diff options
Diffstat (limited to 'net/openvswitch')
| -rw-r--r-- | net/openvswitch/actions.c | 50 | ||||
| -rw-r--r-- | net/openvswitch/datapath.c | 55 | ||||
| -rw-r--r-- | net/openvswitch/datapath.h | 2 | ||||
| -rw-r--r-- | net/openvswitch/vport-internal_dev.c | 24 | ||||
| -rw-r--r-- | net/openvswitch/vport-internal_dev.h | 2 | ||||
| -rw-r--r-- | net/openvswitch/vport-vxlan.c | 5 | ||||
| -rw-r--r-- | net/openvswitch/vport.c | 101 | ||||
| -rw-r--r-- | net/openvswitch/vport.h | 27 |
8 files changed, 224 insertions, 42 deletions
diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c index e70d8b18e962..fe5cda0deb39 100644 --- a/net/openvswitch/actions.c +++ b/net/openvswitch/actions.c @@ -38,7 +38,7 @@ #include "vport.h" static int do_execute_actions(struct datapath *dp, struct sk_buff *skb, - const struct nlattr *attr, int len, bool keep_skb); + const struct nlattr *attr, int len); static int make_writable(struct sk_buff *skb, int write_len) { @@ -434,11 +434,17 @@ static int output_userspace(struct datapath *dp, struct sk_buff *skb, return ovs_dp_upcall(dp, skb, &upcall); } +static bool last_action(const struct nlattr *a, int rem) +{ + return a->nla_len == rem; +} + static int sample(struct datapath *dp, struct sk_buff *skb, const struct nlattr *attr) { const struct nlattr *acts_list = NULL; const struct nlattr *a; + struct sk_buff *sample_skb; int rem; for (a = nla_data(attr), rem = nla_len(attr); rem > 0; @@ -455,8 +461,34 @@ static int sample(struct datapath *dp, struct sk_buff *skb, } } - return do_execute_actions(dp, skb, nla_data(acts_list), - nla_len(acts_list), true); + rem = nla_len(acts_list); + a = nla_data(acts_list); + + /* Actions list is either empty or only contains a single user-space + * action, the latter being a special case as it is the only known + * usage of the sample action. + * In these special cases don't clone the skb as there are no + * side-effects in the nested actions. + * Otherwise, clone in case the nested actions have side effects. + */ + if (likely(rem == 0 || (nla_type(a) == OVS_ACTION_ATTR_USERSPACE && + last_action(a, rem)))) { + sample_skb = skb; + skb_get(skb); + } else { + sample_skb = skb_clone(skb, GFP_ATOMIC); + if (!sample_skb) /* Skip sample action when out of memory. */ + return 0; + } + + /* Note that do_execute_actions() never consumes skb. + * In the case where skb has been cloned above it is the clone that + * is consumed. Otherwise the skb_get(skb) call prevents + * consumption by do_execute_actions(). Thus, it is safe to simply + * return the error code and let the caller (also + * do_execute_actions()) free skb on error. + */ + return do_execute_actions(dp, sample_skb, a, rem); } static int execute_set_action(struct sk_buff *skb, @@ -507,7 +539,7 @@ static int execute_set_action(struct sk_buff *skb, /* Execute a list of actions against 'skb'. */ static int do_execute_actions(struct datapath *dp, struct sk_buff *skb, - const struct nlattr *attr, int len, bool keep_skb) + const struct nlattr *attr, int len) { /* Every output action needs a separate clone of 'skb', but the common * case is just a single output action, so that doing a clone and @@ -562,12 +594,9 @@ static int do_execute_actions(struct datapath *dp, struct