* [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs
@ 2020-10-02 22:44 Guillaume Nault
2020-10-02 22:44 ` [PATCH net-next 1/2] net/sched: act_vlan: Add {POP,PUSH}_ETH actions Guillaume Nault
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Guillaume Nault @ 2020-10-02 22:44 UTC (permalink / raw)
To: David Miller, Jakub Kicinski
Cc: netdev, Jamal Hadi Salim, Cong Wang, Jiri Pirko, Jiri Benc,
Martin Varghese
This patch series adds the necessary TC actions for supporting layer 2
MPLS VPNs (VPLS).
The objective is to give the possibility to add an MPLS header right
before an skb's mac header, then to prepend this MPLS packet with a
new Ethernet header with the MAC address of the next hop.
Patch 1 implements the actions for adding and removing the external
Ethernet header.
Patch 2 adds the possibility to push an MPLS header before the mac
header.
Most of the code already exists as these operations were first
implemented in openvswitch.
Practical example, with encap on Host-A and decap on Host-B:
Host-A# tc filter add dev ethAx ingress matchall \
action mpls mac_push label 20 \
action vlan push_eth dst_mac 02:00:00:00:00:02 \
src_mac 02:00:00:00:00:01 \
action mirred egress redirect dev ethAy
Host-B# tc filter add dev ethBx ingress protocol mpls_uc \
flower mpls_label 20 mpls_bos 1 \
action vlan pop_eth \
action mpls pop proto teb \
action mirred egress redirect dev ethBy
Guillaume Nault (2):
net/sched: act_vlan: Add {POP,PUSH}_ETH actions
net/sched: act_mpls: Add action to push MPLS LSE before Ethernet
header
include/linux/skbuff.h | 3 ++
include/net/tc_act/tc_vlan.h | 2 +
include/uapi/linux/tc_act/tc_mpls.h | 1 +
include/uapi/linux/tc_act/tc_vlan.h | 4 ++
net/core/skbuff.c | 67 +++++++++++++++++++++++++++++
net/openvswitch/actions.c | 28 +++++-------
net/sched/act_mpls.c | 18 ++++++++
net/sched/act_vlan.c | 40 +++++++++++++++++
8 files changed, 145 insertions(+), 18 deletions(-)
--
2.21.3
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH net-next 1/2] net/sched: act_vlan: Add {POP,PUSH}_ETH actions
2020-10-02 22:44 [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs Guillaume Nault
@ 2020-10-02 22:44 ` Guillaume Nault
2020-10-02 22:44 ` [PATCH net-next 2/2] net/sched: act_mpls: Add action to push MPLS LSE before Ethernet header Guillaume Nault
2020-10-04 0:33 ` [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs David Miller
2 siblings, 0 replies; 4+ messages in thread
From: Guillaume Nault @ 2020-10-02 22:44 UTC (permalink / raw)
To: David Miller, Jakub Kicinski
Cc: netdev, Jamal Hadi Salim, Cong Wang, Jiri Pirko, Jiri Benc,
Martin Varghese
Implement TCA_VLAN_ACT_POP_ETH and TCA_VLAN_ACT_PUSH_ETH, to
respectively pop and push a base Ethernet header at the beginning of a
frame.
POP_ETH is just a matter of pulling ETH_HLEN bytes. VLAN tags, if any,
must be stripped before calling POP_ETH.
PUSH_ETH is restricted to skbs with no mac_header, and only the MAC
addresses can be configured. The Ethertype is automatically set from
skb->protocol. These restrictions ensure that all skb's fields remain
consistent, so that this action can't confuse other part of the
networking stack (like GSO).
Since openvswitch already had these actions, consolidate the code in
skbuff.c (like for vlan and mpls push/pop).
