#include "netdev.h"
#include "netdev-vport.h"
#include "netlink.h"
+#include "odp-execute.h"
#include "odp-util.h"
#include "ofp-print.h"
#include "ofpbuf.h"
int port_no; /* Index into dp_netdev's 'ports'. */
struct list node; /* Element in dp_netdev's 'port_list'. */
struct netdev *netdev;
+ struct netdev_saved_flags *sf;
+ struct netdev_rx *rx;
char *type; /* Port type as requested by user. */
};
bool create, struct dpif **);
static int dp_netdev_output_userspace(struct dp_netdev *, const struct ofpbuf *,
int queue_no, const struct flow *,
- uint64_t arg);
+ const struct nlattr *userdata);
static void dp_netdev_execute_actions(struct dp_netdev *,
struct ofpbuf *, struct flow *,
const struct nlattr *actions,
size_t actions_len);
+static void dp_netdev_port_input(struct dp_netdev *dp,
+ struct dp_netdev_port *port,
+ struct ofpbuf *packet, uint32_t skb_priority,
+ uint32_t skb_mark, const struct flow_tnl *tnl);
static struct dpif_netdev *
dpif_netdev_cast(const struct dpif *dpif)
do_add_port(struct dp_netdev *dp, const char *devname, const char *type,
uint32_t port_no)
{
+ struct netdev_saved_flags *sf;
struct dp_netdev_port *port;
struct netdev *netdev;
+ struct netdev_rx *rx;
const char *open_type;
int mtu;
int error;
/* XXX reject loopback devices */
/* XXX reject non-Ethernet devices */
- error = netdev_listen(netdev);
+ error = netdev_rx_open(netdev, &rx);
if (error
&& !(error == EOPNOTSUPP && dpif_netdev_class_is_dummy(dp->class))) {
VLOG_ERR("%s: cannot receive packets on this network device (%s)",
return error;
}
- error = netdev_turn_flags_on(netdev, NETDEV_PROMISC, false);
+ error = netdev_turn_flags_on(netdev, NETDEV_PROMISC, &sf);
if (error) {
+ netdev_rx_close(rx);
netdev_close(netdev);
return error;
}
port = xmalloc(sizeof *port);
port->port_no = port_no;
port->netdev = netdev;
+ port->sf = sf;
+ port->rx = rx;
port->type = xstrdup(type);
error = netdev_get_mtu(netdev, &mtu);
- if (!error) {
+ if (!error && mtu > max_mtu) {
max_mtu = mtu;
}
uint32_t *port_nop)
{
struct dp_netdev *dp = get_dp_netdev(dpif);
+ char namebuf[NETDEV_VPORT_NAME_BUFSIZE];
+ const char *dpif_port;
int port_no;
+ dpif_port = netdev_vport_get_dpif_port(netdev, namebuf, sizeof namebuf);
if (*port_nop != UINT32_MAX) {
if (*port_nop >= MAX_PORTS) {
return EFBIG;
}
port_no = *port_nop;
} else {
- port_no = choose_port(dp, netdev_vport_get_dpif_port(netdev));
+ port_no = choose_port(dp, dpif_port);
}
if (port_no >= 0) {
*port_nop = port_no;
- return do_add_port(dp, netdev_vport_get_dpif_port(netdev),
- netdev_get_type(netdev), port_no);
+ return do_add_port(dp, dpif_port, netdev_get_type(netdev), port_no);
}
return EFBIG;
}
struct dp_netdev_port *port;
LIST_FOR_EACH (port, node, &dp->port_list) {
- if (!strcmp(netdev_vport_get_dpif_port(port->netdev), devname)) {
+ if (!strcmp(netdev_get_name(port->netdev), devname)) {
*portp = port;
return 0;
}
do_del_port(struct dp_netdev *dp, uint32_t port_no)
{
struct dp_netdev_port *port;
- char *name;
int error;