sk_buff *skb, } } - if (prev_port != -1) { - if (keep_skb) - skb = skb_clone(skb, GFP_ATOMIC); - + if (prev_port != -1) do_output(dp, skb, prev_port); - } else if (!keep_skb) + else consume_skb(skb); return 0; @@ -579,6 +608,5 @@ int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb) struct sw_flow_actions *acts = rcu_dereference(OVS_CB(skb)->flow->sf_acts); OVS_CB(skb)->tun_key = NULL; - return do_execute_actions(dp, skb, acts->actions, - acts->actions_len, false); + return do_execute_actions(dp, skb, acts->actions, acts->actions_len); } diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c index 9db4bf6740d1..7ad3f029baae 100644 --- a/net/openvswitch/datapath.c +++ b/net/openvswitch/datapath.c @@ -66,16 +66,16 @@ static struct genl_family dp_packet_genl_family; static struct genl_family dp_flow_genl_family; static struct genl_family dp_datapath_genl_family; -static struct genl_multicast_group ovs_dp_flow_multicast_group = { - .name = OVS_FLOW_MCGROUP +static const struct genl_multicast_group ovs_dp_flow_multicast_group = { + .name = OVS_FLOW_MCGROUP, }; -static struct genl_multicast_group ovs_dp_datapath_multicast_group = { - .name = OVS_DATAPATH_MCGROUP +static const struct genl_multicast_group ovs_dp_datapath_multicast_group = { + .name = OVS_DATAPATH_MCGROUP, }; -struct genl_multicast_group ovs_dp_vport_multicast_group = { - .name = OVS_VPORT_MCGROUP +static const struct genl_multicast_group ovs_dp_vport_multicast_group = { + .name = OVS_VPORT_MCGROUP, }; /* Check if need to build a reply message. @@ -266,7 +266,7 @@ void ovs_dp_process_received_packet(struct vport *p, struct sk_buff *skb) upcall.cmd = OVS_PACKET_CMD_MISS; upcall.key = &key; upcall.userdata = NULL; - upcall.portid = p->upcall_portid; + upcall.portid = ovs_vport_find_upcall_portid(p, skb); ovs_dp_upcall(dp, skb, &upcall); consume_skb(skb); stats_counter = &stats->n_missed; @@ -464,7 +464,8 @@ static int queue_userspace_packet(struct datapath *dp, struct sk_buff *skb, upcall->dp_ifindex = dp_ifindex; nla = nla_nest_start(user_skb, OVS_PACKET_ATTR_KEY); - ovs_nla_put_flow(upcall_info->key, upcall_info->key, user_skb); + err = ovs_nla_put_flow(upcall_info->key, upcall_info->key, user_skb); + BUG_ON(err); nla_nest_end(user_skb, nla); if (upcall_info->userdata) @@ -780,7 +781,7 @@ static struct sk_buff *ovs_flow_cmd_build_info(const struct sw_flow *flow, skb = ovs_flow_cmd_alloc_info(ovsl_dereference(flow->sf_acts), info, always); - if (!skb || IS_ERR(skb)) + if (IS_ERR_OR_NULL(skb)) return skb; retval = ovs_flow_cmd_fill_info(flow, dp_ifindex, skb, @@ -1189,7 +1190,7 @@ static const struct nla_policy flow_policy[OVS_FLOW_ATTR_MAX + 1] = { [OVS_FLOW_ATTR_CLEAR] = { .type = NLA_FLAG }, }; -static struct genl_ops dp_flow_genl_ops[] = { +static const struct genl_ops dp_flow_genl_ops[] = { { .cmd = OVS_FLOW_CMD_NEW, .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */ .policy = flow_policy, @@ -1373,7 +1374,7 @@ static int ovs_dp_cmd_new(struct sk_buff *skb, struct genl_info *info) parms.options = NULL; parms.dp = dp; parms.port_no = OVSP_LOCAL; - parms.upcall_portid = nla_get_u32(a[OVS_DP_ATTR_UPCALL_PID]); + parms.upcall_portids = a[OVS_DP_ATTR_UPCALL_PID]; ovs_dp_change(dp, a); @@ -1577,7 +1578,7 @@ static const struct nla_policy datapath_policy[OVS_DP_ATTR_MAX + 1] = { [OVS_DP_ATTR_USER_FEATURES] = { .type = NLA_U32 }, }; -static struct genl_ops dp_datapath_genl_ops[] = { +static const struct genl_ops dp_datapath_genl_ops[] = { { .cmd = OVS_DP_CMD_NEW, .