Skip to contentSkip to conversation

netfilter: nft_tunnel: reinforce key opts support

PATCH In review Xin Long 3 participants 26 messages Dec 8, 2019 – Dec 17, 2019

Subsystems (3)
lore
Share & tools
lore

Xin Long

author
#0 lore raw

cover netfilter: nft_tunnel: reinforce key opts support

This patchset improves quite a few places to make vxlan/erspan opts in nft_tunnel work with userspace nftables/libnftnl, and also keep consistent with the support for vxlan/erspan opts in act_tunnel_key, cls_flower and ip_tunnel_core.

Meanwhile, add support for geneve opts in nft_tunnel. One patch for nftables and one for libnftnl will be posted here for the testing. With them, nft_tunnel can be set and used by:

  # nft add table ip filter
  # nft add chain ip filter input { type filter hook input priority 0 \; }
  # nft add tunnel filter vxlan_01 { type vxlan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"ffff\"\; }
  # nft add tunnel filter erspan_01 { type erspan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:0:0\"\; }
  # nft add tunnel filter erspan_02 { type erspan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"2:0:1:1\"\; }
  # nft add tunnel filter geneve_01 { type geneve\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:1212121234567890\"\; }
  # nft add tunnel filter geneve_02 { type geneve\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:34567890,2:2:12121212,3:3:1212121234567890\"\; }
  # nft list tunnels table filter
  # nft add rule filter input ip protocol udp tunnel name geneve_02
  # nft add rule filter input meta l4proto udp tunnel id 2 drop
  # nft add rule filter input meta l4proto udp tunnel path 0 drop
  # nft list chain filter input -a

Xin Long (7):
netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8
netfilter: nft_tunnel: parse VXLAN_GBP attr as u32 in nft_tunnel
netfilter: nft_tunnel: no need to call htons() when dumping ports
netfilter: nft_tunnel: also dump ERSPAN_VERSION
netfilter: nft_tunnel: also dump OPTS_ERSPAN/VXLAN
netfilter: nft_tunnel: add the missing nla_nest_cancel()
netfilter: nft_tunnel: add support for geneve opts

 include/uapi/linux/netfilter/nf_tables.h |  10 ++
 net/netfilter/nft_tunnel.c               | 170 +++++++++++++++++++++++++------
 2 files changed, 151 insertions(+), 29 deletions(-)
Signature · 1 line
2.1.0
author
#1 lore raw

PATCH 1/7 netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +3 -2
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+	[NFTA_TUNNEL_KEY_ERSPAN_VERSION]	= { .type = NLA_U8 },
 	[NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]	= { .type = NLA_U32 },
-	[NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]	= { .type = NLA_U8 },
+	[NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]		= { .type = NLA_U8 },
 	[NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]	= { .type = NLA_U8 },
 };
 
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
 	if (err < 0)
 		return err;
 
-	version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+	version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);
 	switch (version) {
 	case ERSPAN_VERSION:
 		if (!tb[NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX])
Signature · 1 line
2.1.0
author
#2 lore raw

PATCH 2/7 netfilter: nft_tunnel: parse VXLAN_GBP attr as u32 in nft_tunnel

Both user and kernel sides want VXLAN_GBP opt as u32, so there's no need to convert it on each side.

---
 net/netfilter/nft_tunnel.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
net/netfilter/nft_tunnel.c +3 -3
index f76cd7d..d9d6c0d 100644
@@ -239,7 +239,7 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 	if (!tb[NFTA_TUNNEL_KEY_VXLAN_GBP])
 		return -EINVAL;
 
-	opts->u.vxlan.gbp = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_VXLAN_GBP]));
+	opts->u.vxlan.gbp = nla_get_u32(tb[NFTA_TUNNEL_KEY_VXLAN_GBP]);
 
 	opts->len	= sizeof(struct vxlan_metadata);
 	opts->flags	= TUNNEL_VXLAN_OPT;
@@ -475,8 +475,8 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 		return -1;
 
 	if (opts->flags & TUNNEL_VXLAN_OPT) {
-		if (nla_put_be32(skb, NFTA_TUNNEL_KEY_VXLAN_GBP,
-				 htonl(opts->u.vxlan.gbp)))
+		if (nla_put_u32(skb, NFTA_TUNNEL_KEY_VXLAN_GBP,
+				opts->u.vxlan.gbp))
 			return -1;
 	} else if (opts->flags & TUNNEL_ERSPAN_OPT) {
 		switch (opts->u.erspan.version) {
Signature · 1 line
2.1.0
author
#3 lore raw

