Netdev Archive on lore.kernel.org
 help / color / Atom feed
* [PATCH net-next 0/3] net/sched: Introduce tc connection tracking
@ 2019-06-11 13:28 Paul Blakey
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
                   ` (2 more replies)
  0 siblings, 3 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-11 13:28 UTC (permalink / raw)
  To: Jiri Pirko, Paul Blakey, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang
  Cc: Rony Efraim, nst-kernel, John Hurley, Simon Horman, Justin Pettit

Hi,

This patch series add connection tracking capabilities in tc sw datapath.
It does so via a new tc action, called act_ct, and new tc flower classifier matching
on conntrack state, mark and label.

Usage is as follows:
$ tc qdisc add dev ens1f0_0 ingress
$ tc qdisc add dev ens1f0_1 ingress

$ tc filter add dev ens1f0_0 ingress \
  prio 1 chain 0 proto ip \
  flower ip_proto tcp ct_state -trk \
  action ct zone 2 pipe \
  action goto chain 2
$ tc filter add dev ens1f0_0 ingress \
  prio 1 chain 2 proto ip \
  flower ct_state +trk+new \
  action ct zone 2 commit mark 0xbb nat src 5.5.5.7 pipe \
  action mirred egress redirect dev ens1f0_1
$ tc filter add dev ens1f0_0 ingress \
  prio 1 chain 2 proto ip \
  flower ct_zone 2 ct_mark 0xbb ct_state +trk+est \
  action ct nat pipe \
  action mirred egress redirect dev ens1f0_1

$ tc filter add dev ens1f0_1 ingress \
  prio 1 chain 0 proto ip \
  flower ip_proto tcp ct_state -trk \
  action ct zone 2 pipe \
  action goto chain 1
$ tc filter add dev ens1f0_1 ingress \
  prio 1 chain 1 proto ip \
  flower ct_zone 2 ct_mark 0xbb ct_state +trk+est \
  action ct nat pipe \
  action mirred egress redirect dev ens1f0_0

The pattern used in the design here closely resembles OvS, as the plan is to also offload
OvS conntrack rules to tc. OvS datapath rules uses it's recirculation mechanism to send
specific packets to conntrack, and return with the new conntrack state (ct_state) on some other recirc_id
to be matched again (we use goto chain for this).

This results in the following OvS datapath rules:

recirc_id(0),in_port(ens1f0_0),ct_state(-trk),... actions:ct(zone=2),recirc(2)
recirc_id(2),in_port(ens1f0_0),ct_state(+new+trk),ct_mark(0xbb),... actions:ct(commit,zone=2,nat(src=5.5.5.7),mark=0xbb),ens1f0_1
recirc_id(2),in_port(ens1f0_0),ct_state(+est+trk),ct_mark(0xbb),... actions:ct(zone=2,nat),ens1f0_1

recirc_id(1),in_port(ens1f0_1),ct_state(-trk),... actions:ct(zone=2),recirc(1)
recirc_id(1),in_port(ens1f0_1),ct_state(+est+trk),... actions:ct(zone=2,nat),ens1f0_0

Paul Blakey (3):
  net/sched: Introduce action ct
  net/flow_dissector: add connection tracking dissection
  net/sched: cls_flower: Add matching on conntrack info

 include/linux/skbuff.h            |  10 +
 include/net/flow_dissector.h      |  15 +
 include/net/flow_offload.h        |   5 +
 include/net/tc_act/tc_ct.h        |  64 +++
 include/uapi/linux/pkt_cls.h      |  18 +
 include/uapi/linux/tc_act/tc_ct.h |  41 ++
 net/core/flow_dissector.c         |  44 ++
 net/sched/Kconfig                 |  11 +
 net/sched/Makefile                |   1 +
 net/sched/act_ct.c                | 900 ++++++++++++++++++++++++++++++++++++++
 net/sched/cls_api.c               |   5 +
 net/sched/cls_flower.c            | 127 +++++-
 12 files changed, 1236 insertions(+), 5 deletions(-)
 create mode 100644 include/net/tc_act/tc_ct.h
 create mode 100644 include/uapi/linux/tc_act/tc_ct.h
 create mode 100644 net/sched/act_ct.c

-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 28+ messages in thread

* [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:28 [PATCH net-next 0/3] net/sched: Introduce tc connection tracking Paul Blakey
@ 2019-06-11 13:28 ` Paul Blakey
  2019-06-11 13:59   ` Toke Høiland-Jørgensen
                     ` (5 more replies)
  2019-06-11 13:28 ` [PATCH net-next 2/3] net/flow_dissector: add connection tracking dissection Paul Blakey
  2019-06-11 13:28 ` [PATCH net-next 3/3] net/sched: cls_flower: Add matching on conntrack info Paul Blakey
  2 siblings, 6 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-11 13:28 UTC (permalink / raw)
  To: Jiri Pirko, Paul Blakey, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang
  Cc: Rony Efraim, nst-kernel, John Hurley, Simon Horman, Justin Pettit

Allow sending a packet to conntrack and set conntrack zone, mark,
labels and nat parameters.

Signed-off-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
 include/net/flow_offload.h        |   5 +
 include/net/tc_act/tc_ct.h        |  64 +++
 include/uapi/linux/pkt_cls.h      |   2 +
 include/uapi/linux/tc_act/tc_ct.h |  41 ++
 net/sched/Kconfig                 |  11 +
 net/sched/Makefile                |   1 +
 net/sched/act_ct.c                | 901 ++++++++++++++++++++++++++++++++++++++
 net/sched/cls_api.c               |   5 +
 8 files changed, 1030 insertions(+)
 create mode 100644 include/net/tc_act/tc_ct.h
 create mode 100644 include/uapi/linux/tc_act/tc_ct.h
 create mode 100644 net/sched/act_ct.c

diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index 36fdb85..5b2c4fa 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h
@@ -123,6 +123,7 @@ enum flow_action_id {
 	FLOW_ACTION_QUEUE,
 	FLOW_ACTION_SAMPLE,
 	FLOW_ACTION_POLICE,
+	FLOW_ACTION_CT,
 };
 
 /* This is mirroring enum pedit_header_type definition for easy mapping between
@@ -172,6 +173,10 @@ struct flow_action_entry {
 			s64			burst;
 			u64			rate_bytes_ps;
 		} police;
+		struct {				/* FLOW_ACTION_CT */
+			int action;
+			u16 zone;
+		} ct;
 	};
 };
 
diff --git a/include/net/tc_act/tc_ct.h b/include/net/tc_act/tc_ct.h
new file mode 100644
index 0000000..59e4f5e
--- /dev/null
+++ b/include/net/tc_act/tc_ct.h
@@ -0,0 +1,64 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __NET_TC_CT_H
+#define __NET_TC_CT_H
+
+#include <net/act_api.h>
+#include <uapi/linux/tc_act/tc_ct.h>
+
+#if IS_ENABLED(CONFIG_NF_CONNTRACK)
+#include <net/netfilter/nf_nat.h>
+#include <net/netfilter/nf_conntrack_labels.h>
+
+struct tcf_ct_params {
+	struct nf_conn *tmpl;
+	u16 zone;
+
+	u32 mark;
+	u32 mark_mask;
+
+	u32 labels[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
+	u32 labels_mask[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
+
+	struct nf_nat_range2 range;
+	bool ipv4_range;
+
+	u16 ct_action;
+
+	struct rcu_head rcu;
+
+};
+
+struct tcf_ct {
+	struct tc_action common;
+	struct tcf_ct_params __rcu *params;
+};
+
+#define to_ct(a) ((struct tcf_ct *)a)
+#define to_ct_params(a) ((struct tcf_ct_params *) \
+			 rtnl_dereference((to_ct(a)->params)))
+
+static inline uint16_t tcf_ct_zone(const struct tc_action *a)
+{
+	return to_ct_params(a)->zone;
+}
+
+static inline int tcf_ct_action(const struct tc_action *a)
+{
+	return to_ct_params(a)->ct_action;
+}
+
+#else
+static inline uint16_t tcf_ct_zone(const struct tc_action *a) { return 0; }
+static inline int tcf_ct_action(const struct tc_action *a) { return 0; }
+#endif /* CONFIG_NF_CONNTRACK */
+
+static inline bool is_tcf_ct(const struct tc_action *a)
+{
+#if defined(CONFIG_NET_CLS_ACT) && IS_ENABLED(CONFIG_NF_CONNTRACK)
+	if (a->ops && a->ops->id == TCA_ID_CT)
+		return true;
+#endif
+	return false;
+}
+
+#endif /* __NET_TC_CT_H */
diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
index a93680f..c5264d7 100644
--- a/include/uapi/linux/pkt_cls.h
+++ b/include/uapi/linux/pkt_cls.h
@@ -83,6 +83,7 @@ enum {
 #define TCA_ACT_SIMP 22
 #define TCA_ACT_IFE 25
 #define TCA_ACT_SAMPLE 26
+#define TCA_ACT_CT 27
 
 /* Action type identifiers*/
 enum tca_id {
@@ -106,6 +107,7 @@ enum tca_id {
 	TCA_ID_SAMPLE = TCA_ACT_SAMPLE,
 	/* other actions go here */
 	TCA_ID_CTINFO,
+	TCA_ID_CT,
 	__TCA_ID_MAX = 255
 };
 
diff --git a/include/uapi/linux/tc_act/tc_ct.h b/include/uapi/linux/tc_act/tc_ct.h
new file mode 100644
index 0000000..5fb1d7a
--- /dev/null
+++ b/include/uapi/linux/tc_act/tc_ct.h
@@ -0,0 +1,41 @@
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+#ifndef __UAPI_TC_CT_H
+#define __UAPI_TC_CT_H
+
+#include <linux/types.h>
+#include <linux/pkt_cls.h>
+
+enum {
+	TCA_CT_UNSPEC,
+	TCA_CT_PARMS,
+	TCA_CT_TM,
+	TCA_CT_ACTION,		/* u16 */
+	TCA_CT_ZONE,		/* u16 */
+	TCA_CT_MARK,		/* u32 */
+	TCA_CT_MARK_MASK,	/* u32 */
+	TCA_CT_LABELS,		/* u128 */
+	TCA_CT_LABELS_MASK,	/* u128 */
+	TCA_CT_NAT_IPV4_MIN,	/* be32 */
+	TCA_CT_NAT_IPV4_MAX,	/* be32 */
+	TCA_CT_NAT_IPV6_MIN,	/* struct in6_addr */
+	TCA_CT_NAT_IPV6_MAX,	/* struct in6_addr */
+	TCA_CT_NAT_PORT_MIN,	/* be16 */
+	TCA_CT_NAT_PORT_MAX,	/* be16 */
+	TCA_CT_PAD,
+	__TCA_CT_MAX
+};
+
+#define TCA_CT_MAX (__TCA_CT_MAX - 1)
+
+#define TCA_CT_ACT_COMMIT	(1 << 0)
+#define TCA_CT_ACT_FORCE	(1 << 1)
+#define TCA_CT_ACT_CLEAR	(1 << 2)
+#define TCA_CT_ACT_NAT		(1 << 3)
+#define TCA_CT_ACT_NAT_SRC	(1 << 4)
+#define TCA_CT_ACT_NAT_DST	(1 << 5)
+
+struct tc_ct {
+	tc_gen;
+};
+
+#endif /* __UAPI_TC_CT_H */
diff --git a/net/sched/Kconfig b/net/sched/Kconfig
index d104f7e..0481a2a 100644
--- a/net/sched/Kconfig
+++ b/net/sched/Kconfig
@@ -929,6 +929,17 @@ config NET_ACT_TUNNEL_KEY
 	  To compile this code as a module, choose M here: the
 	  module will be called act_tunnel_key.
 
+config NET_ACT_CT
+        tristate "connection tracking tc action"
+        depends on NET_CLS_ACT && NF_CONNTRACK
+        help
+	  Say Y here to allow sending the packets to conntrack module.
+
+	  If unsure, say N.
+
+	  To compile this code as a module, choose M here: the
+	  module will be called act_ct.
+
 config NET_IFE_SKBMARK
         tristate "Support to encoding decoding skb mark on IFE action"
         depends on NET_ACT_IFE
diff --git a/net/sched/Makefile b/net/sched/Makefile
index d54bfcb..23d2202 100644
--- a/net/sched/Makefile
+++ b/net/sched/Makefile
@@ -28,6 +28,7 @@ obj-$(CONFIG_NET_IFE_SKBMARK)	+= act_meta_mark.o
 obj-$(CONFIG_NET_IFE_SKBPRIO)	+= act_meta_skbprio.o
 obj-$(CONFIG_NET_IFE_SKBTCINDEX)	+= act_meta_skbtcindex.o
 obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o
+obj-$(CONFIG_NET_ACT_CT)	+= act_ct.o
 obj-$(CONFIG_NET_SCH_FIFO)	+= sch_fifo.o
 obj-$(CONFIG_NET_SCH_CBQ)	+= sch_cbq.o
 obj-$(CONFIG_NET_SCH_HTB)	+= sch_htb.o
diff --git a/net/sched/act_ct.c b/net/sched/act_ct.c
new file mode 100644
index 0000000..4eb0dd7
--- /dev/null
+++ b/net/sched/act_ct.c
@@ -0,0 +1,901 @@
+// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
+/* -
+ * net/sched/act_ct.c  Connection Tracking action
+ *
+ * Authors:   Paul Blakey <paulb@mellanox.com>
+ *            Yossi Kuperman <yossiku@mellanox.com>
+ *            Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/skbuff.h>
+#include <linux/rtnetlink.h>
+#include <linux/pkt_cls.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <net/netlink.h>
+#include <net/pkt_sched.h>
+#include <net/act_api.h>
+#include <uapi/linux/tc_act/tc_ct.h>
+#include <net/tc_act/tc_ct.h>
+
+#include <linux/netfilter/nf_nat.h>
+#include <net/netfilter/nf_conntrack.h>
+#include <net/netfilter/nf_conntrack_core.h>
+#include <net/netfilter/nf_conntrack_zones.h>
+#include <net/netfilter/nf_conntrack_helper.h>
+#include <net/pkt_cls.h>
+
+static struct tc_action_ops act_ct_ops;
+static unsigned int ct_net_id;
+
+struct tc_ct_action_net {
+	struct tc_action_net tn; /* Must be first */
+	bool labels;
+};
+
+/* Determine whether skb->_nfct is equal to the result of conntrack lookup. */
+static bool tcf_ct_skb_nfct_cached(struct net *net, struct sk_buff *skb,
+				   u16 zone_id, bool force)
+{
+	enum ip_conntrack_info ctinfo;
+	struct nf_conn *ct;
+
+	ct = nf_ct_get(skb, &ctinfo);
+	if (!ct)
+		return false;
+	if (!net_eq(net, read_pnet(&ct->ct_net)))
+		return false;
+	if (nf_ct_zone(ct)->id != zone_id)
+		return false;
+
+	/* Force conntrack entry direction. */
+	if (force && CTINFO2DIR(ctinfo) != IP_CT_DIR_ORIGINAL) {
+		nf_conntrack_put(&ct->ct_general);
+		nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
+
+		if (nf_ct_is_confirmed(ct))
+			nf_ct_kill(ct);
+
+		return false;
+	}
+
+	return true;
+}
+
+/* Trim the skb to the length specified by the IP/IPv6 header,
+ * removing any trailing lower-layer padding. This prepares the skb
+ * for higher-layer processing that assumes skb->len excludes padding
+ * (such as nf_ip_checksum). The caller needs to pull the skb to the
+ * network header, and ensure ip_hdr/ipv6_hdr points to valid data.
+ */
+static int tcf_ct_skb_network_trim(struct sk_buff *skb)
+{
+	unsigned int len;
+	int err;
+
+	switch (skb->protocol) {
+	case htons(ETH_P_IP):
+		len = ntohs(ip_hdr(skb)->tot_len);
+		break;
+	case htons(ETH_P_IPV6):
+		len = sizeof(struct ipv6hdr)
+			+ ntohs(ipv6_hdr(skb)->payload_len);
+		break;
+	default:
+		len = skb->len;
+	}
+
+	err = pskb_trim_rcsum(skb, len);
+
+	return err;
+}
+
+static u8 tcf_ct_skb_nf_family(struct sk_buff *skb)
+{
+	u8 family = NFPROTO_UNSPEC;
+
+	switch (skb->protocol) {
+	case htons(ETH_P_IP):
+		family = NFPROTO_IPV4;
+		break;
+	case htons(ETH_P_IPV6):
+		family = NFPROTO_IPV6;
+		break;
+	default:
+	break;
+	}
+
+	return family;
+}
+
+static void tcf_ct_params_free(struct rcu_head *head)
+{
+	struct tcf_ct_params *params = container_of(head,
+						    struct tcf_ct_params, rcu);
+
+	if (params->tmpl)
+		nf_conntrack_put(&params->tmpl->ct_general);
+	kfree(params);
+}
+
+#if IS_ENABLED(CONFIG_NF_NAT)
+/* Modelled after nf_nat_ipv[46]_fn().
+ * range is only used for new, uninitialized NAT state.
+ * Returns either NF_ACCEPT or NF_DROP.
+ */
+static int ct_nat_execute(struct sk_buff *skb, struct nf_conn *ct,
+			  enum ip_conntrack_info ctinfo,
+			  const struct nf_nat_range2 *range,
+			  enum nf_nat_manip_type maniptype)
+{
+	int hooknum, nh_off, err = NF_ACCEPT;
+
+	nh_off = skb_network_offset(skb);
+	skb_pull_rcsum(skb, nh_off);
+
+	/* See HOOK2MANIP(). */
+	if (maniptype == NF_NAT_MANIP_SRC)
+		hooknum = NF_INET_LOCAL_IN; /* Source NAT */
+	else
+		hooknum = NF_INET_LOCAL_OUT; /* Destination NAT */
+
+	switch (ctinfo) {
+	case IP_CT_RELATED:
+	case IP_CT_RELATED_REPLY:
+		if (skb->protocol == htons(ETH_P_IP) &&
+		    ip_hdr(skb)->protocol == IPPROTO_ICMP) {
+			if (!nf_nat_icmp_reply_translation(skb, ct, ctinfo,
+							   hooknum))
+				err = NF_DROP;
+			goto push;
+		} else if (IS_ENABLED(CONFIG_IPV6) &&
+			   skb->protocol == htons(ETH_P_IPV6)) {
+			__be16 frag_off;
+			u8 nexthdr = ipv6_hdr(skb)->nexthdr;
+			int hdrlen = ipv6_skip_exthdr(skb,
+						      sizeof(struct ipv6hdr),
+						      &nexthdr, &frag_off);
+
+			if (hdrlen >= 0 && nexthdr == IPPROTO_ICMPV6) {
+				if (!nf_nat_icmpv6_reply_translation(skb, ct,
+								     ctinfo,
+								     hooknum,
+								     hdrlen))
+					err = NF_DROP;
+				goto push;
+			}
+		}
+		/* Non-ICMP, fall thru to initialize if needed. */
+		/* fall through */
+	case IP_CT_NEW:
+		/* Seen it before?  This can happen for loopback, retrans,
+		 * or local packets.
+		 */
+		if (!nf_nat_initialized(ct, maniptype)) {
+			/* Initialize according to the NAT action. */
+			err = (range && range->flags & NF_NAT_RANGE_MAP_IPS)
+				/* Action is set up to establish a new
+				 * mapping.
+				 */
+				? nf_nat_setup_info(ct, range, maniptype)
+				: nf_nat_alloc_null_binding(ct, hooknum);
+			if (err != NF_ACCEPT)
+				goto push;
+		}
+		break;
+
+	case IP_CT_ESTABLISHED:
+	case IP_CT_ESTABLISHED_REPLY:
+		break;
+
+	default:
+		err = NF_DROP;
+		goto push;
+	}
+
+	err = nf_nat_packet(ct, ctinfo, hooknum, skb);
+push:
+	skb_push(skb, nh_off);
+	skb_postpush_rcsum(skb, skb->data, nh_off);
+
+	return err;
+}
+#endif /* CONFIG_NF_NAT */
+
+static void tcf_ct_act_set_mark(struct nf_conn *ct, u32 mark, u32 mask)
+{
+#if IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)
+	u32 new_mark;
+
+	if (!mask)
+		return;
+
+	new_mark = mark | (ct->mark & ~(mask));
+	if (ct->mark != new_mark) {
+		ct->mark = new_mark;
+		if (nf_ct_is_confirmed(ct))
+			nf_conntrack_event_cache(IPCT_MARK, ct);
+	}
+#endif
+}
+
+static void tcf_ct_act_set_labels(struct nf_conn *ct,
+				  u32 *labels,
+				  u32 *labels_m)
+{
+#if IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)
+	size_t labels_sz = FIELD_SIZEOF(struct tcf_ct_params, labels);
+
+	if (!memchr_inv(labels_m, 0, labels_sz))
+		return;
+
+	nf_connlabels_replace(ct, labels, labels_m, 4);
+#endif
+}
+
+static int tcf_ct_act_nat(struct sk_buff *skb,
+			  struct nf_conn *ct,
+			  enum ip_conntrack_info ctinfo,
+			  int ct_action,
+			  struct nf_nat_range2 *range,
+			  bool commit)
+{
+#if IS_ENABLED(CONFIG_NF_NAT)
+	enum nf_nat_manip_type maniptype;
+
+	if (!(ct_action & TCA_CT_ACT_NAT))
+		return NF_ACCEPT;
+
+	/* Add NAT extension if not confirmed yet. */
+	if (!nf_ct_is_confirmed(ct) && !nf_ct_nat_ext_add(ct))
+		return NF_DROP;   /* Can't NAT. */
+
+	if (ctinfo != IP_CT_NEW && (ct->status & IPS_NAT_MASK) &&
+	    (ctinfo != IP_CT_RELATED || commit)) {
+		/* NAT an established or related connection like before. */
+		if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY)
+			/* This is the REPLY direction for a connection
+			 * for which NAT was applied in the forward
+			 * direction.  Do the reverse NAT.
+			 */
+			maniptype = ct->status & IPS_SRC_NAT
+				? NF_NAT_MANIP_DST : NF_NAT_MANIP_SRC;
+		else
+			maniptype = ct->status & IPS_SRC_NAT
+				? NF_NAT_MANIP_SRC : NF_NAT_MANIP_DST;
+	} else if (ct_action & TCA_CT_ACT_NAT_SRC) {
+		maniptype = NF_NAT_MANIP_SRC;
+	} else if (ct_action & TCA_CT_ACT_NAT_DST) {
+		maniptype = NF_NAT_MANIP_DST;
+	} else {
+		return NF_ACCEPT;
+	}
+
+	return ct_nat_execute(skb, ct, ctinfo, range, maniptype);
+#else
+	return NF_ACCEPT;
+#endif
+}
+
+static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
+		      struct tcf_result *res)
+{
+	struct net *net = dev_net(skb->dev);
+	bool cached, commit, clear, force;
+	enum ip_conntrack_info ctinfo;
+	struct tcf_ct *c = to_ct(a);
+	struct nf_conn *tmpl = NULL;
+	struct nf_hook_state state;
+	struct tcf_ct_params *p;
+	struct nf_conn *ct;
+	int nh_ofs, err;
+	u8 family;
+
+	p = rcu_dereference_bh(c->params);
+
+	commit = p->ct_action & TCA_CT_ACT_COMMIT;
+	clear = p->ct_action & TCA_CT_ACT_CLEAR;
+	force = p->ct_action & TCA_CT_ACT_FORCE;
+	tmpl = p->tmpl;
+
+	if (clear) {
+		ct = nf_ct_get(skb, &ctinfo);
+		if (ct) {
+			nf_conntrack_put(&ct->ct_general);
+			nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
+		}
+
+		goto out;
+	}
+
+	/* The conntrack module expects to be working at L3. */
+	nh_ofs = skb_network_offset(skb);
+	skb_pull_rcsum(skb, nh_ofs);
+
+	err = tcf_ct_skb_network_trim(skb);
+	if (err)
+		goto drop;
+
+	family = tcf_ct_skb_nf_family(skb);
+	if (family == NFPROTO_UNSPEC)
+		goto drop;
+
+	state.hook = NF_INET_PRE_ROUTING;
+	state.net = net;
+	state.pf = family;
+
+	/* If we are recirculating packets to match on ct fields and
+	 * committing with a separate ct action, then we don't need to
+	 * actually run the packet through conntrack twice unless it's for a
+	 * different zone.
+	 */
+	cached = tcf_ct_skb_nfct_cached(net, skb, p->zone, force);
+	if (!cached) {
+		/* Associate skb with specified zone. */
+		if (tmpl) {
+			if (skb_nfct(skb))
+				nf_conntrack_put(skb_nfct(skb));
+			nf_conntrack_get(&tmpl->ct_general);
+			nf_ct_set(skb, tmpl, IP_CT_NEW);
+		}
+
+		err = nf_conntrack_in(skb, &state);
+		if (err != NF_ACCEPT)
+			goto out_push;
+	}
+
+	ct = nf_ct_get(skb, &ctinfo);
+	if (!ct)
+		goto out_push;
+	nf_ct_deliver_cached_events(ct);
+
+	err = tcf_ct_act_nat(skb, ct, ctinfo, p->ct_action, &p->range, commit);
+	if (err != NF_ACCEPT)
+		goto drop;
+
+	if (commit) {
+		tcf_ct_act_set_mark(ct, p->mark, p->mark_mask);
+		tcf_ct_act_set_labels(ct, p->labels, p->labels_mask);
+
+		/* This will take care of sending queued events
+		 * even if the connection is already confirmed.
+		 */
+		nf_conntrack_confirm(skb);
+	}
+
+out_push:
+	skb_push(skb, nh_ofs);
+	skb_postpush_rcsum(skb, skb->data, nh_ofs);
+
+out:
+	bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), skb);
+
+	return c->tcf_action;
+
+drop:
+	qstats_drop_inc(this_cpu_ptr(a->cpu_qstats));
+	return TC_ACT_SHOT;
+}
+
+static const struct nla_policy ct_policy[TCA_CT_MAX + 1] = {
+	[TCA_CT_ACTION] = { .type = NLA_U16 },
+	[TCA_CT_PARMS] = { .type = NLA_EXACT_LEN, .len = sizeof(struct tc_ct) },
+	[TCA_CT_ZONE] = { .type = NLA_U16 },
+	[TCA_CT_MARK] = { .type = NLA_U32 },
+	[TCA_CT_MARK_MASK] = { .type = NLA_U32 },
+	[TCA_CT_LABELS] = { .type = NLA_BINARY,
+			    .len = 128 / BITS_PER_BYTE },
+	[TCA_CT_LABELS_MASK] = { .type = NLA_BINARY,
+				 .len = 128 / BITS_PER_BYTE },
+	[TCA_CT_NAT_IPV4_MIN] = { .type = NLA_U32 },
+	[TCA_CT_NAT_IPV4_MAX] = { .type = NLA_U32 },
+	[TCA_CT_NAT_IPV6_MIN] = { .type = NLA_EXACT_LEN,
+				  .len = sizeof(struct in6_addr) },
+	[TCA_CT_NAT_IPV6_MAX] = { .type = NLA_EXACT_LEN,
+				   .len = sizeof(struct in6_addr) },
+	[TCA_CT_NAT_PORT_MIN] = { .type = NLA_U16 },
+	[TCA_CT_NAT_PORT_MAX] = { .type = NLA_U16 },
+};
+
+static int tcf_ct_fill_params_nat(struct tcf_ct_params *p,
+				  struct tc_ct *parm,
+				  struct nlattr **tb,
+				  struct netlink_ext_ack *extack)
+{
+	struct nf_nat_range2 *range;
+
+	if (!(p->ct_action & TCA_CT_ACT_NAT))
+		return 0;
+
+	if (!IS_ENABLED(CONFIG_NF_NAT)) {
+		NL_SET_ERR_MSG_MOD(extack, "Netfilter nat isn't enabled in kernel");
+		return -EOPNOTSUPP;
+	}
+
+	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
+		return 0;
+
+	if ((p->ct_action & TCA_CT_ACT_NAT_SRC) &&
+	    (p->ct_action & TCA_CT_ACT_NAT_DST)) {
+		NL_SET_ERR_MSG_MOD(extack, "dnat and snat can't be enabled at the same time");
+		return -EOPNOTSUPP;
+	}
+
+	range = &p->range;
+	if (tb[TCA_CT_NAT_IPV4_MIN]) {
+		range->min_addr.ip =
+			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);
+		range->flags |= NF_NAT_RANGE_MAP_IPS;
+		p->ipv4_range = true;
+	}
+	if (tb[TCA_CT_NAT_IPV4_MAX]) {
+		range->max_addr.ip =
+			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]);
+		range->flags |= NF_NAT_RANGE_MAP_IPS;
+		p->ipv4_range = true;
+	} else if (range->min_addr.ip) {
+		range->max_addr.ip = range->min_addr.ip;
+	}
+
+	if (tb[TCA_CT_NAT_IPV6_MIN]) {
+		range->min_addr.in6 =
+			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);
+		range->flags |= NF_NAT_RANGE_MAP_IPS;
+		p->ipv4_range = false;
+	}
+	if (tb[TCA_CT_NAT_IPV6_MAX]) {
+		range->max_addr.in6 =
+			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]);
+		range->flags |= NF_NAT_RANGE_MAP_IPS;
+		p->ipv4_range = false;
+	} else if (memchr_inv(&range->min_addr.in6, 0,
+		   sizeof(range->min_addr.in6))) {
+		range->max_addr.in6 = range->min_addr.in6;
+	}
+
+	if (tb[TCA_CT_NAT_PORT_MIN]) {
+		range->min_proto.all =
+			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MIN]));
+		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
+	}
+	if (tb[TCA_CT_NAT_PORT_MAX]) {
+		range->max_proto.all =
+			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MAX]));
+		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
+	} else if (range->min_proto.all) {
+		range->max_proto.all = range->min_proto.all;
+	}
+
+	return 0;
+}
+
+static void tcf_ct_set_key_val(struct nlattr **tb,
+			       void *val, int val_type,
+			       void *mask, int mask_type,
+			       int len)
+{
+	if (!tb[val_type])
+		return;
+	nla_memcpy(val, tb[val_type], len);
+
+	if (!mask)
+		return;
+
+	if (mask_type == TCA_CT_UNSPEC || !tb[mask_type])
+		memset(mask, 0xff, len);
+	else
+		nla_memcpy(mask, tb[mask_type], len);
+}
+
+static int tcf_ct_fill_params(struct net *net,
+			      struct tcf_ct_params *p,
+			      struct tc_ct *parm,
+			      struct nlattr **tb,
+			      struct netlink_ext_ack *extack)
+{
+	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
+	struct nf_conntrack_zone zone;
+	struct nf_conn *tmpl;
+	int err;
+
+	p->zone = NF_CT_DEFAULT_ZONE_ID;
+
+	tcf_ct_set_key_val(tb,
+			   &p->ct_action, TCA_CT_ACTION,
+			   NULL, TCA_CT_UNSPEC,
+			   sizeof(p->ct_action));
+
+	if (p->ct_action & TCA_CT_ACT_CLEAR)
+		return 0;
+
+	err = tcf_ct_fill_params_nat(p, parm, tb, extack);
+	if (err)
+		return err;
+
+	if (tb[TCA_CT_MARK]) {
+		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)) {
+			NL_SET_ERR_MSG_MOD(extack, "Conntrack mark isn't enabled.");
+			return -EOPNOTSUPP;
+		}
+		tcf_ct_set_key_val(tb,
+				   &p->mark, TCA_CT_MARK,
+				   &p->mark_mask, TCA_CT_MARK_MASK,
+				   sizeof(p->mark));
+	}
+
+	if (tb[TCA_CT_LABELS]) {
+		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)) {
+			NL_SET_ERR_MSG_MOD(extack, "Conntrack labels isn't enabled.");
+			return -EOPNOTSUPP;
+		}
+
+		if (!tn->labels) {
+			NL_SET_ERR_MSG_MOD(extack, "Failed to set connlabel length");
+			return -EOPNOTSUPP;
+		}
+		tcf_ct_set_key_val(tb,
+				   p->labels, TCA_CT_LABELS,
+				   p->labels_mask, TCA_CT_LABELS_MASK,
+				   sizeof(p->labels));
+	}
+
+	if (tb[TCA_CT_ZONE]) {
+		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
+			NL_SET_ERR_MSG_MOD(extack, "Conntrack zones isn't enabled.");
+			return -EOPNOTSUPP;
+		}
+
+		tcf_ct_set_key_val(tb,
+				   &p->zone, TCA_CT_ZONE,
+				   NULL, TCA_CT_UNSPEC,
+				   sizeof(p->zone));
+	}
+
+	if (p->zone == NF_CT_DEFAULT_ZONE_ID)
+		return 0;
+
+	nf_ct_zone_init(&zone, p->zone, NF_CT_DEFAULT_ZONE_DIR, 0);
+	tmpl = nf_ct_tmpl_alloc(net, &zone, GFP_KERNEL);
+	if (!tmpl) {
+		NL_SET_ERR_MSG_MOD(extack, "Failed to allocate conntrack template");
+		return -ENOMEM;
+	}
+	__set_bit(IPS_CONFIRMED_BIT, &tmpl->status);
+	nf_conntrack_get(&tmpl->ct_general);
+	p->tmpl = tmpl;
+
+	return 0;
+}
+
+static int tcf_ct_init(struct net *net, struct nlattr *nla,
+		       struct nlattr *est, struct tc_action **a,
+		       int replace, int bind, bool rtnl_held,
+		       struct tcf_proto *tp,
+		       struct netlink_ext_ack *extack)
+{
+	struct tc_action_net *tn = net_generic(net, ct_net_id);
+	struct tcf_ct_params *params = NULL;
+	struct nlattr *tb[TCA_CT_MAX + 1];
+	struct tcf_chain *goto_ch = NULL;
+	struct tc_ct *parm;
+	struct tcf_ct *c;
+	int err, res = 0;
+
+	if (!nla) {
+		NL_SET_ERR_MSG_MOD(extack, "Ct requires attributes to be passed");
+		return -EINVAL;
+	}
+
+	err = nla_parse_nested(tb, TCA_CT_MAX, nla, ct_policy, extack);
+	if (err < 0)
+		return err;
+
+	if (!tb[TCA_CT_PARMS]) {
+		NL_SET_ERR_MSG_MOD(extack, "Missing required ct parameters");
+		return -EINVAL;
+	}
+	parm = nla_data(tb[TCA_CT_PARMS]);
+
+	err = tcf_idr_check_alloc(tn, &parm->index, a, bind);
+	if (err < 0)
+		return err;
+
+	if (!err) {
+		err = tcf_idr_create(tn, parm->index, est, a,
+				     &act_ct_ops, bind, true);
+		if (err) {
+			tcf_idr_cleanup(tn, parm->index);
+			return err;
+		}
+		res = ACT_P_CREATED;
+	} else {
+		if (bind)
+			return 0;
+
+		if (!replace) {
+			tcf_idr_release(*a, bind);
+			return -EEXIST;
+		}
+	}
+	err = tcf_action_check_ctrlact(parm->action, tp, &goto_ch, extack);
+	if (err < 0)
+		goto cleanup;
+
+	c = to_ct(*a);
+
+	params = kzalloc(sizeof(*params), GFP_KERNEL);
+	if (unlikely(!params)) {
+		err = -ENOMEM;
+		goto cleanup;
+	}
+
+	err = tcf_ct_fill_params(net, params, parm, tb, extack);
+	if (err)
+		goto cleanup;
+
+	spin_lock_bh(&c->tcf_lock);
+	goto_ch = tcf_action_set_ctrlact(*a, parm->action, goto_ch);
+	rcu_swap_protected(c->params, params, lockdep_is_held(&c->tcf_lock));
+	spin_unlock_bh(&c->tcf_lock);
+
+	if (goto_ch)
+		tcf_chain_put_by_act(goto_ch);
+	if (params)
+		kfree_rcu(params, rcu);
+	if (res == ACT_P_CREATED)
+		tcf_idr_insert(tn, *a);
+
+	return res;
+
+cleanup:
+	if (goto_ch)
+		tcf_chain_put_by_act(goto_ch);
+	kfree(params);
+	tcf_idr_release(*a, bind);
+	return err;
+}
+
+static void tcf_ct_cleanup(struct tc_action *a)
+{
+	struct tcf_ct_params *params;
+	struct tcf_ct *c = to_ct(a);
+
+	params = rcu_dereference_protected(c->params, 1);
+	if (params)
+		call_rcu(&params->rcu, tcf_ct_params_free);
+}
+
+static int tcf_ct_dump_key_val(struct sk_buff *skb,
+			       void *val, int val_type,
+			       void *mask, int mask_type,
+			       int len)
+{
+	int err;
+
+	if (mask && !memchr_inv(mask, 0, len))
+		return 0;
+
+	err = nla_put(skb, val_type, len, val);
+	if (err)
+		return err;
+
+	if (mask_type != TCA_CT_UNSPEC) {
+		err = nla_put(skb, mask_type, len, mask);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int tcf_ct_dump_nat(struct sk_buff *skb, struct tcf_ct_params *p)
+{
+	struct nf_nat_range2 *range = &p->range;
+
+	if (!(p->ct_action & TCA_CT_ACT_NAT))
+		return 0;
+
+	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
+		return 0;
+
+	if (range->flags & NF_NAT_RANGE_MAP_IPS) {
+		if (p->ipv4_range) {
+			if (nla_put_in_addr(skb, TCA_CT_NAT_IPV4_MIN,
+					    range->min_addr.ip))
+				return -1;
+			if (nla_put_in_addr(skb, TCA_CT_NAT_IPV4_MAX,
+					    range->max_addr.ip))
+				return -1;
+		} else {
+			if (nla_put_in6_addr(skb, TCA_CT_NAT_IPV6_MIN,
+					     &range->min_addr.in6))
+				return -1;
+			if (nla_put_in6_addr(skb, TCA_CT_NAT_IPV6_MAX,
+					     &range->max_addr.in6))
+				return -1;
+		}
+	}
+
+	if (range->flags & NF_NAT_RANGE_PROTO_SPECIFIED) {
+		if (nla_put_u16(skb, TCA_CT_NAT_PORT_MIN,
+				ntohs(range->min_proto.all)))
+			return -1;
+		if (nla_put_u16(skb, TCA_CT_NAT_PORT_MAX,
+				ntohs(range->max_proto.all)))
+			return -1;
+	}
+
+	return 0;
+}
+
+static inline int tcf_ct_dump(struct sk_buff *skb, struct tc_action *a,
+			      int bind, int ref)
+{
+	unsigned char *b = skb_tail_pointer(skb);
+	struct tcf_ct *c = to_ct(a);
+	struct tcf_ct_params *p;
+
+	struct tc_ct opt = {
+		.index   = c->tcf_index,
+		.refcnt  = refcount_read(&c->tcf_refcnt) - ref,
+		.bindcnt = atomic_read(&c->tcf_bindcnt) - bind,
+	};
+	struct tcf_t t;
+
+	spin_lock_bh(&c->tcf_lock);
+	p = rcu_dereference_protected(c->params,
+				      lockdep_is_held(&c->tcf_lock));
+	opt.action = c->tcf_action;
+
+	if (tcf_ct_dump_key_val(skb,
+				&p->ct_action, TCA_CT_ACTION,
+				NULL, TCA_CT_UNSPEC,
+				sizeof(p->ct_action)))
+		goto nla_put_failure;
+
+	if (p->ct_action & TCA_CT_ACT_CLEAR)
+		goto skip_dump;
+
+	if (IS_ENABLED(CONFIG_NF_CONNTRACK_MARK) &&
+	    tcf_ct_dump_key_val(skb,
+				&p->mark, TCA_CT_MARK,
+				&p->mark_mask, TCA_CT_MARK_MASK,
+				sizeof(p->mark)))
+		goto nla_put_failure;
+
+	if (IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS) &&
+	    tcf_ct_dump_key_val(skb,
+				p->labels, TCA_CT_LABELS,
+				p->labels_mask, TCA_CT_LABELS_MASK,
+				sizeof(p->labels)))
+		goto nla_put_failure;
+
+	if (IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES) &&
+	    tcf_ct_dump_key_val(skb,
+				&p->zone, TCA_CT_ZONE,
+				NULL, TCA_CT_UNSPEC,
+				sizeof(p->zone)))
+		goto nla_put_failure;
+
+	if (tcf_ct_dump_nat(skb, p))
+		goto nla_put_failure;
+
+skip_dump:
+	if (nla_put(skb, TCA_CT_PARMS, sizeof(opt), &opt))
+		goto nla_put_failure;
+
+	tcf_tm_dump(&t, &c->tcf_tm);
+	if (nla_put_64bit(skb, TCA_CT_TM, sizeof(t), &t, TCA_CT_PAD))
+		goto nla_put_failure;
+	spin_unlock_bh(&c->tcf_lock);
+
+	return skb->len;
+nla_put_failure:
+	spin_unlock_bh(&c->tcf_lock);
+	nlmsg_trim(skb, b);
+	return -1;
+}
+
+static int tcf_ct_walker(struct net *net, struct sk_buff *skb,
+			 struct netlink_callback *cb, int type,
+			 const struct tc_action_ops *ops,
+			 struct netlink_ext_ack *extack)
+{
+	struct tc_action_net *tn = net_generic(net, ct_net_id);
+
+	return tcf_generic_walker(tn, skb, cb, type, ops, extack);
+}
+
+static int tcf_ct_search(struct net *net, struct tc_action **a, u32 index)
+{
+	struct tc_action_net *tn = net_generic(net, ct_net_id);
+
+	return tcf_idr_search(tn, a, index);
+}
+
+static void tcf_stats_update(struct tc_action *a, u64 bytes, u32 packets,
+			     u64 lastuse, bool hw)
+{
+	struct tcf_ct *c = to_ct(a);
+
+	_bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), bytes, packets);
+
+	if (hw)
+		_bstats_cpu_update(this_cpu_ptr(a->cpu_bstats_hw),
+				   bytes, packets);
+	c->tcf_tm.lastuse = max_t(u64, c->tcf_tm.lastuse, lastuse);
+}
+
+static struct tc_action_ops act_ct_ops = {
+	.kind		=	"ct",
+	.id		=	TCA_ID_CT,
+	.owner		=	THIS_MODULE,
+	.act		=	tcf_ct_act,
+	.dump		=	tcf_ct_dump,
+	.init		=	tcf_ct_init,
+	.cleanup	=	tcf_ct_cleanup,
+	.walk		=	tcf_ct_walker,
+	.lookup		=	tcf_ct_search,
+	.stats_update	=	tcf_stats_update,
+	.size		=	sizeof(struct tcf_ct),
+};
+
+static __net_init int ct_init_net(struct net *net)
+{
+	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
+	unsigned int n_bits = FIELD_SIZEOF(struct tcf_ct_params, labels) * 8;
+
+	if (nf_connlabels_get(net, n_bits - 1)) {
+		tn->labels = false;
+		pr_err("act_ct: Failed to set connlabels length");
+	} else {
+		tn->labels = true;
+	}
+
+	return tc_action_net_init(&tn->tn, &act_ct_ops);
+}
+
+static void __net_exit ct_exit_net(struct list_head *net_list)
+{
+	struct net *net;
+
+	rtnl_lock();
+	list_for_each_entry(net, net_list, exit_list) {
+		struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
+
+		if (tn->labels)
+			nf_connlabels_put(net);
+	}
+	rtnl_unlock();
+
+	tc_action_net_exit(net_list, ct_net_id);
+}
+
+static struct pernet_operations ct_net_ops = {
+	.init = ct_init_net,
+	.exit_batch = ct_exit_net,
+	.id   = &ct_net_id,
+	.size = sizeof(struct tc_ct_action_net),
+};
+
+static int __init ct_init_module(void)
+{
+	return tcf_register_action(&act_ct_ops, &ct_net_ops);
+}
+
+static void __exit ct_cleanup_module(void)
+{
+	tcf_unregister_action(&act_ct_ops, &ct_net_ops);
+}
+
+module_init(ct_init_module);
+module_exit(ct_cleanup_module);
+MODULE_AUTHOR("Paul Blakey <paulb@mellanox.com>");
+MODULE_AUTHOR("Yossi Kuperman <yossiku@mellanox.com>");
+MODULE_AUTHOR("Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>");
+MODULE_DESCRIPTION("Connection tracking action");
+MODULE_LICENSE("GPL v2");
+
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index ad36bbc..4a7331c 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -35,6 +35,7 @@
 #include <net/tc_act/tc_police.h>
 #include <net/tc_act/tc_sample.h>
 #include <net/tc_act/tc_skbedit.h>
+#include <net/tc_act/tc_ct.h>
 
