* [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary
@ 2020-04-28 5:24 xiangxia.m.yue
2020-04-28 5:24 ` [PATCH net-next 2/3] net/mlx5e: Introduce mlx5e_eswitch_rep_uplink_priv helper xiangxia.m.yue
` (3 more replies)
0 siblings, 4 replies; 8+ messages in thread
From: xiangxia.m.yue @ 2020-04-28 5:24 UTC (permalink / raw)
To: paulb, saeedm, roid, gerlitz.or; +Cc: netdev, Tonghao Zhang
From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
The commit 0a7fcb78cc21 ("net/mlx5e: Support inner header rewrite with
goto action"), will decapitate the tunnel packets if there is a goto
action in chain 0. But in some case, we don't want do that, for example:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower dst_mac 00:11:22:33:44:55 enc_src_ip 2.2.2.200 \
enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
action tunnel_key unset action mirred egress redirect dev enp130s0f0_0
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower dst_mac 00:11:22:33:44:66 enc_src_ip 2.2.2.200 \
enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 200 \
action tunnel_key unset action mirred egress redirect dev enp130s0f0_1
In this patch, if there is a pedit action in chain, do the decapitation action.
if there are pedit and goto actions, do the decapitation and id mapping action.
8 test units:
[1]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action tunnel_key unset \
action mirred egress redirect dev enp130s0f0_0
[2]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:f0 \
action mirred egress redirect dev enp130s0f0_0
[3]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action tunnel_key unset \
action pedit ex munge eth src set 00:11:22:33:44:f0 \
action mirred egress redirect dev enp130s0f0_0
[4]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:ff pipe \
action mirred egress redirect dev enp130s0f0_0
[5]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:ff \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
action tunnel_key unset \
action mirred egress redirect dev enp130s0f0_0
[6]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:ff \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
action tunnel_key unset \
action pedit ex munge eth src set 00:11:22:33:44:f0 \
action mirred egress redirect dev enp130s0f0_0
[7]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:ff \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
action pedit ex munge eth src set 00:11:22:33:44:f0 \
action goto chain 3
$ tc filter add dev vxlan0 protocol ip parent ffff: prio 1 chain 3 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:f0 \
action tunnel_key unset \
action pedit ex munge eth src set 00:11:22:33:44:f1 \
action mirred egress redirect dev enp130s0f0_0
[8]:
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
action goto chain 2
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:f0 \
action goto chain 3
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 3 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:f1 \
action goto chain 4
$ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 4 \
flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
action pedit ex munge eth src set 00:11:22:33:44:f2 \
action mirred egress redirect dev enp130s0f0_0
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
---
.../net/ethernet/mellanox/mlx5/core/en/mapping.c | 24 ++++++
.../net/ethernet/mellanox/mlx5/core/en/mapping.h | 1 +
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 97 +++++++++++++++-------
3 files changed, 92 insertions(+), 30 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
index ea321e528749..90306dde6b60 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
@@ -74,6 +74,30 @@ int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id)
return err;
}
+int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32 *id)
+{
+ struct mapping_item *mi;
+ u32 hash_key;
+
+ mutex_lock(&ctx->lock);
+
+ hash_key = jhash(data, ctx->data_size, 0);
+ hash_for_each_possible(ctx->ht, mi, node, hash_key) {
+ if (!memcmp(data, mi->data, ctx->data_size))
+ goto found;
+ }
+
+ mutex_unlock(&ctx->lock);
+ return -ENOENT;
+
+found:
+ if (id)
+ *id = mi->id;
+
+ mutex_unlock(&ctx->lock);
+ return 0;
+}
+
static void mapping_remove_and_free(struct mapping_ctx *ctx,
struct mapping_item *mi)
{
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
index 285525cc5470..af501c9796b7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
@@ -9,6 +9,7 @@
int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id);
int mapping_remove(struct mapping_ctx *ctx, u32 id);
int mapping_find(struct mapping_ctx *ctx, u32 id, void *data);
+int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32 *id);
/* mapping uses an xarray to map data to ids in add(), and for find().
* For locking, it uses a internal xarray spin lock for add()/remove(),
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index a574c588269a..64f5c3f3dbb3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1786,7 +1786,8 @@ static void mlx5e_tc_del_flow(struct mlx5e_priv *priv,
}
}
-static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
+static int flow_has_tc_action(struct flow_cls_offload *f,
+ enum flow_action_id action)
{
struct flow_rule *rule = flow_cls_offload_flow_rule(f);
struct flow_action *flow_action = &rule->action;
@@ -1794,12 +1795,8 @@ static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
int i;
flow_action_for_each(i, act, flow_action) {
- switch (act->id) {
- case FLOW_ACTION_GOTO:
+ if (act->id == action)
return true;
- default:
- continue;
- }
}
return false;
@@ -1853,10 +1850,37 @@ static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
sizeof(*__dst));\
})
+static void mlx5e_make_tunnel_match_key(struct flow_cls_offload *f,
+ struct net_device *filter_dev,
+ struct tunnel_match_key *tunnel_key)
+{
+ struct flow_rule *rule = flow_cls_offload_flow_rule(f);
+
+ memset(tunnel_key, 0, sizeof(*tunnel_key));
+ COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
+ &tunnel_key->enc_control);
+ if (tunnel_key->enc_control.addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS)
+ COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
+ &tunnel_key->enc_ipv4);
+ else
+ COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
+ &tunnel_key->enc_ipv6);
+
+ COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP, &tunnel_key->enc_ip);
+ COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
+ &tunnel_key->enc_tp);
+ COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
+ &tunnel_key->enc_key_id);
+
+ tunnel_key->filter_ifindex = filter_dev->ifindex;
+}
+
static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
struct flow_cls_offload *f,
- struct net_device *filter_dev)
+ struct net_device *filter_dev,
+ bool sets_mapping,
+ bool needs_mapping)
{
struct flow_rule *rule = flow_cls_offload_flow_rule(f);
struct netlink_ext_ack *extack = f->common.extack;
@@ -1876,22 +1900,7 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
uplink_priv = &uplink_rpriv->uplink_priv;
- memset(&tunnel_key, 0, sizeof(tunnel_key));
- COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
- &tunnel_key.enc_control);
- if (tunnel_key.enc_control.addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS)
- COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
- &tunnel_key.enc_ipv4);
- else
- COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
- &tunnel_key.enc_ipv6);
- COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP, &tunnel_key.enc_ip);
- COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
- &tunnel_key.enc_tp);
- COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
- &tunnel_key.enc_key_id);
- tunnel_key.filter_ifindex = filter_dev->ifindex;
-
+ mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