PATCH 3/7 netfilter: nft_tunnel: no need to call htons() when dumping ports

info->key.tp_src and tp_dst are __be16, when using nla_put_be16() to dump them, htons() is not needed, so remove it in this patch.

---
 net/netfilter/nft_tunnel.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -2
index d9d6c0d..e1184fa 100644
@@ -502,8 +502,8 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 static int nft_tunnel_ports_dump(struct sk_buff *skb,
 				 struct ip_tunnel_info *info)
 {
-	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, htons(info->key.tp_src)) < 0 ||
-	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, htons(info->key.tp_dst)) < 0)
+	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, info->key.tp_src) < 0 ||
+	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, info->key.tp_dst) < 0)
 		return -1;
 
 	return 0;
Signature · 1 line
2.1.0
author
#4 lore raw

PATCH 4/7 netfilter: nft_tunnel: also dump ERSPAN_VERSION

This is not necessary, but it'll be easier to parse in userspace, also given that other places like act_tunnel_key, cls_flower and ip_tunnel_core are also doing so.

---
 net/netfilter/nft_tunnel.c | 3 +++
 1 file changed, 3 insertions(+)
net/netfilter/nft_tunnel.c +3 -0
index e1184fa..576437f 100644
@@ -479,6 +479,9 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 				opts->u.vxlan.gbp))
 			return -1;
 	} else if (opts->flags & TUNNEL_ERSPAN_OPT) {
+		if (nla_put_u8(skb, NFTA_TUNNEL_KEY_ERSPAN_VERSION,
+			       opts->u.erspan.version))
+			return -1;
 		switch (opts->u.erspan.version) {
 		case ERSPAN_VERSION:
 			if (nla_put_be32(skb, NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX,
Signature · 1 line
2.1.0
author
#5 lore raw

PATCH 5/7 netfilter: nft_tunnel: also dump OPTS_ERSPAN/VXLAN

This patch is to add the nest attr OPTS_ERSPAN/VXLAN when dumping KEY_OPTS, and it would be helpful when parsing in userpace. Also, this is needed for supporting multiple geneve opts in the future patches.

---
 net/netfilter/nft_tunnel.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
net/netfilter/nft_tunnel.c +9 -1
index 576437f..e9b94b8 100644
@@ -468,17 +468,24 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 				struct nft_tunnel_obj *priv)
 {
 	struct nft_tunnel_opts *opts = &priv->opts;
-	struct nlattr *nest;
+	struct nlattr *nest, *inner;
 
 	nest = nla_nest_start_noflag(skb, NFTA_TUNNEL_KEY_OPTS);
 	if (!nest)
 		return -1;
 
 	if (opts->flags & TUNNEL_VXLAN_OPT) {
+		inner = nla_nest_start_noflag(skb, NFTA_TUNNEL_KEY_OPTS_VXLAN);
+		if (!inner)
+			return -1;
 		if (nla_put_u32(skb, NFTA_TUNNEL_KEY_VXLAN_GBP,
 				opts->u.vxlan.gbp))
 			return -1;
+		nla_nest_end(skb, inner);
 	} else if (opts->flags & TUNNEL_ERSPAN_OPT) {
+		inner = nla_nest_start_noflag(skb, NFTA_TUNNEL_KEY_OPTS_ERSPAN);
+		if (!inner)
+			return -1;
 		if (nla_put_u8(skb, NFTA_TUNNEL_KEY_ERSPAN_VERSION,
 			       opts->u.erspan.version))
 			return -1;
@@ -496,6 +503,7 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 				return -1;
 			break;
 		}
+		nla_nest_end(skb, inner);
 	}
 	nla_nest_end(skb, nest);
 
Signature · 1 line
2.1.0
author
#6 lore raw

PATCH 6/7 netfilter: nft_tunnel: add the missing nla_nest_cancel()

When nla_put_xxx() fails under nla_nest_start_noflag(), nla_nest_cancel() should be called, so that the skb can be trimmed properly.

