* [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP
@ 2019-12-17 21:57 John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 1/9] nfp: flower: pass flow rule pointer directly to match functions John Hurley
` (9 more replies)
0 siblings, 10 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
Hi,
The following patches add support for IPv6 tunnel offload to the NFP
driver.
Patches 1-2 do some code tidy up and prepare existing code for reuse in
IPv6 tunnels.
Patches 3-4 handle IPv6 tunnel decap (match) rules.
Patches 5-8 handle encap (action) rules.
Patch 9 adds IPv6 support to the merge and pre-tunnel rule functions.
v1->v2:
- fix compiler warning when building without CONFIG_IPV6 set -
Jakub Kicinski (patch 7)
John Hurley (9):
nfp: flower: pass flow rule pointer directly to match functions
nfp: flower: move udp tunnel key match compilation to helper function
nfp: flower: compile match for IPv6 tunnels
nfp: flower: offload list of IPv6 tunnel endpoint addresses
nfp: flower: modify pre-tunnel and set tunnel action for ipv6
nfp: flower: handle ipv6 tunnel no neigh request
nfp: flower: handle notifiers for ipv6 route changes
nfp: flower: support ipv6 tunnel keep-alive messages from fw
nfp: flower: update flow merge code to support IPv6 tunnels
drivers/net/ethernet/netronome/nfp/flower/action.c | 65 ++-
drivers/net/ethernet/netronome/nfp/flower/cmsg.c | 11 +-
drivers/net/ethernet/netronome/nfp/flower/cmsg.h | 106 ++++-
drivers/net/ethernet/netronome/nfp/flower/main.h | 38 +-
drivers/net/ethernet/netronome/nfp/flower/match.c | 260 +++++++----
.../net/ethernet/netronome/nfp/flower/offload.c | 144 ++++--
.../ethernet/netronome/nfp/flower/tunnel_conf.c | 498 ++++++++++++++++++---
7 files changed, 893 insertions(+), 229 deletions(-)
--
2.7.4
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH net-next v2 1/9] nfp: flower: pass flow rule pointer directly to match functions
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 2/9] nfp: flower: move udp tunnel key match compilation to helper function John Hurley
` (8 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
In kernel 5.1, the flow offload API was introduced along with a helper
function to extract the flow_rule from the TC offload struct. Each of the
match helper functions are passed the offload struct and extract the flow
rule to a local variable.
Simplify the code while also removing the extra compat and local variable
calls by extracting the rule once in the main match handler, and passing
a reference to the rule direct to each helper.
This patch does not change driver functionality.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/match.c | 76 ++++++++---------------
1 file changed, 27 insertions(+), 49 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/match.c b/drivers/net/ethernet/netronome/nfp/flower/match.c
index 9cc3ba1..1079f37 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/match.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/match.c
@@ -10,9 +10,8 @@
static void
nfp_flower_compile_meta_tci(struct nfp_flower_meta_tci *ext,
struct nfp_flower_meta_tci *msk,
- struct flow_cls_offload *flow, u8 key_type)
+ struct flow_rule *rule, u8 key_type)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
u16 tmp_tci;
memset(ext, 0, sizeof(struct nfp_flower_meta_tci));
@@ -77,11 +76,8 @@ nfp_flower_compile_port(struct nfp_flower_in_port *frame, u32 cmsg_port,
static void
nfp_flower_compile_mac(struct nfp_flower_mac_mpls *ext,
- struct nfp_flower_mac_mpls *msk,
- struct flow_cls_offload *flow)
+ struct nfp_flower_mac_mpls *msk, struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
memset(ext, 0, sizeof(struct nfp_flower_mac_mpls));
memset(msk, 0, sizeof(struct nfp_flower_mac_mpls));
@@ -130,10 +126,8 @@ nfp_flower_compile_mac(struct nfp_flower_mac_mpls *ext,
static void
nfp_flower_compile_tport(struct nfp_flower_tp_ports *ext,
struct nfp_flower_tp_ports *msk,
- struct flow_cls_offload *flow)
+ struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
memset(ext, 0, sizeof(struct nfp_flower_tp_ports));
memset(msk, 0, sizeof(struct nfp_flower_tp_ports));
@@ -150,11 +144,8 @@ nfp_flower_compile_tport(struct nfp_flower_tp_ports *ext,
static void
nfp_flower_compile_ip_ext(struct nfp_flower_ip_ext *ext,
- struct nfp_flower_ip_ext *msk,
- struct flow_cls_offload *flow)
+ struct nfp_flower_ip_ext *msk, struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_BASIC)) {
struct flow_match_basic match;
@@ -224,10 +215,8 @@ nfp_flower_compile_ip_ext(struct nfp_flower_ip_ext *ext,
static void
nfp_flower_compile_ipv4(struct nfp_flower_ipv4 *ext,
- struct nfp_flower_ipv4 *msk,
- struct flow_cls_offload *flow)
+ struct nfp_flower_ipv4 *msk, struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
struct flow_match_ipv4_addrs match;
memset(ext, 0, sizeof(struct nfp_flower_ipv4));
@@ -241,16 +230,13 @@ nfp_flower_compile_ipv4(struct nfp_flower_ipv4 *ext,
msk->ipv4_dst = match.mask->dst;
}
- nfp_flower_compile_ip_ext(&ext->ip_ext, &msk->ip_ext, flow);
+ nfp_flower_compile_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
}
static void
nfp_flower_compile_ipv6(struct nfp_flower_ipv6 *ext,
- struct nfp_flower_ipv6 *msk,
- struct flow_cls_offload *flow)
+ struct nfp_flower_ipv6 *msk, struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
memset(ext, 0, sizeof(struct nfp_flower_ipv6));
memset(msk, 0, sizeof(struct nfp_flower_ipv6));
@@ -264,16 +250,15 @@ nfp_flower_compile_ipv6(struct nfp_flower_ipv6 *ext,
msk->ipv6_dst = match.mask->dst;
}
- nfp_flower_compile_ip_ext(&ext->ip_ext, &msk->ip_ext, flow);
+ nfp_flower_compile_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
}
static int
-nfp_flower_compile_geneve_opt(void *ext, void *msk,
- struct flow_cls_offload *flow)
+nfp_flower_compile_geneve_opt(void *ext, void *msk, struct flow_rule *rule)
{
struct flow_match_enc_opts match;
- flow_rule_match_enc_opts(flow->rule, &match);
+ flow_rule_match_enc_opts(rule, &match);
memcpy(ext, match.key->data, match.key->len);
memcpy(msk, match.mask->data, match.mask->len);
@@ -283,10 +268,8 @@ nfp_flower_compile_geneve_opt(void *ext, void *msk,
static void
nfp_flower_compile_tun_ipv4_addrs(struct nfp_flower_tun_ipv4 *ext,
struct nfp_flower_tun_ipv4 *msk,
- struct flow_cls_offload *flow)
+ struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS)) {
struct flow_match_ipv4_addrs match;
@@ -301,10 +284,8 @@ nfp_flower_compile_tun_ipv4_addrs(struct nfp_flower_tun_ipv4 *ext,
static void
nfp_flower_compile_tun_ip_ext(struct nfp_flower_tun_ip_ext *ext,
struct nfp_flower_tun_ip_ext *msk,
- struct flow_cls_offload *flow)
+ struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IP)) {
struct flow_match_ip match;
@@ -319,10 +300,8 @@ nfp_flower_compile_tun_ip_ext(struct nfp_flower_tun_ip_ext *ext,
static void
nfp_flower_compile_ipv4_gre_tun(struct nfp_flower_ipv4_gre_tun *ext,
struct nfp_flower_ipv4_gre_tun *msk,
- struct flow_cls_offload *flow)
+ struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
memset(ext, 0, sizeof(struct nfp_flower_ipv4_gre_tun));
memset(msk, 0, sizeof(struct nfp_flower_ipv4_gre_tun));
@@ -341,17 +320,15 @@ nfp_flower_compile_ipv4_gre_tun(struct nfp_flower_ipv4_gre_tun *ext,
msk->tun_flags = cpu_to_be16(NFP_FL_GRE_FLAG_KEY);
}
- nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, flow);
- nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, flow);
+ nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, rule);
+ nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
}
static void
nfp_flower_compile_ipv4_udp_tun(struct nfp_flower_ipv4_udp_tun *ext,
struct nfp_flower_ipv4_udp_tun *msk,
- struct flow_cls_offload *flow)
+ struct flow_rule *rule)
{
- struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
-
memset(ext, 0, sizeof(struct nfp_flower_ipv4_udp_tun));
memset(msk, 0, sizeof(struct nfp_flower_ipv4_udp_tun));
@@ -366,8 +343,8 @@ nfp_flower_compile_ipv4_udp_tun(struct nfp_flower_ipv4_udp_tun *ext,
msk->tun_id = cpu_to_be32(temp_vni);
}
- nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, flow);
- nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, flow);
+ nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, rule);
+ nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
}
int nfp_flower_compile_flow_match(struct nfp_app *app,
@@ -378,6 +355,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
enum nfp_flower_tun_type tun_type,
struct netlink_ext_ack *extack)
{
+ struct flow_rule *rule = flow_cls_offload_flow_rule(flow);
u32 port_id;
int err;
u8 *ext;
@@ -393,7 +371,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
nfp_flower_compile_meta_tci((struct nfp_flower_meta_tci *)ext,
(struct nfp_flower_meta_tci *)msk,
- flow, key_ls->key_layer);
+ rule, key_ls->key_layer);
ext += sizeof(struct nfp_flower_meta_tci);
msk += sizeof(struct nfp_flower_meta_tci);
@@ -425,7 +403,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
if (NFP_FLOWER_LAYER_MAC & key_ls->key_layer) {
nfp_flower_compile_mac((struct nfp_flower_mac_mpls *)ext,
(struct nfp_flower_mac_mpls *)msk,
- flow);
+ rule);
ext += sizeof(struct nfp_flower_mac_mpls);
msk += sizeof(struct nfp_flower_mac_mpls);
}
@@ -433,7 +411,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
if (NFP_FLOWER_LAYER_TP & key_ls->key_layer) {
nfp_flower_compile_tport((struct nfp_flower_tp_ports *)ext,
(struct nfp_flower_tp_ports *)msk,
- flow);
+ rule);
ext += sizeof(struct nfp_flower_tp_ports);
msk += sizeof(struct nfp_flower_tp_ports);
}
@@ -441,7 +419,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
if (NFP_FLOWER_LAYER_IPV4 & key_ls->key_layer) {
nfp_flower_compile_ipv4((struct nfp_flower_ipv4 *)ext,
(struct nfp_flower_ipv4 *)msk,
- flow);
+ rule);
ext += sizeof(struct nfp_flower_ipv4);
msk += sizeof(struct nfp_flower_ipv4);
}
@@ -449,7 +427,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
if (NFP_FLOWER_LAYER_IPV6 & key_ls->key_layer) {
nfp_flower_compile_ipv6((struct nfp_flower_ipv6 *)ext,
(struct nfp_flower_ipv6 *)msk,
- flow);
+ rule);
ext += sizeof(struct nfp_flower_ipv6);
msk += sizeof(struct nfp_flower_ipv6);
}
@@ -457,7 +435,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_GRE) {
__be32 tun_dst;
- nfp_flower_compile_ipv4_gre_tun((void *)ext, (void *)msk, flow);
+ nfp_flower_compile_ipv4_gre_tun((void *)ext, (void *)msk, rule);
tun_dst = ((struct nfp_flower_ipv4_gre_tun *)ext)->ipv4.dst;
ext += sizeof(struct nfp_flower_ipv4_gre_tun);
msk += sizeof(struct nfp_flower_ipv4_gre_tun);
@@ -473,7 +451,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
key_ls->key_layer_two & NFP_FLOWER_LAYER2_GENEVE) {
__be32 tun_dst;
- nfp_flower_compile_ipv4_udp_tun((void *)ext, (void *)msk, flow);
+ nfp_flower_compile_ipv4_udp_tun((void *)ext, (void *)msk, rule);
tun_dst = ((struct nfp_flower_ipv4_udp_tun *)ext)->ipv4.dst;
ext += sizeof(struct nfp_flower_ipv4_udp_tun);
msk += sizeof(struct nfp_flower_ipv4_udp_tun);
@@ -485,7 +463,7 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
nfp_tunnel_add_ipv4_off(app, tun_dst);
if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_GENEVE_OP) {
- err = nfp_flower_compile_geneve_opt(ext, msk, flow);
+ err = nfp_flower_compile_geneve_opt(ext, msk, rule);
if (err)
return err;
}
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 2/9] nfp: flower: move udp tunnel key match compilation to helper function
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 1/9] nfp: flower: pass flow rule pointer directly to match functions John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 3/9] nfp: flower: compile match for IPv6 tunnels John Hurley
` (7 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
IPv4 UDP and GRE tunnel match rule compile helpers share functions for
compiling fields such as IP addresses. However, they handle fields such
tunnel IDs differently.
Create new helper functions for compiling GRE and UDP tunnel key data.
This is in preparation for supporting IPv6 tunnels where these new
functions can be reused.