err = mapping_add(uplink_priv->tunnel_mapping, &tunnel_key, &tun_id);
if (err)
return err;
@@ -1915,10 +1924,10 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
mask = enc_opts_id ? TUNNEL_ID_MASK :
(TUNNEL_ID_MASK & ~ENC_OPTS_BITS_MASK);
- if (attr->chain) {
+ if (needs_mapping) {
mlx5e_tc_match_to_reg_match(&attr->parse_attr->spec,
TUNNEL_TO_REG, value, mask);
- } else {
+ } else if (sets_mapping) {
mod_hdr_acts = &attr->parse_attr->mod_hdr_acts;
err = mlx5e_tc_match_to_reg_set(priv->mdev,
mod_hdr_acts,
@@ -1941,6 +1950,25 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
return err;
}
+static int mlx5e_lookup_flow_tunnel_id(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ struct flow_cls_offload *f,
+ struct net_device *filter_dev,
+ u32 *tun_id)
+{
+ struct mlx5_rep_uplink_priv *uplink_priv;
+ struct mlx5e_rep_priv *uplink_rpriv;
+ struct tunnel_match_key tunnel_key;
+ struct mlx5_eswitch *esw;
+
+ esw = priv->mdev->priv.eswitch;
+ uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+ uplink_priv = &uplink_rpriv->uplink_priv;
+
+ mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
+ return mapping_find_by_data(uplink_priv->tunnel_mapping, &tunnel_key, tun_id);
+}
+
static void mlx5e_put_flow_tunnel_id(struct mlx5e_tc_flow *flow)
{
u32 enc_opts_id = flow->tunnel_id & ENC_OPTS_BITS_MASK;
@@ -1976,14 +2004,22 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
struct netlink_ext_ack *extack = f->common.extack;
bool needs_mapping, sets_mapping;
+ bool pedit_action;
int err;
if (!mlx5e_is_eswitch_flow(flow))
return -EOPNOTSUPP;
- needs_mapping = !!flow->esw_attr->chain;
- sets_mapping = !flow->esw_attr->chain && flow_has_tc_fwd_action(f);
- *match_inner = !needs_mapping;
+ pedit_action = flow_has_tc_action(f, FLOW_ACTION_MANGLE) ||
+ flow_has_tc_action(f, FLOW_ACTION_ADD);
+
+ *match_inner = pedit_action;
+ sets_mapping = pedit_action &&
+ flow_has_tc_action(f, FLOW_ACTION_GOTO);
+
+ needs_mapping = !!flow->esw_attr->chain &&
+ !mlx5e_lookup_flow_tunnel_id(priv, flow, f,
+ filter_dev, NULL);
if ((needs_mapping || sets_mapping) &&
!mlx5_eswitch_reg_c1_loopback_enabled(esw)) {
@@ -1994,7 +2030,7 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
return -EOPNOTSUPP;
}
- if (!flow->esw_attr->chain) {
+ if (*match_inner && !needs_mapping) {
err = mlx5e_tc_tun_parse(filter_dev, priv, spec, f,
match_level);
if (err) {
@@ -2011,7 +2047,8 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
if (!needs_mapping && !sets_mapping)
return 0;
- return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev);
+ return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev,
+ sets_mapping, needs_mapping);
}
static void *get_match_inner_headers_criteria(struct mlx5_flow_spec *spec)
--
1.8.3.1
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH net-next 2/3] net/mlx5e: Introduce mlx5e_eswitch_rep_uplink_priv helper
2020-04-28 5:24 [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary xiangxia.m.yue
@ 2020-04-28 5:24 ` xiangxia.m.yue
2020-04-30 11:17 ` Roi Dayan
2020-04-28 5:24 ` [PATCH net-next 3/3] net/mlx5e: Fix the code style xiangxia.m.yue
` (2 subsequent siblings)
3 siblings, 1 reply; 8+ messages in thread
From: xiangxia.m.yue @ 2020-04-28 5:24 UTC (permalink / raw)
To: paulb, saeedm, roid, gerlitz.or; +Cc: netdev, Tonghao Zhang
From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
Introduce the mlx5e_eswitch_rep_uplink_priv helper
to make the codes readable.
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 4 +--
drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 9 +++++++
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 30 +++++-----------------
3 files changed, 17 insertions(+), 26 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
index 16416eaac39e..c5d5e69ff147 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
@@ -100,10 +100,8 @@ struct mlx5_ct_entry {
{
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *uplink_rpriv;
- uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &uplink_rpriv->uplink_priv;
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
return uplink_priv->ct_priv;
}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
index 6a2337900420..899ffa0872b7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -109,6 +109,15 @@ struct mlx5e_rep_priv *mlx5e_rep_to_rep_priv(struct mlx5_eswitch_rep *rep)
return rep->rep_data[REP_ETH].priv;
}
+static inline struct mlx5_rep_uplink_priv *
+mlx5e_eswitch_rep_uplink_priv(struct mlx5_eswitch *esw)
+{
+ struct mlx5e_rep_priv *priv;
+
+ priv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+ return &priv->uplink_priv;
+}
+
struct mlx5e_neigh {
struct net_device *dev;
union {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 64f5c3f3dbb3..696544e2a25b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1250,12 +1250,10 @@ static void unready_flow_del(struct mlx5e_tc_flow *flow)
static void add_unready_flow(struct mlx5e_tc_flow *flow)
{
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *rpriv;
struct mlx5_eswitch *esw;
esw = flow->priv->mdev->priv.eswitch;
- rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &rpriv->uplink_priv;
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
mutex_lock(&uplink_priv->unready_flows_lock);
unready_flow_add(flow, &uplink_priv->unready_flows);
@@ -1265,12 +1263,10 @@ static void add_unready_flow(struct mlx5e_tc_flow *flow)
static void remove_unready_flow(struct mlx5e_tc_flow *flow)
{
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *rpriv;
struct mlx5_eswitch *esw;
esw = flow->priv->mdev->priv.eswitch;
- rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &rpriv->uplink_priv;
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
mutex_lock(&uplink_priv->unready_flows_lock);
unready_flow_del(flow);
@@ -1888,7 +1884,6 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts;
struct flow_match_enc_opts enc_opts_match;
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *uplink_rpriv;
struct tunnel_match_key tunnel_key;
bool enc_opts_is_dont_care = true;
u32 tun_id, enc_opts_id = 0;
@@ -1897,8 +1892,7 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
int err;
esw = priv->mdev->priv.eswitch;
- uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &uplink_rpriv->uplink_priv;
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
err = mapping_add(uplink_priv->tunnel_mapping, &tunnel_key, &tun_id);
@@ -1957,13 +1951,11 @@ static int mlx5e_lookup_flow_tunnel_id(struct mlx5e_priv *priv,
u32 *tun_id)
{
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *uplink_rpriv;
struct tunnel_match_key tunnel_key;
struct mlx5_eswitch *esw;
esw = priv->mdev->priv.eswitch;
- uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &uplink_rpriv->uplink_priv;
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
return mapping_find_by_data(uplink_priv->tunnel_mapping, &tunnel_key, tun_id);
@@ -1974,12 +1966,10 @@ static void mlx5e_put_flow_tunnel_id(struct mlx5e_tc_flow *flow)
u32 enc_opts_id = flow->tunnel_id & ENC_OPTS_BITS_MASK;
u32 tun_id = flow->tunnel_id >> ENC_OPTS_BITS;
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *uplink_rpriv;
struct mlx5_eswitch *esw;
esw = flow->priv->mdev->priv.eswitch;
- uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &uplink_rpriv->uplink_priv;
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
if (tun_id)
mapping_remove(uplink_priv->tunnel_mapping, tun_id);
@@ -4841,7 +4831,6 @@ static bool mlx5e_restore_tunnel(struct mlx5e_priv *priv, struct sk_buff *skb,
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
struct flow_dissector_key_enc_opts enc_opts = {};
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *uplink_rpriv;
struct metadata_dst *tun_dst;
struct tunnel_match_key key;
u32 tun_id, enc_opts_id;
