@@ -13,6 +13,7 @@ Post-v2.14.0
* Removed support for vhost-user dequeue zero-copy.
* Add support for DPDK 20.11.
* Add hardware offload support for tunnel pop action (experimental).
+ * Add hardware offload support for VXLAN flows (experimental).
- Userspace datapath:
* Add the 'pmd' option to "ovs-appctl dpctl/dump-flows", which
restricts a flow dump to a single PMD thread if set.
@@ -62,6 +62,7 @@ struct ufid_to_rte_flow_data {
struct rte_flow *rte_flow;
bool actions_offloaded;
struct dpif_flow_stats stats;
+ struct ovs_list indirect;
};
static uint32_t
@@ -122,6 +123,7 @@ ufid_to_rte_flow_associate(const ovs_u128 *ufid, struct netdev *netdev,
data->netdev = netdev_ref(netdev);
data->rte_flow = rte_flow;
data->actions_offloaded = actions_offloaded;
+ ovs_list_init(&data->indirect);
cmap_insert(&ufid_to_rte_flow,
CONST_CAST(struct cmap_node *, &data->node), hash);
@@ -136,6 +138,9 @@ ufid_to_rte_flow_disassociate(struct ufid_to_rte_flow_data *data)
cmap_remove(&ufid_to_rte_flow,
CONST_CAST(struct cmap_node *, &data->node), hash);
netdev_close(data->netdev);
+ if (!ovs_list_is_empty(&data->indirect)) {
+ ovs_list_remove(&data->indirect);
+ }
ovsrcu_postpone(free, data);
}
@@ -148,6 +153,8 @@ struct flow_patterns {
struct rte_flow_item *items;
int cnt;
int current_max;
+ uint32_t num_of_tnl_items;
+ struct ds s_tnl;
};
struct flow_actions {
@@ -163,16 +170,20 @@ struct flow_actions {
static void
dump_flow_attr(struct ds *s, struct ds *s_extra,
const struct rte_flow_attr *attr,
+ struct flow_patterns *flow_patterns,
struct flow_actions *flow_actions)
{
if (flow_actions->num_of_tnl_actions) {
ds_clone(s_extra, &flow_actions->s_tnl);
+ } else if (flow_patterns->num_of_tnl_items) {
+ ds_clone(s_extra, &flow_patterns->s_tnl);
}
- ds_put_format(s, "%s%spriority %"PRIu32" group %"PRIu32" %s%s",
+ ds_put_format(s, "%s%spriority %"PRIu32" group %"PRIu32" %s%s%s",
attr->ingress ? "ingress " : "",
attr->egress ? "egress " : "", attr->priority, attr->group,
attr->transfer ? "transfer " : "",
- flow_actions->num_of_tnl_actions ? "tunnel_set 1 " : "");
+ flow_actions->num_of_tnl_actions ? "tunnel_set 1 " : "",
+ flow_patterns->num_of_tnl_items ? "tunnel_match 1 " : "");
}
/* Adds one pattern item 'field' with the 'mask' to dynamic string 's' using
@@ -186,9 +197,18 @@ dump_flow_attr(struct ds *s, struct ds *s_extra,
}
static void
-dump_flow_pattern(struct ds *s, const struct rte_flow_item *item)
+dump_flow_pattern(struct ds *s,
+ struct flow_patterns *flow_patterns,
+ int pattern_index)
{
- if (item->type == RTE_FLOW_ITEM_TYPE_ETH) {
+ const struct rte_flow_item *item = &flow_patterns->items[pattern_index];
+
+ if (item->type == RTE_FLOW_ITEM_TYPE_END) {
+ ds_put_cstr(s, "end ");
+ } else if (flow_patterns->num_of_tnl_items &&
+ pattern_index < flow_patterns->num_of_tnl_items) {
+ return;
+ } else if (item->type == RTE_FLOW_ITEM_TYPE_ETH) {
const struct rte_flow_item_eth *eth_spec = item->spec;
const struct rte_flow_item_eth *eth_mask = item->mask;
@@ -577,19 +597,19 @@ dump_flow_action(struct ds *s, struct ds *s_extra,
static struct ds *
dump_flow(struct ds *s, struct ds *s_extra,
const struct rte_flow_attr *attr,
- const struct rte_flow_item *items,
+ struct flow_patterns *flow_patterns,
struct flow_actions *flow_actions)
{
int i;
if (attr) {
- dump_flow_attr(s, s_extra, attr, flow_actions);
+ dump_flow_attr(s, s_extra, attr, flow_patterns, flow_actions);
}
ds_put_cstr(s, "pattern ");
