X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=datapath%2Fvport-internal_dev.c;h=637d712d8d371a475ab017c1f9f4f609ac4455ab;hb=HEAD;hp=6d52db0f3881dd86ecdd547479d9998c6009c68a;hpb=f2459fe7d91c4c325dfaa3ed18f56200b63ae27e;p=sliver-openvswitch.git diff --git a/datapath/vport-internal_dev.c b/datapath/vport-internal_dev.c index 6d52db0f3..637d712d8 100644 --- a/datapath/vport-internal_dev.c +++ b/datapath/vport-internal_dev.c @@ -1,98 +1,83 @@ /* - * Copyright (c) 2009, 2010 Nicira Networks. - * Distributed under the terms of the GNU GPL version 2. + * Copyright (c) 2007-2012 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 +#include #include #include #include #include -#include -#include -#include +#include #include -#include +#include + +#include +#include #include "datapath.h" +#include "vlan.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; }; -struct vport_ops 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) +/* This function is only called by the kernel network layer.*/ +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,36) +static struct rtnl_link_stats64 *internal_dev_get_stats(struct net_device *netdev, + struct rtnl_link_stats64 *stats) { - struct internal_dev *internal_dev = internal_dev_priv(netdev); - struct net_device_stats *stats; - int i; - - stats = &internal_dev->stats; - memset(stats, 0, sizeof(struct net_device_stats)); - 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) +#else +static struct net_device_stats *internal_dev_sys_stats(struct net_device *netdev) { - struct sockaddr *addr = p; + struct net_device_stats *stats = &netdev->stats; +#endif + struct vport *vport = ovs_internal_dev_get_vport(netdev); + struct ovs_vport_stats vport_stats; + + ovs_vport_get_stats(vport, &vport_stats); + + /* The tx and rx stats need to be swapped because the + * switch and host OS have opposite perspectives. */ + stats->rx_packets = vport_stats.tx_packets; + stats->tx_packets = vport_stats.rx_packets; + stats->rx_bytes = vport_stats.tx_bytes; + stats->tx_bytes = vport_stats.rx_bytes; + stats->rx_errors = vport_stats.tx_errors; + stats->tx_errors = vport_stats.rx_errors; + stats->rx_dropped = vport_stats.tx_dropped; + stats->tx_dropped = vport_stats.rx_dropped; - if (!is_valid_ether_addr(addr->sa_data)) - return -EADDRNOTAVAIL; - memcpy(dev->dev_addr, addr->sa_data, dev->addr_len); - return 0; + return stats; } -/* 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; - - /* We need our own clone. */ - skb = skb_share_check(skb, GFP_ATOMIC); - if (!skb) - return 0; - - lb_stats = per_cpu_ptr(internal_dev->lstats, smp_processor_id()); - lb_stats->tx_packets++; - lb_stats->tx_bytes += skb->len; - - skb_reset_mac_header(skb); - rcu_read_lock_bh(); - vport_receive(vport, skb); - rcu_read_unlock_bh(); - + rcu_read_lock(); + ovs_vport_receive(internal_dev_priv(netdev)->vport, skb, NULL); + rcu_read_unlock(); return 0; } @@ -111,119 +96,85 @@ static int internal_dev_stop(struct net_device *netdev) 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) -{ - struct internal_dev *internal_dev = internal_dev_priv(netdev); - - internal_dev->lstats = alloc_percpu(struct pcpu_lstats); - if (!internal_dev->lstats) - return -ENOMEM; - - return 0; -} - -static void internal_dev_free(struct net_device *netdev) +static void internal_dev_destructor(struct net_device *dev) { - struct internal_dev *internal_dev = internal_dev_priv(netdev); + struct vport *vport = ovs_internal_dev_get_vport(dev); - free_percpu(internal_dev->lstats); - free_netdev(netdev); + ovs_vport_free(vport); + free_netdev(dev); } -static int internal_dev_do_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) -{ - if (dp_ioctl_hook) - return dp_ioctl_hook(dev, ifr, cmd); - return -EOPNOTSUPP; -} - -#ifdef HAVE_NET_DEVICE_OPS static const struct net_device_ops internal_dev_netdev_ops = { - .ndo_init = internal_dev_init, .