Add the rt linux 4.1.3-rt3 as base
[kvmfornfv.git] / kernel / net / netfilter / ipvs / ip_vs_ftp.c
diff --git a/kernel/net/netfilter/ipvs/ip_vs_ftp.c b/kernel/net/netfilter/ipvs/ip_vs_ftp.c
new file mode 100644 (file)
index 0000000..5d3daae
--- /dev/null
@@ -0,0 +1,503 @@
+/*
+ * ip_vs_ftp.c: IPVS ftp application module
+ *
+ * Authors:    Wensong Zhang <wensong@linuxvirtualserver.org>
+ *
+ * Changes:
+ *
+ *
+ *     This program is free software; you can redistribute it and/or
+ *     modify it under the terms of the GNU General Public License
+ *     as published by the Free Software Foundation; either version
+ *     2 of the License, or (at your option) any later version.
+ *
+ * Most code here is taken from ip_masq_ftp.c in kernel 2.2. The difference
+ * is that ip_vs_ftp module handles the reverse direction to ip_masq_ftp.
+ *
+ *             IP_MASQ_FTP ftp masquerading module
+ *
+ * Version:    @(#)ip_masq_ftp.c 0.04   02/05/96
+ *
+ * Author:     Wouter Gadeyne
+ *
+ */
+
+#define KMSG_COMPONENT "IPVS"
+#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/kernel.h>
+#include <linux/skbuff.h>
+#include <linux/in.h>
+#include <linux/ip.h>
+#include <linux/netfilter.h>
+#include <net/netfilter/nf_conntrack.h>
+#include <net/netfilter/nf_conntrack_expect.h>
+#include <net/netfilter/nf_nat.h>
+#include <net/netfilter/nf_nat_helper.h>
+#include <linux/gfp.h>
+#include <net/protocol.h>
+#include <net/tcp.h>
+#include <asm/unaligned.h>
+
+#include <net/ip_vs.h>
+
+
+#define SERVER_STRING "227 "
+#define CLIENT_STRING "PORT"
+
+
+/*
+ * List of ports (up to IP_VS_APP_MAX_PORTS) to be handled by helper
+ * First port is set to the default port.
+ */
+static unsigned int ports_count = 1;
+static unsigned short ports[IP_VS_APP_MAX_PORTS] = {21, 0};
+module_param_array(ports, ushort, &ports_count, 0444);
+MODULE_PARM_DESC(ports, "Ports to monitor for FTP control commands");
+
+
+/*     Dummy variable */
+static int ip_vs_ftp_pasv;
+
+
+static int
+ip_vs_ftp_init_conn(struct ip_vs_app *app, struct ip_vs_conn *cp)
+{
+       /* We use connection tracking for the command connection */
+       cp->flags |= IP_VS_CONN_F_NFCT;
+       return 0;
+}
+
+
+static int
+ip_vs_ftp_done_conn(struct ip_vs_app *app, struct ip_vs_conn *cp)
+{
+       return 0;
+}
+
+
+/*
+ * Get <addr,port> from the string "xxx.xxx.xxx.xxx,ppp,ppp", started
+ * with the "pattern", ignoring before "skip" and terminated with
+ * the "term" character.
+ * <addr,port> is in network order.
+ */
+static int ip_vs_ftp_get_addrport(char *data, char *data_limit,
+                                 const char *pattern, size_t plen,
+                                 char skip, char term,
+                                 __be32 *addr, __be16 *port,
+                                 char **start, char **end)
+{
+       char *s, c;
+       unsigned char p[6];
+       int i = 0;
+
+       if (data_limit - data < plen) {
+               /* check if there is partial match */
+               if (strncasecmp(data, pattern, data_limit - data) == 0)
+                       return -1;
+               else
+                       return 0;
+       }
+
+       if (strncasecmp(data, pattern, plen) != 0) {
+               return 0;
+       }
+       s = data + plen;
+       if (skip) {
+               int found = 0;
+
+               for (;; s++) {
+                       if (s == data_limit)
+                               return -1;
+                       if (!found) {
+                               if (*s == skip)
+                                       found = 1;
+                       } else if (*s != skip) {
+                               break;
+                       }
+               }
+       }
+
+       for (data = s; ; data++) {
+               if (data == data_limit)
+                       return -1;
+               if (*data == term)
+                       break;
+       }
+       *end = data;
+
+       memset(p, 0, sizeof(p));
+       for (data = s; ; data++) {
+               c = *data;
+               if (c == term)
+                       break;
+               if (c >= '0' && c <= '9') {
+                       p[i] = p[i]*10 + c - '0';
+               } else if (c == ',' && i < 5) {
+                       i++;
+               } else {
+                       /* unexpected character */
+                       return -1;
+               }
+       }
+
+       if (i != 5)
+               return -1;
+
+       *start = s;
+       *addr = get_unaligned((__be32 *) p);
+       *port = get_unaligned((__be16 *) (p + 4));
+       return 1;
+}
+
+/*
+ * Look at outgoing ftp packets to catch the response to a PASV command
+ * from the server (inside-to-outside).