- while (items && items->type != RTE_FLOW_ITEM_TYPE_END) {
- dump_flow_pattern(s, items++);
+ for (i = 0; i < flow_patterns->cnt; i++) {
+ dump_flow_pattern(s, flow_patterns, i);
}
- ds_put_cstr(s, "end actions ");
+ ds_put_cstr(s, "actions ");
for (i = 0; i < flow_actions->cnt; i++) {
dump_flow_action(s, s_extra, flow_actions, i);
}
@@ -599,11 +619,12 @@ dump_flow(struct ds *s, struct ds *s_extra,
static struct rte_flow *
netdev_offload_dpdk_flow_create(struct netdev *netdev,
const struct rte_flow_attr *attr,
- const struct rte_flow_item *items,
+ struct flow_patterns *flow_patterns,
struct flow_actions *flow_actions,
struct rte_flow_error *error)
{
const struct rte_flow_action *actions = flow_actions->actions;
+ const struct rte_flow_item *items = flow_patterns->items;
struct ds s_extra = DS_EMPTY_INITIALIZER;
struct ds s = DS_EMPTY_INITIALIZER;
struct rte_flow *flow;
@@ -612,7 +633,7 @@ netdev_offload_dpdk_flow_create(struct netdev *netdev,
flow = netdev_dpdk_rte_flow_create(netdev, attr, items, actions, error);
if (flow) {
if (!VLOG_DROP_DBG(&rl)) {
- dump_flow(&s, &s_extra, attr, items, flow_actions);
+ dump_flow(&s, &s_extra, attr, flow_patterns, flow_actions);
extra_str = ds_cstr(&s_extra);
VLOG_DBG_RL(&rl, "%s: rte_flow 0x%"PRIxPTR" %s flow create %d %s",
netdev_get_name(netdev), (intptr_t) flow, extra_str,
@@ -627,7 +648,7 @@ netdev_offload_dpdk_flow_create(struct netdev *netdev,
VLOG_RL(&rl, level, "%s: rte_flow creation failed: %d (%s).",
netdev_get_name(netdev), error->type, error->message);
if (!vlog_should_drop(&this_module, level, &rl)) {
- dump_flow(&s, &s_extra, attr, items, flow_actions);
+ dump_flow(&s, &s_extra, attr, flow_patterns, flow_actions);
extra_str = ds_cstr(&s_extra);
VLOG_RL(&rl, level, "%s: Failed flow: %s flow create %d %s",
netdev_get_name(netdev), extra_str,
@@ -662,6 +683,28 @@ add_flow_pattern(struct flow_patterns *patterns, enum rte_flow_item_type type,
patterns->cnt++;
}
+static void
+add_flow_tnl_patterns(struct flow_patterns *all_patterns,
+ struct rte_flow_item *tnl_items,
+ uint32_t num_of_tnl_items,
+ struct flow_patterns *flow_patterns)
+{
+ int i;
+
+ all_patterns->num_of_tnl_items = num_of_tnl_items;
+
+ for (i = 0; i < num_of_tnl_items; i++) {
+ add_flow_pattern(all_patterns, tnl_items[i].type, tnl_items[i].spec,
+ tnl_items[i].mask);
+ }
+
+ for (i = 0; i < flow_patterns->cnt; i++) {
+ add_flow_pattern(all_patterns, flow_patterns->items[i].type,
+ flow_patterns->items[i].spec,
+ flow_patterns->items[i].mask);
+ }
+}
+
static void
add_flow_action(struct flow_actions *actions, enum rte_flow_action_type type,
const void *conf)
@@ -1513,7 +1556,6 @@ create_netdev_offload(struct netdev *netdev,
uint32_t flow_mark)
{
struct flow_actions rss_actions = { .s_tnl = DS_EMPTY_INITIALIZER, };
- struct rte_flow_item *items = flow_patterns->items;
struct rte_flow_attr flow_attr = { .ingress = 1 };
struct ufid_to_rte_flow_data *flow_data = NULL;
bool actions_offloaded = true;
@@ -1522,8 +1564,9 @@ create_netdev_offload(struct netdev *netdev,
if (!flow_mark) {
flow_attr.transfer = 1;
- flow = netdev_offload_dpdk_flow_create(netdev, &flow_attr, items,
- flow_actions, &error);
+ flow = netdev_offload_dpdk_flow_create(netdev, &flow_attr,
+ flow_patterns, flow_actions,
+ &error);
}
if (flow_mark || !flow) {
@@ -1532,8 +1575,9 @@ create_netdev_offload(struct netdev *netdev,
*/
actions_offloaded = false;
add_flow_mark_rss_actions(&rss_actions, flow_mark, netdev);