This patch does not change functionality.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/match.c | 57 ++++++++++++++---------
1 file changed, 35 insertions(+), 22 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/match.c b/drivers/net/ethernet/netronome/nfp/flower/match.c
index 1079f37..1cf8eae 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/match.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/match.c
@@ -298,6 +298,38 @@ nfp_flower_compile_tun_ip_ext(struct nfp_flower_tun_ip_ext *ext,
}
static void
+nfp_flower_compile_tun_udp_key(__be32 *key, __be32 *key_msk,
+ struct flow_rule *rule)
+{
+ if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID)) {
+ struct flow_match_enc_keyid match;
+ u32 vni;
+
+ flow_rule_match_enc_keyid(rule, &match);
+ vni = be32_to_cpu(match.key->keyid) << NFP_FL_TUN_VNI_OFFSET;
+ *key = cpu_to_be32(vni);
+ vni = be32_to_cpu(match.mask->keyid) << NFP_FL_TUN_VNI_OFFSET;
+ *key_msk = cpu_to_be32(vni);
+ }
+}
+
+static void
+nfp_flower_compile_tun_gre_key(__be32 *key, __be32 *key_msk, __be16 *flags,
+ __be16 *flags_msk, struct flow_rule *rule)
+{
+ if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID)) {
+ struct flow_match_enc_keyid match;
+
+ flow_rule_match_enc_keyid(rule, &match);
+ *key = match.key->keyid;
+ *key_msk = match.mask->keyid;
+
+ *flags = cpu_to_be16(NFP_FL_GRE_FLAG_KEY);
+ *flags_msk = cpu_to_be16(NFP_FL_GRE_FLAG_KEY);
+ }
+}
+
+static void
nfp_flower_compile_ipv4_gre_tun(struct nfp_flower_ipv4_gre_tun *ext,
struct nfp_flower_ipv4_gre_tun *msk,
struct flow_rule *rule)
@@ -309,19 +341,10 @@ nfp_flower_compile_ipv4_gre_tun(struct nfp_flower_ipv4_gre_tun *ext,
ext->ethertype = cpu_to_be16(ETH_P_TEB);
msk->ethertype = cpu_to_be16(~0);
- if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID)) {
- struct flow_match_enc_keyid match;
-
- flow_rule_match_enc_keyid(rule, &match);
- ext->tun_key = match.key->keyid;
- msk->tun_key = match.mask->keyid;
-
- ext->tun_flags = cpu_to_be16(NFP_FL_GRE_FLAG_KEY);
- msk->tun_flags = cpu_to_be16(NFP_FL_GRE_FLAG_KEY);
- }
-
nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, rule);
nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
+ nfp_flower_compile_tun_gre_key(&ext->tun_key, &msk->tun_key,
+ &ext->tun_flags, &msk->tun_flags, rule);
}
static void
@@ -332,19 +355,9 @@ nfp_flower_compile_ipv4_udp_tun(struct nfp_flower_ipv4_udp_tun *ext,
memset(ext, 0, sizeof(struct nfp_flower_ipv4_udp_tun));
memset(msk, 0, sizeof(struct nfp_flower_ipv4_udp_tun));
- if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID)) {
- struct flow_match_enc_keyid match;
- u32 temp_vni;
-
- flow_rule_match_enc_keyid(rule, &match);
- temp_vni = be32_to_cpu(match.key->keyid) << NFP_FL_TUN_VNI_OFFSET;
- ext->tun_id = cpu_to_be32(temp_vni);
- temp_vni = be32_to_cpu(match.mask->keyid) << NFP_FL_TUN_VNI_OFFSET;
- msk->tun_id = cpu_to_be32(temp_vni);
- }
-
nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, rule);
nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
+ nfp_flower_compile_tun_udp_key(&ext->tun_id, &msk->tun_id, rule);
}
int nfp_flower_compile_flow_match(struct nfp_app *app,
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 3/9] nfp: flower: compile match for IPv6 tunnels
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 1/9] nfp: flower: pass flow rule pointer directly to match functions John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 2/9] nfp: flower: move udp tunnel key match compilation to helper function John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 4/9] nfp: flower: offload list of IPv6 tunnel endpoint addresses John Hurley
` (6 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
IPv6 tunnel matches are now supported by firmware. Modify the NFP driver
to compile these match rules. IPv6 matches are handled similar to IPv4
tunnels with the difference the address length. The type of tunnel is
indicated by the same bitmap that is used in IPv4 with an extra bit
signifying that the IPv6 variation should be used.
Only compile IPv6 tunnel matches when the fw features symbol indicated
that they are compatible with the currently loaded fw.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/cmsg.h | 83 +++++++++++++++
drivers/net/ethernet/netronome/nfp/flower/main.h | 1 +
drivers/net/ethernet/netronome/nfp/flower/match.c | 111 ++++++++++++++++-----
.../net/ethernet/netronome/nfp/flower/offload.c | 102 ++++++++++++++-----
4 files changed, 246 insertions(+), 51 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
index 7eb2ec8..8d12805 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
@@ -26,6 +26,7 @@
#define NFP_FLOWER_LAYER2_GRE BIT(0)
#define NFP_FLOWER_LAYER2_GENEVE BIT(5)
#define NFP_FLOWER_LAYER2_GENEVE_OP BIT(6)
+#define NFP_FLOWER_LAYER2_TUN_IPV6 BIT(7)
#define NFP_FLOWER_MASK_VLAN_PRIO GENMASK(15, 13)
#define NFP_FLOWER_MASK_VLAN_PRESENT BIT(12)
@@ -63,6 +64,7 @@
#define NFP_FL_MAX_GENEVE_OPT_ACT 32
#define NFP_FL_MAX_GENEVE_OPT_CNT 64
#define NFP_FL_MAX_GENEVE_OPT_KEY 32
+#define NFP_FL_MAX_GENEVE_OPT_KEY_V6 8
/* Action opcodes */
#define NFP_FL_ACTION_OPCODE_OUTPUT 0
@@ -387,6 +389,11 @@ struct nfp_flower_tun_ipv4 {
__be32 dst;
};
+struct nfp_flower_tun_ipv6 {
+ struct in6_addr src;
+ struct in6_addr dst;
+};
+
struct nfp_flower_tun_ip_ext {
u8 tos;
u8 ttl;
@@ -416,6 +423,42 @@ struct nfp_flower_ipv4_udp_tun {
__be32 tun_id;
};
+/* Flow Frame IPv6 UDP TUNNEL --> Tunnel details (11W/44B)
+ * -----------------------------------------------------------------
+ * 3 2 1
+ * 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 31 - 0 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 63 - 32 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 95 - 64 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 127 - 96 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 31 - 0 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 63 - 32 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 95 - 64 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 127 - 96 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | Reserved | tos | ttl |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | Reserved |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | VNI | Reserved |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ */
+struct nfp_flower_ipv6_udp_tun {
+ struct nfp_flower_tun_ipv6 ipv6;
+ __be16 reserved1;
+ struct nfp_flower_tun_ip_ext ip_ext;
+ __be32 reserved2;
+ __be32 tun_id;
+};
+
/* Flow Frame GRE TUNNEL --> Tunnel details (6W/24B)
* -----------------------------------------------------------------
* 3 2 1
@@ -445,6 +488,46 @@ struct nfp_flower_ipv4_gre_tun {
__be32 reserved2;
};
+/* Flow Frame GRE TUNNEL V6 --> Tunnel details (12W/48B)
+ * -----------------------------------------------------------------
+ * 3 2 1
+ * 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 31 - 0 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 63 - 32 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 95 - 64 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_src, 127 - 96 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 31 - 0 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 63 - 32 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 95 - 64 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | ipv6_addr_dst, 127 - 96 |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | tun_flags | tos | ttl |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | Reserved | Ethertype |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | Key |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | Reserved |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ */
+struct nfp_flower_ipv6_gre_tun {
+ struct nfp_flower_tun_ipv6 ipv6;
+ __be16 tun_flags;
+ struct nfp_flower_tun_ip_ext ip_ext;
+ __be16 reserved1;
+ __be16 ethertype;
+ __be32 tun_key;
+ __be32 reserved2;
+};
+
struct nfp_flower_geneve_options {
u8 data[NFP_FL_MAX_GENEVE_OPT_KEY];
};
diff --git a/drivers/net/ethernet/netronome/nfp/flower/main.h b/drivers/net/ethernet/netronome/nfp/flower/main.h
index 31d9459..386d629 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/main.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/main.h
@@ -43,6 +43,7 @@ struct nfp_app;
#define NFP_FL_FEATS_VF_RLIM BIT(4)
#define NFP_FL_FEATS_FLOW_MOD BIT(5)
#define NFP_FL_FEATS_PRE_TUN_RULES BIT(6)
+#define NFP_FL_FEATS_IPV6_TUN BIT(7)
#define NFP_FL_FEATS_FLOW_MERGE BIT(30)
#define NFP_FL_FEATS_LAG BIT(31)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/match.c b/drivers/net/ethernet/netronome/nfp/flower/match.c
index 1cf8eae..2410ead 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/match.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/match.c
@@ -282,6 +282,22 @@ nfp_flower_compile_tun_ipv4_addrs(struct nfp_flower_tun_ipv4 *ext,
}
static void
+nfp_flower_compile_tun_ipv6_addrs(struct nfp_flower_tun_ipv6 *ext,
+ struct nfp_flower_tun_ipv6 *msk,
+ struct flow_rule *rule)
+{
+ if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS)) {
+ struct flow_match_ipv6_addrs match;
+
+ flow_rule_match_enc_ipv6_addrs(rule, &match);
+ ext->src = match.key->src;
+ ext->dst = match.key->dst;
+ msk->src = match.mask->src;
+ msk->dst = match.mask->dst;
+ }
+}
+
+static void
nfp_flower_compile_tun_ip_ext(struct nfp_flower_tun_ip_ext *ext,
struct nfp_flower_tun_ip_ext *msk,
struct flow_rule *rule)
@@ -360,6 +376,37 @@ nfp_flower_compile_ipv4_udp_tun(struct nfp_flower_ipv4_udp_tun *ext,
nfp_flower_compile_tun_udp_key(&ext->tun_id, &msk->tun_id, rule);
}
+static void
+nfp_flower_compile_ipv6_udp_tun(struct nfp_flower_ipv6_udp_tun *ext,
+ struct nfp_flower_ipv6_udp_tun *msk,
+ struct flow_rule *rule)
+{
+ memset(ext, 0, sizeof(struct nfp_flower_ipv6_udp_tun));
+ memset(msk, 0, sizeof(struct nfp_flower_ipv6_udp_tun));
+
+ nfp_flower_compile_tun_ipv6_addrs(&ext->ipv6, &msk->ipv6, rule);
+ nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
+ nfp_flower_compile_tun_udp_key(&ext->tun_id, &msk->tun_id, rule);
+}
+
+static void
+nfp_flower_compile_ipv6_gre_tun(struct nfp_flower_ipv6_gre_tun *ext,
+ struct nfp_flower_ipv6_gre_tun *msk,
+ struct flow_rule *rule)
+{
+ memset(ext, 0, sizeof(struct nfp_flower_ipv6_gre_tun));
+ memset(msk, 0, sizeof(struct nfp_flower_ipv6_gre_tun));
+
+ /* NVGRE is the only supported GRE tunnel type */
+ ext->ethertype = cpu_to_be16(ETH_P_TEB);
+ msk->ethertype = cpu_to_be16(~0);
+
+ nfp_flower_compile_tun_ipv6_addrs(&ext->ipv6, &msk->ipv6, rule);
+ nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, rule);
+ nfp_flower_compile_tun_gre_key(&ext->tun_key, &msk->tun_key,
+ &ext->tun_flags, &msk->tun_flags, rule);
+}
+
int nfp_flower_compile_flow_match(struct nfp_app *app,
struct flow_cls_offload *flow,
struct nfp_fl_key_ls *key_ls,
@@ -446,34 +493,50 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
}
if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_GRE) {
- __be32 tun_dst;
-
- nfp_flower_compile_ipv4_gre_tun((void *)ext, (void *)msk, rule);
- tun_dst = ((struct nfp_flower_ipv4_gre_tun *)ext)->ipv4.dst;
- ext += sizeof(struct nfp_flower_ipv4_gre_tun);
- msk += sizeof(struct nfp_flower_ipv4_gre_tun);
-
- /* Store the tunnel destination in the rule data.
- * This must be present and be an exact match.
- */
- nfp_flow->nfp_tun_ipv4_addr = tun_dst;
- nfp_tunnel_add_ipv4_off(app, tun_dst);
+ if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_TUN_IPV6) {
+ nfp_flower_compile_ipv6_gre_tun((void *)ext,
+ (void *)msk, rule);
+ ext += sizeof(struct nfp_flower_ipv6_gre_tun);
+ msk += sizeof(struct nfp_flower_ipv6_gre_tun);
+ } else {
+ __be32 dst;
+
+ nfp_flower_compile_ipv4_gre_tun((void *)ext,
+ (void *)msk, rule);
+ dst = ((struct nfp_flower_ipv4_gre_tun *)ext)->ipv4.dst;
+ ext += sizeof(struct nfp_flower_ipv4_gre_tun);
+ msk += sizeof(struct nfp_flower_ipv4_gre_tun);
+
+ /* Store the tunnel destination in the rule data.
+ * This must be present and be an exact match.