+ * When we see one, we build a connection entry with the client address,
+ * client port 0 (unknown at the moment), the server address and the
+ * server port.  Mark the current connection entry as a control channel
+ * of the new entry. All this work is just to make the data connection
+ * can be scheduled to the right server later.
+ *
+ * The outgoing packet should be something like
+ *   "227 Entering Passive Mode (xxx,xxx,xxx,xxx,ppp,ppp)".
+ * xxx,xxx,xxx,xxx is the server address, ppp,ppp is the server port number.
+ */
+static int ip_vs_ftp_out(struct ip_vs_app *app, struct ip_vs_conn *cp,
+                        struct sk_buff *skb, int *diff)
+{
+       struct iphdr *iph;
+       struct tcphdr *th;
+       char *data, *data_limit;
+       char *start, *end;
+       union nf_inet_addr from;
+       __be16 port;
+       struct ip_vs_conn *n_cp;
+       char buf[24];           /* xxx.xxx.xxx.xxx,ppp,ppp\000 */
+       unsigned int buf_len;
+       int ret = 0;
+       enum ip_conntrack_info ctinfo;
+       struct nf_conn *ct;
+       struct net *net;
+
+       *diff = 0;
+
+#ifdef CONFIG_IP_VS_IPV6
+       /* This application helper doesn't work with IPv6 yet,
+        * so turn this into a no-op for IPv6 packets
+        */
+       if (cp->af == AF_INET6)
+               return 1;
+#endif
+
+       /* Only useful for established sessions */
+       if (cp->state != IP_VS_TCP_S_ESTABLISHED)
+               return 1;
+
+       /* Linear packets are much easier to deal with. */
+       if (!skb_make_writable(skb, skb->len))
+               return 0;
+
+       if (cp->app_data == &ip_vs_ftp_pasv) {
+               iph = ip_hdr(skb);
+               th = (struct tcphdr *)&(((char *)iph)[iph->ihl*4]);
+               data = (char *)th + (th->doff << 2);
+               data_limit = skb_tail_pointer(skb);
+
+               if (ip_vs_ftp_get_addrport(data, data_limit,
+                                          SERVER_STRING,
+                                          sizeof(SERVER_STRING)-1,
+                                          '(', ')',
+                                          &from.ip, &port,
+                                          &start, &end) != 1)
+                       return 1;
+
+               IP_VS_DBG(7, "PASV response (%pI4:%d) -> %pI4:%d detected\n",
+                         &from.ip, ntohs(port), &cp->caddr.ip, 0);
+
+               /*
+                * Now update or create an connection entry for it
+                */
+               {
+                       struct ip_vs_conn_param p;
+                       ip_vs_conn_fill_param(ip_vs_conn_net(cp), AF_INET,
+                                             iph->protocol, &from, port,
+                                             &cp->caddr, 0, &p);
+                       n_cp = ip_vs_conn_out_get(&p);
+               }
+               if (!n_cp) {
+                       struct ip_vs_conn_param p;
+                       ip_vs_conn_fill_param(ip_vs_conn_net(cp),
+                                             AF_INET, IPPROTO_TCP, &cp->caddr,
+                                             0, &cp->vaddr, port, &p);
+                       /* As above, this is ipv4 only */
+                       n_cp = ip_vs_conn_new(&p, AF_INET, &from, port,
+                                             IP_VS_CONN_F_NO_CPORT |
+                                             IP_VS_CONN_F_NFCT,
+                                             cp->dest, skb->mark);
+                       if (!n_cp)
+                               return 0;
+
+                       /* add its controller */
+                       ip_vs_control_add(n_cp, cp);
+               }
+
+               /*
+                * Replace the old passive address with the new one
+                */
+               from.ip = n_cp->vaddr.ip;
+               port = n_cp->vport;
+               snprintf(buf, sizeof(buf), "%u,%u,%u,%u,%u,%u",
+                        ((unsigned char *)&from.ip)[0],
+                        ((unsigned char *)&from.ip)[1],
+                        ((unsigned char *)&from.ip)[2],
+                        ((unsigned char *)&from.ip)[3],
+                        ntohs(port) >> 8,
+                        ntohs(port) & 0xFF);
+
+               buf_len = strlen(buf);
+
+               ct = nf_ct_get(skb, &ctinfo);
+               if (ct && !nf_ct_is_untracked(ct) && nfct_nat(ct)) {
+                       /* If mangling fails this function will return 0
+                        * which will cause the packet to be dropped.
