2
0
mirror of https://github.com/openvswitch/ovs synced 2025-08-22 09:58:01 +00:00
ovs/lib/netdev-dummy.c
David Marchand e99ce7d5df flow: Fix checksum offloads with simple match.
Packets with L4 partial status for a simple match flow would not get L4
checksums offloads applied.

This was not caught in unit tests, because packets from netdev-dummy
(calling miniflow_extract) would get Tx flags set early, before
parse_tcp_flags() got called during packet processing.

Signed-off-by: David Marchand <david.marchand@redhat.com>
Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
2025-05-30 18:00:56 +02:00

2352 lines
68 KiB
C
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

/*
* Copyright (c) 2010, 2011, 2012, 2013, 2015, 2016, 2017 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <config.h>
#include "dummy.h"
#include <errno.h>
#include <unistd.h>
#include "dp-packet.h"
#include "dpif-netdev.h"
#include "flow.h"
#include "netdev-offload-provider.h"
#include "netdev-provider.h"
#include "netdev-vport.h"
#include "odp-util.h"
#include "openvswitch/dynamic-string.h"
#include "openvswitch/list.h"
#include "openvswitch/match.h"
#include "openvswitch/ofp-print.h"
#include "openvswitch/ofpbuf.h"
#include "openvswitch/vlog.h"
#include "ovs-atomic.h"
#include "packets.h"
#include "pcap-file.h"
#include "openvswitch/poll-loop.h"
#include "openvswitch/shash.h"
#include "ovs-router.h"
#include "sset.h"
#include "stream.h"
#include "unaligned.h"
#include "timeval.h"
#include "unixctl.h"
#include "userspace-tso.h"
#include "reconnect.h"
VLOG_DEFINE_THIS_MODULE(netdev_dummy);
#define C_STATS_SIZE 2
struct reconnect;
struct dummy_packet_stream {
struct stream *stream;
struct ovs_list txq;
struct dp_packet rxbuf;
};
enum dummy_packet_conn_type {
NONE, /* No connection is configured. */
PASSIVE, /* Listener. */
ACTIVE /* Connect to listener. */
};
enum dummy_netdev_conn_state {
CONN_STATE_CONNECTED, /* Listener connected. */
CONN_STATE_NOT_CONNECTED, /* Listener not connected. */
CONN_STATE_UNKNOWN, /* No relavent information. */
};
struct dummy_packet_pconn {
struct pstream *pstream;
struct dummy_packet_stream **streams;
size_t n_streams;
};
struct dummy_packet_rconn {
struct dummy_packet_stream *rstream;
struct reconnect *reconnect;
};
struct dummy_packet_conn {
enum dummy_packet_conn_type type;
union {
struct dummy_packet_pconn pconn;
struct dummy_packet_rconn rconn;
};
};
struct pkt_list_node {
struct dp_packet *pkt;
struct ovs_list list_node;
};
struct offloaded_flow {
struct hmap_node node;
ovs_u128 ufid;
struct match match;
uint32_t mark;
};
struct netdev_dummy_q_stats {
uint64_t bytes;
uint64_t packets;
};
/* Protects 'dummy_list'. */
static struct ovs_mutex dummy_list_mutex = OVS_MUTEX_INITIALIZER;
/* Contains all 'struct dummy_dev's. */
static struct ovs_list dummy_list OVS_GUARDED_BY(dummy_list_mutex)
= OVS_LIST_INITIALIZER(&dummy_list);
struct netdev_dummy {
struct netdev up;
/* In dummy_list. */
struct ovs_list list_node OVS_GUARDED_BY(dummy_list_mutex);
/* Protects all members below. */
struct ovs_mutex mutex OVS_ACQ_AFTER(dummy_list_mutex);
struct eth_addr hwaddr OVS_GUARDED;
int mtu OVS_GUARDED;
struct netdev_stats stats OVS_GUARDED;
struct netdev_custom_counter custom_stats[C_STATS_SIZE] OVS_GUARDED;
struct netdev_dummy_q_stats *rxq_stats OVS_GUARDED;
struct netdev_dummy_q_stats *txq_stats OVS_GUARDED;
enum netdev_flags flags OVS_GUARDED;
int ifindex OVS_GUARDED;
int numa_id OVS_GUARDED;
struct dummy_packet_conn conn OVS_GUARDED;
struct pcap_file *tx_pcap, *rxq_pcap OVS_GUARDED;
struct ovs_list addrs OVS_GUARDED;
struct ovs_list rxes OVS_GUARDED; /* List of child "netdev_rxq_dummy"s. */
struct hmap offloaded_flows OVS_GUARDED;
/* The following properties are for dummy-pmd and they cannot be changed
* when a device is running, so we remember the request and update them
* next time netdev_dummy_reconfigure() is called. */
int requested_n_txq OVS_GUARDED;
int requested_n_rxq OVS_GUARDED;
int requested_numa_id OVS_GUARDED;
/* Force IP Rx csum good. */
bool ol_ip_rx_csum_set_good OVS_GUARDED;
/* Force IP Rx csum bad. */
bool ol_ip_rx_csum_set_bad OVS_GUARDED;
/* Announce netdev IP Tx csum offload. */
bool ol_ip_tx_csum OVS_GUARDED;
/* Disable IP Tx csum offload. */
bool ol_ip_tx_csum_disabled OVS_GUARDED;
/* Force L4 Rx csum good. */
bool ol_l4_rx_csum_set_good OVS_GUARDED;
/* Force L4 Rx csum bad. */
bool ol_l4_rx_csum_set_bad OVS_GUARDED;
/* Force L4 Rx csum partial. */
bool ol_l4_rx_csum_set_partial OVS_GUARDED;
/* Announce netdev L4 Tx csum offload. */
bool ol_l4_tx_csum OVS_GUARDED;
/* Disable L4 Tx csum offload. */
bool ol_l4_tx_csum_disabled OVS_GUARDED;
/* Set the segment size for netdev TSO support. */
int ol_tso_segsz OVS_GUARDED;
};
/* Max 'recv_queue_len' in struct netdev_dummy. */
#define NETDEV_DUMMY_MAX_QUEUE 100
struct netdev_rxq_dummy {
struct netdev_rxq up;
struct ovs_list node; /* In netdev_dummy's "rxes" list. */
struct ovs_list recv_queue;
int recv_queue_len; /* ovs_list_size(&recv_queue). */
struct seq *seq; /* Reports newly queued packets. */
};
struct netdev_addr_dummy {
struct in6_addr address;
struct in6_addr netmask;
struct ovs_list node; /* In netdev_dummy's "addrs" list. */
};
static unixctl_cb_func netdev_dummy_set_admin_state;
static int netdev_dummy_construct(struct netdev *);
static void netdev_dummy_queue_packet(struct netdev_dummy *,
struct dp_packet *, struct flow *, int);
static void dummy_packet_stream_close(struct dummy_packet_stream *);
static void pkt_list_delete(struct ovs_list *);
static void addr_list_delete(struct ovs_list *);
static bool
is_dummy_class(const struct netdev_class *class)
{
return class->construct == netdev_dummy_construct;
}
static struct netdev_dummy *
netdev_dummy_cast(const struct netdev *netdev)
{
ovs_assert(is_dummy_class(netdev_get_class(netdev)));
return CONTAINER_OF(netdev, struct netdev_dummy, up);
}
static struct netdev_rxq_dummy *
netdev_rxq_dummy_cast(const struct netdev_rxq *rx)
{
ovs_assert(is_dummy_class(netdev_get_class(rx->netdev)));
return CONTAINER_OF(rx, struct netdev_rxq_dummy, up);
}
static void
dummy_packet_stream_init(struct dummy_packet_stream *s, struct stream *stream)
{
int rxbuf_size = stream ? 2048 : 0;
s->stream = stream;
dp_packet_init(&s->rxbuf, rxbuf_size);
ovs_list_init(&s->txq);
}
static struct dummy_packet_stream *
dummy_packet_stream_create(struct stream *stream)
{
struct dummy_packet_stream *s;
s = xzalloc_cacheline(sizeof *s);
dummy_packet_stream_init(s, stream);
return s;
}
static void
dummy_packet_stream_wait(struct dummy_packet_stream *s)
{
stream_run_wait(s->stream);
if (!ovs_list_is_empty(&s->txq)) {
stream_send_wait(s->stream);
}
stream_recv_wait(s->stream);
}
static void
dummy_packet_stream_send(struct dummy_packet_stream *s, const void *buffer, size_t size)
{
if (ovs_list_size(&s->txq) < NETDEV_DUMMY_MAX_QUEUE) {
