[ovs-dev] [PATCH v12 rebased 3/3] userspace: add non-tap (l3) support to GRE vports
Jiri Benc
jbenc at redhat.com
Wed Oct 19 17:21:19 UTC 2016
From: Simon Horman <simon.horman at netronome.com>
Add support for layer 3 GRE vports (non-tap aka non-VTEP).
This makes use of a vport mode configuration for the existing (tap/VTEP)
GRE vports.
In order to differentiate packets for two different types of GRE vports a
new flow key attribute, OVS_KEY_ATTR_NEXT_BASE_LAYER, is used. It is
intended that this attribute is only used in userspace as there appears to
be no need for it to be used in the kernel datapath.
It is envisaged that this attribute may be used for other encapsulation
protocols that support both layer3 and layer2 inner-packets.
Signed-off-by: Simon Horman <simon.horman at netronome.com>
Signed-off-by: Jiri Benc <jbenc at redhat.com>
---
v12
* Rebase
v11
* Update for new kernel implementation that uses ARPHRD_NONE devices
* Omit match_base_layer from struct tun_port_in, it is not necessary any more
* Correct OVS_KEY_ATTR_NEXT_BASE_LAYER handling in odp_flow_key_from_flow__()
and parse_l2_5_onward() so that the encoding and decoding of the
next_base_layer field of a flow key is consistent both with itself on
encode and decode and with other fields of the flow key.
* Guard OVS_KEY_ATTR_TUNNEL_INFO with #ifndef __KERNEL__ in
openvswitch.h and note that it is only used by the user-space datapath
* Consistently use tabs for indentation
* Correct spelling and capitalisation of documentation
* Append "(layer3)" to when showing layer3 tunnel ports as
they may share the same ODP port as a non-layer3 tunnel port.
This seems simpler than v10 which included an O(n**2) filter on
showing duplicate ports.
v10
* Use a mode for layer3 ports rather than a new port type
* Update BUILD_BUG_ON() call in ovs_key_attr_size()
* Don't update tnl_port_map_lookup() to always match
on next_base_layer: the implementation didn't actually
do that and thus was a lot of code change for no behavioural
change.
v9
* New patch
fix
---
datapath/linux/compat/include/linux/openvswitch.h | 3 ++
include/openvswitch/flow.h | 12 ++++--
lib/flow.c | 34 ++++++++++++----
lib/match.c | 6 ++-
lib/netdev-linux.c | 3 +-
lib/netdev-native-tnl.c | 26 +++++++++---
lib/netdev-vport.c | 22 ++++++++--
lib/netdev.h | 1 +
lib/nx-match.c | 2 +-
lib/odp-execute.c | 2 +
lib/odp-util.c | 22 ++++++++++
lib/odp-util.h | 4 +-
lib/ofp-util.c | 2 +-
lib/tnl-ports.c | 49 +++++++++++++++++------
lib/tnl-ports.h | 3 +-
ofproto/ofproto-dpif-rid.h | 2 +-
ofproto/ofproto-dpif-sflow.c | 1 +
ofproto/ofproto-dpif-xlate.c | 2 +-
ofproto/ofproto-dpif.c | 2 +
ofproto/tunnel.c | 4 +-
tests/tunnel-push-pop-ipv6.at | 12 ++++--
tests/tunnel-push-pop.at | 26 ++++++++++--
vswitchd/vswitch.xml | 13 ++++++
23 files changed, 202 insertions(+), 51 deletions(-)
diff --git a/datapath/linux/compat/include/linux/openvswitch.h b/datapath/linux/compat/include/linux/openvswitch.h
index 67004caaf8b5..eaab35beb83d 100644
--- a/datapath/linux/compat/include/linux/openvswitch.h
+++ b/datapath/linux/compat/include/linux/openvswitch.h
@@ -360,6 +360,9 @@ enum ovs_key_attr {
#ifdef __KERNEL__
/* Only used within kernel data path. */
OVS_KEY_ATTR_TUNNEL_INFO, /* struct ovs_tunnel_info */
+#else
+ /* Only used within user-space data path. */
+ OVS_KEY_ATTR_NEXT_BASE_LAYER, /* base layer of encapsulated packet */
#endif
__OVS_KEY_ATTR_MAX
};
diff --git a/include/openvswitch/flow.h b/include/openvswitch/flow.h
index 93ed37e10ebe..46ef87e83521 100644
--- a/include/openvswitch/flow.h
+++ b/include/openvswitch/flow.h
@@ -23,7 +23,7 @@
/* This sequence number should be incremented whenever anything involving flows
* or the wildcarding of flows changes. This will cause build assertion
* failures in places which likely need to be updated. */
-#define FLOW_WC_SEQ 37
+#define FLOW_WC_SEQ 38
/* Number of Open vSwitch extension 32-bit registers. */
#define FLOW_N_REGS 16
@@ -138,6 +138,10 @@ struct flow {
ovs_be16 tp_dst; /* TCP/UDP/SCTP destination port/ICMP code. */
ovs_be32 igmp_group_ip4; /* IGMP group IPv4 address.
