/*
- * Copyright (c) 2009, 2010 Nicira Networks.
- * Distributed under the terms of the GNU GPL version 2.
+ * Copyright (c) 2007-2015 Nicira, Inc.
*
- * Significant portions of this file may be copied from parts of the Linux
- * kernel, by Linus Torvalds and others.
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of version 2 of the GNU General Public
+ * License as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
+ * 02110-1301, USA
*/
+#include <linux/hardirq.h>
+#include <linux/if_vlan.h>
#include <linux/kernel.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
-#include <linux/percpu.h>
-#include <linux/preempt.h>
-#include <linux/rcupdate.h>
#include <linux/skbuff.h>
-#include <linux/workqueue.h>
+#include <linux/percpu.h>
+#include <linux/u64_stats_sync.h>
+#include <linux/netdev_features.h>
+
+#include <net/dst.h>
+#include <net/xfrm.h>
+#include <net/rtnetlink.h>
#include "datapath.h"
-#include "openvswitch/internal_dev.h"
#include "vport-internal_dev.h"
#include "vport-netdev.h"
-struct pcpu_lstats {
- unsigned long rx_packets;
- unsigned long rx_bytes;
- unsigned long tx_packets;
- unsigned long tx_bytes;
-};
-
struct internal_dev {
struct vport *vport;
-
- struct net_device_stats stats;
- struct pcpu_lstats *lstats;
-
- /* This is warty support for XAPI, which does not support summing bond
- * device statistics itself. 'extra_stats' can be set by userspace via
- * the DP_DEV_SET_STATS ioctl and, if they are, then they are added to
- * the real device stats. */
- struct pcpu_lstats extra_stats;
};
-struct vport_ops internal_vport_ops;
+static struct vport_ops ovs_internal_vport_ops;
-static inline struct internal_dev *internal_dev_priv(struct net_device *netdev)
+static struct internal_dev *internal_dev_priv(struct net_device *netdev)
{
return netdev_priv(netdev);
}
-static struct net_device_stats *internal_dev_get_stats(struct net_device *netdev)
-{
- struct internal_dev *internal_dev = internal_dev_priv(netdev);
- struct net_device_stats *stats;
- int i;
-
- stats = &internal_dev->stats;
- stats->rx_bytes = internal_dev->extra_stats.rx_bytes;
- stats->rx_packets = internal_dev->extra_stats.rx_packets;
- stats->tx_bytes = internal_dev->extra_stats.tx_bytes;
- stats->tx_packets = internal_dev->extra_stats.tx_packets;
- for_each_possible_cpu(i) {
- const struct pcpu_lstats *lb_stats;
-
- lb_stats = per_cpu_ptr(internal_dev->lstats, i);
- stats->rx_bytes += lb_stats->rx_bytes;
- stats->rx_packets += lb_stats->rx_packets;
- stats->tx_bytes += lb_stats->tx_bytes;
- stats->tx_packets += lb_stats->tx_packets;
- }
- return stats;
-}
-
-static int internal_dev_mac_addr(struct net_device *dev, void *p)
-{
- struct sockaddr *addr = p;
-
- if (!is_valid_ether_addr(addr->sa_data))
- return -EADDRNOTAVAIL;
- memcpy(dev->dev_addr, addr->sa_data, dev->addr_len);
- return 0;
-}
-
-/* Not reentrant (because it is called with BHs disabled), but may be called
- * simultaneously on different CPUs. */
+/* Called with rcu_read_lock_bh. */
static int internal_dev_xmit(struct sk_buff *skb, struct net_device *netdev)
{
- struct internal_dev *internal_dev = internal_dev_priv(netdev);