struct dp_packet *b;
struct pkt_list_node *node;
b = dp_packet_clone_data_with_headroom(buffer, size, 2);
put_unaligned_be16(dp_packet_push_uninit(b, 2), htons(size));
node = xmalloc(sizeof *node);
node->pkt = b;
ovs_list_push_back(&s->txq, &node->list_node);
}
}
static int
dummy_packet_stream_run(struct netdev_dummy *dev, struct dummy_packet_stream *s)
{
int error = 0;
size_t n = 0;
stream_run(s->stream);
if (!ovs_list_is_empty(&s->txq)) {
struct pkt_list_node *txbuf_node;
struct dp_packet *txbuf;
int retval;
ASSIGN_CONTAINER(txbuf_node, ovs_list_front(&s->txq), list_node);
txbuf = txbuf_node->pkt;
retval = stream_send(s->stream, dp_packet_data(txbuf), dp_packet_size(txbuf));
if (retval > 0) {
dp_packet_pull(txbuf, retval);
if (!dp_packet_size(txbuf)) {
ovs_list_remove(&txbuf_node->list_node);
free(txbuf_node);
dp_packet_delete(txbuf);
}
} else if (retval != -EAGAIN) {
error = -retval;
}
}
if (!error) {
if (dp_packet_size(&s->rxbuf) < 2) {
n = 2 - dp_packet_size(&s->rxbuf);
} else {
uint16_t frame_len;
frame_len = ntohs(get_unaligned_be16(dp_packet_data(&s->rxbuf)));
if (frame_len < ETH_HEADER_LEN) {
error = EPROTO;
n = 0;
} else {
n = (2 + frame_len) - dp_packet_size(&s->rxbuf);
}
}
}
if (!error) {
int retval;
dp_packet_prealloc_tailroom(&s->rxbuf, n);
retval = stream_recv(s->stream, dp_packet_tail(&s->rxbuf), n);
if (retval > 0) {
dp_packet_set_size(&s->rxbuf, dp_packet_size(&s->rxbuf) + retval);
if (retval == n && dp_packet_size(&s->rxbuf) > 2) {
dp_packet_pull(&s->rxbuf, 2);
netdev_dummy_queue_packet(dev,
dp_packet_clone(&s->rxbuf), NULL, 0);
dp_packet_clear(&s->rxbuf);
}
} else if (retval != -EAGAIN) {
error = (retval < 0 ? -retval
: dp_packet_size(&s->rxbuf) ? EPROTO
: EOF);
}
}
return error;
}
static void
dummy_packet_stream_close(struct dummy_packet_stream *s)
{
stream_close(s->stream);
dp_packet_uninit(&s->rxbuf);
pkt_list_delete(&s->txq);
}
static void
dummy_packet_conn_init(struct dummy_packet_conn *conn)
{
memset(conn, 0, sizeof *conn);
conn->type = NONE;
}
static void
dummy_packet_conn_get_config(struct dummy_packet_conn *conn, struct smap *args)
{
switch (conn->type) {
case PASSIVE:
smap_add(args, "pstream", pstream_get_name(conn->pconn.pstream));
break;
case ACTIVE:
smap_add(args, "stream", stream_get_name(conn->rconn.rstream->stream));
break;
case NONE:
default:
break;
}
}
static void
dummy_packet_conn_close(struct dummy_packet_conn *conn)
{
int i;
struct dummy_packet_pconn *pconn = &conn->pconn;
struct dummy_packet_rconn *rconn = &conn->rconn;
switch (conn->type) {
case PASSIVE:
pstream_close(pconn->pstream);
for (i = 0; i < pconn->n_streams; i++) {
dummy_packet_stream_close(pconn->streams[i]);
free_cacheline(pconn->streams[i]);
}
free(pconn->streams);
pconn->pstream = NULL;
pconn->streams = NULL;
break;
case ACTIVE:
dummy_packet_stream_close(rconn->rstream);
free_cacheline(rconn->rstream);
rconn->rstream = NULL;
reconnect_destroy(rconn->reconnect);
rconn->reconnect = NULL;
break;
case NONE:
default:
break;
}
conn->type = NONE;
memset(conn, 0, sizeof *conn);
}
static void
dummy_packet_conn_set_config(struct dummy_packet_conn *conn,
const struct smap *args)
{
const char *pstream = smap_get(args, "pstream");
const char *stream = smap_get(args, "stream");
if (pstream && stream) {
VLOG_WARN("Open failed: both %s and %s are configured",
pstream, stream);
return;
}
switch (conn->type) {
case PASSIVE:
if (pstream &&
!strcmp(pstream_get_name(conn->pconn.pstream), pstream)) {
return;
}
dummy_packet_conn_close(conn);
break;
case ACTIVE:
if (stream &&
!strcmp(stream_get_name(conn->rconn.rstream->stream), stream)) {
return;
}
dummy_packet_conn_close(conn);
break;
case NONE:
default:
break;
}
if (pstream) {
int error;
error = pstream_open(pstream, &conn->pconn.pstream, DSCP_DEFAULT);
if (error) {
VLOG_WARN("%s: open failed (%s)", pstream, ovs_strerror(error));
} else {
conn->type = PASSIVE;
}
}
if (stream) {
int error;
struct stream *active_stream;
struct reconnect *reconnect;
reconnect = reconnect_create(time_msec());
reconnect_set_name(reconnect, stream);
reconnect_set_passive(reconnect, false, time_msec());
reconnect_enable(reconnect, time_msec());
reconnect_set_backoff(reconnect, 100, INT_MAX);
reconnect_set_probe_interval(reconnect, 0);
conn->rconn.reconnect = reconnect;
conn->type = ACTIVE;
error = stream_open(stream, &active_stream, DSCP_DEFAULT);
conn->rconn.rstream = dummy_packet_stream_create(active_stream);
switch (error) {
case 0:
reconnect_connected(reconnect, time_msec());
break;
case EAGAIN:
reconnect_connecting(reconnect, time_msec());
break;
default:
reconnect_connect_failed(reconnect, time_msec(), error);
stream_close(active_stream);
conn->rconn.rstream->stream = NULL;
break;
}
}
}
static void
dummy_pconn_run(struct netdev_dummy *dev)
OVS_REQUIRES(dev->mutex)
{
struct stream *new_stream;
struct dummy_packet_pconn *pconn = &dev->conn.pconn;
int error;
size_t i;
error = pstream_accept(pconn->pstream, &new_stream);
if (!error) {
struct dummy_packet_stream *s;
pconn->streams = xrealloc(pconn->streams,
((pconn->n_streams + 1)
* sizeof s));
s = xmalloc_cacheline(sizeof *s);
pconn->streams[pconn->n_streams++] = s;
dummy_packet_stream_init(s, new_stream);
} else if (error != EAGAIN) {
VLOG_WARN("%s: accept failed (%s)",
pstream_get_name(pconn->pstream), ovs_strerror(error));
pstream_close(pconn->pstream);
pconn->pstream = NULL;
dev->conn.type = NONE;
}
for (i = 0; i < pconn->n_streams; ) {
struct dummy_packet_stream *s = pconn->streams[i];
error = dummy_packet_stream_run(dev, s);
if (error) {
VLOG_DBG("%s: closing connection (%s)",
stream_get_name(s->stream),
ovs_retval_to_string(error));
dummy_packet_stream_close(s);
free_cacheline(s);
pconn->streams[i] = pconn->streams[--pconn->n_streams];
} else {
i++;
}
}
}
static void
dummy_rconn_run(struct netdev_dummy *dev)
OVS_REQUIRES(dev->mutex)
{
struct dummy_packet_rconn *rconn = &dev->conn.rconn;
switch (reconnect_run(rconn->reconnect, time_msec())) {
case RECONNECT_CONNECT:
{
int error;
if (rconn->rstream->stream) {
error = stream_connect(rconn->rstream->stream);
} else {
error = stream_open(reconnect_get_name(rconn->reconnect),
&rconn->rstream->stream, DSCP_DEFAULT);
}
switch (error) {
case 0:
reconnect_connected(rconn->reconnect, time_msec());
break;
case EAGAIN:
reconnect_connecting(rconn->reconnect, time_msec());
break;
default:
reconnect_connect_failed(rconn->reconnect, time_msec(), error);
stream_close(rconn->rstream->stream);
rconn->rstream->stream = NULL;
break;
}
}
break;
case RECONNECT_DISCONNECT:
case RECONNECT_PROBE:
default:
break;
}
if (reconnect_is_connected(rconn->reconnect)) {
int err;
err = dummy_packet_stream_run(dev, rconn->rstream);
if (err) {
reconnect_disconnected(rconn->reconnect, time_msec(), err);
stream_close(rconn->rstream->stream);
rconn->rstream->stream = NULL;
}
}
}
static void