 extern const struct nla_policy rtm_tca_policy[TCA_MAX + 1];
 
@@ -3266,6 +3267,10 @@ int tc_setup_flow_action(struct flow_action *flow_action,
 			entry->police.burst = tcf_police_tcfp_burst(act);
 			entry->police.rate_bytes_ps =
 				tcf_police_rate_bytes_ps(act);
+		} else if (is_tcf_ct(act)) {
+			entry->id = FLOW_ACTION_CT;
+			entry->ct.action = tcf_ct_action(act);
+			entry->ct.zone = tcf_ct_zone(act);
 		} else {
 			goto err_out;
 		}
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 28+ messages in thread

* [PATCH net-next 2/3] net/flow_dissector: add connection tracking dissection
  2019-06-11 13:28 [PATCH net-next 0/3] net/sched: Introduce tc connection tracking Paul Blakey
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
@ 2019-06-11 13:28 ` Paul Blakey
  2019-06-11 13:28 ` [PATCH net-next 3/3] net/sched: cls_flower: Add matching on conntrack info Paul Blakey
  2 siblings, 0 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-11 13:28 UTC (permalink / raw)
  To: Jiri Pirko, Paul Blakey, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang
  Cc: Rony Efraim, nst-kernel, John Hurley, Simon Horman, Justin Pettit

Retreives connection tracking zone, mark, label, and state from
a SKB.

Signed-off-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
 include/linux/skbuff.h       | 10 ++++++++++
 include/net/flow_dissector.h | 15 +++++++++++++++
 net/core/flow_dissector.c    | 44 ++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 69 insertions(+)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 28bdaf9..e91cc60 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -1320,6 +1320,16 @@ static inline bool skb_flow_dissect_flow_keys(const struct sk_buff *skb,
 				  data, proto, nhoff, hlen, flags);
 }
 
+/* Gets a skb connection tracking info, ctinfo map should be a
+ * a map of mapsize to translate enum ip_conntrack_info states
+ * to user states.
+ */
+void
+skb_flow_dissect_ct(const struct sk_buff *skb,
+		    struct flow_dissector *flow_dissector,
+		    void *target_container,
+		    u16 *ctinfo_map,
+		    size_t mapsize);
 void
 skb_flow_dissect_tunnel_info(const struct sk_buff *skb,
 			     struct flow_dissector *flow_dissector,
diff --git a/include/net/flow_dissector.h b/include/net/flow_dissector.h
index 797e19c..4576098 100644
--- a/include/net/flow_dissector.h
+++ b/include/net/flow_dissector.h
@@ -199,6 +199,20 @@ struct flow_dissector_key_ip {
 	__u8	ttl;
 };
 
+/**
+ * struct flow_dissector_key_ct:
+ * @ct_state: conntrack state after converting with map
+ * @ct_mark: conttrack mark
+ * @ct_zone: conntrack zone
+ * @ct_labels: conntrack labels
+ */
+struct flow_dissector_key_ct {
+	u16	ct_state;
+	u16	ct_zone;
+	u32	ct_mark;
+	u32	ct_labels[4];
+};
+
 enum flow_dissector_key_id {
 	FLOW_DISSECTOR_KEY_CONTROL, /* struct flow_dissector_key_control */
 	FLOW_DISSECTOR_KEY_BASIC, /* struct flow_dissector_key_basic */
@@ -224,6 +238,7 @@ enum flow_dissector_key_id {
 	FLOW_DISSECTOR_KEY_CVLAN, /* struct flow_dissector_key_vlan */
 	FLOW_DISSECTOR_KEY_ENC_IP, /* struct flow_dissector_key_ip */
 	FLOW_DISSECTOR_KEY_ENC_OPTS, /* struct flow_dissector_key_enc_opts */
+	FLOW_DISSECTOR_KEY_CT, /* struct flow_dissector_key_ct */
 
 	FLOW_DISSECTOR_KEY_MAX,
 };
diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index c0559af..293bffc 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -27,6 +27,10 @@
 #include <scsi/fc/fc_fcoe.h>
 #include <uapi/linux/batadv_packet.h>
 #include <linux/bpf.h>
+#if IS_ENABLED(CONFIG_NF_CONNTRACK)
+#include <net/netfilter/nf_conntrack_core.h>
+#include <net/netfilter/nf_conntrack_labels.h>
+#endif
 
 static DEFINE_MUTEX(flow_dissector_mutex);
 
@@ -216,6 +220,46 @@ __be32 __skb_flow_get_ports(const struct sk_buff *skb, int thoff, u8 ip_proto,
 }
 
 void
+skb_flow_dissect_ct(const struct sk_buff *skb,
+		    struct flow_dissector *flow_dissector,
+		    void *target_container,
+		    u16 *ctinfo_map,
+		    size_t mapsize)
+{
+#if IS_ENABLED(CONFIG_NF_CONNTRACK)
+	struct flow_dissector_key_ct *key;
+	enum ip_conntrack_info ctinfo;
+	struct nf_conn_labels *cl;
+	struct nf_conn *ct;
+
+	if (!dissector_uses_key(flow_dissector, FLOW_DISSECTOR_KEY_CT))
+		return;
+
+	ct = nf_ct_get(skb, &ctinfo);
+	if (!ct)
+		return;
+
+	key = skb_flow_dissector_target(flow_dissector,
+					FLOW_DISSECTOR_KEY_CT,
+					target_container);
+
+	if (ctinfo < mapsize)
+		key->ct_state = ctinfo_map[ctinfo];
+#if IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)
+	key->ct_zone = ct->zone.id;
+#endif
+#if IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)
+	key->ct_mark = ct->mark;
+#endif
+
+	cl = nf_ct_labels_find(ct);
+	if (cl)
+		memcpy(key->ct_labels, cl->bits, sizeof(key->ct_labels));
+#endif /* CONFIG_NF_CONNTRACK */
+}
+EXPORT_SYMBOL(skb_flow_dissect_ct);
+
+void
 skb_flow_dissect_tunnel_info(const struct sk_buff *skb,
 			     struct flow_dissector *flow_dissector,
 			     void *target_container)
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 28+ messages in thread

* [PATCH net-next 3/3] net/sched: cls_flower: Add matching on conntrack info
  2019-06-11 13:28 [PATCH net-next 0/3] net/sched: Introduce tc connection tracking Paul Blakey
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
  2019-06-11 13:28 ` [PATCH net-next 2/3] net/flow_dissector: add connection tracking dissection Paul Blakey
@ 2019-06-11 13:28 ` Paul Blakey
  2 siblings, 0 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-11 13:28 UTC (permalink / raw)
  To: Jiri Pirko, Paul Blakey, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang
  Cc: Rony Efraim, nst-kernel, John Hurley, Simon Horman, Justin Pettit

New matches for conntrack mark, label, zone, and state.

Signed-off-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
 include/uapi/linux/pkt_cls.h |  16 ++++++
 net/sched/cls_flower.c       | 127 +++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 138 insertions(+), 5 deletions(-)

diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
index c5264d7..613c9c5 100644
--- a/include/uapi/linux/pkt_cls.h
+++ b/include/uapi/linux/pkt_cls.h
@@ -537,12 +537,28 @@ enum {
 	TCA_FLOWER_KEY_PORT_DST_MIN,	/* be16 */
 	TCA_FLOWER_KEY_PORT_DST_MAX,	/* be16 */
 
+	TCA_FLOWER_KEY_CT_STATE,	/* u16 */
+	TCA_FLOWER_KEY_CT_STATE_MASK,	/* u16 */
+	TCA_FLOWER_KEY_CT_ZONE,		/* u16 */
+	TCA_FLOWER_KEY_CT_ZONE_MASK,	/* u16 */
+	TCA_FLOWER_KEY_CT_MARK,		/* u32 */
+	TCA_FLOWER_KEY_CT_MARK_MASK,	/* u32 */
+	TCA_FLOWER_KEY_CT_LABELS,	/* u128 */
+	TCA_FLOWER_KEY_CT_LABELS_MASK,	/* u128 */
+
 	__TCA_FLOWER_MAX,
 };
 
 #define TCA_FLOWER_MAX (__TCA_FLOWER_MAX - 1)
 
 enum {
+	TCA_FLOWER_KEY_CT_FLAGS_NEW = 1 << 0, /* Beginning of a new connection. */
+	TCA_FLOWER_KEY_CT_FLAGS_ESTABLISHED = 1 << 1, /* Part of an existing connection. */
+	TCA_FLOWER_KEY_CT_FLAGS_RELATED = 1 << 2, /* Related to an established connection. */
+	TCA_FLOWER_KEY_CT_FLAGS_TRACKED = 1 << 3, /* Conntrack has occurred. */
+};
+
+enum {
 	TCA_FLOWER_KEY_ENC_OPTS_UNSPEC,
 	TCA_FLOWER_KEY_ENC_OPTS_GENEVE, /* Nested
 					 * TCA_FLOWER_KEY_ENC_OPT_GENEVE_
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index c388372..3aa7998 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -26,6 +26,8 @@
 #include <net/dst.h>
 #include <net/dst_metadata.h>
 
+#include <uapi/linux/netfilter/nf_conntrack_common.h>
+
 struct fl_flow_key {
 	int	indev_ifindex;
 	struct flow_dissector_key_control control;
@@ -54,6 +56,7 @@ struct fl_flow_key {
 	struct flow_dissector_key_enc_opts enc_opts;
 	struct flow_dissector_key_ports tp_min;
 	struct flow_dissector_key_ports tp_max;
+	struct flow_dissector_key_ct ct;
 } __aligned(BITS_PER_LONG / 8); /* Ensure that we can do comparisons as longs. */
 
 struct fl_flow_mask_range {
@@ -272,14 +275,27 @@ static struct cls_fl_filter *fl_lookup(struct fl_flow_mask *mask,
 	return __fl_lookup(mask, mkey);
 }
 
+static u16 fl_ct_info_to_flower_map[] = {
+	[IP_CT_ESTABLISHED] =		TCA_FLOWER_KEY_CT_FLAGS_TRACKED |
+					TCA_FLOWER_KEY_CT_FLAGS_ESTABLISHED,
+	[IP_CT_RELATED] =		TCA_FLOWER_KEY_CT_FLAGS_TRACKED |
+					TCA_FLOWER_KEY_CT_FLAGS_RELATED,
+	[IP_CT_ESTABLISHED_REPLY] =	TCA_FLOWER_KEY_CT_FLAGS_TRACKED |
+					TCA_FLOWER_KEY_CT_FLAGS_ESTABLISHED,
+	[IP_CT_RELATED_REPLY] =		TCA_FLOWER_KEY_CT_FLAGS_TRACKED |
+					TCA_FLOWER_KEY_CT_FLAGS_RELATED,
+	[IP_CT_NEW] =			TCA_FLOWER_KEY_CT_FLAGS_TRACKED |
+					TCA_FLOWER_KEY_CT_FLAGS_NEW,
+};
+
 static int fl_classify(struct sk_buff *skb, const struct tcf_proto *tp,
 		       struct tcf_result *res)
 {
 	struct cls_fl_head *head = rcu_dereference_bh(tp->root);
-	struct cls_fl_filter *f;
-	struct fl_flow_mask *mask;
-	struct fl_flow_key skb_key;
 	struct fl_flow_key skb_mkey;
+	struct fl_flow_key skb_key;
+	struct fl_flow_mask *mask;
+	struct cls_fl_filter *f;
 
 	list_for_each_entry_rcu(mask, &head->masks, list) {
 		fl_clear_masked_range(&skb_key, mask);
@@ -290,6 +306,9 @@ static int fl_classify(struct sk_buff *skb, const struct tcf_proto *tp,
 		 */
 		skb_key.basic.n_proto = skb->protocol;
 		skb_flow_dissect_tunnel_info(skb, &mask->dissector, &skb_key);
+		skb_flow_dissect_ct(skb, &mask->dissector, &skb_key,
+				    fl_ct_info_to_flower_map,
+				    ARRAY_SIZE(fl_ct_info_to_flower_map));
 		skb_flow_dissect(skb, &mask->dissector, &skb_key, 0);
 
 		fl_set_masked_key(&skb_mkey, &skb_key, mask);
@@ -693,6 +712,16 @@ static void *fl_get(struct tcf_proto *tp, u32 handle)
 	[TCA_FLOWER_KEY_ENC_IP_TTL_MASK] = { .type = NLA_U8 },
 	[TCA_FLOWER_KEY_ENC_OPTS]	= { .type = NLA_NESTED },
 	[TCA_FLOWER_KEY_ENC_OPTS_MASK]	= { .type = NLA_NESTED },
+	[TCA_FLOWER_KEY_CT_STATE]	= { .type = NLA_U16 },
+	[TCA_FLOWER_KEY_CT_STATE_MASK]	= { .type = NLA_U16 },
+	[TCA_FLOWER_KEY_CT_ZONE]	= { .type = NLA_U16 },
+	[TCA_FLOWER_KEY_CT_ZONE_MASK]	= { .type = NLA_U16 },
+	[TCA_FLOWER_KEY_CT_MARK]	= { .type = NLA_U32 },
+	[TCA_FLOWER_KEY_CT_MARK_MASK]	= { .type = NLA_U32 },
+	[TCA_FLOWER_KEY_CT_LABELS]	= { .type = NLA_BINARY,
+					    .len = 128 / BITS_PER_BYTE },
+	[TCA_FLOWER_KEY_CT_LABELS_MASK]	= { .type = NLA_BINARY,
+					    .len = 128 / BITS_PER_BYTE },
 };
 
 static const struct nla_policy
@@ -714,11 +743,11 @@ static void fl_set_key_val(struct nlattr **tb,
 {
 	if (!tb[val_type])
 		return;
-	memcpy(val, nla_data(tb[val_type]), len);
+	nla_memcpy(val, tb[val_type], len);
 	if (mask_type == TCA_FLOWER_UNSPEC || !tb[mask_type])
 		memset(mask, 0xff, len);
 	else
-		memcpy(mask, nla_data(tb[mask_type]), len);
+		nla_memcpy(mask, tb[mask_type], len);
 }
 
 static int fl_set_key_port_range(struct nlattr **tb, struct fl_flow_key *key,
@@ -1004,6 +1033,51 @@ static int fl_set_enc_opt(struct nlattr **tb, struct fl_flow_key *key,
 	return 0;
 }
 
+static int fl_set_key_ct(struct nlattr **tb,
+			 struct flow_dissector_key_ct *key,
+			 struct flow_dissector_key_ct *mask,
+			 struct netlink_ext_ack *extack)
+{
+	if (tb[TCA_FLOWER_KEY_CT_STATE]) {
+		if (!IS_ENABLED(CONFIG_NF_CONNTRACK)) {
+			NL_SET_ERR_MSG(extack, "Conntrack isn't enabled");
+			return -EOPNOTSUPP;
+		}
+		fl_set_key_val(tb, &key->ct_state, TCA_FLOWER_KEY_CT_STATE,
+			       &mask->ct_state, TCA_FLOWER_KEY_CT_STATE_MASK,
+			       sizeof(key->ct_state));
+	}
+	if (tb[TCA_FLOWER_KEY_CT_ZONE]) {
+		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
+			NL_SET_ERR_MSG(extack, "Conntrack zones isn't enabled");
+			return -EOPNOTSUPP;
+		}
+		fl_set_key_val(tb, &key->ct_zone, TCA_FLOWER_KEY_CT_ZONE,
+			       &mask->ct_zone, TCA_FLOWER_KEY_CT_ZONE_MASK,
+			       sizeof(key->ct_zone));
+	}
+	if (tb[TCA_FLOWER_KEY_CT_MARK]) {
+		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)) {
+			NL_SET_ERR_MSG(extack, "Conntrack mark isn't enabled");
+			return -EOPNOTSUPP;
+		}
+		fl_set_key_val(tb, &key->ct_mark, TCA_FLOWER_KEY_CT_MARK,
+			       &mask->ct_mark, TCA_FLOWER_KEY_CT_MARK_MASK,
+			       sizeof(key->ct_mark));
+	}
+	if (tb[TCA_FLOWER_KEY_CT_LABELS]) {
+		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)) {
+			NL_SET_ERR_MSG(extack, "Conntrack labels aren't enabled");
+			return -EOPNOTSUPP;
+		}
+		fl_set_key_val(tb, key->ct_labels, TCA_FLOWER_KEY_CT_LABELS,
+			       mask->ct_labels, TCA_FLOWER_KEY_CT_LABELS_MASK,
+			       sizeof(key->ct_labels));
+	}
+
+	return 0;
+}
+
 static int fl_set_key(struct net *net, struct nlattr **tb,
 		      struct fl_flow_key *key, struct fl_flow_key *mask,
 		      struct netlink_ext_ack *extack)
@@ -1214,6 +1288,10 @@ static int fl_set_key(struct net *net, struct nlattr **tb,
 			return ret;
 	}
 
+	ret = fl_set_key_ct(tb, &key->ct, &mask->ct, extack);
+	if (ret)
+		return ret;
+
 	if (tb[TCA_FLOWER_KEY_FLAGS])
 		ret = fl_set_key_flags(tb, &key->control.flags, &mask->control.flags);
 
@@ -1312,6 +1390,8 @@ static void fl_init_dissector(struct flow_dissector *dissector,
 			     FLOW_DISSECTOR_KEY_ENC_IP, enc_ip);
 	FL_KEY_SET_IF_MASKED(mask, keys, cnt,
 			     FLOW_DISSECTOR_KEY_ENC_OPTS, enc_opts);
+	FL_KEY_SET_IF_MASKED(mask, keys, cnt,
+			     FLOW_DISSECTOR_KEY_CT, ct);
 
 	skb_flow_dissector_init(dissector, keys, cnt);
 }
@@ -2075,6 +2155,40 @@ static int fl_dump_key_geneve_opt(struct sk_buff *skb,
 	return -EMSGSIZE;
 }
 
+static int fl_dump_key_ct(struct sk_buff *skb,
+			  struct flow_dissector_key_ct *key,
+			  struct flow_dissector_key_ct *mask)
+{
+	if (IS_ENABLED(CONFIG_NF_CONNTRACK) &&
+	    fl_dump_key_val(skb, &key->ct_state, TCA_FLOWER_KEY_CT_STATE,
+			    &mask->ct_state, TCA_FLOWER_KEY_CT_STATE_MASK,
+			    sizeof(key->ct_state)))
+		goto nla_put_failure;
+
+	if (IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES) &&
+	    fl_dump_key_val(skb, &key->ct_zone, TCA_FLOWER_KEY_CT_ZONE,
+			    &mask->ct_zone, TCA_FLOWER_KEY_CT_ZONE_MASK,
+			    sizeof(key->ct_zone)))
+		goto nla_put_failure;
+
+	if (IS_ENABLED(CONFIG_NF_CONNTRACK_MARK) &&
+	    fl_dump_key_val(skb, &key->ct_mark, TCA_FLOWER_KEY_CT_MARK,
+			    &mask->ct_mark, TCA_FLOWER_KEY_CT_MARK_MASK,
+			    sizeof(key->ct_mark)))
+		goto nla_put_failure;
+
+	if (IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS) &&
+	    fl_dump_key_val(skb, &key->ct_labels, TCA_FLOWER_KEY_CT_LABELS,
+			    &mask->ct_labels, TCA_FLOWER_KEY_CT_LABELS_MASK,
+			    sizeof(key->ct_labels)))
+		goto nla_put_failure;
+
+	return 0;
+
+nla_put_failure:
+	return -EMSGSIZE;
+}
+
 static int fl_dump_key_options(struct sk_buff *skb, int enc_opt_type,
 			       struct flow_dissector_key_enc_opts *enc_opts)
 {
@@ -2308,6 +2422,9 @@ static int fl_dump_key(struct sk_buff *skb, struct net *net,
 	    fl_dump_key_enc_opt(skb, &key->enc_opts, &mask->enc_opts))
 		goto nla_put_failure;
 
+	if (fl_dump_key_ct(skb, &key->ct, &mask->ct))
+		goto nla_put_failure;
+
 	if (fl_dump_key_flags(skb, key->control.flags, mask->control.flags))
 		goto nla_put_failure;
 
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
@ 2019-06-11 13:59   ` Toke Høiland-Jørgensen
  2019-06-11 14:15     ` Paul Blakey
  2019-06-12 18:04   ` Marcelo Ricardo Leitner
                     ` (4 subsequent siblings)
  5 siblings, 1 reply; 28+ messages in thread
From: Toke Høiland-Jørgensen @ 2019-06-11 13:59 UTC (permalink / raw)
  To: Paul Blakey, Jiri Pirko, Paul Blakey, Roi Dayan, Yossi Kuperman,
	Oz Shlomo, Marcelo Ricardo Leitner, netdev, David Miller,
	Aaron Conole, Zhike Wang
  Cc: Rony Efraim, nst-kernel, John Hurley, Simon Horman,
	Justin Pettit, Kevin Darbyshire-Bryant

Paul Blakey <paulb@mellanox.com> writes:

> Allow sending a packet to conntrack and set conntrack zone, mark,
> labels and nat parameters.

How is this different from the newly merged ctinfo action?