+ */
+ nfp_flow->nfp_tun_ipv4_addr = dst;
+ nfp_tunnel_add_ipv4_off(app, dst);
+ }
}
if (key_ls->key_layer & NFP_FLOWER_LAYER_VXLAN ||
key_ls->key_layer_two & NFP_FLOWER_LAYER2_GENEVE) {
- __be32 tun_dst;
-
- nfp_flower_compile_ipv4_udp_tun((void *)ext, (void *)msk, rule);
- tun_dst = ((struct nfp_flower_ipv4_udp_tun *)ext)->ipv4.dst;
- ext += sizeof(struct nfp_flower_ipv4_udp_tun);
- msk += sizeof(struct nfp_flower_ipv4_udp_tun);
-
- /* Store the tunnel destination in the rule data.
- * This must be present and be an exact match.
- */
- nfp_flow->nfp_tun_ipv4_addr = tun_dst;
- nfp_tunnel_add_ipv4_off(app, tun_dst);
+ if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_TUN_IPV6) {
+ nfp_flower_compile_ipv6_udp_tun((void *)ext,
+ (void *)msk, rule);
+ ext += sizeof(struct nfp_flower_ipv6_udp_tun);
+ msk += sizeof(struct nfp_flower_ipv6_udp_tun);
+ } else {
+ __be32 dst;
+
+ nfp_flower_compile_ipv4_udp_tun((void *)ext,
+ (void *)msk, rule);
+ dst = ((struct nfp_flower_ipv4_udp_tun *)ext)->ipv4.dst;
+ ext += sizeof(struct nfp_flower_ipv4_udp_tun);
+ msk += sizeof(struct nfp_flower_ipv4_udp_tun);
+
+ /* Store the tunnel destination in the rule data.
+ * This must be present and be an exact match.
+ */
+ nfp_flow->nfp_tun_ipv4_addr = dst;
+ nfp_tunnel_add_ipv4_off(app, dst);
+ }
if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_GENEVE_OP) {
err = nfp_flower_compile_geneve_opt(ext, msk, rule);
diff --git a/drivers/net/ethernet/netronome/nfp/flower/offload.c b/drivers/net/ethernet/netronome/nfp/flower/offload.c
index 987ae22..0997350 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/offload.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/offload.c
@@ -54,6 +54,10 @@
(BIT(FLOW_DISSECTOR_KEY_ENC_CONTROL) | \
BIT(FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS))
+#define NFP_FLOWER_WHITELIST_TUN_DISSECTOR_V6_R \
+ (BIT(FLOW_DISSECTOR_KEY_ENC_CONTROL) | \
+ BIT(FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS))
+
#define NFP_FLOWER_MERGE_FIELDS \
(NFP_FLOWER_LAYER_PORT | \
NFP_FLOWER_LAYER_MAC | \
@@ -146,10 +150,11 @@ static bool nfp_flower_check_higher_than_l3(struct flow_cls_offload *f)
static int
nfp_flower_calc_opt_layer(struct flow_dissector_key_enc_opts *enc_opts,
- u32 *key_layer_two, int *key_size,
+ u32 *key_layer_two, int *key_size, bool ipv6,
struct netlink_ext_ack *extack)
{
- if (enc_opts->len > NFP_FL_MAX_GENEVE_OPT_KEY) {
+ if (enc_opts->len > NFP_FL_MAX_GENEVE_OPT_KEY ||
+ (ipv6 && enc_opts->len > NFP_FL_MAX_GENEVE_OPT_KEY_V6)) {
NL_SET_ERR_MSG_MOD(extack, "unsupported offload: geneve options exceed maximum length");
return -EOPNOTSUPP;
}
@@ -167,7 +172,7 @@ nfp_flower_calc_udp_tun_layer(struct flow_dissector_key_ports *enc_ports,
struct flow_dissector_key_enc_opts *enc_op,
u32 *key_layer_two, u8 *key_layer, int *key_size,
struct nfp_flower_priv *priv,
- enum nfp_flower_tun_type *tun_type,
+ enum nfp_flower_tun_type *tun_type, bool ipv6,
struct netlink_ext_ack *extack)
{
int err;
@@ -176,7 +181,15 @@ nfp_flower_calc_udp_tun_layer(struct flow_dissector_key_ports *enc_ports,
case htons(IANA_VXLAN_UDP_PORT):
*tun_type = NFP_FL_TUNNEL_VXLAN;
*key_layer |= NFP_FLOWER_LAYER_VXLAN;
- *key_size += sizeof(struct nfp_flower_ipv4_udp_tun);
+
+ if (ipv6) {
+ *key_layer |= NFP_FLOWER_LAYER_EXT_META;
+ *key_size += sizeof(struct nfp_flower_ext_meta);
+ *key_layer_two |= NFP_FLOWER_LAYER2_TUN_IPV6;
+ *key_size += sizeof(struct nfp_flower_ipv6_udp_tun);
+ } else {
+ *key_size += sizeof(struct nfp_flower_ipv4_udp_tun);
+ }
if (enc_op) {
NL_SET_ERR_MSG_MOD(extack, "unsupported offload: encap options not supported on vxlan tunnels");
@@ -192,7 +205,13 @@ nfp_flower_calc_udp_tun_layer(struct flow_dissector_key_ports *enc_ports,
*key_layer |= NFP_FLOWER_LAYER_EXT_META;
*key_size += sizeof(struct nfp_flower_ext_meta);
*key_layer_two |= NFP_FLOWER_LAYER2_GENEVE;
- *key_size += sizeof(struct nfp_flower_ipv4_udp_tun);
+
+ if (ipv6) {
+ *key_layer_two |= NFP_FLOWER_LAYER2_TUN_IPV6;
+ *key_size += sizeof(struct nfp_flower_ipv6_udp_tun);
+ } else {
+ *key_size += sizeof(struct nfp_flower_ipv4_udp_tun);
+ }
if (!enc_op)
break;
@@ -200,8 +219,8 @@ nfp_flower_calc_udp_tun_layer(struct flow_dissector_key_ports *enc_ports,
NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support geneve option offload");
return -EOPNOTSUPP;
}
- err = nfp_flower_calc_opt_layer(enc_op, key_layer_two,
- key_size, extack);
+ err = nfp_flower_calc_opt_layer(enc_op, key_layer_two, key_size,
+ ipv6, extack);
if (err)
return err;
break;
@@ -237,6 +256,8 @@ nfp_flower_calculate_key_layers(struct nfp_app *app,
/* If any tun dissector is used then the required set must be used. */
if (dissector->used_keys & NFP_FLOWER_WHITELIST_TUN_DISSECTOR &&
+ (dissector->used_keys & NFP_FLOWER_WHITELIST_TUN_DISSECTOR_V6_R)
+ != NFP_FLOWER_WHITELIST_TUN_DISSECTOR_V6_R &&
(dissector->used_keys & NFP_FLOWER_WHITELIST_TUN_DISSECTOR_R)
!= NFP_FLOWER_WHITELIST_TUN_DISSECTOR_R) {
NL_SET_ERR_MSG_MOD(extack, "unsupported offload: tunnel match not supported");
@@ -268,8 +289,10 @@ nfp_flower_calculate_key_layers(struct nfp_app *app,
if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL)) {
struct flow_match_enc_opts enc_op = { NULL, NULL };
struct flow_match_ipv4_addrs ipv4_addrs;
+ struct flow_match_ipv6_addrs ipv6_addrs;
struct flow_match_control enc_ctl;
struct flow_match_ports enc_ports;
+ bool ipv6_tun = false;
flow_rule_match_enc_control(rule, &enc_ctl);
@@ -277,38 +300,62 @@ nfp_flower_calculate_key_layers(struct nfp_app *app,
NL_SET_ERR_MSG_MOD(extack, "unsupported offload: wildcarded protocols on tunnels are not supported");
return -EOPNOTSUPP;
}
- if (enc_ctl.key->addr_type != FLOW_DISSECTOR_KEY_IPV4_ADDRS) {
- NL_SET_ERR_MSG_MOD(extack, "unsupported offload: only IPv4 tunnels are supported");
+
+ ipv6_tun = enc_ctl.key->addr_type ==
+ FLOW_DISSECTOR_KEY_IPV6_ADDRS;
+ if (ipv6_tun &&
+ !(priv->flower_ext_feats & NFP_FL_FEATS_IPV6_TUN)) {
+ NL_SET_ERR_MSG_MOD(extack, "unsupported offload: firmware does not support IPv6 tunnels");
return -EOPNOTSUPP;
}
- /* These fields are already verified as used. */
- flow_rule_match_enc_ipv4_addrs(rule, &ipv4_addrs);
- if (ipv4_addrs.mask->dst != cpu_to_be32(~0)) {
- NL_SET_ERR_MSG_MOD(extack, "unsupported offload: only an exact match IPv4 destination address is supported");
+ if (!ipv6_tun &&
+ enc_ctl.key->addr_type != FLOW_DISSECTOR_KEY_IPV4_ADDRS) {
+ NL_SET_ERR_MSG_MOD(extack, "unsupported offload: tunnel address type not IPv4 or IPv6");
return -EOPNOTSUPP;
}
+ if (ipv6_tun) {
+ flow_rule_match_enc_ipv6_addrs(rule, &ipv6_addrs);
+ if (memchr_inv(&ipv6_addrs.mask->dst, 0xff,
+ sizeof(ipv6_addrs.mask->dst))) {
+ NL_SET_ERR_MSG_MOD(extack, "unsupported offload: only an exact match IPv6 destination address is supported");
+ return -EOPNOTSUPP;
+ }
+ } else {
+ flow_rule_match_enc_ipv4_addrs(rule, &ipv4_addrs);
+ if (ipv4_addrs.mask->dst != cpu_to_be32(~0)) {
+ NL_SET_ERR_MSG_MOD(extack, "unsupported offload: only an exact match IPv4 destination address is supported");
+ return -EOPNOTSUPP;
+ }
+ }
+
if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_OPTS))
flow_rule_match_enc_opts(rule, &enc_op);
-
if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_PORTS)) {
/* check if GRE, which has no enc_ports */
- if (netif_is_gretap(netdev)) {
- *tun_type = NFP_FL_TUNNEL_GRE;
- key_layer |= NFP_FLOWER_LAYER_EXT_META;
- key_size += sizeof(struct nfp_flower_ext_meta);
- key_layer_two |= NFP_FLOWER_LAYER2_GRE;
- key_size +=
- sizeof(struct nfp_flower_ipv4_gre_tun);
+ if (!netif_is_gretap(netdev)) {
+ NL_SET_ERR_MSG_MOD(extack, "unsupported offload: an exact match on L4 destination port is required for non-GRE tunnels");
+ return -EOPNOTSUPP;
+ }
- if (enc_op.key) {
- NL_SET_ERR_MSG_MOD(extack, "unsupported offload: encap options not supported on GRE tunnels");
- return -EOPNOTSUPP;
- }
+ *tun_type = NFP_FL_TUNNEL_GRE;
+ key_layer |= NFP_FLOWER_LAYER_EXT_META;
+ key_size += sizeof(struct nfp_flower_ext_meta);
+ key_layer_two |= NFP_FLOWER_LAYER2_GRE;
+
+ if (ipv6_tun) {
+ key_layer_two |= NFP_FLOWER_LAYER2_TUN_IPV6;
+ key_size +=
+ sizeof(struct nfp_flower_ipv6_udp_tun);
} else {
- NL_SET_ERR_MSG_MOD(extack, "unsupported offload: an exact match on L4 destination port is required for non-GRE tunnels");
+ key_size +=
+ sizeof(struct nfp_flower_ipv4_udp_tun);
+ }
+
+ if (enc_op.key) {
+ NL_SET_ERR_MSG_MOD(extack, "unsupported offload: encap options not supported on GRE tunnels");
return -EOPNOTSUPP;
}
} else {
@@ -323,7 +370,8 @@ nfp_flower_calculate_key_layers(struct nfp_app *app,
&key_layer_two,
&key_layer,
&key_size, priv,
- tun_type, extack);
+ tun_type, ipv6_tun,
+ extack);
if (err)
return err;
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 4/9] nfp: flower: offload list of IPv6 tunnel endpoint addresses
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
` (2 preceding siblings ...)
2019-12-17 21:57 ` [PATCH net-next v2 3/9] nfp: flower: compile match for IPv6 tunnels John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 5/9] nfp: flower: modify pre-tunnel and set tunnel action for ipv6 John Hurley
` (5 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
Fw requires a list of IPv6 addresses that are used as tunnel endpoints to
enable correct decap of tunneled packets.
Store a list of IPv6 endpoints used in rules with a ref counter to track
how many times it is in use. Offload the entire list any time a new IPv6
address is added or when an address is removed (ref count is 0).
