X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=datapath%2Factions.c;h=faa6a0023df87d67adab6e09afff9eae7ed54d60;hb=72e8bf28bb38e8816435c64859fb350215b6a9e6;hp=470e617b97cb26c51b2f1898fc1d11eb23329ab0;hpb=98403001ece61cbf783297c467a06032b200b8d0;p=sliver-openvswitch.git diff --git a/datapath/actions.c b/datapath/actions.c index 470e617b9..faa6a0023 100644 --- a/datapath/actions.c +++ b/datapath/actions.c @@ -1,36 +1,45 @@ /* - * Distributed under the terms of the GNU GPL version 2. - * Copyright (c) 2007, 2008, 2009, 2010, 2011 Nicira Networks. + * 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 */ -/* Functions for executing flow actions. */ - #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt #include #include #include +#include #include #include #include #include #include -#include #include +#include #include +#include -#include "actions.h" #include "checksum.h" #include "datapath.h" -#include "openvswitch/datapath-protocol.h" #include "vlan.h" #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, + struct ovs_key_ipv4_tunnel *tun_key, bool keep_skb); static int make_writable(struct sk_buff *skb, int write_len) { @@ -40,11 +49,10 @@ static int make_writable(struct sk_buff *skb, int write_len) return pskb_expand_head(skb, 0, 0, GFP_ATOMIC); } -/* remove VLAN header from packet and update csum accrodingly. */ +/* remove VLAN header from packet and update csum accordingly. */ static int __pop_vlan_tci(struct sk_buff *skb, __be16 *current_tci) { - struct ethhdr *eh; - struct vlan_ethhdr *veth; + struct vlan_hdr *vhdr; int err; err = make_writable(skb, VLAN_ETH_HLEN); @@ -55,15 +63,15 @@ static int __pop_vlan_tci(struct sk_buff *skb, __be16 *current_tci) skb->csum = csum_sub(skb->csum, csum_partial(skb->data + ETH_HLEN, VLAN_HLEN, 0)); - veth = (struct vlan_ethhdr *) skb->data; - *current_tci = veth->h_vlan_TCI; + vhdr = (struct vlan_hdr *)(skb->data + ETH_HLEN); + *current_tci = vhdr->h_vlan_TCI; memmove(skb->data + VLAN_HLEN, skb->data, 2 * ETH_ALEN); + __skb_pull(skb, VLAN_HLEN); - eh = (struct ethhdr *)__skb_pull(skb, VLAN_HLEN); - - skb->protocol = eh->h_proto; + vlan_set_encap_proto(skb, vhdr); skb->mac_header += VLAN_HLEN; + skb_reset_mac_len(skb); return 0; } @@ -77,7 +85,7 @@ static int pop_vlan(struct sk_buff *skb) vlan_set_tci(skb, 0); } else { if (unlikely(skb->protocol != htons(ETH_P_8021Q) || - skb->len < VLAN_ETH_HLEN)) + skb->len < VLAN_ETH_HLEN)) return 0; err = __pop_vlan_tci(skb, &tci); @@ -86,7 +94,7 @@ static int pop_vlan(struct sk_buff *skb) } /* move next vlan tag to hw accel tag */ if (likely(skb->protocol != htons(ETH_P_8021Q) || - skb->len < VLAN_ETH_HLEN)) + skb->len < VLAN_ETH_HLEN)) return 0; err = __pop_vlan_tci(skb, &tci); @@ -97,7 +105,7 @@ static int pop_vlan(struct sk_buff *skb) return 0; } -static int push_vlan(struct sk_buff *skb, __be16 new_tci) +static int push_vlan(struct sk_buff *skb, const struct ovs_action_push_vlan *vlan) { if (unlikely(vlan_tx_tag_present(skb))) { u16 current_tag; @@ -113,117 +121,234 @@ static int push_vlan(struct sk_buff *skb, __be16 new_tci) + ETH_HLEN, VLAN_HLEN, 0)); } - __vlan_hwaccel_put_tag(skb, ntohs(new_tci)); + __vlan_hwaccel_put_tag(skb, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT); + return 0; +} + +static int