This commit relaxes the assumption that all packets have an Ethernet
header, and adds support for layer 3 flows. For each packet received on
the Linux kernel datapath the l2 and l3 members of struct ofpbuf are
intialized appropriately, and some functions now expect this (notable
flow_extract()), in order to differentiate between layer 2 and layer 3
packets. struct flow has now a new 'noeth' member, because we cannot
assume that a flow has no Ethernet header when eth_src and eth_dst are
0. For layer 3 packets, the protocol type is still stored in the
eth_type member.
Switching L2->L3 and L3->L2 are both implemented by adding the pop_eth
and push_eth actions respectively when a transition is detected. The
push_eth action puts 0s on both source and destination MACs. These
addresses can be modified with mod_dl_dst and mod_dl_src actions.
Signed-off-by: Lorand Jakab <loja...@cisco.com>
---
lib/bfd.c | 1 +
lib/dpif-linux.c | 7 +++++++
lib/dpif-netdev.c | 12 +++++-------
lib/dpif.c | 6 ++++--
lib/flow.c | 36 ++++++++++++++++++++++++++++++++----
lib/flow.h | 3 ++-
lib/match.c | 9 ++++++---
lib/odp-util.c | 15 ++++++++++-----
lib/ofp-print.c | 14 ++++++++++----
lib/ofp-print.h | 3 ++-
ofproto/ofproto-dpif-xlate.c | 17 +++++++++++++++--
ofproto/ofproto-dpif-xlate.h | 3 ++-
ofproto/ofproto-dpif.c | 3 ++-
ofproto/ofproto.c | 1 +
14 files changed, 99 insertions(+), 31 deletions(-)
diff --git a/lib/bfd.c b/lib/bfd.c
index 740f4fc..21b70b1 100644
--- a/lib/bfd.c
+++ b/lib/bfd.c
@@ -540,6 +540,7 @@ bfd_put_packet(struct bfd *bfd, struct ofpbuf *p,
ovs_assert(!(bfd->flags & FLAG_POLL) || !(bfd->flags & FLAG_FINAL));
ofpbuf_reserve(p, 2); /* Properly align after the ethernet header. */
+ p->l2 = p->data;
eth = ofpbuf_put_uninit(p, sizeof *eth);
memcpy(eth->eth_src, eth_src, ETH_ADDR_LEN);
memcpy(eth->eth_dst, bfd->eth_dst, ETH_ADDR_LEN);
diff --git a/lib/dpif-linux.c b/lib/dpif-linux.c
index 42958d3..777fdee 100644
--- a/lib/dpif-linux.c
+++ b/lib/dpif-linux.c
@@ -1383,6 +1383,13 @@ parse_odp_packet(struct ofpbuf *buf, struct dpif_upcall
*upcall,
upcall->key = CONST_CAST(struct nlattr *,
nl_attr_get(a[OVS_PACKET_ATTR_KEY]));
upcall->key_len = nl_attr_get_size(a[OVS_PACKET_ATTR_KEY]);
+ if (nl_attr_find__(upcall->key, upcall->key_len, OVS_KEY_ATTR_ETHERNET)) {
+ upcall->packet->l2 = upcall->packet->data;
+ upcall->packet->l3 = NULL;
+ } else {
+ upcall->packet->l2 = NULL;
+ upcall->packet->l3 = upcall->packet->data;
+ }
upcall->userdata = a[OVS_PACKET_ATTR_USERDATA];
*dp_ifindex = ovs_header->dp_ifindex;
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 1b0039c..9c07016 100644
--- a/lib/dpif-netdev.c
+++ b/lib/dpif-netdev.c
@@ -1030,7 +1030,7 @@ static int
dpif_netdev_execute(struct dpif *dpif, const struct dpif_execute *execute)
{
struct dp_netdev *dp = get_dp_netdev(dpif);
- struct ofpbuf copy;
+ struct ofpbuf *copy;
struct flow key;
int error;
@@ -1040,21 +1040,19 @@ dpif_netdev_execute(struct dpif *dpif, const struct
dpif_execute *execute)
}
/* Make a deep copy of 'packet', because we might modify its data. */
- ofpbuf_init(©, DP_NETDEV_HEADROOM + execute->packet->size);