@@ -4854,9 +4843,7 @@ static bool mlx5e_restore_tunnel(struct mlx5e_priv *priv, struct sk_buff *skb,
if (!tun_id)
return true;
- uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &uplink_rpriv->uplink_priv;
-
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
err = mapping_find(uplink_priv->tunnel_mapping, tun_id, &key);
if (err) {
WARN_ON_ONCE(true);
@@ -4920,7 +4907,6 @@ bool mlx5e_tc_rep_update_skb(struct mlx5_cqe64 *cqe,
#if IS_ENABLED(CONFIG_NET_TC_SKB_EXT)
u32 chain = 0, reg_c0, reg_c1, tunnel_id, tuple_id;
struct mlx5_rep_uplink_priv *uplink_priv;
- struct mlx5e_rep_priv *uplink_rpriv;
struct tc_skb_ext *tc_skb_ext;
struct mlx5_eswitch *esw;
struct mlx5e_priv *priv;
@@ -4956,9 +4942,7 @@ bool mlx5e_tc_rep_update_skb(struct mlx5_cqe64 *cqe,
tc_skb_ext->chain = chain;
tuple_id = reg_c1 & TUPLE_ID_MAX;
-
- uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
- uplink_priv = &uplink_rpriv->uplink_priv;
+ uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
if (!mlx5e_tc_ct_restore_flow(uplink_priv, skb, tuple_id))
return false;
}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH net-next 3/3] net/mlx5e: Fix the code style
2020-04-28 5:24 [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary xiangxia.m.yue
2020-04-28 5:24 ` [PATCH net-next 2/3] net/mlx5e: Introduce mlx5e_eswitch_rep_uplink_priv helper xiangxia.m.yue
@ 2020-04-28 5:24 ` xiangxia.m.yue
2020-04-30 11:19 ` Roi Dayan
2020-04-28 19:57 ` [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary Saeed Mahameed
2020-04-30 15:26 ` Roi Dayan
3 siblings, 1 reply; 8+ messages in thread
From: xiangxia.m.yue @ 2020-04-28 5:24 UTC (permalink / raw)
To: paulb, saeedm, roid, gerlitz.or; +Cc: netdev, Tonghao Zhang
From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 55457f268495..6b68f47e7024 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -1367,7 +1367,7 @@ static bool mlx5e_rep_has_offload_stats(const struct net_device *dev, int attr_i
{
switch (attr_id) {
case IFLA_OFFLOAD_XSTATS_CPU_HIT:
- return true;
+ return true;
}
return false;
--
1.8.3.1
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary
2020-04-28 5:24 [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary xiangxia.m.yue
2020-04-28 5:24 ` [PATCH net-next 2/3] net/mlx5e: Introduce mlx5e_eswitch_rep_uplink_priv helper xiangxia.m.yue
2020-04-28 5:24 ` [PATCH net-next 3/3] net/mlx5e: Fix the code style xiangxia.m.yue
@ 2020-04-28 19:57 ` Saeed Mahameed
2020-04-30 15:26 ` Roi Dayan
3 siblings, 0 replies; 8+ messages in thread
From: Saeed Mahameed @ 2020-04-28 19:57 UTC (permalink / raw)
To: Roi Dayan, Paul Blakey, Oz Shlomo, xiangxia.m.yue, gerlitz.or; +Cc: netdev
On Tue, 2020-04-28 at 13:24 +0800, xiangxia.m.yue@gmail.com wrote:
> From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
>
> The commit 0a7fcb78cc21 ("net/mlx5e: Support inner header rewrite
> with
> goto action"), will decapitate the tunnel packets if there is a goto
> action in chain 0. But in some case, we don't want do that, for
> example:
>
Roi, Paul, Oz, please review..
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower dst_mac 00:11:22:33:44:55 enc_src_ip 2.2.2.200
> \
> enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100
> \
> action tunnel_key unset action mirred egress redirect dev
> enp130s0f0_0
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower dst_mac 00:11:22:33:44:66 enc_src_ip 2.2.2.200
> \
> enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 200
> \
> action tunnel_key unset action mirred egress redirect dev
> enp130s0f0_1
>
> In this patch, if there is a pedit action in chain, do the
> decapitation action.
> if there are pedit and goto actions, do the decapitation and id
> mapping action.
>
> 8 test units:
> [1]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action tunnel_key unset \
> action mirred egress redirect dev enp130s0f0_0
> [2]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action pedit ex munge eth src set 00:11:22:33:44:f0
> \
> action mirred egress redirect dev enp130s0f0_0
> [3]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action tunnel_key unset \
> action pedit ex munge eth src set 00:11:22:33:44:f0
> \
> action mirred egress redirect dev enp130s0f0_0
> [4]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789
> \
> enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:ff pipe \
> action mirred egress redirect dev enp130s0f0_0
> [5]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action pedit ex munge eth src set 00:11:22:33:44:ff
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff
> \
> action tunnel_key unset \
> action mirred egress redirect dev enp130s0f0_0
> [6]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action pedit ex munge eth src set 00:11:22:33:44:ff
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff
> \
> action tunnel_key unset \
> action pedit ex munge eth src set 00:11:22:33:44:f0
> \
> action mirred egress redirect dev enp130s0f0_0
> [7]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action pedit ex munge eth src set 00:11:22:33:44:ff
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff
> \
> action pedit ex munge eth src set 00:11:22:33:44:f0
> \
> action goto chain 3
> $ tc filter add dev vxlan0 protocol ip parent ffff: prio 1 chain 3
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:f0
> \
> action tunnel_key unset \
> action pedit ex munge eth src set 00:11:22:33:44:f1
> \
> action mirred egress redirect dev enp130s0f0_0
> [8]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0
> \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789
> \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action pedit ex munge eth src set 00:11:22:33:44:f0
> \
> action goto chain 3
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 3
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action pedit ex munge eth src set 00:11:22:33:44:f1
> \
> action goto chain 4
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 4
> \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55
> \
> action pedit ex munge eth src set 00:11:22:33:44:f2
> \
> action mirred egress redirect dev enp130s0f0_0
>
> Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
> ---
> .../net/ethernet/mellanox/mlx5/core/en/mapping.c | 24 ++++++
> .../net/ethernet/mellanox/mlx5/core/en/mapping.h | 1 +
> drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 97
> +++++++++++++++-------
> 3 files changed, 92 insertions(+), 30 deletions(-)
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> index ea321e528749..90306dde6b60 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> @@ -74,6 +74,30 @@ int mapping_add(struct mapping_ctx *ctx, void
> *data, u32 *id)
> return err;
> }
>
> +int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32
> *id)
> +{
> + struct mapping_item *mi;
> + u32 hash_key;
> +
> + mutex_lock(&ctx->lock);
> +
> + hash_key = jhash(data, ctx->data_size, 0);
> + hash_for_each_possible(ctx->ht, mi, node, hash_key) {
> + if (!memcmp(data, mi->data, ctx->data_size))
> + goto found;
> + }
> +
> + mutex_unlock(&ctx->lock);
> + return -ENOENT;
> +
> +found:
> + if (id)
> + *id = mi->id;
> +
> + mutex_unlock(&ctx->lock);
> + return 0;
> +}
> +
> static void mapping_remove_and_free(struct mapping_ctx *ctx,
> struct mapping_item *mi)
> {
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> index 285525cc5470..af501c9796b7 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> @@ -9,6 +9,7 @@
> int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id);
> int mapping_remove(struct mapping_ctx *ctx, u32 id);
> int mapping_find(struct mapping_ctx *ctx, u32 id, void *data);
> +int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32
> *id);
>
> /* mapping uses an xarray to map data to ids in add(), and for
> find().