dummy_packet_conn_run(struct netdev_dummy *dev)
OVS_REQUIRES(dev->mutex)
{
switch (dev->conn.type) {
case PASSIVE:
dummy_pconn_run(dev);
break;
case ACTIVE:
dummy_rconn_run(dev);
break;
case NONE:
default:
break;
}
}
static void
dummy_packet_conn_wait(struct dummy_packet_conn *conn)
{
int i;
switch (conn->type) {
case PASSIVE:
pstream_wait(conn->pconn.pstream);
for (i = 0; i < conn->pconn.n_streams; i++) {
struct dummy_packet_stream *s = conn->pconn.streams[i];
dummy_packet_stream_wait(s);
}
break;
case ACTIVE:
if (reconnect_is_connected(conn->rconn.reconnect)) {
dummy_packet_stream_wait(conn->rconn.rstream);
}
break;
case NONE:
default:
break;
}
}
static void
dummy_packet_conn_send(struct dummy_packet_conn *conn,
const void *buffer, size_t size)
{
int i;
switch (conn->type) {
case PASSIVE:
for (i = 0; i < conn->pconn.n_streams; i++) {
struct dummy_packet_stream *s = conn->pconn.streams[i];
dummy_packet_stream_send(s, buffer, size);
pstream_wait(conn->pconn.pstream);
}
break;
case ACTIVE:
if (reconnect_is_connected(conn->rconn.reconnect)) {
dummy_packet_stream_send(conn->rconn.rstream, buffer, size);
dummy_packet_stream_wait(conn->rconn.rstream);
}
break;
case NONE:
default:
break;
}
}
static enum dummy_netdev_conn_state
dummy_netdev_get_conn_state(struct dummy_packet_conn *conn)
{
enum dummy_netdev_conn_state state;
if (conn->type == ACTIVE) {
if (reconnect_is_connected(conn->rconn.reconnect)) {
state = CONN_STATE_CONNECTED;
} else {
state = CONN_STATE_NOT_CONNECTED;
}
} else {
state = CONN_STATE_UNKNOWN;
}
return state;
}
static void
netdev_dummy_run(const struct netdev_class *netdev_class)
{
struct netdev_dummy *dev;
ovs_mutex_lock(&dummy_list_mutex);
LIST_FOR_EACH (dev, list_node, &dummy_list) {
if (netdev_get_class(&dev->up) != netdev_class) {
continue;
}
ovs_mutex_lock(&dev->mutex);
dummy_packet_conn_run(dev);
ovs_mutex_unlock(&dev->mutex);
}
ovs_mutex_unlock(&dummy_list_mutex);
}
static void
netdev_dummy_wait(const struct netdev_class *netdev_class)
{
struct netdev_dummy *dev;
ovs_mutex_lock(&dummy_list_mutex);
LIST_FOR_EACH (dev, list_node, &dummy_list) {
if (netdev_get_class(&dev->up) != netdev_class) {
continue;
}
ovs_mutex_lock(&dev->mutex);
dummy_packet_conn_wait(&dev->conn);
ovs_mutex_unlock(&dev->mutex);
}
ovs_mutex_unlock(&dummy_list_mutex);
}
static struct netdev *
netdev_dummy_alloc(void)
{
struct netdev_dummy *netdev = xzalloc(sizeof *netdev);
return &netdev->up;
}
static int
netdev_dummy_construct(struct netdev *netdev_)
{
static atomic_count next_n = ATOMIC_COUNT_INIT(0xaa550000);
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
unsigned int n;
n = atomic_count_inc(&next_n);
ovs_mutex_init(&netdev->mutex);
ovs_mutex_lock(&netdev->mutex);
netdev->hwaddr.ea[0] = 0xaa;
netdev->hwaddr.ea[1] = 0x55;
netdev->hwaddr.ea[2] = n >> 24;
netdev->hwaddr.ea[3] = n >> 16;
netdev->hwaddr.ea[4] = n >> 8;
netdev->hwaddr.ea[5] = n;
netdev->mtu = 1500;
netdev->flags = NETDEV_UP;
netdev->ifindex = -EOPNOTSUPP;
netdev->requested_n_rxq = netdev_->n_rxq;
netdev->requested_n_txq = netdev_->n_txq;
netdev->numa_id = 0;
memset(&netdev->custom_stats, 0, sizeof(netdev->custom_stats));
ovs_strlcpy(netdev->custom_stats[0].name,
"rx_custom_packets_1", NETDEV_CUSTOM_STATS_NAME_SIZE);
ovs_strlcpy(netdev->custom_stats[1].name,
"rx_custom_packets_2", NETDEV_CUSTOM_STATS_NAME_SIZE);
netdev->rxq_stats = xcalloc(netdev->up.n_rxq, sizeof *netdev->rxq_stats);
netdev->txq_stats = xcalloc(netdev->up.n_rxq, sizeof *netdev->txq_stats);
dummy_packet_conn_init(&netdev->conn);
ovs_list_init(&netdev->rxes);
ovs_list_init(&netdev->addrs);
hmap_init(&netdev->offloaded_flows);
ovs_mutex_unlock(&netdev->mutex);
ovs_mutex_lock(&dummy_list_mutex);
ovs_list_push_back(&dummy_list, &netdev->list_node);
ovs_mutex_unlock(&dummy_list_mutex);
return 0;
}
static void
netdev_dummy_destruct(struct netdev *netdev_)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
struct offloaded_flow *off_flow;
ovs_mutex_lock(&dummy_list_mutex);
ovs_list_remove(&netdev->list_node);
ovs_mutex_unlock(&dummy_list_mutex);
ovs_mutex_lock(&netdev->mutex);
free(netdev->rxq_stats);
free(netdev->txq_stats);
if (netdev->rxq_pcap) {
ovs_pcap_close(netdev->rxq_pcap);
}
if (netdev->tx_pcap && netdev->tx_pcap != netdev->rxq_pcap) {
ovs_pcap_close(netdev->tx_pcap);
}
dummy_packet_conn_close(&netdev->conn);
netdev->conn.type = NONE;
HMAP_FOR_EACH_POP (off_flow, node, &netdev->offloaded_flows) {
free(off_flow);
}
hmap_destroy(&netdev->offloaded_flows);
addr_list_delete(&netdev->addrs);
ovs_mutex_unlock(&netdev->mutex);
ovs_mutex_destroy(&netdev->mutex);
}
static void
netdev_dummy_dealloc(struct netdev *netdev_)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
free(netdev);
}
static int
netdev_dummy_get_config(const struct netdev *dev, struct smap *args)
{
struct netdev_dummy *netdev = netdev_dummy_cast(dev);
ovs_mutex_lock(&netdev->mutex);
if (netdev->ifindex >= 0) {
smap_add_format(args, "ifindex", "%d", netdev->ifindex);
}
dummy_packet_conn_get_config(&netdev->conn, args);
/* pcap, rxq_pcap and tx_pcap cannot be recovered because filenames have
* been discarded after opening file descriptors */
if (netdev->ol_ip_rx_csum_set_good) {
smap_add_format(args, "ol_ip_rx_csum_set_good", "%s", "true");
}
if (netdev->ol_ip_rx_csum_set_bad) {
smap_add_format(args, "ol_ip_rx_csum_set_bad", "%s", "true");
}
if (netdev->ol_ip_tx_csum) {
smap_add_format(args, "ol_ip_tx_csum", "%s", "true");
if (netdev->ol_ip_tx_csum_disabled) {
smap_add_format(args, "ol_ip_tx_csum_disabled", "%s", "true");
}
}
if (netdev->ol_l4_rx_csum_set_good) {
smap_add_format(args, "ol_l4_rx_csum_set_good", "%s", "true");
}
if (netdev->ol_l4_rx_csum_set_bad) {
smap_add_format(args, "ol_l4_rx_csum_set_bad", "%s", "true");
}
if (netdev->ol_l4_rx_csum_set_partial) {
smap_add_format(args, "ol_l4_rx_csum_set_partial", "%s", "true");
}
if (netdev->ol_l4_tx_csum) {
smap_add_format(args, "ol_l4_tx_csum", "%s", "true");
if (netdev->ol_l4_tx_csum_disabled) {
smap_add_format(args, "ol_l4_tx_csum_disabled", "%s", "true");
}
}
if (netdev->ol_tso_segsz && userspace_tso_enabled()) {
smap_add_format(args, "ol_tso_segsz", "%d", netdev->ol_tso_segsz);
}
/* 'dummy-pmd' specific config. */
if (!netdev_is_pmd(dev)) {
goto exit;
}
smap_add_format(args, "n_rxq", "%d", netdev->requested_n_rxq);
smap_add_format(args, "n_txq", "%d", netdev->requested_n_txq);
smap_add_format(args, "numa_id", "%d", netdev->requested_numa_id);
exit:
ovs_mutex_unlock(&netdev->mutex);
return 0;
}
static int
netdev_dummy_get_addr_list(const struct netdev *netdev_, struct in6_addr **paddr,
struct in6_addr **pmask, int *n_addr)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
int cnt = 0, i = 0, err = 0;
struct in6_addr *addr, *mask;
struct netdev_addr_dummy *addr_dummy;
ovs_mutex_lock(&netdev->mutex);
cnt = ovs_list_size(&netdev->addrs);
if (!cnt) {