-Toke

> Signed-off-by: Paul Blakey <paulb@mellanox.com>
> Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
> Acked-by: Jiri Pirko <jiri@mellanox.com>
> ---
>  include/net/flow_offload.h        |   5 +
>  include/net/tc_act/tc_ct.h        |  64 +++
>  include/uapi/linux/pkt_cls.h      |   2 +
>  include/uapi/linux/tc_act/tc_ct.h |  41 ++
>  net/sched/Kconfig                 |  11 +
>  net/sched/Makefile                |   1 +
>  net/sched/act_ct.c                | 901 ++++++++++++++++++++++++++++++++++++++
>  net/sched/cls_api.c               |   5 +
>  8 files changed, 1030 insertions(+)
>  create mode 100644 include/net/tc_act/tc_ct.h
>  create mode 100644 include/uapi/linux/tc_act/tc_ct.h
>  create mode 100644 net/sched/act_ct.c
>
> diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
> index 36fdb85..5b2c4fa 100644
> --- a/include/net/flow_offload.h
> +++ b/include/net/flow_offload.h
> @@ -123,6 +123,7 @@ enum flow_action_id {
>  	FLOW_ACTION_QUEUE,
>  	FLOW_ACTION_SAMPLE,
>  	FLOW_ACTION_POLICE,
> +	FLOW_ACTION_CT,
>  };
>  
>  /* This is mirroring enum pedit_header_type definition for easy mapping between
> @@ -172,6 +173,10 @@ struct flow_action_entry {
>  			s64			burst;
>  			u64			rate_bytes_ps;
>  		} police;
> +		struct {				/* FLOW_ACTION_CT */
> +			int action;
> +			u16 zone;
> +		} ct;
>  	};
>  };
>  
> diff --git a/include/net/tc_act/tc_ct.h b/include/net/tc_act/tc_ct.h
> new file mode 100644
> index 0000000..59e4f5e
> --- /dev/null
> +++ b/include/net/tc_act/tc_ct.h
> @@ -0,0 +1,64 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +#ifndef __NET_TC_CT_H
> +#define __NET_TC_CT_H
> +
> +#include <net/act_api.h>
> +#include <uapi/linux/tc_act/tc_ct.h>
> +
> +#if IS_ENABLED(CONFIG_NF_CONNTRACK)
> +#include <net/netfilter/nf_nat.h>
> +#include <net/netfilter/nf_conntrack_labels.h>
> +
> +struct tcf_ct_params {
> +	struct nf_conn *tmpl;
> +	u16 zone;
> +
> +	u32 mark;
> +	u32 mark_mask;
> +
> +	u32 labels[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
> +	u32 labels_mask[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
> +
> +	struct nf_nat_range2 range;
> +	bool ipv4_range;
> +
> +	u16 ct_action;
> +
> +	struct rcu_head rcu;
> +
> +};
> +
> +struct tcf_ct {
> +	struct tc_action common;
> +	struct tcf_ct_params __rcu *params;
> +};
> +
> +#define to_ct(a) ((struct tcf_ct *)a)
> +#define to_ct_params(a) ((struct tcf_ct_params *) \
> +			 rtnl_dereference((to_ct(a)->params)))
> +
> +static inline uint16_t tcf_ct_zone(const struct tc_action *a)
> +{
> +	return to_ct_params(a)->zone;
> +}
> +
> +static inline int tcf_ct_action(const struct tc_action *a)
> +{
> +	return to_ct_params(a)->ct_action;
> +}
> +
> +#else
> +static inline uint16_t tcf_ct_zone(const struct tc_action *a) { return 0; }
> +static inline int tcf_ct_action(const struct tc_action *a) { return 0; }
> +#endif /* CONFIG_NF_CONNTRACK */
> +
> +static inline bool is_tcf_ct(const struct tc_action *a)
> +{
> +#if defined(CONFIG_NET_CLS_ACT) && IS_ENABLED(CONFIG_NF_CONNTRACK)
> +	if (a->ops && a->ops->id == TCA_ID_CT)
> +		return true;
> +#endif
> +	return false;
> +}
> +
> +#endif /* __NET_TC_CT_H */
> diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
> index a93680f..c5264d7 100644
> --- a/include/uapi/linux/pkt_cls.h
> +++ b/include/uapi/linux/pkt_cls.h
> @@ -83,6 +83,7 @@ enum {
>  #define TCA_ACT_SIMP 22
>  #define TCA_ACT_IFE 25
>  #define TCA_ACT_SAMPLE 26
> +#define TCA_ACT_CT 27
>  
>  /* Action type identifiers*/
>  enum tca_id {
> @@ -106,6 +107,7 @@ enum tca_id {
>  	TCA_ID_SAMPLE = TCA_ACT_SAMPLE,
>  	/* other actions go here */
>  	TCA_ID_CTINFO,
> +	TCA_ID_CT,
>  	__TCA_ID_MAX = 255
>  };
>  
> diff --git a/include/uapi/linux/tc_act/tc_ct.h b/include/uapi/linux/tc_act/tc_ct.h
> new file mode 100644
> index 0000000..5fb1d7a
> --- /dev/null
> +++ b/include/uapi/linux/tc_act/tc_ct.h
> @@ -0,0 +1,41 @@
> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
> +#ifndef __UAPI_TC_CT_H
> +#define __UAPI_TC_CT_H
> +
> +#include <linux/types.h>
> +#include <linux/pkt_cls.h>
> +
> +enum {
> +	TCA_CT_UNSPEC,
> +	TCA_CT_PARMS,
> +	TCA_CT_TM,
> +	TCA_CT_ACTION,		/* u16 */
> +	TCA_CT_ZONE,		/* u16 */
> +	TCA_CT_MARK,		/* u32 */
> +	TCA_CT_MARK_MASK,	/* u32 */
> +	TCA_CT_LABELS,		/* u128 */
> +	TCA_CT_LABELS_MASK,	/* u128 */
> +	TCA_CT_NAT_IPV4_MIN,	/* be32 */
> +	TCA_CT_NAT_IPV4_MAX,	/* be32 */
> +	TCA_CT_NAT_IPV6_MIN,	/* struct in6_addr */
> +	TCA_CT_NAT_IPV6_MAX,	/* struct in6_addr */
> +	TCA_CT_NAT_PORT_MIN,	/* be16 */
> +	TCA_CT_NAT_PORT_MAX,	/* be16 */
> +	TCA_CT_PAD,
> +	__TCA_CT_MAX
> +};
> +
> +#define TCA_CT_MAX (__TCA_CT_MAX - 1)
> +
> +#define TCA_CT_ACT_COMMIT	(1 << 0)
> +#define TCA_CT_ACT_FORCE	(1 << 1)
> +#define TCA_CT_ACT_CLEAR	(1 << 2)
> +#define TCA_CT_ACT_NAT		(1 << 3)
> +#define TCA_CT_ACT_NAT_SRC	(1 << 4)
> +#define TCA_CT_ACT_NAT_DST	(1 << 5)
> +
> +struct tc_ct {
> +	tc_gen;
> +};
> +
> +#endif /* __UAPI_TC_CT_H */
> diff --git a/net/sched/Kconfig b/net/sched/Kconfig
> index d104f7e..0481a2a 100644
> --- a/net/sched/Kconfig
> +++ b/net/sched/Kconfig
> @@ -929,6 +929,17 @@ config NET_ACT_TUNNEL_KEY
>  	  To compile this code as a module, choose M here: the
>  	  module will be called act_tunnel_key.
>  
> +config NET_ACT_CT
> +        tristate "connection tracking tc action"
> +        depends on NET_CLS_ACT && NF_CONNTRACK
> +        help
> +	  Say Y here to allow sending the packets to conntrack module.
> +
> +	  If unsure, say N.
> +
> +	  To compile this code as a module, choose M here: the
> +	  module will be called act_ct.
> +
>  config NET_IFE_SKBMARK
>          tristate "Support to encoding decoding skb mark on IFE action"
>          depends on NET_ACT_IFE
> diff --git a/net/sched/Makefile b/net/sched/Makefile
> index d54bfcb..23d2202 100644
> --- a/net/sched/Makefile
> +++ b/net/sched/Makefile
> @@ -28,6 +28,7 @@ obj-$(CONFIG_NET_IFE_SKBMARK)	+= act_meta_mark.o
>  obj-$(CONFIG_NET_IFE_SKBPRIO)	+= act_meta_skbprio.o
>  obj-$(CONFIG_NET_IFE_SKBTCINDEX)	+= act_meta_skbtcindex.o
>  obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o
> +obj-$(CONFIG_NET_ACT_CT)	+= act_ct.o
>  obj-$(CONFIG_NET_SCH_FIFO)	+= sch_fifo.o
>  obj-$(CONFIG_NET_SCH_CBQ)	+= sch_cbq.o
>  obj-$(CONFIG_NET_SCH_HTB)	+= sch_htb.o
> diff --git a/net/sched/act_ct.c b/net/sched/act_ct.c
> new file mode 100644
> index 0000000..4eb0dd7
> --- /dev/null
> +++ b/net/sched/act_ct.c
> @@ -0,0 +1,901 @@
> +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
> +/* -
> + * net/sched/act_ct.c  Connection Tracking action
> + *
> + * Authors:   Paul Blakey <paulb@mellanox.com>
> + *            Yossi Kuperman <yossiku@mellanox.com>
> + *            Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> + */
> +
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/kernel.h>
> +#include <linux/skbuff.h>
> +#include <linux/rtnetlink.h>
> +#include <linux/pkt_cls.h>
> +#include <linux/ip.h>
> +#include <linux/ipv6.h>
> +#include <net/netlink.h>
> +#include <net/pkt_sched.h>
> +#include <net/act_api.h>
> +#include <uapi/linux/tc_act/tc_ct.h>
> +#include <net/tc_act/tc_ct.h>
> +
> +#include <linux/netfilter/nf_nat.h>
> +#include <net/netfilter/nf_conntrack.h>
> +#include <net/netfilter/nf_conntrack_core.h>
> +#include <net/netfilter/nf_conntrack_zones.h>
> +#include <net/netfilter/nf_conntrack_helper.h>
> +#include <net/pkt_cls.h>
> +
> +static struct tc_action_ops act_ct_ops;
> +static unsigned int ct_net_id;
> +
> +struct tc_ct_action_net {
> +	struct tc_action_net tn; /* Must be first */
> +	bool labels;
> +};
> +
> +/* Determine whether skb->_nfct is equal to the result of conntrack lookup. */
> +static bool tcf_ct_skb_nfct_cached(struct net *net, struct sk_buff *skb,
> +				   u16 zone_id, bool force)
> +{
> +	enum ip_conntrack_info ctinfo;
> +	struct nf_conn *ct;
> +
> +	ct = nf_ct_get(skb, &ctinfo);
> +	if (!ct)
> +		return false;
> +	if (!net_eq(net, read_pnet(&ct->ct_net)))
> +		return false;
> +	if (nf_ct_zone(ct)->id != zone_id)
> +		return false;
> +
> +	/* Force conntrack entry direction. */
> +	if (force && CTINFO2DIR(ctinfo) != IP_CT_DIR_ORIGINAL) {
> +		nf_conntrack_put(&ct->ct_general);
> +		nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
> +
> +		if (nf_ct_is_confirmed(ct))
> +			nf_ct_kill(ct);
> +
> +		return false;
> +	}
> +
> +	return true;
> +}
> +
> +/* Trim the skb to the length specified by the IP/IPv6 header,
> + * removing any trailing lower-layer padding. This prepares the skb
> + * for higher-layer processing that assumes skb->len excludes padding
> + * (such as nf_ip_checksum). The caller needs to pull the skb to the
> + * network header, and ensure ip_hdr/ipv6_hdr points to valid data.
> + */
> +static int tcf_ct_skb_network_trim(struct sk_buff *skb)
> +{
> +	unsigned int len;
> +	int err;
> +
> +	switch (skb->protocol) {
> +	case htons(ETH_P_IP):
> +		len = ntohs(ip_hdr(skb)->tot_len);
> +		break;
> +	case htons(ETH_P_IPV6):
> +		len = sizeof(struct ipv6hdr)
> +			+ ntohs(ipv6_hdr(skb)->payload_len);
> +		break;
> +	default:
> +		len = skb->len;
> +	}
> +
> +	err = pskb_trim_rcsum(skb, len);
> +
> +	return err;
> +}
> +
> +static u8 tcf_ct_skb_nf_family(struct sk_buff *skb)
> +{
> +	u8 family = NFPROTO_UNSPEC;
> +
> +	switch (skb->protocol) {
> +	case htons(ETH_P_IP):
> +		family = NFPROTO_IPV4;
> +		break;
> +	case htons(ETH_P_IPV6):
> +		family = NFPROTO_IPV6;
> +		break;
> +	default:
> +	break;
> +	}
> +
> +	return family;
> +}
> +
> +static void tcf_ct_params_free(struct rcu_head *head)
> +{
> +	struct tcf_ct_params *params = container_of(head,
> +						    struct tcf_ct_params, rcu);
> +
> +	if (params->tmpl)
> +		nf_conntrack_put(&params->tmpl->ct_general);
> +	kfree(params);
> +}
> +
> +#if IS_ENABLED(CONFIG_NF_NAT)
> +/* Modelled after nf_nat_ipv[46]_fn().
> + * range is only used for new, uninitialized NAT state.
> + * Returns either NF_ACCEPT or NF_DROP.
> + */
> +static int ct_nat_execute(struct sk_buff *skb, struct nf_conn *ct,
> +			  enum ip_conntrack_info ctinfo,
> +			  const struct nf_nat_range2 *range,
> +			  enum nf_nat_manip_type maniptype)
> +{
> +	int hooknum, nh_off, err = NF_ACCEPT;
> +
> +	nh_off = skb_network_offset(skb);
> +	skb_pull_rcsum(skb, nh_off);
> +
> +	/* See HOOK2MANIP(). */
> +	if (maniptype == NF_NAT_MANIP_SRC)
> +		hooknum = NF_INET_LOCAL_IN; /* Source NAT */
> +	else
> +		hooknum = NF_INET_LOCAL_OUT; /* Destination NAT */
> +
> +	switch (ctinfo) {
> +	case IP_CT_RELATED:
> +	case IP_CT_RELATED_REPLY:
> +		if (skb->protocol == htons(ETH_P_IP) &&
> +		    ip_hdr(skb)->protocol == IPPROTO_ICMP) {
> +			if (!nf_nat_icmp_reply_translation(skb, ct, ctinfo,
> +							   hooknum))
> +				err = NF_DROP;
> +			goto push;
> +		} else if (IS_ENABLED(CONFIG_IPV6) &&
> +			   skb->protocol == htons(ETH_P_IPV6)) {
> +			__be16 frag_off;
> +			u8 nexthdr = ipv6_hdr(skb)->nexthdr;
> +			int hdrlen = ipv6_skip_exthdr(skb,
> +						      sizeof(struct ipv6hdr),
> +						      &nexthdr, &frag_off);
> +
> +			if (hdrlen >= 0 && nexthdr == IPPROTO_ICMPV6) {
> +				if (!nf_nat_icmpv6_reply_translation(skb, ct,
> +								     ctinfo,
> +								     hooknum,
> +								     hdrlen))
> +					err = NF_DROP;
> +				goto push;
> +			}
> +		}
> +		/* Non-ICMP, fall thru to initialize if needed. */
> +		/* fall through */
> +	case IP_CT_NEW:
> +		/* Seen it before?  This can happen for loopback, retrans,
> +		 * or local packets.
> +		 */
> +		if (!nf_nat_initialized(ct, maniptype)) {
> +			/* Initialize according to the NAT action. */
> +			err = (range && range->flags & NF_NAT_RANGE_MAP_IPS)
> +				/* Action is set up to establish a new
> +				 * mapping.
> +				 */
> +				? nf_nat_setup_info(ct, range, maniptype)
> +				: nf_nat_alloc_null_binding(ct, hooknum);
> +			if (err != NF_ACCEPT)
> +				goto push;
> +		}
> +		break;
> +
> +	case IP_CT_ESTABLISHED:
> +	case IP_CT_ESTABLISHED_REPLY:
> +		break;
> +
> +	default:
> +		err = NF_DROP;
> +		goto push;
> +	}
> +
> +	err = nf_nat_packet(ct, ctinfo, hooknum, skb);
> +push:
> +	skb_push(skb, nh_off);
> +	skb_postpush_rcsum(skb, skb->data, nh_off);
> +
> +	return err;
> +}
> +#endif /* CONFIG_NF_NAT */
> +
> +static void tcf_ct_act_set_mark(struct nf_conn *ct, u32 mark, u32 mask)
> +{
> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)
> +	u32 new_mark;
> +
> +	if (!mask)
> +		return;
> +
> +	new_mark = mark | (ct->mark & ~(mask));
> +	if (ct->mark != new_mark) {
> +		ct->mark = new_mark;
> +		if (nf_ct_is_confirmed(ct))
> +			nf_conntrack_event_cache(IPCT_MARK, ct);
> +	}
> +#endif
> +}
> +
> +static void tcf_ct_act_set_labels(struct nf_conn *ct,
> +				  u32 *labels,
> +				  u32 *labels_m)
> +{
> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)
> +	size_t labels_sz = FIELD_SIZEOF(struct tcf_ct_params, labels);
> +
> +	if (!memchr_inv(labels_m, 0, labels_sz))
> +		return;
> +
> +	nf_connlabels_replace(ct, labels, labels_m, 4);
> +#endif
> +}
> +
> +static int tcf_ct_act_nat(struct sk_buff *skb,
> +			  struct nf_conn *ct,
> +			  enum ip_conntrack_info ctinfo,
> +			  int ct_action,
> +			  struct nf_nat_range2 *range,
> +			  bool commit)
> +{
> +#if IS_ENABLED(CONFIG_NF_NAT)
> +	enum nf_nat_manip_type maniptype;
> +
> +	if (!(ct_action & TCA_CT_ACT_NAT))
> +		return NF_ACCEPT;
> +
> +	/* Add NAT extension if not confirmed yet. */
> +	if (!nf_ct_is_confirmed(ct) && !nf_ct_nat_ext_add(ct))
> +		return NF_DROP;   /* Can't NAT. */
> +
> +	if (ctinfo != IP_CT_NEW && (ct->status & IPS_NAT_MASK) &&
> +	    (ctinfo != IP_CT_RELATED || commit)) {
> +		/* NAT an established or related connection like before. */
> +		if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY)
> +			/* This is the REPLY direction for a connection
> +			 * for which NAT was applied in the forward
> +			 * direction.  Do the reverse NAT.
> +			 */
> +			maniptype = ct->status & IPS_SRC_NAT
> +				? NF_NAT_MANIP_DST : NF_NAT_MANIP_SRC;
> +		else
> +			maniptype = ct->status & IPS_SRC_NAT
> +				? NF_NAT_MANIP_SRC : NF_NAT_MANIP_DST;
> +	} else if (ct_action & TCA_CT_ACT_NAT_SRC) {
> +		maniptype = NF_NAT_MANIP_SRC;
> +	} else if (ct_action & TCA_CT_ACT_NAT_DST) {
> +		maniptype = NF_NAT_MANIP_DST;
> +	} else {
> +		return NF_ACCEPT;
> +	}
> +
> +	return ct_nat_execute(skb, ct, ctinfo, range, maniptype);
> +#else
> +	return NF_ACCEPT;
> +#endif
> +}
> +
> +static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
> +		      struct tcf_result *res)
> +{
> +	struct net *net = dev_net(skb->dev);
> +	bool cached, commit, clear, force;
> +	enum ip_conntrack_info ctinfo;
> +	struct tcf_ct *c = to_ct(a);
> +	struct nf_conn *tmpl = NULL;
> +	struct nf_hook_state state;
> +	struct tcf_ct_params *p;
> +	struct nf_conn *ct;
> +	int nh_ofs, err;
> +	u8 family;
> +
> +	p = rcu_dereference_bh(c->params);
> +
> +	commit = p->ct_action & TCA_CT_ACT_COMMIT;
> +	clear = p->ct_action & TCA_CT_ACT_CLEAR;
> +	force = p->ct_action & TCA_CT_ACT_FORCE;
> +	tmpl = p->tmpl;
> +
> +	if (clear) {
> +		ct = nf_ct_get(skb, &ctinfo);
> +		if (ct) {
> +			nf_conntrack_put(&ct->ct_general);
> +			nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
> +		}
> +
> +		goto out;
> +	}
> +
> +	/* The conntrack module expects to be working at L3. */
> +	nh_ofs = skb_network_offset(skb);
> +	skb_pull_rcsum(skb, nh_ofs);
> +
> +	err = tcf_ct_skb_network_trim(skb);
> +	if (err)
> +		goto drop;
> +
> +	family = tcf_ct_skb_nf_family(skb);
> +	if (family == NFPROTO_UNSPEC)
> +		goto drop;
> +
> +	state.hook = NF_INET_PRE_ROUTING;
> +	state.net = net;
> +	state.pf = family;
> +
> +	/* If we are recirculating packets to match on ct fields and
> +	 * committing with a separate ct action, then we don't need to
> +	 * actually run the packet through conntrack twice unless it's for a
> +	 * different zone.
> +	 */
> +	cached = tcf_ct_skb_nfct_cached(net, skb, p->zone, force);
> +	if (!cached) {
> +		/* Associate skb with specified zone. */
> +		if (tmpl) {
> +			if (skb_nfct(skb))
> +				nf_conntrack_put(skb_nfct(skb));
> +			nf_conntrack_get(&tmpl->ct_general);
> +			nf_ct_set(skb, tmpl, IP_CT_NEW);
> +		}
> +
> +		err = nf_conntrack_in(skb, &state);
> +		if (err != NF_ACCEPT)
> +			goto out_push;
> +	}
> +
> +	ct = nf_ct_get(skb, &ctinfo);
> +	if (!ct)
> +		goto out_push;
> +	nf_ct_deliver_cached_events(ct);
> +
> +	err = tcf_ct_act_nat(skb, ct, ctinfo, p->ct_action, &p->range, commit);
> +	if (err != NF_ACCEPT)
> +		goto drop;
> +
> +	if (commit) {
> +		tcf_ct_act_set_mark(ct, p->mark, p->mark_mask);
> +		tcf_ct_act_set_labels(ct, p->labels, p->labels_mask);
> +
> +		/* This will take care of sending queued events
> +		 * even if the connection is already confirmed.
> +		 */
> +		nf_conntrack_confirm(skb);
> +	}
> +
> +out_push:
> +	skb_push(skb, nh_ofs);
> +	skb_postpush_rcsum(skb, skb->data, nh_ofs);
> +
> +out:
> +	bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), skb);
> +
> +	return c->tcf_action;
> +
> +drop:
> +	qstats_drop_inc(this_cpu_ptr(a->cpu_qstats));
> +	return TC_ACT_SHOT;
> +}
> +
> +static const struct nla_policy ct_policy[TCA_CT_MAX + 1] = {
> +	[TCA_CT_ACTION] = { .type = NLA_U16 },
> +	[TCA_CT_PARMS] = { .type = NLA_EXACT_LEN, .len = sizeof(struct tc_ct) },
> +	[TCA_CT_ZONE] = { .type = NLA_U16 },
> +	[TCA_CT_MARK] = { .type = NLA_U32 },
> +	[TCA_CT_MARK_MASK] = { .type = NLA_U32 },
> +	[TCA_CT_LABELS] = { .type = NLA_BINARY,
> +			    .len = 128 / BITS_PER_BYTE },
> +	[TCA_CT_LABELS_MASK] = { .type = NLA_BINARY,
> +				 .len = 128 / BITS_PER_BYTE },
> +	[TCA_CT_NAT_IPV4_MIN] = { .type = NLA_U32 },
> +	[TCA_CT_NAT_IPV4_MAX] = { .type = NLA_U32 },
> +	[TCA_CT_NAT_IPV6_MIN] = { .type = NLA_EXACT_LEN,
> +				  .len = sizeof(struct in6_addr) },
> +	[TCA_CT_NAT_IPV6_MAX] = { .type = NLA_EXACT_LEN,
> +				   .len = sizeof(struct in6_addr) },
> +	[TCA_CT_NAT_PORT_MIN] = { .type = NLA_U16 },
> +	[TCA_CT_NAT_PORT_MAX] = { .type = NLA_U16 },
> +};
> +
> +static int tcf_ct_fill_params_nat(struct tcf_ct_params *p,
> +				  struct tc_ct *parm,
> +				  struct nlattr **tb,
> +				  struct netlink_ext_ack *extack)
> +{
> +	struct nf_nat_range2 *range;
> +
> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
> +		return 0;
> +
> +	if (!IS_ENABLED(CONFIG_NF_NAT)) {
> +		NL_SET_ERR_MSG_MOD(extack, "Netfilter nat isn't enabled in kernel");
> +		return -EOPNOTSUPP;
> +	}
> +
> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
> +		return 0;
> +
> +	if ((p->ct_action & TCA_CT_ACT_NAT_SRC) &&
> +	    (p->ct_action & TCA_CT_ACT_NAT_DST)) {
> +		NL_SET_ERR_MSG_MOD(extack, "dnat and snat can't be enabled at the same time");
> +		return -EOPNOTSUPP;
> +	}
> +
> +	range = &p->range;
> +	if (tb[TCA_CT_NAT_IPV4_MIN]) {
> +		range->min_addr.ip =
> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = true;
> +	}
> +	if (tb[TCA_CT_NAT_IPV4_MAX]) {
> +		range->max_addr.ip =
> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = true;
> +	} else if (range->min_addr.ip) {
> +		range->max_addr.ip = range->min_addr.ip;
> +	}
> +
> +	if (tb[TCA_CT_NAT_IPV6_MIN]) {
> +		range->min_addr.in6 =
> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = false;
> +	}
> +	if (tb[TCA_CT_NAT_IPV6_MAX]) {
> +		range->max_addr.in6 =
> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = false;
> +	} else if (memchr_inv(&range->min_addr.in6, 0,
> +		   sizeof(range->min_addr.in6))) {
> +		range->max_addr.in6 = range->min_addr.in6;
> +	}
> +
> +	if (tb[TCA_CT_NAT_PORT_MIN]) {
> +		range->min_proto.all =
> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MIN]));
> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
> +	}
> +	if (tb[TCA_CT_NAT_PORT_MAX]) {
> +		range->max_proto.all =
> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MAX]));
> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
> +	} else if (range->min_proto.all) {
> +		range->max_proto.all = range->min_proto.all;
> +	}
> +
> +	return 0;
> +}
> +
> +static void tcf_ct_set_key_val(struct nlattr **tb,
> +			       void *val, int val_type,
> +			       void *mask, int mask_type,
> +			       int len)
> +{
> +	if (!tb[val_type])
> +		return;
> +	nla_memcpy(val, tb[val_type], len);
> +
> +	if (!mask)
> +		return;
> +
> +	if (mask_type == TCA_CT_UNSPEC || !tb[mask_type])
> +		memset(mask, 0xff, len);
> +	else
> +		nla_memcpy(mask, tb[mask_type], len);
> +}
> +
> +static int tcf_ct_fill_params(struct net *net,
> +			      struct tcf_ct_params *p,
> +			      struct tc_ct *parm,
> +			      struct nlattr **tb,
> +			      struct netlink_ext_ack *extack)
> +{
> +	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
> +	struct nf_conntrack_zone zone;
> +	struct nf_conn *tmpl;
> +	int err;
> +
> +	p->zone = NF_CT_DEFAULT_ZONE_ID;
> +
> +	tcf_ct_set_key_val(tb,
> +			   &p->ct_action, TCA_CT_ACTION,
> +			   NULL, TCA_CT_UNSPEC,
> +			   sizeof(p->ct_action));
> +
> +	if (p->ct_action & TCA_CT_ACT_CLEAR)
> +		return 0;
> +
> +	err = tcf_ct_fill_params_nat(p, parm, tb, extack);
> +	if (err)
> +		return err;
> +
> +	if (tb[TCA_CT_MARK]) {
> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)) {
> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack mark isn't enabled.");
> +			return -EOPNOTSUPP;
> +		}
> +		tcf_ct_set_key_val(tb,
> +				   &p->mark, TCA_CT_MARK,
> +				   &p->mark_mask, TCA_CT_MARK_MASK,
> +				   sizeof(p->mark));
> +	}
> +
> +	if (tb[TCA_CT_LABELS]) {
> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)) {
> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack labels isn't enabled.");
> +			return -EOPNOTSUPP;
> +		}
> +
> +		if (!tn->labels) {
> +			NL_SET_ERR_MSG_MOD(extack, "Failed to set connlabel length");
> +			return -EOPNOTSUPP;
> +		}
> +		tcf_ct_set_key_val(tb,
> +				   p->labels, TCA_CT_LABELS,
> +				   p->labels_mask, TCA_CT_LABELS_MASK,
> +				   sizeof(p->labels));
> +	}
> +
> +	if (tb[TCA_CT_ZONE]) {
> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack zones isn't enabled.");
> +			return -EOPNOTSUPP;
> +		}
> +
> +		tcf_ct_set_key_val(tb,
> +				   &p->zone, TCA_CT_ZONE,
> +				   NULL, TCA_CT_UNSPEC,
> +				   sizeof(p->zone));
> +	}
> +
> +	if (p->zone == NF_CT_DEFAULT_ZONE_ID)
> +		return 0;
> +
> +	nf_ct_zone_init(&zone, p->zone, NF_CT_DEFAULT_ZONE_DIR, 0);
> +	tmpl = nf_ct_tmpl_alloc(net, &zone, GFP_KERNEL);
> +	if (!tmpl) {
> +		NL_SET_ERR_MSG_MOD(extack, "Failed to allocate conntrack template");
> +		return -ENOMEM;
> +	}
> +	__set_bit(IPS_CONFIRMED_BIT, &tmpl->status);
> +	nf_conntrack_get(&tmpl->ct_general);
> +	p->tmpl = tmpl;
> +
> +	return 0;
> +}
> +
> +static int tcf_ct_init(struct net *net, struct nlattr *nla,
> +		       struct nlattr *est, struct tc_action **a,
> +		       int replace, int bind, bool rtnl_held,
> +		       struct tcf_proto *tp,
> +		       struct netlink_ext_ack *extack)
> +{
> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
> +	struct tcf_ct_params *params = NULL;
> +	struct nlattr *tb[TCA_CT_MAX + 1];
> +	struct tcf_chain *goto_ch = NULL;
> +	struct tc_ct *parm;
> +	struct tcf_ct *c;
> +	int err, res = 0;
> +
> +	if (!nla) {
> +		NL_SET_ERR_MSG_MOD(extack, "Ct requires attributes to be passed");
> +		return -EINVAL;
> +	}
> +
> +	err = nla_parse_nested(tb, TCA_CT_MAX, nla, ct_policy, extack);
> +	if (err < 0)
> +		return err;
> +
> +	if (!tb[TCA_CT_PARMS]) {
> +		NL_SET_ERR_MSG_MOD(extack, "Missing required ct parameters");
> +		return -EINVAL;
> +	}
> +	parm = nla_data(tb[TCA_CT_PARMS]);
> +
> +	err = tcf_idr_check_alloc(tn, &parm->index, a, bind);
> +	if (err < 0)
> +		return err;
> +
> +	if (!err) {
> +		err = tcf_idr_create(tn, parm->index, est, a,
> +				     &act_ct_ops, bind, true);
> +		if (err) {
> +			tcf_idr_cleanup(tn, parm->index);
> +			return err;
> +		}
> +		res = ACT_P_CREATED;
> +	} else {
> +		if (bind)
> +			return 0;
> +
> +		if (!replace) {
> +			tcf_idr_release(*a, bind);
> +			return -EEXIST;
> +		}
> +	}
> +	err = tcf_action_check_ctrlact(parm->action, tp, &goto_ch, extack);
> +	if (err < 0)
> +		goto cleanup;
> +
> +	c = to_ct(*a);
> +
> +	params = kzalloc(sizeof(*params), GFP_KERNEL);
> +	if (unlikely(!params)) {
> +		err = -ENOMEM;
> +		goto cleanup;
> +	}
> +
> +	err = tcf_ct_fill_params(net, params, parm, tb, extack);
> +	if (err)
> +		goto cleanup;
> +
> +	spin_lock_bh(&c->tcf_lock);
> +	goto_ch = tcf_action_set_ctrlact(*a, parm->action, goto_ch);
> +	rcu_swap_protected(c->params, params, lockdep_is_held(&c->tcf_lock));
> +	spin_unlock_bh(&c->tcf_lock);
> +
> +	if (goto_ch)
> +		tcf_chain_put_by_act(goto_ch);
> +	if (params)
> +		kfree_rcu(params, rcu);
> +	if (res == ACT_P_CREATED)
> +		tcf_idr_insert(tn, *a);
> +
> +	return res;
> +
> +cleanup:
> +	if (goto_ch)
> +		tcf_chain_put_by_act(goto_ch);
> +	kfree(params);
> +	tcf_idr_release(*a, bind);
> +	return err;
> +}
> +
> +static void tcf_ct_cleanup(struct tc_action *a)
> +{
> +	struct tcf_ct_params *params;
> +	struct tcf_ct *c = to_ct(a);
> +
> +	params = rcu_dereference_protected(c->params, 1);
> +	if (params)
> +		call_rcu(&params->rcu, tcf_ct_params_free);
> +}
> +
> +static int tcf_ct_dump_key_val(struct sk_buff *skb,
> +			       void *val, int val_type,
> +			       void *mask, int mask_type,
> +			       int len)
> +{
> +	int err;
> +
> +	if (mask && !memchr_inv(mask, 0, len))
> +		return 0;
> +
> +	err = nla_put(skb, val_type, len, val);
> +	if (err)
> +		return err;
> +
> +	if (mask_type != TCA_CT_UNSPEC) {
> +		err = nla_put(skb, mask_type, len, mask);
> +		if (err)
> +			return err;
> +	}
> +
> +	return 0;
> +}
> +
> +static int tcf_ct_dump_nat(struct sk_buff *skb, struct tcf_ct_params *p)
> +{
> +	struct nf_nat_range2 *range = &p->range;
> +
> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
> +		return 0;
> +
> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
> +		return 0;
> +
> +	if (range->flags & NF_NAT_RANGE_MAP_IPS) {
> +		if (p->ipv4_range) {
> +			if (nla_put_in_addr(skb, TCA_CT_NAT_IPV4_MIN,
> +					    range->min_addr.ip))
> +				return -1;
> +			if (nla_put_in_addr(skb, TCA_CT_NAT_IPV4_MAX,
> +					    range->max_addr.ip))
> +				return -1;
> +		} else {
> +			if (nla_put_in6_addr(skb, TCA_CT_NAT_IPV6_MIN,
> +					     &range->min_addr.in6))
> +				return -1;
> +			if (nla_put_in6_addr(skb, TCA_CT_NAT_IPV6_MAX,
> +					     &range->max_addr.in6))
> +				return -1;
> +		}
> +	}
> +
> +	if (range->flags & NF_NAT_RANGE_PROTO_SPECIFIED) {
> +		if (nla_put_u16(skb, TCA_CT_NAT_PORT_MIN,
> +				ntohs(range->min_proto.all)))
> +			return -1;
> +		if (nla_put_u16(skb, TCA_CT_NAT_PORT_MAX,
> +				ntohs(range->max_proto.all)))
> +			return -1;
> +	}
> +
> +	return 0;
> +}
> +
> +static inline int tcf_ct_dump(struct sk_buff *skb, struct tc_action *a,
> +			      int bind, int ref)
> +{
> +	unsigned char *b = skb_tail_pointer(skb);
> +	struct tcf_ct *c = to_ct(a);
> +	struct tcf_ct_params *p;
> +
> +	struct tc_ct opt = {
> +		.index   = c->tcf_index,
> +		.refcnt  = refcount_read(&c->tcf_refcnt) - ref,
> +		.bindcnt = atomic_read(&c->tcf_bindcnt) - bind,
> +	};
> +	struct tcf_t t;
> +
> +	spin_lock_bh(&c->tcf_lock);
> +	p = rcu_dereference_protected(c->params,
> +				      lockdep_is_held(&c->tcf_lock));
> +	opt.action = c->tcf_action;
> +
> +	if (tcf_ct_dump_key_val(skb,
> +				&p->ct_action, TCA_CT_ACTION,
> +				NULL, TCA_CT_UNSPEC,
> +				sizeof(p->ct_action)))
> +		goto nla_put_failure;
> +
> +	if (p->ct_action & TCA_CT_ACT_CLEAR)
> +		goto skip_dump;
> +
> +	if (IS_ENABLED(CONFIG_NF_CONNTRACK_MARK) &&
> +	    tcf_ct_dump_key_val(skb,
> +				&p->mark, TCA_CT_MARK,
> +				&p->mark_mask, TCA_CT_MARK_MASK,
> +				sizeof(p->mark)))
> +		goto nla_put_failure;
> +
> +	if (IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS) &&
> +	    tcf_ct_dump_key_val(skb,
> +				p->labels, TCA_CT_LABELS,
> +				p->labels_mask, TCA_CT_LABELS_MASK,
> +				sizeof(p->labels)))
> +		goto nla_put_failure;
> +
> +	if (IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES) &&
> +	    tcf_ct_dump_key_val(skb,
> +				&p->zone, TCA_CT_ZONE,
> +				NULL, TCA_CT_UNSPEC,
> +				sizeof(p->zone)))
> +		goto nla_put_failure;
> +
> +	if (tcf_ct_dump_nat(skb, p))
> +		goto nla_put_failure;
> +
> +skip_dump:
> +	if (nla_put(skb, TCA_CT_PARMS, sizeof(opt), &opt))
> +		goto nla_put_failure;
> +
> +	tcf_tm_dump(&t, &c->tcf_tm);
> +	if (nla_put_64bit(skb, TCA_CT_TM, sizeof(t), &t, TCA_CT_PAD))
> +		goto nla_put_failure;
> +	spin_unlock_bh(&c->tcf_lock);
> +
> +	return skb->len;
> +nla_put_failure:
> +	spin_unlock_bh(&c->tcf_lock);
> +	nlmsg_trim(skb, b);
> +	return -1;
> +}
> +
> +static int tcf_ct_walker(struct net *net, struct sk_buff *skb,
> +			 struct netlink_callback *cb, int type,
> +			 const struct tc_action_ops *ops,
> +			 struct netlink_ext_ack *extack)
> +{
> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
> +
> +	return tcf_generic_walker(tn, skb, cb, type, ops, extack);
> +}
> +
> +static int tcf_ct_search(struct net *net, struct tc_action **a, u32 index)
> +{
> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
> +
> +	return tcf_idr_search(tn, a, index);
> +}
> +
> +static void tcf_stats_update(struct tc_action *a, u64 bytes, u32 packets,
> +			     u64 lastuse, bool hw)
> +{
> +	struct tcf_ct *c = to_ct(a);
> +
> +	_bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), bytes, packets);
> +
> +	if (hw)
> +		_bstats_cpu_update(this_cpu_ptr(a->cpu_bstats_hw),
> +				   bytes, packets);
> +	c->tcf_tm.lastuse = max_t(u64, c->tcf_tm.lastuse, lastuse);
> +}
> +
> +static struct tc_action_ops act_ct_ops = {
> +	.kind		=	"ct",
> +	.id		=	TCA_ID_CT,
> +	.owner		=	THIS_MODULE,
> +	.act		=	tcf_ct_act,
> +	.dump		=	tcf_ct_dump,
> +	.init		=	tcf_ct_init,
> +	.cleanup	=	tcf_ct_cleanup,
> +	.walk		=	tcf_ct_walker,
> +	.lookup		=	tcf_ct_search,
> +	.stats_update	=	tcf_stats_update,
> +	.size		=	sizeof(struct tcf_ct),
> +};
> +
> +static __net_init int ct_init_net(struct net *net)
> +{
> +	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
> +	unsigned int n_bits = FIELD_SIZEOF(struct tcf_ct_params, labels) * 8;
> +
> +	if (nf_connlabels_get(net, n_bits - 1)) {
> +		tn->labels = false;
> +		pr_err("act_ct: Failed to set connlabels length");
> +	} else {
> +		tn->labels = true;
> +	}
> +
> +	return tc_action_net_init(&tn->tn, &act_ct_ops);
> +}
> +
> +static void __net_exit ct_exit_net(struct list_head *net_list)
> +{
> +	struct net *net;
> +
> +	rtnl_lock();
> +	list_for_each_entry(net, net_list, exit_list) {
> +		struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
> +
> +		if (tn->labels)
> +			nf_connlabels_put(net);
> +	}
> +	rtnl_unlock();
> +
> +	tc_action_net_exit(net_list, ct_net_id);
> +}
> +
> +static struct pernet_operations ct_net_ops = {
> +	.init = ct_init_net,
> +	.exit_batch = ct_exit_net,
> +	.id   = &ct_net_id,
> +	.size = sizeof(struct tc_ct_action_net),
> +};
> +
> +static int __init ct_init_module(void)
> +{
> +	return tcf_register_action(&act_ct_ops, &ct_net_ops);
> +}
> +
> +static void __exit ct_cleanup_module(void)
> +{
> +	tcf_unregister_action(&act_ct_ops, &ct_net_ops);
> +}
> +
> +module_init(ct_init_module);
> +module_exit(ct_cleanup_module);
> +MODULE_AUTHOR("Paul Blakey <paulb@mellanox.com>");
> +MODULE_AUTHOR("Yossi Kuperman <yossiku@mellanox.com>");
> +MODULE_AUTHOR("Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>");
> +MODULE_DESCRIPTION("Connection tracking action");
> +MODULE_LICENSE("GPL v2");
> +
> diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
> index ad36bbc..4a7331c 100644
> --- a/net/sched/cls_api.c
> +++ b/net/sched/cls_api.c
> @@ -35,6 +35,7 @@
>  #include <net/tc_act/tc_police.h>
>  #include <net/tc_act/tc_sample.h>
>  #include <net/tc_act/tc_skbedit.h>
> +#include <net/tc_act/tc_ct.h>
>  
>  extern const struct nla_policy rtm_tca_policy[TCA_MAX + 1];
>  
> @@ -3266,6 +3267,10 @@ int tc_setup_flow_action(struct flow_action *flow_action,
>  			entry->police.burst = tcf_police_tcfp_burst(act);
>  			entry->police.rate_bytes_ps =
>  				tcf_police_rate_bytes_ps(act);
> +		} else if (is_tcf_ct(act)) {
> +			entry->id = FLOW_ACTION_CT;
> +			entry->ct.action = tcf_ct_action(act);
> +			entry->ct.zone = tcf_ct_zone(act);
>  		} else {
>  			goto err_out;
>  		}
> -- 
> 1.8.3.1