Signed-off-by: Guillaume Nault <gnault@redhat.com>
---
include/linux/skbuff.h | 3 ++
include/net/tc_act/tc_vlan.h | 2 +
include/uapi/linux/tc_act/tc_vlan.h | 4 ++
net/core/skbuff.c | 67 +++++++++++++++++++++++++++++
net/openvswitch/actions.c | 28 +++++-------
net/sched/act_vlan.c | 40 +++++++++++++++++
6 files changed, 126 insertions(+), 18 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 3d0cf3722bb4..42131e325e27 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -3573,6 +3573,9 @@ int skb_ensure_writable(struct sk_buff *skb, int write_len);
int __skb_vlan_pop(struct sk_buff *skb, u16 *vlan_tci);
int skb_vlan_pop(struct sk_buff *skb);
int skb_vlan_push(struct sk_buff *skb, __be16 vlan_proto, u16 vlan_tci);
+int skb_eth_pop(struct sk_buff *skb);
+int skb_eth_push(struct sk_buff *skb, const unsigned char *dst,
+ const unsigned char *src);
int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto,
int mac_len, bool ethernet);
int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len,
diff --git a/include/net/tc_act/tc_vlan.h b/include/net/tc_act/tc_vlan.h
index 4e2502408c31..f051046ba034 100644
--- a/include/net/tc_act/tc_vlan.h
+++ b/include/net/tc_act/tc_vlan.h
@@ -11,6 +11,8 @@
struct tcf_vlan_params {
int tcfv_action;
+ unsigned char tcfv_push_dst[ETH_ALEN];
+ unsigned char tcfv_push_src[ETH_ALEN];
u16 tcfv_push_vid;
__be16 tcfv_push_proto;
u8 tcfv_push_prio;
diff --git a/include/uapi/linux/tc_act/tc_vlan.h b/include/uapi/linux/tc_act/tc_vlan.h
index 168995b54a70..5b306fe815cc 100644
--- a/include/uapi/linux/tc_act/tc_vlan.h
+++ b/include/uapi/linux/tc_act/tc_vlan.h
@@ -16,6 +16,8 @@
#define TCA_VLAN_ACT_POP 1
#define TCA_VLAN_ACT_PUSH 2
#define TCA_VLAN_ACT_MODIFY 3
+#define TCA_VLAN_ACT_POP_ETH 4
+#define TCA_VLAN_ACT_PUSH_ETH 5
struct tc_vlan {
tc_gen;
@@ -30,6 +32,8 @@ enum {
TCA_VLAN_PUSH_VLAN_PROTOCOL,
TCA_VLAN_PAD,
TCA_VLAN_PUSH_VLAN_PRIORITY,
+ TCA_VLAN_PUSH_ETH_DST,
+ TCA_VLAN_PUSH_ETH_SRC,
__TCA_VLAN_MAX,
};
#define TCA_VLAN_MAX (__TCA_VLAN_MAX - 1)
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index e0774471f56d..75b043accddb 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -5558,6 +5558,73 @@ int skb_vlan_push(struct sk_buff *skb, __be16 vlan_proto, u16 vlan_tci)
}
EXPORT_SYMBOL(skb_vlan_push);
+/**
+ * skb_eth_pop() - Drop the Ethernet header at the head of a packet
+ *
+ * @skb: Socket buffer to modify
+ *
+ * Drop the Ethernet header of @skb.
+ *
+ * Expects that skb->data points to the mac header and that no VLAN tags are
+ * present.
+ *
+ * Returns 0 on success, -errno otherwise.
+ */
+int skb_eth_pop(struct sk_buff *skb)
+{
+ if (!pskb_may_pull(skb, ETH_HLEN) || skb_vlan_tagged(skb) ||
+ skb_network_offset(skb) < ETH_HLEN)
+ return -EPROTO;
+
+ skb_pull_rcsum(skb, ETH_HLEN);
+ skb_reset_mac_header(skb);
+ skb_reset_mac_len(skb);
+
+ return 0;
+}
+EXPORT_SYMBOL(skb_eth_pop);
+
+/**
+ * skb_eth_push() - Add a new Ethernet header at the head of a packet
+ *
+ * @skb: Socket buffer to modify
+ * @dst: Destination MAC address of the new header
+ * @src: Source MAC address of the new header
+ *
+ * Prepend @skb with a new Ethernet header.
+ *
+ * Expects that skb->data points to the mac header, which must be empty.
+ *
+ * Returns 0 on success, -errno otherwise.