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */ .policy = datapath_policy, @@ -1632,8 +1633,8 @@ static int ovs_vport_cmd_fill_info(struct vport *vport, struct sk_buff *skb, if (nla_put_u32(skb, OVS_VPORT_ATTR_PORT_NO, vport->port_no) || nla_put_u32(skb, OVS_VPORT_ATTR_TYPE, vport->ops->type) || - nla_put_string(skb, OVS_VPORT_ATTR_NAME, vport->ops->get_name(vport)) || - nla_put_u32(skb, OVS_VPORT_ATTR_UPCALL_PID, vport->upcall_portid)) + nla_put_string(skb, OVS_VPORT_ATTR_NAME, + vport->ops->get_name(vport))) goto nla_put_failure; ovs_vport_get_stats(vport, &vport_stats); @@ -1641,6 +1642,9 @@ static int ovs_vport_cmd_fill_info(struct vport *vport, struct sk_buff *skb, &vport_stats)) goto nla_put_failure; + if (ovs_vport_get_upcall_portids(vport, skb)) + goto nla_put_failure; + err = ovs_vport_get_options(vport, skb); if (err == -EMSGSIZE) goto error; @@ -1762,7 +1766,7 @@ static int ovs_vport_cmd_new(struct sk_buff *skb, struct genl_info *info) parms.options = a[OVS_VPORT_ATTR_OPTIONS]; parms.dp = dp; parms.port_no = port_no; - parms.upcall_portid = nla_get_u32(a[OVS_VPORT_ATTR_UPCALL_PID]); + parms.upcall_portids = a[OVS_VPORT_ATTR_UPCALL_PID]; vport = new_vport(&parms); err = PTR_ERR(vport); @@ -1812,8 +1816,14 @@ static int ovs_vport_cmd_set(struct sk_buff *skb, struct genl_info *info) goto exit_unlock_free; } - if (a[OVS_VPORT_ATTR_UPCALL_PID]) - vport->upcall_portid = nla_get_u32(a[OVS_VPORT_ATTR_UPCALL_PID]); + + if (a[OVS_VPORT_ATTR_UPCALL_PID]) { + struct nlattr *ids = a[OVS_VPORT_ATTR_UPCALL_PID]; + + err = ovs_vport_set_upcall_portids(vport, ids); + if (err) + goto exit_unlock_free; + } err = ovs_vport_cmd_fill_info(vport, reply, info->snd_portid, info->snd_seq, 0, OVS_VPORT_CMD_NEW); @@ -1944,7 +1954,7 @@ static const struct nla_policy vport_policy[OVS_VPORT_ATTR_MAX + 1] = { [OVS_VPORT_ATTR_OPTIONS] = { .type = NLA_NESTED }, }; -static struct genl_ops dp_vport_genl_ops[] = { +static const struct genl_ops dp_vport_genl_ops[] = { { .cmd = OVS_VPORT_CMD_NEW, .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */ .policy = vport_policy, @@ -2053,10 +2063,14 @@ static int __init dp_init(void) pr_info("Open vSwitch switching datapath\n"); - err = ovs_flow_init(); + err = ovs_internal_dev_rtnl_link_register(); if (err) goto error; + err = ovs_flow_init(); + if (err) + goto error_unreg_rtnl_link; + err = ovs_vport_init(); if (err) goto error_flow_exit; @@ -2083,6 +2097,8 @@ error_vport_exit: ovs_vport_exit(); error_flow_exit: ovs_flow_exit(); +error_unreg_rtnl_link: + ovs_internal_dev_rtnl_link_unregister(); error: return err; } @@ -2095,6 +2111,7 @@ static void dp_cleanup(void) rcu_barrier(); ovs_vport_exit(); ovs_flow_exit(); + ovs_internal_dev_rtnl_link_unregister(); } module_init(dp_init); diff --git a/net/openvswitch/datapath.h b/net/openvswitch/datapath.h index 7ede507500d7..701b5738c38a 100644 --- a/net/openvswitch/datapath.h +++ b/net/openvswitch/datapath.h @@ -144,7 +144,7 @@ int lockdep_ovsl_is_held(void); #define