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:59   ` Toke Høiland-Jørgensen
@ 2019-06-11 14:15     ` Paul Blakey
  2019-06-11 15:34       ` Toke Høiland-Jørgensen
  2019-06-13 11:00       ` Kevin Darbyshire-Bryant
  0 siblings, 2 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-11 14:15 UTC (permalink / raw)
  To: Toke Høiland-Jørgensen, Jiri Pirko, Roi Dayan,
	Yossi Kuperman, Oz Shlomo, Marcelo Ricardo Leitner, netdev,
	David Miller, Aaron Conole, Zhike Wang
  Cc: Rony Efraim, nst-kernel, John Hurley, Simon Horman,
	Justin Pettit, Kevin Darbyshire-Bryant


On 6/11/2019 4:59 PM, Toke Høiland-Jørgensen wrote:
> Paul Blakey <paulb@mellanox.com> writes:
>
>> Allow sending a packet to conntrack and set conntrack zone, mark,
>> labels and nat parameters.
> How is this different from the newly merged ctinfo action?
>
> -Toke

Hi,

ctinfo does one of two very specific things,

1) copies DSCP values that have been placed in the firewall conntrack 
mark back into the IPv4/v6 diffserv field

2) copies the firewall conntrack mark to the skb's mark field (like 
act_connmark)

Originally ctinfo action was named conndscp (then conntrack, which is 
what our ct shorthand stands for).

We also talked about merging both at some point, but they seem only 
coincidentally related.

don't know how it was then be agreed to be named ctinfo suggesting it 
does something else but the above.

This action sends packets to conntrack, configures nat, and doesn't get 
"info" from conntrack, while the ctinfo already expects packets to be 
passed conntrack

by some other kernel mechanism.



>
>> Signed-off-by: Paul Blakey <paulb@mellanox.com>
>> Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
>> Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
>> Acked-by: Jiri Pirko <jiri@mellanox.com>
>> ---
>>   include/net/flow_offload.h        |   5 +
>>   include/net/tc_act/tc_ct.h        |  64 +++
>>   include/uapi/linux/pkt_cls.h      |   2 +
>>   include/uapi/linux/tc_act/tc_ct.h |  41 ++
>>   net/sched/Kconfig                 |  11 +
>>   net/sched/Makefile                |   1 +
>>   net/sched/act_ct.c                | 901 ++++++++++++++++++++++++++++++++++++++
>>   net/sched/cls_api.c               |   5 +
>>   8 files changed, 1030 insertions(+)
>>   create mode 100644 include/net/tc_act/tc_ct.h
>>   create mode 100644 include/uapi/linux/tc_act/tc_ct.h
>>   create mode 100644 net/sched/act_ct.c
>>
>> diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
>> index 36fdb85..5b2c4fa 100644
>> --- a/include/net/flow_offload.h
>> +++ b/include/net/flow_offload.h
>> @@ -123,6 +123,7 @@ enum flow_action_id {
>>   	FLOW_ACTION_QUEUE,
>>   	FLOW_ACTION_SAMPLE,
>>   	FLOW_ACTION_POLICE,
>> +	FLOW_ACTION_CT,
>>   };
>>   
>>   /* This is mirroring enum pedit_header_type definition for easy mapping between
>> @@ -172,6 +173,10 @@ struct flow_action_entry {
>>   			s64			burst;
>>   			u64			rate_bytes_ps;
>>   		} police;
>> +		struct {				/* FLOW_ACTION_CT */
>> +			int action;
>> +			u16 zone;
>> +		} ct;
>>   	};
>>   };
>>   
>> diff --git a/include/net/tc_act/tc_ct.h b/include/net/tc_act/tc_ct.h
>> new file mode 100644
>> index 0000000..59e4f5e
>> --- /dev/null
>> +++ b/include/net/tc_act/tc_ct.h
>> @@ -0,0 +1,64 @@
>> +/* SPDX-License-Identifier: GPL-2.0 */
>> +#ifndef __NET_TC_CT_H
>> +#define __NET_TC_CT_H
>> +
>> +#include <net/act_api.h>
>> +#include <uapi/linux/tc_act/tc_ct.h>
>> +
>> +#if IS_ENABLED(CONFIG_NF_CONNTRACK)
>> +#include <net/netfilter/nf_nat.h>
>> +#include <net/netfilter/nf_conntrack_labels.h>
>> +
>> +struct tcf_ct_params {
>> +	struct nf_conn *tmpl;
>> +	u16 zone;
>> +
>> +	u32 mark;
>> +	u32 mark_mask;
>> +
>> +	u32 labels[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
>> +	u32 labels_mask[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
>> +
>> +	struct nf_nat_range2 range;
>> +	bool ipv4_range;
>> +
>> +	u16 ct_action;
>> +
>> +	struct rcu_head rcu;
>> +
>> +};
>> +
>> +struct tcf_ct {
>> +	struct tc_action common;
>> +	struct tcf_ct_params __rcu *params;
>> +};
>> +
>> +#define to_ct(a) ((struct tcf_ct *)a)
>> +#define to_ct_params(a) ((struct tcf_ct_params *) \
>> +			 rtnl_dereference((to_ct(a)->params)))
>> +
>> +static inline uint16_t tcf_ct_zone(const struct tc_action *a)
>> +{
>> +	return to_ct_params(a)->zone;
>> +}
>> +
>> +static inline int tcf_ct_action(const struct tc_action *a)
>> +{
>> +	return to_ct_params(a)->ct_action;
>> +}
>> +
>> +#else
>> +static inline uint16_t tcf_ct_zone(const struct tc_action *a) { return 0; }
>> +static inline int tcf_ct_action(const struct tc_action *a) { return 0; }
>> +#endif /* CONFIG_NF_CONNTRACK */
>> +
>> +static inline bool is_tcf_ct(const struct tc_action *a)
>> +{
>> +#if defined(CONFIG_NET_CLS_ACT) && IS_ENABLED(CONFIG_NF_CONNTRACK)
>> +	if (a->ops && a->ops->id == TCA_ID_CT)
>> +		return true;
>> +#endif
>> +	return false;
>> +}
>> +
>> +#endif /* __NET_TC_CT_H */
>> diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
>> index a93680f..c5264d7 100644
>> --- a/include/uapi/linux/pkt_cls.h
>> +++ b/include/uapi/linux/pkt_cls.h
>> @@ -83,6 +83,7 @@ enum {
>>   #define TCA_ACT_SIMP 22
>>   #define TCA_ACT_IFE 25
>>   #define TCA_ACT_SAMPLE 26
>> +#define TCA_ACT_CT 27
>>   
>>   /* Action type identifiers*/
>>   enum tca_id {
>> @@ -106,6 +107,7 @@ enum tca_id {
>>   	TCA_ID_SAMPLE = TCA_ACT_SAMPLE,
>>   	/* other actions go here */
>>   	TCA_ID_CTINFO,
>> +	TCA_ID_CT,
>>   	__TCA_ID_MAX = 255
>>   };
>>   
>> diff --git a/include/uapi/linux/tc_act/tc_ct.h b/include/uapi/linux/tc_act/tc_ct.h
>> new file mode 100644
>> index 0000000..5fb1d7a
>> --- /dev/null
>> +++ b/include/uapi/linux/tc_act/tc_ct.h
>> @@ -0,0 +1,41 @@
>> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
>> +#ifndef __UAPI_TC_CT_H
>> +#define __UAPI_TC_CT_H
>> +
>> +#include <linux/types.h>
>> +#include <linux/pkt_cls.h>
>> +
>> +enum {
>> +	TCA_CT_UNSPEC,
>> +	TCA_CT_PARMS,
>> +	TCA_CT_TM,
>> +	TCA_CT_ACTION,		/* u16 */
>> +	TCA_CT_ZONE,		/* u16 */
>> +	TCA_CT_MARK,		/* u32 */
>> +	TCA_CT_MARK_MASK,	/* u32 */
>> +	TCA_CT_LABELS,		/* u128 */
>> +	TCA_CT_LABELS_MASK,	/* u128 */
>> +	TCA_CT_NAT_IPV4_MIN,	/* be32 */
>> +	TCA_CT_NAT_IPV4_MAX,	/* be32 */
>> +	TCA_CT_NAT_IPV6_MIN,	/* struct in6_addr */
>> +	TCA_CT_NAT_IPV6_MAX,	/* struct in6_addr */
>> +	TCA_CT_NAT_PORT_MIN,	/* be16 */
>> +	TCA_CT_NAT_PORT_MAX,	/* be16 */
>> +	TCA_CT_PAD,
>> +	__TCA_CT_MAX
>> +};
>> +
>> +#define TCA_CT_MAX (__TCA_CT_MAX - 1)
>> +
>> +#define TCA_CT_ACT_COMMIT	(1 << 0)
>> +#define TCA_CT_ACT_FORCE	(1 << 1)
>> +#define TCA_CT_ACT_CLEAR	(1 << 2)
>> +#define TCA_CT_ACT_NAT		(1 << 3)
>> +#define TCA_CT_ACT_NAT_SRC	(1 << 4)
>> +#define TCA_CT_ACT_NAT_DST	(1 << 5)
>> +
>> +struct tc_ct {
>> +	tc_gen;
>> +};
>> +
>> +#endif /* __UAPI_TC_CT_H */
>> diff --git a/net/sched/Kconfig b/net/sched/Kconfig
>> index d104f7e..0481a2a 100644
>> --- a/net/sched/Kconfig
>> +++ b/net/sched/Kconfig
>> @@ -929,6 +929,17 @@ config NET_ACT_TUNNEL_KEY
>>   	  To compile this code as a module, choose M here: the
>>   	  module will be called act_tunnel_key.
>>   
>> +config NET_ACT_CT
>> +        tristate "connection tracking tc action"
>> +        depends on NET_CLS_ACT && NF_CONNTRACK
>> +        help
>> +	  Say Y here to allow sending the packets to conntrack module.
>> +
>> +	  If unsure, say N.
>> +
>> +	  To compile this code as a module, choose M here: the
>> +	  module will be called act_ct.
>> +
>>   config NET_IFE_SKBMARK
>>           tristate "Support to encoding decoding skb mark on IFE action"
>>           depends on NET_ACT_IFE
>> diff --git a/net/sched/Makefile b/net/sched/Makefile
>> index d54bfcb..23d2202 100644
>> --- a/net/sched/Makefile
>> +++ b/net/sched/Makefile
>> @@ -28,6 +28,7 @@ obj-$(CONFIG_NET_IFE_SKBMARK)	+= act_meta_mark.o
>>   obj-$(CONFIG_NET_IFE_SKBPRIO)	+= act_meta_skbprio.o
>>   obj-$(CONFIG_NET_IFE_SKBTCINDEX)	+= act_meta_skbtcindex.o
>>   obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o
>> +obj-$(CONFIG_NET_ACT_CT)	+= act_ct.o
>>   obj-$(CONFIG_NET_SCH_FIFO)	+= sch_fifo.o
>>   obj-$(CONFIG_NET_SCH_CBQ)	+= sch_cbq.o
>>   obj-$(CONFIG_NET_SCH_HTB)	+= sch_htb.o
>> diff --git a/net/sched/act_ct.c b/net/sched/act_ct.c
>> new file mode 100644
>> index 0000000..4eb0dd7
>> --- /dev/null
>> +++ b/net/sched/act_ct.c
>> @@ -0,0 +1,901 @@
>> +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
>> +/* -
>> + * net/sched/act_ct.c  Connection Tracking action
>> + *
>> + * Authors:   Paul Blakey <paulb@mellanox.com>
>> + *            Yossi Kuperman <yossiku@mellanox.com>
>> + *            Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
>> + */
>> +
>> +#include <linux/module.h>
>> +#include <linux/init.h>
>> +#include <linux/kernel.h>
>> +#include <linux/skbuff.h>
>> +#include <linux/rtnetlink.h>
>> +#include <linux/pkt_cls.h>
>> +#include <linux/ip.h>
>> +#include <linux/ipv6.h>
>> +#include <net/netlink.h>
>> +#include <net/pkt_sched.h>
>> +#include <net/act_api.h>
>> +#include <uapi/linux/tc_act/tc_ct.h>
>> +#include <net/tc_act/tc_ct.h>
>> +
>> +#include <linux/netfilter/nf_nat.h>
>> +#include <net/netfilter/nf_conntrack.h>
>> +#include <net/netfilter/nf_conntrack_core.h>
>> +#include <net/netfilter/nf_conntrack_zones.h>
>> +#include <net/netfilter/nf_conntrack_helper.h>
>> +#include <net/pkt_cls.h>
>> +
>> +static struct tc_action_ops act_ct_ops;
>> +static unsigned int ct_net_id;
>> +
>> +struct tc_ct_action_net {
>> +	struct tc_action_net tn; /* Must be first */
>> +	bool labels;
>> +};
>> +
>> +/* Determine whether skb->_nfct is equal to the result of conntrack lookup. */
>> +static bool tcf_ct_skb_nfct_cached(struct net *net, struct sk_buff *skb,
>> +				   u16 zone_id, bool force)
>> +{
>> +	enum ip_conntrack_info ctinfo;
>> +	struct nf_conn *ct;
>> +
>> +	ct = nf_ct_get(skb, &ctinfo);
>> +	if (!ct)
>> +		return false;
>> +	if (!net_eq(net, read_pnet(&ct->ct_net)))
>> +		return false;
>> +	if (nf_ct_zone(ct)->id != zone_id)
>> +		return false;
>> +
>> +	/* Force conntrack entry direction. */
>> +	if (force && CTINFO2DIR(ctinfo) != IP_CT_DIR_ORIGINAL) {
>> +		nf_conntrack_put(&ct->ct_general);
>> +		nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
>> +
>> +		if (nf_ct_is_confirmed(ct))
>> +			nf_ct_kill(ct);
>> +
>> +		return false;
>> +	}
>> +
>> +	return true;
>> +}
>> +
>> +/* Trim the skb to the length specified by the IP/IPv6 header,
>> + * removing any trailing lower-layer padding. This prepares the skb
>> + * for higher-layer processing that assumes skb->len excludes padding
>> + * (such as nf_ip_checksum). The caller needs to pull the skb to the
>> + * network header, and ensure ip_hdr/ipv6_hdr points to valid data.
>> + */
>> +static int tcf_ct_skb_network_trim(struct sk_buff *skb)
>> +{
>> +	unsigned int len;
>> +	int err;
>> +
>> +	switch (skb->protocol) {
>> +	case htons(ETH_P_IP):
>> +		len = ntohs(ip_hdr(skb)->tot_len);
>> +		break;
>> +	case htons(ETH_P_IPV6):
>> +		len = sizeof(struct ipv6hdr)
>> +			+ ntohs(ipv6_hdr(skb)->payload_len);
>> +		break;
>> +	default:
>> +		len = skb->len;
>> +	}
>> +
>> +	err = pskb_trim_rcsum(skb, len);
>> +
>> +	return err;
>> +}
>> +
>> +static u8 tcf_ct_skb_nf_family(struct sk_buff *skb)
>> +{
>> +	u8 family = NFPROTO_UNSPEC;
>> +
>> +	switch (skb->protocol) {
>> +	case htons(ETH_P_IP):
>> +		family = NFPROTO_IPV4;
>> +		break;
>> +	case htons(ETH_P_IPV6):
>> +		family = NFPROTO_IPV6;
>> +		break;
>> +	default:
>> +	break;
>> +	}
>> +
>> +	return family;
>> +}
>> +
>> +static void tcf_ct_params_free(struct rcu_head *head)
>> +{
>> +	struct tcf_ct_params *params = container_of(head,
>> +						    struct tcf_ct_params, rcu);
>> +
>> +	if (params->tmpl)
>> +		nf_conntrack_put(&params->tmpl->ct_general);
>> +	kfree(params);
>> +}
>> +
>> +#if IS_ENABLED(CONFIG_NF_NAT)
>> +/* Modelled after nf_nat_ipv[46]_fn().
>> + * range is only used for new, uninitialized NAT state.
>> + * Returns either NF_ACCEPT or NF_DROP.
>> + */
>> +static int ct_nat_execute(struct sk_buff *skb, struct nf_conn *ct,
>> +			  enum ip_conntrack_info ctinfo,
>> +			  const struct nf_nat_range2 *range,
>> +			  enum nf_nat_manip_type maniptype)
>> +{
>> +	int hooknum, nh_off, err = NF_ACCEPT;
>> +
>> +	nh_off = skb_network_offset(skb);
>> +	skb_pull_rcsum(skb, nh_off);
>> +
>> +	/* See HOOK2MANIP(). */
>> +	if (maniptype == NF_NAT_MANIP_SRC)
>> +		hooknum = NF_INET_LOCAL_IN; /* Source NAT */
>> +	else
>> +		hooknum = NF_INET_LOCAL_OUT; /* Destination NAT */
>> +
>> +	switch (ctinfo) {
>> +	case IP_CT_RELATED:
>> +	case IP_CT_RELATED_REPLY:
>> +		if (skb->protocol == htons(ETH_P_IP) &&
>> +		    ip_hdr(skb)->protocol == IPPROTO_ICMP) {
>> +			if (!nf_nat_icmp_reply_translation(skb, ct, ctinfo,
>> +							   hooknum))
>> +				err = NF_DROP;
>> +			goto push;
>> +		} else if (IS_ENABLED(CONFIG_IPV6) &&
>> +			   skb->protocol == htons(ETH_P_IPV6)) {
>> +			__be16 frag_off;
>> +			u8 nexthdr = ipv6_hdr(skb)->nexthdr;
>> +			int hdrlen = ipv6_skip_exthdr(skb,
>> +						      sizeof(struct ipv6hdr),
>> +						      &nexthdr, &frag_off);
>> +
>> +			if (hdrlen >= 0 && nexthdr == IPPROTO_ICMPV6) {
>> +				if (!nf_nat_icmpv6_reply_translation(skb, ct,
>> +								     ctinfo,
>> +								     hooknum,
>> +								     hdrlen))
>> +					err = NF_DROP;
>> +				goto push;
>> +			}
>> +		}
>> +		/* Non-ICMP, fall thru to initialize if needed. */
>> +		/* fall through */
>> +	case IP_CT_NEW:
>> +		/* Seen it before?  This can happen for loopback, retrans,
>> +		 * or local packets.
>> +		 */
>> +		if (!nf_nat_initialized(ct, maniptype)) {
>> +			/* Initialize according to the NAT action. */
>> +			err = (range && range->flags & NF_NAT_RANGE_MAP_IPS)
>> +				/* Action is set up to establish a new
>> +				 * mapping.
>> +				 */
>> +				? nf_nat_setup_info(ct, range, maniptype)
>> +				: nf_nat_alloc_null_binding(ct, hooknum);
>> +			if (err != NF_ACCEPT)
>> +				goto push;
>> +		}
>> +		break;
>> +
>> +	case IP_CT_ESTABLISHED:
>> +	case IP_CT_ESTABLISHED_REPLY:
>> +		break;
>> +
>> +	default:
>> +		err = NF_DROP;
>> +		goto push;
>> +	}
>> +
>> +	err = nf_nat_packet(ct, ctinfo, hooknum, skb);
>> +push:
>> +	skb_push(skb, nh_off);
>> +	skb_postpush_rcsum(skb, skb->data, nh_off);
>> +
>> +	return err;
>> +}
>> +#endif /* CONFIG_NF_NAT */
>> +
>> +static void tcf_ct_act_set_mark(struct nf_conn *ct, u32 mark, u32 mask)
>> +{
>> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)
>> +	u32 new_mark;
>> +
>> +	if (!mask)
>> +		return;
>> +
>> +	new_mark = mark | (ct->mark & ~(mask));
>> +	if (ct->mark != new_mark) {
>> +		ct->mark = new_mark;
>> +		if (nf_ct_is_confirmed(ct))
>> +			nf_conntrack_event_cache(IPCT_MARK, ct);
>> +	}
>> +#endif
>> +}
>> +
>> +static void tcf_ct_act_set_labels(struct nf_conn *ct,
>> +				  u32 *labels,
>> +				  u32 *labels_m)
>> +{
>> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)
>> +	size_t labels_sz = FIELD_SIZEOF(struct tcf_ct_params, labels);
>> +
>> +	if (!memchr_inv(labels_m, 0, labels_sz))
>> +		return;
>> +
>> +	nf_connlabels_replace(ct, labels, labels_m, 4);
>> +#endif
>> +}
>> +
>> +static int tcf_ct_act_nat(struct sk_buff *skb,
>> +			  struct nf_conn *ct,
>> +			  enum ip_conntrack_info ctinfo,
>> +			  int ct_action,
>> +			  struct nf_nat_range2 *range,
>> +			  bool commit)
>> +{
>> +#if IS_ENABLED(CONFIG_NF_NAT)
>> +	enum nf_nat_manip_type maniptype;
>> +
>> +	if (!(ct_action & TCA_CT_ACT_NAT))
>> +		return NF_ACCEPT;
>> +
>> +	/* Add NAT extension if not confirmed yet. */
>> +	if (!nf_ct_is_confirmed(ct) && !nf_ct_nat_ext_add(ct))
>> +		return NF_DROP;   /* Can't NAT. */
>> +
>> +	if (ctinfo != IP_CT_NEW && (ct->status & IPS_NAT_MASK) &&
>> +	    (ctinfo != IP_CT_RELATED || commit)) {
>> +		/* NAT an established or related connection like before. */
>> +		if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY)
>> +			/* This is the REPLY direction for a connection
>> +			 * for which NAT was applied in the forward
>> +			 * direction.  Do the reverse NAT.
>> +			 */
>> +			maniptype = ct->status & IPS_SRC_NAT
>> +				? NF_NAT_MANIP_DST : NF_NAT_MANIP_SRC;
>> +		else
>> +			maniptype = ct->status & IPS_SRC_NAT
>> +				? NF_NAT_MANIP_SRC : NF_NAT_MANIP_DST;
>> +	} else if (ct_action & TCA_CT_ACT_NAT_SRC) {
>> +		maniptype = NF_NAT_MANIP_SRC;
>> +	} else if (ct_action & TCA_CT_ACT_NAT_DST) {
>> +		maniptype = NF_NAT_MANIP_DST;
>> +	} else {
>> +		return NF_ACCEPT;
>> +	}
>> +
>> +	return ct_nat_execute(skb, ct, ctinfo, range, maniptype);
>> +#else
>> +	return NF_ACCEPT;
>> +#endif
>> +}
>> +
>> +static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
>> +		      struct tcf_result *res)
>> +{
>> +	struct net *net = dev_net(skb->dev);
>> +	bool cached, commit, clear, force;
>> +	enum ip_conntrack_info ctinfo;
>> +	struct tcf_ct *c = to_ct(a);
>> +	struct nf_conn *tmpl = NULL;
>> +	struct nf_hook_state state;
>> +	struct tcf_ct_params *p;
>> +	struct nf_conn *ct;
>> +	int nh_ofs, err;
>> +	u8 family;
>> +
>> +	p = rcu_dereference_bh(c->params);
>> +
>> +	commit = p->ct_action & TCA_CT_ACT_COMMIT;
>> +	clear = p->ct_action & TCA_CT_ACT_CLEAR;
>> +	force = p->ct_action & TCA_CT_ACT_FORCE;
>> +	tmpl = p->tmpl;
>> +
>> +	if (clear) {
>> +		ct = nf_ct_get(skb, &ctinfo);
>> +		if (ct) {
>> +			nf_conntrack_put(&ct->ct_general);
>> +			nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
>> +		}
>> +
>> +		goto out;
>> +	}
>> +
>> +	/* The conntrack module expects to be working at L3. */
>> +	nh_ofs = skb_network_offset(skb);
>> +	skb_pull_rcsum(skb, nh_ofs);
>> +
>> +	err = tcf_ct_skb_network_trim(skb);
>> +	if (err)
>> +		goto drop;
>> +
>> +	family = tcf_ct_skb_nf_family(skb);
>> +	if (family == NFPROTO_UNSPEC)
>> +		goto drop;
>> +
>> +	state.hook = NF_INET_PRE_ROUTING;
>> +	state.net = net;
>> +	state.pf = family;
>> +
>> +	/* If we are recirculating packets to match on ct fields and
>> +	 * committing with a separate ct action, then we don't need to
>> +	 * actually run the packet through conntrack twice unless it's for a
>> +	 * different zone.
>> +	 */
>> +	cached = tcf_ct_skb_nfct_cached(net, skb, p->zone, force);
>> +	if (!cached) {
>> +		/* Associate skb with specified zone. */
>> +		if (tmpl) {
>> +			if (skb_nfct(skb))
>> +				nf_conntrack_put(skb_nfct(skb));
>> +			nf_conntrack_get(&tmpl->ct_general);
>> +			nf_ct_set(skb, tmpl, IP_CT_NEW);
>> +		}
>> +
>> +		err = nf_conntrack_in(skb, &state);
>> +		if (err != NF_ACCEPT)
>> +			goto out_push;
>> +	}
>> +
>> +	ct = nf_ct_get(skb, &ctinfo);
>> +	if (!ct)
>> +		goto out_push;
>> +	nf_ct_deliver_cached_events(ct);
>> +
>> +	err = tcf_ct_act_nat(skb, ct, ctinfo, p->ct_action, &p->range, commit);
>> +	if (err != NF_ACCEPT)
>> +		goto drop;
>> +
>> +	if (commit) {
>> +		tcf_ct_act_set_mark(ct, p->mark, p->mark_mask);
>> +		tcf_ct_act_set_labels(ct, p->labels, p->labels_mask);
>> +
>> +		/* This will take care of sending queued events
>> +		 * even if the connection is already confirmed.
>> +		 */
>> +		nf_conntrack_confirm(skb);
>> +	}
>> +
>> +out_push:
>> +	skb_push(skb, nh_ofs);
>> +	skb_postpush_rcsum(skb, skb->data, nh_ofs);
>> +
>> +out:
>> +	bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), skb);
>> +
>> +	return c->tcf_action;
>> +
>> +drop:
>> +	qstats_drop_inc(this_cpu_ptr(a->cpu_qstats));
>> +	return TC_ACT_SHOT;
>> +}
>> +
>> +static const struct nla_policy ct_policy[TCA_CT_MAX + 1] = {
>> +	[TCA_CT_ACTION] = { .type = NLA_U16 },
>> +	[TCA_CT_PARMS] = { .type = NLA_EXACT_LEN, .len = sizeof(struct tc_ct) },
>> +	[TCA_CT_ZONE] = { .type = NLA_U16 },
>> +	[TCA_CT_MARK] = { .type = NLA_U32 },
>> +	[TCA_CT_MARK_MASK] = { .type = NLA_U32 },
>> +	[TCA_CT_LABELS] = { .type = NLA_BINARY,
>> +			    .len = 128 / BITS_PER_BYTE },
>> +	[TCA_CT_LABELS_MASK] = { .type = NLA_BINARY,
>> +				 .len = 128 / BITS_PER_BYTE },
>> +	[TCA_CT_NAT_IPV4_MIN] = { .type = NLA_U32 },
>> +	[TCA_CT_NAT_IPV4_MAX] = { .type = NLA_U32 },
>> +	[TCA_CT_NAT_IPV6_MIN] = { .type = NLA_EXACT_LEN,
>> +				  .len = sizeof(struct in6_addr) },
>> +	[TCA_CT_NAT_IPV6_MAX] = { .type = NLA_EXACT_LEN,
>> +				   .len = sizeof(struct in6_addr) },
>> +	[TCA_CT_NAT_PORT_MIN] = { .type = NLA_U16 },
>> +	[TCA_CT_NAT_PORT_MAX] = { .type = NLA_U16 },
>> +};
>> +
>> +static int tcf_ct_fill_params_nat(struct tcf_ct_params *p,
>> +				  struct tc_ct *parm,
>> +				  struct nlattr **tb,
>> +				  struct netlink_ext_ack *extack)
>> +{
>> +	struct nf_nat_range2 *range;
>> +
>> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
>> +		return 0;
>> +
>> +	if (!IS_ENABLED(CONFIG_NF_NAT)) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Netfilter nat isn't enabled in kernel");
>> +		return -EOPNOTSUPP;
>> +	}
>> +
>> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
>> +		return 0;
>> +
>> +	if ((p->ct_action & TCA_CT_ACT_NAT_SRC) &&
>> +	    (p->ct_action & TCA_CT_ACT_NAT_DST)) {
>> +		NL_SET_ERR_MSG_MOD(extack, "dnat and snat can't be enabled at the same time");
>> +		return -EOPNOTSUPP;
>> +	}
>> +
>> +	range = &p->range;
>> +	if (tb[TCA_CT_NAT_IPV4_MIN]) {
>> +		range->min_addr.ip =
>> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = true;
>> +	}
>> +	if (tb[TCA_CT_NAT_IPV4_MAX]) {
>> +		range->max_addr.ip =
>> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = true;
>> +	} else if (range->min_addr.ip) {
>> +		range->max_addr.ip = range->min_addr.ip;
>> +	}
>> +
>> +	if (tb[TCA_CT_NAT_IPV6_MIN]) {
>> +		range->min_addr.in6 =
>> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = false;
>> +	}
>> +	if (tb[TCA_CT_NAT_IPV6_MAX]) {
>> +		range->max_addr.in6 =
>> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = false;
>> +	} else if (memchr_inv(&range->min_addr.in6, 0,
>> +		   sizeof(range->min_addr.in6))) {
>> +		range->max_addr.in6 = range->min_addr.in6;
>> +	}
>> +
>> +	if (tb[TCA_CT_NAT_PORT_MIN]) {
>> +		range->min_proto.all =
>> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MIN]));
>> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
>> +	}
>> +	if (tb[TCA_CT_NAT_PORT_MAX]) {
>> +		range->max_proto.all =
>> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MAX]));
>> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
>> +	} else if (range->min_proto.all) {
>> +		range->max_proto.all = range->min_proto.all;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static void tcf_ct_set_key_val(struct nlattr **tb,
>> +			       void *val, int val_type,
>> +			       void *mask, int mask_type,
>> +			       int len)
>> +{
>> +	if (!tb[val_type])
>> +		return;
>> +	nla_memcpy(val, tb[val_type], len);
>> +
>> +	if (!mask)
>> +		return;
>> +
>> +	if (mask_type == TCA_CT_UNSPEC || !tb[mask_type])
>> +		memset(mask, 0xff, len);
>> +	else
>> +		nla_memcpy(mask, tb[mask_type], len);
>> +}
>> +
>> +static int tcf_ct_fill_params(struct net *net,
>> +			      struct tcf_ct_params *p,
>> +			      struct tc_ct *parm,
>> +			      struct nlattr **tb,
>> +			      struct netlink_ext_ack *extack)
>> +{
>> +	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
>> +	struct nf_conntrack_zone zone;
>> +	struct nf_conn *tmpl;
>> +	int err;
>> +
>> +	p->zone = NF_CT_DEFAULT_ZONE_ID;
>> +
>> +	tcf_ct_set_key_val(tb,
>> +			   &p->ct_action, TCA_CT_ACTION,
>> +			   NULL, TCA_CT_UNSPEC,
>> +			   sizeof(p->ct_action));
>> +
>> +	if (p->ct_action & TCA_CT_ACT_CLEAR)
>> +		return 0;
>> +
>> +	err = tcf_ct_fill_params_nat(p, parm, tb, extack);
>> +	if (err)
>> +		return err;
>> +
>> +	if (tb[TCA_CT_MARK]) {
>> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack mark isn't enabled.");
>> +			return -EOPNOTSUPP;
>> +		}
>> +		tcf_ct_set_key_val(tb,
>> +				   &p->mark, TCA_CT_MARK,
>> +				   &p->mark_mask, TCA_CT_MARK_MASK,
>> +				   sizeof(p->mark));
>> +	}
>> +
>> +	if (tb[TCA_CT_LABELS]) {
>> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack labels isn't enabled.");
>> +			return -EOPNOTSUPP;
>> +		}
>> +
>> +		if (!tn->labels) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Failed to set connlabel length");
>> +			return -EOPNOTSUPP;
>> +		}
>> +		tcf_ct_set_key_val(tb,
>> +				   p->labels, TCA_CT_LABELS,
>> +				   p->labels_mask, TCA_CT_LABELS_MASK,
>> +				   sizeof(p->labels));
>> +	}
>> +
>> +	if (tb[TCA_CT_ZONE]) {
>> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack zones isn't enabled.");
>> +			return -EOPNOTSUPP;
>> +		}
>> +
>> +		tcf_ct_set_key_val(tb,
>> +				   &p->zone, TCA_CT_ZONE,
>> +				   NULL, TCA_CT_UNSPEC,
>> +				   sizeof(p->zone));
>> +	}
>> +
>> +	if (p->zone == NF_CT_DEFAULT_ZONE_ID)
>> +		return 0;
>> +
>> +	nf_ct_zone_init(&zone, p->zone, NF_CT_DEFAULT_ZONE_DIR, 0);
>> +	tmpl = nf_ct_tmpl_alloc(net, &zone, GFP_KERNEL);
>> +	if (!tmpl) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Failed to allocate conntrack template");
>> +		return -ENOMEM;
>> +	}
>> +	__set_bit(IPS_CONFIRMED_BIT, &tmpl->status);
>> +	nf_conntrack_get(&tmpl->ct_general);
>> +	p->tmpl = tmpl;
>> +
>> +	return 0;
>> +}
>> +
>> +static int tcf_ct_init(struct net *net, struct nlattr *nla,
>> +		       struct nlattr *est, struct tc_action **a,
>> +		       int replace, int bind, bool rtnl_held,
>> +		       struct tcf_proto *tp,
>> +		       struct netlink_ext_ack *extack)
>> +{
>> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
>> +	struct tcf_ct_params *params = NULL;
>> +	struct nlattr *tb[TCA_CT_MAX + 1];
>> +	struct tcf_chain *goto_ch = NULL;
>> +	struct tc_ct *parm;
>> +	struct tcf_ct *c;
>> +	int err, res = 0;
>> +
>> +	if (!nla) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Ct requires attributes to be passed");
>> +		return -EINVAL;
>> +	}
>> +
>> +	err = nla_parse_nested(tb, TCA_CT_MAX, nla, ct_policy, extack);
>> +	if (err < 0)
>> +		return err;
>> +
>> +	if (!tb[TCA_CT_PARMS]) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Missing required ct parameters");
>> +		return -EINVAL;
>> +	}
>> +	parm = nla_data(tb[TCA_CT_PARMS]);
>> +
>> +	err = tcf_idr_check_alloc(tn, &parm->index, a, bind);
>> +	if (err < 0)
>> +		return err;
>> +
>> +	if (!err) {
>> +		err = tcf_idr_create(tn, parm->index, est, a,
>> +				     &act_ct_ops, bind, true);
>> +		if (err) {
>> +			tcf_idr_cleanup(tn, parm->index);
>> +			return err;
>> +		}
>> +		res = ACT_P_CREATED;
>> +	} else {
>> +		if (bind)
>> +			return 0;
>> +
>> +		if (!replace) {
>> +			tcf_idr_release(*a, bind);
>> +			return -EEXIST;
>> +		}
>> +	}
>> +	err = tcf_action_check_ctrlact(parm->action, tp, &goto_ch, extack);
>> +	if (err < 0)
>> +		goto cleanup;
>> +
>> +	c = to_ct(*a);
>> +
>> +	params = kzalloc(sizeof(*params), GFP_KERNEL);
>> +	if (unlikely(!params)) {
>> +		err = -ENOMEM;
>> +		goto cleanup;
>> +	}
>> +
>> +	err = tcf_ct_fill_params(net, params, parm, tb, extack);
>> +	if (err)
>> +		goto cleanup;
>> +
>> +	spin_lock_bh(&c->tcf_lock);
>> +	goto_ch = tcf_action_set_ctrlact(*a, parm->action, goto_ch);
>> +	rcu_swap_protected(c->params, params, lockdep_is_held(&c->tcf_lock));
>> +	spin_unlock_bh(&c->tcf_lock);
>> +
>> +	if (goto_ch)
>> +		tcf_chain_put_by_act(goto_ch);
>> +	if (params)
>> +		kfree_rcu(params, rcu);
>> +	if (res == ACT_P_CREATED)
>> +		tcf_idr_insert(tn, *a);
>> +
>> +	return res;
>> +
>> +cleanup:
>> +	if (goto_ch)
>> +		tcf_chain_put_by_act(goto_ch);
>> +	kfree(params);
>> +	tcf_idr_release(*a, bind);
>> +	return err;
>> +}
>> +
>> +static void tcf_ct_cleanup(struct tc_action *a)
>> +{
>> +	struct tcf_ct_params *params;
>> +	struct tcf_ct *c = to_ct(a);
>> +
>> +	params = rcu_dereference_protected(c->params, 1);
>> +	if (params)
>> +		call_rcu(&params->rcu, tcf_ct_params_free);
>> +}
>> +
>> +static int tcf_ct_dump_key_val(struct sk_buff *skb,
>> +			       void *val, int val_type,
>> +			       void *mask, int mask_type,
>> +			       int len)
>> +{
>> +	int err;
>> +
>> +	if (mask && !memchr_inv(mask, 0, len))
>> +		return 0;
>> +
>> +	err = nla_put(skb, val_type, len, val);
>> +	if (err)
>> +		return err;
>> +
>> +	if (mask_type != TCA_CT_UNSPEC) {
>> +		err = nla_put(skb, mask_type, len, mask);
>> +		if (err)
>> +			return err;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static int tcf_ct_dump_nat(struct sk_buff *skb, struct tcf_ct_params *p)
>> +{
>> +	struct nf_nat_range2 *range = &p->range;
>> +
>> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
>> +		return 0;
>> +
>> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
>> +		return 0;
>> +
>> +	if (range->flags & NF_NAT_RANGE_MAP_IPS) {
>> +		if (p->ipv4_range) {
>> +			if (nla_put_in_addr(skb, TCA_CT_NAT_IPV4_MIN,
>> +					    range->min_addr.ip))
>> +				return -1;
>> +			if (nla_put_in_addr(skb, TCA_CT_NAT_IPV4_MAX,
>> +					    range->max_addr.ip))
>> +				return -1;
>> +		} else {
>> +			if (nla_put_in6_addr(skb, TCA_CT_NAT_IPV6_MIN,
>> +					     &range->min_addr.in6))
>> +				return -1;
>> +			if (nla_put_in6_addr(skb, TCA_CT_NAT_IPV6_MAX,
>> +					     &range->max_addr.in6))
>> +				return -1;
>> +		}
>> +	}
>> +
>> +	if (range->flags & NF_NAT_RANGE_PROTO_SPECIFIED) {
>> +		if (nla_put_u16(skb, TCA_CT_NAT_PORT_MIN,
>> +				ntohs(range->min_proto.all)))
>> +			return -1;
>> +		if (nla_put_u16(skb, TCA_CT_NAT_PORT_MAX,
>> +				ntohs(range->max_proto.all)))
>> +			return -1;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static inline int tcf_ct_dump(struct sk_buff *skb, struct tc_action *a,
>> +			      int bind, int ref)
>> +{
>> +	unsigned char *b = skb_tail_pointer(skb);
>> +	struct tcf_ct *c = to_ct(a);
>> +	struct tcf_ct_params *p;
>> +
>> +	struct tc_ct opt = {
>> +		.index   = c->tcf_index,
>> +		.refcnt  = refcount_read(&c->tcf_refcnt) - ref,
>> +		.bindcnt = atomic_read(&c->tcf_bindcnt) - bind,
>> +	};
>> +	struct tcf_t t;
>> +
>> +	spin_lock_bh(&c->tcf_lock);
>> +	p = rcu_dereference_protected(c->params,
>> +				      lockdep_is_held(&c->tcf_lock));
>> +	opt.action = c->tcf_action;
>> +
>> +	if (tcf_ct_dump_key_val(skb,
>> +				&p->ct_action, TCA_CT_ACTION,
>> +				NULL, TCA_CT_UNSPEC,
>> +				sizeof(p->ct_action)))
>> +		goto nla_put_failure;
>> +
>> +	if (p->ct_action & TCA_CT_ACT_CLEAR)
>> +		goto skip_dump;
>> +
>> +	if (IS_ENABLED(CONFIG_NF_CONNTRACK_MARK) &&
>> +	    tcf_ct_dump_key_val(skb,
>> +				&p->mark, TCA_CT_MARK,
>> +				&p->mark_mask, TCA_CT_MARK_MASK,
>> +				sizeof(p->mark)))
>> +		goto nla_put_failure;
>> +
>> +	if (IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS) &&
>> +	    tcf_ct_dump_key_val(skb,
>> +				p->labels, TCA_CT_LABELS,
>> +				p->labels_mask, TCA_CT_LABELS_MASK,
>> +				sizeof(p->labels)))
>> +		goto nla_put_failure;
>> +
>> +	if (IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES) &&
>> +	    tcf_ct_dump_key_val(skb,
>> +				&p->zone, TCA_CT_ZONE,
>> +				NULL, TCA_CT_UNSPEC,
>> +				sizeof(p->zone)))
>> +		goto nla_put_failure;
>> +
>> +	if (tcf_ct_dump_nat(skb, p))
>> +		goto nla_put_failure;
>> +
>> +skip_dump:
>> +	if (nla_put(skb, TCA_CT_PARMS, sizeof(opt), &opt))
>> +		goto nla_put_failure;
>> +
>> +	tcf_tm_dump(&t, &c->tcf_tm);
>> +	if (nla_put_64bit(skb, TCA_CT_TM, sizeof(t), &t, TCA_CT_PAD))
>> +		goto nla_put_failure;
>> +	spin_unlock_bh(&c->tcf_lock);
>> +
>> +	return skb->len;
>> +nla_put_failure:
>> +	spin_unlock_bh(&c->tcf_lock);
>> +	nlmsg_trim(skb, b);
>> +	return -1;
>> +}
>> +
>> +static int tcf_ct_walker(struct net *net, struct sk_buff *skb,
>> +			 struct netlink_callback *cb, int type,
>> +			 const struct tc_action_ops *ops,
>> +			 struct netlink_ext_ack *extack)
>> +{
>> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
>> +
>> +	return tcf_generic_walker(tn, skb, cb, type, ops, extack);
>> +}
>> +
>> +static int tcf_ct_search(struct net *net, struct tc_action **a, u32 index)
>> +{
>> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
>> +
>> +	return tcf_idr_search(tn, a, index);
>> +}
>> +
>> +static void tcf_stats_update(struct tc_action *a, u64 bytes, u32 packets,
>> +			     u64 lastuse, bool hw)
>> +{
>> +	struct tcf_ct *c = to_ct(a);
>> +
>> +	_bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), bytes, packets);
>> +
>> +	if (hw)
>> +		_bstats_cpu_update(this_cpu_ptr(a->cpu_bstats_hw),
>> +				   bytes, packets);
>> +	c->tcf_tm.lastuse = max_t(u64, c->tcf_tm.lastuse, lastuse);
>> +}
>> +
>> +static struct tc_action_ops act_ct_ops = {
>> +	.kind		=	"ct",
>> +	.id		=	TCA_ID_CT,
>> +	.owner		=	THIS_MODULE,
>> +	.act		=	tcf_ct_act,
>> +	.dump		=	tcf_ct_dump,
>> +	.init		=	tcf_ct_init,
>> +	.cleanup	=	tcf_ct_cleanup,
>> +	.walk		=	tcf_ct_walker,
>> +	.lookup		=	tcf_ct_search,
>> +	.stats_update	=	tcf_stats_update,
>> +	.size		=	sizeof(struct tcf_ct),
>> +};
>> +
>> +static __net_init int ct_init_net(struct net *net)
>> +{
>> +	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
>> +	unsigned int n_bits = FIELD_SIZEOF(struct tcf_ct_params, labels) * 8;
>> +
>> +	if (nf_connlabels_get(net, n_bits - 1)) {
>> +		tn->labels = false;
>> +		pr_err("act_ct: Failed to set connlabels length");
>> +	} else {
>> +		tn->labels = true;
>> +	}
>> +
>> +	return tc_action_net_init(&tn->tn, &act_ct_ops);
>> +}
>> +
>> +static void __net_exit ct_exit_net(struct list_head *net_list)
>> +{
>> +	struct net *net;
>> +
>> +	rtnl_lock();
>> +	list_for_each_entry(net, net_list, exit_list) {
>> +		struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
>> +
>> +		if (tn->labels)
>> +			nf_connlabels_put(net);
>> +	}
>> +	rtnl_unlock();
>> +
>> +	tc_action_net_exit(net_list, ct_net_id);
>> +}
>> +
>> +static struct pernet_operations ct_net_ops = {
>> +	.init = ct_init_net,
>> +	.exit_batch = ct_exit_net,
>> +	.id   = &ct_net_id,
>> +	.size = sizeof(struct tc_ct_action_net),
>> +};
>> +
>> +static int __init ct_init_module(void)
>> +{
>> +	return tcf_register_action(&act_ct_ops, &ct_net_ops);
>> +}
>> +
>> +static void __exit ct_cleanup_module(void)
>> +{
>> +	tcf_unregister_action(&act_ct_ops, &ct_net_ops);
>> +}
>> +
>> +module_init(ct_init_module);
>> +module_exit(ct_cleanup_module);
>> +MODULE_AUTHOR("Paul Blakey <paulb@mellanox.com>");
>> +MODULE_AUTHOR("Yossi Kuperman <yossiku@mellanox.com>");
>> +MODULE_AUTHOR("Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>");
>> +MODULE_DESCRIPTION("Connection tracking action");
>> +MODULE_LICENSE("GPL v2");
>> +
>> diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
>> index ad36bbc..4a7331c 100644
>> --- a/net/sched/cls_api.c
>> +++ b/net/sched/cls_api.c
>> @@ -35,6 +35,7 @@
>>   #include <net/tc_act/tc_police.h>
>>   #include <net/tc_act/tc_sample.h>
>>   #include <net/tc_act/tc_skbedit.h>
>> +#include <net/tc_act/tc_ct.h>
>>   
>>   extern const struct nla_policy rtm_tca_policy[TCA_MAX + 1];
>>   
>> @@ -3266,6 +3267,10 @@ int tc_setup_flow_action(struct flow_action *flow_action,
>>   			entry->police.burst = tcf_police_tcfp_burst(act);
>>   			entry->police.rate_bytes_ps =
>>   				tcf_police_rate_bytes_ps(act);
>> +		} else if (is_tcf_ct(act)) {
>> +			entry->id = FLOW_ACTION_CT;
>> +			entry->ct.action = tcf_ct_action(act);
>> +			entry->ct.zone = tcf_ct_zone(act);
>>   		} else {
>>   			goto err_out;
>>   		}
>> -- 
>> 1.8.3.1