- ofpbuf_reserve(©, DP_NETDEV_HEADROOM);
- ofpbuf_put(©, execute->packet->data, execute->packet->size);
+ copy = ofpbuf_clone_with_headroom(execute->packet, DP_NETDEV_HEADROOM);
- flow_extract(©, 0, 0, NULL, NULL, &key);
+ flow_extract(copy, 0, 0, NULL, NULL, &key);
error = dpif_netdev_flow_from_nlattrs(execute->key, execute->key_len,
&key);
if (!error) {
ovs_mutex_lock(&dp_netdev_mutex);
- dp_netdev_execute_actions(dp, ©, &key,
+ dp_netdev_execute_actions(dp, copy, &key,
execute->actions, execute->actions_len);
ovs_mutex_unlock(&dp_netdev_mutex);
}
- ofpbuf_uninit(©);
+ ofpbuf_uninit(copy);
return error;
}
diff --git a/lib/dpif.c b/lib/dpif.c
index 783a7cb..8a51b08 100644
--- a/lib/dpif.c
+++ b/lib/dpif.c
@@ -1328,7 +1328,8 @@ dpif_recv(struct dpif *dpif, struct dpif_upcall *upcall,
struct ofpbuf *buf)
char *packet;
packet = ofp_packet_to_string(upcall->packet->data,
- upcall->packet->size);
+ upcall->packet->size,
+ upcall->packet->l3);
ds_init(&flow);
odp_flow_key_format(upcall->key, upcall->key_len, &flow);
@@ -1530,7 +1531,8 @@ log_execute_message(struct dpif *dpif, const struct
dpif_execute *execute,
char *packet;
packet = ofp_packet_to_string(execute->packet->data,
- execute->packet->size);
+ execute->packet->size,
+ execute->packet->l3);
ds_put_format(&ds, "%s: execute ", dpif_name(dpif));
format_odp_actions(&ds, execute->actions, execute->actions_len);
if (error) {
diff --git a/lib/flow.c b/lib/flow.c
index 8c336f6..fa7faa1 100644
--- a/lib/flow.c
+++ b/lib/flow.c
@@ -41,6 +41,21 @@
COVERAGE_DEFINE(flow_extract);
COVERAGE_DEFINE(miniflow_malloc);
+static ovs_be16
+get_l3_eth_type(struct ofpbuf *packet)
+{
+ struct ip_header *ip = packet->l3;
+ int ip_ver = IP_VER(ip->ip_ihl_ver);
+ switch (ip_ver) {
+ case 4:
+ return htons(ETH_TYPE_IP);
+ case 6:
+ return htons(ETH_TYPE_IPV6);
+ default:
+ return 0;
+ }
+}
+
static struct arp_eth_header *
pull_arp(struct ofpbuf *packet)
{
@@ -383,6 +398,8 @@ flow_extract(struct ofpbuf *packet, uint32_t skb_priority,
uint32_t pkt_mark,
COVERAGE_INC(flow_extract);
+ ovs_assert(packet->l2 != NULL || packet->l3 != NULL);
+
memset(flow, 0, sizeof *flow);
if (tnl) {
@@ -395,11 +412,21 @@ flow_extract(struct ofpbuf *packet, uint32_t
skb_priority, uint32_t pkt_mark,
flow->skb_priority = skb_priority;
flow->pkt_mark = pkt_mark;
- packet->l2 = b.data;
- packet->l2_5 = NULL;
- packet->l3 = NULL;
- packet->l4 = NULL;
packet->l7 = NULL;
+ packet->l4 = NULL;
+
+ if (packet->l3) {
+ packet->l2_5 = NULL;
+ packet->l2 = NULL;
+ flow->noeth = true;
+ /* We assume L3 packets are either IPv4 or IPv6 */
+ flow->dl_type = get_l3_eth_type(packet);
+ goto layer3;
+ }
+
+ packet->l3 = NULL;
+ packet->l2_5 = NULL;
+ packet->l2 = b.data;
if (b.size < sizeof *eth) {
return;
@@ -425,6 +452,7 @@ flow_extract(struct ofpbuf *packet, uint32_t skb_priority,
uint32_t pkt_mark,
/* Network layer. */
packet->l3 = b.data;
+layer3:
if (flow->dl_type == htons(ETH_TYPE_IP)) {
const struct ip_header *nh = pull_ip(&b);
if (nh) {
diff --git a/lib/flow.h b/lib/flow.h
index 093d509..5d7c948 100644
--- a/lib/flow.h
+++ b/lib/flow.h
@@ -90,6 +90,7 @@ union flow_in_port {
* a 32-bit datapath port number.