---
 net/netfilter/nft_tunnel.c | 38 ++++++++++++++++++++++++++------------
 1 file changed, 26 insertions(+), 12 deletions(-)
net/netfilter/nft_tunnel.c +26 -12
index e9b94b8..32263dc 100644
@@ -443,10 +443,15 @@ static int nft_tunnel_ip_dump(struct sk_buff *skb, struct ip_tunnel_info *info)
 		if (!nest)
 			return -1;
 
-		if (nla_put_in6_addr(skb, NFTA_TUNNEL_KEY_IP6_SRC, &info->key.u.ipv6.src) < 0 ||
-		    nla_put_in6_addr(skb, NFTA_TUNNEL_KEY_IP6_DST, &info->key.u.ipv6.dst) < 0 ||
-		    nla_put_be32(skb, NFTA_TUNNEL_KEY_IP6_FLOWLABEL, info->key.label))
+		if (nla_put_in6_addr(skb, NFTA_TUNNEL_KEY_IP6_SRC,
+				     &info->key.u.ipv6.src) < 0 ||
+		    nla_put_in6_addr(skb, NFTA_TUNNEL_KEY_IP6_DST,
+				     &info->key.u.ipv6.dst) < 0 ||
+		    nla_put_be32(skb, NFTA_TUNNEL_KEY_IP6_FLOWLABEL,
+				 info->key.label)) {
+			nla_nest_cancel(skb, nest);
 			return -1;
+		}
 
 		nla_nest_end(skb, nest);
 	} else {
@@ -454,9 +459,13 @@ static int nft_tunnel_ip_dump(struct sk_buff *skb, struct ip_tunnel_info *info)
 		if (!nest)
 			return -1;
 
-		if (nla_put_in_addr(skb, NFTA_TUNNEL_KEY_IP_SRC, info->key.u.ipv4.src) < 0 ||
-		    nla_put_in_addr(skb, NFTA_TUNNEL_KEY_IP_DST, info->key.u.ipv4.dst) < 0)
+		if (nla_put_in_addr(skb, NFTA_TUNNEL_KEY_IP_SRC,
+				    info->key.u.ipv4.src) < 0 ||
+		    nla_put_in_addr(skb, NFTA_TUNNEL_KEY_IP_DST,
+				    info->key.u.ipv4.dst) < 0) {
+			nla_nest_cancel(skb, nest);
 			return -1;
+		}
 
 		nla_nest_end(skb, nest);
 	}
@@ -477,37 +486,42 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 	if (opts->flags & TUNNEL_VXLAN_OPT) {
 		inner = nla_nest_start_noflag(skb, NFTA_TUNNEL_KEY_OPTS_VXLAN);
 		if (!inner)
-			return -1;
+			goto failure;
 		if (nla_put_u32(skb, NFTA_TUNNEL_KEY_VXLAN_GBP,
 				opts->u.vxlan.gbp))
-			return -1;
+			goto inner_failure;
 		nla_nest_end(skb, inner);
 	} else if (opts->flags & TUNNEL_ERSPAN_OPT) {
 		inner = nla_nest_start_noflag(skb, NFTA_TUNNEL_KEY_OPTS_ERSPAN);
 		if (!inner)
-			return -1;
+			goto failure;
 		if (nla_put_u8(skb, NFTA_TUNNEL_KEY_ERSPAN_VERSION,
 			       opts->u.erspan.version))
-			return -1;
+			goto inner_failure;
 		switch (opts->u.erspan.version) {
 		case ERSPAN_VERSION:
 			if (nla_put_be32(skb, NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX,
 					 opts->u.erspan.u.index))
-				return -1;
+				goto inner_failure;
 			break;
 		case ERSPAN_VERSION2:
 			if (nla_put_u8(skb, NFTA_TUNNEL_KEY_ERSPAN_V2_HWID,
 				       get_hwid(&opts->u.erspan.u.md2)) ||
 			    nla_put_u8(skb, NFTA_TUNNEL_KEY_ERSPAN_V2_DIR,
 				       opts->u.erspan.u.md2.dir))
-				return -1;
+				goto inner_failure;
 			break;
 		}
 		nla_nest_end(skb, inner);
 	}
 	nla_nest_end(skb, nest);
-
 	return 0;
+
+inner_failure:
+	nla_nest_cancel(skb, inner);
+failure:
+	nla_nest_cancel(skb, nest);
+	return -1;
 }
 