- struct vport *vport = internal_dev_get_vport(netdev);
- struct pcpu_lstats *lb_stats;
+ int len, err;
- /* We need our own clone. */
- skb = skb_share_check(skb, GFP_ATOMIC);
- if (!skb)
- return 0;
+ len = skb->len;
+ rcu_read_lock();
+ err = ovs_vport_receive(internal_dev_priv(netdev)->vport, skb, NULL);
+ rcu_read_unlock();
- lb_stats = per_cpu_ptr(internal_dev->lstats, smp_processor_id());
- lb_stats->tx_packets++;
- lb_stats->tx_bytes += skb->len;
+ if (likely(!err)) {
+#ifdef HAVE_DEV_TSTATS
+ struct pcpu_sw_netstats *tstats;
- skb_reset_mac_header(skb);
- rcu_read_lock_bh();
- vport_receive(vport, skb);
- rcu_read_unlock_bh();
+ tstats = this_cpu_ptr((struct pcpu_sw_netstats __percpu *)netdev->tstats);
+ u64_stats_update_begin(&tstats->syncp);
+ tstats->tx_bytes += len;
+ tstats->tx_packets++;
+ u64_stats_update_end(&tstats->syncp);
+#endif
+ } else {
+ netdev->stats.tx_errors++;
+ }
return 0;
}
static void internal_dev_getinfo(struct net_device *netdev,
struct ethtool_drvinfo *info)
{
- struct dp_port *dp_port = vport_get_dp_port(internal_dev_get_vport(netdev));
-
- strcpy(info->driver, "openvswitch");
- if (dp_port)
- sprintf(info->bus_info, "%d.%d", dp_port->dp->dp_idx, dp_port->port_no);
+ strlcpy(info->driver, "openvswitch", sizeof(info->driver));
}
-static struct ethtool_ops internal_dev_ethtool_ops = {
- .get_drvinfo = internal_dev_getinfo,
- .get_link = ethtool_op_get_link,
- .get_sg = ethtool_op_get_sg,
- .get_tx_csum = ethtool_op_get_tx_csum,
- .get_tso = ethtool_op_get_tso,
+static const struct ethtool_ops internal_dev_ethtool_ops = {
+ .get_drvinfo = internal_dev_getinfo,
+ .get_link = ethtool_op_get_link,
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,39)
+ .get_sg = ethtool_op_get_sg,
+ .set_sg = ethtool_op_set_sg,
+ .get_tx_csum = ethtool_op_get_tx_csum,
+ .set_tx_csum = ethtool_op_set_tx_hw_csum,
+ .get_tso = ethtool_op_get_tso,
+ .set_tso = ethtool_op_set_tso,
+#endif
};
static int internal_dev_change_mtu(struct net_device *netdev, int new_mtu)
{
- struct dp_port *dp_port = vport_get_dp_port(internal_dev_get_vport(netdev));
-
if (new_mtu < 68)
return -EINVAL;
- if (dp_port) {
- int min_mtu;
-
- mutex_lock(&dp_port->dp->mutex);
- min_mtu = dp_min_mtu(dp_port->dp);
- mutex_unlock(&dp_port->dp->mutex);
-
- if (new_mtu > min_mtu)
- return -EINVAL;
- }
-
netdev->mtu = new_mtu;
return 0;
}
-static int internal_dev_init(struct net_device *netdev)
+static void internal_dev_destructor(struct net_device *dev)
{
- struct internal_dev *internal_dev = internal_dev_priv(netdev);
-
- internal_dev->lstats = alloc_percpu(struct pcpu_lstats);
- if (!internal_dev->lstats)
- return -ENOMEM;
+ struct vport *vport = ovs_internal_dev_get_vport(dev);
- return 0;
+ ovs_vport_free(vport);
+ free_netdev(dev);
}
-static void internal_dev_free(struct net_device *netdev)
+#ifdef HAVE_DEV_TSTATS
+static int internal_dev_init(struct net_device *dev)
{
- struct internal_dev *internal_dev = internal_dev_priv(netdev);
-
- free_percpu(internal_dev->lstats);
- free_netdev(netdev);
+ dev->tstats = (typeof(dev->tstats)) netdev_alloc_pcpu_stats(struct pcpu_sw_netstats);
+ if (!dev->tstats)
+ return -ENOMEM;
+ return 0;