* Keep last for BUILD_ASSERT_DECL below. */
+
+ uint8_t next_base_layer; /* Fields of encapsulated packet, if any,
+ * start at this layer */
+ uint8_t pad4[7];
};
BUILD_ASSERT_DECL(sizeof(struct flow) % sizeof(uint64_t) == 0);
BUILD_ASSERT_DECL(sizeof(struct flow_tnl) % sizeof(uint64_t) == 0);
@@ -145,9 +149,9 @@ BUILD_ASSERT_DECL(sizeof(struct flow_tnl) % sizeof(uint64_t) == 0);
#define FLOW_U64S (sizeof(struct flow) / sizeof(uint64_t))
/* Remember to update FLOW_WC_SEQ when changing 'struct flow'. */
-BUILD_ASSERT_DECL(offsetof(struct flow, igmp_group_ip4) + sizeof(uint32_t)
- == sizeof(struct flow_tnl) + 256
- && FLOW_WC_SEQ == 37);
+BUILD_ASSERT_DECL(OFFSETOFEND(struct flow, pad4)
+ == sizeof(struct flow_tnl) + 264
+ && FLOW_WC_SEQ == 38);
/* Incremental points at which flow classification may be performed in
* segments.
diff --git a/lib/flow.c b/lib/flow.c
index d87c02ad1545..09c9fff35c0b 100644
--- a/lib/flow.c
+++ b/lib/flow.c
@@ -125,7 +125,7 @@ struct mf_ctx {
* away. Some GCC versions gave warnings on ALWAYS_INLINE, so these are
* defined as macros. */
-#if (FLOW_WC_SEQ != 36)
+#if (FLOW_WC_SEQ != 37)
#define MINIFLOW_ASSERT(X) ovs_assert(X)
BUILD_MESSAGE("FLOW_WC_SEQ changed: miniflow_extract() will have runtime "
"assertions enabled. Consider updating FLOW_WC_SEQ after "
@@ -846,6 +846,20 @@ miniflow_extract(struct dp_packet *packet, struct miniflow *dst)
miniflow_push_be16(mf, tp_dst, htons(icmp->icmp6_code));
miniflow_pad_to_64(mf, tp_dst);
}
+ } else if (OVS_LIKELY(nw_proto == IPPROTO_GRE)) {
+ if (OVS_LIKELY(size >= sizeof(struct gre_base_hdr))) {
+ const struct gre_base_hdr *gre = data_pull(&data, &size,
+ sizeof *gre);
+ if (gre->protocol == htons(ETH_TYPE_TEB)) {
+ /* No need to store a zero value for next_base_layer
+ * in the miniflow which would cost an extra word of
+ * storage. */
+ BUILD_ASSERT(LAYER_2 == 0);
+ } else {
+ miniflow_push_uint8(mf, next_base_layer, LAYER_3);
+ miniflow_pad_to_64(mf, next_base_layer);
+ }
+ }
}
}
out:
@@ -894,7 +908,7 @@ flow_get_metadata(const struct flow *flow, struct match *flow_metadata)
{
int i;
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
match_init_catchall(flow_metadata);
if (flow->tunnel.tun_id != htonll(0)) {
@@ -1304,7 +1318,7 @@ void flow_wildcards_init_for_packet(struct flow_wildcards *wc,
memset(&wc->masks, 0x0, sizeof wc->masks);
/* Update this function whenever struct flow changes. */
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
if (flow_tnl_dst_is_set(&flow->tunnel)) {
if (flow->tunnel.flags & FLOW_TNL_F_KEY) {
@@ -1425,7 +1439,7 @@ void
flow_wc_map(const struct flow *flow, struct flowmap *map)
{
/* Update this function whenever struct flow changes. */
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
flowmap_init(map);
@@ -1473,6 +1487,8 @@ flow_wc_map(const struct flow *flow, struct flowmap *map)
if (OVS_UNLIKELY(flow->nw_proto == IPPROTO_IGMP)) {
FLOWMAP_SET(map, igmp_group_ip4);
+ } else if (OVS_UNLIKELY(flow->nw_proto == IPPROTO_GRE)) {
+ FLOWMAP_SET(map, next_base_layer);
} else {
FLOWMAP_SET(map, tcp_flags);
}
@@ -1491,6 +1507,8 @@ flow_wc_map(const struct flow *flow, struct flowmap *map)
FLOWMAP_SET(map, nd_target);
FLOWMAP_SET(map, arp_sha);
FLOWMAP_SET(map, arp_tha);
+ } else if (OVS_UNLIKELY(flow->nw_proto == IPPROTO_GRE)) {
+ FLOWMAP_SET(map, next_base_layer);
} else {
FLOWMAP_SET(map, tcp_flags);
}
@@ -1512,7 +1530,7 @@ void
flow_wildcards_clear_non_packet_fields(struct flow_wildcards *wc)
{
/* Update this function whenever struct flow changes. */
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
memset(&wc->masks.metadata, 0, sizeof wc->masks.metadata);
memset(&wc->masks.regs, 0, sizeof wc->masks.regs);
@@ -1657,7 +1675,7 @@ flow_wildcards_set_xxreg_mask(struct flow_wildcards *wc, int idx,
uint32_t
miniflow_hash_5tuple(const struct miniflow *flow, uint32_t basis)
{
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
uint32_t hash = basis;
if (flow) {
@@ -1704,7 +1722,7 @@ ASSERT_SEQUENTIAL(ipv6_src, ipv6_dst);
uint32_t
flow_hash_5tuple(const struct flow *flow, uint32_t basis)
{
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
uint32_t hash = basis;
if (flow) {
@@ -2171,7 +2189,7 @@ flow_push_mpls(struct flow *flow, int n, ovs_be16 mpls_eth_type,
flow->mpls_lse[0] = set_mpls_lse_values(ttl, tc, 1, htonl(label));
/* Clear all L3 and L4 fields and dp_hash. */
- BUILD_ASSERT(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT(FLOW_WC_SEQ == 38);
memset((char *) flow + FLOW_SEGMENT_2_ENDS_AT, 0,
sizeof(struct flow) - FLOW_SEGMENT_2_ENDS_AT);
flow->dp_hash = 0;
diff --git a/lib/match.c b/lib/match.c