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/cmsg.h | 1 +
drivers/net/ethernet/netronome/nfp/flower/main.h | 21 +++++
drivers/net/ethernet/netronome/nfp/flower/match.c | 24 ++++++
.../net/ethernet/netronome/nfp/flower/offload.c | 6 ++
.../ethernet/netronome/nfp/flower/tunnel_conf.c | 90 +++++++++++++++++++++-
5 files changed, 141 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
index 8d12805..cbb94cf 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
@@ -568,6 +568,7 @@ enum nfp_flower_cmsg_type_port {
NFP_FLOWER_CMSG_TYPE_QOS_DEL = 19,
NFP_FLOWER_CMSG_TYPE_QOS_STATS = 20,
NFP_FLOWER_CMSG_TYPE_PRE_TUN_RULE = 21,
+ NFP_FLOWER_CMSG_TYPE_TUN_IPS_V6 = 22,
NFP_FLOWER_CMSG_TYPE_MAX = 32,
};
diff --git a/drivers/net/ethernet/netronome/nfp/flower/main.h b/drivers/net/ethernet/netronome/nfp/flower/main.h
index 386d629..bd288e2 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/main.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/main.h
@@ -63,8 +63,10 @@ struct nfp_fl_stats_id {
* struct nfp_fl_tunnel_offloads - priv data for tunnel offloads
* @offloaded_macs: Hashtable of the offloaded MAC addresses
* @ipv4_off_list: List of IPv4 addresses to offload
+ * @ipv6_off_list: List of IPv6 addresses to offload
* @neigh_off_list: List of neighbour offloads
* @ipv4_off_lock: Lock for the IPv4 address list
+ * @ipv6_off_lock: Lock for the IPv6 address list
* @neigh_off_lock: Lock for the neighbour address list
* @mac_off_ids: IDA to manage id assignment for offloaded MACs
* @neigh_nb: Notifier to monitor neighbour state
@@ -72,8 +74,10 @@ struct nfp_fl_stats_id {
struct nfp_fl_tunnel_offloads {
struct rhashtable offloaded_macs;
struct list_head ipv4_off_list;
+ struct list_head ipv6_off_list;
struct list_head neigh_off_list;
struct mutex ipv4_off_lock;
+ struct mutex ipv6_off_lock;
spinlock_t neigh_off_lock;
struct ida mac_off_ids;
struct notifier_block neigh_nb;
@@ -274,12 +278,25 @@ struct nfp_fl_stats {
u64 used;
};
+/**
+ * struct nfp_ipv6_addr_entry - cached IPv6 addresses
+ * @ipv6_addr: IP address
+ * @ref_count: number of rules currently using this IP
+ * @list: list pointer
+ */
+struct nfp_ipv6_addr_entry {
+ struct in6_addr ipv6_addr;
+ int ref_count;
+ struct list_head list;
+};
+
struct nfp_fl_payload {
struct nfp_fl_rule_metadata meta;
unsigned long tc_flower_cookie;
struct rhash_head fl_node;
struct rcu_head rcu;
__be32 nfp_tun_ipv4_addr;
+ struct nfp_ipv6_addr_entry *nfp_tun_ipv6;
struct net_device *ingress_dev;
char *unmasked_data;
char *mask_data;
@@ -397,6 +414,10 @@ int nfp_tunnel_mac_event_handler(struct nfp_app *app,
unsigned long event, void *ptr);
void nfp_tunnel_del_ipv4_off(struct nfp_app *app, __be32 ipv4);
void nfp_tunnel_add_ipv4_off(struct nfp_app *app, __be32 ipv4);
+void
+nfp_tunnel_put_ipv6_off(struct nfp_app *app, struct nfp_ipv6_addr_entry *entry);
+struct nfp_ipv6_addr_entry *
+nfp_tunnel_add_ipv6_off(struct nfp_app *app, struct in6_addr *ipv6);
void nfp_tunnel_request_route(struct nfp_app *app, struct sk_buff *skb);
void nfp_tunnel_keep_alive(struct nfp_app *app, struct sk_buff *skb);
void nfp_flower_lag_init(struct nfp_fl_lag *lag);
diff --git a/drivers/net/ethernet/netronome/nfp/flower/match.c b/drivers/net/ethernet/netronome/nfp/flower/match.c
index 2410ead..546bc01 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/match.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/match.c
@@ -494,10 +494,22 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_GRE) {
if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_TUN_IPV6) {
+ struct nfp_flower_ipv6_gre_tun *gre_match;
+ struct nfp_ipv6_addr_entry *entry;
+ struct in6_addr *dst;
+
nfp_flower_compile_ipv6_gre_tun((void *)ext,
(void *)msk, rule);
+ gre_match = (struct nfp_flower_ipv6_gre_tun *)ext;
+ dst = &gre_match->ipv6.dst;
ext += sizeof(struct nfp_flower_ipv6_gre_tun);
msk += sizeof(struct nfp_flower_ipv6_gre_tun);
+
+ entry = nfp_tunnel_add_ipv6_off(app, dst);
+ if (!entry)
+ return -EOPNOTSUPP;
+
+ nfp_flow->nfp_tun_ipv6 = entry;
} else {
__be32 dst;
@@ -518,10 +530,22 @@ int nfp_flower_compile_flow_match(struct nfp_app *app,
if (key_ls->key_layer & NFP_FLOWER_LAYER_VXLAN ||
key_ls->key_layer_two & NFP_FLOWER_LAYER2_GENEVE) {
if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_TUN_IPV6) {
+ struct nfp_flower_ipv6_udp_tun *udp_match;
+ struct nfp_ipv6_addr_entry *entry;
+ struct in6_addr *dst;
+
nfp_flower_compile_ipv6_udp_tun((void *)ext,
(void *)msk, rule);
+ udp_match = (struct nfp_flower_ipv6_udp_tun *)ext;
+ dst = &udp_match->ipv6.dst;
ext += sizeof(struct nfp_flower_ipv6_udp_tun);
msk += sizeof(struct nfp_flower_ipv6_udp_tun);
+
+ entry = nfp_tunnel_add_ipv6_off(app, dst);
+ if (!entry)
+ return -EOPNOTSUPP;
+
+ nfp_flow->nfp_tun_ipv6 = entry;
} else {
__be32 dst;
diff --git a/drivers/net/ethernet/netronome/nfp/flower/offload.c b/drivers/net/ethernet/netronome/nfp/flower/offload.c
index 0997350..83ada1b 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/offload.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/offload.c
@@ -539,6 +539,7 @@ nfp_flower_allocate_new(struct nfp_fl_key_ls *key_layer)
goto err_free_mask;
flow_pay->nfp_tun_ipv4_addr = 0;
+ flow_pay->nfp_tun_ipv6 = NULL;
flow_pay->meta.flags = 0;
INIT_LIST_HEAD(&flow_pay->linked_flows);
flow_pay->in_hw = false;
@@ -1243,6 +1244,8 @@ nfp_flower_add_offload(struct nfp_app *app, struct net_device *netdev,
err_release_metadata:
nfp_modify_flow_metadata(app, flow_pay);
err_destroy_flow:
+ if (flow_pay->nfp_tun_ipv6)
+ nfp_tunnel_put_ipv6_off(app, flow_pay->nfp_tun_ipv6);
kfree(flow_pay->action_data);
kfree(flow_pay->mask_data);
kfree(flow_pay->unmasked_data);
@@ -1359,6 +1362,9 @@ nfp_flower_del_offload(struct nfp_app *app, struct net_device *netdev,
if (nfp_flow->nfp_tun_ipv4_addr)
nfp_tunnel_del_ipv4_off(app, nfp_flow->nfp_tun_ipv4_addr);
+ if (nfp_flow->nfp_tun_ipv6)
+ nfp_tunnel_put_ipv6_off(app, nfp_flow->nfp_tun_ipv6);
+
if (!nfp_flow->in_hw) {
err = 0;
goto err_free_merge_flow;
diff --git a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
index 2600ce4..eddb52d 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
@@ -116,6 +116,18 @@ struct nfp_ipv4_addr_entry {
struct list_head list;
};
+#define NFP_FL_IPV6_ADDRS_MAX 4
+
+/**
+ * struct nfp_tun_ipv6_addr - set the IP address list on the NFP
+ * @count: number of IPs populated in the array
+ * @ipv6_addr: array of IPV6_ADDRS_MAX 128 bit IPv6 addresses
+ */
+struct nfp_tun_ipv6_addr {
+ __be32 count;
+ struct in6_addr ipv6_addr[NFP_FL_IPV6_ADDRS_MAX];
+};
+
#define NFP_TUN_MAC_OFFLOAD_DEL_FLAG 0x2
/**
@@ -502,6 +514,78 @@ void nfp_tunnel_del_ipv4_off(struct nfp_app *app, __be32 ipv4)
nfp_tun_write_ipv4_list(app);
}
+static void nfp_tun_write_ipv6_list(struct nfp_app *app)
+{
+ struct nfp_flower_priv *priv = app->priv;
+ struct nfp_ipv6_addr_entry *entry;
+ struct nfp_tun_ipv6_addr payload;
+ int count = 0;
+
+ memset(&payload, 0, sizeof(struct nfp_tun_ipv6_addr));
+ mutex_lock(&priv->tun.ipv6_off_lock);
+ list_for_each_entry(entry, &priv->tun.ipv6_off_list, list) {
+ if (count >= NFP_FL_IPV6_ADDRS_MAX) {
+ nfp_flower_cmsg_warn(app, "Too many IPv6 tunnel endpoint addresses, some cannot be offloaded.\n");
+ break;
+ }
+ payload.ipv6_addr[count++] = entry->ipv6_addr;
+ }
+ mutex_unlock(&priv->tun.ipv6_off_lock);
+ payload.count = cpu_to_be32(count);
+
+ nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_TUN_IPS_V6,
+ sizeof(struct nfp_tun_ipv6_addr),
+ &payload, GFP_KERNEL);
+}
+
+struct nfp_ipv6_addr_entry *
+nfp_tunnel_add_ipv6_off(struct nfp_app *app, struct in6_addr *ipv6)
+{
+ struct nfp_flower_priv *priv = app->priv;
+ struct nfp_ipv6_addr_entry *entry;
+
+ mutex_lock(&priv->tun.ipv6_off_lock);
+ list_for_each_entry(entry, &priv->tun.ipv6_off_list, list)
+ if (!memcmp(&entry->ipv6_addr, ipv6, sizeof(*ipv6))) {
+ entry->ref_count++;
+ mutex_unlock(&priv->tun.ipv6_off_lock);
+ return entry;
+ }
+
+ entry = kmalloc(sizeof(*entry), GFP_KERNEL);
+ if (!entry) {
+ mutex_unlock(&priv->tun.ipv6_off_lock);
+ nfp_flower_cmsg_warn(app, "Mem error when offloading IP address.\n");
+ return NULL;
+ }
+ entry->ipv6_addr = *ipv6;
+ entry->ref_count = 1;
+ list_add_tail(&entry->list, &priv->tun.ipv6_off_list);
+ mutex_unlock(&priv->tun.ipv6_off_lock);
+
+ nfp_tun_write_ipv6_list(app);
+
+ return entry;
+}
+
+void
+nfp_tunnel_put_ipv6_off(struct nfp_app *app, struct nfp_ipv6_addr_entry *entry)
+{
+ struct nfp_flower_priv *priv = app->priv;
+ bool freed = false;
+
+ mutex_lock(&priv->tun.ipv6_off_lock);
+ if (!--entry->ref_count) {
+ list_del(&entry->list);
+ kfree(entry);
+ freed = true;
+ }
+ mutex_unlock(&priv->tun.ipv6_off_lock);
+
+ if (freed)
+ nfp_tun_write_ipv6_list(app);
+}
+
static int
__nfp_tunnel_offload_mac(struct nfp_app *app, u8 *mac, u16 idx, bool del)
{
@@ -1013,9 +1097,11 @@ int nfp_tunnel_config_start(struct nfp_app *app)
ida_init(&priv->tun.mac_off_ids);
- /* Initialise priv data for IPv4 offloading. */
+ /* Initialise priv data for IPv4/v6 offloading. */
mutex_init(&priv->tun.ipv4_off_lock);
INIT_LIST_HEAD(&priv->tun.ipv4_off_list);
+ mutex_init(&priv->tun.ipv6_off_lock);
+ INIT_LIST_HEAD(&priv->tun.ipv6_off_list);
/* Initialise priv data for neighbour offloading. */
spin_lock_init(&priv->tun.neigh_off_lock);
@@ -1050,6 +1136,8 @@ void nfp_tunnel_config_stop(struct nfp_app *app)
kfree(ip_entry);
}
+ mutex_destroy(&priv->tun.ipv6_off_lock);
+
/* Free any memory that may be occupied by the route list. */
list_for_each_safe(ptr, storage, &priv->tun.neigh_off_list) {
route_entry = list_entry(ptr, struct nfp_ipv4_route_entry,
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 5/9] nfp: flower: modify pre-tunnel and set tunnel action for ipv6
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
` (3 preceding siblings ...)
2019-12-17 21:57 ` [PATCH net-next v2 4/9] nfp: flower: offload list of IPv6 tunnel endpoint addresses John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 6/9] nfp: flower: handle ipv6 tunnel no neigh request John Hurley
` (4 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
The IPv4 set tunnel action allows the setting of tunnel metadata such as
the TTL and ToS values. The pre-tunnel action includes the destination IP
address and is used to calculate the next hop from from the neighbour
table.
Much of the IPv4 tunnel actions can be reused for IPv6 tunnels. Change the
names of associated functions and structs to remove the IPv4 identifier
and make minor modifcations to support IPv6 tunnel actions.