err = EADDRNOTAVAIL;
goto out;
}
addr = xmalloc(sizeof *addr * cnt);
mask = xmalloc(sizeof *mask * cnt);
LIST_FOR_EACH (addr_dummy, node, &netdev->addrs) {
memcpy(&addr[i], &addr_dummy->address, sizeof *addr);
memcpy(&mask[i], &addr_dummy->netmask, sizeof *mask);
i++;
}
if (paddr) {
*paddr = addr;
*pmask = mask;
*n_addr = cnt;
} else {
free(addr);
free(mask);
}
out:
ovs_mutex_unlock(&netdev->mutex);
return err;
}
static int
netdev_dummy_add_in4(struct netdev *netdev_, struct in_addr address,
struct in_addr netmask)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
struct netdev_addr_dummy *addr_dummy = xmalloc(sizeof *addr_dummy);
ovs_mutex_lock(&netdev->mutex);
in6_addr_set_mapped_ipv4(&addr_dummy->address, address.s_addr);
in6_addr_set_mapped_ipv4(&addr_dummy->netmask, netmask.s_addr);
ovs_list_push_back(&netdev->addrs, &addr_dummy->node);
netdev_change_seq_changed(netdev_);
ovs_mutex_unlock(&netdev->mutex);
return 0;
}
static int
netdev_dummy_add_in6(struct netdev *netdev_, struct in6_addr *in6,
struct in6_addr *mask)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
struct netdev_addr_dummy *addr_dummy = xmalloc(sizeof *addr_dummy);
ovs_mutex_lock(&netdev->mutex);
addr_dummy->address = *in6;
addr_dummy->netmask = *mask;
ovs_list_push_back(&netdev->addrs, &addr_dummy->node);
netdev_change_seq_changed(netdev_);
ovs_mutex_unlock(&netdev->mutex);
return 0;
}
#define DUMMY_MAX_QUEUES_PER_PORT 1024
static int
netdev_dummy_set_config(struct netdev *netdev_, const struct smap *args,
char **errp OVS_UNUSED)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
const char *pcap;
int new_n_rxq, new_n_txq, new_numa_id;
ovs_mutex_lock(&netdev->mutex);
netdev->ifindex = smap_get_int(args, "ifindex", -EOPNOTSUPP);
dummy_packet_conn_set_config(&netdev->conn, args);
if (netdev->rxq_pcap) {
ovs_pcap_close(netdev->rxq_pcap);
}
if (netdev->tx_pcap && netdev->tx_pcap != netdev->rxq_pcap) {
ovs_pcap_close(netdev->tx_pcap);
}
netdev->rxq_pcap = netdev->tx_pcap = NULL;
pcap = smap_get(args, "pcap");
if (pcap) {
netdev->rxq_pcap = netdev->tx_pcap = ovs_pcap_open(pcap, "ab");
} else {
const char *rxq_pcap = smap_get(args, "rxq_pcap");
const char *tx_pcap = smap_get(args, "tx_pcap");
if (rxq_pcap) {
netdev->rxq_pcap = ovs_pcap_open(rxq_pcap, "ab");
}
if (tx_pcap) {
netdev->tx_pcap = ovs_pcap_open(tx_pcap, "ab");
}
}
netdev->ol_ip_rx_csum_set_good =
smap_get_bool(args, "ol_ip_rx_csum_set_good", false);
netdev->ol_ip_rx_csum_set_bad =
smap_get_bool(args, "ol_ip_rx_csum_set_bad", false);
netdev->ol_ip_tx_csum = smap_get_bool(args, "ol_ip_tx_csum", false);
if (netdev->ol_ip_tx_csum) {
netdev_->ol_flags |= NETDEV_TX_OFFLOAD_IPV4_CKSUM;
netdev->ol_ip_tx_csum_disabled =
smap_get_bool(args, "ol_ip_tx_csum_disabled", false);
} else {
netdev_->ol_flags &= ~NETDEV_TX_OFFLOAD_IPV4_CKSUM;
netdev->ol_ip_tx_csum_disabled = true;
}
netdev->ol_l4_rx_csum_set_good =
smap_get_bool(args, "ol_l4_rx_csum_set_good", false);
netdev->ol_l4_rx_csum_set_bad =
smap_get_bool(args, "ol_l4_rx_csum_set_bad", false);
netdev->ol_l4_rx_csum_set_partial =
smap_get_bool(args, "ol_l4_rx_csum_set_partial", false);
netdev->ol_l4_tx_csum = smap_get_bool(args, "ol_l4_tx_csum", false);
if (netdev->ol_l4_tx_csum) {
netdev_->ol_flags |= NETDEV_TX_OFFLOAD_TCP_CKSUM;
netdev_->ol_flags |= NETDEV_TX_OFFLOAD_UDP_CKSUM;
netdev->ol_l4_tx_csum_disabled =
smap_get_bool(args, "ol_l4_tx_csum_disabled", false);
} else {
netdev_->ol_flags &= ~NETDEV_TX_OFFLOAD_TCP_CKSUM;
netdev_->ol_flags &= ~NETDEV_TX_OFFLOAD_UDP_CKSUM;
netdev->ol_l4_tx_csum_disabled = true;
}
if (userspace_tso_enabled()) {
netdev->ol_tso_segsz = smap_get_int(args, "ol_tso_segsz", 0);
if (netdev->ol_tso_segsz) {
netdev_->ol_flags |= (NETDEV_TX_OFFLOAD_TCP_TSO
| NETDEV_TX_OFFLOAD_TCP_CKSUM);
}
}
netdev_change_seq_changed(netdev_);
/* 'dummy-pmd' specific config. */
if (!netdev_->netdev_class->is_pmd) {
goto exit;
}
new_n_rxq = MAX(smap_get_int(args, "n_rxq", NR_QUEUE), 1);
new_n_txq = MAX(smap_get_int(args, "n_txq", NR_QUEUE), 1);
if (new_n_rxq > DUMMY_MAX_QUEUES_PER_PORT ||
new_n_txq > DUMMY_MAX_QUEUES_PER_PORT) {
VLOG_WARN("The one or both of interface %s queues"
"(rxq: %d, txq: %d) exceed %d. Sets it %d.\n",
netdev_get_name(netdev_),
new_n_rxq,
new_n_txq,
DUMMY_MAX_QUEUES_PER_PORT,
DUMMY_MAX_QUEUES_PER_PORT);
new_n_rxq = MIN(DUMMY_MAX_QUEUES_PER_PORT, new_n_rxq);
new_n_txq = MIN(DUMMY_MAX_QUEUES_PER_PORT, new_n_txq);
}
new_numa_id = smap_get_int(args, "numa_id", 0);
if (new_n_rxq != netdev->requested_n_rxq
|| new_n_txq != netdev->requested_n_txq
|| new_numa_id != netdev->requested_numa_id) {
netdev->requested_n_rxq = new_n_rxq;
netdev->requested_n_txq = new_n_txq;
netdev->requested_numa_id = new_numa_id;
netdev_request_reconfigure(netdev_);
}
exit:
ovs_mutex_unlock(&netdev->mutex);
return 0;
}
static int
netdev_dummy_get_numa_id(const struct netdev *netdev_)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
ovs_mutex_lock(&netdev->mutex);
int numa_id = netdev->numa_id;
ovs_mutex_unlock(&netdev->mutex);
return numa_id;
}
/* Sets the number of tx queues and rx queues for the dummy PMD interface. */
static int
netdev_dummy_reconfigure(struct netdev *netdev_)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
int old_n_txq = netdev_->n_txq;
int old_n_rxq = netdev_->n_rxq;
ovs_mutex_lock(&netdev->mutex);
netdev_->n_txq = netdev->requested_n_txq;
netdev_->n_rxq = netdev->requested_n_rxq;
netdev->numa_id = netdev->requested_numa_id;
if (netdev_->n_txq != old_n_txq || netdev_->n_rxq != old_n_rxq) {
/* Resize the per queue stats arrays. */
netdev->txq_stats = xrealloc(netdev->txq_stats,
netdev_->n_txq *
sizeof *netdev->txq_stats);
netdev->rxq_stats = xrealloc(netdev->rxq_stats,
netdev_->n_rxq *
sizeof *netdev->rxq_stats);
/* Reset all stats for consistency between per-queue and global
* counters. */
memset(&netdev->stats, 0, sizeof netdev->stats);
netdev->custom_stats[0].value = 0;
netdev->custom_stats[1].value = 0;
memset(netdev->txq_stats, 0,
netdev_->n_txq * sizeof *netdev->txq_stats);
memset(netdev->rxq_stats, 0,
netdev_->n_rxq * sizeof *netdev->rxq_stats);
}
ovs_mutex_unlock(&netdev->mutex);
return 0;
}
static struct netdev_rxq *
netdev_dummy_rxq_alloc(void)
{
struct netdev_rxq_dummy *rx = xzalloc(sizeof *rx);
return &rx->up;
}
static int
netdev_dummy_rxq_construct(struct netdev_rxq *rxq_)
{
struct netdev_rxq_dummy *rx = netdev_rxq_dummy_cast(rxq_);
struct netdev_dummy *netdev = netdev_dummy_cast(rx->up.netdev);
ovs_mutex_lock(&netdev->mutex);
ovs_list_push_back(&netdev->rxes, &rx->node);
ovs_list_init(&rx->recv_queue);
rx->recv_queue_len = 0;
rx->seq = seq_create();
ovs_mutex_unlock(&netdev->mutex);
return 0;
}
static void
netdev_dummy_rxq_destruct(struct netdev_rxq *rxq_)
{
struct netdev_rxq_dummy *rx = netdev_rxq_dummy_cast(rxq_);
struct netdev_dummy *netdev = netdev_dummy_cast(rx->up.netdev);