index c551e578582b..65d5b8e5e96d 100644
--- a/lib/match.c
+++ b/lib/match.c
@@ -1082,7 +1082,7 @@ match_format(const struct match *match, struct ds *s, int priority)
int i;
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
if (priority != OFP_DEFAULT_PRIORITY) {
ds_put_format(s, "%spriority=%s%d,",
@@ -1358,6 +1358,10 @@ match_format(const struct match *match, struct ds *s, int priority)
TCP_FLAGS(OVS_BE16_MAX));
}
+ if (wc->masks.next_base_layer) {
+ ds_put_format(s, "next_base_layer=%"PRIu8",", f->next_base_layer);
+ }
+
if (s->length > start_len) {
ds_chomp(s, ',');
}
diff --git a/lib/netdev-linux.c b/lib/netdev-linux.c
index 30189b47bad6..786e7274a417 100644
--- a/lib/netdev-linux.c
+++ b/lib/netdev-linux.c
@@ -5507,7 +5507,8 @@ get_etheraddr(const char *netdev_name, struct eth_addr *ea)
return error;
}
hwaddr_family = ifr.ifr_hwaddr.sa_family;
- if (hwaddr_family != AF_UNSPEC && hwaddr_family != ARPHRD_ETHER) {
+ if (hwaddr_family != AF_UNSPEC && hwaddr_family != ARPHRD_ETHER &&
+ hwaddr_family != ARPHRD_NONE) {
VLOG_INFO("%s device has unknown hardware address family %d",
netdev_name, hwaddr_family);
return EINVAL;
diff --git a/lib/netdev-native-tnl.c b/lib/netdev-native-tnl.c
index ce2582f136c2..bf52e9d8ab4b 100644
--- a/lib/netdev-native-tnl.c
+++ b/lib/netdev-native-tnl.c
@@ -153,6 +153,9 @@ netdev_tnl_push_ip_header(struct dp_packet *packet,
memcpy(eth, header, size);
+ dp_packet_reset_offsets(packet);
+ packet->l3_ofs = sizeof (struct eth_header);
+
if (netdev_tnl_is_header_ipv6(header)) {
ip6 = netdev_tnl_ipv6_hdr(eth);
*ip_tot_size -= IPV6_HEADER_LEN;
@@ -350,10 +353,6 @@ parse_gre_header(struct dp_packet *packet,
return -EINVAL;
}
- if (greh->protocol != htons(ETH_TYPE_TEB)) {
- return -EINVAL;
- }
-
hlen = ulen + gre_header_len(greh->flags);
if (hlen > dp_packet_size(packet)) {
return -EINVAL;
@@ -383,6 +382,12 @@ parse_gre_header(struct dp_packet *packet,
options++;
}
+ if (greh->protocol == htons(ETH_TYPE_TEB)) {
+ packet->md.packet_ethertype = htons(0);
+ } else {
+ packet->md.packet_ethertype = greh->protocol;
+ }
+
return hlen;
}
@@ -408,6 +413,12 @@ netdev_gre_pop_header(struct dp_packet *packet)
dp_packet_reset_packet(packet, hlen);
+ if (eth_type_mpls(packet->md.packet_ethertype)) {
+ packet->l2_5_ofs = 0;
+ } else if (packet->md.packet_ethertype) {
+ packet->l3_ofs = 0;
+ }
+
return packet;
err:
dp_packet_delete(packet);
@@ -446,7 +457,12 @@ netdev_gre_build_header(const struct netdev *netdev,
greh = netdev_tnl_ip_build_header(data, params, IPPROTO_GRE);
- greh->protocol = htons(ETH_TYPE_TEB);
+ if (tnl_cfg->is_layer3) {
+ greh->protocol = params->flow->dl_type;
+ } else {
+ greh->protocol = htons(ETH_TYPE_TEB);
+ }
+
greh->flags = 0;
options = (ovs_16aligned_be32 *) (greh + 1);
diff --git a/lib/netdev-vport.c b/lib/netdev-vport.c
index 02a246affcc0..9fcdc649cd22 100644
--- a/lib/netdev-vport.c
+++ b/lib/netdev-vport.c
@@ -96,9 +96,13 @@ netdev_vport_is_patch(const struct netdev *netdev)
bool
netdev_vport_is_layer3(const struct netdev *dev)
{
- const char *type = netdev_get_type(dev);
+ if (is_vport_class(netdev_get_class(dev))) {
+ struct netdev_vport *vport = netdev_vport_cast(dev);
+
+ return vport->tnl_cfg.is_layer3;
+ }
- return (!strcmp("lisp", type));
+ return false;
}
static bool
@@ -402,13 +406,14 @@ set_tunnel_config(struct netdev *dev_, const struct smap *args)
struct netdev_vport *dev = netdev_vport_cast(dev_);
const char *name = netdev_get_name(dev_);
const char *type = netdev_get_type(dev_);
- bool needs_dst_port, has_csum;
+ bool needs_dst_port, has_csum, optional_layer3;
uint16_t dst_proto = 0, src_proto = 0;
struct netdev_tunnel_config tnl_cfg;
struct smap_node *node;
has_csum = strstr(type, "gre") || strstr(type, "geneve") ||
strstr(type, "stt") || strstr(type, "vxlan");
+ optional_layer3 = !strcmp(type, "gre");
memset(&tnl_cfg, 0, sizeof tnl_cfg);
/* Add a default destination port for tunnel ports if none specified. */
@@ -422,6 +427,7 @@ set_tunnel_config(struct netdev *dev_, const struct smap *args)
if (!strcmp(type, "lisp")) {
tnl_cfg.dst_port = htons(LISP_DST_PORT);
+ tnl_cfg.is_layer3 = true;
}
if (!strcmp(type, "stt")) {
@@ -505,6 +511,10 @@ set_tunnel_config(struct netdev *dev_, const struct smap *args)
}
free(str);
+ } else if (!strcmp(node->key, "layer3") && optional_layer3) {
+ if (!strcmp(node->value, "true")) {
+ tnl_cfg.is_layer3 = true;
+ }
} else {
VLOG_WARN("%s: unknown %s argument '%s'", name, type, node->key);
}
@@ -552,6 +562,7 @@ static int
get_tunnel_config(const struct netdev *dev, struct smap *args)
{
struct netdev_vport *netdev = netdev_vport_cast(dev);
