net/openvswitch/vport-netdev.c

   1 /*
   2  * Copyright (c) 2007-2012 Nicira, Inc.
   3  *
   4  * This program is free software; you can redistribute it and/or
   5  * modify it under the terms of version 2 of the GNU General Public
   6  * License as published by the Free Software Foundation.
   7  *
   8  * This program is distributed in the hope that it will be useful, but
   9  * WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  11  * General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License
  14  * along with this program; if not, write to the Free Software
  15  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  16  * 02110-1301, USA
  17  */
  18
  19 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
  20
  21 #include <linux/if_arp.h>
  22 #include <linux/if_bridge.h>
  23 #include <linux/if_vlan.h>
  24 #include <linux/kernel.h>
  25 #include <linux/llc.h>
  26 #include <linux/rtnetlink.h>
  27 #include <linux/skbuff.h>
  28 #include <linux/openvswitch.h>
  29
  30 #include <net/llc.h>
  31
  32 #include "datapath.h"
  33 #include "vport-internal_dev.h"
  34 #include "vport-netdev.h"
  35
  36 static struct vport_ops ovs_netdev_vport_ops;
  37
  38 /* Must be called with rcu_read_lock. */
  39 static void netdev_port_receive(struct vport *vport, struct sk_buff *skb)
  40 {
  41         if (unlikely(!vport))
  42                 goto error;
  43
  44         if (unlikely(skb_warn_if_lro(skb)))
  45                 goto error;
  46
  47         /* Make our own copy of the packet.  Otherwise we will mangle the
  48          * packet for anyone who came before us (e.g. tcpdump via AF_PACKET).
  49          */
  50         skb = skb_share_check(skb, GFP_ATOMIC);
  51         if (unlikely(!skb))
  52                 return;
  53
  54         skb_push(skb, ETH_HLEN);
  55         ovs_skb_postpush_rcsum(skb, skb->data, ETH_HLEN);
  56
  57         ovs_vport_receive(vport, skb, NULL);
  58         return;
  59
  60 error:
  61         kfree_skb(skb);
  62 }
  63
  64 /* Called with rcu_read_lock and bottom-halves disabled. */
  65 static rx_handler_result_t netdev_frame_hook(struct sk_buff **pskb)
  66 {
  67         struct sk_buff *skb = *pskb;
  68         struct vport *vport;
  69
  70         if (unlikely(skb->pkt_type == PACKET_LOOPBACK))
  71                 return RX_HANDLER_PASS;
  72
  73         vport = ovs_netdev_get_vport(skb->dev);
  74
  75         netdev_port_receive(vport, skb);
  76
  77         return RX_HANDLER_CONSUMED;
  78 }
  79
  80 static struct net_device *get_dpdev(const struct datapath *dp)
  81 {
  82         struct vport *local;
  83
  84         local = ovs_vport_ovsl(dp, OVSP_LOCAL);
  85         BUG_ON(!local);
  86         return netdev_vport_priv(local)->dev;
  87 }
  88
  89 static struct vport *netdev_create(const struct vport_parms *parms)
  90 {
  91         struct vport *vport;
  92         struct netdev_vport *netdev_vport;
  93         int err;
  94
  95         vport = ovs_vport_alloc(sizeof(struct netdev_vport),
  96                                 &ovs_netdev_vport_ops, parms);
  97         if (IS_ERR(vport)) {
  98                 err = PTR_ERR(vport);
  99                 goto error;
 100         }
 101
 102         netdev_vport = netdev_vport_priv(vport);
 103
 104         netdev_vport->dev = dev_get_by_name(ovs_dp_get_net(vport->dp), parms->name);
 105         if (!netdev_vport->dev) {
 106                 err = -ENODEV;
 107                 goto error_free_vport;
 108         }
 109
 110         if (netdev_vport->dev->flags & IFF_LOOPBACK ||
 111             netdev_vport->dev->type != ARPHRD_ETHER ||
 112             ovs_is_internal_dev(netdev_vport->dev)) {
 113                 err = -EINVAL;
 114                 goto error_put;
 115         }
 116
 117         rtnl_lock();
 118         err = netdev_master_upper_dev_link(netdev_vport->dev,
 119                                            get_dpdev(vport->dp));
 120         if (err)
 121                 goto error_unlock;
 122
 123         err = netdev_rx_handler_register(netdev_vport->dev, netdev_frame_hook,