+                        * Mangling can only fail under memory pressure,
+                        * hopefully it will succeed on the retransmitted
+                        * packet.
+                        */
+                       rcu_read_lock();
+                       ret = nf_nat_mangle_tcp_packet(skb, ct, ctinfo,
+                                                      iph->ihl * 4,
+                                                      start-data, end-start,
+                                                      buf, buf_len);
+                       rcu_read_unlock();
+                       if (ret) {
+                               ip_vs_nfct_expect_related(skb, ct, n_cp,
+                                                         IPPROTO_TCP, 0, 0);
+                               if (skb->ip_summed == CHECKSUM_COMPLETE)
+                                       skb->ip_summed = CHECKSUM_UNNECESSARY;
+                               /* csum is updated */
+                               ret = 1;
+                       }
+               }
+
+               /*
+                * Not setting 'diff' is intentional, otherwise the sequence
+                * would be adjusted twice.
+                */
+
+               net = skb_net(skb);
+               cp->app_data = NULL;
+               ip_vs_tcp_conn_listen(net, n_cp);
+               ip_vs_conn_put(n_cp);
+               return ret;
+       }
+       return 1;
+}
+
+
+/*
+ * Look at incoming ftp packets to catch the PASV/PORT command
+ * (outside-to-inside).
+ *
+ * The incoming packet having the PORT command should be something like
+ *      "PORT xxx,xxx,xxx,xxx,ppp,ppp\n".
+ * xxx,xxx,xxx,xxx is the client address, ppp,ppp is the client port number.
+ * In this case, we create a connection entry using the client address and
+ * port, so that the active ftp data connection from the server can reach
+ * the client.
+ */
+static int ip_vs_ftp_in(struct ip_vs_app *app, struct ip_vs_conn *cp,
+                       struct sk_buff *skb, int *diff)
+{
+       struct iphdr *iph;
+       struct tcphdr *th;
+       char *data, *data_start, *data_limit;
+       char *start, *end;
+       union nf_inet_addr to;
+       __be16 port;
+       struct ip_vs_conn *n_cp;
+       struct net *net;
+
+       /* no diff required for incoming packets */
+       *diff = 0;
+
+#ifdef CONFIG_IP_VS_IPV6
+       /* This application helper doesn't work with IPv6 yet,
+        * so turn this into a no-op for IPv6 packets
+        */
+       if (cp->af == AF_INET6)
+               return 1;
+#endif
+
+       /* Only useful for established sessions */
+       if (cp->state != IP_VS_TCP_S_ESTABLISHED)
+               return 1;
+
+       /* Linear packets are much easier to deal with. */
+       if (!skb_make_writable(skb, skb->len))
+               return 0;
+
+       /*
+        * Detecting whether it is passive
+        */
+       iph = ip_hdr(skb);
+       th = (struct tcphdr *)&(((char *)iph)[iph->ihl*4]);
+
+       /* Since there may be OPTIONS in the TCP packet and the HLEN is
+          the length of the header in 32-bit multiples, it is accurate
+          to calculate data address by th+HLEN*4 */
+       data = data_start = (char *)th + (th->doff << 2);
+       data_limit = skb_tail_pointer(skb);
+
+       while (data <= data_limit - 6) {
+               if (strncasecmp(data, "PASV\r\n", 6) == 0) {
+                       /* Passive mode on */
+                       IP_VS_DBG(7, "got PASV at %td of %td\n",
+                                 data - data_start,
+                                 data_limit - data_start);
+                       cp->app_data = &ip_vs_ftp_pasv;
+                       return 1;
+               }
+               data++;
+       }
+
+       /*
+        * To support virtual FTP server, the scenerio is as follows:
+        *       FTP client ----> Load Balancer ----> FTP server
+        * First detect the port number in the application data,
+        * then create a new connection entry for the coming data
+        * connection.