set_eth_addr(struct sk_buff *skb, + const struct ovs_key_ethernet *eth_key) +{ + int err; + err = make_writable(skb, ETH_HLEN); + if (unlikely(err)) + return err; + + memcpy(eth_hdr(skb)->h_source, eth_key->eth_src, ETH_ALEN); + memcpy(eth_hdr(skb)->h_dest, eth_key->eth_dst, ETH_ALEN); + return 0; } -static bool is_ip(struct sk_buff *skb) +static void set_ip_addr(struct sk_buff *skb, struct iphdr *nh, + __be32 *addr, __be32 new_addr) { - return (OVS_CB(skb)->flow->key.eth.type == htons(ETH_P_IP) && - skb->transport_header > skb->network_header); + int transport_len = skb->len - skb_transport_offset(skb); + + if (nh->protocol == IPPROTO_TCP) { + if (likely(transport_len >= sizeof(struct tcphdr))) + inet_proto_csum_replace4(&tcp_hdr(skb)->check, skb, + *addr, new_addr, 1); + } else if (nh->protocol == IPPROTO_UDP) { + if (likely(transport_len >= sizeof(struct udphdr))) { + struct udphdr *uh = udp_hdr(skb); + + if (uh->check || + get_ip_summed(skb) == OVS_CSUM_PARTIAL) { + inet_proto_csum_replace4(&uh->check, skb, + *addr, new_addr, 1); + if (!uh->check) + uh->check = CSUM_MANGLED_0; + } + } + } + + csum_replace4(&nh->check, *addr, new_addr); + skb_clear_rxhash(skb); + *addr = new_addr; } -static __sum16 *get_l4_checksum(struct sk_buff *skb) +static void update_ipv6_checksum(struct sk_buff *skb, u8 l4_proto, + __be32 addr[4], const __be32 new_addr[4]) { - u8 nw_proto = OVS_CB(skb)->flow->key.ip.proto; int transport_len = skb->len - skb_transport_offset(skb); - if (nw_proto == IPPROTO_TCP) { + + if (l4_proto == IPPROTO_TCP) { if (likely(transport_len >= sizeof(struct tcphdr))) - return &tcp_hdr(skb)->check; - } else if (nw_proto == IPPROTO_UDP) { - if (likely(transport_len >= sizeof(struct udphdr))) - return &udp_hdr(skb)->check; + inet_proto_csum_replace16(&tcp_hdr(skb)->check, skb, + addr, new_addr, 1); + } else if (l4_proto == IPPROTO_UDP) { + if (likely(transport_len >= sizeof(struct udphdr))) { + struct udphdr *uh = udp_hdr(skb); + + if (uh->check || + get_ip_summed(skb) == OVS_CSUM_PARTIAL) { + inet_proto_csum_replace16(&uh->check, skb, + addr, new_addr, 1); + if (!uh->check) + uh->check = CSUM_MANGLED_0; + } + } } - return NULL; } -static int set_nw_addr(struct sk_buff *skb, const struct nlattr *a) +static void set_ipv6_addr(struct sk_buff *skb, u8 l4_proto, + __be32 addr[4], const __be32 new_addr[4], + bool recalculate_csum) +{ + if (recalculate_csum) + update_ipv6_checksum(skb, l4_proto, addr, new_addr); + + skb_clear_rxhash(skb); + memcpy(addr, new_addr, sizeof(__be32[4])); +} + +static void set_ipv6_tc(struct ipv6hdr *nh, u8 tc) +{ + nh->priority = tc >> 4; + nh->flow_lbl[0] = (nh->flow_lbl[0] & 0x0F) | ((tc & 0x0F) << 4); +} + +static void set_ipv6_fl(struct ipv6hdr *nh, u32 fl) +{ + nh->flow_lbl[0] = (nh->flow_lbl[0] & 0xF0) | (fl & 0x000F0000) >> 16; + nh->flow_lbl[1] = (fl & 0x0000FF00) >> 8; + nh->flow_lbl[2] = fl & 0x000000FF; +} + +static void set_ip_ttl(struct sk_buff *skb, struct iphdr *nh, u8 new_ttl) +{ + csum_replace2(&nh->check, htons(nh->ttl << 8), htons(new_ttl << 8)); + nh->ttl = new_ttl; +} + +static int set_ipv4(struct sk_buff *skb, const struct ovs_key_ipv4 *ipv4_key) { - __be32 new_nwaddr = nla_get_be32(a); struct iphdr *nh; - __sum16 *check; - __be32 *nwaddr; int err; - if (unlikely(!is_ip(skb))) - return 