+ const char *type = netdev_get_type(dev);
struct netdev_tunnel_config tnl_cfg;
ovs_mutex_lock(&netdev->mutex);
@@ -605,7 +616,6 @@ get_tunnel_config(const struct netdev *dev, struct smap *args)
if (tnl_cfg.dst_port) {
uint16_t dst_port = ntohs(tnl_cfg.dst_port);
- const char *type = netdev_get_type(dev);
if ((!strcmp("geneve", type) && dst_port != GENEVE_DST_PORT) ||
(!strcmp("vxlan", type) && dst_port != VXLAN_DST_PORT) ||
@@ -619,6 +629,10 @@ get_tunnel_config(const struct netdev *dev, struct smap *args)
smap_add(args, "csum", "true");
}
+ if (tnl_cfg.is_layer3 && !strcmp("gre", type)) {
+ smap_add(args, "layer3", "true");
+ }
+
if (!tnl_cfg.dont_fragment) {
smap_add(args, "df_default", "false");
}
diff --git a/lib/netdev.h b/lib/netdev.h
index bad28c4c1473..12e009f72843 100644
--- a/lib/netdev.h
+++ b/lib/netdev.h
@@ -98,6 +98,7 @@ struct netdev_tunnel_config {
bool csum;
bool dont_fragment;
+ bool is_layer3;
};
void netdev_run(void);
diff --git a/lib/nx-match.c b/lib/nx-match.c
index da2919f6639e..a6b7e2ff883d 100644
--- a/lib/nx-match.c
+++ b/lib/nx-match.c
@@ -930,7 +930,7 @@ nx_put_raw(struct ofpbuf *b, enum ofp_version oxm, const struct match *match,
int match_len;
int i;
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
/* Metadata. */
if (match->wc.masks.dp_hash) {
diff --git a/lib/odp-execute.c b/lib/odp-execute.c
index 5fe6a867d5b8..4dbbea50e9a1 100644
--- a/lib/odp-execute.c
+++ b/lib/odp-execute.c
@@ -336,6 +336,7 @@ odp_execute_set_action(struct dp_packet *packet, const struct nlattr *a)
case OVS_KEY_ATTR_CT_ZONE:
case OVS_KEY_ATTR_CT_MARK:
case OVS_KEY_ATTR_CT_LABELS:
+ case OVS_KEY_ATTR_NEXT_BASE_LAYER:
case __OVS_KEY_ATTR_MAX:
default:
OVS_NOT_REACHED();
@@ -435,6 +436,7 @@ odp_execute_masked_set_action(struct dp_packet *packet,
case OVS_KEY_ATTR_ICMP:
case OVS_KEY_ATTR_ICMPV6:
case OVS_KEY_ATTR_TCP_FLAGS:
+ case OVS_KEY_ATTR_NEXT_BASE_LAYER:
case __OVS_KEY_ATTR_MAX:
default:
OVS_NOT_REACHED();
diff --git a/lib/odp-util.c b/lib/odp-util.c
index bdb9ba785c64..c6b9536a6e20 100644
--- a/lib/odp-util.c
+++ b/lib/odp-util.c
@@ -166,6 +166,7 @@ ovs_key_attr_to_string(enum ovs_key_attr attr, char *namebuf, size_t bufsize)
case OVS_KEY_ATTR_MPLS: return "mpls";
case OVS_KEY_ATTR_DP_HASH: return "dp_hash";
case OVS_KEY_ATTR_RECIRC_ID: return "recirc_id";
+ case OVS_KEY_ATTR_NEXT_BASE_LAYER: return "next_base_layer";
case __OVS_KEY_ATTR_MAX:
default:
@@ -1841,6 +1842,7 @@ static const struct attr_len_tbl ovs_flow_key_attr_lens[OVS_KEY_ATTR_MAX + 1] =
[OVS_KEY_ATTR_CT_ZONE] = { .len = 2 },
[OVS_KEY_ATTR_CT_MARK] = { .len = 4 },
[OVS_KEY_ATTR_CT_LABELS] = { .len = sizeof(struct ovs_key_ct_labels) },
+ [OVS_KEY_ATTR_NEXT_BASE_LAYER] = { .len = 1 },
};
/* Returns the correct length of the payload for a flow key attribute of the
@@ -2965,6 +2967,13 @@ format_odp_key_attr(const struct nlattr *a, const struct nlattr *ma,
ds_chomp(ds, ',');
break;
}
+
+ case OVS_KEY_ATTR_NEXT_BASE_LAYER: {
+ const uint8_t *mask = ma ? nl_attr_get(ma) : NULL;
+ format_u8u(ds, "type", nl_attr_get_u8(a), mask, verbose);
+ break;
+ }
+
case OVS_KEY_ATTR_UNSPEC:
case __OVS_KEY_ATTR_MAX:
default:
@@ -4444,6 +4453,11 @@ odp_flow_key_from_flow__(const struct odp_flow_key_parms *parms,
sctp_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_SCTP,
sizeof *sctp_key);
get_tp_key(data, sctp_key);
+ } else if (flow->nw_proto == IPPROTO_GRE) {
+ if (parms->support.next_base_layer) {
+ nl_msg_put_u8(buf, OVS_KEY_ATTR_NEXT_BASE_LAYER,
+ data->next_base_layer);
+ }
} else if (flow->dl_type == htons(ETH_TYPE_IP)
&& flow->nw_proto == IPPROTO_ICMP) {
struct ovs_key_icmp *icmp_key;
@@ -5008,6 +5022,14 @@ parse_l2_5_onward(const struct nlattr *attrs[OVS_KEY_ATTR_MAX + 1],
put_tp_key(sctp_key, flow);
expected_bit = OVS_KEY_ATTR_SCTP;
}
+ } else if (src_flow->nw_proto == IPPROTO_GRE
+ && (src_flow->dl_type == htons(ETH_TYPE_IP) ||
+ src_flow->dl_type == htons(ETH_TYPE_IPV6))
+ && !(src_flow->nw_frag & FLOW_NW_FRAG_LATER)) {
+ if (present_attrs & (UINT64_C(1) << OVS_KEY_ATTR_NEXT_BASE_LAYER)) {
+ flow->next_base_layer = nl_attr_get_u8(attrs[OVS_KEY_ATTR_NEXT_BASE_LAYER]);
+ expected_attrs |= UINT64_C(1) << OVS_KEY_ATTR_NEXT_BASE_LAYER;
+ }
} else if (src_flow->nw_proto == IPPROTO_ICMP
&& src_flow->dl_type == htons(ETH_TYPE_IP)
&& !(src_flow->nw_frag & FLOW_NW_FRAG_LATER)) {
diff --git a/lib/odp-util.h b/lib/odp-util.h
index 0f17175a61df..e5cf82455596 100644
--- a/lib/odp-util.h
+++ b/lib/odp-util.h
@@ -141,7 +141,7 @@ void odp_portno_names_destroy(struct hmap *portno_names);
* add another field and forget to adjust this value.