*/
struct flow {
+ bool noeth; /* Flow has no Ethernet header */
struct flow_tnl tunnel; /* Encapsulating tunnel parameters. */
ovs_be64 metadata; /* OpenFlow Metadata. */
struct in6_addr ipv6_src; /* IPv6 source address. */
@@ -124,7 +125,7 @@ BUILD_ASSERT_DECL(sizeof(struct flow) % 4 == 0);
/* Remember to update FLOW_WC_SEQ when changing 'struct flow'. */
BUILD_ASSERT_DECL(offsetof(struct flow, nw_frag) + 1
- == sizeof(struct flow_tnl) + 154
+ == sizeof(struct flow_tnl) + 162
&& FLOW_WC_SEQ == 22);
/* Represents the metadata fields of struct flow. */
diff --git a/lib/match.c b/lib/match.c
index 0f674b0..0e7cd6d 100644
--- a/lib/match.c
+++ b/lib/match.c
@@ -81,9 +81,12 @@ match_wc_init(struct match *match, const struct flow *flow)
memset(&wc->masks.metadata, 0xff, sizeof wc->masks.metadata);
memset(&wc->masks.in_port, 0xff, sizeof wc->masks.in_port);
- memset(&wc->masks.vlan_tci, 0xff, sizeof wc->masks.vlan_tci);
- memset(&wc->masks.dl_src, 0xff, sizeof wc->masks.dl_src);
- memset(&wc->masks.dl_dst, 0xff, sizeof wc->masks.dl_dst);
+
+ if (!(flow->noeth)) {
+ memset(&wc->masks.vlan_tci, 0xff, sizeof wc->masks.vlan_tci);
+ memset(&wc->masks.dl_src, 0xff, sizeof wc->masks.dl_src);
+ memset(&wc->masks.dl_dst, 0xff, sizeof wc->masks.dl_dst);
+ }
if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
memset(&wc->masks.ipv6_src, 0xff, sizeof wc->masks.ipv6_src);
diff --git a/lib/odp-util.c b/lib/odp-util.c
index d584d3f..c3b5db4 100644
--- a/lib/odp-util.c
+++ b/lib/odp-util.c
@@ -3250,12 +3250,10 @@ odp_flow_key_to_flow__(const struct nlattr *key, size_t
key_len,
eth_key = nl_attr_get(attrs[OVS_KEY_ATTR_ETHERNET]);
memcpy(flow->dl_src, eth_key->eth_src, ETH_ADDR_LEN);
memcpy(flow->dl_dst, eth_key->eth_dst, ETH_ADDR_LEN);
- if (is_mask) {
- expected_attrs |= UINT64_C(1) << OVS_KEY_ATTR_ETHERNET;
- }
- }
- if (!is_mask) {
+ flow->noeth = false;
expected_attrs |= UINT64_C(1) << OVS_KEY_ATTR_ETHERNET;
+ } else {
+ flow->noeth = true;
}
/* Get Ethertype or 802.1Q TPID or FLOW_DL_TYPE_NONE. */
@@ -3452,6 +3450,13 @@ commit_set_ether_addr_action(const struct flow *flow,
struct flow *base,
return;
}
+ /* If we have a L3 --> L2 flow, the push_eth action takes care of setting
+ * the appropriate MAC source and destination addresses, no need to add a
+ * set action
+ */
+ if (base->noeth && !flow->noeth)
+ return;
+
memset(&wc->masks.dl_src, 0xff, sizeof wc->masks.dl_src);
memset(&wc->masks.dl_dst, 0xff, sizeof wc->masks.dl_dst);
diff --git a/lib/ofp-print.c b/lib/ofp-print.c
index 37e1f4f..9bcfa7c 100644
--- a/lib/ofp-print.c
+++ b/lib/ofp-print.c
@@ -55,13 +55,19 @@ static void ofp_print_error(struct ds *, enum ofperr);
/* Returns a string that represents the contents of the Ethernet frame in the
* 'len' bytes starting at 'data'. The caller must free the returned string.*/
char *
-ofp_packet_to_string(const void *data, size_t len)
+ofp_packet_to_string(const void *data, size_t len, bool is_layer3)