ovs_mutex_lock(&netdev->mutex);
ovs_list_remove(&rx->node);
pkt_list_delete(&rx->recv_queue);
ovs_mutex_unlock(&netdev->mutex);
seq_destroy(rx->seq);
}
static void
netdev_dummy_rxq_dealloc(struct netdev_rxq *rxq_)
{
struct netdev_rxq_dummy *rx = netdev_rxq_dummy_cast(rxq_);
free(rx);
}
static int
netdev_dummy_rxq_recv(struct netdev_rxq *rxq_, struct dp_packet_batch *batch,
int *qfill)
{
struct netdev_rxq_dummy *rx = netdev_rxq_dummy_cast(rxq_);
struct netdev_dummy *netdev = netdev_dummy_cast(rx->up.netdev);
struct dp_packet *packet;
ovs_mutex_lock(&netdev->mutex);
if (!ovs_list_is_empty(&rx->recv_queue)) {
struct pkt_list_node *pkt_node;
ASSIGN_CONTAINER(pkt_node, ovs_list_pop_front(&rx->recv_queue), list_node);
packet = pkt_node->pkt;
free(pkt_node);
rx->recv_queue_len--;
} else {
packet = NULL;
}
ovs_mutex_unlock(&netdev->mutex);
if (!packet) {
if (netdev_is_pmd(&netdev->up)) {
/* If 'netdev' is a PMD device, this is called as part of the PMD
* thread busy loop. We yield here (without quiescing) for two
* reasons:
*
* - To reduce the CPU utilization during the testsuite
* - To give valgrind a chance to switch thread. According
* to the valgrind documentation, there's a big lock that
* prevents multiple thread from being executed at the same
* time. On my system, without this sleep, the pmd threads
* testcases fail under valgrind, because ovs-vswitchd becomes
* unresponsive. */
sched_yield();
}
return EAGAIN;
}
ovs_mutex_lock(&netdev->mutex);
netdev->stats.rx_packets++;
netdev->rxq_stats[rxq_->queue_id].packets++;
netdev->stats.rx_bytes += dp_packet_size(packet);
netdev->rxq_stats[rxq_->queue_id].bytes += dp_packet_size(packet);
netdev->custom_stats[0].value++;
netdev->custom_stats[1].value++;
*dp_packet_ol_flags_ptr(packet) &= ~DP_PACKET_OL_TX_ANY_CKSUM;
if (netdev->ol_ip_rx_csum_set_good) {
dp_packet_ol_set_ip_csum_good(packet);
} else if (netdev->ol_ip_rx_csum_set_bad) {
dp_packet_ol_set_ip_csum_bad(packet);
} else {
dp_packet_ol_reset_ip_csum_good(packet);
}
if (netdev->ol_l4_rx_csum_set_good) {
dp_packet_ol_set_l4_csum_good(packet);
} else if (netdev->ol_l4_rx_csum_set_bad) {
dp_packet_ol_set_l4_csum_bad(packet);
} else if (netdev->ol_l4_rx_csum_set_partial) {
dp_packet_ol_set_l4_csum_partial(packet);
} else {
dp_packet_ol_reset_l4_csum_good(packet);
}
if (userspace_tso_enabled() && netdev->ol_tso_segsz) {
dp_packet_set_tso_segsz(packet, netdev->ol_tso_segsz);
dp_packet_hwol_set_tcp_seg(packet);
}
if (VLOG_IS_DBG_ENABLED()) {
bool ip_csum_good;
bool l4_csum_good;
bool ip_csum_bad;
bool l4_csum_bad;
ip_csum_good = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_IP_CKSUM_GOOD);
ip_csum_bad = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_IP_CKSUM_BAD);
l4_csum_good = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_L4_CKSUM_GOOD);
l4_csum_bad = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_L4_CKSUM_BAD);
VLOG_DBG("Rx: packet with csum IP %s, L4 %s, segsz %"PRIu16,
ip_csum_good ? (ip_csum_bad ? "good+bad" : "good")
: (ip_csum_bad ? "bad" : "unknown"),
l4_csum_good ? (l4_csum_bad ? "partial" : "good")
: (l4_csum_bad ? "bad" : "unknown"),
dp_packet_get_tso_segsz(packet));
}
ovs_mutex_unlock(&netdev->mutex);
dp_packet_batch_init_packet(batch, packet);
if (qfill) {
*qfill = -ENOTSUP;
}
return 0;
}
static void
netdev_dummy_rxq_wait(struct netdev_rxq *rxq_)
{
struct netdev_rxq_dummy *rx = netdev_rxq_dummy_cast(rxq_);
struct netdev_dummy *netdev = netdev_dummy_cast(rx->up.netdev);
uint64_t seq = seq_read(rx->seq);
ovs_mutex_lock(&netdev->mutex);
if (!ovs_list_is_empty(&rx->recv_queue)) {
poll_immediate_wake();
} else {
seq_wait(rx->seq, seq);
}
ovs_mutex_unlock(&netdev->mutex);
}
static int
netdev_dummy_rxq_drain(struct netdev_rxq *rxq_)
{
struct netdev_rxq_dummy *rx = netdev_rxq_dummy_cast(rxq_);
struct netdev_dummy *netdev = netdev_dummy_cast(rx->up.netdev);
ovs_mutex_lock(&netdev->mutex);
pkt_list_delete(&rx->recv_queue);
rx->recv_queue_len = 0;
ovs_mutex_unlock(&netdev->mutex);
seq_change(rx->seq);
return 0;
}
static int
netdev_dummy_send(struct netdev *netdev, int qid,
struct dp_packet_batch *batch,
bool concurrent_txq OVS_UNUSED)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
int error = 0;
struct dp_packet *packet;
DP_PACKET_BATCH_FOR_EACH(i, packet, batch) {
const void *buffer = dp_packet_data(packet);
size_t size = dp_packet_size(packet);
uint64_t flags;
bool is_tso;
ovs_mutex_lock(&dev->mutex);
flags = netdev->ol_flags;
if (!dev->ol_ip_tx_csum_disabled) {
flags &= ~NETDEV_TX_OFFLOAD_IPV4_CKSUM;
}
if (!dev->ol_l4_tx_csum_disabled) {
flags &= ~NETDEV_TX_OFFLOAD_TCP_CKSUM;
flags &= ~NETDEV_TX_OFFLOAD_UDP_CKSUM;
}
is_tso = userspace_tso_enabled() && dev->ol_tso_segsz &&
dp_packet_hwol_is_tso(packet);
ovs_mutex_unlock(&dev->mutex);
if (!dp_packet_is_eth(packet)) {
error = EPFNOSUPPORT;
break;
}
if (size < ETH_HEADER_LEN) {
error = EMSGSIZE;
break;
} else {
const struct eth_header *eth = buffer;
int max_size;
ovs_mutex_lock(&dev->mutex);
max_size = dev->mtu + ETH_HEADER_LEN;
ovs_mutex_unlock(&dev->mutex);
if (eth->eth_type == htons(ETH_TYPE_VLAN)) {
max_size += VLAN_HEADER_LEN;
}
if (size > max_size && !is_tso) {
error = EMSGSIZE;
break;
}
}
if (VLOG_IS_DBG_ENABLED()) {
bool ip_csum_good;
bool l4_csum_good;
bool ip_csum_bad;
bool l4_csum_bad;
ip_csum_good = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_IP_CKSUM_GOOD);
ip_csum_bad = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_IP_CKSUM_BAD);
l4_csum_good = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_L4_CKSUM_GOOD);
l4_csum_bad = !!(*dp_packet_ol_flags_ptr(packet)
& DP_PACKET_OL_RX_L4_CKSUM_BAD);
VLOG_DBG("Tx: packet with csum IP %s, L4 %s, segsz %"PRIu16
", Tx flags %s, %s, %s",
ip_csum_good ? (ip_csum_bad ? "good+bad" : "good")
: (ip_csum_bad ? "bad" : "unknown"),
l4_csum_good ? (l4_csum_bad ? "partial" : "good")
: (l4_csum_bad ? "bad" : "unknown"),
dp_packet_get_tso_segsz(packet),
dp_packet_hwol_tx_ip_csum(packet) ? "ip_csum" : "none",
dp_packet_hwol_tx_l4_checksum(packet) ? "l4_csum"
: "none",
dp_packet_hwol_is_tso(packet) ? "tso" : "none");
}
if (dp_packet_hwol_tx_ip_csum(packet)
&& !dp_packet_ip_checksum_good(packet)) {
dp_packet_ol_send_prepare(packet, flags);
}
if (dp_packet_hwol_l4_is_tcp(packet)
&& !dp_packet_l4_checksum_good(packet)) {
dp_packet_ol_send_prepare(packet, flags);
}
if (dp_packet_hwol_l4_is_udp(packet)
&& !dp_packet_l4_checksum_good(packet)) {
dp_packet_ol_send_prepare(packet, flags);
}
ovs_mutex_lock(&dev->mutex);
dev->stats.tx_packets++;
dev->txq_stats[qid].packets++;
dev->stats.tx_bytes += size;
dev->txq_stats[qid].bytes += size;
dummy_packet_conn_send(&dev->conn, buffer, size);
/* Reply to ARP requests for 'dev''s assigned IP address. */
struct netdev_addr_dummy *addr_dummy;
LIST_FOR_EACH (addr_dummy, node, &dev->addrs) {
ovs_be32 address = in6_addr_get_mapped_ipv4(&addr_dummy->address);