*/
#define ODPUTIL_FLOW_KEY_BYTES 640
-BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
/* A buffer with sufficient size and alignment to hold an nlattr-formatted flow
* key. An array of "struct nlattr" might not, in theory, be sufficiently
@@ -184,6 +184,8 @@ struct odp_support {
* 'ct_state'. The above 'ct_state' member must be true for this
* to make sense */
bool ct_state_nat;
+
+ bool next_base_layer;
};
struct odp_flow_key_parms {
diff --git a/lib/ofp-util.c b/lib/ofp-util.c
index 6d2722e39012..cf071c549d4a 100644
--- a/lib/ofp-util.c
+++ b/lib/ofp-util.c
@@ -101,7 +101,7 @@ ofputil_netmask_to_wcbits(ovs_be32 netmask)
void
ofputil_wildcard_from_ofpfw10(uint32_t ofpfw, struct flow_wildcards *wc)
{
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
/* Initialize most of wc. */
flow_wildcards_init_catchall(wc);
diff --git a/lib/tnl-ports.c b/lib/tnl-ports.c
index ffa13899ac95..8f82ac96060e 100644
--- a/lib/tnl-ports.c
+++ b/lib/tnl-ports.c
@@ -27,6 +27,7 @@
#include "hash.h"
#include "openvswitch/list.h"
#include "netdev.h"
+#include "netdev-vport.h"
#include "openvswitch/ofpbuf.h"
#include "ovs-thread.h"
#include "odp-util.h"
@@ -53,6 +54,7 @@ struct tnl_port {
odp_port_t port;
ovs_be16 tp_port;
uint8_t nw_proto;
+ bool is_layer3;
char dev_name[IFNAMSIZ];
struct ovs_list node;
};
@@ -83,7 +85,8 @@ tnl_port_free(struct tnl_port_in *p)
static void
tnl_port_init_flow(struct flow *flow, struct eth_addr mac,
- struct in6_addr *addr, uint8_t nw_proto, ovs_be16 tp_port)
+ struct in6_addr *addr, uint8_t nw_proto, ovs_be16 tp_port,
+ bool is_layer3)
{
memset(flow, 0, sizeof *flow);
@@ -98,18 +101,20 @@ tnl_port_init_flow(struct flow *flow, struct eth_addr mac,
flow->nw_proto = nw_proto;
flow->tp_dst = tp_port;
+ flow->next_base_layer = is_layer3 ? LAYER_3 : LAYER_2;
}
static void
map_insert(odp_port_t port, struct eth_addr mac, struct in6_addr *addr,
- uint8_t nw_proto, ovs_be16 tp_port, const char dev_name[])
+ uint8_t nw_proto, ovs_be16 tp_port, const char dev_name[],
+ bool is_layer3)
{
const struct cls_rule *cr;
struct tnl_port_in *p;
struct match match;
memset(&match, 0, sizeof match);
- tnl_port_init_flow(&match.flow, mac, addr, nw_proto, tp_port);
+ tnl_port_init_flow(&match.flow, mac, addr, nw_proto, tp_port, is_layer3);
do {
cr = classifier_lookup(&cls, OVS_VERSION_MAX, &match.flow, NULL);
@@ -130,6 +135,11 @@ map_insert(odp_port_t port, struct eth_addr mac, struct in6_addr *addr,
* doesn't make sense to match on UDP port numbers. */
if (tp_port) {
match.wc.masks.tp_dst = OVS_BE16_MAX;
+ } else {
+ /* Match base layer for GRE tunnels as it may
+ * be used to differentiate them.