error = get_port_by_number(dp, port_no, &port);
dp->ports[port->port_no] = NULL;
dp->serial++;
- name = xstrdup(netdev_vport_get_dpif_port(port->netdev));
netdev_close(port->netdev);
+ netdev_restore_flags(port->sf);
+ netdev_rx_close(port->rx);
free(port->type);
-
- free(name);
free(port);
return 0;
answer_port_query(const struct dp_netdev_port *port,
struct dpif_port *dpif_port)
{
- dpif_port->name = xstrdup(netdev_vport_get_dpif_port(port->netdev));
+ dpif_port->name = xstrdup(netdev_get_name(port->netdev));
dpif_port->type = xstrdup(port->type);
dpif_port->port_no = port->port_no;
}
struct dp_netdev_port *port = dp->ports[port_no];
if (port) {
free(state->name);
- state->name = xstrdup(netdev_vport_get_dpif_port(port->netdev));
+ state->name = xstrdup(netdev_get_name(port->netdev));
dpif_port->name = state->name;
dpif_port->type = port->type;
dpif_port->port_no = port->port_no;
static void
dp_netdev_port_input(struct dp_netdev *dp, struct dp_netdev_port *port,
- struct ofpbuf *packet)
+ struct ofpbuf *packet, uint32_t skb_priority,
+ uint32_t skb_mark, const struct flow_tnl *tnl)
{
struct dp_netdev_flow *flow;
struct flow key;
if (packet->size < ETH_HEADER_LEN) {
return;
}
- flow_extract(packet, 0, 0, NULL, port->port_no, &key);
+ flow_extract(packet, skb_priority, skb_mark, tnl, port->port_no, &key);
flow = dp_netdev_lookup_flow(dp, &key);
if (flow) {
dp_netdev_flow_used(flow, packet);
dp->n_hit++;
} else {
dp->n_missed++;
- dp_netdev_output_userspace(dp, packet, DPIF_UC_MISS, &key, 0);
+ dp_netdev_output_userspace(dp, packet, DPIF_UC_MISS, &key, NULL);
}
}
ofpbuf_clear(&packet);
ofpbuf_reserve(&packet, DP_NETDEV_HEADROOM);
- error = netdev_recv(port->netdev, &packet);
+ error = port->rx ? netdev_rx_recv(port->rx, &packet) : EOPNOTSUPP;
if (!error) {
- dp_netdev_port_input(dp, port, &packet);
+ dp_netdev_port_input(dp, port, &packet, 0, 0, NULL);
} else if (error != EAGAIN && error != EOPNOTSUPP) {
static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5);
+
VLOG_ERR_RL(&rl, "error receiving data from %s: %s",
- netdev_vport_get_dpif_port(port->netdev),
- strerror(error));
+ netdev_get_name(port->netdev), strerror(error));
}
}
ofpbuf_uninit(&packet);
struct dp_netdev_port *port;
LIST_FOR_EACH (port, node, &dp->port_list) {
- netdev_recv_wait(port->netdev);
+ if (port->rx) {
+ netdev_rx_wait(port->rx);
+ }
}
}
static void
-dp_netdev_set_dl(struct ofpbuf *packet, const struct ovs_key_ethernet *eth_key)
-{
- struct eth_header *eh = packet->l2;
-
- memcpy(eh->eth_src, eth_key->eth_src, sizeof eh->eth_src);
- memcpy(eh->eth_dst, eth_key->eth_dst, sizeof eh->eth_dst);
-}
-
-static void
-dp_netdev_output_port(struct dp_netdev *dp, struct ofpbuf *packet,
- uint32_t out_port)
+dp_netdev_output_port(void *dp_, struct ofpbuf *packet, uint32_t out_port)
{
+ struct dp_netdev *dp = dp_;
struct dp_netdev_port *p = dp->ports[out_port];
if (p) {
netdev_send(p->netdev, packet);
static int
dp_netdev_output_userspace(struct dp_netdev *dp, const struct ofpbuf *packet,