Ensure the pre-tunnel action contains the IPv6 address along with an
identifying flag when an IPv6 tunnel action is required.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/action.c | 65 ++++++++++++++++------
drivers/net/ethernet/netronome/nfp/flower/cmsg.h | 19 ++++---
.../net/ethernet/netronome/nfp/flower/offload.c | 8 +--
3 files changed, 62 insertions(+), 30 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/action.c b/drivers/net/ethernet/netronome/nfp/flower/action.c
index 1b019fd..c06600f 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/action.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/action.c
@@ -22,8 +22,9 @@
#define NFP_FL_TUNNEL_CSUM cpu_to_be16(0x01)
#define NFP_FL_TUNNEL_KEY cpu_to_be16(0x04)
#define NFP_FL_TUNNEL_GENEVE_OPT cpu_to_be16(0x0800)
-#define NFP_FL_SUPPORTED_TUNNEL_INFO_FLAGS IP_TUNNEL_INFO_TX
-#define NFP_FL_SUPPORTED_IPV4_UDP_TUN_FLAGS (NFP_FL_TUNNEL_CSUM | \
+#define NFP_FL_SUPPORTED_TUNNEL_INFO_FLAGS (IP_TUNNEL_INFO_TX | \
+ IP_TUNNEL_INFO_IPV6)
+#define NFP_FL_SUPPORTED_UDP_TUN_FLAGS (NFP_FL_TUNNEL_CSUM | \
NFP_FL_TUNNEL_KEY | \
NFP_FL_TUNNEL_GENEVE_OPT)
@@ -394,19 +395,26 @@ nfp_fl_push_geneve_options(struct nfp_fl_payload *nfp_fl, int *list_len,
}
static int
-nfp_fl_set_ipv4_tun(struct nfp_app *app, struct nfp_fl_set_ipv4_tun *set_tun,
- const struct flow_action_entry *act,
- struct nfp_fl_pre_tunnel *pre_tun,
- enum nfp_flower_tun_type tun_type,
- struct net_device *netdev, struct netlink_ext_ack *extack)
+nfp_fl_set_tun(struct nfp_app *app, struct nfp_fl_set_tun *set_tun,
+ const struct flow_action_entry *act,
+ struct nfp_fl_pre_tunnel *pre_tun,
+ enum nfp_flower_tun_type tun_type,
+ struct net_device *netdev, struct netlink_ext_ack *extack)
{
- size_t act_size = sizeof(struct nfp_fl_set_ipv4_tun);
const struct ip_tunnel_info *ip_tun = act->tunnel;
+ bool ipv6 = ip_tunnel_info_af(ip_tun) == AF_INET6;
+ size_t act_size = sizeof(struct nfp_fl_set_tun);
struct nfp_flower_priv *priv = app->priv;
u32 tmp_set_ip_tun_type_index = 0;
/* Currently support one pre-tunnel so index is always 0. */
int pretun_idx = 0;
+ if (!IS_ENABLED(CONFIG_IPV6) && ipv6)
+ return -EOPNOTSUPP;
+
+ if (ipv6 && !(priv->flower_ext_feats & NFP_FL_FEATS_IPV6_TUN))
+ return -EOPNOTSUPP;
+
BUILD_BUG_ON(NFP_FL_TUNNEL_CSUM != TUNNEL_CSUM ||
NFP_FL_TUNNEL_KEY != TUNNEL_KEY ||
NFP_FL_TUNNEL_GENEVE_OPT != TUNNEL_GENEVE_OPT);
@@ -417,19 +425,35 @@ nfp_fl_set_ipv4_tun(struct nfp_app *app, struct nfp_fl_set_ipv4_tun *set_tun,
return -EOPNOTSUPP;
}
- set_tun->head.jump_id = NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL;
+ set_tun->head.jump_id = NFP_FL_ACTION_OPCODE_SET_TUNNEL;
set_tun->head.len_lw = act_size >> NFP_FL_LW_SIZ;
/* Set tunnel type and pre-tunnel index. */
tmp_set_ip_tun_type_index |=
- FIELD_PREP(NFP_FL_IPV4_TUNNEL_TYPE, tun_type) |
- FIELD_PREP(NFP_FL_IPV4_PRE_TUN_INDEX, pretun_idx);
+ FIELD_PREP(NFP_FL_TUNNEL_TYPE, tun_type) |
+ FIELD_PREP(NFP_FL_PRE_TUN_INDEX, pretun_idx);
set_tun->tun_type_index = cpu_to_be32(tmp_set_ip_tun_type_index);
set_tun->tun_id = ip_tun->key.tun_id;
if (ip_tun->key.ttl) {
set_tun->ttl = ip_tun->key.ttl;
+#ifdef CONFIG_IPV6
+ } else if (ipv6) {
+ struct net *net = dev_net(netdev);
+ struct flowi6 flow = {};
+ struct dst_entry *dst;
+
+ flow.daddr = ip_tun->key.u.ipv6.dst;
+ flow.flowi4_proto = IPPROTO_UDP;
+ dst = ipv6_stub->ipv6_dst_lookup_flow(net, NULL, &flow, NULL);
+ if (!IS_ERR(dst)) {
+ set_tun->ttl = ip6_dst_hoplimit(dst);
+ dst_release(dst);
+ } else {
+ set_tun->ttl = net->ipv6.devconf_all->hop_limit;
+ }
+#endif
} else {
struct net *net = dev_net(netdev);
struct flowi4 flow = {};
@@ -455,7 +479,7 @@ nfp_fl_set_ipv4_tun(struct nfp_app *app, struct nfp_fl_set_ipv4_tun *set_tun,
set_tun->tos = ip_tun->key.tos;
if (!(ip_tun->key.tun_flags & NFP_FL_TUNNEL_KEY) ||
- ip_tun->key.tun_flags & ~NFP_FL_SUPPORTED_IPV4_UDP_TUN_FLAGS) {
+ ip_tun->key.tun_flags & ~NFP_FL_SUPPORTED_UDP_TUN_FLAGS) {
NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support tunnel flag offload");
return -EOPNOTSUPP;
}
@@ -467,7 +491,12 @@ nfp_fl_set_ipv4_tun(struct nfp_app *app, struct nfp_fl_set_ipv4_tun *set_tun,
}
/* Complete pre_tunnel action. */
- pre_tun->ipv4_dst = ip_tun->key.u.ipv4.dst;
+ if (ipv6) {
+ pre_tun->flags |= cpu_to_be16(NFP_FL_PRE_TUN_IPV6);
+ pre_tun->ipv6_dst = ip_tun->key.u.ipv6.dst;
+ } else {
+ pre_tun->ipv4_dst = ip_tun->key.u.ipv4.dst;
+ }
return 0;
}
@@ -956,8 +985,8 @@ nfp_flower_loop_action(struct nfp_app *app, const struct flow_action_entry *act,
struct nfp_flower_pedit_acts *set_act, bool *pkt_host,
struct netlink_ext_ack *extack, int act_idx)
{
- struct nfp_fl_set_ipv4_tun *set_tun;
struct nfp_fl_pre_tunnel *pre_tun;
+ struct nfp_fl_set_tun *set_tun;
struct nfp_fl_push_vlan *psh_v;
struct nfp_fl_push_mpls *psh_m;
struct nfp_fl_pop_vlan *pop_v;
@@ -1032,7 +1061,7 @@ nfp_flower_loop_action(struct nfp_app *app, const struct flow_action_entry *act,
* If none, the packet falls back before applying other actions.
*/
if (*a_len + sizeof(struct nfp_fl_pre_tunnel) +
- sizeof(struct nfp_fl_set_ipv4_tun) > NFP_FL_MAX_A_SIZ) {
+ sizeof(struct nfp_fl_set_tun) > NFP_FL_MAX_A_SIZ) {
NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at tunnel encap");
return -EOPNOTSUPP;
}
@@ -1046,11 +1075,11 @@ nfp_flower_loop_action(struct nfp_app *app, const struct flow_action_entry *act,
return err;
set_tun = (void *)&nfp_fl->action_data[*a_len];
- err = nfp_fl_set_ipv4_tun(app, set_tun, act, pre_tun,
- *tun_type, netdev, extack);
+ err = nfp_fl_set_tun(app, set_tun, act, pre_tun, *tun_type,
+ netdev, extack);
if (err)
return err;
- *a_len += sizeof(struct nfp_fl_set_ipv4_tun);
+ *a_len += sizeof(struct nfp_fl_set_tun);
}
break;
case FLOW_ACTION_TUNNEL_DECAP:
diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
index cbb94cf..7537695 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
@@ -72,7 +72,7 @@
#define NFP_FL_ACTION_OPCODE_POP_VLAN 2
#define NFP_FL_ACTION_OPCODE_PUSH_MPLS 3
#define NFP_FL_ACTION_OPCODE_POP_MPLS 4
-#define NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL 6
+#define NFP_FL_ACTION_OPCODE_SET_TUNNEL 6
#define NFP_FL_ACTION_OPCODE_SET_ETHERNET 7
#define NFP_FL_ACTION_OPCODE_SET_MPLS 8
#define NFP_FL_ACTION_OPCODE_SET_IPV4_ADDRS 9
@@ -101,8 +101,8 @@
/* Tunnel ports */
#define NFP_FL_PORT_TYPE_TUN 0x50000000
-#define NFP_FL_IPV4_TUNNEL_TYPE GENMASK(7, 4)
-#define NFP_FL_IPV4_PRE_TUN_INDEX GENMASK(2, 0)
+#define NFP_FL_TUNNEL_TYPE GENMASK(7, 4)
+#define NFP_FL_PRE_TUN_INDEX GENMASK(2, 0)
#define NFP_FLOWER_WORKQ_MAX_SKBS 30000
@@ -208,13 +208,16 @@ struct nfp_fl_pre_lag {
struct nfp_fl_pre_tunnel {
struct nfp_fl_act_head head;
- __be16 reserved;
- __be32 ipv4_dst;
- /* reserved for use with IPv6 addresses */
- __be32 extra[3];
+ __be16 flags;
+ union {
+ __be32 ipv4_dst;
+ struct in6_addr ipv6_dst;
+ };
};
-struct nfp_fl_set_ipv4_tun {
+#define NFP_FL_PRE_TUN_IPV6 BIT(0)
+
+struct nfp_fl_set_tun {
struct nfp_fl_act_head head;
__be16 reserved;
__be64 tun_id __packed;
diff --git a/drivers/net/ethernet/netronome/nfp/flower/offload.c b/drivers/net/ethernet/netronome/nfp/flower/offload.c
index 83ada1b..d512105 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/offload.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/offload.c
@@ -591,7 +591,7 @@ nfp_flower_update_merge_with_actions(struct nfp_fl_payload *flow,
case NFP_FL_ACTION_OPCODE_POP_VLAN:
merge->tci = cpu_to_be16(0);
break;
- case NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL:
+ case NFP_FL_ACTION_OPCODE_SET_TUNNEL:
/* New tunnel header means l2 to l4 can be matched. */
eth_broadcast_addr(&merge->l2.mac_dst[0]);
eth_broadcast_addr(&merge->l2.mac_src[0]);
@@ -814,15 +814,15 @@ nfp_fl_verify_post_tun_acts(char *acts, int len, struct nfp_fl_push_vlan **vlan)
static int
nfp_fl_push_vlan_after_tun(char *acts, int len, struct nfp_fl_push_vlan *vlan)
{
- struct nfp_fl_set_ipv4_tun *tun;
+ struct nfp_fl_set_tun *tun;
struct nfp_fl_act_head *a;
unsigned int act_off = 0;
while (act_off < len) {
a = (struct nfp_fl_act_head *)&acts[act_off];
- if (a->jump_id == NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL) {
- tun = (struct nfp_fl_set_ipv4_tun *)a;
+ if (a->jump_id == NFP_FL_ACTION_OPCODE_SET_TUNNEL) {
+ tun = (struct nfp_fl_set_tun *)a;
tun->outer_vlan_tpid = vlan->vlan_tpid;
tun->outer_vlan_tci = vlan->vlan_tci;
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 6/9] nfp: flower: handle ipv6 tunnel no neigh request
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
` (4 preceding siblings ...)
2019-12-17 21:57 ` [PATCH net-next v2 5/9] nfp: flower: modify pre-tunnel and set tunnel action for ipv6 John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 7/9] nfp: flower: handle notifiers for ipv6 route changes John Hurley
` (3 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
When fw does not know the next hop for an IPv6 tunnel, it sends a request
to the driver.
Handle this request by doing a route lookup on the IPv6 address and
offloading the next hop to the fw neighbour table.
