vport-internal_dev.c revision f6eec614d2252a99b861e288b6301599d2d58da4 
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#include <linux/hardirq.h>
20#include <linux/if_vlan.h>
21#include <linux/kernel.h>
22#include <linux/netdevice.h>
23#include <linux/etherdevice.h>
24#include <linux/ethtool.h>
25#include <linux/skbuff.h>
26
27#include <net/dst.h>
28#include <net/xfrm.h>
29#include <net/rtnetlink.h>
30
31#include "datapath.h"
32#include "vport-internal_dev.h"
33#include "vport-netdev.h"
34
35struct internal_dev {
36	struct vport *vport;
37};
38
39static struct internal_dev *internal_dev_priv(struct net_device *netdev)
40{
41	return netdev_priv(netdev);
42}
43
44/* This function is only called by the kernel network layer.*/
45static struct rtnl_link_stats64 *internal_dev_get_stats(struct net_device *netdev,
46							struct rtnl_link_stats64 *stats)
47{
48	struct vport *vport = ovs_internal_dev_get_vport(netdev);
49	struct ovs_vport_stats vport_stats;
50
51	ovs_vport_get_stats(vport, &vport_stats);
52
53	/* The tx and rx stats need to be swapped because the
54	 * switch and host OS have opposite perspectives. */
55	stats->rx_packets	= vport_stats.tx_packets;
56	stats->tx_packets	= vport_stats.rx_packets;
57	stats->rx_bytes		= vport_stats.tx_bytes;
58	stats->tx_bytes		= vport_stats.rx_bytes;
59	stats->rx_errors	= vport_stats.tx_errors;
60	stats->tx_errors	= vport_stats.rx_errors;
61	stats->rx_dropped	= vport_stats.tx_dropped;
62	stats->tx_dropped	= vport_stats.rx_dropped;
63
64	return stats;
65}
66
67/* Called with rcu_read_lock_bh. */
68static int internal_dev_xmit(struct sk_buff *skb, struct net_device *netdev)
69{
70	rcu_read_lock();
71	ovs_vport_receive(internal_dev_priv(netdev)->vport, skb, NULL);
72	rcu_read_unlock();
73	return 0;
74}
75
76static int internal_dev_open(struct net_device *netdev)
77{
78	netif_start_queue(netdev);
79	return 0;
80}
81
82static int internal_dev_stop(struct net_device *netdev)
83{
84	netif_stop_queue(netdev);
85	return 0;
86}
87
88static void internal_dev_getinfo(struct net_device *netdev,
89				 struct ethtool_drvinfo *info)
90{
91	strlcpy(info->driver, "openvswitch", sizeof(info->driver));
92}
93
94static const struct ethtool_ops internal_dev_ethtool_ops = {
95	.get_drvinfo	= internal_dev_getinfo,
96	.get_link	= ethtool_op_get_link,
97};
98
99static int internal_dev_change_mtu(struct net_device *netdev, int new_mtu)
100{
101	if (new_mtu < 68)
102		return -EINVAL;
103
104	netdev->mtu = new_mtu;
105	return 0;
106}
107
108static void internal_dev_destructor(struct net_device *dev)
109{
110	struct vport *vport = ovs_internal_dev_get_vport(dev);
111
112	ovs_vport_free(vport);
113	free_netdev(dev);
114}
115
116static const struct net_device_ops internal_dev_netdev_ops = {
117	.ndo_open = internal_dev_open,
118	.ndo_stop = internal_dev_stop,
119	.ndo_start_xmit = internal_dev_xmit,
120	.ndo_set_mac_address = eth_mac_addr,
121	.ndo_change_mtu = internal_dev_change_mtu,
122	.ndo_get_stats64 = internal_dev_get_stats,
123};
124
125static struct rtnl_link_ops internal_dev_link_ops __read_mostly = {
126	.kind = "openvswitch",
127};
128
129static void do_setup(struct net_device *netdev)
130{
131	ether_setup(netdev);
132
133	netdev->netdev_ops = &internal_dev_netdev_ops;
134
135	netdev->priv_flags &= ~IFF_TX_SKB_SHARING;
136	netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
137	netdev->destructor = internal_dev_destructor;
138	netdev->ethtool_ops = &internal_dev_ethtool_ops;