- int queue_no, const struct flow *flow, uint64_t arg)
+ int queue_no, const struct flow *flow,
+ const struct nlattr *userdata)
{
struct dp_netdev_queue *q = &dp->queues[queue_no];
- struct dp_netdev_upcall *u;
- struct dpif_upcall *upcall;
- struct ofpbuf *buf;
- size_t key_len;
-
- if (q->head - q->tail >= MAX_QUEUE_LEN) {
- dp->n_lost++;
- return ENOBUFS;
- }
-
- u = &q->upcalls[q->head++ & QUEUE_MASK];
-
- buf = &u->buf;
- ofpbuf_init(buf, ODPUTIL_FLOW_KEY_BYTES + 2 + packet->size);
- odp_flow_key_from_flow(buf, flow, flow->in_port);
- key_len = buf->size;
- ofpbuf_pull(buf, key_len);
- ofpbuf_reserve(buf, 2);
- ofpbuf_put(buf, packet->data, packet->size);
-
- upcall = &u->upcall;
- upcall->type = queue_no;
- upcall->packet = buf;
- upcall->key = buf->base;
- upcall->key_len = key_len;
- upcall->userdata = arg;
-
- return 0;
-}
-
-static void
-dp_netdev_sample(struct dp_netdev *dp,
- struct ofpbuf *packet, struct flow *key,
- const struct nlattr *action)
-{
- const struct nlattr *subactions = NULL;
- const struct nlattr *a;
- size_t left;
+ if (q->head - q->tail < MAX_QUEUE_LEN) {
+ struct dp_netdev_upcall *u = &q->upcalls[q->head++ & QUEUE_MASK];
+ struct dpif_upcall *upcall = &u->upcall;
+ struct ofpbuf *buf = &u->buf;
+ size_t buf_size;
+
+ upcall->type = queue_no;
+
+ /* Allocate buffer big enough for everything. */
+ buf_size = ODPUTIL_FLOW_KEY_BYTES + 2 + packet->size;
+ if (userdata) {
+ buf_size += NLA_ALIGN(userdata->nla_len);
+ }
+ ofpbuf_init(buf, buf_size);
- NL_NESTED_FOR_EACH_UNSAFE (a, left, action) {
- int type = nl_attr_type(a);
+ /* Put ODP flow. */
+ odp_flow_key_from_flow(buf, flow, flow->in_port);
+ upcall->key = buf->data;
+ upcall->key_len = buf->size;
- switch ((enum ovs_sample_attr) type) {
- case OVS_SAMPLE_ATTR_PROBABILITY:
- if (random_uint32() >= nl_attr_get_u32(a)) {
- return;
- }
- break;
+ /* Put userdata. */
+ if (userdata) {
+ upcall->userdata = ofpbuf_put(buf, userdata,
+ NLA_ALIGN(userdata->nla_len));
+ }
- case OVS_SAMPLE_ATTR_ACTIONS:
- subactions = a;
- break;
+ /* Put packet.
+ *
+ * We adjust 'data' and 'size' in 'buf' so that only the packet itself
+ * is visible in 'upcall->packet'. The ODP flow and (if present)
+ * userdata become part of the headroom. */
+ ofpbuf_put_zeros(buf, 2);
+ buf->data = ofpbuf_put(buf, packet->data, packet->size);
+ buf->size = packet->size;
+ upcall->packet = buf;
- case OVS_SAMPLE_ATTR_UNSPEC:
- case __OVS_SAMPLE_ATTR_MAX:
- default:
- NOT_REACHED();
- }
+ return 0;
+ } else {
+ dp->n_lost++;
+ return ENOBUFS;
}
-
- dp_netdev_execute_actions(dp, packet, key, nl_attr_get(subactions),
- nl_attr_get_size(subactions));
}
static void
-dp_netdev_action_userspace(struct dp_netdev *dp,
- struct ofpbuf *packet, struct flow *key,
- const struct nlattr *a)
+dp_netdev_action_userspace(void *dp, struct ofpbuf *packet,
+ const struct flow *key,
+ const struct nlattr *userdata)
{
- const struct nlattr *userdata_attr;
- uint64_t userdata;
-
- userdata_attr = nl_attr_find_nested(a, OVS_USERSPACE_ATTR_USERDATA);