- flow = netdev_offload_dpdk_flow_create(netdev, &flow_attr, items,
- &rss_actions, &error);
+ flow = netdev_offload_dpdk_flow_create(netdev, &flow_attr,
+ flow_patterns, &rss_actions,
+ &error);
}
if (flow) {
@@ -1548,6 +1592,88 @@ create_netdev_offload(struct netdev *netdev,
return flow_data;
}
+static struct ufid_to_rte_flow_data *
+create_vport_offload(struct netdev *vport,
+ const ovs_u128 *ufid,
+ struct flow_patterns *flow_patterns,
+ struct flow_actions *flow_actions)
+{
+ struct flow_patterns all_patterns = { .items = NULL, .cnt = 0 };
+ struct ufid_to_rte_flow_data *flows_data = NULL;
+ struct netdev_flow_dump **netdev_dumps;
+ struct rte_flow_item *tnl_items;
+ struct rte_flow_tunnel tunnel;
+ struct rte_flow_error error;
+ uint32_t num_of_tnl_items;
+ struct ovs_list indirect;
+ struct netdev *physdev;
+ bool actions_offloaded;
+ int num_ports = 0;
+ int err;
+ int i;
+
+ ovs_list_init(&indirect);
+ netdev_dumps = netdev_ports_flow_dump_create(vport->dpif_type, &num_ports,
+ false);
+ actions_offloaded = false;
+ for (i = 0; i < num_ports; i++) {
+ physdev = netdev_dumps[i]->netdev;
+ if (!netdev_dpdk_is_uplink_port(physdev)) {
+ continue;
+ }
+
+ ds_init(&all_patterns.s_tnl);
+ if (vport_to_rte_tunnel(vport, &tunnel, physdev,
+ &all_patterns.s_tnl)) {
+ break;
+ }
+ if (netdev_dpdk_rte_flow_tunnel_match(physdev, &tunnel, &tnl_items,
+ &num_of_tnl_items, &error)) {
+ VLOG_DBG_RL(&rl, "%s: netdev_dpdk_rte_flow_tunnel_match failed: "
+ "%d (%s).", netdev_get_name(physdev), error.type,
+ error.message);
+ ds_destroy(&all_patterns.s_tnl);
+ continue;
+ }
+ add_flow_tnl_patterns(&all_patterns, tnl_items, num_of_tnl_items,
+ flow_patterns);
+ flows_data = create_netdev_offload(physdev, ufid, &all_patterns,
+ flow_actions, 0);
+ if (netdev_dpdk_rte_flow_tunnel_item_release(physdev, tnl_items,
+ num_of_tnl_items,
+ &error)) {
+ VLOG_DBG_RL(&rl, "%s: netdev_dpdk_rte_flow_tunnel_item_release "
+ "failed: %d (%s).", netdev_get_name(physdev),
+ error.type, error.message);
+ }
+ ds_destroy(&all_patterns.s_tnl);
+ all_patterns.cnt = 0;
+ if (flows_data) {
+ ovs_list_push_back(&indirect, &flows_data->indirect);
+ actions_offloaded |= flows_data->actions_offloaded;
+ }
+ }
+ free_flow_patterns(&all_patterns);
+
+ for (i = 0; i < num_ports; i++) {
+ err = netdev_flow_dump_destroy(netdev_dumps[i]);
+ if (err != 0 && err != EOPNOTSUPP) {
+ VLOG_ERR("failed dumping netdev: %s", ovs_strerror(err));
+ }
+ }
+
+ if (!ovs_list_is_empty(&indirect)) {
+ flows_data = ufid_to_rte_flow_associate(ufid, vport, NULL,
+ actions_offloaded);
+ ovs_list_push_back(&indirect, &flows_data->indirect);
+ ovs_list_remove(&indirect);
+ VLOG_DBG("%s: installed flow for ufid "UUID_FMT,
+ netdev_get_name(vport), UUID_ARGS((struct uuid *) ufid));
+ }
+
+ return flows_data;
+}
+
static struct ufid_to_rte_flow_data *
netdev_offload_dpdk_add_flow(struct netdev *netdev,
struct match *match,
@@ -1569,8 +1695,12 @@ netdev_offload_dpdk_add_flow(struct netdev *netdev,
err = parse_flow_actions(netdev, &actions, nl_actions, actions_len);
- flows_data = create_netdev_offload(netdev, ufid, &patterns, &actions,
- err ? info->flow_mark : 0);
+ if (netdev_vport_is_vport_class(netdev->netdev_class)) {
+ flows_data = create_vport_offload(netdev, ufid, &patterns, &actions);
+ } else {
+ flows_data = create_netdev_offload(netdev, ufid, &patterns, &actions,
+ err ? info->flow_mark : 0);
+ }
out:
free_flow_patterns(&patterns);
@@ -1581,17 +1711,30 @@ out:
static int
netdev_offload_dpdk_flow_destroy(struct ufid_to_rte_flow_data *rte_flow_data)
{
+ struct ufid_to_rte_flow_data *phys, *next;
struct rte_flow_error error;
struct rte_flow *rte_flow;
struct netdev *netdev;
ovs_u128 *ufid;
- int ret;
+ int ret = 0;
- rte_flow = rte_flow_data->rte_flow;
netdev = rte_flow_data->netdev;
+
+ /* For vport, follow indirect rules (recursive call). */
+ if (netdev_vport_is_vport_class(netdev->netdev_class)) {
+ LIST_FOR_EACH_SAFE (phys, next, indirect, &rte_flow_data->indirect) {
+ if (netdev_offload_dpdk_flow_destroy(phys)) {
+ ret = -1;
+ }
+ }
+ }
+
+ rte_flow = rte_flow_data->rte_flow;
ufid = &rte_flow_data->ufid;
- ret = netdev_dpdk_rte_flow_destroy(netdev, rte_flow, &error);
+ if (rte_flow) {
+ ret = netdev_dpdk_rte_flow_destroy(netdev, rte_flow, &error);
+ }
if (ret == 0) {
ufid_to_rte_flow_disassociate(rte_flow_data);
@@ -1657,7 +1800,7 @@ netdev_offload_dpdk_flow_del(struct netdev *netdev OVS_UNUSED,
struct ufid_to_rte_flow_data *rte_flow_data;
rte_flow_data = ufid_to_rte_flow_data_find(ufid, netdev, true);
- if (!rte_flow_data || !rte_flow_data->rte_flow) {
+ if (!rte_flow_data) {
return -1;
}
@@ -1691,11 +1834,14 @@ netdev_offload_dpdk_flow_get(struct netdev *netdev,
{
struct rte_flow_query_count query = { .reset = 1 };
struct ufid_to_rte_flow_data *rte_flow_data;
+ struct ufid_to_rte_flow_data *phys_data;
+ struct dpif_flow_stats phys_stats;
+ struct dpif_flow_attrs phys_attrs;
struct rte_flow_error error;
int ret = 0;
rte_flow_data = ufid_to_rte_flow_data_find(ufid, netdev, false);
- if (!rte_flow_data || !rte_flow_data->rte_flow) {
+ if (!rte_flow_data) {
ret = -1;
goto out;
}
@@ -1707,6 +1853,29 @@ netdev_offload_dpdk_flow_get(struct netdev *netdev,
goto out;
}
attrs->dp_layer = "dpdk";
+ /* For vport, follow indirect rules (recursive call). */
+ if (netdev_vport_is_vport_class(netdev->netdev_class)) {
+ memset(&rte_flow_data->stats, 0, sizeof rte_flow_data->stats);
+ LIST_FOR_EACH (phys_data, indirect, &rte_flow_data->indirect) {
+ if (netdev_offload_dpdk_flow_get(phys_data->netdev, match, actions,
+ ufid, &phys_stats, &phys_attrs,
+ buf)) {
+ ret = -1;
+ } else {
+ rte_flow_data->stats.n_packets += phys_stats.n_packets;
+ rte_flow_data->stats.n_bytes += phys_stats.n_bytes;
+ if (phys_stats.used > rte_flow_data->stats.used) {
+ rte_flow_data->stats.used = phys_stats.used;
+ }
+ }
+ }
+ goto out;
+ }
+
+ if (!rte_flow_data->rte_flow) {
+ ret = -1;
+ goto out;
+ }
ret = netdev_dpdk_rte_flow_query_count(netdev, rte_flow_data->rte_flow,
&query, &error);
if (ret) {
@@ -1720,9 +1889,11 @@ netdev_offload_dpdk_flow_get(struct netdev *netdev,
if (query.hits_set && query.hits) {
rte_flow_data->stats.used = time_msec();
}
- memcpy(stats, &rte_flow_data->stats, sizeof *stats);
out:
- attrs->dp_extra_info = NULL;
+ if (!ret) {
+ memcpy(stats, &rte_flow_data->stats, sizeof *stats);
+ attrs->dp_extra_info = NULL;
+ }
return ret;
}
Vports are virtual, OVS only logical devices, so rte_flows cannot be applied as is on them. Instead, apply the rules on uplink ports. This create a one-to-many relation between an OVS flow to HW rules. Link them with an indirect list, and support such rules with such indirection. Signed-off-by: Eli Britstein <elibr@nvidia.com> --- NEWS | 1 + lib/netdev-offload-dpdk.c | 223 +++++++++++++++++++++++++++++++++----- 2 files changed, 198 insertions(+), 26 deletions(-)