+ */
+int skb_eth_push(struct sk_buff *skb, const unsigned char *dst,
+ const unsigned char *src)
+{
+ struct ethhdr *eth;
+ int err;
+
+ if (skb_network_offset(skb) || skb_vlan_tag_present(skb))
+ return -EPROTO;
+
+ err = skb_cow_head(skb, sizeof(*eth));
+ if (err < 0)
+ return err;
+
+ skb_push(skb, sizeof(*eth));
+ skb_reset_mac_header(skb);
+ skb_reset_mac_len(skb);
+
+ eth = eth_hdr(skb);
+ ether_addr_copy(eth->h_dest, dst);
+ ether_addr_copy(eth->h_source, src);
+ eth->h_proto = skb->protocol;
+
+ skb_postpush_rcsum(skb, eth, sizeof(*eth));
+
+ return 0;
+}
+EXPORT_SYMBOL(skb_eth_push);
+
/* Update the ethertype of hdr and the skb csum value if required. */
static void skb_mod_eth_type(struct sk_buff *skb, struct ethhdr *hdr,
__be16 ethertype)
diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c
index 855f2c155956..b87bfc82f44f 100644
--- a/net/openvswitch/actions.c
+++ b/net/openvswitch/actions.c
@@ -277,9 +277,11 @@ static int set_eth_addr(struct sk_buff *skb, struct sw_flow_key *flow_key,
*/
static int pop_eth(struct sk_buff *skb, struct sw_flow_key *key)
{
- skb_pull_rcsum(skb, ETH_HLEN);
- skb_reset_mac_header(skb);
- skb_reset_mac_len(skb);
+ int err;
+
+ err = skb_eth_pop(skb);
+ if (err)
+ return err;
/* safe right before invalidate_flow_key */
key->mac_proto = MAC_PROTO_NONE;
@@ -290,22 +292,12 @@ static int pop_eth(struct sk_buff *skb, struct sw_flow_key *key)
static int push_eth(struct sk_buff *skb, struct sw_flow_key *key,
const struct ovs_action_push_eth *ethh)
{
- struct ethhdr *hdr;
-
- /* Add the new Ethernet header */
- if (skb_cow_head(skb, ETH_HLEN) < 0)
- return -ENOMEM;
-
- skb_push(skb, ETH_HLEN);
- skb_reset_mac_header(skb);
- skb_reset_mac_len(skb);
-
- hdr = eth_hdr(skb);
- ether_addr_copy(hdr->h_source, ethh->addresses.eth_src);
- ether_addr_copy(hdr->h_dest, ethh->addresses.eth_dst);
- hdr->h_proto = skb->protocol;
+ int err;
- skb_postpush_rcsum(skb, hdr, ETH_HLEN);
+ err = skb_eth_push(skb, ethh->addresses.eth_dst,
+ ethh->addresses.eth_src);
+ if (err)
+ return err;
/* safe right before invalidate_flow_key */
key->mac_proto = MAC_PROTO_ETHERNET;
diff --git a/net/sched/act_vlan.c b/net/sched/act_vlan.c
index a5ff9f68ab02..8758bd2a78fa 100644
--- a/net/sched/act_vlan.c
+++ b/net/sched/act_vlan.c
@@ -77,6 +77,16 @@ static int tcf_vlan_act(struct sk_buff *skb, const struct tc_action *a,
/* put updated tci as hwaccel tag */
__vlan_hwaccel_put_tag(skb, p->tcfv_push_proto, tci);
break;
+ case TCA_VLAN_ACT_POP_ETH:
+ err = skb_eth_pop(skb);
+ if (err)
+ goto drop;
+ break;
+ case TCA_VLAN_ACT_PUSH_ETH:
+ err = skb_eth_push(skb, p->tcfv_push_dst, p->tcfv_push_src);
+ if (err)
+ goto drop;
+ break;
default:
BUG();
}
@@ -93,10 +103,13 @@ static int tcf_vlan_act(struct sk_buff *skb, const struct tc_action *a,
}
static const struct nla_policy vlan_policy[TCA_VLAN_MAX + 1] = {