- userdata = userdata_attr ? nl_attr_get_u64(userdata_attr) : 0;
dp_netdev_output_userspace(dp, packet, DPIF_UC_ACTION, key, userdata);
}
-static void
-execute_set_action(struct ofpbuf *packet, const struct nlattr *a)
-{
- enum ovs_key_attr type = nl_attr_type(a);
- const struct ovs_key_ipv4 *ipv4_key;
- const struct ovs_key_ipv6 *ipv6_key;
- const struct ovs_key_tcp *tcp_key;
- const struct ovs_key_udp *udp_key;
-
- switch (type) {
- case OVS_KEY_ATTR_TUN_ID:
- case OVS_KEY_ATTR_PRIORITY:
- case OVS_KEY_ATTR_SKB_MARK:
- case OVS_KEY_ATTR_TUNNEL:
- /* not implemented */
- break;
-
- case OVS_KEY_ATTR_ETHERNET:
- dp_netdev_set_dl(packet,
- nl_attr_get_unspec(a, sizeof(struct ovs_key_ethernet)));
- break;
-
- case OVS_KEY_ATTR_IPV4:
- ipv4_key = nl_attr_get_unspec(a, sizeof(struct ovs_key_ipv4));
- packet_set_ipv4(packet, ipv4_key->ipv4_src, ipv4_key->ipv4_dst,
- ipv4_key->ipv4_tos, ipv4_key->ipv4_ttl);
- break;
-
- case OVS_KEY_ATTR_IPV6:
- ipv6_key = nl_attr_get_unspec(a, sizeof(struct ovs_key_ipv6));
- packet_set_ipv6(packet, ipv6_key->ipv6_proto, ipv6_key->ipv6_src,
- ipv6_key->ipv6_dst, ipv6_key->ipv6_tclass,
- ipv6_key->ipv6_label, ipv6_key->ipv6_hlimit);
- break;
-
- case OVS_KEY_ATTR_TCP:
- tcp_key = nl_attr_get_unspec(a, sizeof(struct ovs_key_tcp));
- packet_set_tcp_port(packet, tcp_key->tcp_src, tcp_key->tcp_dst);
- break;
-
- case OVS_KEY_ATTR_UDP:
- udp_key = nl_attr_get_unspec(a, sizeof(struct ovs_key_udp));
- packet_set_udp_port(packet, udp_key->udp_src, udp_key->udp_dst);
- break;
-
- case OVS_KEY_ATTR_UNSPEC:
- case OVS_KEY_ATTR_ENCAP:
- case OVS_KEY_ATTR_ETHERTYPE:
- case OVS_KEY_ATTR_IN_PORT:
- case OVS_KEY_ATTR_VLAN:
- case OVS_KEY_ATTR_ICMP:
- case OVS_KEY_ATTR_ICMPV6:
- case OVS_KEY_ATTR_ARP:
- case OVS_KEY_ATTR_ND:
- case __OVS_KEY_ATTR_MAX:
- default:
- NOT_REACHED();
- }
-}
-
static void
dp_netdev_execute_actions(struct dp_netdev *dp,
struct ofpbuf *packet, struct flow *key,
const struct nlattr *actions,
size_t actions_len)
{
- const struct nlattr *a;
- unsigned int left;
-
- NL_ATTR_FOR_EACH_UNSAFE (a, left, actions, actions_len) {
- int type = nl_attr_type(a);
-
- switch ((enum ovs_action_attr) type) {
- case OVS_ACTION_ATTR_OUTPUT:
- dp_netdev_output_port(dp, packet, nl_attr_get_u32(a));
- break;
-
- case OVS_ACTION_ATTR_USERSPACE:
- dp_netdev_action_userspace(dp, packet, key, a);
- break;
-
- case OVS_ACTION_ATTR_PUSH_VLAN: {
- const struct ovs_action_push_vlan *vlan = nl_attr_get(a);
- eth_push_vlan(packet, vlan->vlan_tci);
- break;
- }
-
- case OVS_ACTION_ATTR_POP_VLAN:
- eth_pop_vlan(packet);
- break;
-
- case OVS_ACTION_ATTR_SET:
- execute_set_action(packet, nl_attr_get(a));
- break;
-
- case OVS_ACTION_ATTR_SAMPLE:
- dp_netdev_sample(dp, packet, key, a);
- break;
-
- case OVS_ACTION_ATTR_UNSPEC:
- case __OVS_ACTION_ATTR_MAX:
- NOT_REACHED();
- }
- }
+ odp_execute_actions(dp, packet, key, actions, actions_len,
+ dp_netdev_output_port, dp_netdev_action_userspace);
}
const struct dpif_class dpif_netdev_class = {