139	netdev->rtnl_link_ops = &internal_dev_link_ops;
140	netdev->tx_queue_len = 0;
141
142	netdev->features = NETIF_F_LLTX | NETIF_F_SG | NETIF_F_FRAGLIST |
143			   NETIF_F_HIGHDMA | NETIF_F_HW_CSUM |
144			   NETIF_F_GSO_SOFTWARE | NETIF_F_GSO_ENCAP_ALL;
145
146	netdev->vlan_features = netdev->features;
147	netdev->hw_enc_features = netdev->features;
148	netdev->features |= NETIF_F_HW_VLAN_CTAG_TX;
149	netdev->hw_features = netdev->features & ~NETIF_F_LLTX;
150
151	eth_hw_addr_random(netdev);
152}
153
154static struct vport *internal_dev_create(const struct vport_parms *parms)
155{
156	struct vport *vport;
157	struct netdev_vport *netdev_vport;
158	struct internal_dev *internal_dev;
159	int err;
160
161	vport = ovs_vport_alloc(sizeof(struct netdev_vport),
162				&ovs_internal_vport_ops, parms);
163	if (IS_ERR(vport)) {
164		err = PTR_ERR(vport);
165		goto error;
166	}
167
168	netdev_vport = netdev_vport_priv(vport);
169
170	netdev_vport->dev = alloc_netdev(sizeof(struct internal_dev),
171					 parms->name, NET_NAME_UNKNOWN,
172					 do_setup);
173	if (!netdev_vport->dev) {
174		err = -ENOMEM;
175		goto error_free_vport;
176	}
177
178	dev_net_set(netdev_vport->dev, ovs_dp_get_net(vport->dp));
179	internal_dev = internal_dev_priv(netdev_vport->dev);
180	internal_dev->vport = vport;
181
182	/* Restrict bridge port to current netns. */
183	if (vport->port_no == OVSP_LOCAL)
184		netdev_vport->dev->features |= NETIF_F_NETNS_LOCAL;
185
186	rtnl_lock();
187	err = register_netdevice(netdev_vport->dev);
188	if (err)
189		goto error_free_netdev;
190
191	dev_set_promiscuity(netdev_vport->dev, 1);
192	rtnl_unlock();
193	netif_start_queue(netdev_vport->dev);
194
195	return vport;
196
197error_free_netdev:
198	rtnl_unlock();
199	free_netdev(netdev_vport->dev);
200error_free_vport:
201	ovs_vport_free(vport);
202error:
203	return ERR_PTR(err);
204}
205
206static void internal_dev_destroy(struct vport *vport)
207{
208	struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
209
210	netif_stop_queue(netdev_vport->dev);
211	rtnl_lock();
212	dev_set_promiscuity(netdev_vport->dev, -1);
213
214	/* unregister_netdevice() waits for an RCU grace period. */
215	unregister_netdevice(netdev_vport->dev);
216
217	rtnl_unlock();
218}
219
220static int internal_dev_recv(struct vport *vport, struct sk_buff *skb)
221{
222	struct net_device *netdev = netdev_vport_priv(vport)->dev;
223	int len;
224
225	len = skb->len;
226
227	skb_dst_drop(skb);
228	nf_reset(skb);
229	secpath_reset(skb);
230
231	skb->dev = netdev;
232	skb->pkt_type = PACKET_HOST;
233	skb->protocol = eth_type_trans(skb, netdev);
234	skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
235
236	netif_rx(skb);
237
238	return len;
239}
240
241const struct vport_ops ovs_internal_vport_ops = {
242	.type		= OVS_VPORT_TYPE_INTERNAL,
243	.create		= internal_dev_create,
244	.destroy	= internal_dev_destroy,
245	.get_name	= ovs_netdev_get_name,
246	.send		= internal_dev_recv,
247};
248
249int ovs_is_internal_dev(const struct net_device *netdev)
250{
251	return netdev->netdev_ops == &internal_dev_netdev_ops;
252}
253
254struct vport *ovs_internal_dev_get_vport(struct net_device *netdev)
255{
256	if (!ovs_is_internal_dev(netdev))
257		return NULL;
258
259	return internal_dev_priv(netdev)->vport;
260}
261
262int ovs_internal_dev_rtnl_link_register(void)
263{
264	return rtnl_link_register(&internal_dev_link_ops);
265}
266
267void ovs_internal_dev_rtnl_link_unregister(void)
268{
269	rtnl_link_unregister(&internal_dev_link_ops);
270}
271