+ [TCA_VLAN_UNSPEC] = { .strict_start_type = TCA_VLAN_PUSH_ETH_DST },
[TCA_VLAN_PARMS] = { .len = sizeof(struct tc_vlan) },
[TCA_VLAN_PUSH_VLAN_ID] = { .type = NLA_U16 },
[TCA_VLAN_PUSH_VLAN_PROTOCOL] = { .type = NLA_U16 },
[TCA_VLAN_PUSH_VLAN_PRIORITY] = { .type = NLA_U8 },
+ [TCA_VLAN_PUSH_ETH_DST] = NLA_POLICY_ETH_ADDR,
+ [TCA_VLAN_PUSH_ETH_SRC] = NLA_POLICY_ETH_ADDR,
};
static int tcf_vlan_init(struct net *net, struct nlattr *nla,
@@ -179,6 +192,17 @@ static int tcf_vlan_init(struct net *net, struct nlattr *nla,
if (tb[TCA_VLAN_PUSH_VLAN_PRIORITY])
push_prio = nla_get_u8(tb[TCA_VLAN_PUSH_VLAN_PRIORITY]);
break;
+ case TCA_VLAN_ACT_POP_ETH:
+ break;
+ case TCA_VLAN_ACT_PUSH_ETH:
+ if (!tb[TCA_VLAN_PUSH_ETH_DST] || !tb[TCA_VLAN_PUSH_ETH_SRC]) {
+ if (exists)
+ tcf_idr_release(*a, bind);
+ else
+ tcf_idr_cleanup(tn, index);
+ return -EINVAL;
+ }
+ break;
default:
if (exists)
tcf_idr_release(*a, bind);
@@ -219,6 +243,13 @@ static int tcf_vlan_init(struct net *net, struct nlattr *nla,
p->tcfv_push_prio = push_prio;
p->tcfv_push_proto = push_proto;
+ if (action == TCA_VLAN_ACT_PUSH_ETH) {
+ nla_memcpy(&p->tcfv_push_dst, tb[TCA_VLAN_PUSH_ETH_DST],
+ ETH_ALEN);
+ nla_memcpy(&p->tcfv_push_src, tb[TCA_VLAN_PUSH_ETH_SRC],
+ ETH_ALEN);
+ }
+
spin_lock_bh(&v->tcf_lock);
goto_ch = tcf_action_set_ctrlact(*a, parm->action, goto_ch);
p = rcu_replace_pointer(v->vlan_p, p, lockdep_is_held(&v->tcf_lock));
@@ -279,6 +310,15 @@ static int tcf_vlan_dump(struct sk_buff *skb, struct tc_action *a,
p->tcfv_push_prio))))
goto nla_put_failure;
+ if (p->tcfv_action == TCA_VLAN_ACT_PUSH_ETH) {
+ if (nla_put(skb, TCA_VLAN_PUSH_ETH_DST, ETH_ALEN,
+ p->tcfv_push_dst))
+ goto nla_put_failure;
+ if (nla_put(skb, TCA_VLAN_PUSH_ETH_SRC, ETH_ALEN,
+ p->tcfv_push_src))
+ goto nla_put_failure;
+ }
+
tcf_tm_dump(&t, &v->tcf_tm);
if (nla_put_64bit(skb, TCA_VLAN_TM, sizeof(t), &t, TCA_VLAN_PAD))
goto nla_put_failure;
--
2.21.3
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH net-next 2/2] net/sched: act_mpls: Add action to push MPLS LSE before Ethernet header
2020-10-02 22:44 [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs Guillaume Nault
2020-10-02 22:44 ` [PATCH net-next 1/2] net/sched: act_vlan: Add {POP,PUSH}_ETH actions Guillaume Nault
@ 2020-10-02 22:44 ` Guillaume Nault
2020-10-04 0:33 ` [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs David Miller
2 siblings, 0 replies; 4+ messages in thread
From: Guillaume Nault @ 2020-10-02 22:44 UTC (permalink / raw)
To: David Miller, Jakub Kicinski
Cc: netdev, Jamal Hadi Salim, Cong Wang, Jiri Pirko, Jiri Benc,
Martin Varghese
Define the MAC_PUSH action which pushes an MPLS LSE before the mac
header (instead of between the mac and the network headers as the
plain PUSH action does).
The only special case is when the skb has an offloaded VLAN. In that
case, it has to be inlined before pushing the MPLS header.