struct dp_packet dp;
struct flow flow;
dp_packet_use_const(&dp, buffer, size);
flow_extract(&dp, &flow);
if (flow.dl_type == htons(ETH_TYPE_ARP)
&& flow.nw_proto == ARP_OP_REQUEST
&& flow.nw_dst == address) {
struct dp_packet *reply = dp_packet_new(0);
compose_arp(reply, ARP_OP_REPLY, dev->hwaddr, flow.dl_src,
false, flow.nw_dst, flow.nw_src);
netdev_dummy_queue_packet(dev, reply, NULL, 0);
break;
}
}
if (dev->tx_pcap) {
struct dp_packet dp;
dp_packet_use_const(&dp, buffer, size);
ovs_pcap_write(dev->tx_pcap, &dp);
}
ovs_mutex_unlock(&dev->mutex);
}
dp_packet_delete_batch(batch, true);
return error;
}
static int
netdev_dummy_set_etheraddr(struct netdev *netdev, const struct eth_addr mac)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dev->mutex);
if (!eth_addr_equals(dev->hwaddr, mac)) {
dev->hwaddr = mac;
netdev_change_seq_changed(netdev);
}
ovs_mutex_unlock(&dev->mutex);
return 0;
}
static int
netdev_dummy_get_etheraddr(const struct netdev *netdev, struct eth_addr *mac)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dev->mutex);
*mac = dev->hwaddr;
ovs_mutex_unlock(&dev->mutex);
return 0;
}
static int
netdev_dummy_get_mtu(const struct netdev *netdev, int *mtup)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dev->mutex);
*mtup = dev->mtu;
ovs_mutex_unlock(&dev->mutex);
return 0;
}
#define DUMMY_MIN_MTU 68
#define DUMMY_MAX_MTU 65535
static int
netdev_dummy_set_mtu(struct netdev *netdev, int mtu)
{
if (mtu < DUMMY_MIN_MTU || mtu > DUMMY_MAX_MTU) {
return EINVAL;
}
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dev->mutex);
if (dev->mtu != mtu) {
dev->mtu = mtu;
netdev_change_seq_changed(netdev);
}
ovs_mutex_unlock(&dev->mutex);
return 0;
}
static int
netdev_dummy_get_stats(const struct netdev *netdev, struct netdev_stats *stats)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dev->mutex);
/* Passing only collected counters */
stats->tx_packets = dev->stats.tx_packets;
stats->tx_bytes = dev->stats.tx_bytes;
stats->rx_packets = dev->stats.rx_packets;
stats->rx_bytes = dev->stats.rx_bytes;
ovs_mutex_unlock(&dev->mutex);
return 0;
}
static int
netdev_dummy_get_custom_stats(const struct netdev *netdev,
struct netdev_custom_stats *custom_stats)
{
int i, j;
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dev->mutex);
#define DUMMY_Q_STATS \
DUMMY_Q_STAT(bytes) \
DUMMY_Q_STAT(packets)
custom_stats->size = C_STATS_SIZE;
#define DUMMY_Q_STAT(NAME) + netdev->n_rxq
custom_stats->size += DUMMY_Q_STATS;
#undef DUMMY_Q_STAT
#define DUMMY_Q_STAT(NAME) + netdev->n_txq
custom_stats->size += DUMMY_Q_STATS;
#undef DUMMY_Q_STAT
custom_stats->counters = xcalloc(custom_stats->size,
sizeof(struct netdev_custom_counter));
j = 0;
for (i = 0 ; i < C_STATS_SIZE ; i++) {
custom_stats->counters[j].value = dev->custom_stats[i].value;
ovs_strlcpy(custom_stats->counters[j++].name,
dev->custom_stats[i].name,
NETDEV_CUSTOM_STATS_NAME_SIZE);
}
for (i = 0; i < netdev->n_rxq; i++) {
#define DUMMY_Q_STAT(NAME) \
snprintf(custom_stats->counters[j].name, \
NETDEV_CUSTOM_STATS_NAME_SIZE, "rx_q%d_"#NAME, i); \
custom_stats->counters[j++].value = dev->rxq_stats[i].NAME;
DUMMY_Q_STATS
#undef DUMMY_Q_STAT
}
for (i = 0; i < netdev->n_txq; i++) {
#define DUMMY_Q_STAT(NAME) \
snprintf(custom_stats->counters[j].name, \
NETDEV_CUSTOM_STATS_NAME_SIZE, "tx_q%d_"#NAME, i); \
custom_stats->counters[j++].value = dev->txq_stats[i].NAME;
DUMMY_Q_STATS
#undef DUMMY_Q_STAT
}
ovs_mutex_unlock(&dev->mutex);
return 0;
}
static int
netdev_dummy_get_queue(const struct netdev *netdev OVS_UNUSED,
unsigned int queue_id, struct smap *details OVS_UNUSED)
{
if (queue_id == 0) {
return 0;
} else {
return EINVAL;
}
}
static void
netdev_dummy_init_queue_stats(struct netdev_queue_stats *stats)
{
*stats = (struct netdev_queue_stats) {
.tx_bytes = UINT64_MAX,
.tx_packets = UINT64_MAX,
.tx_errors = UINT64_MAX,
.created = LLONG_MIN,
};
}
static int
netdev_dummy_get_queue_stats(const struct netdev *netdev OVS_UNUSED,
unsigned int queue_id,
struct netdev_queue_stats *stats)
{
if (queue_id == 0) {
netdev_dummy_init_queue_stats(stats);
return 0;
} else {
return EINVAL;
}
}
struct netdev_dummy_queue_state {
unsigned int next_queue;
};
static int
netdev_dummy_queue_dump_start(const struct netdev *netdev OVS_UNUSED,
void **statep)
{
struct netdev_dummy_queue_state *state = xmalloc(sizeof *state);
state->next_queue = 0;
*statep = state;
return 0;
}
static int
netdev_dummy_queue_dump_next(const struct netdev *netdev OVS_UNUSED,
void *state_,
unsigned int *queue_id,
struct smap *details OVS_UNUSED)
{
struct netdev_dummy_queue_state *state = state_;
if (state->next_queue == 0) {
*queue_id = 0;
state->next_queue++;
return 0;
} else {
return EOF;
}
}
static int
netdev_dummy_queue_dump_done(const struct netdev *netdev OVS_UNUSED,
void *state)
{
free(state);
return 0;
}
static int
netdev_dummy_dump_queue_stats(const struct netdev *netdev OVS_UNUSED,
void (*cb)(unsigned int queue_id,
struct netdev_queue_stats *,
void *aux),
void *aux)
{
struct netdev_queue_stats stats;
netdev_dummy_init_queue_stats(&stats);
cb(0, &stats, aux);
return 0;
}
static int
netdev_dummy_get_ifindex(const struct netdev *netdev)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
int ifindex;
ovs_mutex_lock(&dev->mutex);
ifindex = dev->ifindex;
ovs_mutex_unlock(&dev->mutex);
return ifindex;
}
static int
netdev_dummy_update_flags__(struct netdev_dummy *netdev,
enum netdev_flags off, enum netdev_flags on,
enum netdev_flags *old_flagsp)
OVS_REQUIRES(netdev->mutex)
{
if ((off | on) & ~(NETDEV_UP | NETDEV_PROMISC)) {
return EINVAL;
}
*old_flagsp = netdev->flags;
netdev->flags |= on;
netdev->flags &= ~off;
if (*old_flagsp != netdev->flags) {
netdev_change_seq_changed(&netdev->up);
}
return 0;
}
static int
netdev_dummy_update_flags(struct netdev *netdev_,
enum netdev_flags off, enum netdev_flags on,
enum netdev_flags *old_flagsp)
{
struct netdev_dummy *netdev = netdev_dummy_cast(netdev_);
int error;
ovs_mutex_lock(&netdev->mutex);
error = netdev_dummy_update_flags__(netdev, off, on, old_flagsp);
ovs_mutex_unlock(&netdev->mutex);
return error;
}
/* Flow offload API. */
static uint32_t
netdev_dummy_flow_hash(const ovs_u128 *ufid)
{
return ufid->u32[0];
}
static struct offloaded_flow *
find_offloaded_flow(const struct hmap *offloaded_flows, const ovs_u128 *ufid)
{
uint32_t hash = netdev_dummy_flow_hash(ufid);
struct offloaded_flow *data;
HMAP_FOR_EACH_WITH_HASH (data, node, hash, offloaded_flows) {
if (ovs_u128_equals(*ufid, data->ufid)) {
return data;
}
}
return NULL;
}
static int
netdev_dummy_flow_put(struct netdev *netdev, struct match *match,
struct nlattr *actions OVS_UNUSED,
size_t actions_len OVS_UNUSED,
const ovs_u128 *ufid, struct offload_info *info,
struct dpif_flow_stats *stats)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
struct offloaded_flow *off_flow;
bool modify = true;