Similar functions already exist to handle IPv4 no neighbour requests. To
avoid confusion, append these functions with the _ipv4 tag. There is no
change in functionality with this.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/cmsg.c | 8 +-
drivers/net/ethernet/netronome/nfp/flower/cmsg.h | 2 +
drivers/net/ethernet/netronome/nfp/flower/main.h | 3 +-
.../ethernet/netronome/nfp/flower/tunnel_conf.c | 111 ++++++++++++++++++++-
4 files changed, 116 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.c b/drivers/net/ethernet/netronome/nfp/flower/cmsg.c
index 05981b5..00b904e 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.c
@@ -270,7 +270,10 @@ nfp_flower_cmsg_process_one_rx(struct nfp_app *app, struct sk_buff *skb)
}
goto err_default;
case NFP_FLOWER_CMSG_TYPE_NO_NEIGH:
- nfp_tunnel_request_route(app, skb);
+ nfp_tunnel_request_route_v4(app, skb);
+ break;
+ case NFP_FLOWER_CMSG_TYPE_NO_NEIGH_V6:
+ nfp_tunnel_request_route_v6(app, skb);
break;
case NFP_FLOWER_CMSG_TYPE_ACTIVE_TUNS:
nfp_tunnel_keep_alive(app, skb);
@@ -361,7 +364,8 @@ void nfp_flower_cmsg_rx(struct nfp_app *app, struct sk_buff *skb)
nfp_flower_process_mtu_ack(app, skb)) {
/* Handle MTU acks outside wq to prevent RTNL conflict. */
dev_consume_skb_any(skb);
- } else if (cmsg_hdr->type == NFP_FLOWER_CMSG_TYPE_TUN_NEIGH) {
+ } else if (cmsg_hdr->type == NFP_FLOWER_CMSG_TYPE_TUN_NEIGH ||
+ cmsg_hdr->type == NFP_FLOWER_CMSG_TYPE_TUN_NEIGH_V6) {
/* Acks from the NFP that the route is added - ignore. */
dev_consume_skb_any(skb);
} else if (cmsg_hdr->type == NFP_FLOWER_CMSG_TYPE_PORT_REIFY) {
diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
index 7537695..5a8ad58 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
@@ -572,6 +572,8 @@ enum nfp_flower_cmsg_type_port {
NFP_FLOWER_CMSG_TYPE_QOS_STATS = 20,
NFP_FLOWER_CMSG_TYPE_PRE_TUN_RULE = 21,
NFP_FLOWER_CMSG_TYPE_TUN_IPS_V6 = 22,
+ NFP_FLOWER_CMSG_TYPE_NO_NEIGH_V6 = 23,
+ NFP_FLOWER_CMSG_TYPE_TUN_NEIGH_V6 = 24,
NFP_FLOWER_CMSG_TYPE_MAX = 32,
};
diff --git a/drivers/net/ethernet/netronome/nfp/flower/main.h b/drivers/net/ethernet/netronome/nfp/flower/main.h
index bd288e2..01e1886 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/main.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/main.h
@@ -418,7 +418,8 @@ void
nfp_tunnel_put_ipv6_off(struct nfp_app *app, struct nfp_ipv6_addr_entry *entry);
struct nfp_ipv6_addr_entry *
nfp_tunnel_add_ipv6_off(struct nfp_app *app, struct in6_addr *ipv6);
-void nfp_tunnel_request_route(struct nfp_app *app, struct sk_buff *skb);
+void nfp_tunnel_request_route_v4(struct nfp_app *app, struct sk_buff *skb);
+void nfp_tunnel_request_route_v6(struct nfp_app *app, struct sk_buff *skb);
void nfp_tunnel_keep_alive(struct nfp_app *app, struct sk_buff *skb);
void nfp_flower_lag_init(struct nfp_fl_lag *lag);
void nfp_flower_lag_cleanup(struct nfp_fl_lag *lag);
diff --git a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
index eddb52d..09807b3 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
@@ -71,6 +71,22 @@ struct nfp_tun_neigh {
};
/**
+ * struct nfp_tun_neigh_v6 - neighbour/route entry on the NFP
+ * @dst_ipv6: destination IPv6 address
+ * @src_ipv6: source IPv6 address
+ * @dst_addr: destination MAC address
+ * @src_addr: source MAC address
+ * @port_id: NFP port to output packet on - associated with source IPv6
+ */
+struct nfp_tun_neigh_v6 {
+ struct in6_addr dst_ipv6;
+ struct in6_addr src_ipv6;
+ u8 dst_addr[ETH_ALEN];
+ u8 src_addr[ETH_ALEN];
+ __be32 port_id;
+};
+
+/**
* struct nfp_tun_req_route_ipv4 - NFP requests a route/neighbour lookup
* @ingress_port: ingress port of packet that signalled request
* @ipv4_addr: destination ipv4 address for route
@@ -83,6 +99,16 @@ struct nfp_tun_req_route_ipv4 {
};
/**
+ * struct nfp_tun_req_route_ipv6 - NFP requests an IPv6 route/neighbour lookup
+ * @ingress_port: ingress port of packet that signalled request
+ * @ipv6_addr: destination ipv6 address for route
+ */
+struct nfp_tun_req_route_ipv6 {
+ __be32 ingress_port;
+ struct in6_addr ipv6_addr;
+};
+
+/**
* struct nfp_ipv4_route_entry - routes that are offloaded to the NFP
* @ipv4_addr: destination of route
* @list: list pointer
@@ -299,8 +325,8 @@ static void nfp_tun_del_route_from_cache(struct nfp_app *app, __be32 ipv4_addr)
}
static void
-nfp_tun_write_neigh(struct net_device *netdev, struct nfp_app *app,
- struct flowi4 *flow, struct neighbour *neigh, gfp_t flag)
+nfp_tun_write_neigh_v4(struct net_device *netdev, struct nfp_app *app,
+ struct flowi4 *flow, struct neighbour *neigh, gfp_t flag)
{
struct nfp_tun_neigh payload;
u32 port_id;
@@ -334,6 +360,39 @@ nfp_tun_write_neigh(struct net_device *netdev, struct nfp_app *app,
(unsigned char *)&payload, flag);
}
+static void
+nfp_tun_write_neigh_v6(struct net_device *netdev, struct nfp_app *app,
+ struct flowi6 *flow, struct neighbour *neigh, gfp_t flag)
+{
+ struct nfp_tun_neigh_v6 payload;
+ u32 port_id;
+
+ port_id = nfp_flower_get_port_id_from_netdev(app, netdev);
+ if (!port_id)
+ return;
+
+ memset(&payload, 0, sizeof(struct nfp_tun_neigh_v6));
+ payload.dst_ipv6 = flow->daddr;
+
+ /* If entry has expired send dst IP with all other fields 0. */
+ if (!(neigh->nud_state & NUD_VALID) || neigh->dead) {
+ /* Trigger probe to verify invalid neighbour state. */
+ neigh_event_send(neigh, NULL);
+ goto send_msg;
+ }
+
+ /* Have a valid neighbour so populate rest of entry. */
+ payload.src_ipv6 = flow->saddr;
+ ether_addr_copy(payload.src_addr, netdev->dev_addr);
+ neigh_ha_snapshot(payload.dst_addr, neigh, netdev);
+ payload.port_id = cpu_to_be32(port_id);
+
+send_msg:
+ nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_TUN_NEIGH_V6,
+ sizeof(struct nfp_tun_neigh_v6),
+ (unsigned char *)&payload, flag);
+}
+
static int
nfp_tun_neigh_event_handler(struct notifier_block *nb, unsigned long event,
void *ptr)
@@ -384,12 +443,12 @@ nfp_tun_neigh_event_handler(struct notifier_block *nb, unsigned long event,
#endif
flow.flowi4_proto = IPPROTO_UDP;
- nfp_tun_write_neigh(n->dev, app, &flow, n, GFP_ATOMIC);
+ nfp_tun_write_neigh_v4(n->dev, app, &flow, n, GFP_ATOMIC);
return NOTIFY_OK;
}
-void nfp_tunnel_request_route(struct nfp_app *app, struct sk_buff *skb)
+void nfp_tunnel_request_route_v4(struct nfp_app *app, struct sk_buff *skb)
{
struct nfp_tun_req_route_ipv4 *payload;
struct net_device *netdev;
@@ -423,7 +482,7 @@ void nfp_tunnel_request_route(struct nfp_app *app, struct sk_buff *skb)
ip_rt_put(rt);
if (!n)
goto fail_rcu_unlock;
- nfp_tun_write_neigh(n->dev, app, &flow, n, GFP_ATOMIC);
+ nfp_tun_write_neigh_v4(n->dev, app, &flow, n, GFP_ATOMIC);
neigh_release(n);
rcu_read_unlock();
return;
@@ -433,6 +492,48 @@ void nfp_tunnel_request_route(struct nfp_app *app, struct sk_buff *skb)
nfp_flower_cmsg_warn(app, "Requested route not found.\n");
}
+void nfp_tunnel_request_route_v6(struct nfp_app *app, struct sk_buff *skb)
+{
+ struct nfp_tun_req_route_ipv6 *payload;
+ struct net_device *netdev;
+ struct flowi6 flow = {};
+ struct dst_entry *dst;
+ struct neighbour *n;
+
+ payload = nfp_flower_cmsg_get_data(skb);
+
+ rcu_read_lock();
+ netdev = nfp_app_dev_get(app, be32_to_cpu(payload->ingress_port), NULL);
+ if (!netdev)
+ goto fail_rcu_unlock;
+
+ flow.daddr = payload->ipv6_addr;
+ flow.flowi6_proto = IPPROTO_UDP;
+
+#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
+ dst = ipv6_stub->ipv6_dst_lookup_flow(dev_net(netdev), NULL, &flow,
+ NULL);
+ if (IS_ERR(dst))
+ goto fail_rcu_unlock;
+#else
+ goto fail_rcu_unlock;
+#endif
+
+ n = dst_neigh_lookup(dst, &flow.daddr);
+ dst_release(dst);
+ if (!n)
+ goto fail_rcu_unlock;
+
+ nfp_tun_write_neigh_v6(n->dev, app, &flow, n, GFP_ATOMIC);
+ neigh_release(n);
+ rcu_read_unlock();
+ return;
+
+fail_rcu_unlock:
+ rcu_read_unlock();
+ nfp_flower_cmsg_warn(app, "Requested IPv6 route not found.\n");
+}
+
static void nfp_tun_write_ipv4_list(struct nfp_app *app)
{
struct nfp_flower_priv *priv = app->priv;
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 7/9] nfp: flower: handle notifiers for ipv6 route changes
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
` (5 preceding siblings ...)
2019-12-17 21:57 ` [PATCH net-next v2 6/9] nfp: flower: handle ipv6 tunnel no neigh request John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 8/9] nfp: flower: support ipv6 tunnel keep-alive messages from fw John Hurley
` (2 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
A notifier is used to track route changes in the kernel. If a change is
made to a route that is offloaded to fw then an update is sent to the NIC.
The driver tracks all routes that are offloaded to determine if a kernel
change is of interest.