+ */
+ match.wc.masks.next_base_layer = UINT8_MAX;
}
if (IN6_IS_ADDR_V4MAPPED(addr)) {
match.wc.masks.nw_dst = OVS_BE32_MAX;
@@ -149,14 +159,15 @@ map_insert(odp_port_t port, struct eth_addr mac, struct in6_addr *addr,
static void
map_insert_ipdev__(struct ip_device *ip_dev, char dev_name[],
- odp_port_t port, uint8_t nw_proto, ovs_be16 tp_port)
+ odp_port_t port, uint8_t nw_proto, ovs_be16 tp_port,
+ bool is_layer3)
{
if (ip_dev->n_addr) {
int i;
for (i = 0; i < ip_dev->n_addr; i++) {
map_insert(port, ip_dev->mac, &ip_dev->addr[i],
- nw_proto, tp_port, dev_name);
+ nw_proto, tp_port, dev_name, is_layer3);
}
}
}
@@ -181,7 +192,7 @@ tnl_type_to_nw_proto(const char type[])
void
tnl_port_map_insert(odp_port_t port, ovs_be16 tp_port,
- const char dev_name[], const char type[])
+ const char dev_name[], const char type[], bool is_layer3)
{
struct tnl_port *p;
struct ip_device *ip_dev;
@@ -194,7 +205,8 @@ tnl_port_map_insert(odp_port_t port, ovs_be16 tp_port,
ovs_mutex_lock(&mutex);
LIST_FOR_EACH(p, node, &port_list) {
- if (tp_port == p->tp_port && p->nw_proto == nw_proto) {
+ if (tp_port == p->tp_port && p->nw_proto == nw_proto &&
+ p->is_layer3 == is_layer3) {
goto out;
}
}
@@ -203,11 +215,13 @@ tnl_port_map_insert(odp_port_t port, ovs_be16 tp_port,
p->port = port;
p->tp_port = tp_port;
p->nw_proto = nw_proto;
+ p->is_layer3 = is_layer3;
ovs_strlcpy(p->dev_name, dev_name, sizeof p->dev_name);
ovs_list_insert(&port_list, &p->node);
LIST_FOR_EACH(ip_dev, node, &addr_list) {
- map_insert_ipdev__(ip_dev, p->dev_name, p->port, p->nw_proto, p->tp_port);
+ map_insert_ipdev__(ip_dev, p->dev_name, p->port, p->nw_proto,
+ p->tp_port, p->is_layer3);
}
out:
@@ -228,12 +242,12 @@ tnl_port_unref(const struct cls_rule *cr)
static void
map_delete(struct eth_addr mac, struct in6_addr *addr,
- ovs_be16 tp_port, uint8_t nw_proto)
+ ovs_be16 tp_port, uint8_t nw_proto, bool is_layer3)
{
const struct cls_rule *cr;
struct flow flow;
- tnl_port_init_flow(&flow, mac, addr, nw_proto, tp_port);
+ tnl_port_init_flow(&flow, mac, addr, nw_proto, tp_port, is_layer3);
cr = classifier_lookup(&cls, OVS_VERSION_MAX, &flow, NULL);
tnl_port_unref(cr);
@@ -242,11 +256,14 @@ map_delete(struct eth_addr mac, struct in6_addr *addr,
static void
ipdev_map_delete(struct ip_device *ip_dev, ovs_be16 tp_port, uint8_t nw_proto)
{
+ bool is_layer3 = netdev_vport_is_layer3(ip_dev->dev);
+
if (ip_dev->n_addr) {
int i;
for (i = 0; i < ip_dev->n_addr; i++) {
- map_delete(ip_dev->mac, &ip_dev->addr[i], tp_port, nw_proto);
+ map_delete(ip_dev->mac, &ip_dev->addr[i], tp_port, nw_proto,
+ is_layer3);
}
}
}
@@ -352,7 +369,12 @@ tnl_port_show(struct unixctl_conn *conn, int argc OVS_UNUSED,
}
LIST_FOR_EACH(p, node, &port_list) {
- ds_put_format(&ds, "%s (%"PRIu32")\n", p->dev_name, p->port);
+ /* A layer3 and non-layer3 tunnel port may share the same ODP port.
+ * To allow differentiation and avoid displaying otherwise
+ * duplicated ouput append " (layer3)" when showing layer-3
+ * tunnel ports. */
+ ds_put_format(&ds, "%s (%"PRIu32")%s\n", p->dev_name, p->port,
+ p->is_layer3 ? " (layer3)" : "");
}
out:
@@ -367,7 +389,8 @@ map_insert_ipdev(struct ip_device *ip_dev)
struct tnl_port *p;
LIST_FOR_EACH(p, node, &port_list) {
- map_insert_ipdev__(ip_dev, p->dev_name, p->port, p->nw_proto, p->tp_port);
+ map_insert_ipdev__(ip_dev, p->dev_name, p->port, p->nw_proto,
+ p->tp_port, p->is_layer3);
}
}
diff --git a/lib/tnl-ports.h b/lib/tnl-ports.h
index 58b048a9c63e..fb576733a0c8 100644
--- a/lib/tnl-ports.h
+++ b/lib/tnl-ports.h
@@ -27,7 +27,8 @@
odp_port_t tnl_port_map_lookup(struct flow *flow, struct flow_wildcards *wc);
void tnl_port_map_insert(odp_port_t port, ovs_be16 udp_port,
- const char dev_name[], const char type[]);
+ const char dev_name[], const char type[],
+ bool is_layer3);
void tnl_port_map_delete(ovs_be16 udp_port, const char type[]);
void tnl_port_map_insert_ipdev(const char dev[]);
diff --git a/ofproto/ofproto-dpif-rid.h b/ofproto/ofproto-dpif-rid.h
index f62227855586..8ccd68b056b2 100644
--- a/ofproto/ofproto-dpif-rid.h
+++ b/ofproto/ofproto-dpif-rid.h
@@ -99,7 +99,7 @@ struct rule;
/* Metadata for restoring pipeline context after recirculation. Helpers
* are inlined below to keep them together with the definition for easier
* updates. */
-BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
struct frozen_metadata {
/* Metadata in struct flow. */