 124                                          vport);
 125         if (err)
 126                 goto error_master_upper_dev_unlink;
 127
 128         dev_disable_lro(netdev_vport->dev);
 129         dev_set_promiscuity(netdev_vport->dev, 1);
 130         netdev_vport->dev->priv_flags |= IFF_OVS_DATAPATH;
 131         rtnl_unlock();
 132
 133         return vport;
 134
 135 error_master_upper_dev_unlink:
 136         netdev_upper_dev_unlink(netdev_vport->dev, get_dpdev(vport->dp));
 137 error_unlock:
 138         rtnl_unlock();
 139 error_put:
 140         dev_put(netdev_vport->dev);
 141 error_free_vport:
 142         ovs_vport_free(vport);
 143 error:
 144         return ERR_PTR(err);
 145 }
 146
 147 static void free_port_rcu(struct rcu_head *rcu)
 148 {
 149         struct netdev_vport *netdev_vport = container_of(rcu,
 150                                         struct netdev_vport, rcu);
 151
 152         dev_put(netdev_vport->dev);
 153         ovs_vport_free(vport_from_priv(netdev_vport));
 154 }
 155
 156 void ovs_netdev_detach_dev(struct vport *vport)
 157 {
 158         struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
 159
 160         ASSERT_RTNL();
 161         netdev_vport->dev->priv_flags &= ~IFF_OVS_DATAPATH;
 162         netdev_rx_handler_unregister(netdev_vport->dev);
 163         netdev_upper_dev_unlink(netdev_vport->dev,
 164                                 netdev_master_upper_dev_get(netdev_vport->dev));
 165         dev_set_promiscuity(netdev_vport->dev, -1);
 166 }
 167
 168 static void netdev_destroy(struct vport *vport)
 169 {
 170         struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
 171
 172         rtnl_lock();
 173         if (netdev_vport->dev->priv_flags & IFF_OVS_DATAPATH)
 174                 ovs_netdev_detach_dev(vport);
 175         rtnl_unlock();
 176
 177         call_rcu(&netdev_vport->rcu, free_port_rcu);
 178 }
 179
 180 const char *ovs_netdev_get_name(const struct vport *vport)
 181 {
 182         const struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
 183         return netdev_vport->dev->name;
 184 }
 185
 186 static unsigned int packet_length(const struct sk_buff *skb)
 187 {
 188         unsigned int length = skb->len - ETH_HLEN;
 189
 190         if (skb->protocol == htons(ETH_P_8021Q))
 191                 length -= VLAN_HLEN;
 192
 193         return length;
 194 }
 195
 196 static int netdev_send(struct vport *vport, struct sk_buff *skb)
 197 {
 198         struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
 199         int mtu = netdev_vport->dev->mtu;
 200         int len;
 201
 202         if (unlikely(packet_length(skb) > mtu && !skb_is_gso(skb))) {
 203                 net_warn_ratelimited("%s: dropped over-mtu packet: %d > %d\n",
 204                                      netdev_vport->dev->name,
 205                                      packet_length(skb), mtu);
 206                 goto drop;
 207         }
 208
 209         skb->dev = netdev_vport->dev;
 210         len = skb->len;
 211         dev_queue_xmit(skb);
 212
 213         return len;
 214
 215 drop:
 216         kfree_skb(skb);
 217         return 0;
 218 }
 219
 220 /* Returns null if this device is not attached to a datapath. */
 221 struct vport *ovs_netdev_get_vport(struct net_device *dev)
 222 {
 223         if (likely(dev->priv_flags & IFF_OVS_DATAPATH))
 224                 return (struct vport *)
 225                         rcu_dereference_rtnl(dev->rx_handler_data);
 226         else
 227                 return NULL;
 228 }
 229
 230 static struct vport_ops ovs_netdev_vport_ops = {
 231         .type           = OVS_VPORT_TYPE_NETDEV,
 232         .create         = netdev_create,
 233         .destroy        = netdev_destroy,
 234         .get_name       = ovs_netdev_get_name,
 235         .send           = netdev_send,
 236 };
 237
 238 int __init ovs_netdev_init(void)
 239 {
 240         return ovs_vport_ops_register(&ovs_netdev_vport_ops);
 241 }
 242
 243 void ovs_netdev_exit(void)
 244 {
 245         ovs_vport_ops_unregister(&ovs_netdev_vport_ops);
 246 }