> * For locking, it uses a internal xarray spin lock for
> add()/remove(),
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> index a574c588269a..64f5c3f3dbb3 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> @@ -1786,7 +1786,8 @@ static void mlx5e_tc_del_flow(struct mlx5e_priv
> *priv,
> }
> }
>
> -static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
> +static int flow_has_tc_action(struct flow_cls_offload *f,
> + enum flow_action_id action)
> {
> struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> struct flow_action *flow_action = &rule->action;
> @@ -1794,12 +1795,8 @@ static int flow_has_tc_fwd_action(struct
> flow_cls_offload *f)
> int i;
>
> flow_action_for_each(i, act, flow_action) {
> - switch (act->id) {
> - case FLOW_ACTION_GOTO:
> + if (act->id == action)
> return true;
> - default:
> - continue;
> - }
> }
>
> return false;
> @@ -1853,10 +1850,37 @@ static int flow_has_tc_fwd_action(struct
> flow_cls_offload *f)
> sizeof(*__dst));\
> })
>
> +static void mlx5e_make_tunnel_match_key(struct flow_cls_offload *f,
> + struct net_device *filter_dev,
> + struct tunnel_match_key
> *tunnel_key)
> +{
> + struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> +
> + memset(tunnel_key, 0, sizeof(*tunnel_key));
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
> + &tunnel_key->enc_control);
> + if (tunnel_key->enc_control.addr_type ==
> FLOW_DISSECTOR_KEY_IPV4_ADDRS)
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
> + &tunnel_key->enc_ipv4);
> + else
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
> + &tunnel_key->enc_ipv6);
> +
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP, &tunnel_key-
> >enc_ip);
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
> + &tunnel_key->enc_tp);
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
> + &tunnel_key->enc_key_id);
> +
> + tunnel_key->filter_ifindex = filter_dev->ifindex;
> +}
> +
> static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> struct mlx5e_tc_flow *flow,
> struct flow_cls_offload *f,
> - struct net_device *filter_dev)
> + struct net_device *filter_dev,
> + bool sets_mapping,
> + bool needs_mapping)
> {
> struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> struct netlink_ext_ack *extack = f->common.extack;
> @@ -1876,22 +1900,7 @@ static int mlx5e_get_flow_tunnel_id(struct
> mlx5e_priv *priv,
> uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> uplink_priv = &uplink_rpriv->uplink_priv;
>
> - memset(&tunnel_key, 0, sizeof(tunnel_key));
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
> - &tunnel_key.enc_control);
> - if (tunnel_key.enc_control.addr_type ==
> FLOW_DISSECTOR_KEY_IPV4_ADDRS)
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
> - &tunnel_key.enc_ipv4);
> - else
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
> - &tunnel_key.enc_ipv6);
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP,
> &tunnel_key.enc_ip);
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
> - &tunnel_key.enc_tp);
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
> - &tunnel_key.enc_key_id);
> - tunnel_key.filter_ifindex = filter_dev->ifindex;
> -
> + mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> err = mapping_add(uplink_priv->tunnel_mapping, &tunnel_key,
> &tun_id);
> if (err)
> return err;
> @@ -1915,10 +1924,10 @@ static int mlx5e_get_flow_tunnel_id(struct
> mlx5e_priv *priv,
> mask = enc_opts_id ? TUNNEL_ID_MASK :
> (TUNNEL_ID_MASK & ~ENC_OPTS_BITS_MASK);
>
> - if (attr->chain) {
> + if (needs_mapping) {
> mlx5e_tc_match_to_reg_match(&attr->parse_attr->spec,
> TUNNEL_TO_REG, value,
> mask);
> - } else {
> + } else if (sets_mapping) {
> mod_hdr_acts = &attr->parse_attr->mod_hdr_acts;
> err = mlx5e_tc_match_to_reg_set(priv->mdev,
> mod_hdr_acts,
> @@ -1941,6 +1950,25 @@ static int mlx5e_get_flow_tunnel_id(struct
> mlx5e_priv *priv,
> return err;
> }
>
> +static int mlx5e_lookup_flow_tunnel_id(struct mlx5e_priv *priv,
> + struct mlx5e_tc_flow *flow,
> + struct flow_cls_offload *f,
> + struct net_device *filter_dev,
> + u32 *tun_id)
> +{
> + struct mlx5_rep_uplink_priv *uplink_priv;
> + struct mlx5e_rep_priv *uplink_rpriv;
> + struct tunnel_match_key tunnel_key;
> + struct mlx5_eswitch *esw;
> +
> + esw = priv->mdev->priv.eswitch;
> + uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> + uplink_priv = &uplink_rpriv->uplink_priv;
> +
> + mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> + return mapping_find_by_data(uplink_priv->tunnel_mapping,
> &tunnel_key, tun_id);
> +}
> +
> static void mlx5e_put_flow_tunnel_id(struct mlx5e_tc_flow *flow)
> {
> u32 enc_opts_id = flow->tunnel_id & ENC_OPTS_BITS_MASK;
> @@ -1976,14 +2004,22 @@ static int parse_tunnel_attr(struct
> mlx5e_priv *priv,
> struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
> struct netlink_ext_ack *extack = f->common.extack;
> bool needs_mapping, sets_mapping;
> + bool pedit_action;
> int err;
>
> if (!mlx5e_is_eswitch_flow(flow))
> return -EOPNOTSUPP;
>
> - needs_mapping = !!flow->esw_attr->chain;
> - sets_mapping = !flow->esw_attr->chain &&
> flow_has_tc_fwd_action(f);
> - *match_inner = !needs_mapping;
> + pedit_action = flow_has_tc_action(f, FLOW_ACTION_MANGLE) ||
> + flow_has_tc_action(f, FLOW_ACTION_ADD);
> +
> + *match_inner = pedit_action;
> + sets_mapping = pedit_action &&
> + flow_has_tc_action(f, FLOW_ACTION_GOTO);
> +
> + needs_mapping = !!flow->esw_attr->chain &&
> + !mlx5e_lookup_flow_tunnel_id(priv, flow, f,
> + filter_dev, NULL);
>
> if ((needs_mapping || sets_mapping) &&
> !mlx5_eswitch_reg_c1_loopback_enabled(esw)) {
> @@ -1994,7 +2030,7 @@ static int parse_tunnel_attr(struct mlx5e_priv
> *priv,
> return -EOPNOTSUPP;
> }
>
> - if (!flow->esw_attr->chain) {
> + if (*match_inner && !needs_mapping) {
> err = mlx5e_tc_tun_parse(filter_dev, priv, spec, f,
> match_level);
> if (err) {
> @@ -2011,7 +2047,8 @@ static int parse_tunnel_attr(struct mlx5e_priv
> *priv,
> if (!needs_mapping && !sets_mapping)
> return 0;
>
> - return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev);
> + return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev,
> + sets_mapping, needs_mapping);
> }
>
> static void *get_match_inner_headers_criteria(struct mlx5_flow_spec
> *spec)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH net-next 2/3] net/mlx5e: Introduce mlx5e_eswitch_rep_uplink_priv helper
2020-04-28 5:24 ` [PATCH net-next 2/3] net/mlx5e: Introduce mlx5e_eswitch_rep_uplink_priv helper xiangxia.m.yue
@ 2020-04-30 11:17 ` Roi Dayan
0 siblings, 0 replies; 8+ messages in thread
From: Roi Dayan @ 2020-04-30 11:17 UTC (permalink / raw)
To: xiangxia.m.yue, paulb, saeedm, gerlitz.or; +Cc: netdev
On 2020-04-28 8:24 AM, xiangxia.m.yue@gmail.com wrote:
> From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
>
> Introduce the mlx5e_eswitch_rep_uplink_priv helper
> to make the codes readable.