diff --git a/ofproto/ofproto-dpif-sflow.c b/ofproto/ofproto-dpif-sflow.c
index 5ba5ec933554..e206a3c042b0 100644
--- a/ofproto/ofproto-dpif-sflow.c
+++ b/ofproto/ofproto-dpif-sflow.c
@@ -1023,6 +1023,7 @@ sflow_read_set_action(const struct nlattr *attr,
case OVS_KEY_ATTR_CT_MARK:
case OVS_KEY_ATTR_CT_LABELS:
case OVS_KEY_ATTR_UNSPEC:
+ case OVS_KEY_ATTR_NEXT_BASE_LAYER:
case __OVS_KEY_ATTR_MAX:
default:
break;
diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c
index a04e634a4972..4f7cc1ddd088 100644
--- a/ofproto/ofproto-dpif-xlate.c
+++ b/ofproto/ofproto-dpif-xlate.c
@@ -2864,7 +2864,7 @@ compose_output_action__(struct xlate_ctx *ctx, ofp_port_t ofp_port,
/* If 'struct flow' gets additional metadata, we'll need to zero it out
* before traversing a patch port. */
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 37);
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 38);
memset(&flow_tnl, 0, sizeof flow_tnl);
if (!xport) {
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index 916e0c1c4e48..6da39ff6b0c6 100644
--- a/ofproto/ofproto-dpif.c
+++ b/ofproto/ofproto-dpif.c
@@ -1323,6 +1323,8 @@ check_support(struct dpif_backer *backer)
backer->support.odp.ct_label = check_ct_label(backer);
backer->support.odp.ct_state_nat = check_ct_state_nat(backer);
+
+ backer->support.odp.next_base_layer = backer->support.tnl_push_pop;
}
static int
diff --git a/ofproto/tunnel.c b/ofproto/tunnel.c
index 97de59e0d0b7..2cad840bfa4f 100644
--- a/ofproto/tunnel.c
+++ b/ofproto/tunnel.c
@@ -26,6 +26,7 @@
#include "hash.h"
#include "openvswitch/hmap.h"
#include "netdev.h"
+#include "netdev-vport.h"
#include "odp-util.h"
#include "openvswitch/ofpbuf.h"
#include "packets.h"
@@ -192,7 +193,8 @@ tnl_port_add__(const struct ofport_dpif *ofport, const struct netdev *netdev,
const char *type;
type = netdev_get_type(netdev);
- tnl_port_map_insert(odp_port, cfg->dst_port, name, type);
+ tnl_port_map_insert(odp_port, cfg->dst_port, name, type,
+ cfg->is_layer3);
}
return true;
diff --git a/tests/tunnel-push-pop-ipv6.at b/tests/tunnel-push-pop-ipv6.at
index 3f3d5ee3ae86..0e4cb915bf55 100644
--- a/tests/tunnel-push-pop-ipv6.at
+++ b/tests/tunnel-push-pop-ipv6.at
@@ -12,6 +12,8 @@ AT_CHECK([ovs-vsctl add-port int-br t2 -- set Interface t2 type=vxlan \
options:remote_ip=2001:cafe::93 options:out_key=flow options:csum=true ofport_request=4\
-- add-port int-br t4 -- set Interface t4 type=geneve \
options:remote_ip=flow options:key=123 ofport_request=5\
+ -- add-port int-br t5 -- set Interface t5 type=gre \
+ options:remote_ip=2001:cafe::92 options:key=455 options:layer3=true ofport_request=6\
], [0])
AT_CHECK([ovs-appctl dpif/show], [0], [dnl
@@ -25,6 +27,7 @@ dummy at ovs-dummy: hit:0 missed:0
t2 2/4789: (vxlan: key=123, remote_ip=2001:cafe::92)
t3 4/4789: (vxlan: csum=true, out_key=flow, remote_ip=2001:cafe::93)
t4 5/6081: (geneve: key=123, remote_ip=flow)
+ t5 6/3: (gre: key=455, layer3=true, remote_ip=2001:cafe::92)
])
dnl First setup dummy interface IP address, then add the route
@@ -65,6 +68,7 @@ AT_CHECK([ovs-appctl tnl/ports/show |sort], [0], [dnl
Listening ports:
genev_sys_6081 (6081)
gre_sys (3)
+gre_sys (3) (layer3)
vxlan_sys_4789 (4789)
])
@@ -130,12 +134,12 @@ AT_CHECK([ovs-ofctl dump-ports int-br | grep 'port 3'], [0], [dnl
port 3: rx pkts=1, bytes=98, drop=?, errs=?, frame=?, over=?, crc=?
])
-dnl Check GRE only accepts encapsulated Ethernet frames
-AT_CHECK([ovs-appctl netdev-dummy/receive p0 'aa55aa550000001b213cab6486dd60000000006a2f402001cafe0000000000000000000000922001cafe00000000000000000000008820000800000001c8fe71d883724fbeb6f4e1494a080045000054ba200000400184861e0000011e00000200004227e75400030af3195500000000f265010000000000101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f3031323334353637'])
+dnl Check decapsulation of L3GRE packet
+AT_CHECK([ovs-appctl netdev-dummy/receive p0 'aa55aa550000001b213cab6486dd60000000005a2f402001cafe0000000000000000000000922001cafe00000000000000000000008820000800000001c745000054ba200000400184861e0000011e00000200004227e75400030af3195500000000f265010000000000101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f3031323334353637'])
ovs-appctl time/warp 1000
-AT_CHECK([ovs-ofctl dump-ports int-br | grep 'port 3'], [0], [dnl
- port 3: rx pkts=1, bytes=98, drop=?, errs=?, frame=?, over=?, crc=?
+AT_CHECK([ovs-ofctl dump-ports int-br | grep 'port 6'], [0], [dnl
+ port 6: rx pkts=1, bytes=84, drop=?, errs=?, frame=?, over=?, crc=?