}
-static int internal_dev_do_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
+static void internal_dev_uninit(struct net_device *dev)
{
- struct internal_dev *internal_dev = internal_dev_priv(dev);
-
- if (cmd == INTERNAL_DEV_SET_STATS) {
- struct internal_dev_stats stats;
-
- if (copy_from_user(&stats, ifr->ifr_data, sizeof(stats)))
- return -EFAULT;
-
- internal_dev->extra_stats.rx_bytes = stats.rx_bytes;
- internal_dev->extra_stats.rx_packets = stats.rx_packets;
- internal_dev->extra_stats.tx_bytes = stats.tx_bytes;
- internal_dev->extra_stats.tx_packets = stats.tx_packets;
-
- return 0;
- }
-
- if (dp_ioctl_hook)
- return dp_ioctl_hook(dev, ifr, cmd);
- return -EOPNOTSUPP;
+ free_percpu(dev->tstats);
}
+#endif
-#ifdef HAVE_NET_DEVICE_OPS
static const struct net_device_ops internal_dev_netdev_ops = {
+#ifdef HAVE_DEV_TSTATS
.ndo_init = internal_dev_init,
+ .ndo_uninit = internal_dev_uninit,
+ .ndo_get_stats64 = ip_tunnel_get_stats64,
+#endif
.ndo_open = internal_dev_open,
.ndo_stop = internal_dev_stop,
.ndo_start_xmit = internal_dev_xmit,
- .ndo_set_mac_address = internal_dev_mac_addr,
- .ndo_do_ioctl = internal_dev_do_ioctl,
+ .ndo_set_mac_address = eth_mac_addr,
.ndo_change_mtu = internal_dev_change_mtu,
- .ndo_get_stats = internal_dev_get_stats,
};
-#endif
-static void
-do_setup(struct net_device *netdev)
+static struct rtnl_link_ops internal_dev_link_ops __read_mostly = {
+ .kind = "openvswitch",
+};
+
+static void do_setup(struct net_device *netdev)
{
ether_setup(netdev);
-#ifdef HAVE_NET_DEVICE_OPS
netdev->netdev_ops = &internal_dev_netdev_ops;
-#else
- netdev->do_ioctl = internal_dev_do_ioctl;
- netdev->get_stats = internal_dev_get_stats;
- netdev->hard_start_xmit = internal_dev_xmit;
- netdev->open = internal_dev_open;
- netdev->stop = internal_dev_stop;
- netdev->set_mac_address = internal_dev_mac_addr;
- netdev->change_mtu = internal_dev_change_mtu;
- netdev->init = internal_dev_init;
-#endif
- netdev->destructor = internal_dev_free;
- SET_ETHTOOL_OPS(netdev, &internal_dev_ethtool_ops);
+ netdev->priv_flags &= ~IFF_TX_SKB_SHARING;
+ netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE | IFF_OPENVSWITCH;
+ netdev->destructor = internal_dev_destructor;
+ netdev->ethtool_ops = &internal_dev_ethtool_ops;
+ netdev->rtnl_link_ops = &internal_dev_link_ops;
netdev->tx_queue_len = 0;
- netdev->flags = IFF_BROADCAST | IFF_MULTICAST;
- netdev->features = NETIF_F_LLTX; /* XXX other features? */
+ netdev->features = NETIF_F_LLTX | NETIF_F_SG | NETIF_F_FRAGLIST |
+ NETIF_F_HIGHDMA | NETIF_F_HW_CSUM |
+ NETIF_F_GSO_SOFTWARE | NETIF_F_GSO_ENCAP_ALL;
+
+ netdev->vlan_features = netdev->features;
+ netdev->features |= NETIF_F_HW_VLAN_CTAG_TX;
- vport_gen_ether_addr(netdev->dev_addr);
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(3,8,0)
+ netdev->hw_enc_features = netdev->features;
+#endif
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,39)
+ netdev->hw_features = netdev->features & ~NETIF_F_LLTX;
+#endif
+ eth_hw_addr_random(netdev);
}
-static struct vport *
-internal_dev_create(const char *name, const void __user *config)
+static struct vport *internal_dev_create(const struct vport_parms *parms)
