X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=datapath%2Fvport-netdev.c;h=78f149341ad3353defa2aaf264319ed2cb6e52a9;hb=5ca1ba484bd9ade5116a49cf241cb98219d7d696;hp=2583566b1afd884005e084542122ffc2840fbbbb;hpb=431488e6638d3fbb53d215fa10d2ff2d8f1c2824;p=sliver-openvswitch.git diff --git a/datapath/vport-netdev.c b/datapath/vport-netdev.c index 2583566b1..78f149341 100644 --- a/datapath/vport-netdev.c +++ b/datapath/vport-netdev.c @@ -1,11 +1,23 @@ /* - * Copyright (c) 2010, 2011 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 */ +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + #include #include #include @@ -23,22 +35,35 @@ #include "vport-netdev.h" #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,37) && \ - !defined(HAVE_VLAN_BUG_WORKAROUND) + !defined(HAVE_VLAN_BUG_WORKAROUND) #include -static int vlan_tso __read_mostly = 0; +static int vlan_tso __read_mostly; module_param(vlan_tso, int, 0644); MODULE_PARM_DESC(vlan_tso, "Enable TSO for VLAN packets"); #else #define vlan_tso true #endif -/* If the native device stats aren't 64 bit use the vport stats tracking instead. */ -#define USE_VPORT_STATS (sizeof(((struct net_device_stats *)0)->rx_bytes) < sizeof(u64)) - static void netdev_port_receive(struct vport *vport, struct sk_buff *skb); -#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,36) +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,39) +/* Called with rcu_read_lock and bottom-halves disabled. */ +static rx_handler_result_t netdev_frame_hook(struct sk_buff **pskb) +{ + struct sk_buff *skb = *pskb; + struct vport *vport; + + if (unlikely(skb->pkt_type == PACKET_LOOPBACK)) + return RX_HANDLER_PASS; + + vport = ovs_netdev_get_vport(skb->dev); + + netdev_port_receive(vport, skb); + + return RX_HANDLER_CONSUMED; +} +#elif LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,36) /* Called with rcu_read_lock and bottom-halves disabled. */ static struct sk_buff *netdev_frame_hook(struct sk_buff *skb) { @@ -47,7 +72,7 @@ static struct sk_buff *netdev_frame_hook(struct sk_buff *skb) if (unlikely(skb->pkt_type == PACKET_LOOPBACK)) return skb; - vport = netdev_get_vport(skb->dev); + vport = ovs_netdev_get_vport(skb->dev); netdev_port_receive(vport, skb); @@ -105,7 +130,8 @@ static struct vport *netdev_create(const struct vport_parms *parms) struct netdev_vport *netdev_vport; int err; - vport = vport_alloc(sizeof(struct netdev_vport), &netdev_vport_ops, parms); + vport = ovs_vport_alloc(sizeof(struct netdev_vport), + &ovs_netdev_vport_ops, parms); if (IS_ERR(vport)) { err = PTR_ERR(vport); goto error; @@ -113,7 +139,7 @@ static struct vport *netdev_create(const struct vport_parms *parms) netdev_vport = netdev_vport_priv(vport); - netdev_vport->dev = dev_get_by_name(&init_net, parms->name); + netdev_vport->dev = dev_get_by_name(ovs_dp_get_net(vport->dp), parms->name); if (!netdev_vport->dev) { err = -ENODEV; goto error_free_vport; @@ -121,28 +147,20 @@ static struct vport *netdev_create(const struct vport_parms *parms) if (netdev_vport->dev->flags & IFF_LOOPBACK || netdev_vport->dev->type != ARPHRD_ETHER || - is_internal_dev(netdev_vport->dev)) { + ovs_is_internal_dev(netdev_vport->dev)) { err = -EINVAL; goto error_put; } - /* If we are using the vport stats layer initialize it to the current - * values so we are roughly consistent with the device stats. */ - if (USE_VPORT_STATS) { - struct rtnl_link_stats64 stats; - - err = netdev_get_stats(vport, &stats); - if (!err) - vport_set_stats(vport, &stats); - } - err = netdev_rx_handler_register(netdev_vport->dev, netdev_frame_hook, vport); if (err) goto error_put; dev_set_promiscuity(netdev_vport->dev, 1); +#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,24) dev_disable_lro(netdev_vport->dev); +#endif netdev_vport->dev->priv_flags |= IFF_OVS_DATAPATH; return vport; @@ -150,34 +168,32 @@ static struct vport *netdev_create(const struct vport_parms *parms) error_put: dev_put(netdev_vport->dev); error_free_vport: - vport_free(vport); + ovs_vport_free(vport); error: return ERR_PTR(err); } -static int netdev_destroy(struct vport *vport) +static void free_port_rcu(struct rcu_head *rcu) { - struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - - netdev_vport->dev->priv_flags &= ~IFF_OVS_DATAPATH; - netdev_rx_handler_unregister(netdev_vport->dev); - dev_set_promiscuity(netdev_vport->dev, -1); - - synchronize_rcu(); + struct netdev_vport *netdev_vport = container_of(rcu, + struct netdev_vport, rcu); dev_put(netdev_vport->dev); - vport_free(vport); - - return 0; + ovs_vport_free(vport_from_priv(netdev_vport)); } -int netdev_set_mtu(struct vport *vport, int mtu) +static void netdev_destroy(struct vport *vport) { struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return dev_set_mtu(netdev_vport->dev, mtu); + + netdev_vport->dev->priv_flags &= ~IFF_OVS_DATAPATH; + netdev_rx_handler_unregister(netdev_vport->dev); + dev_set_promiscuity(netdev_vport->dev, -1); + + call_rcu(&netdev_vport->rcu, free_port_rcu); } -int netdev_set_addr(struct vport *vport, const unsigned char *addr) +int ovs_netdev_set_addr(struct vport *vport, const unsigned char *addr) { struct netdev_vport *netdev_vport = netdev_vport_priv(vport); struct sockaddr sa; @@ -188,70 +204,32 @@ int netdev_set_addr(struct vport *vport, const unsigned char *addr) return dev_set_mac_address(netdev_vport->dev, &sa); } -const char *netdev_get_name(const struct vport *vport) +const char *ovs_netdev_get_name(const struct vport *vport) { const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); return netdev_vport->dev->name; } -const unsigned char *netdev_get_addr(const struct vport *vport) +const unsigned char *ovs_netdev_get_addr(const struct vport *vport) { const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); return netdev_vport->dev->dev_addr; } -struct kobject *netdev_get_kobj(const struct vport *vport) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return &netdev_vport->dev->NETDEV_DEV_MEMBER.kobj; -} - -int netdev_get_stats(const struct vport *vport, struct rtnl_link_stats64 *stats) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - dev_get_stats(netdev_vport->dev, stats); - return 0; -} - -unsigned netdev_get_dev_flags(const struct vport *vport) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return dev_get_flags(netdev_vport->dev); -} - -int netdev_is_running(const struct vport *vport) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return netif_running(netdev_vport->dev); -} - -unsigned char netdev_get_operstate(const struct vport *vport) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return netdev_vport->dev->operstate; -} - -int netdev_get_ifindex(const struct vport *vport) +int ovs_netdev_get_ifindex(const struct vport *vport) { const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); return netdev_vport->dev->ifindex; } -int netdev_get_iflink(const struct vport *vport) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return netdev_vport->dev->iflink; -} - -int netdev_get_mtu(const struct vport *vport) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return netdev_vport->dev->mtu; -} - /* Must be called with rcu_read_lock. */ static void netdev_port_receive(struct vport *vport, struct sk_buff *skb) { + if (unlikely(!vport)) { + kfree_skb(skb); + return; + } + /* Make our own copy of the packet. Otherwise we will mangle the * packet for anyone who came before us (e.g. tcpdump via AF_PACKET). * (No one comes after us, since we tell handle_bridge() that we took @@ -260,13 +238,25 @@ static void netdev_port_receive(struct vport *vport, struct sk_buff *skb) if (unlikely(!skb)) return; - skb_warn_if_lro(skb); - skb_push(skb, ETH_HLEN); - compute_ip_summed(skb, false); + + if (unlikely(compute_ip_summed(skb, false))) { + kfree_skb(skb); + return; + } vlan_copy_skb_tci(skb); - vport_receive(vport, skb); + ovs_vport_receive(vport, skb); +} + +static unsigned int packet_length(const struct sk_buff *skb) +{ + unsigned int length = skb->len - ETH_HLEN; + + if (skb->protocol == htons(ETH_P_8021Q)) + length -= VLAN_HLEN; + + return length; } static bool dev_supports_vlan_tx(struct net_device *dev) @@ -285,32 +275,32 @@ static bool dev_supports_vlan_tx(struct net_device *dev) static int netdev_send(struct vport *vport, struct sk_buff *skb) { struct netdev_vport *netdev_vport = netdev_vport_priv(vport); + int mtu = netdev_vport->dev->mtu; int len; + if (unlikely(packet_length(skb) > mtu && !skb_is_gso(skb))) { + net_warn_ratelimited("%s: dropped over-mtu packet: %d > %d\n", + netdev_vport->dev->name, + packet_length(skb), mtu); + goto error; + } + + if (unlikely(skb_warn_if_lro(skb))) + goto error; + skb->dev = netdev_vport->dev; - forward_ip_summed(skb); + forward_ip_summed(skb, true); if (vlan_tx_tag_present(skb) && !dev_supports_vlan_tx(skb->dev)) { - int err; - int features = 0; + int features; -#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,26) - features = skb->dev->features & skb->dev->vlan_features; -#endif - - err = vswitch_skb_checksum_setup(skb); - if (unlikely(err)) { - kfree_skb(skb); - return 0; - } + features = netif_skb_features(skb); if (!vlan_tso) features &= ~(NETIF_F_TSO | NETIF_F_TSO6 | NETIF_F_UFO | NETIF_F_FSO); - if (skb_is_gso(skb) && - (!skb_gso_ok(skb, features) || - unlikely(skb->ip_summed != CHECKSUM_PARTIAL))) { + if (netif_needs_gso(skb, features)) { struct sk_buff *nskb; nskb = skb_gso_segment(skb, features); @@ -325,10 +315,12 @@ static int netdev_send(struct vport *vport, struct sk_buff *skb) goto tag; } - kfree_skb(skb); - skb = nskb; - if (IS_ERR(skb)) + if (IS_ERR(nskb)) { + kfree_skb(skb); return 0; + } + consume_skb(skb); + skb = nskb; len = 0; do { @@ -359,16 +351,21 @@ tag: dev_queue_xmit(skb); return len; + +error: + kfree_skb(skb); + ovs_vport_record_error(vport, VPORT_E_TX_DROPPED); + return 0; } /* Returns null if this device is not attached to a datapath. */ -struct vport *netdev_get_vport(struct net_device *dev) +struct vport *ovs_netdev_get_vport(struct net_device *dev) { -#ifdef IFF_BRIDGE_PORT -#if IFF_BRIDGE_PORT != IFF_OVS_DATAPATH +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,36) +#if IFF_OVS_DATAPATH != 0 if (likely(dev->priv_flags & IFF_OVS_DATAPATH)) #else - if (likely(rcu_access_pointer(dev->rx_handler) == netdev_frame_hook)) + if (likely(rcu_access_pointer(dev->rx_handler) == netdev_frame_hook)) #endif return (struct vport *)rcu_dereference_rtnl(dev->rx_handler_data); else @@ -378,40 +375,31 @@ struct vport *netdev_get_vport(struct net_device *dev) #endif } -const struct vport_ops netdev_vport_ops = { - .type = ODP_VPORT_TYPE_NETDEV, - .flags = (VPORT_F_REQUIRED | - (USE_VPORT_STATS ? VPORT_F_GEN_STATS : 0)), +const struct vport_ops ovs_netdev_vport_ops = { + .type = OVS_VPORT_TYPE_NETDEV, + .flags = VPORT_F_REQUIRED, .init = netdev_init, .exit = netdev_exit, .create = netdev_create, .destroy = netdev_destroy, - .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, + .set_addr = ovs_netdev_set_addr, + .get_name = ovs_netdev_get_name, + .get_addr = ovs_netdev_get_addr, + .get_ifindex = ovs_netdev_get_ifindex, .send = netdev_send, }; #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,36) /* - * In kernels earlier than 2.6.36, Open vSwitch cannot safely coexist with - * the Linux bridge module on any released version of Linux, because there - * is only a single bridge hook function and only a single br_port member - * in struct net_device. + * Enforces, mutual exclusion with the Linux bridge module, by declaring and + * exporting br_should_route_hook. Because the bridge module also exports the + * same symbol, the module loader will refuse to load both modules at the same + * time (e.g. "bridge: exports duplicate symbol br_should_route_hook (owned by + * openvswitch)"). * - * Declaring and exporting this symbol enforces mutual exclusion. The bridge - * module also exports the same symbol, so the module loader will refuse to - * load both modules at the same time (e.g. "bridge: exports duplicate symbol - * br_should_route_hook (owned by openvswitch_mod)"). + * Before Linux 2.6.36, Open vSwitch cannot safely coexist with the Linux + * bridge module, so openvswitch uses this macro in those versions. In + * Linux 2.6.36 and later, Open vSwitch can coexist with the bridge module. * * The use of "typeof" here avoids the need to track changes in the type of * br_should_route_hook over various kernel versions.