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 14:15     ` Paul Blakey
@ 2019-06-11 15:34       ` Toke Høiland-Jørgensen
  2019-06-11 15:53         ` Marcelo Ricardo Leitner
  2019-06-13 11:00       ` Kevin Darbyshire-Bryant
  1 sibling, 1 reply; 28+ messages in thread
From: Toke Høiland-Jørgensen @ 2019-06-11 15:34 UTC (permalink / raw)
  To: Paul Blakey, Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev\,
	David Miller, Aaron Conole, Zhike Wang
  Cc: Rony Efraim, nst-kernel\,
	John Hurley, Simon Horman, Justin Pettit,
	Kevin Darbyshire-Bryant

Paul Blakey <paulb@mellanox.com> writes:

> On 6/11/2019 4:59 PM, Toke Høiland-Jørgensen wrote:
>> Paul Blakey <paulb@mellanox.com> writes:
>>
>>> Allow sending a packet to conntrack and set conntrack zone, mark,
>>> labels and nat parameters.
>> How is this different from the newly merged ctinfo action?
>>
>> -Toke
>
> Hi,
>
> ctinfo does one of two very specific things,
>
> 1) copies DSCP values that have been placed in the firewall conntrack 
> mark back into the IPv4/v6 diffserv field
>
> 2) copies the firewall conntrack mark to the skb's mark field (like 
> act_connmark)
>
> Originally ctinfo action was named conndscp (then conntrack, which is 
> what our ct shorthand stands for).
>
> We also talked about merging both at some point, but they seem only 
> coincidentally related.

Well, I'm predicting it will create some confusion to have them so
closely named... Not sure what the best way to fix that is, though...?

-Toke

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 15:34       ` Toke Høiland-Jørgensen
@ 2019-06-11 15:53         ` Marcelo Ricardo Leitner
  2019-06-11 20:23           ` Toke Høiland-Jørgensen
  2019-06-14 18:07           ` Cong Wang
  0 siblings, 2 replies; 28+ messages in thread
From: Marcelo Ricardo Leitner @ 2019-06-11 15:53 UTC (permalink / raw)
  To: Toke Høiland-Jørgensen
  Cc: Paul Blakey, Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	netdev, David Miller, Aaron Conole, Zhike Wang, Rony Efraim,
	nst-kernel, John Hurley, Simon Horman, Justin Pettit,
	Kevin Darbyshire-Bryant

On Tue, Jun 11, 2019 at 05:34:50PM +0200, Toke Høiland-Jørgensen wrote:
> Paul Blakey <paulb@mellanox.com> writes:
> 
> > On 6/11/2019 4:59 PM, Toke Høiland-Jørgensen wrote:
> >> Paul Blakey <paulb@mellanox.com> writes:
> >>
> >>> Allow sending a packet to conntrack and set conntrack zone, mark,
> >>> labels and nat parameters.
> >> How is this different from the newly merged ctinfo action?
> >>
> >> -Toke
> >
> > Hi,
> >
> > ctinfo does one of two very specific things,
> >
> > 1) copies DSCP values that have been placed in the firewall conntrack 
> > mark back into the IPv4/v6 diffserv field
> >
> > 2) copies the firewall conntrack mark to the skb's mark field (like 
> > act_connmark)
> >
> > Originally ctinfo action was named conndscp (then conntrack, which is 
> > what our ct shorthand stands for).
> >
> > We also talked about merging both at some point, but they seem only 
> > coincidentally related.
> 
> Well, I'm predicting it will create some confusion to have them so
> closely named... Not sure what the best way to fix that is, though...?

I had suggested to let act_ct handle the above as well, as there is a
big chunk of code on both that is pretty similar. There is quite some
boilerplate for interfacing with conntrack which is duplicated.
But it was considered that the end actions are unrelated, and ctinfo
went ahead. (I'm still not convinced of that, btw)

Other than this, which is not an option anymore, I don't see a way to
avoid confusion here. Seems anything we pick now will be confusing
because ctinfo is a generic name, and we also need one here.

  Marcelo

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 15:53         ` Marcelo Ricardo Leitner
@ 2019-06-11 20:23           ` Toke Høiland-Jørgensen
  2019-06-13 13:46             ` Paul Blakey
  2019-06-14 18:07           ` Cong Wang
  1 sibling, 1 reply; 28+ messages in thread
From: Toke Høiland-Jørgensen @ 2019-06-11 20:23 UTC (permalink / raw)
  To: Marcelo Ricardo Leitner
  Cc: Paul Blakey, Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	netdev\,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel\,
	John Hurley, Simon Horman, Justin Pettit,
	Kevin Darbyshire-Bryant

Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> writes:

> On Tue, Jun 11, 2019 at 05:34:50PM +0200, Toke Høiland-Jørgensen wrote:
>> Paul Blakey <paulb@mellanox.com> writes:
>> 
>> > On 6/11/2019 4:59 PM, Toke Høiland-Jørgensen wrote:
>> >> Paul Blakey <paulb@mellanox.com> writes:
>> >>
>> >>> Allow sending a packet to conntrack and set conntrack zone, mark,
>> >>> labels and nat parameters.
>> >> How is this different from the newly merged ctinfo action?
>> >>
>> >> -Toke
>> >
>> > Hi,
>> >
>> > ctinfo does one of two very specific things,
>> >
>> > 1) copies DSCP values that have been placed in the firewall conntrack 
>> > mark back into the IPv4/v6 diffserv field
>> >
>> > 2) copies the firewall conntrack mark to the skb's mark field (like 
>> > act_connmark)
>> >
>> > Originally ctinfo action was named conndscp (then conntrack, which is 
>> > what our ct shorthand stands for).
>> >
>> > We also talked about merging both at some point, but they seem only 
>> > coincidentally related.
>> 
>> Well, I'm predicting it will create some confusion to have them so
>> closely named... Not sure what the best way to fix that is, though...?
>
> I had suggested to let act_ct handle the above as well, as there is a
> big chunk of code on both that is pretty similar. There is quite some
> boilerplate for interfacing with conntrack which is duplicated.
> But it was considered that the end actions are unrelated, and ctinfo
> went ahead. (I'm still not convinced of that, btw)
>
> Other than this, which is not an option anymore, I don't see a way to
> avoid confusion here. Seems anything we pick now will be confusing
> because ctinfo is a generic name, and we also need one here.

Hmm, yeah, dunno if I have any better ideas for naming that would avoid
this. act_runct ? Meh...

