddr, pptr[1], &p); cp = ip_vs_conn_new(&p, dest->af, &dest->addr, dest->port ? dest->port : pptr[1], flags, dest, skb->mark); if (!cp) { *ignored = -1; return NULL; } } IP_VS_DBG_BUF(6, "Schedule fwd:%c c:%s:%u v:%s:%u " "d:%s:%u conn->flags:%X conn->refcnt:%d\n", ip_vs_fwd_tag(cp), IP_VS_DBG_ADDR(cp->af, &cp->caddr), ntohs(cp->cport), IP_VS_DBG_ADDR(cp->af, &cp->vaddr), ntohs(cp->vport), IP_VS_DBG_ADDR(cp->daf, &cp->daddr), ntohs(cp->dport), cp->flags, atomic_read(&cp->refcnt)); ip_vs_conn_stats(cp, svc); return cp; } /* * Pass or drop the packet. * Called by ip_vs_in, when the virtual service is available but * no destination is available for a new connection. */ int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb, struct ip_vs_proto_data *pd, struct ip_vs_iphdr *iph) { __be16 _ports[2], *pptr; #ifdef CONFIG_SYSCTL struct net *net; struct netns_ipvs *ipvs; int unicast; #endif pptr = frag_safe_skb_hp(skb, iph->len, sizeof(_ports), _ports, iph); if (pptr == NULL) { return NF_DROP; } #ifdef CONFIG_SYSCTL net = skb_net(skb); #ifdef CONFIG_IP_VS_IPV6 if (svc->af == AF_INET6) unicast = ipv6_addr_type(&iph->daddr.in6) & IPV6_ADDR_UNICAST; else #endif unicast = (inet_addr_type(net, iph->daddr.ip) == RTN_UNICAST); /* if it is fwmark-based service, the cache_bypass sysctl is up and the destination is a non-local unicast, then create a cache_bypass connection entry */ ipvs = net_ipvs(net); if (ipvs->sysctl_cache_bypass && svc->fwmark && unicast) { int ret; struct ip_vs_conn *cp; unsigned int flags = (svc->flags & IP_VS_SVC_F_ONEPACKET && iph->protocol == IPPROTO_UDP) ? IP_VS_CONN_F_ONE_PACKET : 0; union nf_inet_addr daddr = { .all = { 0, 0, 0, 0 } }; /* create a new connection entry */ IP_VS_DBG(6, "%s(): create a cache_bypass entry\n", __func__); { struct ip_vs_conn_param p; ip_vs_conn_fill_param(svc->net, svc->af, iph->protocol, &iph->saddr, pptr[0], &iph->daddr, pptr[1], &p); cp = ip_vs_conn_new(&p, svc->af, &daddr, 0, IP_VS_CONN_F_BYPASS | flags, NULL, skb->mark); if (!cp) return NF_DROP; } /* statistics */ ip_vs_in_stats(cp, skb); /* set state */ ip_vs_set_state(cp, IP_VS_DIR_INPUT, skb, pd); /* transmit the first SYN packet */ ret = cp->packet_xmit(skb, cp, pd->pp, iph); /* do not touch skb anymore */ atomic_inc(&cp->in_pkts); ip_vs_conn_put(cp); return ret; }