>
> Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
> ---
> drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 4 +--
> drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 9 +++++++
> drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 30 +++++-----------------
> 3 files changed, 17 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
> index 16416eaac39e..c5d5e69ff147 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
> @@ -100,10 +100,8 @@ struct mlx5_ct_entry {
> {
> struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *uplink_rpriv;
>
> - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &uplink_rpriv->uplink_priv;
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
> return uplink_priv->ct_priv;
> }
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
> index 6a2337900420..899ffa0872b7 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
> @@ -109,6 +109,15 @@ struct mlx5e_rep_priv *mlx5e_rep_to_rep_priv(struct mlx5_eswitch_rep *rep)
> return rep->rep_data[REP_ETH].priv;
> }
>
> +static inline struct mlx5_rep_uplink_priv *
> +mlx5e_eswitch_rep_uplink_priv(struct mlx5_eswitch *esw)
> +{
> + struct mlx5e_rep_priv *priv;
> +
> + priv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> + return &priv->uplink_priv;
> +}
> +
since its in en_rep.h, please use "mlx5e_rep_" prefix.
thanks
> struct mlx5e_neigh {
> struct net_device *dev;
> union {
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> index 64f5c3f3dbb3..696544e2a25b 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> @@ -1250,12 +1250,10 @@ static void unready_flow_del(struct mlx5e_tc_flow *flow)
> static void add_unready_flow(struct mlx5e_tc_flow *flow)
> {
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *rpriv;
> struct mlx5_eswitch *esw;
>
> esw = flow->priv->mdev->priv.eswitch;
> - rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &rpriv->uplink_priv;
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
>
> mutex_lock(&uplink_priv->unready_flows_lock);
> unready_flow_add(flow, &uplink_priv->unready_flows);
> @@ -1265,12 +1263,10 @@ static void add_unready_flow(struct mlx5e_tc_flow *flow)
> static void remove_unready_flow(struct mlx5e_tc_flow *flow)
> {
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *rpriv;
> struct mlx5_eswitch *esw;
>
> esw = flow->priv->mdev->priv.eswitch;
> - rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &rpriv->uplink_priv;
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
>
> mutex_lock(&uplink_priv->unready_flows_lock);
> unready_flow_del(flow);
> @@ -1888,7 +1884,6 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts;
> struct flow_match_enc_opts enc_opts_match;
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *uplink_rpriv;
> struct tunnel_match_key tunnel_key;
> bool enc_opts_is_dont_care = true;
> u32 tun_id, enc_opts_id = 0;
> @@ -1897,8 +1892,7 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> int err;
>
> esw = priv->mdev->priv.eswitch;
> - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &uplink_rpriv->uplink_priv;
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
>
> mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> err = mapping_add(uplink_priv->tunnel_mapping, &tunnel_key, &tun_id);
> @@ -1957,13 +1951,11 @@ static int mlx5e_lookup_flow_tunnel_id(struct mlx5e_priv *priv,
> u32 *tun_id)
> {
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *uplink_rpriv;
> struct tunnel_match_key tunnel_key;
> struct mlx5_eswitch *esw;
>
> esw = priv->mdev->priv.eswitch;
> - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &uplink_rpriv->uplink_priv;
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
>
> mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> return mapping_find_by_data(uplink_priv->tunnel_mapping, &tunnel_key, tun_id);
> @@ -1974,12 +1966,10 @@ static void mlx5e_put_flow_tunnel_id(struct mlx5e_tc_flow *flow)
> u32 enc_opts_id = flow->tunnel_id & ENC_OPTS_BITS_MASK;
> u32 tun_id = flow->tunnel_id >> ENC_OPTS_BITS;
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *uplink_rpriv;
> struct mlx5_eswitch *esw;
>
> esw = flow->priv->mdev->priv.eswitch;
> - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &uplink_rpriv->uplink_priv;
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
>
> if (tun_id)
> mapping_remove(uplink_priv->tunnel_mapping, tun_id);
> @@ -4841,7 +4831,6 @@ static bool mlx5e_restore_tunnel(struct mlx5e_priv *priv, struct sk_buff *skb,
> struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
> struct flow_dissector_key_enc_opts enc_opts = {};
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *uplink_rpriv;
> struct metadata_dst *tun_dst;
> struct tunnel_match_key key;
> u32 tun_id, enc_opts_id;
> @@ -4854,9 +4843,7 @@ static bool mlx5e_restore_tunnel(struct mlx5e_priv *priv, struct sk_buff *skb,
> if (!tun_id)
> return true;
>
> - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &uplink_rpriv->uplink_priv;
> -
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
> err = mapping_find(uplink_priv->tunnel_mapping, tun_id, &key);
> if (err) {
> WARN_ON_ONCE(true);
> @@ -4920,7 +4907,6 @@ bool mlx5e_tc_rep_update_skb(struct mlx5_cqe64 *cqe,
> #if IS_ENABLED(CONFIG_NET_TC_SKB_EXT)
> u32 chain = 0, reg_c0, reg_c1, tunnel_id, tuple_id;
> struct mlx5_rep_uplink_priv *uplink_priv;
> - struct mlx5e_rep_priv *uplink_rpriv;
> struct tc_skb_ext *tc_skb_ext;
> struct mlx5_eswitch *esw;
> struct mlx5e_priv *priv;
> @@ -4956,9 +4942,7 @@ bool mlx5e_tc_rep_update_skb(struct mlx5_cqe64 *cqe,
> tc_skb_ext->chain = chain;
>
> tuple_id = reg_c1 & TUPLE_ID_MAX;
> -
> - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> - uplink_priv = &uplink_rpriv->uplink_priv;
> + uplink_priv = mlx5e_eswitch_rep_uplink_priv(esw);
> if (!mlx5e_tc_ct_restore_flow(uplink_priv, skb, tuple_id))
> return false;
> }
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH net-next 3/3] net/mlx5e: Fix the code style
2020-04-28 5:24 ` [PATCH net-next 3/3] net/mlx5e: Fix the code style xiangxia.m.yue
@ 2020-04-30 11:19 ` Roi Dayan
0 siblings, 0 replies; 8+ messages in thread
From: Roi Dayan @ 2020-04-30 11:19 UTC (permalink / raw)
To: xiangxia.m.yue, paulb, saeedm, gerlitz.or; +Cc: netdev
On 2020-04-28 8:24 AM, xiangxia.m.yue@gmail.com wrote:
> From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
>
> Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
> ---
> drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> index 55457f268495..6b68f47e7024 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
> @@ -1367,7 +1367,7 @@ static bool mlx5e_rep_has_offload_stats(const struct net_device *dev, int attr_i
> {
> switch (attr_id) {
> case IFLA_OFFLOAD_XSTATS_CPU_HIT:
> - return true;
> + return true;
> }
>
> return false;
>
Reviewed-by: Roi Dayan <roid@mellanox.com>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary
2020-04-28 5:24 [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary xiangxia.m.yue
` (2 preceding siblings ...)
2020-04-28 19:57 ` [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary Saeed Mahameed
@ 2020-04-30 15:26 ` Roi Dayan
2020-05-06 1:20 ` Tonghao Zhang
3 siblings, 1 reply; 8+ messages in thread
From: Roi Dayan @ 2020-04-30 15:26 UTC (permalink / raw)
To: xiangxia.m.yue, paulb, saeedm, gerlitz.or; +Cc: netdev
On 2020-04-28 8:24 AM, xiangxia.m.yue@gmail.com wrote:
> From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
>
> The commit 0a7fcb78cc21 ("net/mlx5e: Support inner header rewrite with
> goto action"), will decapitate the tunnel packets if there is a goto
> action in chain 0. But in some case, we don't want do that, for example:
Hi Zhang,
Thanks for your commit. i'll run more tests so i might have some comments later.
Can you use decapsulate instead of decapitate please.
>
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower dst_mac 00:11:22:33:44:55 enc_src_ip 2.2.2.200 \
> enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
> action tunnel_key unset action mirred egress redirect dev enp130s0f0_0
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower dst_mac 00:11:22:33:44:66 enc_src_ip 2.2.2.200 \
> enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 200 \
> action tunnel_key unset action mirred egress redirect dev enp130s0f0_1
>
> In this patch, if there is a pedit action in chain, do the decapitation action.
decapsulation
> if there are pedit and goto actions, do the decapitation and id mapping action.