ovs_mutex_lock(&dev->mutex);
off_flow = find_offloaded_flow(&dev->offloaded_flows, ufid);
if (!off_flow) {
/* Create new offloaded flow. */
off_flow = xzalloc(sizeof *off_flow);
memcpy(&off_flow->ufid, ufid, sizeof *ufid);
hmap_insert(&dev->offloaded_flows, &off_flow->node,
netdev_dummy_flow_hash(ufid));
modify = false;
}
off_flow->mark = info->flow_mark;
memcpy(&off_flow->match, match, sizeof *match);
/* As we have per-netdev 'offloaded_flows', we don't need to match
* the 'in_port' for received packets. This will also allow offloading for
* packets passed to 'receive' command without specifying the 'in_port'. */
off_flow->match.wc.masks.in_port.odp_port = 0;
ovs_mutex_unlock(&dev->mutex);
if (VLOG_IS_DBG_ENABLED()) {
struct ds ds = DS_EMPTY_INITIALIZER;
ds_put_format(&ds, "%s: flow put[%s]: ", netdev_get_name(netdev),
modify ? "modify" : "create");
odp_format_ufid(ufid, &ds);
ds_put_cstr(&ds, " flow match: ");
match_format(match, NULL, &ds, OFP_DEFAULT_PRIORITY);
ds_put_format(&ds, ", mark: %"PRIu32, info->flow_mark);
VLOG_DBG("%s", ds_cstr(&ds));
ds_destroy(&ds);
}
if (stats) {
memset(stats, 0, sizeof *stats);
}
return 0;
}
static int
netdev_dummy_flow_del(struct netdev *netdev, const ovs_u128 *ufid,
struct dpif_flow_stats *stats)
{
struct netdev_dummy *dev = netdev_dummy_cast(netdev);
struct offloaded_flow *off_flow;
const char *error = NULL;
uint32_t mark;
ovs_mutex_lock(&dev->mutex);
off_flow = find_offloaded_flow(&dev->offloaded_flows, ufid);
if (!off_flow) {
error = "No such flow.";
goto exit;
}
mark = off_flow->mark;
hmap_remove(&dev->offloaded_flows, &off_flow->node);
free(off_flow);
exit:
ovs_mutex_unlock(&dev->mutex);
if (error || VLOG_IS_DBG_ENABLED()) {
struct ds ds = DS_EMPTY_INITIALIZER;
ds_put_format(&ds, "%s: ", netdev_get_name(netdev));
if (error) {
ds_put_cstr(&ds, "failed to ");
}
ds_put_cstr(&ds, "flow del: ");
odp_format_ufid(ufid, &ds);
if (error) {
ds_put_format(&ds, " error: %s", error);
} else {
ds_put_format(&ds, " mark: %"PRIu32, mark);
}
VLOG(error ? VLL_WARN : VLL_DBG, "%s", ds_cstr(&ds));
ds_destroy(&ds);
}
if (stats) {
memset(stats, 0, sizeof *stats);
}
return error ? -1 : 0;
}
#define NETDEV_DUMMY_CLASS_COMMON \
.run = netdev_dummy_run, \
.wait = netdev_dummy_wait, \
.alloc = netdev_dummy_alloc, \
.construct = netdev_dummy_construct, \
.destruct = netdev_dummy_destruct, \
.dealloc = netdev_dummy_dealloc, \
.get_config = netdev_dummy_get_config, \
.set_config = netdev_dummy_set_config, \
.get_numa_id = netdev_dummy_get_numa_id, \
.send = netdev_dummy_send, \
.set_etheraddr = netdev_dummy_set_etheraddr, \
.get_etheraddr = netdev_dummy_get_etheraddr, \
.get_mtu = netdev_dummy_get_mtu, \
.set_mtu = netdev_dummy_set_mtu, \
.get_ifindex = netdev_dummy_get_ifindex, \
.get_stats = netdev_dummy_get_stats, \
.get_custom_stats = netdev_dummy_get_custom_stats, \
.get_queue = netdev_dummy_get_queue, \
.get_queue_stats = netdev_dummy_get_queue_stats, \
.queue_dump_start = netdev_dummy_queue_dump_start, \
.queue_dump_next = netdev_dummy_queue_dump_next, \
.queue_dump_done = netdev_dummy_queue_dump_done, \
.dump_queue_stats = netdev_dummy_dump_queue_stats, \
.get_addr_list = netdev_dummy_get_addr_list, \
.update_flags = netdev_dummy_update_flags, \
.rxq_alloc = netdev_dummy_rxq_alloc, \
.rxq_construct = netdev_dummy_rxq_construct, \
.rxq_destruct = netdev_dummy_rxq_destruct, \
.rxq_dealloc = netdev_dummy_rxq_dealloc, \
.rxq_recv = netdev_dummy_rxq_recv, \
.rxq_wait = netdev_dummy_rxq_wait, \
.rxq_drain = netdev_dummy_rxq_drain
static const struct netdev_class dummy_class = {
NETDEV_DUMMY_CLASS_COMMON,
.type = "dummy"
};
static const struct netdev_class dummy_internal_class = {
NETDEV_DUMMY_CLASS_COMMON,
.type = "dummy-internal"
};
static const struct netdev_class dummy_pmd_class = {
NETDEV_DUMMY_CLASS_COMMON,
.type = "dummy-pmd",
.is_pmd = true,
.reconfigure = netdev_dummy_reconfigure
};
static int
netdev_dummy_offloads_init_flow_api(struct netdev *netdev)
{
return is_dummy_class(netdev->netdev_class) ? 0 : EOPNOTSUPP;
}
static const struct netdev_flow_api netdev_offload_dummy = {
.type = "dummy",
.flow_put = netdev_dummy_flow_put,
.flow_del = netdev_dummy_flow_del,
.init_flow_api = netdev_dummy_offloads_init_flow_api,
};
/* Helper functions. */
static void
pkt_list_delete(struct ovs_list *l)
{
struct pkt_list_node *pkt;
LIST_FOR_EACH_POP(pkt, list_node, l) {
dp_packet_delete(pkt->pkt);
free(pkt);
}
}
static void
addr_list_delete(struct ovs_list *l)
{
struct netdev_addr_dummy *addr_dummy;
LIST_FOR_EACH_POP (addr_dummy, node, l) {
free(addr_dummy);
}
}
static struct dp_packet *
eth_from_packet(const char *s)
{
struct dp_packet *packet;
eth_from_hex(s, &packet);
return packet;
}
static struct dp_packet *
eth_from_flow_str(const char *s, size_t packet_size,
struct flow *flow, char **errorp)
{
*errorp = NULL;
enum odp_key_fitness fitness;
struct dp_packet *packet;
struct ofpbuf odp_key;
int error;
/* Convert string to datapath key.
*
* It would actually be nicer to parse an OpenFlow-like flow key here, but
* the code for that currently calls exit() on parse error. We have to
* settle for parsing a datapath key for now.
*/
ofpbuf_init(&odp_key, 0);
error = odp_flow_from_string(s, NULL, &odp_key, NULL, errorp);
if (error) {
ofpbuf_uninit(&odp_key);
return NULL;
}
/* Convert odp_key to flow. */
fitness = odp_flow_key_to_flow(odp_key.data, odp_key.size, flow, errorp);
if (fitness == ODP_FIT_ERROR) {
ofpbuf_uninit(&odp_key);
return NULL;
}
packet = dp_packet_new(0);
if (packet_size) {
flow_compose(packet, flow, NULL, 0, false);
if (dp_packet_size(packet) < packet_size) {
packet_expand(packet, flow, packet_size);
} else if (dp_packet_size(packet) > packet_size){
dp_packet_delete(packet);
packet = NULL;
}
} else {
flow_compose(packet, flow, NULL, 64, false);
}
ofpbuf_uninit(&odp_key);
return packet;
}
static void
netdev_dummy_queue_packet__(struct netdev_rxq_dummy *rx, struct dp_packet *packet)
{
struct pkt_list_node *pkt_node = xmalloc(sizeof *pkt_node);
pkt_node->pkt = packet;
ovs_list_push_back(&rx->recv_queue, &pkt_node->list_node);
rx->recv_queue_len++;
seq_change(rx->seq);
}
static void
netdev_dummy_queue_packet(struct netdev_dummy *dummy, struct dp_packet *packet,
struct flow *flow, int queue_id)
OVS_REQUIRES(dummy->mutex)
{
struct netdev_rxq_dummy *rx, *prev;
struct offloaded_flow *data;
struct flow packet_flow;
if (dummy->rxq_pcap) {
ovs_pcap_write(dummy->rxq_pcap, packet);
}
if (!flow) {
flow = &packet_flow;
flow_extract(packet, flow);
}
HMAP_FOR_EACH (data, node, &dummy->offloaded_flows) {
if (flow_equal_except(flow, &data->match.flow, &data->match.wc)) {
dp_packet_set_flow_mark(packet, data->mark);
if (VLOG_IS_DBG_ENABLED()) {
struct ds ds = DS_EMPTY_INITIALIZER;
ds_put_format(&ds, "%s: packet: ",
netdev_get_name(&dummy->up));
/* 'flow' does not contain proper port number here.