lockdep_ovsl_is_held() 1 #endif -#define ASSERT_OVSL() WARN_ON(unlikely(!lockdep_ovsl_is_held())) +#define ASSERT_OVSL() WARN_ON(!lockdep_ovsl_is_held()) #define ovsl_dereference(p) \ rcu_dereference_protected(p, lockdep_ovsl_is_held()) #define rcu_dereference_ovsl(p) \ diff --git a/net/openvswitch/vport-internal_dev.c b/net/openvswitch/vport-internal_dev.c index 789af9280e77..84516126e5f3 100644 --- a/net/openvswitch/vport-internal_dev.c +++ b/net/openvswitch/vport-internal_dev.c @@ -26,6 +26,7 @@ #include <net/dst.h> #include <net/xfrm.h> +#include <net/rtnetlink.h> #include "datapath.h" #include "vport-internal_dev.h" @@ -121,6 +122,10 @@ static const struct net_device_ops internal_dev_netdev_ops = { .ndo_get_stats64 = internal_dev_get_stats, }; +static struct rtnl_link_ops internal_dev_link_ops __read_mostly = { + .kind = "openvswitch", +}; + static void do_setup(struct net_device *netdev) { ether_setup(netdev); @@ -131,14 +136,18 @@ static void do_setup(struct net_device *netdev) netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE; 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->features = NETIF_F_LLTX | NETIF_F_SG | NETIF_F_FRAGLIST | - NETIF_F_HIGHDMA | NETIF_F_HW_CSUM | NETIF_F_GSO_SOFTWARE; + NETIF_F_HIGHDMA | NETIF_F_HW_CSUM | + NETIF_F_GSO_SOFTWARE | NETIF_F_GSO_ENCAP_ALL; netdev->vlan_features = netdev->features; + netdev->hw_enc_features = netdev->features; netdev->features |= NETIF_F_HW_VLAN_CTAG_TX; netdev->hw_features = netdev->features & ~NETIF_F_LLTX; + eth_hw_addr_random(netdev); } @@ -159,7 +168,8 @@ static struct vport *internal_dev_create(const struct vport_parms *parms) netdev_vport = netdev_vport_priv(vport); netdev_vport->dev = alloc_netdev(sizeof(struct internal_dev), - parms->name, do_setup); + parms->name, NET_NAME_UNKNOWN, + do_setup); if (!netdev_vport->dev) { err = -ENOMEM; goto error_free_vport; @@ -248,3 +258,13 @@ struct vport *ovs_internal_dev_get_vport(struct net_device *netdev) return internal_dev_priv(netdev)->vport; } + +int ovs_internal_dev_rtnl_link_register(void) +{ + return rtnl_link_register(&internal_dev_link_ops); +} + +void ovs_internal_dev_rtnl_link_unregister(void) +{ + rtnl_link_unregister(&internal_dev_link_ops); +} diff --git a/net/openvswitch/vport-internal_dev.h b/net/openvswitch/vport-internal_dev.h index 9a7d30ecc6a2..1b179a190cff 100644 --- a/net/openvswitch/vport-internal_dev.h +++ b/net/openvswitch/vport-internal_dev.h @@ -24,5 +24,7 @@ int ovs_is_internal_dev(const struct net_device *); struct vport *ovs_internal_dev_get_vport(struct net_device *); +int ovs_internal_dev_rtnl_link_register(void); +void ovs_internal_dev_rtnl_link_unregister(void); #endif /* vport-internal_dev.h */ diff --git a/net/openvswitch/vport-vxlan.c b/net/openvswitch/vport-vxlan.c index 0edbd95c60e7..d8b7e247bebf 100644 --- a/net/openvswitch/vport-vxlan.c +++ b/net/openvswitch/vport-vxlan.c @@ -143,8 +143,6 @@ static int vxlan_tnl_send(struct vport *vport, struct sk_buff *skb) struct rtable *rt; struct flowi4 fl; __be16 src_port; - int port_min; - int port_max; __be16 df; int err; @@ -172,8 +170,7 @@ static int vxlan_tnl_send(struct vport *vport, struct sk_buff *skb) skb->ignore_df = 