>
> 8 test units:
> [1]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action tunnel_key unset \
> action mirred egress redirect dev enp130s0f0_0
> [2]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:f0 \
> action mirred egress redirect dev enp130s0f0_0
> [3]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action tunnel_key unset \
> action pedit ex munge eth src set 00:11:22:33:44:f0 \
> action mirred egress redirect dev enp130s0f0_0
> [4]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:ff pipe \
> action mirred egress redirect dev enp130s0f0_0
what about the case of only chain 0 without pedit?
I think now we skip matching tun? see comment below in parse_tunnel_attr().
> [5]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:ff \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
> action tunnel_key unset \
> action mirred egress redirect dev enp130s0f0_0
> [6]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:ff \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
> action tunnel_key unset \
> action pedit ex munge eth src set 00:11:22:33:44:f0 \
> action mirred egress redirect dev enp130s0f0_0
> [7]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:ff \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
> action pedit ex munge eth src set 00:11:22:33:44:f0 \
> action goto chain 3
> $ tc filter add dev vxlan0 protocol ip parent ffff: prio 1 chain 3 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:f0 \
> action tunnel_key unset \
> action pedit ex munge eth src set 00:11:22:33:44:f1 \
> action mirred egress redirect dev enp130s0f0_0
> [8]:
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> action goto chain 2
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:f0 \
> action goto chain 3
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 3 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:f1 \
> action goto chain 4
> $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 4 \
> flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> action pedit ex munge eth src set 00:11:22:33:44:f2 \
> action mirred egress redirect dev enp130s0f0_0
>
> Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
> ---
> .../net/ethernet/mellanox/mlx5/core/en/mapping.c | 24 ++++++
> .../net/ethernet/mellanox/mlx5/core/en/mapping.h | 1 +
> drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 97 +++++++++++++++-------
> 3 files changed, 92 insertions(+), 30 deletions(-)
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> index ea321e528749..90306dde6b60 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> @@ -74,6 +74,30 @@ int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id)
> return err;
> }
>
> +int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32 *id)
> +{
> + struct mapping_item *mi;
> + u32 hash_key;
> +
> + mutex_lock(&ctx->lock);
> +
> + hash_key = jhash(data, ctx->data_size, 0);
> + hash_for_each_possible(ctx->ht, mi, node, hash_key) {
> + if (!memcmp(data, mi->data, ctx->data_size))
> + goto found;
> + }
> +
> + mutex_unlock(&ctx->lock);
> + return -ENOENT;
> +
> +found:
> + if (id)
> + *id = mi->id;
> +
> + mutex_unlock(&ctx->lock);
> + return 0;
> +}
> +
> static void mapping_remove_and_free(struct mapping_ctx *ctx,
> struct mapping_item *mi)
> {
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> index 285525cc5470..af501c9796b7 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> @@ -9,6 +9,7 @@
> int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id);
> int mapping_remove(struct mapping_ctx *ctx, u32 id);
> int mapping_find(struct mapping_ctx *ctx, u32 id, void *data);
> +int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32 *id);
>
> /* mapping uses an xarray to map data to ids in add(), and for find().
> * For locking, it uses a internal xarray spin lock for add()/remove(),
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> index a574c588269a..64f5c3f3dbb3 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> @@ -1786,7 +1786,8 @@ static void mlx5e_tc_del_flow(struct mlx5e_priv *priv,
> }
> }
>
> -static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
> +static int flow_has_tc_action(struct flow_cls_offload *f,
> + enum flow_action_id action)
> {
> struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> struct flow_action *flow_action = &rule->action;
> @@ -1794,12 +1795,8 @@ static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
> int i;
>
> flow_action_for_each(i, act, flow_action) {
> - switch (act->id) {
> - case FLOW_ACTION_GOTO:
> + if (act->id == action)
> return true;
> - default:
> - continue;
> - }
> }
>
> return false;
> @@ -1853,10 +1850,37 @@ static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
> sizeof(*__dst));\
> })
>
> +static void mlx5e_make_tunnel_match_key(struct flow_cls_offload *f,
> + struct net_device *filter_dev,
> + struct tunnel_match_key *tunnel_key)
> +{
> + struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> +
> + memset(tunnel_key, 0, sizeof(*tunnel_key));
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
> + &tunnel_key->enc_control);
> + if (tunnel_key->enc_control.addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS)
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
> + &tunnel_key->enc_ipv4);
> + else
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
> + &tunnel_key->enc_ipv6);
> +
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP, &tunnel_key->enc_ip);
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
> + &tunnel_key->enc_tp);
> + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
> + &tunnel_key->enc_key_id);
> +
> + tunnel_key->filter_ifindex = filter_dev->ifindex;
> +}
> +
> static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> struct mlx5e_tc_flow *flow,
> struct flow_cls_offload *f,
> - struct net_device *filter_dev)
> + struct net_device *filter_dev,
> + bool sets_mapping,
> + bool needs_mapping)
> {
> struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> struct netlink_ext_ack *extack = f->common.extack;
> @@ -1876,22 +1900,7 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> uplink_priv = &uplink_rpriv->uplink_priv;
>
> - memset(&tunnel_key, 0, sizeof(tunnel_key));
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
> - &tunnel_key.enc_control);
> - if (tunnel_key.enc_control.addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS)
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
> - &tunnel_key.enc_ipv4);
> - else
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
> - &tunnel_key.enc_ipv6);
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP, &tunnel_key.enc_ip);
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
> - &tunnel_key.enc_tp);
> - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
> - &tunnel_key.enc_key_id);
> - tunnel_key.filter_ifindex = filter_dev->ifindex;
> -
> + mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> err = mapping_add(uplink_priv->tunnel_mapping, &tunnel_key, &tun_id);
> if (err)
> return err;
> @@ -1915,10 +1924,10 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> mask = enc_opts_id ? TUNNEL_ID_MASK :
> (TUNNEL_ID_MASK & ~ENC_OPTS_BITS_MASK);
>
> - if (attr->chain) {
> + if (needs_mapping) {
> mlx5e_tc_match_to_reg_match(&attr->parse_attr->spec,
> TUNNEL_TO_REG, value, mask);
> - } else {
> + } else if (sets_mapping) {
> mod_hdr_acts = &attr->parse_attr->mod_hdr_acts;
> err = mlx5e_tc_match_to_reg_set(priv->mdev,
> mod_hdr_acts,
> @@ -1941,6 +1950,25 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> return err;
> }
>
> +static int mlx5e_lookup_flow_tunnel_id(struct mlx5e_priv *priv,
> + struct mlx5e_tc_flow *flow,
> + struct flow_cls_offload *f,
> + struct net_device *filter_dev,
> + u32 *tun_id)
> +{
> + struct mlx5_rep_uplink_priv *uplink_priv;
> + struct mlx5e_rep_priv *uplink_rpriv;
> + struct tunnel_match_key tunnel_key;
> + struct mlx5_eswitch *esw;
> +
> + esw = priv->mdev->priv.eswitch;
> + uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> + uplink_priv = &uplink_rpriv->uplink_priv;
> +
> + mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> + return mapping_find_by_data(uplink_priv->tunnel_mapping, &tunnel_key, tun_id);
> +}
> +
> static void mlx5e_put_flow_tunnel_id(struct mlx5e_tc_flow *flow)
> {
> u32 enc_opts_id = flow->tunnel_id & ENC_OPTS_BITS_MASK;
> @@ -1976,14 +2004,22 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
> struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
> struct netlink_ext_ack *extack = f->common.extack;
> bool needs_mapping, sets_mapping;
> + bool pedit_action;
> int err;
>
> if (!mlx5e_is_eswitch_flow(flow))
> return -EOPNOTSUPP;
>
> - needs_mapping = !!flow->esw_attr->chain;
> - sets_mapping = !flow->esw_attr->chain && flow_has_tc_fwd_action(f);
> - *match_inner = !needs_mapping;
> + pedit_action = flow_has_tc_action(f, FLOW_ACTION_MANGLE) ||
> + flow_has_tc_action(f, FLOW_ACTION_ADD);
> +
> + *match_inner = pedit_action;
> + sets_mapping = pedit_action &&
> + flow_has_tc_action(f, FLOW_ACTION_GOTO);
> +
> + needs_mapping = !!flow->esw_attr->chain &&
> + !mlx5e_lookup_flow_tunnel_id(priv, flow, f,
> + filter_dev, NULL);
>
> if ((needs_mapping || sets_mapping) &&
> !mlx5_eswitch_reg_c1_loopback_enabled(esw)) {
> @@ -1994,7 +2030,7 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
> return -EOPNOTSUPP;
> }
>
> - if (!flow->esw_attr->chain) {
> + if (*match_inner && !needs_mapping) {
why you only get inside for match_inner?