-Toke

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
  2019-06-11 13:59   ` Toke Høiland-Jørgensen
@ 2019-06-12 18:04   ` Marcelo Ricardo Leitner
  2019-06-13 13:44     ` Paul Blakey
  2019-06-12 19:55   ` Florian Westphal
                     ` (3 subsequent siblings)
  5 siblings, 1 reply; 28+ messages in thread
From: Marcelo Ricardo Leitner @ 2019-06-12 18:04 UTC (permalink / raw)
  To: Paul Blakey
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit

On Tue, Jun 11, 2019 at 04:28:31PM +0300, Paul Blakey wrote:
> Allow sending a packet to conntrack and set conntrack zone, mark,
> labels and nat parameters.
> 
> Signed-off-by: Paul Blakey <paulb@mellanox.com>
> Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
> Acked-by: Jiri Pirko <jiri@mellanox.com>
> ---
>  include/net/flow_offload.h        |   5 +
>  include/net/tc_act/tc_ct.h        |  64 +++
>  include/uapi/linux/pkt_cls.h      |   2 +
>  include/uapi/linux/tc_act/tc_ct.h |  41 ++
>  net/sched/Kconfig                 |  11 +
>  net/sched/Makefile                |   1 +
>  net/sched/act_ct.c                | 901 ++++++++++++++++++++++++++++++++++++++
>  net/sched/cls_api.c               |   5 +
>  8 files changed, 1030 insertions(+)
>  create mode 100644 include/net/tc_act/tc_ct.h
>  create mode 100644 include/uapi/linux/tc_act/tc_ct.h
>  create mode 100644 net/sched/act_ct.c
> 
> diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
> index 36fdb85..5b2c4fa 100644
> --- a/include/net/flow_offload.h
> +++ b/include/net/flow_offload.h
> @@ -123,6 +123,7 @@ enum flow_action_id {
>  	FLOW_ACTION_QUEUE,
>  	FLOW_ACTION_SAMPLE,
>  	FLOW_ACTION_POLICE,
> +	FLOW_ACTION_CT,
>  };
>  
>  /* This is mirroring enum pedit_header_type definition for easy mapping between
> @@ -172,6 +173,10 @@ struct flow_action_entry {
>  			s64			burst;
>  			u64			rate_bytes_ps;
>  		} police;
> +		struct {				/* FLOW_ACTION_CT */
> +			int action;
> +			u16 zone;
> +		} ct;
>  	};
>  };
>  
> diff --git a/include/net/tc_act/tc_ct.h b/include/net/tc_act/tc_ct.h
> new file mode 100644
> index 0000000..59e4f5e
> --- /dev/null
> +++ b/include/net/tc_act/tc_ct.h
> @@ -0,0 +1,64 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +#ifndef __NET_TC_CT_H
> +#define __NET_TC_CT_H
> +
> +#include <net/act_api.h>
> +#include <uapi/linux/tc_act/tc_ct.h>
> +
> +#if IS_ENABLED(CONFIG_NF_CONNTRACK)
> +#include <net/netfilter/nf_nat.h>
> +#include <net/netfilter/nf_conntrack_labels.h>
> +
> +struct tcf_ct_params {
> +	struct nf_conn *tmpl;
> +	u16 zone;
> +
> +	u32 mark;
> +	u32 mark_mask;
> +
> +	u32 labels[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
> +	u32 labels_mask[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
> +
> +	struct nf_nat_range2 range;
> +	bool ipv4_range;
> +
> +	u16 ct_action;
> +
> +	struct rcu_head rcu;
> +
> +};
> +
> +struct tcf_ct {
> +	struct tc_action common;
> +	struct tcf_ct_params __rcu *params;
> +};
> +
> +#define to_ct(a) ((struct tcf_ct *)a)
> +#define to_ct_params(a) ((struct tcf_ct_params *) \
> +			 rtnl_dereference((to_ct(a)->params)))
> +
> +static inline uint16_t tcf_ct_zone(const struct tc_action *a)
> +{
> +	return to_ct_params(a)->zone;
> +}
> +
> +static inline int tcf_ct_action(const struct tc_action *a)
> +{
> +	return to_ct_params(a)->ct_action;
> +}
> +
> +#else
> +static inline uint16_t tcf_ct_zone(const struct tc_action *a) { return 0; }
> +static inline int tcf_ct_action(const struct tc_action *a) { return 0; }
> +#endif /* CONFIG_NF_CONNTRACK */
> +
> +static inline bool is_tcf_ct(const struct tc_action *a)
> +{
> +#if defined(CONFIG_NET_CLS_ACT) && IS_ENABLED(CONFIG_NF_CONNTRACK)
> +	if (a->ops && a->ops->id == TCA_ID_CT)
> +		return true;
> +#endif
> +	return false;
> +}
> +
> +#endif /* __NET_TC_CT_H */
> diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
> index a93680f..c5264d7 100644
> --- a/include/uapi/linux/pkt_cls.h
> +++ b/include/uapi/linux/pkt_cls.h
> @@ -83,6 +83,7 @@ enum {
>  #define TCA_ACT_SIMP 22
>  #define TCA_ACT_IFE 25
>  #define TCA_ACT_SAMPLE 26
> +#define TCA_ACT_CT 27
>  
>  /* Action type identifiers*/
>  enum tca_id {
> @@ -106,6 +107,7 @@ enum tca_id {
>  	TCA_ID_SAMPLE = TCA_ACT_SAMPLE,
>  	/* other actions go here */
>  	TCA_ID_CTINFO,
> +	TCA_ID_CT,
>  	__TCA_ID_MAX = 255
>  };
>  
> diff --git a/include/uapi/linux/tc_act/tc_ct.h b/include/uapi/linux/tc_act/tc_ct.h
> new file mode 100644
> index 0000000..5fb1d7a
> --- /dev/null
> +++ b/include/uapi/linux/tc_act/tc_ct.h
> @@ -0,0 +1,41 @@
> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
> +#ifndef __UAPI_TC_CT_H
> +#define __UAPI_TC_CT_H
> +
> +#include <linux/types.h>
> +#include <linux/pkt_cls.h>
> +
> +enum {
> +	TCA_CT_UNSPEC,
> +	TCA_CT_PARMS,
> +	TCA_CT_TM,
> +	TCA_CT_ACTION,		/* u16 */
> +	TCA_CT_ZONE,		/* u16 */
> +	TCA_CT_MARK,		/* u32 */
> +	TCA_CT_MARK_MASK,	/* u32 */
> +	TCA_CT_LABELS,		/* u128 */
> +	TCA_CT_LABELS_MASK,	/* u128 */
> +	TCA_CT_NAT_IPV4_MIN,	/* be32 */
> +	TCA_CT_NAT_IPV4_MAX,	/* be32 */
> +	TCA_CT_NAT_IPV6_MIN,	/* struct in6_addr */
> +	TCA_CT_NAT_IPV6_MAX,	/* struct in6_addr */
> +	TCA_CT_NAT_PORT_MIN,	/* be16 */
> +	TCA_CT_NAT_PORT_MAX,	/* be16 */
> +	TCA_CT_PAD,
> +	__TCA_CT_MAX
> +};
> +
> +#define TCA_CT_MAX (__TCA_CT_MAX - 1)
> +
> +#define TCA_CT_ACT_COMMIT	(1 << 0)
> +#define TCA_CT_ACT_FORCE	(1 << 1)
> +#define TCA_CT_ACT_CLEAR	(1 << 2)
> +#define TCA_CT_ACT_NAT		(1 << 3)
> +#define TCA_CT_ACT_NAT_SRC	(1 << 4)
> +#define TCA_CT_ACT_NAT_DST	(1 << 5)
> +
> +struct tc_ct {
> +	tc_gen;
> +};
> +
> +#endif /* __UAPI_TC_CT_H */
> diff --git a/net/sched/Kconfig b/net/sched/Kconfig
> index d104f7e..0481a2a 100644
> --- a/net/sched/Kconfig
> +++ b/net/sched/Kconfig
> @@ -929,6 +929,17 @@ config NET_ACT_TUNNEL_KEY
>  	  To compile this code as a module, choose M here: the
>  	  module will be called act_tunnel_key.
>  
> +config NET_ACT_CT
> +        tristate "connection tracking tc action"
> +        depends on NET_CLS_ACT && NF_CONNTRACK
> +        help
> +	  Say Y here to allow sending the packets to conntrack module.
> +
> +	  If unsure, say N.
> +
> +	  To compile this code as a module, choose M here: the
> +	  module will be called act_ct.
> +
>  config NET_IFE_SKBMARK
>          tristate "Support to encoding decoding skb mark on IFE action"
>          depends on NET_ACT_IFE
> diff --git a/net/sched/Makefile b/net/sched/Makefile
> index d54bfcb..23d2202 100644
> --- a/net/sched/Makefile
> +++ b/net/sched/Makefile
> @@ -28,6 +28,7 @@ obj-$(CONFIG_NET_IFE_SKBMARK)	+= act_meta_mark.o
>  obj-$(CONFIG_NET_IFE_SKBPRIO)	+= act_meta_skbprio.o
>  obj-$(CONFIG_NET_IFE_SKBTCINDEX)	+= act_meta_skbtcindex.o
>  obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o
> +obj-$(CONFIG_NET_ACT_CT)	+= act_ct.o
>  obj-$(CONFIG_NET_SCH_FIFO)	+= sch_fifo.o
>  obj-$(CONFIG_NET_SCH_CBQ)	+= sch_cbq.o
>  obj-$(CONFIG_NET_SCH_HTB)	+= sch_htb.o
> diff --git a/net/sched/act_ct.c b/net/sched/act_ct.c
> new file mode 100644
> index 0000000..4eb0dd7
> --- /dev/null
> +++ b/net/sched/act_ct.c
> @@ -0,0 +1,901 @@
> +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
> +/* -
> + * net/sched/act_ct.c  Connection Tracking action
> + *
> + * Authors:   Paul Blakey <paulb@mellanox.com>
> + *            Yossi Kuperman <yossiku@mellanox.com>
> + *            Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> + */
> +
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/kernel.h>
> +#include <linux/skbuff.h>
> +#include <linux/rtnetlink.h>
> +#include <linux/pkt_cls.h>
> +#include <linux/ip.h>
> +#include <linux/ipv6.h>
> +#include <net/netlink.h>
> +#include <net/pkt_sched.h>
> +#include <net/act_api.h>
> +#include <uapi/linux/tc_act/tc_ct.h>
> +#include <net/tc_act/tc_ct.h>
> +
> +#include <linux/netfilter/nf_nat.h>
> +#include <net/netfilter/nf_conntrack.h>
> +#include <net/netfilter/nf_conntrack_core.h>
> +#include <net/netfilter/nf_conntrack_zones.h>
> +#include <net/netfilter/nf_conntrack_helper.h>
> +#include <net/pkt_cls.h>
> +
> +static struct tc_action_ops act_ct_ops;
> +static unsigned int ct_net_id;
> +
> +struct tc_ct_action_net {
> +	struct tc_action_net tn; /* Must be first */
> +	bool labels;
> +};
> +
> +/* Determine whether skb->_nfct is equal to the result of conntrack lookup. */
> +static bool tcf_ct_skb_nfct_cached(struct net *net, struct sk_buff *skb,
> +				   u16 zone_id, bool force)
> +{
> +	enum ip_conntrack_info ctinfo;
> +	struct nf_conn *ct;
> +
> +	ct = nf_ct_get(skb, &ctinfo);
> +	if (!ct)
> +		return false;
> +	if (!net_eq(net, read_pnet(&ct->ct_net)))
> +		return false;
> +	if (nf_ct_zone(ct)->id != zone_id)
> +		return false;
> +
> +	/* Force conntrack entry direction. */
> +	if (force && CTINFO2DIR(ctinfo) != IP_CT_DIR_ORIGINAL) {
> +		nf_conntrack_put(&ct->ct_general);
> +		nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
> +
> +		if (nf_ct_is_confirmed(ct))
> +			nf_ct_kill(ct);
> +
> +		return false;
> +	}
> +
> +	return true;
> +}
> +
> +/* Trim the skb to the length specified by the IP/IPv6 header,
> + * removing any trailing lower-layer padding. This prepares the skb
> + * for higher-layer processing that assumes skb->len excludes padding
> + * (such as nf_ip_checksum). The caller needs to pull the skb to the
> + * network header, and ensure ip_hdr/ipv6_hdr points to valid data.
> + */
> +static int tcf_ct_skb_network_trim(struct sk_buff *skb)
> +{
> +	unsigned int len;
> +	int err;
> +
> +	switch (skb->protocol) {
> +	case htons(ETH_P_IP):
> +		len = ntohs(ip_hdr(skb)->tot_len);
> +		break;
> +	case htons(ETH_P_IPV6):
> +		len = sizeof(struct ipv6hdr)
> +			+ ntohs(ipv6_hdr(skb)->payload_len);
> +		break;
> +	default:
> +		len = skb->len;
> +	}
> +
> +	err = pskb_trim_rcsum(skb, len);
> +
> +	return err;
> +}
> +
> +static u8 tcf_ct_skb_nf_family(struct sk_buff *skb)
> +{
> +	u8 family = NFPROTO_UNSPEC;
> +
> +	switch (skb->protocol) {
> +	case htons(ETH_P_IP):
> +		family = NFPROTO_IPV4;
> +		break;
> +	case htons(ETH_P_IPV6):
> +		family = NFPROTO_IPV6;
> +		break;
> +	default:
> +	break;
> +	}
> +
> +	return family;
> +}
> +
> +static void tcf_ct_params_free(struct rcu_head *head)
> +{
> +	struct tcf_ct_params *params = container_of(head,
> +						    struct tcf_ct_params, rcu);
> +
> +	if (params->tmpl)
> +		nf_conntrack_put(&params->tmpl->ct_general);
> +	kfree(params);
> +}
> +
> +#if IS_ENABLED(CONFIG_NF_NAT)
> +/* Modelled after nf_nat_ipv[46]_fn().
> + * range is only used for new, uninitialized NAT state.
> + * Returns either NF_ACCEPT or NF_DROP.
> + */
> +static int ct_nat_execute(struct sk_buff *skb, struct nf_conn *ct,
> +			  enum ip_conntrack_info ctinfo,
> +			  const struct nf_nat_range2 *range,
> +			  enum nf_nat_manip_type maniptype)
> +{
> +	int hooknum, nh_off, err = NF_ACCEPT;
> +
> +	nh_off = skb_network_offset(skb);
> +	skb_pull_rcsum(skb, nh_off);
> +
> +	/* See HOOK2MANIP(). */
> +	if (maniptype == NF_NAT_MANIP_SRC)
> +		hooknum = NF_INET_LOCAL_IN; /* Source NAT */
> +	else
> +		hooknum = NF_INET_LOCAL_OUT; /* Destination NAT */
> +
> +	switch (ctinfo) {
> +	case IP_CT_RELATED:
> +	case IP_CT_RELATED_REPLY:
> +		if (skb->protocol == htons(ETH_P_IP) &&
> +		    ip_hdr(skb)->protocol == IPPROTO_ICMP) {
> +			if (!nf_nat_icmp_reply_translation(skb, ct, ctinfo,
> +							   hooknum))
> +				err = NF_DROP;
> +			goto push;
> +		} else if (IS_ENABLED(CONFIG_IPV6) &&
> +			   skb->protocol == htons(ETH_P_IPV6)) {
> +			__be16 frag_off;
> +			u8 nexthdr = ipv6_hdr(skb)->nexthdr;
> +			int hdrlen = ipv6_skip_exthdr(skb,
> +						      sizeof(struct ipv6hdr),
> +						      &nexthdr, &frag_off);
> +
> +			if (hdrlen >= 0 && nexthdr == IPPROTO_ICMPV6) {
> +				if (!nf_nat_icmpv6_reply_translation(skb, ct,
> +								     ctinfo,
> +								     hooknum,
> +								     hdrlen))
> +					err = NF_DROP;
> +				goto push;
> +			}
> +		}
> +		/* Non-ICMP, fall thru to initialize if needed. */
> +		/* fall through */
> +	case IP_CT_NEW:
> +		/* Seen it before?  This can happen for loopback, retrans,
> +		 * or local packets.
> +		 */
> +		if (!nf_nat_initialized(ct, maniptype)) {
> +			/* Initialize according to the NAT action. */
> +			err = (range && range->flags & NF_NAT_RANGE_MAP_IPS)
> +				/* Action is set up to establish a new
> +				 * mapping.
> +				 */
> +				? nf_nat_setup_info(ct, range, maniptype)
> +				: nf_nat_alloc_null_binding(ct, hooknum);
> +			if (err != NF_ACCEPT)
> +				goto push;
> +		}
> +		break;
> +
> +	case IP_CT_ESTABLISHED:
> +	case IP_CT_ESTABLISHED_REPLY:
> +		break;
> +
> +	default:
> +		err = NF_DROP;
> +		goto push;
> +	}
> +
> +	err = nf_nat_packet(ct, ctinfo, hooknum, skb);
> +push:
> +	skb_push(skb, nh_off);
> +	skb_postpush_rcsum(skb, skb->data, nh_off);
> +
> +	return err;
> +}
> +#endif /* CONFIG_NF_NAT */
> +
> +static void tcf_ct_act_set_mark(struct nf_conn *ct, u32 mark, u32 mask)
> +{
> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)
> +	u32 new_mark;
> +
> +	if (!mask)
> +		return;
> +
> +	new_mark = mark | (ct->mark & ~(mask));
> +	if (ct->mark != new_mark) {
> +		ct->mark = new_mark;
> +		if (nf_ct_is_confirmed(ct))
> +			nf_conntrack_event_cache(IPCT_MARK, ct);
> +	}
> +#endif
> +}
> +
> +static void tcf_ct_act_set_labels(struct nf_conn *ct,
> +				  u32 *labels,
> +				  u32 *labels_m)
> +{
> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)
> +	size_t labels_sz = FIELD_SIZEOF(struct tcf_ct_params, labels);
> +
> +	if (!memchr_inv(labels_m, 0, labels_sz))
> +		return;
> +
> +	nf_connlabels_replace(ct, labels, labels_m, 4);
> +#endif
> +}
> +
> +static int tcf_ct_act_nat(struct sk_buff *skb,
> +			  struct nf_conn *ct,
> +			  enum ip_conntrack_info ctinfo,
> +			  int ct_action,
> +			  struct nf_nat_range2 *range,
> +			  bool commit)
> +{
> +#if IS_ENABLED(CONFIG_NF_NAT)
> +	enum nf_nat_manip_type maniptype;
> +
> +	if (!(ct_action & TCA_CT_ACT_NAT))
> +		return NF_ACCEPT;
> +
> +	/* Add NAT extension if not confirmed yet. */
> +	if (!nf_ct_is_confirmed(ct) && !nf_ct_nat_ext_add(ct))
> +		return NF_DROP;   /* Can't NAT. */
> +
> +	if (ctinfo != IP_CT_NEW && (ct->status & IPS_NAT_MASK) &&
> +	    (ctinfo != IP_CT_RELATED || commit)) {
> +		/* NAT an established or related connection like before. */
> +		if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY)
> +			/* This is the REPLY direction for a connection
> +			 * for which NAT was applied in the forward
> +			 * direction.  Do the reverse NAT.
> +			 */
> +			maniptype = ct->status & IPS_SRC_NAT
> +				? NF_NAT_MANIP_DST : NF_NAT_MANIP_SRC;
> +		else
> +			maniptype = ct->status & IPS_SRC_NAT
> +				? NF_NAT_MANIP_SRC : NF_NAT_MANIP_DST;
> +	} else if (ct_action & TCA_CT_ACT_NAT_SRC) {
> +		maniptype = NF_NAT_MANIP_SRC;
> +	} else if (ct_action & TCA_CT_ACT_NAT_DST) {
> +		maniptype = NF_NAT_MANIP_DST;
> +	} else {
> +		return NF_ACCEPT;
> +	}
> +
> +	return ct_nat_execute(skb, ct, ctinfo, range, maniptype);
> +#else
> +	return NF_ACCEPT;
> +#endif
> +}
> +
> +static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
> +		      struct tcf_result *res)
> +{
> +	struct net *net = dev_net(skb->dev);
> +	bool cached, commit, clear, force;
> +	enum ip_conntrack_info ctinfo;
> +	struct tcf_ct *c = to_ct(a);
> +	struct nf_conn *tmpl = NULL;
> +	struct nf_hook_state state;
> +	struct tcf_ct_params *p;
> +	struct nf_conn *ct;
> +	int nh_ofs, err;
> +	u8 family;
> +
> +	p = rcu_dereference_bh(c->params);
> +
> +	commit = p->ct_action & TCA_CT_ACT_COMMIT;
> +	clear = p->ct_action & TCA_CT_ACT_CLEAR;
> +	force = p->ct_action & TCA_CT_ACT_FORCE;
> +	tmpl = p->tmpl;
> +
> +	if (clear) {
> +		ct = nf_ct_get(skb, &ctinfo);
> +		if (ct) {
> +			nf_conntrack_put(&ct->ct_general);
> +			nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
> +		}
> +
> +		goto out;
> +	}
> +
> +	/* The conntrack module expects to be working at L3. */
> +	nh_ofs = skb_network_offset(skb);
> +	skb_pull_rcsum(skb, nh_ofs);
> +
> +	err = tcf_ct_skb_network_trim(skb);
> +	if (err)
> +		goto drop;
> +
> +	family = tcf_ct_skb_nf_family(skb);
> +	if (family == NFPROTO_UNSPEC)
> +		goto drop;
> +
> +	state.hook = NF_INET_PRE_ROUTING;
> +	state.net = net;
> +	state.pf = family;
> +
> +	/* If we are recirculating packets to match on ct fields and
> +	 * committing with a separate ct action, then we don't need to
> +	 * actually run the packet through conntrack twice unless it's for a
> +	 * different zone.
> +	 */
> +	cached = tcf_ct_skb_nfct_cached(net, skb, p->zone, force);
> +	if (!cached) {
> +		/* Associate skb with specified zone. */
> +		if (tmpl) {
> +			if (skb_nfct(skb))
> +				nf_conntrack_put(skb_nfct(skb));
> +			nf_conntrack_get(&tmpl->ct_general);
> +			nf_ct_set(skb, tmpl, IP_CT_NEW);
> +		}
> +
> +		err = nf_conntrack_in(skb, &state);
> +		if (err != NF_ACCEPT)
> +			goto out_push;
> +	}
> +
> +	ct = nf_ct_get(skb, &ctinfo);
> +	if (!ct)
> +		goto out_push;
> +	nf_ct_deliver_cached_events(ct);
> +
> +	err = tcf_ct_act_nat(skb, ct, ctinfo, p->ct_action, &p->range, commit);
> +	if (err != NF_ACCEPT)
> +		goto drop;
> +
> +	if (commit) {
> +		tcf_ct_act_set_mark(ct, p->mark, p->mark_mask);
> +		tcf_ct_act_set_labels(ct, p->labels, p->labels_mask);
> +
> +		/* This will take care of sending queued events
> +		 * even if the connection is already confirmed.
> +		 */
> +		nf_conntrack_confirm(skb);
> +	}
> +
> +out_push:
> +	skb_push(skb, nh_ofs);
> +	skb_postpush_rcsum(skb, skb->data, nh_ofs);
> +
> +out:
> +	bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), skb);
> +
> +	return c->tcf_action;
> +
> +drop:
> +	qstats_drop_inc(this_cpu_ptr(a->cpu_qstats));
> +	return TC_ACT_SHOT;
> +}
> +
> +static const struct nla_policy ct_policy[TCA_CT_MAX + 1] = {
> +	[TCA_CT_ACTION] = { .type = NLA_U16 },
> +	[TCA_CT_PARMS] = { .type = NLA_EXACT_LEN, .len = sizeof(struct tc_ct) },
> +	[TCA_CT_ZONE] = { .type = NLA_U16 },
> +	[TCA_CT_MARK] = { .type = NLA_U32 },
> +	[TCA_CT_MARK_MASK] = { .type = NLA_U32 },
> +	[TCA_CT_LABELS] = { .type = NLA_BINARY,
> +			    .len = 128 / BITS_PER_BYTE },
> +	[TCA_CT_LABELS_MASK] = { .type = NLA_BINARY,
> +				 .len = 128 / BITS_PER_BYTE },
> +	[TCA_CT_NAT_IPV4_MIN] = { .type = NLA_U32 },
> +	[TCA_CT_NAT_IPV4_MAX] = { .type = NLA_U32 },
> +	[TCA_CT_NAT_IPV6_MIN] = { .type = NLA_EXACT_LEN,
> +				  .len = sizeof(struct in6_addr) },
> +	[TCA_CT_NAT_IPV6_MAX] = { .type = NLA_EXACT_LEN,
> +				   .len = sizeof(struct in6_addr) },
> +	[TCA_CT_NAT_PORT_MIN] = { .type = NLA_U16 },
> +	[TCA_CT_NAT_PORT_MAX] = { .type = NLA_U16 },
> +};
> +
> +static int tcf_ct_fill_params_nat(struct tcf_ct_params *p,
> +				  struct tc_ct *parm,
> +				  struct nlattr **tb,
> +				  struct netlink_ext_ack *extack)
> +{
> +	struct nf_nat_range2 *range;
> +
> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
> +		return 0;
> +
> +	if (!IS_ENABLED(CONFIG_NF_NAT)) {
> +		NL_SET_ERR_MSG_MOD(extack, "Netfilter nat isn't enabled in kernel");
> +		return -EOPNOTSUPP;
> +	}
> +
> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
> +		return 0;
> +
> +	if ((p->ct_action & TCA_CT_ACT_NAT_SRC) &&
> +	    (p->ct_action & TCA_CT_ACT_NAT_DST)) {
> +		NL_SET_ERR_MSG_MOD(extack, "dnat and snat can't be enabled at the same time");
> +		return -EOPNOTSUPP;
> +	}
> +
> +	range = &p->range;
> +	if (tb[TCA_CT_NAT_IPV4_MIN]) {
> +		range->min_addr.ip =
> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = true;
> +	}
> +	if (tb[TCA_CT_NAT_IPV4_MAX]) {
> +		range->max_addr.ip =
> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = true;
> +	} else if (range->min_addr.ip) {
> +		range->max_addr.ip = range->min_addr.ip;
> +	}
> +
> +	if (tb[TCA_CT_NAT_IPV6_MIN]) {
> +		range->min_addr.in6 =
> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = false;
> +	}
> +	if (tb[TCA_CT_NAT_IPV6_MAX]) {
> +		range->max_addr.in6 =
> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = false;
> +	} else if (memchr_inv(&range->min_addr.in6, 0,
> +		   sizeof(range->min_addr.in6))) {
> +		range->max_addr.in6 = range->min_addr.in6;
> +	}
> +
> +	if (tb[TCA_CT_NAT_PORT_MIN]) {
> +		range->min_proto.all =
> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MIN]));
> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
> +	}
> +	if (tb[TCA_CT_NAT_PORT_MAX]) {
> +		range->max_proto.all =
> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MAX]));
> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
> +	} else if (range->min_proto.all) {
> +		range->max_proto.all = range->min_proto.all;
> +	}
> +
> +	return 0;
> +}
> +
> +static void tcf_ct_set_key_val(struct nlattr **tb,
> +			       void *val, int val_type,
> +			       void *mask, int mask_type,
> +			       int len)
> +{
> +	if (!tb[val_type])
> +		return;
> +	nla_memcpy(val, tb[val_type], len);
> +
> +	if (!mask)
> +		return;
> +
> +	if (mask_type == TCA_CT_UNSPEC || !tb[mask_type])
> +		memset(mask, 0xff, len);
> +	else
> +		nla_memcpy(mask, tb[mask_type], len);
> +}
> +
> +static int tcf_ct_fill_params(struct net *net,
> +			      struct tcf_ct_params *p,
> +			      struct tc_ct *parm,
> +			      struct nlattr **tb,
> +			      struct netlink_ext_ack *extack)
> +{
> +	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
> +	struct nf_conntrack_zone zone;
> +	struct nf_conn *tmpl;
> +	int err;
> +
> +	p->zone = NF_CT_DEFAULT_ZONE_ID;
> +
> +	tcf_ct_set_key_val(tb,
> +			   &p->ct_action, TCA_CT_ACTION,
> +			   NULL, TCA_CT_UNSPEC,
> +			   sizeof(p->ct_action));
> +
> +	if (p->ct_action & TCA_CT_ACT_CLEAR)
> +		return 0;
> +
> +	err = tcf_ct_fill_params_nat(p, parm, tb, extack);
> +	if (err)
> +		return err;
> +
> +	if (tb[TCA_CT_MARK]) {
> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)) {
> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack mark isn't enabled.");
> +			return -EOPNOTSUPP;
> +		}
> +		tcf_ct_set_key_val(tb,
> +				   &p->mark, TCA_CT_MARK,
> +				   &p->mark_mask, TCA_CT_MARK_MASK,
> +				   sizeof(p->mark));
> +	}
> +
> +	if (tb[TCA_CT_LABELS]) {
> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)) {
> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack labels isn't enabled.");
> +			return -EOPNOTSUPP;
> +		}
> +
> +		if (!tn->labels) {
> +			NL_SET_ERR_MSG_MOD(extack, "Failed to set connlabel length");
> +			return -EOPNOTSUPP;
> +		}
> +		tcf_ct_set_key_val(tb,
> +				   p->labels, TCA_CT_LABELS,
> +				   p->labels_mask, TCA_CT_LABELS_MASK,
> +				   sizeof(p->labels));
> +	}
> +
> +	if (tb[TCA_CT_ZONE]) {
> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack zones isn't enabled.");
> +			return -EOPNOTSUPP;
> +		}
> +
> +		tcf_ct_set_key_val(tb,
> +				   &p->zone, TCA_CT_ZONE,
> +				   NULL, TCA_CT_UNSPEC,
> +				   sizeof(p->zone));
> +	}
> +
> +	if (p->zone == NF_CT_DEFAULT_ZONE_ID)
> +		return 0;
> +
> +	nf_ct_zone_init(&zone, p->zone, NF_CT_DEFAULT_ZONE_DIR, 0);
> +	tmpl = nf_ct_tmpl_alloc(net, &zone, GFP_KERNEL);
> +	if (!tmpl) {
> +		NL_SET_ERR_MSG_MOD(extack, "Failed to allocate conntrack template");
> +		return -ENOMEM;
> +	}
> +	__set_bit(IPS_CONFIRMED_BIT, &tmpl->status);
> +	nf_conntrack_get(&tmpl->ct_general);
> +	p->tmpl = tmpl;
> +
> +	return 0;
> +}
> +
> +static int tcf_ct_init(struct net *net, struct nlattr *nla,
> +		       struct nlattr *est, struct tc_action **a,
> +		       int replace, int bind, bool rtnl_held,
> +		       struct tcf_proto *tp,
> +		       struct netlink_ext_ack *extack)
> +{
> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
> +	struct tcf_ct_params *params = NULL;
> +	struct nlattr *tb[TCA_CT_MAX + 1];
> +	struct tcf_chain *goto_ch = NULL;
> +	struct tc_ct *parm;
> +	struct tcf_ct *c;
> +	int err, res = 0;
> +
> +	if (!nla) {
> +		NL_SET_ERR_MSG_MOD(extack, "Ct requires attributes to be passed");
> +		return -EINVAL;
> +	}
> +
> +	err = nla_parse_nested(tb, TCA_CT_MAX, nla, ct_policy, extack);

I know Paul is aware of this already, but for others:
Please see my reply to
[PATCH net-next v6] net: sched: Introduce act_ctinfo action
regarding the usage of nla_parse_nested() here. Thanks.


^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
  2019-06-11 13:59   ` Toke Høiland-Jørgensen
  2019-06-12 18:04   ` Marcelo Ricardo Leitner
@ 2019-06-12 19:55   ` Florian Westphal
  2019-06-13 13:39     ` Paul Blakey
  2019-06-13 10:16   ` Davide Caratti
                     ` (2 subsequent siblings)
  5 siblings, 1 reply; 28+ messages in thread
From: Florian Westphal @ 2019-06-12 19:55 UTC (permalink / raw)
  To: Paul Blakey
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang, Rony Efraim, nst-kernel, John Hurley, Simon Horman,
	Justin Pettit, pablo

Paul Blakey <paulb@mellanox.com> wrote:
> +	/* The conntrack module expects to be working at L3. */

It also expects that IP stack has validated ip(v6)
headers and has pulled the ip header into linear area.

What are your plans wrt. IP fragments? AFAICS right now they will
not match which means they won't be NATed either.  Is that ok?

For offloading connection tracking and NAT, I think the flowtable
infrastructure is much better: it will allow any device to push packets
that it can't deal with (fragmented, too large mtu, changed route, etc)
to the software path and conntrack will be aware its dealing with a flow
that was offloaded, e.g. it will elide sequence number checks.

For connection tracking on L2, Pablo recently added conntrack for
classic bridge (without the 'call-iptables' infrastructure), see
net/bridge/netfilter/nf_conntrack_bridge.c (especially the defrag/refrag
and header validation its doing).

I suspect parts of that are also needed in the conntrack action (you
might be able to reuse/export some of the functionality I think).

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
                     ` (2 preceding siblings ...)
  2019-06-12 19:55   ` Florian Westphal
@ 2019-06-13 10:16   ` Davide Caratti
  2019-06-13 10:49     ` Davide Caratti
  2019-06-14 18:02   ` Cong Wang
  2019-06-19 18:33   ` Marcelo Ricardo Leitner
  5 siblings, 1 reply; 28+ messages in thread
From: Davide Caratti @ 2019-06-13 10:16 UTC (permalink / raw)
  To: Paul Blakey, Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang
  Cc: Rony Efraim, nst-kernel, John Hurley, Simon Horman, Justin Pettit

hello Paul!

On Tue, 2019-06-11 at 16:28 +0300, Paul Blakey wrote:

> +#endif /* __NET_TC_CT_H */
> diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
> index a93680f..c5264d7 100644
> --- a/include/uapi/linux/pkt_cls.h
> +++ b/include/uapi/linux/pkt_cls.h
> @@ -83,6 +83,7 @@ enum {
>  #define TCA_ACT_SIMP 22
>  #define TCA_ACT_IFE 25
>  #define TCA_ACT_SAMPLE 26
> +#define TCA_ACT_CT 27

^^  I think you can't use 27 (act_ctinfo forgot to explicitly define it),
or the uAPI will break. See below:

>  /* Action type identifiers*/
>  enum tca_id {
> @@ -106,6 +107,7 @@ enum tca_id {
>  	TCA_ID_SAMPLE = TCA_ACT_SAMPLE,
>  	/* other actions go here */
>  	TCA_ID_CTINFO,
> +	TCA_ID_CT,
>  	__TCA_ID_MAX = 255
>  };

and (minor ntit) the comment 

/* other actions go here */

should be interpreted like 

/* other actions go right above this comment */

otherwise the comment will float up as we continue adding new TC actions.

After your patch CTinfo and CT will have the same ID:

[dcaratti@localhost tmp.j4kzBzv3oe]$ cat prova.c 
#include <stdio.h>

#define UNO 1
#define DUE 2
#define TRE 3
#define QUATTRO 4

enum {
        TCA_ID_UNO = UNO,
        TCA_ID_DUE = DUE,
        TCA_ID_TRE = TRE,
        TCA_ID_CTINFO,
        TCA_ID_QUATTRO = QUATTRO,
        TCA_ID_MAX
};

int main (int argc, const char *argv[])
{
        printf("%d %d %d %d %d %d\n", TCA_ID_UNO, TCA_ID_DUE, TCA_ID_TRE,
TCA_ID_CTINFO, TCA_ID_QUATTRO, TCA_ID_MAX);
        return 0;
}
[dcaratti@localhost tmp.j4kzBzv3oe]$ gcc prova.c -o prova
[dcaratti@localhost tmp.j4kzBzv3oe]$ ./prova 
1 2 3 4 4 5
[dcaratti@localhost tmp.j4kzBzv3oe]$ 

so, I think you should use 28. And I will send a patch for net-next now
that adds the missing define for TCA_ID_CTINFO. Ok?

-- 
davide


^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-13 10:16   ` Davide Caratti
@ 2019-06-13 10:49     ` Davide Caratti
  2019-06-13 13:40       ` Paul Blakey
  0 siblings, 1 reply; 28+ messages in thread
From: Davide Caratti @ 2019-06-13 10:49 UTC (permalink / raw)
  To: Paul Blakey, Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang
  Cc: Justin Pettit, John Hurley, Rony Efraim, nst-kernel, Simon Horman

On Thu, 2019-06-13 at 12:16 +0200, Davide Caratti wrote:
> hello Paul!
> 
> On Tue, 2019-06-11 at 16:28 +0300, Paul Blakey wrote:
> 
> > +#endif /* __NET_TC_CT_H */
> > diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
> > index a93680f..c5264d7 100644
> > --- a/include/uapi/linux/pkt_cls.h
> > +++ b/include/uapi/linux/pkt_cls.h
> > @@ -83,6 +83,7 @@ enum {
> >  #define TCA_ACT_SIMP 22
> >  #define TCA_ACT_IFE 25
> >  #define TCA_ACT_SAMPLE 26
> > +#define TCA_ACT_CT 27
> 
> ^^  I think you can't use 27 (act_ctinfo forgot to explicitly define it),
> or the uAPI will break. See below:

[...]