Extend the notifier to track IPv6 route changes and create a new list that
stores offloaded IPv6 routes. Modify the IPv4 route helper functions to
accept varying address lengths. This way, the same core functions can be
used to handle IPv4 and IPv6.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/main.h | 12 +-
.../ethernet/netronome/nfp/flower/tunnel_conf.c | 237 +++++++++++++++------
2 files changed, 181 insertions(+), 68 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/main.h b/drivers/net/ethernet/netronome/nfp/flower/main.h
index 01e1886..7a46032 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/main.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/main.h
@@ -64,10 +64,12 @@ struct nfp_fl_stats_id {
* @offloaded_macs: Hashtable of the offloaded MAC addresses
* @ipv4_off_list: List of IPv4 addresses to offload
* @ipv6_off_list: List of IPv6 addresses to offload
- * @neigh_off_list: List of neighbour offloads
+ * @neigh_off_list_v4: List of IPv4 neighbour offloads
+ * @neigh_off_list_v6: List of IPv6 neighbour offloads
* @ipv4_off_lock: Lock for the IPv4 address list
* @ipv6_off_lock: Lock for the IPv6 address list
- * @neigh_off_lock: Lock for the neighbour address list
+ * @neigh_off_lock_v4: Lock for the IPv4 neighbour address list
+ * @neigh_off_lock_v6: Lock for the IPv6 neighbour address list
* @mac_off_ids: IDA to manage id assignment for offloaded MACs
* @neigh_nb: Notifier to monitor neighbour state
*/
@@ -75,10 +77,12 @@ struct nfp_fl_tunnel_offloads {
struct rhashtable offloaded_macs;
struct list_head ipv4_off_list;
struct list_head ipv6_off_list;
- struct list_head neigh_off_list;
+ struct list_head neigh_off_list_v4;
+ struct list_head neigh_off_list_v6;
struct mutex ipv4_off_lock;
struct mutex ipv6_off_lock;
- spinlock_t neigh_off_lock;
+ spinlock_t neigh_off_lock_v4;
+ spinlock_t neigh_off_lock_v6;
struct ida mac_off_ids;
struct notifier_block neigh_nb;
};
diff --git a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
index 09807b3..b179b63 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
@@ -109,13 +109,13 @@ struct nfp_tun_req_route_ipv6 {
};
/**
- * struct nfp_ipv4_route_entry - routes that are offloaded to the NFP
- * @ipv4_addr: destination of route
+ * struct nfp_offloaded_route - routes that are offloaded to the NFP
* @list: list pointer
+ * @ip_add: destination of route - can be IPv4 or IPv6
*/
-struct nfp_ipv4_route_entry {
- __be32 ipv4_addr;
+struct nfp_offloaded_route {
struct list_head list;
+ u8 ip_add[];
};
#define NFP_FL_IPV4_ADDRS_MAX 32
@@ -262,66 +262,121 @@ nfp_flower_xmit_tun_conf(struct nfp_app *app, u8 mtype, u16 plen, void *pdata,
return 0;
}
-static bool nfp_tun_has_route(struct nfp_app *app, __be32 ipv4_addr)
+static bool
+__nfp_tun_has_route(struct list_head *route_list, spinlock_t *list_lock,
+ void *add, int add_len)
{
- struct nfp_flower_priv *priv = app->priv;
- struct nfp_ipv4_route_entry *entry;
- struct list_head *ptr, *storage;
+ struct nfp_offloaded_route *entry;
- spin_lock_bh(&priv->tun.neigh_off_lock);
- list_for_each_safe(ptr, storage, &priv->tun.neigh_off_list) {
- entry = list_entry(ptr, struct nfp_ipv4_route_entry, list);
- if (entry->ipv4_addr == ipv4_addr) {
- spin_unlock_bh(&priv->tun.neigh_off_lock);
+ spin_lock_bh(list_lock);
+ list_for_each_entry(entry, route_list, list)
+ if (!memcmp(entry->ip_add, add, add_len)) {
+ spin_unlock_bh(list_lock);
return true;
}
- }
- spin_unlock_bh(&priv->tun.neigh_off_lock);
+ spin_unlock_bh(list_lock);
return false;
}
-static void nfp_tun_add_route_to_cache(struct nfp_app *app, __be32 ipv4_addr)
+static int
+__nfp_tun_add_route_to_cache(struct list_head *route_list,
+ spinlock_t *list_lock, void *add, int add_len)
{
- struct nfp_flower_priv *priv = app->priv;
- struct nfp_ipv4_route_entry *entry;
- struct list_head *ptr, *storage;
+ struct nfp_offloaded_route *entry;
- spin_lock_bh(&priv->tun.neigh_off_lock);
- list_for_each_safe(ptr, storage, &priv->tun.neigh_off_list) {
- entry = list_entry(ptr, struct nfp_ipv4_route_entry, list);
- if (entry->ipv4_addr == ipv4_addr) {
- spin_unlock_bh(&priv->tun.neigh_off_lock);
- return;
+ spin_lock_bh(list_lock);
+ list_for_each_entry(entry, route_list, list)
+ if (!memcmp(entry->ip_add, add, add_len)) {
+ spin_unlock_bh(list_lock);
+ return 0;
}
- }
- entry = kmalloc(sizeof(*entry), GFP_ATOMIC);
+
+ entry = kmalloc(sizeof(*entry) + add_len, GFP_ATOMIC);
if (!entry) {
- spin_unlock_bh(&priv->tun.neigh_off_lock);
- nfp_flower_cmsg_warn(app, "Mem error when storing new route.\n");
- return;
+ spin_unlock_bh(list_lock);
+ return -ENOMEM;
}
- entry->ipv4_addr = ipv4_addr;
- list_add_tail(&entry->list, &priv->tun.neigh_off_list);
- spin_unlock_bh(&priv->tun.neigh_off_lock);
+ memcpy(entry->ip_add, add, add_len);
+ list_add_tail(&entry->list, route_list);
+ spin_unlock_bh(list_lock);
+
+ return 0;
}
-static void nfp_tun_del_route_from_cache(struct nfp_app *app, __be32 ipv4_addr)
+static void
+__nfp_tun_del_route_from_cache(struct list_head *route_list,
+ spinlock_t *list_lock, void *add, int add_len)
{
- struct nfp_flower_priv *priv = app->priv;
- struct nfp_ipv4_route_entry *entry;
- struct list_head *ptr, *storage;
+ struct nfp_offloaded_route *entry;
- spin_lock_bh(&priv->tun.neigh_off_lock);
- list_for_each_safe(ptr, storage, &priv->tun.neigh_off_list) {
- entry = list_entry(ptr, struct nfp_ipv4_route_entry, list);
- if (entry->ipv4_addr == ipv4_addr) {
+ spin_lock_bh(list_lock);
+ list_for_each_entry(entry, route_list, list)
+ if (!memcmp(entry->ip_add, add, add_len)) {
list_del(&entry->list);
kfree(entry);
break;
}
- }
- spin_unlock_bh(&priv->tun.neigh_off_lock);
+ spin_unlock_bh(list_lock);
+}
+
+static bool nfp_tun_has_route_v4(struct nfp_app *app, __be32 *ipv4_addr)
+{
+ struct nfp_flower_priv *priv = app->priv;
+
+ return __nfp_tun_has_route(&priv->tun.neigh_off_list_v4,
+ &priv->tun.neigh_off_lock_v4, ipv4_addr,
+ sizeof(*ipv4_addr));
+}
+
+static bool
+nfp_tun_has_route_v6(struct nfp_app *app, struct in6_addr *ipv6_addr)
+{
+ struct nfp_flower_priv *priv = app->priv;
+
+ return __nfp_tun_has_route(&priv->tun.neigh_off_list_v6,
+ &priv->tun.neigh_off_lock_v6, ipv6_addr,
+ sizeof(*ipv6_addr));
+}
+
+static void
+nfp_tun_add_route_to_cache_v4(struct nfp_app *app, __be32 *ipv4_addr)
+{
+ struct nfp_flower_priv *priv = app->priv;
+
+ __nfp_tun_add_route_to_cache(&priv->tun.neigh_off_list_v4,
+ &priv->tun.neigh_off_lock_v4, ipv4_addr,
+ sizeof(*ipv4_addr));
+}
+
+static void
+nfp_tun_add_route_to_cache_v6(struct nfp_app *app, struct in6_addr *ipv6_addr)
+{
+ struct nfp_flower_priv *priv = app->priv;
+
+ __nfp_tun_add_route_to_cache(&priv->tun.neigh_off_list_v6,
+ &priv->tun.neigh_off_lock_v6, ipv6_addr,
+ sizeof(*ipv6_addr));
+}
+
+static void
+nfp_tun_del_route_from_cache_v4(struct nfp_app *app, __be32 *ipv4_addr)
+{
+ struct nfp_flower_priv *priv = app->priv;
+
+ __nfp_tun_del_route_from_cache(&priv->tun.neigh_off_list_v4,
+ &priv->tun.neigh_off_lock_v4, ipv4_addr,
+ sizeof(*ipv4_addr));
+}
+
+static void
+nfp_tun_del_route_from_cache_v6(struct nfp_app *app, struct in6_addr *ipv6_addr)
+{
+ struct nfp_flower_priv *priv = app->priv;
+
+ __nfp_tun_del_route_from_cache(&priv->tun.neigh_off_list_v6,
+ &priv->tun.neigh_off_lock_v6, ipv6_addr,
+ sizeof(*ipv6_addr));
}
static void
@@ -340,7 +395,7 @@ nfp_tun_write_neigh_v4(struct net_device *netdev, struct nfp_app *app,
/* If entry has expired send dst IP with all other fields 0. */
if (!(neigh->nud_state & NUD_VALID) || neigh->dead) {
- nfp_tun_del_route_from_cache(app, payload.dst_ipv4);
+ nfp_tun_del_route_from_cache_v4(app, &payload.dst_ipv4);
/* Trigger ARP to verify invalid neighbour state. */
neigh_event_send(neigh, NULL);
goto send_msg;
@@ -352,7 +407,7 @@ nfp_tun_write_neigh_v4(struct net_device *netdev, struct nfp_app *app,
neigh_ha_snapshot(payload.dst_addr, neigh, netdev);
payload.port_id = cpu_to_be32(port_id);
/* Add destination of new route to NFP cache. */
- nfp_tun_add_route_to_cache(app, payload.dst_ipv4);
+ nfp_tun_add_route_to_cache_v4(app, &payload.dst_ipv4);
send_msg:
nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_TUN_NEIGH,
@@ -376,6 +431,7 @@ nfp_tun_write_neigh_v6(struct net_device *netdev, struct nfp_app *app,
/* If entry has expired send dst IP with all other fields 0. */
if (!(neigh->nud_state & NUD_VALID) || neigh->dead) {
+ nfp_tun_del_route_from_cache_v6(app, &payload.dst_ipv6);
/* Trigger probe to verify invalid neighbour state. */
neigh_event_send(neigh, NULL);
goto send_msg;
@@ -386,6 +442,8 @@ nfp_tun_write_neigh_v6(struct net_device *netdev, struct nfp_app *app,
ether_addr_copy(payload.src_addr, netdev->dev_addr);
neigh_ha_snapshot(payload.dst_addr, neigh, netdev);
payload.port_id = cpu_to_be32(port_id);
+ /* Add destination of new route to NFP cache. */
+ nfp_tun_add_route_to_cache_v6(app, &payload.dst_ipv6);
send_msg:
nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_TUN_NEIGH_V6,
@@ -399,10 +457,12 @@ nfp_tun_neigh_event_handler(struct notifier_block *nb, unsigned long event,
{
struct nfp_flower_priv *app_priv;
struct netevent_redirect *redir;
- struct flowi4 flow = {};
+ struct flowi4 flow4 = {};
+ struct flowi6 flow6 = {};
struct neighbour *n;
struct nfp_app *app;
struct rtable *rt;
+ bool ipv6 = false;
int err;
switch (event) {
@@ -417,7 +477,13 @@ nfp_tun_neigh_event_handler(struct notifier_block *nb, unsigned long event,
return NOTIFY_DONE;
}
- flow.daddr = *(__be32 *)n->primary_key;
+ if (n->tbl->family == AF_INET6)
+ ipv6 = true;
+
+ if (ipv6)
+ flow6.daddr = *(struct in6_addr *)n->primary_key;
+ else
+ flow4.daddr = *(__be32 *)n->primary_key;
app_priv = container_of(nb, struct nfp_flower_priv, tun.neigh_nb);
app = app_priv->app;
@@ -427,23 +493,41 @@ nfp_tun_neigh_event_handler(struct notifier_block *nb, unsigned long event,
return NOTIFY_DONE;
/* Only concerned with changes to routes already added to NFP. */
- if (!nfp_tun_has_route(app, flow.daddr))
+ if ((ipv6 && !nfp_tun_has_route_v6(app, &flow6.daddr)) ||
+ (!ipv6 && !nfp_tun_has_route_v4(app, &flow4.daddr)))
return NOTIFY_DONE;
#if IS_ENABLED(CONFIG_INET)
- /* Do a route lookup to populate flow data. */
- rt = ip_route_output_key(dev_net(n->dev), &flow);
- err = PTR_ERR_OR_ZERO(rt);
- if (err)
+ if (ipv6) {
+#if IS_ENABLED(CONFIG_IPV6)
+ struct dst_entry *dst;
+
+ dst = ipv6_stub->ipv6_dst_lookup_flow(dev_net(n->dev), NULL,
+ &flow6, NULL);
+ if (IS_ERR(dst))
+ return NOTIFY_DONE;
+
+ dst_release(dst);
+ flow6.flowi6_proto = IPPROTO_UDP;
+ nfp_tun_write_neigh_v6(n->dev, app, &flow6, n, GFP_ATOMIC);
+#else
return NOTIFY_DONE;
+#endif /* CONFIG_IPV6 */
+ } else {
+ /* Do a route lookup to populate flow data. */
+ rt = ip_route_output_key(dev_net(n->dev), &flow4);
+ err = PTR_ERR_OR_ZERO(rt);
+ if (err)
+ return NOTIFY_DONE;
- ip_rt_put(rt);
+ ip_rt_put(rt);
+
+ flow4.flowi4_proto = IPPROTO_UDP;
+ nfp_tun_write_neigh_v4(n->dev, app, &flow4, n, GFP_ATOMIC);
+ }
#else
return NOTIFY_DONE;
-#endif
-
- flow.flowi4_proto = IPPROTO_UDP;
- nfp_tun_write_neigh_v4(n->dev, app, &flow, n, GFP_ATOMIC);
+#endif /* CONFIG_INET */
return NOTIFY_OK;
}
@@ -1205,8 +1289,10 @@ int nfp_tunnel_config_start(struct nfp_app *app)
INIT_LIST_HEAD(&priv->tun.ipv6_off_list);
/* Initialise priv data for neighbour offloading. */
- spin_lock_init(&priv->tun.neigh_off_lock);
- INIT_LIST_HEAD(&priv->tun.neigh_off_list);
+ spin_lock_init(&priv->tun.neigh_off_lock_v4);
+ INIT_LIST_HEAD(&priv->tun.neigh_off_list_v4);
+ spin_lock_init(&priv->tun.neigh_off_lock_v6);
+ INIT_LIST_HEAD(&priv->tun.neigh_off_list_v6);
priv->tun.neigh_nb.notifier_call = nfp_tun_neigh_event_handler;
err = register_netevent_notifier(&priv->tun.neigh_nb);
@@ -1221,9 +1307,11 @@ int nfp_tunnel_config_start(struct nfp_app *app)
void nfp_tunnel_config_stop(struct nfp_app *app)
{
+ struct nfp_offloaded_route *route_entry, *temp;
struct nfp_flower_priv *priv = app->priv;
- struct nfp_ipv4_route_entry *route_entry;
struct nfp_ipv4_addr_entry *ip_entry;
+ struct nfp_tun_neigh_v6 ipv6_route;
+ struct nfp_tun_neigh ipv4_route;
struct list_head *ptr, *storage;
unregister_netevent_notifier(&priv->tun.neigh_nb);
@@ -1239,12 +1327,33 @@ void nfp_tunnel_config_stop(struct nfp_app *app)
mutex_destroy(&priv->tun.ipv6_off_lock);
- /* Free any memory that may be occupied by the route list. */
- list_for_each_safe(ptr, storage, &priv->tun.neigh_off_list) {
- route_entry = list_entry(ptr, struct nfp_ipv4_route_entry,
- list);
+ /* Free memory in the route list and remove entries from fw cache. */
+ list_for_each_entry_safe(route_entry, temp,
+ &priv->tun.neigh_off_list_v4, list) {
+ memset(&ipv4_route, 0, sizeof(ipv4_route));
+ memcpy(&ipv4_route.dst_ipv4, &route_entry->ip_add,
+ sizeof(ipv4_route.dst_ipv4));
list_del(&route_entry->list);
kfree(route_entry);
+
+ nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_TUN_NEIGH,
+ sizeof(struct nfp_tun_neigh),
+ (unsigned char *)&ipv4_route,
+ GFP_KERNEL);
+ }
+
+ list_for_each_entry_safe(route_entry, temp,
+ &priv->tun.neigh_off_list_v6, list) {
+ memset(&ipv6_route, 0, sizeof(ipv6_route));
+ memcpy(&ipv6_route.dst_ipv6, &route_entry->ip_add,
+ sizeof(ipv6_route.dst_ipv6));
+ list_del(&route_entry->list);
+ kfree(route_entry);
+
+ nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_TUN_NEIGH_V6,
+ sizeof(struct nfp_tun_neigh),
+ (unsigned char *)&ipv6_route,
+ GFP_KERNEL);
}
/* Destroy rhash. Entries should be cleaned on netdev notifier unreg. */
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 8/9] nfp: flower: support ipv6 tunnel keep-alive messages from fw
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
` (6 preceding siblings ...)