what about the case of matching tunnel on chain 0 without pedit action?
we need to call mlx5e_tc_tun_parse()
> err = mlx5e_tc_tun_parse(filter_dev, priv, spec, f,
> match_level);
> if (err) {
> @@ -2011,7 +2047,8 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
> if (!needs_mapping && !sets_mapping)
> return 0;
>
> - return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev);
> + return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev,
> + sets_mapping, needs_mapping);
> }
>
> static void *get_match_inner_headers_criteria(struct mlx5_flow_spec *spec)
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary
2020-04-30 15:26 ` Roi Dayan
@ 2020-05-06 1:20 ` Tonghao Zhang
0 siblings, 0 replies; 8+ messages in thread
From: Tonghao Zhang @ 2020-05-06 1:20 UTC (permalink / raw)
To: Roi Dayan
Cc: Paul Blakey, Saeed Mahameed, Or Gerlitz, Linux Kernel Network Developers
On Thu, Apr 30, 2020 at 11:26 PM Roi Dayan <roid@mellanox.com> wrote:
>
>
>
> On 2020-04-28 8:24 AM, xiangxia.m.yue@gmail.com wrote:
> > From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
> >
> > The commit 0a7fcb78cc21 ("net/mlx5e: Support inner header rewrite with
> > goto action"), will decapitate the tunnel packets if there is a goto
> > action in chain 0. But in some case, we don't want do that, for example:
>
> Hi Zhang,
>
> Thanks for your commit. i'll run more tests so i might have some comments later.
> Can you use decapsulate instead of decapitate please.
>
> >
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower dst_mac 00:11:22:33:44:55 enc_src_ip 2.2.2.200 \
> > enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
> > action tunnel_key unset action mirred egress redirect dev enp130s0f0_0
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower dst_mac 00:11:22:33:44:66 enc_src_ip 2.2.2.200 \
> > enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 200 \
> > action tunnel_key unset action mirred egress redirect dev enp130s0f0_1
> >
> > In this patch, if there is a pedit action in chain, do the decapitation action.
>
> decapsulation
>
> > if there are pedit and goto actions, do the decapitation and id mapping action.
> >
> > 8 test units:
> > [1]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action tunnel_key unset \
> > action mirred egress redirect dev enp130s0f0_0
> > [2]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:f0 \
> > action mirred egress redirect dev enp130s0f0_0
> > [3]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 enc_key_id 100 \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action tunnel_key unset \
> > action pedit ex munge eth src set 00:11:22:33:44:f0 \
> > action mirred egress redirect dev enp130s0f0_0
> > [4]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> > enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:ff pipe \
> > action mirred egress redirect dev enp130s0f0_0
>
> what about the case of only chain 0 without pedit?
> I think now we skip matching tun? see comment below in parse_tunnel_attr().
>
> > [5]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:ff \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
> > action tunnel_key unset \
> > action mirred egress redirect dev enp130s0f0_0
> > [6]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:ff \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
> > action tunnel_key unset \
> > action pedit ex munge eth src set 00:11:22:33:44:f0 \
> > action mirred egress redirect dev enp130s0f0_0
> > [7]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:ff \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:ff \
> > action pedit ex munge eth src set 00:11:22:33:44:f0 \
> > action goto chain 3
> > $ tc filter add dev vxlan0 protocol ip parent ffff: prio 1 chain 3 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 src_mac 00:11:22:33:44:f0 \
> > action tunnel_key unset \
> > action pedit ex munge eth src set 00:11:22:33:44:f1 \
> > action mirred egress redirect dev enp130s0f0_0
> > [8]:
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 0 \
> > flower enc_dst_ip 2.2.2.100 enc_dst_port 4789 \
> > action goto chain 2
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 2 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:f0 \
> > action goto chain 3
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 3 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:f1 \
> > action goto chain 4
> > $ tc filter add dev $VXLAN protocol ip parent ffff: prio 1 chain 4 \
> > flower enc_src_ip 2.2.2.200 enc_dst_ip 2.2.2.100 \
> > enc_dst_port 4789 enc_key_id 100 dst_mac 00:11:22:33:44:55 \
> > action pedit ex munge eth src set 00:11:22:33:44:f2 \
> > action mirred egress redirect dev enp130s0f0_0
> >
> > Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
> > ---
> > .../net/ethernet/mellanox/mlx5/core/en/mapping.c | 24 ++++++
> > .../net/ethernet/mellanox/mlx5/core/en/mapping.h | 1 +
> > drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 97 +++++++++++++++-------
> > 3 files changed, 92 insertions(+), 30 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> > index ea321e528749..90306dde6b60 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.c
> > @@ -74,6 +74,30 @@ int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id)
> > return err;
> > }
> >
> > +int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32 *id)
> > +{
> > + struct mapping_item *mi;
> > + u32 hash_key;
> > +
> > + mutex_lock(&ctx->lock);
> > +
> > + hash_key = jhash(data, ctx->data_size, 0);
> > + hash_for_each_possible(ctx->ht, mi, node, hash_key) {
> > + if (!memcmp(data, mi->data, ctx->data_size))
> > + goto found;
> > + }
> > +
> > + mutex_unlock(&ctx->lock);
> > + return -ENOENT;
> > +
> > +found:
> > + if (id)
> > + *id = mi->id;
> > +
> > + mutex_unlock(&ctx->lock);
> > + return 0;
> > +}
> > +
> > static void mapping_remove_and_free(struct mapping_ctx *ctx,
> > struct mapping_item *mi)
> > {
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> > index 285525cc5470..af501c9796b7 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h
> > @@ -9,6 +9,7 @@
> > int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id);
> > int mapping_remove(struct mapping_ctx *ctx, u32 id);
> > int mapping_find(struct mapping_ctx *ctx, u32 id, void *data);
> > +int mapping_find_by_data(struct mapping_ctx *ctx, void *data, u32 *id);
> >
> > /* mapping uses an xarray to map data to ids in add(), and for find().