{
struct ds ds = DS_EMPTY_INITIALIZER;
struct ofpbuf buf;
struct flow flow;
ofpbuf_use_const(&buf, data, len);
+
+ if (is_layer3)
+ buf.l3 = buf.data;
+ else
+ buf.l2 = buf.data;
+
flow_extract(&buf, 0, 0, NULL, NULL, &flow);
flow_format(&ds, &flow);
@@ -157,7 +163,7 @@ ofp_print_packet_in(struct ds *string, const struct
ofp_header *oh,
ds_put_char(string, '\n');
if (verbosity > 0) {
- char *packet = ofp_packet_to_string(pin.packet, pin.packet_len);
+ char *packet = ofp_packet_to_string(pin.packet, pin.packet_len, false);
ds_put_cstr(string, packet);
free(packet);
}
@@ -191,7 +197,7 @@ ofp_print_packet_out(struct ds *string, const struct
ofp_header *oh,
if (po.buffer_id == UINT32_MAX) {
ds_put_format(string, " data_len=%zu", po.packet_len);
if (verbosity > 0 && po.packet_len > 0) {
- char *packet = ofp_packet_to_string(po.packet, po.packet_len);
+ char *packet = ofp_packet_to_string(po.packet, po.packet_len,
false);
ds_put_char(string, '\n');
ds_put_cstr(string, packet);
free(packet);
@@ -2811,5 +2817,5 @@ ofp_print(FILE *stream, const void *oh, size_t len, int
verbosity)
void
ofp_print_packet(FILE *stream, const void *data, size_t len)
{
- print_and_free(stream, ofp_packet_to_string(data, len));
+ print_and_free(stream, ofp_packet_to_string(data, len, false));
}
diff --git a/lib/ofp-print.h b/lib/ofp-print.h
index 825e139..15aa196 100644
--- a/lib/ofp-print.h
+++ b/lib/ofp-print.h
@@ -21,6 +21,7 @@
#include <stdint.h>
#include <stdio.h>
+#include <stdbool.h>
struct ds;
struct ofp10_match;
@@ -39,7 +40,7 @@ void ofp10_match_print(struct ds *, const struct ofp10_match
*, int verbosity);
char *ofp_to_string(const void *, size_t, int verbosity);
char *ofp10_match_to_string(const struct ofp10_match *, int verbosity);
-char *ofp_packet_to_string(const void *data, size_t len);
+char *ofp_packet_to_string(const void *data, size_t len, bool is_layer3);
void ofp_print_flow_stats(struct ds *, struct ofputil_flow_stats *);
void ofp_print_version(const struct ofp_header *, struct ds *);
diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c
index 367dd88..b3c9c31 100644
--- a/ofproto/ofproto-dpif-xlate.c
+++ b/ofproto/ofproto-dpif-xlate.c
@@ -134,6 +134,7 @@ struct xport {
bool may_enable; /* May be enabled in bonds. */
bool is_tunnel; /* Is a tunnel port. */
+ bool is_layer3; /* Is a layer 3 port. */
struct cfm *cfm; /* CFM handle or null. */
struct bfd *bfd; /* BFD handle or null. */
@@ -400,7 +401,7 @@ xlate_ofport_set(struct ofproto_dpif *ofproto, struct
ofbundle *ofbundle,
const struct ofproto_port_queue *qdscp_list, size_t n_qdscp,
enum ofputil_port_config config,
enum ofputil_port_state state, bool is_tunnel,
- bool may_enable)
+ bool may_enable, bool is_layer3)
{
struct xport *xport = xport_lookup(ofport);
size_t i;
@@ -424,6 +425,7 @@ xlate_ofport_set(struct ofproto_dpif *ofproto, struct
ofbundle *ofbundle,
xport->stp_port_no = stp_port_no;