Nevermind, I just read the comment above the definition of TCA_ACT_GACT.
> > 
> so, I think you should use 28. And I will send a patch for net-next now
> that adds the missing define for TCA_ID_CTINFO. Ok?

we don't need to patch pkt_cls.h. Just avoid 

#define TCA_ACT_CT 27 

and the assignment in the enum, that should be sufficient.
-- 
davide


^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 14:15     ` Paul Blakey
  2019-06-11 15:34       ` Toke Høiland-Jørgensen
@ 2019-06-13 11:00       ` Kevin Darbyshire-Bryant
  1 sibling, 0 replies; 28+ messages in thread
From: Kevin Darbyshire-Bryant @ 2019-06-13 11:00 UTC (permalink / raw)
  To: Paul Blakey
  Cc: Toke Høiland-Jørgensen, Jiri Pirko, Roi Dayan,
	Yossi Kuperman, Oz Shlomo, Marcelo Ricardo Leitner, netdev,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit

[-- Attachment #1: Type: text/plain, Size: 2202 bytes --]



> On 11 Jun 2019, at 16:15, Paul Blakey <paulb@mellanox.com> wrote:
> 
> 
> On 6/11/2019 4:59 PM, Toke Høiland-Jørgensen wrote:
>> Paul Blakey <paulb@mellanox.com> writes:
>> 
>>> Allow sending a packet to conntrack and set conntrack zone, mark,
>>> labels and nat parameters.
>> How is this different from the newly merged ctinfo action?
>> 
>> -Toke
> 
> Hi,
> 
> ctinfo does one of two very specific things,
> 
> 1) copies DSCP values that have been placed in the firewall conntrack
> mark back into the IPv4/v6 diffserv field
> 
> 2) copies the firewall conntrack mark to the skb's mark field (like
> act_connmark)

It can do both at the same time if required, taking advantage of the single
conntrack entry lookup for both packet/skb mangling operations, but this isn’t
relevant to the discussion really.

> 
> Originally ctinfo action was named conndscp (then conntrack, which is
> what our ct shorthand stands for).
> 
> We also talked about merging both at some point, but they seem only
> coincidentally related.
> 
> don't know how it was then be agreed to be named ctinfo suggesting it
> does something else but the above.

I’m a newbie around here so trying to fit in.  conndscp did one thing, then it
suggested that as it was doing a similar lookup to act_connmark that the connmark
functionality could also be integrated.  There was a brief flirtation with a
new ‘act ct’ it sort of ‘fell out’ that they were only semi-related in function
by name only.

conndscp was clearly the wrong name for what act_ctinfo had become, amalgamating
two functions, so I thought it’s a “conntrack information lookup/user/extractor/mangler’
and thought ‘ctinfo’ was as good as anything - and nobody screamed and AFAIK no
kittens died :-)

But as a newbie around here I’m happy to fit in with whatever consensus is reached
as long as it is reached.

> 
> This action sends packets to conntrack, configures nat, and doesn't get
> "info" from conntrack, while the ctinfo already expects packets to be
> passed conntrack
> 
> by some other kernel mechanism.
> 

Yeah, one is pulling, the other is pushing :-)

Kevin


[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-12 19:55   ` Florian Westphal
@ 2019-06-13 13:39     ` Paul Blakey
  0 siblings, 0 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-13 13:39 UTC (permalink / raw)
  To: Florian Westphal
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang, Rony Efraim, nst-kernel, John Hurley, Simon Horman,
	Justin Pettit, pablo


On 6/12/2019 10:55 PM, Florian Westphal wrote:
> Paul Blakey <paulb@mellanox.com> wrote:
>> +	/* The conntrack module expects to be working at L3. */
> It also expects that IP stack has validated ip(v6)
> headers and has pulled the ip header into linear area.
>
> What are your plans wrt. IP fragments? AFAICS right now they will
> not match which means they won't be NATed either.  Is that ok?

Right, I'll add defragment via nf_ct_ipv4_gather_frags and 
nf_ct_frag6_gather, which

is basically what conntrack_bridge and openvswitch/conntrack.c does with 
it's calls.

> For offloading connection tracking and NAT, I think the flowtable
> infrastructure is much better: it will allow any device to push packets
> that it can't deal with (fragmented, too large mtu, changed route, etc)
> to the software path and conntrack will be aware its dealing with a flow
> that was offloaded, e.g. it will elide sequence number checks.

That still be the case, Later, hardware offload for this tc rules will 
not offload fragmented packets (unless they somehow support it)

and they will be defragmented in act_ct.c and only then will pass this 
action.

Such offloaded connections will be marked as such and so netfilter can 
skip seq number checking.

> For connection tracking on L2, Pablo recently added conntrack for
> classic bridge (without the 'call-iptables' infrastructure), see
> net/bridge/netfilter/nf_conntrack_bridge.c (especially the defrag/refrag
> and header validation its doing).
>
> I suspect parts of that are also needed in the conntrack action (you
> might be able to reuse/export some of the functionality I think).

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-13 10:49     ` Davide Caratti
@ 2019-06-13 13:40       ` Paul Blakey
  0 siblings, 0 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-13 13:40 UTC (permalink / raw)
  To: Davide Caratti, Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, netdev, David Miller, Aaron Conole,
	Zhike Wang
  Cc: Justin Pettit, John Hurley, Rony Efraim, nst-kernel, Simon Horman


On 6/13/2019 1:49 PM, Davide Caratti wrote:
> On Thu, 2019-06-13 at 12:16 +0200, Davide Caratti wrote:
>> hello Paul!
>>
>> On Tue, 2019-06-11 at 16:28 +0300, Paul Blakey wrote:
>>
>>> +#endif /* __NET_TC_CT_H */
>>> diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
>>> index a93680f..c5264d7 100644
>>> --- a/include/uapi/linux/pkt_cls.h
>>> +++ b/include/uapi/linux/pkt_cls.h
>>> @@ -83,6 +83,7 @@ enum {
>>>   #define TCA_ACT_SIMP 22
>>>   #define TCA_ACT_IFE 25
>>>   #define TCA_ACT_SAMPLE 26
>>> +#define TCA_ACT_CT 27
>> ^^  I think you can't use 27 (act_ctinfo forgot to explicitly define it),
>> or the uAPI will break. See below:
> [...]
>
> Nevermind, I just read the comment above the definition of TCA_ACT_GACT.
>> so, I think you should use 28. And I will send a patch for net-next now
>> that adds the missing define for TCA_ID_CTINFO. Ok?
> we don't need to patch pkt_cls.h. Just avoid
>
> #define TCA_ACT_CT 27
>
> and the assignment in the enum, that should be sufficient.


Right, thanks.


^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-12 18:04   ` Marcelo Ricardo Leitner
@ 2019-06-13 13:44     ` Paul Blakey
  0 siblings, 0 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-13 13:44 UTC (permalink / raw)
  To: Marcelo Ricardo Leitner
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit


On 6/12/2019 9:04 PM, Marcelo Ricardo Leitner wrote:
> On Tue, Jun 11, 2019 at 04:28:31PM +0300, Paul Blakey wrote:
>> Allow sending a packet to conntrack and set conntrack zone, mark,
>> labels and nat parameters.
>>
>> Signed-off-by: Paul Blakey <paulb@mellanox.com>
>> Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
>> Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
>> Acked-by: Jiri Pirko <jiri@mellanox.com>
>> ---
>>   include/net/flow_offload.h        |   5 +
>>   include/net/tc_act/tc_ct.h        |  64 +++
>>   include/uapi/linux/pkt_cls.h      |   2 +
>>   include/uapi/linux/tc_act/tc_ct.h |  41 ++
>>   net/sched/Kconfig                 |  11 +
>>   net/sched/Makefile                |   1 +
>>   net/sched/act_ct.c                | 901 ++++++++++++++++++++++++++++++++++++++
>>   net/sched/cls_api.c               |   5 +
>>   8 files changed, 1030 insertions(+)
>>   create mode 100644 include/net/tc_act/tc_ct.h
>>   create mode 100644 include/uapi/linux/tc_act/tc_ct.h
>>   create mode 100644 net/sched/act_ct.c
>>
>> diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
>> index 36fdb85..5b2c4fa 100644
>> --- a/include/net/flow_offload.h
>> +++ b/include/net/flow_offload.h
>> @@ -123,6 +123,7 @@ enum flow_action_id {
>>   	FLOW_ACTION_QUEUE,
>>   	FLOW_ACTION_SAMPLE,
>>   	FLOW_ACTION_POLICE,
>> +	FLOW_ACTION_CT,
>>   };
>>   
>>   /* This is mirroring enum pedit_header_type definition for easy mapping between
>> @@ -172,6 +173,10 @@ struct flow_action_entry {
>>   			s64			burst;
>>   			u64			rate_bytes_ps;
>>   		} police;
>> +		struct {				/* FLOW_ACTION_CT */
>> +			int action;
>> +			u16 zone;
>> +		} ct;
>>   	};
>>   };
>>   
>> diff --git a/include/net/tc_act/tc_ct.h b/include/net/tc_act/tc_ct.h
>> new file mode 100644
>> index 0000000..59e4f5e
>> --- /dev/null
>> +++ b/include/net/tc_act/tc_ct.h
>> @@ -0,0 +1,64 @@
>> +/* SPDX-License-Identifier: GPL-2.0 */
>> +#ifndef __NET_TC_CT_H
>> +#define __NET_TC_CT_H
>> +
>> +#include <net/act_api.h>
>> +#include <uapi/linux/tc_act/tc_ct.h>
>> +
>> +#if IS_ENABLED(CONFIG_NF_CONNTRACK)
>> +#include <net/netfilter/nf_nat.h>
>> +#include <net/netfilter/nf_conntrack_labels.h>
>> +
>> +struct tcf_ct_params {
>> +	struct nf_conn *tmpl;
>> +	u16 zone;
>> +
>> +	u32 mark;
>> +	u32 mark_mask;
>> +
>> +	u32 labels[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
>> +	u32 labels_mask[NF_CT_LABELS_MAX_SIZE / sizeof(u32)];
>> +
>> +	struct nf_nat_range2 range;
>> +	bool ipv4_range;
>> +
>> +	u16 ct_action;
>> +
>> +	struct rcu_head rcu;
>> +
>> +};
>> +
>> +struct tcf_ct {
>> +	struct tc_action common;
>> +	struct tcf_ct_params __rcu *params;
>> +};
>> +
>> +#define to_ct(a) ((struct tcf_ct *)a)
>> +#define to_ct_params(a) ((struct tcf_ct_params *) \
>> +			 rtnl_dereference((to_ct(a)->params)))
>> +
>> +static inline uint16_t tcf_ct_zone(const struct tc_action *a)
>> +{
>> +	return to_ct_params(a)->zone;
>> +}
>> +
>> +static inline int tcf_ct_action(const struct tc_action *a)
>> +{
>> +	return to_ct_params(a)->ct_action;
>> +}
>> +
>> +#else
>> +static inline uint16_t tcf_ct_zone(const struct tc_action *a) { return 0; }
>> +static inline int tcf_ct_action(const struct tc_action *a) { return 0; }
>> +#endif /* CONFIG_NF_CONNTRACK */
>> +
>> +static inline bool is_tcf_ct(const struct tc_action *a)
>> +{
>> +#if defined(CONFIG_NET_CLS_ACT) && IS_ENABLED(CONFIG_NF_CONNTRACK)
>> +	if (a->ops && a->ops->id == TCA_ID_CT)
>> +		return true;
>> +#endif
>> +	return false;
>> +}
>> +
>> +#endif /* __NET_TC_CT_H */
>> diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
>> index a93680f..c5264d7 100644
>> --- a/include/uapi/linux/pkt_cls.h
>> +++ b/include/uapi/linux/pkt_cls.h
>> @@ -83,6 +83,7 @@ enum {
>>   #define TCA_ACT_SIMP 22
>>   #define TCA_ACT_IFE 25
>>   #define TCA_ACT_SAMPLE 26
>> +#define TCA_ACT_CT 27
>>   
>>   /* Action type identifiers*/
>>   enum tca_id {
>> @@ -106,6 +107,7 @@ enum tca_id {
>>   	TCA_ID_SAMPLE = TCA_ACT_SAMPLE,
>>   	/* other actions go here */
>>   	TCA_ID_CTINFO,
>> +	TCA_ID_CT,
>>   	__TCA_ID_MAX = 255
>>   };
>>   
>> diff --git a/include/uapi/linux/tc_act/tc_ct.h b/include/uapi/linux/tc_act/tc_ct.h
>> new file mode 100644
>> index 0000000..5fb1d7a
>> --- /dev/null
>> +++ b/include/uapi/linux/tc_act/tc_ct.h
>> @@ -0,0 +1,41 @@
>> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
>> +#ifndef __UAPI_TC_CT_H
>> +#define __UAPI_TC_CT_H
>> +
>> +#include <linux/types.h>
>> +#include <linux/pkt_cls.h>
>> +
>> +enum {
>> +	TCA_CT_UNSPEC,
>> +	TCA_CT_PARMS,
>> +	TCA_CT_TM,
>> +	TCA_CT_ACTION,		/* u16 */
>> +	TCA_CT_ZONE,		/* u16 */
>> +	TCA_CT_MARK,		/* u32 */
>> +	TCA_CT_MARK_MASK,	/* u32 */
>> +	TCA_CT_LABELS,		/* u128 */
>> +	TCA_CT_LABELS_MASK,	/* u128 */
>> +	TCA_CT_NAT_IPV4_MIN,	/* be32 */
>> +	TCA_CT_NAT_IPV4_MAX,	/* be32 */
>> +	TCA_CT_NAT_IPV6_MIN,	/* struct in6_addr */
>> +	TCA_CT_NAT_IPV6_MAX,	/* struct in6_addr */
>> +	TCA_CT_NAT_PORT_MIN,	/* be16 */
>> +	TCA_CT_NAT_PORT_MAX,	/* be16 */
>> +	TCA_CT_PAD,
>> +	__TCA_CT_MAX
>> +};
>> +
>> +#define TCA_CT_MAX (__TCA_CT_MAX - 1)
>> +
>> +#define TCA_CT_ACT_COMMIT	(1 << 0)
>> +#define TCA_CT_ACT_FORCE	(1 << 1)
>> +#define TCA_CT_ACT_CLEAR	(1 << 2)
>> +#define TCA_CT_ACT_NAT		(1 << 3)
>> +#define TCA_CT_ACT_NAT_SRC	(1 << 4)
>> +#define TCA_CT_ACT_NAT_DST	(1 << 5)
>> +
>> +struct tc_ct {
>> +	tc_gen;
>> +};
>> +
>> +#endif /* __UAPI_TC_CT_H */
>> diff --git a/net/sched/Kconfig b/net/sched/Kconfig
>> index d104f7e..0481a2a 100644
>> --- a/net/sched/Kconfig
>> +++ b/net/sched/Kconfig
>> @@ -929,6 +929,17 @@ config NET_ACT_TUNNEL_KEY
>>   	  To compile this code as a module, choose M here: the
>>   	  module will be called act_tunnel_key.
>>   
>> +config NET_ACT_CT
>> +        tristate "connection tracking tc action"
>> +        depends on NET_CLS_ACT && NF_CONNTRACK
>> +        help
>> +	  Say Y here to allow sending the packets to conntrack module.
>> +
>> +	  If unsure, say N.
>> +
>> +	  To compile this code as a module, choose M here: the
>> +	  module will be called act_ct.
>> +
>>   config NET_IFE_SKBMARK
>>           tristate "Support to encoding decoding skb mark on IFE action"
>>           depends on NET_ACT_IFE
>> diff --git a/net/sched/Makefile b/net/sched/Makefile
>> index d54bfcb..23d2202 100644
>> --- a/net/sched/Makefile
>> +++ b/net/sched/Makefile
>> @@ -28,6 +28,7 @@ obj-$(CONFIG_NET_IFE_SKBMARK)	+= act_meta_mark.o
>>   obj-$(CONFIG_NET_IFE_SKBPRIO)	+= act_meta_skbprio.o
>>   obj-$(CONFIG_NET_IFE_SKBTCINDEX)	+= act_meta_skbtcindex.o
>>   obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o
>> +obj-$(CONFIG_NET_ACT_CT)	+= act_ct.o
>>   obj-$(CONFIG_NET_SCH_FIFO)	+= sch_fifo.o
>>   obj-$(CONFIG_NET_SCH_CBQ)	+= sch_cbq.o
>>   obj-$(CONFIG_NET_SCH_HTB)	+= sch_htb.o
>> diff --git a/net/sched/act_ct.c b/net/sched/act_ct.c
>> new file mode 100644
>> index 0000000..4eb0dd7
>> --- /dev/null
>> +++ b/net/sched/act_ct.c
>> @@ -0,0 +1,901 @@
>> +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
>> +/* -
>> + * net/sched/act_ct.c  Connection Tracking action
>> + *
>> + * Authors:   Paul Blakey <paulb@mellanox.com>
>> + *            Yossi Kuperman <yossiku@mellanox.com>
>> + *            Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
>> + */
>> +
>> +#include <linux/module.h>
>> +#include <linux/init.h>
>> +#include <linux/kernel.h>
>> +#include <linux/skbuff.h>
>> +#include <linux/rtnetlink.h>
>> +#include <linux/pkt_cls.h>
>> +#include <linux/ip.h>
>> +#include <linux/ipv6.h>
>> +#include <net/netlink.h>
>> +#include <net/pkt_sched.h>
>> +#include <net/act_api.h>
>> +#include <uapi/linux/tc_act/tc_ct.h>
>> +#include <net/tc_act/tc_ct.h>
>> +
>> +#include <linux/netfilter/nf_nat.h>
>> +#include <net/netfilter/nf_conntrack.h>
>> +#include <net/netfilter/nf_conntrack_core.h>
>> +#include <net/netfilter/nf_conntrack_zones.h>
>> +#include <net/netfilter/nf_conntrack_helper.h>
>> +#include <net/pkt_cls.h>
>> +
>> +static struct tc_action_ops act_ct_ops;
>> +static unsigned int ct_net_id;
>> +
>> +struct tc_ct_action_net {
>> +	struct tc_action_net tn; /* Must be first */
>> +	bool labels;
>> +};
>> +
>> +/* Determine whether skb->_nfct is equal to the result of conntrack lookup. */
>> +static bool tcf_ct_skb_nfct_cached(struct net *net, struct sk_buff *skb,
>> +				   u16 zone_id, bool force)
>> +{
>> +	enum ip_conntrack_info ctinfo;
>> +	struct nf_conn *ct;
>> +
>> +	ct = nf_ct_get(skb, &ctinfo);
>> +	if (!ct)
>> +		return false;
>> +	if (!net_eq(net, read_pnet(&ct->ct_net)))
>> +		return false;
>> +	if (nf_ct_zone(ct)->id != zone_id)
>> +		return false;
>> +
>> +	/* Force conntrack entry direction. */
>> +	if (force && CTINFO2DIR(ctinfo) != IP_CT_DIR_ORIGINAL) {
>> +		nf_conntrack_put(&ct->ct_general);
>> +		nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
>> +
>> +		if (nf_ct_is_confirmed(ct))
>> +			nf_ct_kill(ct);
>> +
>> +		return false;
>> +	}
>> +
>> +	return true;
>> +}
>> +
>> +/* Trim the skb to the length specified by the IP/IPv6 header,
>> + * removing any trailing lower-layer padding. This prepares the skb
>> + * for higher-layer processing that assumes skb->len excludes padding
>> + * (such as nf_ip_checksum). The caller needs to pull the skb to the
>> + * network header, and ensure ip_hdr/ipv6_hdr points to valid data.
>> + */
>> +static int tcf_ct_skb_network_trim(struct sk_buff *skb)
>> +{
>> +	unsigned int len;
>> +	int err;
>> +
>> +	switch (skb->protocol) {
>> +	case htons(ETH_P_IP):
>> +		len = ntohs(ip_hdr(skb)->tot_len);
>> +		break;
>> +	case htons(ETH_P_IPV6):
>> +		len = sizeof(struct ipv6hdr)
>> +			+ ntohs(ipv6_hdr(skb)->payload_len);
>> +		break;
>> +	default:
>> +		len = skb->len;
>> +	}
>> +
>> +	err = pskb_trim_rcsum(skb, len);
>> +
>> +	return err;
>> +}
>> +
>> +static u8 tcf_ct_skb_nf_family(struct sk_buff *skb)
>> +{
>> +	u8 family = NFPROTO_UNSPEC;
>> +
>> +	switch (skb->protocol) {
>> +	case htons(ETH_P_IP):
>> +		family = NFPROTO_IPV4;
>> +		break;
>> +	case htons(ETH_P_IPV6):
>> +		family = NFPROTO_IPV6;
>> +		break;
>> +	default:
>> +	break;
>> +	}
>> +
>> +	return family;
>> +}
>> +
>> +static void tcf_ct_params_free(struct rcu_head *head)
>> +{
>> +	struct tcf_ct_params *params = container_of(head,
>> +						    struct tcf_ct_params, rcu);
>> +
>> +	if (params->tmpl)
>> +		nf_conntrack_put(&params->tmpl->ct_general);
>> +	kfree(params);
>> +}
>> +
>> +#if IS_ENABLED(CONFIG_NF_NAT)
>> +/* Modelled after nf_nat_ipv[46]_fn().
>> + * range is only used for new, uninitialized NAT state.
>> + * Returns either NF_ACCEPT or NF_DROP.
>> + */
>> +static int ct_nat_execute(struct sk_buff *skb, struct nf_conn *ct,
>> +			  enum ip_conntrack_info ctinfo,
>> +			  const struct nf_nat_range2 *range,
>> +			  enum nf_nat_manip_type maniptype)
>> +{
>> +	int hooknum, nh_off, err = NF_ACCEPT;
>> +
>> +	nh_off = skb_network_offset(skb);
>> +	skb_pull_rcsum(skb, nh_off);
>> +
>> +	/* See HOOK2MANIP(). */
>> +	if (maniptype == NF_NAT_MANIP_SRC)
>> +		hooknum = NF_INET_LOCAL_IN; /* Source NAT */
>> +	else
>> +		hooknum = NF_INET_LOCAL_OUT; /* Destination NAT */
>> +
>> +	switch (ctinfo) {
>> +	case IP_CT_RELATED:
>> +	case IP_CT_RELATED_REPLY:
>> +		if (skb->protocol == htons(ETH_P_IP) &&
>> +		    ip_hdr(skb)->protocol == IPPROTO_ICMP) {
>> +			if (!nf_nat_icmp_reply_translation(skb, ct, ctinfo,
>> +							   hooknum))
>> +				err = NF_DROP;
>> +			goto push;
>> +		} else if (IS_ENABLED(CONFIG_IPV6) &&
>> +			   skb->protocol == htons(ETH_P_IPV6)) {
>> +			__be16 frag_off;
>> +			u8 nexthdr = ipv6_hdr(skb)->nexthdr;
>> +			int hdrlen = ipv6_skip_exthdr(skb,
>> +						      sizeof(struct ipv6hdr),
>> +						      &nexthdr, &frag_off);
>> +
>> +			if (hdrlen >= 0 && nexthdr == IPPROTO_ICMPV6) {
>> +				if (!nf_nat_icmpv6_reply_translation(skb, ct,
>> +								     ctinfo,
>> +								     hooknum,
>> +								     hdrlen))
>> +					err = NF_DROP;
>> +				goto push;
>> +			}
>> +		}
>> +		/* Non-ICMP, fall thru to initialize if needed. */
>> +		/* fall through */
>> +	case IP_CT_NEW:
>> +		/* Seen it before?  This can happen for loopback, retrans,
>> +		 * or local packets.
>> +		 */
>> +		if (!nf_nat_initialized(ct, maniptype)) {
>> +			/* Initialize according to the NAT action. */
>> +			err = (range && range->flags & NF_NAT_RANGE_MAP_IPS)
>> +				/* Action is set up to establish a new
>> +				 * mapping.
>> +				 */
>> +				? nf_nat_setup_info(ct, range, maniptype)
>> +				: nf_nat_alloc_null_binding(ct, hooknum);
>> +			if (err != NF_ACCEPT)
>> +				goto push;
>> +		}
>> +		break;
>> +
>> +	case IP_CT_ESTABLISHED:
>> +	case IP_CT_ESTABLISHED_REPLY:
>> +		break;
>> +
>> +	default:
>> +		err = NF_DROP;
>> +		goto push;
>> +	}
>> +
>> +	err = nf_nat_packet(ct, ctinfo, hooknum, skb);
>> +push:
>> +	skb_push(skb, nh_off);
>> +	skb_postpush_rcsum(skb, skb->data, nh_off);
>> +
>> +	return err;
>> +}
>> +#endif /* CONFIG_NF_NAT */
>> +
>> +static void tcf_ct_act_set_mark(struct nf_conn *ct, u32 mark, u32 mask)
>> +{
>> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)
>> +	u32 new_mark;
>> +
>> +	if (!mask)
>> +		return;
>> +
>> +	new_mark = mark | (ct->mark & ~(mask));
>> +	if (ct->mark != new_mark) {
>> +		ct->mark = new_mark;
>> +		if (nf_ct_is_confirmed(ct))
>> +			nf_conntrack_event_cache(IPCT_MARK, ct);
>> +	}
>> +#endif
>> +}
>> +
>> +static void tcf_ct_act_set_labels(struct nf_conn *ct,
>> +				  u32 *labels,
>> +				  u32 *labels_m)
>> +{
>> +#if IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)
>> +	size_t labels_sz = FIELD_SIZEOF(struct tcf_ct_params, labels);
>> +
>> +	if (!memchr_inv(labels_m, 0, labels_sz))
>> +		return;
>> +
>> +	nf_connlabels_replace(ct, labels, labels_m, 4);
>> +#endif
>> +}
>> +
>> +static int tcf_ct_act_nat(struct sk_buff *skb,
>> +			  struct nf_conn *ct,
>> +			  enum ip_conntrack_info ctinfo,
>> +			  int ct_action,
>> +			  struct nf_nat_range2 *range,
>> +			  bool commit)
>> +{
>> +#if IS_ENABLED(CONFIG_NF_NAT)
>> +	enum nf_nat_manip_type maniptype;
>> +
>> +	if (!(ct_action & TCA_CT_ACT_NAT))
>> +		return NF_ACCEPT;
>> +
>> +	/* Add NAT extension if not confirmed yet. */
>> +	if (!nf_ct_is_confirmed(ct) && !nf_ct_nat_ext_add(ct))
>> +		return NF_DROP;   /* Can't NAT. */
>> +
>> +	if (ctinfo != IP_CT_NEW && (ct->status & IPS_NAT_MASK) &&
>> +	    (ctinfo != IP_CT_RELATED || commit)) {
>> +		/* NAT an established or related connection like before. */
>> +		if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY)
>> +			/* This is the REPLY direction for a connection
>> +			 * for which NAT was applied in the forward
>> +			 * direction.  Do the reverse NAT.
>> +			 */
>> +			maniptype = ct->status & IPS_SRC_NAT
>> +				? NF_NAT_MANIP_DST : NF_NAT_MANIP_SRC;
>> +		else
>> +			maniptype = ct->status & IPS_SRC_NAT
>> +				? NF_NAT_MANIP_SRC : NF_NAT_MANIP_DST;
>> +	} else if (ct_action & TCA_CT_ACT_NAT_SRC) {
>> +		maniptype = NF_NAT_MANIP_SRC;
>> +	} else if (ct_action & TCA_CT_ACT_NAT_DST) {
>> +		maniptype = NF_NAT_MANIP_DST;
>> +	} else {
>> +		return NF_ACCEPT;
>> +	}
>> +
>> +	return ct_nat_execute(skb, ct, ctinfo, range, maniptype);
>> +#else
>> +	return NF_ACCEPT;
>> +#endif
>> +}
>> +
>> +static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
>> +		      struct tcf_result *res)
>> +{
>> +	struct net *net = dev_net(skb->dev);
>> +	bool cached, commit, clear, force;
>> +	enum ip_conntrack_info ctinfo;
>> +	struct tcf_ct *c = to_ct(a);
>> +	struct nf_conn *tmpl = NULL;
>> +	struct nf_hook_state state;
>> +	struct tcf_ct_params *p;
>> +	struct nf_conn *ct;
>> +	int nh_ofs, err;
>> +	u8 family;
>> +
>> +	p = rcu_dereference_bh(c->params);
>> +
>> +	commit = p->ct_action & TCA_CT_ACT_COMMIT;
>> +	clear = p->ct_action & TCA_CT_ACT_CLEAR;
>> +	force = p->ct_action & TCA_CT_ACT_FORCE;
>> +	tmpl = p->tmpl;
>> +
>> +	if (clear) {
>> +		ct = nf_ct_get(skb, &ctinfo);
>> +		if (ct) {
>> +			nf_conntrack_put(&ct->ct_general);
>> +			nf_ct_set(skb, NULL, IP_CT_UNTRACKED);
>> +		}
>> +
>> +		goto out;
>> +	}
>> +
>> +	/* The conntrack module expects to be working at L3. */
>> +	nh_ofs = skb_network_offset(skb);
>> +	skb_pull_rcsum(skb, nh_ofs);
>> +
>> +	err = tcf_ct_skb_network_trim(skb);
>> +	if (err)
>> +		goto drop;
>> +
>> +	family = tcf_ct_skb_nf_family(skb);
>> +	if (family == NFPROTO_UNSPEC)
>> +		goto drop;
>> +
>> +	state.hook = NF_INET_PRE_ROUTING;
>> +	state.net = net;
>> +	state.pf = family;
>> +
>> +	/* If we are recirculating packets to match on ct fields and
>> +	 * committing with a separate ct action, then we don't need to
>> +	 * actually run the packet through conntrack twice unless it's for a
>> +	 * different zone.
>> +	 */
>> +	cached = tcf_ct_skb_nfct_cached(net, skb, p->zone, force);
>> +	if (!cached) {
>> +		/* Associate skb with specified zone. */
>> +		if (tmpl) {
>> +			if (skb_nfct(skb))
>> +				nf_conntrack_put(skb_nfct(skb));
>> +			nf_conntrack_get(&tmpl->ct_general);
>> +			nf_ct_set(skb, tmpl, IP_CT_NEW);
>> +		}
>> +
>> +		err = nf_conntrack_in(skb, &state);
>> +		if (err != NF_ACCEPT)
>> +			goto out_push;
>> +	}
>> +
>> +	ct = nf_ct_get(skb, &ctinfo);
>> +	if (!ct)
>> +		goto out_push;
>> +	nf_ct_deliver_cached_events(ct);
>> +
>> +	err = tcf_ct_act_nat(skb, ct, ctinfo, p->ct_action, &p->range, commit);
>> +	if (err != NF_ACCEPT)
>> +		goto drop;
>> +
>> +	if (commit) {
>> +		tcf_ct_act_set_mark(ct, p->mark, p->mark_mask);
>> +		tcf_ct_act_set_labels(ct, p->labels, p->labels_mask);
>> +
>> +		/* This will take care of sending queued events
>> +		 * even if the connection is already confirmed.
>> +		 */
>> +		nf_conntrack_confirm(skb);
>> +	}
>> +
>> +out_push:
>> +	skb_push(skb, nh_ofs);
>> +	skb_postpush_rcsum(skb, skb->data, nh_ofs);
>> +
>> +out:
>> +	bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), skb);
>> +
>> +	return c->tcf_action;
>> +
>> +drop:
>> +	qstats_drop_inc(this_cpu_ptr(a->cpu_qstats));
>> +	return TC_ACT_SHOT;
>> +}
>> +
>> +static const struct nla_policy ct_policy[TCA_CT_MAX + 1] = {
>> +	[TCA_CT_ACTION] = { .type = NLA_U16 },
>> +	[TCA_CT_PARMS] = { .type = NLA_EXACT_LEN, .len = sizeof(struct tc_ct) },
>> +	[TCA_CT_ZONE] = { .type = NLA_U16 },
>> +	[TCA_CT_MARK] = { .type = NLA_U32 },
>> +	[TCA_CT_MARK_MASK] = { .type = NLA_U32 },
>> +	[TCA_CT_LABELS] = { .type = NLA_BINARY,
>> +			    .len = 128 / BITS_PER_BYTE },
>> +	[TCA_CT_LABELS_MASK] = { .type = NLA_BINARY,
>> +				 .len = 128 / BITS_PER_BYTE },
>> +	[TCA_CT_NAT_IPV4_MIN] = { .type = NLA_U32 },
>> +	[TCA_CT_NAT_IPV4_MAX] = { .type = NLA_U32 },
>> +	[TCA_CT_NAT_IPV6_MIN] = { .type = NLA_EXACT_LEN,
>> +				  .len = sizeof(struct in6_addr) },
>> +	[TCA_CT_NAT_IPV6_MAX] = { .type = NLA_EXACT_LEN,
>> +				   .len = sizeof(struct in6_addr) },
>> +	[TCA_CT_NAT_PORT_MIN] = { .type = NLA_U16 },
>> +	[TCA_CT_NAT_PORT_MAX] = { .type = NLA_U16 },
>> +};
>> +
>> +static int tcf_ct_fill_params_nat(struct tcf_ct_params *p,
>> +				  struct tc_ct *parm,
>> +				  struct nlattr **tb,
>> +				  struct netlink_ext_ack *extack)
>> +{
>> +	struct nf_nat_range2 *range;
>> +
>> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
>> +		return 0;
>> +
>> +	if (!IS_ENABLED(CONFIG_NF_NAT)) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Netfilter nat isn't enabled in kernel");
>> +		return -EOPNOTSUPP;
>> +	}
>> +
>> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
>> +		return 0;
>> +
>> +	if ((p->ct_action & TCA_CT_ACT_NAT_SRC) &&
>> +	    (p->ct_action & TCA_CT_ACT_NAT_DST)) {
>> +		NL_SET_ERR_MSG_MOD(extack, "dnat and snat can't be enabled at the same time");
>> +		return -EOPNOTSUPP;
>> +	}
>> +
>> +	range = &p->range;
>> +	if (tb[TCA_CT_NAT_IPV4_MIN]) {
>> +		range->min_addr.ip =
>> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = true;
>> +	}
>> +	if (tb[TCA_CT_NAT_IPV4_MAX]) {
>> +		range->max_addr.ip =
>> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = true;
>> +	} else if (range->min_addr.ip) {
>> +		range->max_addr.ip = range->min_addr.ip;
>> +	}
>> +
>> +	if (tb[TCA_CT_NAT_IPV6_MIN]) {
>> +		range->min_addr.in6 =
>> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = false;
>> +	}
>> +	if (tb[TCA_CT_NAT_IPV6_MAX]) {
>> +		range->max_addr.in6 =
>> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = false;
>> +	} else if (memchr_inv(&range->min_addr.in6, 0,
>> +		   sizeof(range->min_addr.in6))) {
>> +		range->max_addr.in6 = range->min_addr.in6;
>> +	}
>> +
>> +	if (tb[TCA_CT_NAT_PORT_MIN]) {
>> +		range->min_proto.all =
>> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MIN]));
>> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
>> +	}
>> +	if (tb[TCA_CT_NAT_PORT_MAX]) {
>> +		range->max_proto.all =
>> +			htons(nla_get_u16(tb[TCA_CT_NAT_PORT_MAX]));
>> +		range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
>> +	} else if (range->min_proto.all) {
>> +		range->max_proto.all = range->min_proto.all;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static void tcf_ct_set_key_val(struct nlattr **tb,
>> +			       void *val, int val_type,
>> +			       void *mask, int mask_type,
>> +			       int len)
>> +{
>> +	if (!tb[val_type])
>> +		return;
>> +	nla_memcpy(val, tb[val_type], len);
>> +
>> +	if (!mask)
>> +		return;
>> +
>> +	if (mask_type == TCA_CT_UNSPEC || !tb[mask_type])
>> +		memset(mask, 0xff, len);
>> +	else
>> +		nla_memcpy(mask, tb[mask_type], len);
>> +}
>> +
>> +static int tcf_ct_fill_params(struct net *net,
>> +			      struct tcf_ct_params *p,
>> +			      struct tc_ct *parm,
>> +			      struct nlattr **tb,
>> +			      struct netlink_ext_ack *extack)
>> +{
>> +	struct tc_ct_action_net *tn = net_generic(net, ct_net_id);
>> +	struct nf_conntrack_zone zone;
>> +	struct nf_conn *tmpl;
>> +	int err;
>> +
>> +	p->zone = NF_CT_DEFAULT_ZONE_ID;
>> +
>> +	tcf_ct_set_key_val(tb,
>> +			   &p->ct_action, TCA_CT_ACTION,
>> +			   NULL, TCA_CT_UNSPEC,
>> +			   sizeof(p->ct_action));
>> +
>> +	if (p->ct_action & TCA_CT_ACT_CLEAR)
>> +		return 0;
>> +
>> +	err = tcf_ct_fill_params_nat(p, parm, tb, extack);
>> +	if (err)
>> +		return err;
>> +
>> +	if (tb[TCA_CT_MARK]) {
>> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_MARK)) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack mark isn't enabled.");
>> +			return -EOPNOTSUPP;
>> +		}
>> +		tcf_ct_set_key_val(tb,
>> +				   &p->mark, TCA_CT_MARK,
>> +				   &p->mark_mask, TCA_CT_MARK_MASK,
>> +				   sizeof(p->mark));
>> +	}
>> +
>> +	if (tb[TCA_CT_LABELS]) {
>> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_LABELS)) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack labels isn't enabled.");
>> +			return -EOPNOTSUPP;
>> +		}
>> +
>> +		if (!tn->labels) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Failed to set connlabel length");
>> +			return -EOPNOTSUPP;
>> +		}
>> +		tcf_ct_set_key_val(tb,
>> +				   p->labels, TCA_CT_LABELS,
>> +				   p->labels_mask, TCA_CT_LABELS_MASK,
>> +				   sizeof(p->labels));
>> +	}
>> +
>> +	if (tb[TCA_CT_ZONE]) {
>> +		if (!IS_ENABLED(CONFIG_NF_CONNTRACK_ZONES)) {
>> +			NL_SET_ERR_MSG_MOD(extack, "Conntrack zones isn't enabled.");
>> +			return -EOPNOTSUPP;
>> +		}
>> +
>> +		tcf_ct_set_key_val(tb,
>> +				   &p->zone, TCA_CT_ZONE,
>> +				   NULL, TCA_CT_UNSPEC,
>> +				   sizeof(p->zone));
>> +	}
>> +
>> +	if (p->zone == NF_CT_DEFAULT_ZONE_ID)
>> +		return 0;
>> +
>> +	nf_ct_zone_init(&zone, p->zone, NF_CT_DEFAULT_ZONE_DIR, 0);
>> +	tmpl = nf_ct_tmpl_alloc(net, &zone, GFP_KERNEL);
>> +	if (!tmpl) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Failed to allocate conntrack template");
>> +		return -ENOMEM;
>> +	}
>> +	__set_bit(IPS_CONFIRMED_BIT, &tmpl->status);
>> +	nf_conntrack_get(&tmpl->ct_general);
>> +	p->tmpl = tmpl;
>> +
>> +	return 0;
>> +}
>> +
>> +static int tcf_ct_init(struct net *net, struct nlattr *nla,
>> +		       struct nlattr *est, struct tc_action **a,
>> +		       int replace, int bind, bool rtnl_held,
>> +		       struct tcf_proto *tp,
>> +		       struct netlink_ext_ack *extack)
>> +{
>> +	struct tc_action_net *tn = net_generic(net, ct_net_id);
>> +	struct tcf_ct_params *params = NULL;
>> +	struct nlattr *tb[TCA_CT_MAX + 1];
>> +	struct tcf_chain *goto_ch = NULL;
>> +	struct tc_ct *parm;
>> +	struct tcf_ct *c;
>> +	int err, res = 0;
>> +
>> +	if (!nla) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Ct requires attributes to be passed");
>> +		return -EINVAL;
>> +	}
>> +
>> +	err = nla_parse_nested(tb, TCA_CT_MAX, nla, ct_policy, extack);
> I know Paul is aware of this already, but for others:
> Please see my reply to
> [PATCH net-next v6] net: sched: Introduce act_ctinfo action
> regarding the usage of nla_parse_nested() here. Thanks.