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, -}; +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,36) + .ndo_get_stats64 = internal_dev_get_stats, +#else + .ndo_get_stats = internal_dev_sys_stats, #endif +}; -static void -do_setup(struct net_device *netdev) +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; + netdev->priv_flags &= ~IFF_TX_SKB_SHARING; + netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE; + netdev->destructor = internal_dev_destructor; SET_ETHTOOL_OPS(netdev, &internal_dev_ethtool_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; - vport_gen_ether_addr(netdev->dev_addr); + netdev->vlan_features = netdev->features; + netdev->features |= NETIF_F_HW_VLAN_CTAG_TX; + +#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(sizeof(struct netdev_vport), + &ovs_internal_vport_ops, parms); if (IS_ERR(vport)) { err = PTR_ERR(vport); goto error; @@ -231,138 +182,109 @@ internal_dev_create(const char *name, const void __user *config) netdev_vport = netdev_vport_priv(vport); - netdev_vport->dev = alloc_netdev(sizeof(struct internal_dev), name, do_setup); + netdev_vport->dev = alloc_netdev(sizeof(struct internal_dev), + parms->name, do_setup); if (!netdev_vport->dev) { err = -ENOMEM; goto error_free_vport; } + dev_net_set(netdev_vport->dev, ovs_dp_get_net(vport->dp)); internal_dev = internal_dev_priv(netdev_vport->dev); internal_dev->vport = vport; + /* Restrict bridge port to current netns. */ + if (vport->port_no == OVSP_LOCAL) + netdev_vport->dev->features |= NETIF_F_NETNS_LOCAL; + + rtnl_lock(); err = register_netdevice(netdev_vport->dev); if (err) goto error_free_netdev; + dev_set_promiscuity(netdev_vport->dev, 1); + rtnl_unlock(); + netif_start_queue(netdev_vport->dev); + return vport; error_free_netdev: + rtnl_unlock(); free_netdev(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(netdev_vport->dev); + rtnl_lock(); + dev_set_promiscuity(netdev_vport->dev, -1); + + /* unregister_netdevice() waits for an RCU grace period. */ unregister_netdevice(netdev_vport->dev); - vport_free(vport); - return 0; + rtnl_unlock(); } -static int -internal_dev_attach(struct vport *vport) +static int internal_dev_recv(struct vport *vport, 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 = netdev_vport_priv(vport)->dev; + int len; -static int -internal_dev_detach(struct vport *vport) -{ - struct netdev_vport *netdev_vport = netdev_vport_priv(vport); +#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,37) + if (vlan_tx_tag_present(skb)) { + if (unlikely(!__vlan_put_tag(skb, + skb->vlan_proto, + vlan_tx_tag_get(skb)))) + return 0; - dev_set_promiscuity(netdev_vport->dev, -1); + if (skb->ip_summed == CHECKSUM_COMPLETE) + skb->csum = csum_add(skb->csum, + csum_partial(skb->data + (2 * ETH_ALEN), + VLAN_HLEN, 0)); - /* 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); + vlan_set_tci(skb, 0); + } +#endif - return 0; -} + len = skb->len; -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); - 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(); + netif_rx(skb); return len; } -struct vport_ops internal_vport_ops = { - .type = "internal", - .flags = VPORT_F_REQUIRED, +const 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, + .get_name = ovs_netdev_get_name, .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) { - return vport->ops == &internal_vport_ops; -} + if (!ovs_is_internal_dev(netdev)) + return NULL; -struct vport * -internal_dev_get_vport(struct net_device *netdev) -{ - struct internal_dev *internal_dev = internal_dev_priv(netdev); - return rcu_dereference(internal_dev->vport); + return internal_dev_priv(netdev)->vport; }