{
struct vport *vport;
- struct netdev_vport *netdev_vport;
struct internal_dev *internal_dev;
int err;
- vport = vport_alloc(sizeof(struct netdev_vport), &internal_vport_ops);
+ vport = ovs_vport_alloc(0, &ovs_internal_vport_ops, parms);
if (IS_ERR(vport)) {
err = PTR_ERR(vport);
goto error;
}
- netdev_vport = netdev_vport_priv(vport);
-
- netdev_vport->dev = alloc_netdev(sizeof(struct internal_dev), name, do_setup);
- if (!netdev_vport->dev) {
+ vport->dev = alloc_netdev(sizeof(struct internal_dev),
+ parms->name, NET_NAME_UNKNOWN, do_setup);
+ if (!vport->dev) {
err = -ENOMEM;
goto error_free_vport;
}
- internal_dev = internal_dev_priv(netdev_vport->dev);
+ dev_net_set(vport->dev, ovs_dp_get_net(vport->dp));
+ internal_dev = internal_dev_priv(vport->dev);
internal_dev->vport = vport;
- err = register_netdevice(netdev_vport->dev);
+ /* Restrict bridge port to current netns. */
+ if (vport->port_no == OVSP_LOCAL)
+ vport->dev->features |= NETIF_F_NETNS_LOCAL;
+
+ rtnl_lock();
+ err = register_netdevice(vport->dev);
if (err)
goto error_free_netdev;
+ dev_set_promiscuity(vport->dev, 1);
+ rtnl_unlock();
+ netif_start_queue(vport->dev);
+
return vport;
error_free_netdev:
- free_netdev(netdev_vport->dev);
+ rtnl_unlock();
+ free_netdev(vport->dev);
error_free_vport:
- vport_free(vport);
+ ovs_vport_free(vport);
error:
return ERR_PTR(err);
}
-static int
-internal_dev_destroy(struct vport *vport)
+static void internal_dev_destroy(struct vport *vport)
{
- struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
+ netif_stop_queue(vport->dev);
+ rtnl_lock();
+ dev_set_promiscuity(vport->dev, -1);
- unregister_netdevice(netdev_vport->dev);
- vport_free(vport);
+ /* unregister_netdevice() waits for an RCU grace period. */
+ unregister_netdevice(vport->dev);
- return 0;
+ rtnl_unlock();
}
-static int
-internal_dev_attach(struct vport *vport)
+static netdev_tx_t internal_dev_recv(struct sk_buff *skb)
{
- struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
-
- dev_set_promiscuity(netdev_vport->dev, 1);
-
- /* It would make sense to assign dev->br_port here too, but
- * that causes packets received on internal ports to get caught
- * in netdev_frame_hook(). In turn netdev_frame_hook() can reject them
- * back to the network stack, but that's a waste of time. */
-
- return 0;
-}
+ struct net_device *netdev = skb->dev;
+#ifdef HAVE_DEV_TSTATS
+ struct pcpu_sw_netstats *stats;
+#endif
-static int
-internal_dev_detach(struct vport *vport)
-{
- struct netdev_vport *netdev_vport = netdev_vport_priv(vport);
+ if (unlikely(!(netdev->flags & IFF_UP))) {
+ kfree_skb(skb);
+ netdev->stats.rx_dropped++;
+ return NETDEV_TX_OK;
+ }
- dev_set_promiscuity(netdev_vport->dev, -1);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,37)
+ if (skb_vlan_tag_present(skb)) {
+ if (unlikely(!vlan_insert_tag_set_proto(skb,
+ skb->vlan_proto,
+ skb_vlan_tag_get(skb))))
+ return NETDEV_TX_OK;
- /* Make sure that no packets arrive from now on, since
- * internal_dev_xmit() will try to find itself through
- * p->dp->ports[], and we're about to set that to null. */
- netif_tx_disable(netdev_vport->dev);