> > * For locking, it uses a internal xarray spin lock for add()/remove(),
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> > index a574c588269a..64f5c3f3dbb3 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> > @@ -1786,7 +1786,8 @@ static void mlx5e_tc_del_flow(struct mlx5e_priv *priv,
> > }
> > }
> >
> > -static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
> > +static int flow_has_tc_action(struct flow_cls_offload *f,
> > + enum flow_action_id action)
> > {
> > struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> > struct flow_action *flow_action = &rule->action;
> > @@ -1794,12 +1795,8 @@ static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
> > int i;
> >
> > flow_action_for_each(i, act, flow_action) {
> > - switch (act->id) {
> > - case FLOW_ACTION_GOTO:
> > + if (act->id == action)
> > return true;
> > - default:
> > - continue;
> > - }
> > }
> >
> > return false;
> > @@ -1853,10 +1850,37 @@ static int flow_has_tc_fwd_action(struct flow_cls_offload *f)
> > sizeof(*__dst));\
> > })
> >
> > +static void mlx5e_make_tunnel_match_key(struct flow_cls_offload *f,
> > + struct net_device *filter_dev,
> > + struct tunnel_match_key *tunnel_key)
> > +{
> > + struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> > +
> > + memset(tunnel_key, 0, sizeof(*tunnel_key));
> > + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
> > + &tunnel_key->enc_control);
> > + if (tunnel_key->enc_control.addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS)
> > + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
> > + &tunnel_key->enc_ipv4);
> > + else
> > + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
> > + &tunnel_key->enc_ipv6);
> > +
> > + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP, &tunnel_key->enc_ip);
> > + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
> > + &tunnel_key->enc_tp);
> > + COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
> > + &tunnel_key->enc_key_id);
> > +
> > + tunnel_key->filter_ifindex = filter_dev->ifindex;
> > +}
> > +
> > static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> > struct mlx5e_tc_flow *flow,
> > struct flow_cls_offload *f,
> > - struct net_device *filter_dev)
> > + struct net_device *filter_dev,
> > + bool sets_mapping,
> > + bool needs_mapping)
> > {
> > struct flow_rule *rule = flow_cls_offload_flow_rule(f);
> > struct netlink_ext_ack *extack = f->common.extack;
> > @@ -1876,22 +1900,7 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> > uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> > uplink_priv = &uplink_rpriv->uplink_priv;
> >
> > - memset(&tunnel_key, 0, sizeof(tunnel_key));
> > - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL,
> > - &tunnel_key.enc_control);
> > - if (tunnel_key.enc_control.addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS)
> > - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
> > - &tunnel_key.enc_ipv4);
> > - else
> > - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
> > - &tunnel_key.enc_ipv6);
> > - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_IP, &tunnel_key.enc_ip);
> > - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_PORTS,
> > - &tunnel_key.enc_tp);
> > - COPY_DISSECTOR(rule, FLOW_DISSECTOR_KEY_ENC_KEYID,
> > - &tunnel_key.enc_key_id);
> > - tunnel_key.filter_ifindex = filter_dev->ifindex;
> > -
> > + mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> > err = mapping_add(uplink_priv->tunnel_mapping, &tunnel_key, &tun_id);
> > if (err)
> > return err;
> > @@ -1915,10 +1924,10 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> > mask = enc_opts_id ? TUNNEL_ID_MASK :
> > (TUNNEL_ID_MASK & ~ENC_OPTS_BITS_MASK);
> >
> > - if (attr->chain) {
> > + if (needs_mapping) {
> > mlx5e_tc_match_to_reg_match(&attr->parse_attr->spec,
> > TUNNEL_TO_REG, value, mask);
> > - } else {
> > + } else if (sets_mapping) {
> > mod_hdr_acts = &attr->parse_attr->mod_hdr_acts;
> > err = mlx5e_tc_match_to_reg_set(priv->mdev,
> > mod_hdr_acts,
> > @@ -1941,6 +1950,25 @@ static int mlx5e_get_flow_tunnel_id(struct mlx5e_priv *priv,
> > return err;
> > }
> >
> > +static int mlx5e_lookup_flow_tunnel_id(struct mlx5e_priv *priv,
> > + struct mlx5e_tc_flow *flow,
> > + struct flow_cls_offload *f,
> > + struct net_device *filter_dev,
> > + u32 *tun_id)
> > +{
> > + struct mlx5_rep_uplink_priv *uplink_priv;
> > + struct mlx5e_rep_priv *uplink_rpriv;
> > + struct tunnel_match_key tunnel_key;
> > + struct mlx5_eswitch *esw;
> > +
> > + esw = priv->mdev->priv.eswitch;
> > + uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
> > + uplink_priv = &uplink_rpriv->uplink_priv;
> > +
> > + mlx5e_make_tunnel_match_key(f, filter_dev, &tunnel_key);
> > + return mapping_find_by_data(uplink_priv->tunnel_mapping, &tunnel_key, tun_id);
> > +}
> > +
> > static void mlx5e_put_flow_tunnel_id(struct mlx5e_tc_flow *flow)
> > {
> > u32 enc_opts_id = flow->tunnel_id & ENC_OPTS_BITS_MASK;
> > @@ -1976,14 +2004,22 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
> > struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
> > struct netlink_ext_ack *extack = f->common.extack;
> > bool needs_mapping, sets_mapping;
> > + bool pedit_action;
> > int err;
> >
> > if (!mlx5e_is_eswitch_flow(flow))
> > return -EOPNOTSUPP;
> >
> > - needs_mapping = !!flow->esw_attr->chain;
> > - sets_mapping = !flow->esw_attr->chain && flow_has_tc_fwd_action(f);
> > - *match_inner = !needs_mapping;
> > + pedit_action = flow_has_tc_action(f, FLOW_ACTION_MANGLE) ||
> > + flow_has_tc_action(f, FLOW_ACTION_ADD);
> > +
> > + *match_inner = pedit_action;
> > + sets_mapping = pedit_action &&
> > + flow_has_tc_action(f, FLOW_ACTION_GOTO);
> > +
> > + needs_mapping = !!flow->esw_attr->chain &&
> > + !mlx5e_lookup_flow_tunnel_id(priv, flow, f,
> > + filter_dev, NULL);
> >
> > if ((needs_mapping || sets_mapping) &&
> > !mlx5_eswitch_reg_c1_loopback_enabled(esw)) {
> > @@ -1994,7 +2030,7 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
> > return -EOPNOTSUPP;
> > }
> >
> > - if (!flow->esw_attr->chain) {
> > + if (*match_inner && !needs_mapping) {
>
> why you only get inside for match_inner?
> what about the case of matching tunnel on chain 0 without pedit action?
Oh, sorry, I forget that case, I change the codes:
+ tunnel_decap = flow_has_tc_action(f, FLOW_ACTION_TUNNEL_DECAP);
+
+ *match_inner = pedit_action || tunnel_decap;
and all cases were tested, and I will sent v2. Thanks
> we need to call mlx5e_tc_tun_parse()
>
> > err = mlx5e_tc_tun_parse(filter_dev, priv, spec, f,
> > match_level);
> > if (err) {
> > @@ -2011,7 +2047,8 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
> > if (!needs_mapping && !sets_mapping)
> > return 0;
> >
> > - return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev);
> > + return mlx5e_get_flow_tunnel_id(priv, flow, f, filter_dev,
> > + sets_mapping, needs_mapping);
> > }
> >
> > static void *get_match_inner_headers_criteria(struct mlx5_flow_spec *spec)
> >
--
Best regards, Tonghao
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2020-05-06 1:20 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-04-28 5:24 [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary xiangxia.m.yue
2020-04-28 5:24 ` [PATCH net-next 2/3] net/mlx5e: Introduce mlx5e_eswitch_rep_uplink_priv helper xiangxia.m.yue
2020-04-30 11:17 ` Roi Dayan
2020-04-28 5:24 ` [PATCH net-next 3/3] net/mlx5e: Fix the code style xiangxia.m.yue
2020-04-30 11:19 ` Roi Dayan
2020-04-28 19:57 ` [PATCH net-next 1/3] net/mlx5e: Implicitly decap the tunnel packet when necessary Saeed Mahameed
2020-04-30 15:26 ` Roi Dayan
2020-05-06 1:20 ` Tonghao Zhang
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).