Signed-off-by: Guillaume Nault <gnault@redhat.com>
---
include/uapi/linux/tc_act/tc_mpls.h | 1 +
net/sched/act_mpls.c | 18 ++++++++++++++++++
2 files changed, 19 insertions(+)
diff --git a/include/uapi/linux/tc_act/tc_mpls.h b/include/uapi/linux/tc_act/tc_mpls.h
index 9360e95273c7..9e4e8f52a779 100644
--- a/include/uapi/linux/tc_act/tc_mpls.h
+++ b/include/uapi/linux/tc_act/tc_mpls.h
@@ -10,6 +10,7 @@
#define TCA_MPLS_ACT_PUSH 2
#define TCA_MPLS_ACT_MODIFY 3
#define TCA_MPLS_ACT_DEC_TTL 4
+#define TCA_MPLS_ACT_MAC_PUSH 5
struct tc_mpls {
tc_gen; /* generic TC action fields. */
diff --git a/net/sched/act_mpls.c b/net/sched/act_mpls.c
index 8118e2640979..bb6b715636db 100644
--- a/net/sched/act_mpls.c
+++ b/net/sched/act_mpls.c
@@ -87,6 +87,23 @@ static int tcf_mpls_act(struct sk_buff *skb, const struct tc_action *a,
skb->dev && skb->dev->type == ARPHRD_ETHER))
goto drop;
break;
+ case TCA_MPLS_ACT_MAC_PUSH:
+ if (skb_vlan_tag_present(skb)) {
+ if (__vlan_insert_inner_tag(skb, skb->vlan_proto,
+ skb_vlan_tag_get(skb),
+ ETH_HLEN) < 0)
+ goto drop;
+
+ skb->protocol = skb->vlan_proto;
+ __vlan_hwaccel_clear_tag(skb);
+ }
+
+ new_lse = tcf_mpls_get_lse(NULL, p, mac_len ||
+ !eth_p_mpls(skb->protocol));
+
+ if (skb_mpls_push(skb, new_lse, p->tcfm_proto, 0, false))
+ goto drop;
+ break;
case TCA_MPLS_ACT_MODIFY:
new_lse = tcf_mpls_get_lse(mpls_hdr(skb), p, false);
if (skb_mpls_update_lse(skb, new_lse))
@@ -188,6 +205,7 @@ static int tcf_mpls_init(struct net *net, struct nlattr *nla,
}
break;
case TCA_MPLS_ACT_PUSH:
+ case TCA_MPLS_ACT_MAC_PUSH:
if (!tb[TCA_MPLS_LABEL]) {
NL_SET_ERR_MSG_MOD(extack, "Label is required for MPLS push");
return -EINVAL;
--
2.21.3
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs
2020-10-02 22:44 [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs Guillaume Nault
2020-10-02 22:44 ` [PATCH net-next 1/2] net/sched: act_vlan: Add {POP,PUSH}_ETH actions Guillaume Nault
2020-10-02 22:44 ` [PATCH net-next 2/2] net/sched: act_mpls: Add action to push MPLS LSE before Ethernet header Guillaume Nault
@ 2020-10-04 0:33 ` David Miller
2 siblings, 0 replies; 4+ messages in thread
From: David Miller @ 2020-10-04 0:33 UTC (permalink / raw)
To: gnault; +Cc: kuba, netdev, jhs, xiyou.wangcong, jiri, jbenc, martin.varghese
From: Guillaume Nault <gnault@redhat.com>
Date: Sat, 3 Oct 2020 00:44:25 +0200
> This patch series adds the necessary TC actions for supporting layer 2
> MPLS VPNs (VPLS).
...
Series applied, thank you.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2020-10-04 0:33 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-02 22:44 [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs Guillaume Nault
2020-10-02 22:44 ` [PATCH net-next 1/2] net/sched: act_vlan: Add {POP,PUSH}_ETH actions Guillaume Nault
2020-10-02 22:44 ` [PATCH net-next 2/2] net/sched: act_mpls: Add action to push MPLS LSE before Ethernet header Guillaume Nault
2020-10-04 0:33 ` [PATCH net-next 0/2] net/sched: Add actions for MPLS L2 VPNs David Miller
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.