2019-12-17 21:57 ` [PATCH net-next v2 7/9] nfp: flower: handle notifiers for ipv6 route changes John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 9/9] nfp: flower: update flow merge code to support IPv6 tunnels John Hurley
2019-12-20 1:37 ` [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP David Miller
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
FW sends an update of IPv6 tunnels that are active in a given period. Use
this information to update the kernel table so that neighbour entries do
not time out when active on the NIC.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/flower/cmsg.c | 3 ++
drivers/net/ethernet/netronome/nfp/flower/cmsg.h | 1 +
drivers/net/ethernet/netronome/nfp/flower/main.h | 1 +
.../ethernet/netronome/nfp/flower/tunnel_conf.c | 62 ++++++++++++++++++++++
4 files changed, 67 insertions(+)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.c b/drivers/net/ethernet/netronome/nfp/flower/cmsg.c
index 00b904e..a595ddb 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.c
@@ -278,6 +278,9 @@ nfp_flower_cmsg_process_one_rx(struct nfp_app *app, struct sk_buff *skb)
case NFP_FLOWER_CMSG_TYPE_ACTIVE_TUNS:
nfp_tunnel_keep_alive(app, skb);
break;
+ case NFP_FLOWER_CMSG_TYPE_ACTIVE_TUNS_V6:
+ nfp_tunnel_keep_alive_v6(app, skb);
+ break;
case NFP_FLOWER_CMSG_TYPE_QOS_STATS:
nfp_flower_stats_rlim_reply(app, skb);
break;
diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
index 5a8ad58..9b50d76 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h
@@ -574,6 +574,7 @@ enum nfp_flower_cmsg_type_port {
NFP_FLOWER_CMSG_TYPE_TUN_IPS_V6 = 22,
NFP_FLOWER_CMSG_TYPE_NO_NEIGH_V6 = 23,
NFP_FLOWER_CMSG_TYPE_TUN_NEIGH_V6 = 24,
+ NFP_FLOWER_CMSG_TYPE_ACTIVE_TUNS_V6 = 25,
NFP_FLOWER_CMSG_TYPE_MAX = 32,
};
diff --git a/drivers/net/ethernet/netronome/nfp/flower/main.h b/drivers/net/ethernet/netronome/nfp/flower/main.h
index 7a46032..ddd7b7f 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/main.h
+++ b/drivers/net/ethernet/netronome/nfp/flower/main.h
@@ -425,6 +425,7 @@ nfp_tunnel_add_ipv6_off(struct nfp_app *app, struct in6_addr *ipv6);
void nfp_tunnel_request_route_v4(struct nfp_app *app, struct sk_buff *skb);
void nfp_tunnel_request_route_v6(struct nfp_app *app, struct sk_buff *skb);
void nfp_tunnel_keep_alive(struct nfp_app *app, struct sk_buff *skb);
+void nfp_tunnel_keep_alive_v6(struct nfp_app *app, struct sk_buff *skb);
void nfp_flower_lag_init(struct nfp_fl_lag *lag);
void nfp_flower_lag_cleanup(struct nfp_fl_lag *lag);
int nfp_flower_lag_reset(struct nfp_fl_lag *lag);
diff --git a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
index b179b63..2df3dee 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c
@@ -55,6 +55,25 @@ struct nfp_tun_active_tuns {
};
/**
+ * struct nfp_tun_active_tuns_v6 - periodic message of active IPv6 tunnels
+ * @seq: sequence number of the message
+ * @count: number of tunnels report in message
+ * @flags: options part of the request
+ * @tun_info.ipv6: dest IPv6 address of active route
+ * @tun_info.egress_port: port the encapsulated packet egressed
+ * @tun_info: tunnels that have sent traffic in reported period
+ */
+struct nfp_tun_active_tuns_v6 {
+ __be32 seq;
+ __be32 count;
+ __be32 flags;
+ struct route_ip_info_v6 {
+ struct in6_addr ipv6;
+ __be32 egress_port;
+ } tun_info[];
+};
+
+/**
* struct nfp_tun_neigh - neighbour/route entry on the NFP
* @dst_ipv4: destination IPv4 address
* @src_ipv4: source IPv4 address
@@ -244,6 +263,49 @@ void nfp_tunnel_keep_alive(struct nfp_app *app, struct sk_buff *skb)
rcu_read_unlock();
}
+void nfp_tunnel_keep_alive_v6(struct nfp_app *app, struct sk_buff *skb)
+{
+#if IS_ENABLED(CONFIG_IPV6)
+ struct nfp_tun_active_tuns_v6 *payload;
+ struct net_device *netdev;
+ int count, i, pay_len;
+ struct neighbour *n;
+ void *ipv6_add;
+ u32 port;
+
+ payload = nfp_flower_cmsg_get_data(skb);
+ count = be32_to_cpu(payload->count);
+ if (count > NFP_FL_IPV6_ADDRS_MAX) {
+ nfp_flower_cmsg_warn(app, "IPv6 tunnel keep-alive request exceeds max routes.\n");
+ return;
+ }
+
+ pay_len = nfp_flower_cmsg_get_data_len(skb);
+ if (pay_len != struct_size(payload, tun_info, count)) {
+ nfp_flower_cmsg_warn(app, "Corruption in tunnel keep-alive message.\n");
+ return;
+ }
+
+ rcu_read_lock();
+ for (i = 0; i < count; i++) {
+ ipv6_add = &payload->tun_info[i].ipv6;
+ port = be32_to_cpu(payload->tun_info[i].egress_port);
+ netdev = nfp_app_dev_get(app, port, NULL);
+ if (!netdev)
+ continue;
+
+ n = neigh_lookup(&nd_tbl, ipv6_add, netdev);
+ if (!n)
+ continue;
+
+ /* Update the used timestamp of neighbour */
+ neigh_event_send(n, NULL);
+ neigh_release(n);
+ }
+ rcu_read_unlock();
+#endif
+}
+
static int
nfp_flower_xmit_tun_conf(struct nfp_app *app, u8 mtype, u16 plen, void *pdata,
gfp_t flag)
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH net-next v2 9/9] nfp: flower: update flow merge code to support IPv6 tunnels
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
` (7 preceding siblings ...)
2019-12-17 21:57 ` [PATCH net-next v2 8/9] nfp: flower: support ipv6 tunnel keep-alive messages from fw John Hurley
@ 2019-12-17 21:57 ` John Hurley
2019-12-20 1:37 ` [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP David Miller
9 siblings, 0 replies; 11+ messages in thread
From: John Hurley @ 2019-12-17 21:57 UTC (permalink / raw)
To: netdev; +Cc: simon.horman, jakub.kicinski, oss-drivers, John Hurley
Both pre-tunnel match rules and flow merge functions parse compiled
match/action fields for validation.
Update these validation functions to include IPv6 match and action fields.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
.../net/ethernet/netronome/nfp/flower/offload.c | 28 ++++++++++++++++++----
1 file changed, 23 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/flower/offload.c b/drivers/net/ethernet/netronome/nfp/flower/offload.c
index d512105..7ca5c1b 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/offload.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/offload.c
@@ -68,7 +68,8 @@
#define NFP_FLOWER_PRE_TUN_RULE_FIELDS \
(NFP_FLOWER_LAYER_PORT | \
NFP_FLOWER_LAYER_MAC | \
- NFP_FLOWER_LAYER_IPV4)
+ NFP_FLOWER_LAYER_IPV4 | \
+ NFP_FLOWER_LAYER_IPV6)
struct nfp_flower_merge_check {
union {
@@ -566,10 +567,12 @@ nfp_flower_update_merge_with_actions(struct nfp_fl_payload *flow,
struct nfp_fl_set_ip4_addrs *ipv4_add;
struct nfp_fl_set_ipv6_addr *ipv6_add;
struct nfp_fl_push_vlan *push_vlan;
+ struct nfp_fl_pre_tunnel *pre_tun;
struct nfp_fl_set_tport *tport;
struct nfp_fl_set_eth *eth;
struct nfp_fl_act_head *a;
unsigned int act_off = 0;
+ bool ipv6_tun = false;
u8 act_id = 0;
u8 *ports;
int i;
@@ -597,8 +600,12 @@ nfp_flower_update_merge_with_actions(struct nfp_fl_payload *flow,
eth_broadcast_addr(&merge->l2.mac_src[0]);
memset(&merge->l4, 0xff,
sizeof(struct nfp_flower_tp_ports));
- memset(&merge->ipv4, 0xff,
- sizeof(struct nfp_flower_ipv4));
+ if (ipv6_tun)
+ memset(&merge->ipv6, 0xff,
+ sizeof(struct nfp_flower_ipv6));
+ else
+ memset(&merge->ipv4, 0xff,
+ sizeof(struct nfp_flower_ipv4));
break;
case NFP_FL_ACTION_OPCODE_SET_ETHERNET:
eth = (struct nfp_fl_set_eth *)a;
@@ -646,6 +653,10 @@ nfp_flower_update_merge_with_actions(struct nfp_fl_payload *flow,
ports[i] |= tport->tp_port_mask[i];
break;
case NFP_FL_ACTION_OPCODE_PRE_TUNNEL:
+ pre_tun = (struct nfp_fl_pre_tunnel *)a;
+ ipv6_tun = be16_to_cpu(pre_tun->flags) &
+ NFP_FL_PRE_TUN_IPV6;
+ break;
case NFP_FL_ACTION_OPCODE_PRE_LAG:
case NFP_FL_ACTION_OPCODE_PUSH_GENEVE:
break;
@@ -1107,15 +1118,22 @@ nfp_flower_validate_pre_tun_rule(struct nfp_app *app,
return -EOPNOTSUPP;
}
- if (key_layer & NFP_FLOWER_LAYER_IPV4) {
+ if (key_layer & NFP_FLOWER_LAYER_IPV4 ||
+ key_layer & NFP_FLOWER_LAYER_IPV6) {
+ /* Flags and proto fields have same offset in IPv4 and IPv6. */
int ip_flags = offsetof(struct nfp_flower_ipv4, ip_ext.flags);
int ip_proto = offsetof(struct nfp_flower_ipv4, ip_ext.proto);
+ int size;
int i;
+ size = key_layer & NFP_FLOWER_LAYER_IPV4 ?
+ sizeof(struct nfp_flower_ipv4) :
+ sizeof(struct nfp_flower_ipv6);
+
mask += sizeof(struct nfp_flower_mac_mpls);
/* Ensure proto and flags are the only IP layer fields. */
- for (i = 0; i < sizeof(struct nfp_flower_ipv4); i++)
+ for (i = 0; i < size; i++)
if (mask[i] && i != ip_flags && i != ip_proto) {
NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: only flags and proto can be matched in ip header");
return -EOPNOTSUPP;
--
2.7.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
` (8 preceding siblings ...)
2019-12-17 21:57 ` [PATCH net-next v2 9/9] nfp: flower: update flow merge code to support IPv6 tunnels John Hurley
@ 2019-12-20 1:37 ` David Miller
9 siblings, 0 replies; 11+ messages in thread
From: David Miller @ 2019-12-20 1:37 UTC (permalink / raw)
To: john.hurley; +Cc: netdev, simon.horman, jakub.kicinski, oss-drivers
From: John Hurley <john.hurley@netronome.com>
Date: Tue, 17 Dec 2019 21:57:15 +0000
> The following patches add support for IPv6 tunnel offload to the NFP
> driver.
>
> Patches 1-2 do some code tidy up and prepare existing code for reuse in
> IPv6 tunnels.
> Patches 3-4 handle IPv6 tunnel decap (match) rules.
> Patches 5-8 handle encap (action) rules.
> Patch 9 adds IPv6 support to the merge and pre-tunnel rule functions.
>
> v1->v2:
> - fix compiler warning when building without CONFIG_IPV6 set -
> Jakub Kicinski (patch 7)
Series applied, thanks.
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2019-12-20 1:37 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-12-17 21:57 [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 1/9] nfp: flower: pass flow rule pointer directly to match functions John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 2/9] nfp: flower: move udp tunnel key match compilation to helper function John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 3/9] nfp: flower: compile match for IPv6 tunnels John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 4/9] nfp: flower: offload list of IPv6 tunnel endpoint addresses John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 5/9] nfp: flower: modify pre-tunnel and set tunnel action for ipv6 John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 6/9] nfp: flower: handle ipv6 tunnel no neigh request John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 7/9] nfp: flower: handle notifiers for ipv6 route changes John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 8/9] nfp: flower: support ipv6 tunnel keep-alive messages from fw John Hurley
2019-12-17 21:57 ` [PATCH net-next v2 9/9] nfp: flower: update flow merge code to support IPv6 tunnels John Hurley
2019-12-20 1:37 ` [PATCH net-next v2 0/9] Add ipv6 tunnel support to NFP David Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).