* Let's just clear it as it wildcarded anyway. */
flow->in_port.ofp_port = 0;
flow_format(&ds, flow, NULL);
ds_put_cstr(&ds, " matches with flow: ");
odp_format_ufid(&data->ufid, &ds);
ds_put_cstr(&ds, " ");
match_format(&data->match, NULL, &ds, OFP_DEFAULT_PRIORITY);
ds_put_format(&ds, " with mark: %"PRIu32, data->mark);
VLOG_DBG("%s", ds_cstr(&ds));
ds_destroy(&ds);
}
break;
}
}
prev = NULL;
LIST_FOR_EACH (rx, node, &dummy->rxes) {
if (rx->up.queue_id == queue_id &&
rx->recv_queue_len < NETDEV_DUMMY_MAX_QUEUE) {
if (prev) {
netdev_dummy_queue_packet__(prev, dp_packet_clone(packet));
}
prev = rx;
}
}
if (prev) {
netdev_dummy_queue_packet__(prev, packet);
} else {
dp_packet_delete(packet);
}
}
static void
netdev_dummy_receive(struct unixctl_conn *conn,
int argc, const char *argv[], void *aux OVS_UNUSED)
{
struct netdev_dummy *dummy_dev;
struct netdev *netdev;
int i, k = 1, rx_qid = 0;
netdev = netdev_from_name(argv[k++]);
if (!netdev || !is_dummy_class(netdev->netdev_class)) {
unixctl_command_reply_error(conn, "no such dummy netdev");
goto exit_netdev;
}
dummy_dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dummy_dev->mutex);
if (argc > k + 1 && !strcmp(argv[k], "--qid")) {
rx_qid = strtol(argv[k + 1], NULL, 10);
if (rx_qid < 0 || rx_qid >= netdev->n_rxq) {
unixctl_command_reply_error(conn, "bad rx queue id.");
goto exit;
}
k += 2;
}
for (i = k; i < argc; i++) {
struct dp_packet *packet;
struct flow flow;
/* Try to parse 'argv[i]' as packet in hex. */
packet = eth_from_packet(argv[i]);
if (!packet) {
int packet_size = 0;
const char *flow_str = argv[i];
/* Parse optional --len argument immediately follows a 'flow'. */
if (argc >= i + 2 && !strcmp(argv[i + 1], "--len")) {
packet_size = strtol(argv[i + 2], NULL, 10);
if (packet_size < ETH_TOTAL_MIN) {
unixctl_command_reply_error(conn, "too small packet len");
goto exit;
}
i += 2;
}
/* Try parse 'argv[i]' as odp flow. */
char *error_s;
packet = eth_from_flow_str(flow_str, packet_size, &flow, &error_s);
if (!packet) {
unixctl_command_reply_error(conn, error_s);
free(error_s);
goto exit;
}
} else {
flow_extract(packet, &flow);
}
netdev_dummy_queue_packet(dummy_dev, packet, &flow, rx_qid);
}
unixctl_command_reply(conn, NULL);
exit:
ovs_mutex_unlock(&dummy_dev->mutex);
exit_netdev:
netdev_close(netdev);
}
static void
netdev_dummy_set_admin_state__(struct netdev_dummy *dev, bool admin_state)
OVS_REQUIRES(dev->mutex)
{
enum netdev_flags old_flags;
if (admin_state) {
netdev_dummy_update_flags__(dev, 0, NETDEV_UP, &old_flags);
} else {
netdev_dummy_update_flags__(dev, NETDEV_UP, 0, &old_flags);
}
}
static void
netdev_dummy_set_admin_state(struct unixctl_conn *conn, int argc,
const char *argv[], void *aux OVS_UNUSED)
{
bool up;
if (!strcasecmp(argv[argc - 1], "up")) {
up = true;
} else if ( !strcasecmp(argv[argc - 1], "down")) {
up = false;
} else {
unixctl_command_reply_error(conn, "Invalid Admin State");
return;
}
if (argc > 2) {
struct netdev *netdev = netdev_from_name(argv[1]);
if (netdev && is_dummy_class(netdev->netdev_class)) {
struct netdev_dummy *dummy_dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dummy_dev->mutex);
netdev_dummy_set_admin_state__(dummy_dev, up);
ovs_mutex_unlock(&dummy_dev->mutex);
netdev_close(netdev);
} else {
unixctl_command_reply_error(conn, "Unknown Dummy Interface");
netdev_close(netdev);
return;
}
} else {
struct netdev_dummy *netdev;
ovs_mutex_lock(&dummy_list_mutex);
LIST_FOR_EACH (netdev, list_node, &dummy_list) {
ovs_mutex_lock(&netdev->mutex);
netdev_dummy_set_admin_state__(netdev, up);
ovs_mutex_unlock(&netdev->mutex);
}
ovs_mutex_unlock(&dummy_list_mutex);
}
unixctl_command_reply(conn, "OK");
}
static void
display_conn_state__(struct ds *s, const char *name,
enum dummy_netdev_conn_state state)
{
ds_put_format(s, "%s: ", name);
switch (state) {
case CONN_STATE_CONNECTED:
ds_put_cstr(s, "connected\n");
break;
case CONN_STATE_NOT_CONNECTED:
ds_put_cstr(s, "disconnected\n");
break;
case CONN_STATE_UNKNOWN:
default:
ds_put_cstr(s, "unknown\n");
break;
};
}
static void
netdev_dummy_conn_state(struct unixctl_conn *conn, int argc,
const char *argv[], void *aux OVS_UNUSED)
{
enum dummy_netdev_conn_state state = CONN_STATE_UNKNOWN;
struct ds s;
ds_init(&s);
if (argc > 1) {
const char *dev_name = argv[1];
struct netdev *netdev = netdev_from_name(dev_name);
if (netdev && is_dummy_class(netdev->netdev_class)) {
struct netdev_dummy *dummy_dev = netdev_dummy_cast(netdev);
ovs_mutex_lock(&dummy_dev->mutex);
state = dummy_netdev_get_conn_state(&dummy_dev->conn);
ovs_mutex_unlock(&dummy_dev->mutex);
netdev_close(netdev);
}
display_conn_state__(&s, dev_name, state);
} else {
struct netdev_dummy *netdev;
ovs_mutex_lock(&dummy_list_mutex);
LIST_FOR_EACH (netdev, list_node, &dummy_list) {
ovs_mutex_lock(&netdev->mutex);
state = dummy_netdev_get_conn_state(&netdev->conn);
ovs_mutex_unlock(&netdev->mutex);
if (state != CONN_STATE_UNKNOWN) {
display_conn_state__(&s, netdev->up.name, state);
}
}
ovs_mutex_unlock(&dummy_list_mutex);
}
unixctl_command_reply(conn, ds_cstr(&s));
ds_destroy(&s);
}
static void
netdev_dummy_ip4addr(struct unixctl_conn *conn, int argc OVS_UNUSED,
const char *argv[], void *aux OVS_UNUSED)
{
struct netdev *netdev = netdev_from_name(argv[1]);
if (netdev && is_dummy_class(netdev->netdev_class)) {
struct in_addr ip, mask;
struct in6_addr ip6;
uint32_t plen;
char *error;
error = ip_parse_cidr(argv[2], &ip.s_addr, &plen);
if (!error) {
mask.s_addr = be32_prefix_mask(plen);
netdev_dummy_add_in4(netdev, ip, mask);
/* Insert local route entry for the new address. */
in6_addr_set_mapped_ipv4(&ip6, ip.s_addr);
ovs_router_force_insert(0, &ip6, plen + 96, true, argv[1],
&in6addr_any, &ip6);
unixctl_command_reply(conn, "OK");
} else {
unixctl_command_reply_error(conn, error);
free(error);
}
} else {
unixctl_command_reply_error(conn, "Unknown Dummy Interface");
}
netdev_close(netdev);
}
static void
netdev_dummy_ip6addr(struct unixctl_conn *conn, int argc OVS_UNUSED,
const char *argv[], void *aux OVS_UNUSED)
{
struct netdev *netdev = netdev_from_name(argv[1]);
if (netdev && is_dummy_class(netdev->netdev_class)) {
struct in6_addr ip6;
char *error;
uint32_t plen;
error = ipv6_parse_cidr(argv[2], &ip6, &plen);
if (!error) {
struct in6_addr mask;
mask = ipv6_create_mask(plen);
netdev_dummy_add_in6(netdev, &ip6, &mask);
/* Insert local route entry for the new address. */
ovs_router_force_insert(0, &ip6, plen, true, argv[1],
&in6addr_any, &ip6);
unixctl_command_reply(conn, "OK");
} else {
unixctl_command_reply_error(conn, error);
free(error);
}
} else {
unixctl_command_reply_error(conn, "Unknown Dummy Interface");
}
netdev_close(netdev);
}
static void
netdev_dummy_override(const char *type)
{
if (!netdev_unregister_provider(type)) {
struct netdev_class *class;
int error;
class = xmemdup(&dummy_class, sizeof dummy_class);
class->type = xstrdup(type);
error = netdev_register_provider(class);
if (error) {
VLOG_ERR("%s: failed to register netdev provider (%s)",
type, ovs_strerror(error));
free(CONST_CAST(char *, class->type));
free(class);
}
}
}
void
netdev_dummy_register(enum dummy_level level)
{
unixctl_command_register("netdev-dummy/receive",
"name [--qid queue_id] packet|flow [--len packet_len]",
2, INT_MAX, netdev_dummy_receive, NULL);
unixctl_command_register("netdev-dummy/set-admin-state",
"[netdev] up|down", 1, 2,
netdev_dummy_set_admin_state, NULL);
unixctl_command_register("netdev-dummy/conn-state",
"[netdev]", 0, 1,
netdev_dummy_conn_state, NULL);
unixctl_command_register("netdev-dummy/ip4addr",
"[netdev] ipaddr/mask-prefix-len", 2, 2,
netdev_dummy_ip4addr, NULL);
unixctl_command_register("netdev-dummy/ip6addr",
"[netdev] ip6addr", 2, 2,
netdev_dummy_ip6addr, NULL);
if (level == DUMMY_OVERRIDE_ALL) {
struct sset types;
const char *type;
sset_init(&types);
netdev_enumerate_types(&types);
SSET_FOR_EACH (type, &types) {
if (strcmp(type, "patch")) {
netdev_dummy_override(type);
}
}
sset_destroy(&types);
} else if (level == DUMMY_OVERRIDE_SYSTEM) {
netdev_dummy_override("system");
}
netdev_register_provider(&dummy_class);
netdev_register_provider(&dummy_internal_class);
netdev_register_provider(&dummy_pmd_class);
netdev_register_flow_api_provider(&netdev_offload_dummy);
netdev_vport_tunnel_register();
}