1; - inet_get_local_port_range(net, &port_min, &port_max); - src_port = vxlan_src_port(port_min, port_max, skb); + src_port = udp_flow_src_port(net, skb, 0, 0, true); err = vxlan_xmit_skb(vxlan_port->vs, rt, skb, fl.saddr, OVS_CB(skb)->tun_key->ipv4_dst, diff --git a/net/openvswitch/vport.c b/net/openvswitch/vport.c index 42c0f4a0b78c..702fb21bfe15 100644 --- a/net/openvswitch/vport.c +++ b/net/openvswitch/vport.c @@ -134,10 +134,12 @@ struct vport *ovs_vport_alloc(int priv_size, const struct vport_ops *ops, vport->dp = parms->dp; vport->port_no = parms->port_no; - vport->upcall_portid = parms->upcall_portid; vport->ops = ops; INIT_HLIST_NODE(&vport->dp_hash_node); + if (ovs_vport_set_upcall_portids(vport, parms->upcall_portids)) + return ERR_PTR(-EINVAL); + vport->percpu_stats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats); if (!vport->percpu_stats) { kfree(vport); @@ -161,6 +163,10 @@ struct vport *ovs_vport_alloc(int priv_size, const struct vport_ops *ops, */ void ovs_vport_free(struct vport *vport) { + /* vport is freed from RCU callback or error path, Therefore + * it is safe to use raw dereference. + */ + kfree(rcu_dereference_raw(vport->upcall_portids)); free_percpu(vport->percpu_stats); kfree(vport); } @@ -327,6 +333,99 @@ int ovs_vport_get_options(const struct vport *vport, struct sk_buff *skb) } /** + * ovs_vport_set_upcall_portids - set upcall portids of @vport. + * + * @vport: vport to modify. + * @ids: new configuration, an array of port ids. + * + * Sets the vport's upcall_portids to @ids. + * + * Returns 0 if successful, -EINVAL if @ids is zero length or cannot be parsed + * as an array of U32. + * + * Must be called with ovs_mutex. + */ +int ovs_vport_set_upcall_portids(struct vport *vport, struct nlattr *ids) +{ + struct vport_portids *old, *vport_portids; + + if (!nla_len(ids) || nla_len(ids) % sizeof(u32)) + return -EINVAL; + + old = ovsl_dereference(vport->upcall_portids); + + vport_portids = kmalloc(sizeof(*vport_portids) + nla_len(ids), + GFP_KERNEL); + if (!vport_portids) + return -ENOMEM; + + vport_portids->n_ids = nla_len(ids) / sizeof(u32); + vport_portids->rn_ids = reciprocal_value(vport_portids->n_ids); + nla_memcpy(vport_portids->ids, ids, nla_len(ids)); + + rcu_assign_pointer(vport->upcall_portids, vport_portids); + + if (old) + kfree_rcu(old, rcu); + return 0; +} + +/** + * ovs_vport_get_upcall_portids - get the upcall_portids of @vport. + * + * @vport: vport from which to retrieve the portids. + * @skb: sk_buff where portids should be appended. + * + * Retrieves the configuration of the given vport, appending the + * %OVS_VPORT_ATTR_UPCALL_PID attribute which is the array of upcall + * portids to @skb. + * + * Returns 0 if successful, -EMSGSIZE if @skb has insufficient room. + * If an error occurs, @skb is left unmodified. Must be called with + * ovs_mutex or rcu_read_lock. + */ +int ovs_vport_get_upcall_portids(const struct vport *vport, + struct sk_buff *skb) +{ + struct vport_portids *ids; + + ids = rcu_dereference_ovsl(vport->upcall_portids); + + if (vport->dp->user_features & OVS_DP_F_VPORT_PIDS) + return nla_put(skb, OVS_VPORT_ATTR_UPCALL_PID, + ids->n_ids * sizeof(u32), (void *)ids->ids); + else + return nla_put_u32(skb, OVS_VPORT_ATTR_UPCALL_PID, ids->ids[0]); +} + +/** + * ovs_vport_find_upcall_portid - find the upcall portid to send upcall. + * + * @vport: vport from which the missed packet is received. + * @skb: skb that the missed packet was received. + * + * Uses the skb_get_hash() to select the upcall portid to send the + * upcall. + * + * Returns the portid of the target socket. Must be called with rcu_read_lock. + */ +u32 ovs_vport_find_upcall_portid(const struct vport *p, struct sk_buff *skb) +{ + struct vport_portids *ids; + u32 ids_index; + u32 hash; + + ids = rcu_dereference(p->upcall_portids); + + if (ids->n_ids == 1 && ids->ids[0] == 0) + return 0; + + hash = skb_get_hash(skb); + ids_index = hash - ids->n_ids * reciprocal_divide(hash, ids->rn_ids); + return ids->ids[ids_index]; +} + +/** * ovs_vport_receive - pass up received packet to the datapath for processing * * @vport: vport that received the packet diff --git a/net/openvswitch/vport.h b/net/openvswitch/vport.h index 8d721e62f388..35f89d84b45e 100644 --- a/net/openvswitch/vport.h +++ b/net/openvswitch/vport.h @@ -23,6 +23,7 @@ #include <linux/list.h> #include <linux/netlink.h> #include <linux/openvswitch.h> +#include <linux/reciprocal_div.h> #include <linux/skbuff.h> #include <linux/spinlock.h> #include <linux/u64_stats_sync.h> @@ -52,6 +53,10 @@ void ovs_vport_get_stats(struct vport *, struct ovs_vport_stats *); int ovs_vport_set_options(struct vport *, struct nlattr *options); int ovs_vport_get_options(const struct vport *, struct sk_buff *); +int ovs_vport_set_upcall_portids(struct vport *, struct nlattr *pids); +int ovs_vport_get_upcall_portids(const struct vport *, struct sk_buff *); +u32 ovs_vport_find_upcall_portid(const struct vport *, struct sk_buff *); + int ovs_vport_send(struct vport *, struct sk_buff *); /* The following definitions are for implementers of vport devices: */ @@ -62,13 +67,27 @@ struct vport_err_stats { u64 tx_dropped; u64 tx_errors; }; +/** + * struct vport_portids - array of netlink portids of a vport. + * must be protected by rcu. + * @rn_ids: The reciprocal value of @n_ids. + * @rcu: RCU callback head for deferred destruction. + * @n_ids: Size of @ids array. + * @ids: Array storing the Netlink socket pids to be used for packets received + * on this port that miss the flow table. + */ +struct vport_portids { + struct reciprocal_value rn_ids; + struct rcu_head rcu; + u32 n_ids; + u32 ids[]; +}; /** * struct vport - one port within a datapath * @rcu: RCU callback head for deferred destruction. * @dp: Datapath to which this port belongs. - * @upcall_portid: The Netlink port to use for packets received on this port that - * miss the flow table. + * @upcall_portids: RCU protected 'struct vport_portids'. * @port_no: Index into @dp's @ports array. * @hash_node: Element in @dev_table hash table in vport.c. * @dp_hash_node: Element in @datapath->ports hash table in datapath.c. @@ -80,7 +99,7 @@ struct vport_err_stats { struct vport { struct rcu_head rcu; struct datapath *dp; - u32 upcall_portid; + struct vport_portids __rcu *upcall_portids; u16 port_no; struct hlist_node hash_node; @@ -111,7 +130,7 @@ struct vport_parms { /* For ovs_vport_alloc(). */ struct datapath *dp; u16 port_no; - u32 upcall_portid; + struct nlattr *upcall_portids; }; /** |