+ if (skb->ip_summed == CHECKSUM_COMPLETE)
+ skb->csum = csum_add(skb->csum,
+ csum_partial(skb->data + (2 * ETH_ALEN),
+ VLAN_HLEN, 0));
- return 0;
-}
+ vlan_set_tci(skb, 0);
+ }
+#endif
-static int
-internal_dev_recv(struct vport *vport, struct sk_buff *skb)
-{
- struct net_device *netdev = netdev_vport_priv(vport)->dev;
- struct internal_dev *internal_dev = internal_dev_priv(netdev);
- struct pcpu_lstats *lb_stats;
- int len;
+ skb_dst_drop(skb);
+ nf_reset(skb);
+ secpath_reset(skb);
- skb->dev = netdev;
- len = skb->len;
skb->pkt_type = PACKET_HOST;
skb->protocol = eth_type_trans(skb, netdev);
+ skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
+
+#ifdef HAVE_DEV_TSTATS
+ stats = this_cpu_ptr((struct pcpu_sw_netstats __percpu *)netdev->tstats);
+ u64_stats_update_begin(&stats->syncp);
+ stats->rx_packets++;
+ stats->rx_bytes += skb->len;
+ u64_stats_update_end(&stats->syncp);
+#endif
- if (in_interrupt())
- netif_rx(skb);
- else
- netif_rx_ni(skb);
- netdev->last_rx = jiffies;
-
- preempt_disable();
- lb_stats = per_cpu_ptr(internal_dev->lstats, smp_processor_id());
- lb_stats->rx_packets++;
- lb_stats->rx_bytes += len;
- preempt_enable();
-
- return len;
+ netif_rx(skb);
+ return NETDEV_TX_OK;
}
-struct vport_ops internal_vport_ops = {
- .type = "internal",
- .flags = VPORT_F_REQUIRED,
+static struct vport_ops ovs_internal_vport_ops = {
+ .type = OVS_VPORT_TYPE_INTERNAL,
.create = internal_dev_create,
.destroy = internal_dev_destroy,
- .attach = internal_dev_attach,
- .detach = internal_dev_detach,
- .set_mtu = netdev_set_mtu,
- .set_addr = netdev_set_addr,
- .get_name = netdev_get_name,
- .get_addr = netdev_get_addr,
- .get_kobj = netdev_get_kobj,
- .get_stats = netdev_get_stats,
- .get_dev_flags = netdev_get_dev_flags,
- .is_running = netdev_is_running,
- .get_operstate = netdev_get_operstate,
- .get_ifindex = netdev_get_ifindex,
- .get_iflink = netdev_get_iflink,
- .get_mtu = netdev_get_mtu,
.send = internal_dev_recv,
};
-int is_internal_dev(const struct net_device *netdev)
+int ovs_is_internal_dev(const struct net_device *netdev)
{
-#ifdef HAVE_NET_DEVICE_OPS
return netdev->netdev_ops == &internal_dev_netdev_ops;
-#else
- return netdev->open == internal_dev_open;
-#endif
}
-int
-is_internal_vport(const struct vport *vport)
+struct vport *ovs_internal_dev_get_vport(struct net_device *netdev)
+{
+ if (!ovs_is_internal_dev(netdev))
+ return NULL;
+
+ return internal_dev_priv(netdev)->vport;
+}
+
+int ovs_internal_dev_rtnl_link_register(void)
{
- return vport->ops == &internal_vport_ops;
+ int err;
+
+ err = rtnl_link_register(&internal_dev_link_ops);
+ if (err < 0)
+ return err;
+
+ err = ovs_vport_ops_register(&ovs_internal_vport_ops);
+ if (err < 0)
+ rtnl_link_unregister(&internal_dev_link_ops);
+
+ return err;
}
-struct vport *
-internal_dev_get_vport(struct net_device *netdev)
+void ovs_internal_dev_rtnl_link_unregister(void)
{
- struct internal_dev *internal_dev = internal_dev_priv(netdev);
- return rcu_dereference(internal_dev->vport);
+ ovs_vport_ops_unregister(&ovs_internal_vport_ops);
+ rtnl_link_unregister(&internal_dev_link_ops);
}
projects / cascardo / ovs.git / blobdiff