0; - err = make_writable(skb, skb_network_offset(skb) + sizeof(struct iphdr)); if (unlikely(err)) return err; nh = ip_hdr(skb); - nwaddr = nla_type(a) == OVS_ACTION_ATTR_SET_NW_SRC ? &nh->saddr : &nh->daddr; - check = get_l4_checksum(skb); - if (likely(check)) - inet_proto_csum_replace4(check, skb, *nwaddr, new_nwaddr, 1); - csum_replace4(&nh->check, *nwaddr, new_nwaddr); + if (ipv4_key->ipv4_src != nh->saddr) + set_ip_addr(skb, nh, &nh->saddr, ipv4_key->ipv4_src); - skb_clear_rxhash(skb); + if (ipv4_key->ipv4_dst != nh->daddr) + set_ip_addr(skb, nh, &nh->daddr, ipv4_key->ipv4_dst); - *nwaddr = new_nwaddr; + if (ipv4_key->ipv4_tos != nh->tos) + ipv4_change_dsfield(nh, 0, ipv4_key->ipv4_tos); + + if (ipv4_key->ipv4_ttl != nh->ttl) + set_ip_ttl(skb, nh, ipv4_key->ipv4_ttl); return 0; } -static int set_nw_tos(struct sk_buff *skb, u8 nw_tos) +static int set_ipv6(struct sk_buff *skb, const struct ovs_key_ipv6 *ipv6_key) { - struct iphdr *nh = ip_hdr(skb); - u8 old, new; + struct ipv6hdr *nh; int err; - - if (unlikely(!is_ip(skb))) - return 0; + __be32 *saddr; + __be32 *daddr; err = make_writable(skb, skb_network_offset(skb) + - sizeof(struct iphdr)); + sizeof(struct ipv6hdr)); if (unlikely(err)) return err; - /* Set the DSCP bits and preserve the ECN bits. */ - old = nh->tos; - new = nw_tos | (nh->tos & INET_ECN_MASK); - csum_replace4(&nh->check, (__force __be32)old, - (__force __be32)new); - nh->tos = new; + nh = ipv6_hdr(skb); + saddr = (__be32 *)&nh->saddr; + daddr = (__be32 *)&nh->daddr; + + if (memcmp(ipv6_key->ipv6_src, saddr, sizeof(ipv6_key->ipv6_src))) + set_ipv6_addr(skb, ipv6_key->ipv6_proto, saddr, + ipv6_key->ipv6_src, true); + + if (memcmp(ipv6_key->ipv6_dst, daddr, sizeof(ipv6_key->ipv6_dst))) { + unsigned int offset = 0; + int flags = OVS_IP6T_FH_F_SKIP_RH; + bool recalc_csum = true; + + if (ipv6_ext_hdr(nh->nexthdr)) + recalc_csum = ipv6_find_hdr(skb, &offset, + NEXTHDR_ROUTING, NULL, + &flags) != NEXTHDR_ROUTING; + + set_ipv6_addr(skb, ipv6_key->ipv6_proto, daddr, + ipv6_key->ipv6_dst, recalc_csum); + } + + set_ipv6_tc(nh, ipv6_key->ipv6_tclass); + set_ipv6_fl(nh, ntohl(ipv6_key->ipv6_label)); + nh->hop_limit = ipv6_key->ipv6_hlimit; return 0; } -static int set_tp_port(struct sk_buff *skb, const struct nlattr *a) +/* Must follow make_writable() since that can move the skb data. */ +static void set_tp_port(struct sk_buff *skb, __be16 *port, + __be16 new_port, __sum16 *check) +{ + inet_proto_csum_replace2(check, skb, *port, new_port, 0); + *port = new_port; + skb_clear_rxhash(skb); +} + +static void set_udp_port(struct sk_buff *skb, __be16 *port, __be16 new_port) +{ + struct udphdr *uh = udp_hdr(skb); + + if (uh->check && get_ip_summed(skb) != OVS_CSUM_PARTIAL) { + set_tp_port(skb, port, new_port, &uh->check); + + if (!uh->check) + uh->check = CSUM_MANGLED_0; + } else { + *port = new_port; + skb_clear_rxhash(skb); + } +} + +static int set_udp(struct sk_buff *skb, const struct ovs_key_udp *udp_port_key) { - struct udphdr *th; - __sum16 *check; - __be16 *port; + struct udphdr *uh; int err; - if (unlikely(!is_ip(skb))) - return 0; + err = make_writable(skb, skb_transport_offset(skb) + + sizeof(struct udphdr)); + if (unlikely(err)) + return err; + + uh = udp_hdr(skb); + if (udp_port_key->udp_src != uh->source) + set_udp_port(skb, &uh->source, udp_port_key->udp_src); + + if (udp_port_key->udp_dst != uh->dest) + set_udp_port(skb, &uh->dest, udp_port_key->udp_dst); + + return 0; +} + +static int set_tcp(struct sk_buff *skb, const struct ovs_key_tcp *tcp_port_key) +{ + struct tcphdr *th; + int err; err = make_writable(skb, skb_transport_offset(skb) + sizeof(struct tcphdr)); if (unlikely(err)) return err; - /* Must follow make_writable() since that can move the skb data. */ - check = get_l4_checksum(skb); - if (unlikely(!check)) - return 0; + th = tcp_hdr(skb); + if (tcp_port_key->tcp_src != th->source) + set_tp_port(skb, &th->source, tcp_port_key->tcp_src, &th->check); - /* - * Update port and checksum. - * - * This is OK because source and destination port numbers are at the - * same offsets in both UDP and TCP headers, and get_l4_checksum() only - * supports those protocols. - */ - th = udp_hdr(skb); - port = nla_type(a) == OVS_ACTION_ATTR_SET_TP_SRC ? &th->source : &th->dest; - inet_proto_csum_replace2(check, skb, *port, nla_get_be16(a), 0); - *port = nla_get_be16(a); - skb_clear_rxhash(skb); + if (tcp_port_key->tcp_dst != th->dest) + set_tp_port(skb, &th->dest, tcp_port_key->tcp_dst, &th->check); return 0; } @@ -235,13 +360,13 @@ static int do_output(struct datapath *dp, struct sk_buff *skb, int out_port) if (unlikely(!skb)) return -ENOMEM; - vport = rcu_dereference(dp->ports[out_port]); + vport = ovs_vport_rcu(dp, out_port); if (unlikely(!vport)) { kfree_skb(skb); return -ENODEV; } - vport_send(vport, skb); + ovs_vport_send(vport, skb); return 0; } @@ -270,11 +395,12 @@ static int output_userspace(struct datapath *dp, struct sk_buff *skb, } } - return dp_upcall(dp, skb, &upcall); + return ovs_dp_upcall(dp, skb, &upcall); } static int sample(struct datapath *dp, struct sk_buff *skb, - const struct nlattr *attr) + const struct nlattr *attr, + struct ovs_key_ipv4_tunnel *tun_key) { const struct nlattr *acts_list = NULL; const struct nlattr *a; @@ -295,19 +421,78 @@ 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); + nla_len(acts_list), tun_key, true); +} + +static int execute_set_action(struct sk_buff *skb, + const struct nlattr *nested_attr, + struct ovs_key_ipv4_tunnel *tun_key) +{ + int err = 0; + + switch (nla_type(nested_attr)) { + case OVS_KEY_ATTR_PRIORITY: + skb->priority = nla_get_u32(nested_attr); + break; + + case OVS_KEY_ATTR_SKB_MARK: + skb_set_mark(skb, nla_get_u32(nested_attr)); + break; + + case OVS_KEY_ATTR_TUN_ID: + /* If we're only using the TUN_ID action, store the value in a + * temporary instance of struct ovs_key_ipv4_tunnel on the stack. + * If both IPV4_TUNNEL and TUN_ID are being used together we + * can't write into the IPV4_TUNNEL action, so make a copy and + * write into that version. + */ + if (!OVS_CB(skb)->tun_key) + memset(tun_key, 0, sizeof(*tun_key)); + else if (OVS_CB(skb)->tun_key != tun_key) + memcpy(tun_key, OVS_CB(skb)->tun_key, sizeof(*tun_key)); + OVS_CB(skb)->tun_key = tun_key; + + OVS_CB(skb)->tun_key->tun_id = nla_get_be64(nested_attr); + break; + + case OVS_KEY_ATTR_IPV4_TUNNEL: + OVS_CB(skb)->tun_key = nla_data(nested_attr); + break; + + case OVS_KEY_ATTR_ETHERNET: + err = set_eth_addr(skb, nla_data(nested_attr)); + break; + + case OVS_KEY_ATTR_IPV4: + err = set_ipv4(skb, nla_data(nested_attr)); + break; + + case OVS_KEY_ATTR_IPV6: + err = set_ipv6(skb, nla_data(nested_attr)); + break; + + case OVS_KEY_ATTR_TCP: + err = set_tcp(skb, nla_data(nested_attr)); + break; + + case OVS_KEY_ATTR_UDP: + err = set_udp(skb, nla_data(nested_attr)); + break; + } + + return err; } /* 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, + struct ovs_key_ipv4_tunnel *tun_key, bool keep_skb) { /* 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 * then freeing the original skbuff is wasteful. So the following code * is slightly obscure just to avoid that. */ int prev_port = -1; - u32 priority = skb->priority; const struct nlattr *a; int rem; @@ -329,13 +514,9 @@ static int do_execute_actions(struct datapath *dp, struct sk_buff *skb, output_userspace(dp, skb, a); break; - case OVS_ACTION_ATTR_SET_TUNNEL: - OVS_CB(skb)->tun_id = nla_get_be64(a); - break; - case OVS_ACTION_ATTR_PUSH_VLAN: - err = push_vlan(skb, nla_get_be16(a)); - if (unlikely(err)) /* skb already freed */ + err = push_vlan(skb, nla_data(a)); + if (unlikely(err)) /* skb already freed. */ return err; break; @@ -343,45 +524,15 @@ static int do_execute_actions(struct datapath *dp, struct sk_buff *skb, err = pop_vlan(skb); break; - case OVS_ACTION_ATTR_SET_DL_SRC: - err = make_writable(skb, ETH_HLEN); - if (likely(!err)) - memcpy(eth_hdr(skb)->h_source, nla_data(a), ETH_ALEN); - break; - - case OVS_ACTION_ATTR_SET_DL_DST: - err = make_writable(skb, ETH_HLEN); - if (likely(!err)) - memcpy(eth_hdr(skb)->h_dest, nla_data(a), ETH_ALEN); - break; - - case OVS_ACTION_ATTR_SET_NW_SRC: - case OVS_ACTION_ATTR_SET_NW_DST: - err = set_nw_addr(skb, a); - break; - - case OVS_ACTION_ATTR_SET_NW_TOS: - err = set_nw_tos(skb, nla_get_u8(a)); - break; - - case OVS_ACTION_ATTR_SET_TP_SRC: - case OVS_ACTION_ATTR_SET_TP_DST: - err = set_tp_port(skb, a); - break; - - case OVS_ACTION_ATTR_SET_PRIORITY: - skb->priority = nla_get_u32(a); - break; - - case OVS_ACTION_ATTR_POP_PRIORITY: - skb->priority = priority; + case OVS_ACTION_ATTR_SET: + err = execute_set_action(skb, nla_data(a), tun_key); break; case OVS_ACTION_ATTR_SAMPLE: - err = sample(dp, skb, a); + err = sample(dp, skb, a, tun_key); break; - } + if (unlikely(err)) { kfree_skb(skb); return err; @@ -414,17 +565,18 @@ static int loop_suppress(struct datapath *dp, struct sw_flow_actions *actions) { if (net_ratelimit()) pr_warn("%s: flow looped %d times, dropping\n", - dp_name(dp), MAX_LOOPS); + ovs_dp_name(dp), MAX_LOOPS); actions->actions_len = 0; return -ELOOP; } /* Execute a list of actions against 'skb'. */ -int execute_actions(struct datapath *dp, struct sk_buff *skb) +int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb) { struct sw_flow_actions *acts = rcu_dereference(OVS_CB(skb)->flow->sf_acts); struct loop_counter *loop; int error; + struct ovs_key_ipv4_tunnel tun_key; /* Check whether we've looped too much. */ loop = &__get_cpu_var(loop_counters); @@ -436,9 +588,9 @@ int execute_actions(struct datapath *dp, struct sk_buff *skb) goto out_loop; } - OVS_CB(skb)->tun_id = 0; + OVS_CB(skb)->tun_key = NULL; error = do_execute_actions(dp, skb, acts->actions, - acts->actions_len, false); + acts->actions_len, &tun_key, false); /* Check whether sub-actions looped too much. */ if (unlikely(loop->looping))