What Marcelo talks about here is that  nla_parse_nested checks has been 
made stricter (compared to nla_parse_nested _deprecated) by checking

the netlink attribute (nla variable here) has the NLA_F_NESTED flag, 
that is suppose to be supplied by the userspace caller (e.g iproute2)

Since this is a new tc action, I'll make sure that before the 
corresponding userspace iproute2 patches merge, we will

send this flag for all tc actions, or at least for this specific function.



^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 20:23           ` Toke Høiland-Jørgensen
@ 2019-06-13 13:46             ` Paul Blakey
  2019-06-13 14:08               ` Toke Høiland-Jørgensen
  0 siblings, 1 reply; 28+ messages in thread
From: Paul Blakey @ 2019-06-13 13:46 UTC (permalink / raw)
  To: Toke Høiland-Jørgensen, Marcelo Ricardo Leitner
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit,
	Kevin Darbyshire-Bryant


On 6/11/2019 11:23 PM, Toke Høiland-Jørgensen wrote:
> Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> writes:
>
>> On Tue, Jun 11, 2019 at 05:34:50PM +0200, Toke Høiland-Jørgensen wrote:
>>> Paul Blakey <paulb@mellanox.com> writes:
>>>
>>>> On 6/11/2019 4:59 PM, Toke Høiland-Jørgensen wrote:
>>>>> Paul Blakey <paulb@mellanox.com> writes:
>>>>>
>>>>>> Allow sending a packet to conntrack and set conntrack zone, mark,
>>>>>> labels and nat parameters.
>>>>> How is this different from the newly merged ctinfo action?
>>>>>
>>>>> -Toke
>>>> Hi,
>>>>
>>>> ctinfo does one of two very specific things,
>>>>
>>>> 1) copies DSCP values that have been placed in the firewall conntrack
>>>> mark back into the IPv4/v6 diffserv field
>>>>
>>>> 2) copies the firewall conntrack mark to the skb's mark field (like
>>>> act_connmark)
>>>>
>>>> Originally ctinfo action was named conndscp (then conntrack, which is
>>>> what our ct shorthand stands for).
>>>>
>>>> We also talked about merging both at some point, but they seem only
>>>> coincidentally related.
>>> Well, I'm predicting it will create some confusion to have them so
>>> closely named... Not sure what the best way to fix that is, though...?
>> I had suggested to let act_ct handle the above as well, as there is a
>> big chunk of code on both that is pretty similar. There is quite some
>> boilerplate for interfacing with conntrack which is duplicated.
>> But it was considered that the end actions are unrelated, and ctinfo
>> went ahead. (I'm still not convinced of that, btw)
>>
>> Other than this, which is not an option anymore, I don't see a way to
>> avoid confusion here. Seems anything we pick now will be confusing
>> because ctinfo is a generic name, and we also need one here.
> Hmm, yeah, dunno if I have any better ideas for naming that would avoid
> this. act_runct ? Meh...
>
> -Toke


If it's fine with you guys, can we keep the name act_ct ? :)


^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-13 13:46             ` Paul Blakey
@ 2019-06-13 14:08               ` Toke Høiland-Jørgensen
  0 siblings, 0 replies; 28+ messages in thread
From: Toke Høiland-Jørgensen @ 2019-06-13 14:08 UTC (permalink / raw)
  To: Paul Blakey, Marcelo Ricardo Leitner
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev\,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel\,
	John Hurley, Simon Horman, Justin Pettit,
	Kevin Darbyshire-Bryant

Paul Blakey <paulb@mellanox.com> writes:

> On 6/11/2019 11:23 PM, Toke Høiland-Jørgensen wrote:
>> Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> writes:
>>
>>> On Tue, Jun 11, 2019 at 05:34:50PM +0200, Toke Høiland-Jørgensen wrote:
>>>> Paul Blakey <paulb@mellanox.com> writes:
>>>>
>>>>> On 6/11/2019 4:59 PM, Toke Høiland-Jørgensen wrote:
>>>>>> Paul Blakey <paulb@mellanox.com> writes:
>>>>>>
>>>>>>> Allow sending a packet to conntrack and set conntrack zone, mark,
>>>>>>> labels and nat parameters.
>>>>>> How is this different from the newly merged ctinfo action?
>>>>>>
>>>>>> -Toke
>>>>> Hi,
>>>>>
>>>>> ctinfo does one of two very specific things,
>>>>>
>>>>> 1) copies DSCP values that have been placed in the firewall conntrack
>>>>> mark back into the IPv4/v6 diffserv field
>>>>>
>>>>> 2) copies the firewall conntrack mark to the skb's mark field (like
>>>>> act_connmark)
>>>>>
>>>>> Originally ctinfo action was named conndscp (then conntrack, which is
>>>>> what our ct shorthand stands for).
>>>>>
>>>>> We also talked about merging both at some point, but they seem only
>>>>> coincidentally related.
>>>> Well, I'm predicting it will create some confusion to have them so
>>>> closely named... Not sure what the best way to fix that is, though...?
>>> I had suggested to let act_ct handle the above as well, as there is a
>>> big chunk of code on both that is pretty similar. There is quite some
>>> boilerplate for interfacing with conntrack which is duplicated.
>>> But it was considered that the end actions are unrelated, and ctinfo
>>> went ahead. (I'm still not convinced of that, btw)
>>>
>>> Other than this, which is not an option anymore, I don't see a way to
>>> avoid confusion here. Seems anything we pick now will be confusing
>>> because ctinfo is a generic name, and we also need one here.
>> Hmm, yeah, dunno if I have any better ideas for naming that would avoid
>> this. act_runct ? Meh...
>>
>> -Toke
>
>
> If it's fine with you guys, can we keep the name act_ct ? :)

Sure, let's just keep the colour of this particular bike shed :)

-Toke

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
                     ` (3 preceding siblings ...)
  2019-06-13 10:16   ` Davide Caratti
@ 2019-06-14 18:02   ` Cong Wang
  2019-06-16  7:29     ` Paul Blakey
  2019-06-19 18:33   ` Marcelo Ricardo Leitner
  5 siblings, 1 reply; 28+ messages in thread
From: Cong Wang @ 2019-06-14 18:02 UTC (permalink / raw)
  To: Paul Blakey
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, Linux Kernel Network Developers,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit

On Tue, Jun 11, 2019 at 7:05 AM Paul Blakey <paulb@mellanox.com> wrote:
>
> Allow sending a packet to conntrack and set conntrack zone, mark,
> labels and nat parameters.
>

This is too short to justify why you want to play with L3 stuff in L2.
Please be as specific as you can.

Also, please document its use case too.

Thanks.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 15:53         ` Marcelo Ricardo Leitner
  2019-06-11 20:23           ` Toke Høiland-Jørgensen
@ 2019-06-14 18:07           ` Cong Wang
  2019-06-14 19:24             ` Marcelo Ricardo Leitner
  1 sibling, 1 reply; 28+ messages in thread
From: Cong Wang @ 2019-06-14 18:07 UTC (permalink / raw)
  To: Marcelo Ricardo Leitner
  Cc: Toke Høiland-Jørgensen, Paul Blakey, Jiri Pirko,
	Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev, David Miller,
	Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel, John Hurley,
	Simon Horman, Justin Pettit, Kevin Darbyshire-Bryant

On Tue, Jun 11, 2019 at 9:44 AM Marcelo Ricardo Leitner
<marcelo.leitner@gmail.com> wrote:
> I had suggested to let act_ct handle the above as well, as there is a
> big chunk of code on both that is pretty similar. There is quite some
> boilerplate for interfacing with conntrack which is duplicated.

Why do you want to mix retrieving conntrack info with executing
conntrack?

They are totally different things to me, act_ctinfo merely retrieves
information from conntrack, while this one, act_ct, is supposed to
move packets to conntrack.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-14 18:07           ` Cong Wang
@ 2019-06-14 19:24             ` Marcelo Ricardo Leitner
  2019-06-18 16:03               ` Cong Wang
  0 siblings, 1 reply; 28+ messages in thread
From: Marcelo Ricardo Leitner @ 2019-06-14 19:24 UTC (permalink / raw)
  To: Cong Wang
  Cc: Toke Høiland-Jørgensen, Paul Blakey, Jiri Pirko,
	Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev, David Miller,
	Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel, John Hurley,
	Simon Horman, Justin Pettit, Kevin Darbyshire-Bryant

On Fri, Jun 14, 2019 at 11:07:37AM -0700, Cong Wang wrote:
> On Tue, Jun 11, 2019 at 9:44 AM Marcelo Ricardo Leitner
> <marcelo.leitner@gmail.com> wrote:
> > I had suggested to let act_ct handle the above as well, as there is a
> > big chunk of code on both that is pretty similar. There is quite some
> > boilerplate for interfacing with conntrack which is duplicated.
> 
> Why do you want to mix retrieving conntrack info with executing
> conntrack?

To save on the heavy boilerplate for interfacing with conntrack.

> 
> They are totally different things to me, act_ctinfo merely retrieves
> information from conntrack, while this one, act_ct, is supposed to
> move packets to conntrack.

Seems we have a different understanding for "move packets to
conntrack": conntrack will not consume the packets after this.
But after act_ct is executed, if not with the clear flag, skb will now
have the skb->_nfct entry available, on which flower then will be able
to match. So in essence, it is also fetching information from
conntrack.

I see act_ctinfo is a subset of what act_ct is doing.

  Marcelo

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-14 18:02   ` Cong Wang
@ 2019-06-16  7:29     ` Paul Blakey
  0 siblings, 0 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-16  7:29 UTC (permalink / raw)
  To: Cong Wang
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo,
	Marcelo Ricardo Leitner, Linux Kernel Network Developers,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit


On 6/14/2019 9:02 PM, Cong Wang wrote:
> On Tue, Jun 11, 2019 at 7:05 AM Paul Blakey <paulb@mellanox.com> wrote:
>> Allow sending a packet to conntrack and set conntrack zone, mark,
>> labels and nat parameters.
>>
> This is too short to justify why you want to play with L3 stuff in L2.
> Please be as specific as you can.
>
> Also, please document its use case too.
>
> Thanks.

Sure I'll add that.


^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-14 19:24             ` Marcelo Ricardo Leitner
@ 2019-06-18 16:03               ` Cong Wang
  2019-06-20  7:32                 ` Paul Blakey
  0 siblings, 1 reply; 28+ messages in thread
From: Cong Wang @ 2019-06-18 16:03 UTC (permalink / raw)
  To: Marcelo Ricardo Leitner
  Cc: Toke Høiland-Jørgensen, Paul Blakey, Jiri Pirko,
	Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev, David Miller,
	Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel, John Hurley,
	Simon Horman, Justin Pettit, Kevin Darbyshire-Bryant

On Fri, Jun 14, 2019 at 12:24 PM Marcelo Ricardo Leitner
<marcelo.leitner@gmail.com> wrote:
>
> On Fri, Jun 14, 2019 at 11:07:37AM -0700, Cong Wang wrote:
> > On Tue, Jun 11, 2019 at 9:44 AM Marcelo Ricardo Leitner
> > <marcelo.leitner@gmail.com> wrote:
> > > I had suggested to let act_ct handle the above as well, as there is a
> > > big chunk of code on both that is pretty similar. There is quite some
> > > boilerplate for interfacing with conntrack which is duplicated.
> >
> > Why do you want to mix retrieving conntrack info with executing
> > conntrack?
>
> To save on the heavy boilerplate for interfacing with conntrack.
>
> >
> > They are totally different things to me, act_ctinfo merely retrieves
> > information from conntrack, while this one, act_ct, is supposed to
> > move packets to conntrack.
>
> Seems we have a different understanding for "move packets to
> conntrack": conntrack will not consume the packets after this.
> But after act_ct is executed, if not with the clear flag, skb will now
> have the skb->_nfct entry available, on which flower then will be able
> to match. So in essence, it is also fetching information from
> conntrack.

Interesting. Is it because cls_flower uses conntrack for flow dissection?
What's the reason behind?

Again, I am still not convinced to do L3 operations in L2, skb->_nfct
belongs to conntrack which is L3, no matter the packet is consumed
or not.

Thanks.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
                     ` (4 preceding siblings ...)
  2019-06-14 18:02   ` Cong Wang
@ 2019-06-19 18:33   ` Marcelo Ricardo Leitner
  2019-06-20  7:07     ` Paul Blakey
  5 siblings, 1 reply; 28+ messages in thread
From: Marcelo Ricardo Leitner @ 2019-06-19 18:33 UTC (permalink / raw)
  To: Paul Blakey
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit

On Tue, Jun 11, 2019 at 04:28:31PM +0300, Paul Blakey wrote:
...
> +static int tcf_ct_fill_params_nat(struct tcf_ct_params *p,
> +				  struct tc_ct *parm,
> +				  struct nlattr **tb,
> +				  struct netlink_ext_ack *extack)
> +{
> +	struct nf_nat_range2 *range;
> +
> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
> +		return 0;
> +
> +	if (!IS_ENABLED(CONFIG_NF_NAT)) {
> +		NL_SET_ERR_MSG_MOD(extack, "Netfilter nat isn't enabled in kernel");
> +		return -EOPNOTSUPP;
> +	}
> +
> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
> +		return 0;
> +
> +	if ((p->ct_action & TCA_CT_ACT_NAT_SRC) &&
> +	    (p->ct_action & TCA_CT_ACT_NAT_DST)) {
> +		NL_SET_ERR_MSG_MOD(extack, "dnat and snat can't be enabled at the same time");
> +		return -EOPNOTSUPP;
> +	}
> +
> +	range = &p->range;
> +	if (tb[TCA_CT_NAT_IPV4_MIN]) {
> +		range->min_addr.ip =
> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = true;
> +	}
> +	if (tb[TCA_CT_NAT_IPV4_MAX]) {
> +		range->max_addr.ip =
> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = true;
> +	} else if (range->min_addr.ip) {
> +		range->max_addr.ip = range->min_addr.ip;
> +	}
> +
> +	if (tb[TCA_CT_NAT_IPV6_MIN]) {
> +		range->min_addr.in6 =
> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = false;
> +	}
> +	if (tb[TCA_CT_NAT_IPV6_MAX]) {
> +		range->max_addr.in6 =
> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]);
> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
> +		p->ipv4_range = false;
> +	} else if (memchr_inv(&range->min_addr.in6, 0,
> +		   sizeof(range->min_addr.in6))) {
> +		range->max_addr.in6 = range->min_addr.in6;

This will overwrite ipv4_max if it was used, as min/max_addr are
unions.
What about having the _MAX handling (for both ipv4/6) inside the
 if (.._MIN) { }  block ?

> +	}
> +

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-19 18:33   ` Marcelo Ricardo Leitner
@ 2019-06-20  7:07     ` Paul Blakey
  0 siblings, 0 replies; 28+ messages in thread
From: Paul Blakey @ 2019-06-20  7:07 UTC (permalink / raw)
  To: Marcelo Ricardo Leitner
  Cc: Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit


On 6/19/2019 9:33 PM, Marcelo Ricardo Leitner wrote:
> On Tue, Jun 11, 2019 at 04:28:31PM +0300, Paul Blakey wrote:
> ...
>> +static int tcf_ct_fill_params_nat(struct tcf_ct_params *p,
>> +				  struct tc_ct *parm,
>> +				  struct nlattr **tb,
>> +				  struct netlink_ext_ack *extack)
>> +{
>> +	struct nf_nat_range2 *range;
>> +
>> +	if (!(p->ct_action & TCA_CT_ACT_NAT))
>> +		return 0;
>> +
>> +	if (!IS_ENABLED(CONFIG_NF_NAT)) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Netfilter nat isn't enabled in kernel");
>> +		return -EOPNOTSUPP;
>> +	}
>> +
>> +	if (!(p->ct_action & (TCA_CT_ACT_NAT_SRC | TCA_CT_ACT_NAT_DST)))
>> +		return 0;
>> +
>> +	if ((p->ct_action & TCA_CT_ACT_NAT_SRC) &&
>> +	    (p->ct_action & TCA_CT_ACT_NAT_DST)) {
>> +		NL_SET_ERR_MSG_MOD(extack, "dnat and snat can't be enabled at the same time");
>> +		return -EOPNOTSUPP;
>> +	}
>> +
>> +	range = &p->range;
>> +	if (tb[TCA_CT_NAT_IPV4_MIN]) {
>> +		range->min_addr.ip =
>> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = true;
>> +	}
>> +	if (tb[TCA_CT_NAT_IPV4_MAX]) {
>> +		range->max_addr.ip =
>> +			nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = true;
>> +	} else if (range->min_addr.ip) {
>> +		range->max_addr.ip = range->min_addr.ip;
>> +	}
>> +
>> +	if (tb[TCA_CT_NAT_IPV6_MIN]) {
>> +		range->min_addr.in6 =
>> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = false;
>> +	}
>> +	if (tb[TCA_CT_NAT_IPV6_MAX]) {
>> +		range->max_addr.in6 =
>> +			nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]);
>> +		range->flags |= NF_NAT_RANGE_MAP_IPS;
>> +		p->ipv4_range = false;
>> +	} else if (memchr_inv(&range->min_addr.in6, 0,
>> +		   sizeof(range->min_addr.in6))) {
>> +		range->max_addr.in6 = range->min_addr.in6;
> This will overwrite ipv4_max if it was used, as min/max_addr are
> unions.
> What about having the _MAX handling (for both ipv4/6) inside the
>   if (.._MIN) { }  block ?

Yes that what I planned on doing:

         range = &p->range;
         if (tb[TCA_CT_NAT_IPV4_MIN]) {
                 p->ipv4_range = true;
                 range->flags |= NF_NAT_RANGE_MAP_IPS;
                 range->min_addr.ip =
                         nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MIN]);

                 range->max_addr.ip = tb[TCA_CT_NAT_IPV4_MAX] ?
nla_get_in_addr(tb[TCA_CT_NAT_IPV4_MAX]) :
                                      range->min_addr.ip;
         } else if (tb[TCA_CT_NAT_IPV6_MIN]) {
                 p->ipv4_range = false;
                 range->flags |= NF_NAT_RANGE_MAP_IPS;
                 range->min_addr.in6 =
                         nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MIN]);

                 range->max_addr.in6 = tb[TCA_CT_NAT_IPV6_MAX] ?
nla_get_in6_addr(tb[TCA_CT_NAT_IPV6_MAX]) :
                                       range->min_addr.in6;
         }

         if (tb[TCA_CT_NAT_PORT_MIN]) {
                 range->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
                 range->min_proto.all = 
nla_get_be16(tb[TCA_CT_NAT_PORT_MIN]);

                 range->max_proto.all = tb[TCA_CT_NAT_PORT_MAX]?
nla_get_be16(tb[TCA_CT_NAT_PORT_MAX]) :
                                        range->min_proto.all;





>> +	}
>> +

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-18 16:03               ` Cong Wang
@ 2019-06-20  7:32                 ` Paul Blakey
  2019-06-24 17:46                   ` Cong Wang
  0 siblings, 1 reply; 28+ messages in thread
From: Paul Blakey @ 2019-06-20  7:32 UTC (permalink / raw)
  To: Cong Wang, Marcelo Ricardo Leitner
  Cc: Toke Høiland-Jørgensen, Jiri Pirko, Roi Dayan,
	Yossi Kuperman, Oz Shlomo, netdev, David Miller, Aaron Conole,
	Zhike Wang, Rony Efraim, nst-kernel, John Hurley, Simon Horman,
	Justin Pettit, Kevin Darbyshire-Bryant


On 6/18/2019 7:03 PM, Cong Wang wrote:
> On Fri, Jun 14, 2019 at 12:24 PM Marcelo Ricardo Leitner
> <marcelo.leitner@gmail.com> wrote:
>> On Fri, Jun 14, 2019 at 11:07:37AM -0700, Cong Wang wrote:
>>> On Tue, Jun 11, 2019 at 9:44 AM Marcelo Ricardo Leitner
>>> <marcelo.leitner@gmail.com> wrote:
>>>> I had suggested to let act_ct handle the above as well, as there is a
>>>> big chunk of code on both that is pretty similar. There is quite some
>>>> boilerplate for interfacing with conntrack which is duplicated.
>>> Why do you want to mix retrieving conntrack info with executing
>>> conntrack?
>> To save on the heavy boilerplate for interfacing with conntrack.
>>
>>> They are totally different things to me, act_ctinfo merely retrieves
>>> information from conntrack, while this one, act_ct, is supposed to
>>> move packets to conntrack.
>> Seems we have a different understanding for "move packets to
>> conntrack": conntrack will not consume the packets after this.
>> But after act_ct is executed, if not with the clear flag, skb will now
>> have the skb->_nfct entry available, on which flower then will be able
>> to match. So in essence, it is also fetching information from
>> conntrack.
> Interesting. Is it because cls_flower uses conntrack for flow dissection?
> What's the reason behind?
>
> Again, I am still not convinced to do L3 operations in L2, skb->_nfct
> belongs to conntrack which is L3, no matter the packet is consumed
> or not.
>
> Thanks.

I'm not sure what you mean, the reason behind what?

We use conntrack to track, mark the packet with conntrack info, and 
execute nat, then we push the

headers back to continue processing the next action. This action will 
probably be followed by

goto chain or reclassify and then cls_flower can be used to match on 
conntrack state and metadata via the new flow dissector change.








^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [PATCH net-next 1/3] net/sched: Introduce action ct
  2019-06-20  7:32                 ` Paul Blakey
@ 2019-06-24 17:46                   ` Cong Wang
  0 siblings, 0 replies; 28+ messages in thread
From: Cong Wang @ 2019-06-24 17:46 UTC (permalink / raw)
  To: Paul Blakey
  Cc: Marcelo Ricardo Leitner, Toke Høiland-Jørgensen,
	Jiri Pirko, Roi Dayan, Yossi Kuperman, Oz Shlomo, netdev,
	David Miller, Aaron Conole, Zhike Wang, Rony Efraim, nst-kernel,
	John Hurley, Simon Horman, Justin Pettit,
	Kevin Darbyshire-Bryant

On Thu, Jun 20, 2019 at 12:32 AM Paul Blakey <paulb@mellanox.com> wrote:
>
>
> On 6/18/2019 7:03 PM, Cong Wang wrote:
> > On Fri, Jun 14, 2019 at 12:24 PM Marcelo Ricardo Leitner
> > <marcelo.leitner@gmail.com> wrote:
> >> On Fri, Jun 14, 2019 at 11:07:37AM -0700, Cong Wang wrote:
> >>> On Tue, Jun 11, 2019 at 9:44 AM Marcelo Ricardo Leitner
> >>> <marcelo.leitner@gmail.com> wrote:
> >>>> I had suggested to let act_ct handle the above as well, as there is a
> >>>> big chunk of code on both that is pretty similar. There is quite some
> >>>> boilerplate for interfacing with conntrack which is duplicated.
> >>> Why do you want to mix retrieving conntrack info with executing
> >>> conntrack?
> >> To save on the heavy boilerplate for interfacing with conntrack.
> >>
> >>> They are totally different things to me, act_ctinfo merely retrieves
> >>> information from conntrack, while this one, act_ct, is supposed to
> >>> move packets to conntrack.
> >> Seems we have a different understanding for "move packets to
> >> conntrack": conntrack will not consume the packets after this.
> >> But after act_ct is executed, if not with the clear flag, skb will now
> >> have the skb->_nfct entry available, on which flower then will be able
> >> to match. So in essence, it is also fetching information from
> >> conntrack.
> > Interesting. Is it because cls_flower uses conntrack for flow dissection?
> > What's the reason behind?
> >
> > Again, I am still not convinced to do L3 operations in L2, skb->_nfct
> > belongs to conntrack which is L3, no matter the packet is consumed
> > or not.
> >
> > Thanks.
>
> I'm not sure what you mean, the reason behind what?


Yes, which should be the most important info in changelog.


>
> We use conntrack to track, mark the packet with conntrack info, and
> execute nat, then we push the
>
> headers back to continue processing the next action. This action will
> probably be followed by
>
> goto chain or reclassify and then cls_flower can be used to match on
> conntrack state and metadata via the new flow dissector change.
>

Sounds cool, but again why do we have to do this in L2?

Also, I am not sure if cls_flower really matches packets with any
conntrack state, from my quick glance of its code. Is this feature
merged in upstream?

Is this for ingress only? For egress, packets coming down from L3
so they should already have conntrack state as long as it is enabled?

Sorry for asking many questions here, because your changelog is too
short. :-/

Thanks.

^ permalink raw reply	[flat|nested] 28+ messages in thread

end of thread, back to index

Thread overview: 28+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-06-11 13:28 [PATCH net-next 0/3] net/sched: Introduce tc connection tracking Paul Blakey
2019-06-11 13:28 ` [PATCH net-next 1/3] net/sched: Introduce action ct Paul Blakey
2019-06-11 13:59   ` Toke Høiland-Jørgensen
2019-06-11 14:15     ` Paul Blakey
2019-06-11 15:34       ` Toke Høiland-Jørgensen
2019-06-11 15:53         ` Marcelo Ricardo Leitner
2019-06-11 20:23           ` Toke Høiland-Jørgensen
2019-06-13 13:46             ` Paul Blakey
2019-06-13 14:08               ` Toke Høiland-Jørgensen
2019-06-14 18:07           ` Cong Wang
2019-06-14 19:24             ` Marcelo Ricardo Leitner
2019-06-18 16:03               ` Cong Wang
2019-06-20  7:32                 ` Paul Blakey
2019-06-24 17:46                   ` Cong Wang
2019-06-13 11:00       ` Kevin Darbyshire-Bryant
2019-06-12 18:04   ` Marcelo Ricardo Leitner
2019-06-13 13:44     ` Paul Blakey
2019-06-12 19:55   ` Florian Westphal
2019-06-13 13:39     ` Paul Blakey
2019-06-13 10:16   ` Davide Caratti
2019-06-13 10:49     ` Davide Caratti
2019-06-13 13:40       ` Paul Blakey
2019-06-14 18:02   ` Cong Wang
2019-06-16  7:29     ` Paul Blakey
2019-06-19 18:33   ` Marcelo Ricardo Leitner
2019-06-20  7:07     ` Paul Blakey
2019-06-11 13:28 ` [PATCH net-next 2/3] net/flow_dissector: add connection tracking dissection Paul Blakey
2019-06-11 13:28 ` [PATCH net-next 3/3] net/sched: cls_flower: Add matching on conntrack info Paul Blakey

Netdev Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/netdev/0 netdev/git/0.git
	git clone --mirror https://lore.kernel.org/netdev/1 netdev/git/1.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 netdev netdev/ https://lore.kernel.org/netdev \
		netdev@vger.kernel.org netdev@archiver.kernel.org
	public-inbox-index netdev


Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.netdev


AGPL code for this site: git clone https://public-inbox.org/ public-inbox