xport->is_tunnel = is_tunnel;
xport->may_enable = may_enable;
+ xport->is_layer3 = is_layer3;
xport->odp_port = odp_port;
if (xport->netdev != netdev) {
@@ -1410,7 +1412,7 @@ xlate_normal(struct xlate_ctx *ctx)
}
/* Learn source MAC. */
- if (ctx->xin->may_learn) {
+ if (ctx->xin->may_learn && !(in_port->is_layer3)) {
update_learning_table(ctx->xbridge, flow, wc, vlan, in_xbundle);
}
@@ -1655,6 +1657,7 @@ compose_output_action__(struct xlate_ctx *ctx, ofp_port_t
ofp_port,
const struct xport *xport = get_ofp_port(ctx->xbridge, ofp_port);
struct flow_wildcards *wc = &ctx->xout->wc;
struct flow *flow = &ctx->xin->flow;
+ const struct xport *in_xport = get_ofp_port(ctx->xbridge,
flow->in_port.ofp_port);
ovs_be16 flow_vlan_tci;
uint32_t flow_pkt_mark;
uint8_t flow_nw_tos;
@@ -1681,6 +1684,16 @@ compose_output_action__(struct xlate_ctx *ctx,
ofp_port_t ofp_port,
xport->xbundle);
}
+ if ((in_xport) && !(in_xport->is_layer3) && xport->is_layer3) {
+ odp_put_pop_eth_action(&ctx->xout->odp_actions);
+ }
+
+ if (flow->noeth && !(xport->is_layer3)) {
+ flow->noeth = false;
+ odp_put_push_eth_action(&ctx->xout->odp_actions, flow->dl_src,
+ flow->dl_dst, flow->dl_type);
+ }
+
if (xport->peer) {
const struct xport *peer = xport->peer;
struct flow old_flow = ctx->xin->flow;
diff --git a/ofproto/ofproto-dpif-xlate.h b/ofproto/ofproto-dpif-xlate.h
index 7dd3534..0f4ac70 100644
--- a/ofproto/ofproto-dpif-xlate.h
+++ b/ofproto/ofproto-dpif-xlate.h
@@ -140,7 +140,8 @@ void xlate_ofport_set(struct ofproto_dpif *, struct
ofbundle *,
int stp_port_no, const struct ofproto_port_queue *qdscp,
size_t n_qdscp, enum ofputil_port_config,
enum ofputil_port_state, bool is_tunnel,
- bool may_enable) OVS_REQ_WRLOCK(xlate_rwlock);
+ bool may_enable, bool is_layer3)
+ OVS_REQ_WRLOCK(xlate_rwlock);
void xlate_ofport_remove(struct ofport_dpif *) OVS_REQ_WRLOCK(xlate_rwlock);
int xlate_receive(const struct dpif_backer *, struct ofpbuf *packet,
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index 4b029f9..56f000e 100644
--- a/ofproto/ofproto-dpif.c
+++ b/ofproto/ofproto-dpif.c
@@ -828,7 +828,8 @@ type_run(const char *type)
ofport->bfd, ofport->peer, stp_port,
ofport->qdscp, ofport->n_qdscp,
ofport->up.pp.config, ofport->up.pp.state,
- ofport->is_tunnel, ofport->may_enable);
+ ofport->is_tunnel, ofport->may_enable,
+ ofport->is_layer3);
}
ovs_rwlock_unlock(&xlate_rwlock);
diff --git a/ofproto/ofproto.c b/ofproto/ofproto.c
index 2ccbcee..f031c9d 100644
--- a/ofproto/ofproto.c
+++ b/ofproto/ofproto.c
@@ -2936,6 +2936,7 @@ handle_packet_out(struct ofconn *ofconn, const struct
ofp_header *oh)
} else {
/* Ensure that the L3 header is 32-bit aligned. */
payload = ofpbuf_clone_data_with_headroom(po.packet, po.packet_len, 2);
+ payload->l2 = payload->data;
}
/* Verify actions against packet, then send packet if successful. */
--
1.8.3.4 (Apple Git-47)
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