])
dnl Check decapsulation of Geneve packet with options
diff --git a/tests/tunnel-push-pop.at b/tests/tunnel-push-pop.at
index 0e596f2dc8c7..f1f956d54d74 100644
--- a/tests/tunnel-push-pop.at
+++ b/tests/tunnel-push-pop.at
@@ -12,6 +12,8 @@ AT_CHECK([ovs-vsctl add-port int-br t2 -- set Interface t2 type=vxlan \
options:remote_ip=1.1.2.93 options:out_key=flow options:csum=true ofport_request=4\
-- add-port int-br t4 -- set Interface t4 type=geneve \
options:remote_ip=flow options:key=123 ofport_request=5\
+ -- add-port int-br t5 -- set Interface t5 type=gre \
+ options:remote_ip=1.1.2.92 options:key=455 options:layer3=true ofport_request=6\
], [0])
AT_CHECK([ovs-appctl dpif/show], [0], [dnl
@@ -25,6 +27,7 @@ dummy at ovs-dummy: hit:0 missed:0
t2 2/4789: (vxlan: key=123, remote_ip=1.1.2.92)
t3 4/4789: (vxlan: csum=true, out_key=flow, remote_ip=1.1.2.93)
t4 5/6081: (geneve: key=123, remote_ip=flow)
+ t5 6/3: (gre: key=455, layer3=true, remote_ip=1.1.2.92)
])
dnl First setup dummy interface IP address, then add the route
@@ -70,6 +73,7 @@ AT_CHECK([ovs-appctl tnl/ports/show |sort], [0], [dnl
Listening ports:
genev_sys_6081 (6081)
gre_sys (3)
+gre_sys (3) (layer3)
vxlan_sys_4789 (4789)
])
@@ -108,8 +112,14 @@ AT_CHECK([tail -1 stdout], [0],
dnl Check GRE tunnel push
AT_CHECK([ovs-ofctl add-flow int-br action=3])
AT_CHECK([ovs-appctl ofproto/trace ovs-dummy 'in_port(2),eth(src=f8:bc:12:44:34:b6,dst=aa:55:aa:55:00:00),eth_type(0x0800),ipv4(src=1.1.3.88,dst=1.1.3.112,proto=47,tos=0,ttl=64,frag=no)'], [0], [stdout])
+AT_CHECK([tail -1 stdout], [0], [Datapath actions: tnl_push(tnl_port(3),header(size=42,type=3,eth(dst=f8:bc:12:44:34:b6,src=aa:55:aa:55:00:00,dl_type=0x0800),ipv4(src=1.1.2.88,dst=1.1.2.92,proto=47,tos=0,ttl=64,frag=0x4000),gre((flags=0x2000,proto=0x6558),key=0x1c8)),out_port(100))
+])
+
+dnl Check L3GRE tunnel push
+AT_CHECK([ovs-ofctl add-flow int-br action=6])
+AT_CHECK([ovs-appctl ofproto/trace ovs-dummy 'in_port(2),eth(src=f8:bc:12:44:34:b6,dst=aa:55:aa:55:00:00),eth_type(0x0800),ipv4(src=1.1.3.88,dst=1.1.3.112,proto=47,tos=0,ttl=64,frag=no)'], [0], [stdout])
AT_CHECK([tail -1 stdout], [0],
- [Datapath actions: tnl_push(tnl_port(3),header(size=42,type=3,eth(dst=f8:bc:12:44:34:b6,src=aa:55:aa:55:00:00,dl_type=0x0800),ipv4(src=1.1.2.88,dst=1.1.2.92,proto=47,tos=0,ttl=64,frag=0x4000),gre((flags=0x2000,proto=0x6558),key=0x1c8)),out_port(100))
+ [Datapath actions: pop_eth,tnl_push(tnl_port(3),header(size=42,type=3,eth(dst=f8:bc:12:44:34:b6,src=aa:55:aa:55:00:00,dl_type=0x0800),ipv4(src=1.1.2.88,dst=1.1.2.92,proto=47,tos=0,ttl=64,frag=0x4000),gre((flags=0x2000,proto=0x800),key=0x1c7)),out_port(100))
])
dnl Check Geneve tunnel push
@@ -135,12 +145,20 @@ AT_CHECK([ovs-ofctl dump-ports int-br | grep 'port 3'], [0], [dnl
port 3: rx pkts=1, bytes=98, drop=?, errs=?, frame=?, over=?, crc=?
])
-dnl Check GRE only accepts encapsulated Ethernet frames
-AT_CHECK([ovs-appctl netdev-dummy/receive p0 'aa55aa550000001b213cab6408004500007e79464000402fba550101025c0101025820000800000001c8fe71d883724fbeb6f4e1494a080045000054ba200000400184861e0000011e00000200004227e75400030af3195500000000f265010000000000101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f3031323334353637'])
+dnl Check decapsulation of L3GRE packet
+AT_CHECK([ovs-appctl netdev-dummy/receive p0 'aa55aa550000001b213cab6408004500007079464000402fba630101025c0101025820000800000001c745000054ba200000400184861e0000011e00000200004227e75400030af3195500000000f265010000000000101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f3031323334353637'])
ovs-appctl time/warp 1000
-AT_CHECK([ovs-ofctl dump-ports int-br | grep 'port 3'], [0], [dnl
+AT_CHECK([ovs-ofctl dump-ports int-br | grep 'port 6'], [0], [dnl
+ port 6: rx pkts=1, bytes=84, drop=?, errs=?, frame=?, over=?, crc=?
+])
+
+dnl Check GREL3 only accepts non-fragmented packets?
+AT_CHECK([ovs-appctl netdev-dummy/receive p0 'aa55aa550000001b213cab6408004500007e79464000402fba550101025c0101025820000800000001c7fe71d883724fbeb6f4e1494a080045000054ba200000400184861e0000011e00000200004227e75400030af3195500000000f265010000000000101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f3031323334353637'])
+
+AT_CHECK([ovs-ofctl dump-ports int-br | grep 'port [[36]]' | sort], [0], [dnl
port 3: rx pkts=1, bytes=98, drop=?, errs=?, frame=?, over=?, crc=?
+ port 6: rx pkts=1, bytes=84, drop=?, errs=?, frame=?, over=?, crc=?
])
dnl Check decapsulation of Geneve packet with options
diff --git a/vswitchd/vswitch.xml b/vswitchd/vswitch.xml
index 99c14766c46b..ed48ac340843 100644
--- a/vswitchd/vswitch.xml
+++ b/vswitchd/vswitch.xml
@@ -2244,6 +2244,19 @@
</group>
+ <group title="Tunnel Options: gre only">
+ <p>
+ <code>gre</code> interfaces support these options.
+ </p>
+
+ <column name="options" key="layer3" type='{"type": "boolean"}'>
+ <p>
+ Optional. Packets are sent and received without an Ethernet
+ header present.
+ </p>
+ </column>
+ </group>
+
<group title="Tunnel Options: gre, geneve, and vxlan">
<p>
<code>gre</code>, <code>geneve</code>, and
--
1.8.3.1
More information about the dev
mailing list