+        */
+       if (ip_vs_ftp_get_addrport(data_start, data_limit,
+                                  CLIENT_STRING, sizeof(CLIENT_STRING)-1,
+                                  ' ', '\r', &to.ip, &port,
+                                  &start, &end) != 1)
+               return 1;
+
+       IP_VS_DBG(7, "PORT %pI4:%d detected\n", &to.ip, ntohs(port));
+
+       /* Passive mode off */
+       cp->app_data = NULL;
+
+       /*
+        * Now update or create a connection entry for it
+        */
+       IP_VS_DBG(7, "protocol %s %pI4:%d %pI4:%d\n",
+                 ip_vs_proto_name(iph->protocol),
+                 &to.ip, ntohs(port), &cp->vaddr.ip, 0);
+
+       {
+               struct ip_vs_conn_param p;
+               ip_vs_conn_fill_param(ip_vs_conn_net(cp), AF_INET,
+                                     iph->protocol, &to, port, &cp->vaddr,
+                                     htons(ntohs(cp->vport)-1), &p);
+               n_cp = ip_vs_conn_in_get(&p);
+               if (!n_cp) {
+                       /* This is ipv4 only */
+                       n_cp = ip_vs_conn_new(&p, AF_INET, &cp->daddr,
+                                             htons(ntohs(cp->dport)-1),
+                                             IP_VS_CONN_F_NFCT, cp->dest,
+                                             skb->mark);
+                       if (!n_cp)
+                               return 0;
+
+                       /* add its controller */
+                       ip_vs_control_add(n_cp, cp);
+               }
+       }
+
+       /*
+        *      Move tunnel to listen state
+        */
+       net = skb_net(skb);
+       ip_vs_tcp_conn_listen(net, n_cp);
+       ip_vs_conn_put(n_cp);
+
+       return 1;
+}
+
+
+static struct ip_vs_app ip_vs_ftp = {
+       .name =         "ftp",
+       .type =         IP_VS_APP_TYPE_FTP,
+       .protocol =     IPPROTO_TCP,
+       .module =       THIS_MODULE,
+       .incs_list =    LIST_HEAD_INIT(ip_vs_ftp.incs_list),
+       .init_conn =    ip_vs_ftp_init_conn,
+       .done_conn =    ip_vs_ftp_done_conn,
+       .bind_conn =    NULL,
+       .unbind_conn =  NULL,
+       .pkt_out =      ip_vs_ftp_out,
+       .pkt_in =       ip_vs_ftp_in,
+};
+
+/*
+ *     per netns ip_vs_ftp initialization
+ */
+static int __net_init __ip_vs_ftp_init(struct net *net)
+{
+       int i, ret;
+       struct ip_vs_app *app;
+       struct netns_ipvs *ipvs = net_ipvs(net);
+
+       if (!ipvs)
+               return -ENOENT;
+
+       app = register_ip_vs_app(net, &ip_vs_ftp);
+       if (IS_ERR(app))
+               return PTR_ERR(app);
+
+       for (i = 0; i < ports_count; i++) {
+               if (!ports[i])
+                       continue;
+               ret = register_ip_vs_app_inc(net, app, app->protocol, ports[i]);
+               if (ret)
+                       goto err_unreg;
+               pr_info("%s: loaded support on port[%d] = %d\n",
+                       app->name, i, ports[i]);
+       }
+       return 0;
+
+err_unreg:
+       unregister_ip_vs_app(net, &ip_vs_ftp);
+       return ret;
+}
+/*
+ *     netns exit
+ */
+static void __ip_vs_ftp_exit(struct net *net)
+{
+       unregister_ip_vs_app(net, &ip_vs_ftp);
+}
+
+static struct pernet_operations ip_vs_ftp_ops = {
+       .init = __ip_vs_ftp_init,
+       .exit = __ip_vs_ftp_exit,
+};
+
+static int __init ip_vs_ftp_init(void)
+{
+       int rv;
+
+       rv = register_pernet_subsys(&ip_vs_ftp_ops);
+       /* rcu_barrier() is called by netns on error */
+       return rv;
+}
+
+/*
+ *     ip_vs_ftp finish.
+ */
+static void __exit ip_vs_ftp_exit(void)
+{
+       unregister_pernet_subsys(&ip_vs_ftp_ops);
+       /* rcu_barrier() is called by netns */
+}
+
+
+module_init(ip_vs_ftp_init);
+module_exit(ip_vs_ftp_exit);
+MODULE_LICENSE("GPL");