 static int nft_tunnel_ports_dump(struct sk_buff *skb,
Signature · 1 line
2.1.0
author
#7 lore raw

PATCH 7/7 netfilter: nft_tunnel: add support for geneve opts

Like vxlan and erspan opts, geneve opts should also be supported in nft_tunnel. The difference is geneve RFC (draft-ietf-nvo3-geneve-14) allows a geneve packet to carry multiple geneve opts. So with this patch, nftables/libnftnl would do:

  # nft add table ip filter
  # nft add chain ip filter input { type filter hook input priority 0 \; }
  # nft add tunnel filter geneve_02 { type geneve\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:34567890,2:2:12121212,3:3:1212121234567890\"\; }
  # nft list tunnels table filter
    table ip filter {
    	tunnel geneve_02 {
    		id 2
    		ip saddr 192.168.1.1
    		ip daddr 192.168.1.2
    		sport 9000
    		dport 9001
    		tos 18
    		ttl 64
    		flags 1
    		geneve opts 1:1:34567890,2:2:12121212,3:3:1212121234567890
    	}
    }
---
 include/uapi/linux/netfilter/nf_tables.h |  10 +++
 net/netfilter/nft_tunnel.c               | 110 +++++++++++++++++++++++++++----
 2 files changed, 108 insertions(+), 12 deletions(-)
include/uapi/linux/netfilter/nf_tables.h +10 -0
index bb9b049..f74b957 100644
@@ -1729,6 +1729,7 @@ enum nft_tunnel_opts_attributes {
 	NFTA_TUNNEL_KEY_OPTS_UNSPEC,
 	NFTA_TUNNEL_KEY_OPTS_VXLAN,
 	NFTA_TUNNEL_KEY_OPTS_ERSPAN,
+	NFTA_TUNNEL_KEY_OPTS_GENEVE,
 	__NFTA_TUNNEL_KEY_OPTS_MAX
 };
 #define NFTA_TUNNEL_KEY_OPTS_MAX	(__NFTA_TUNNEL_KEY_OPTS_MAX - 1)
@@ -1750,6 +1751,15 @@ enum nft_tunnel_opts_erspan_attributes {
 };
 #define NFTA_TUNNEL_KEY_ERSPAN_MAX	(__NFTA_TUNNEL_KEY_ERSPAN_MAX - 1)
 
+enum nft_tunnel_opts_geneve_attributes {
+	NFTA_TUNNEL_KEY_GENEVE_UNSPEC,
+	NFTA_TUNNEL_KEY_GENEVE_CLASS,
+	NFTA_TUNNEL_KEY_GENEVE_TYPE,
+	NFTA_TUNNEL_KEY_GENEVE_DATA,
+	__NFTA_TUNNEL_KEY_GENEVE_MAX
+};
+#define NFTA_TUNNEL_KEY_GENEVE_MAX	(__NFTA_TUNNEL_KEY_GENEVE_MAX - 1)
+
 enum nft_tunnel_flags {
 	NFT_TUNNEL_F_ZERO_CSUM_TX	= (1 << 0),
 	NFT_TUNNEL_F_DONT_FRAGMENT	= (1 << 1),
net/netfilter/nft_tunnel.c +98 -12

147 lines not on this page · raw message

Signature · 1 line
2.1.0
author
#23 lore raw
In reply to Xin Long

The 1st patch file for libnftnl is
Based on:
git://git.netfilter.org/libnftnl
The 2nd patch file for nftables is:

    Based on:
      git://git.netfilter.org/nftables: d42e9a1 (tag: v0.9.2) build:
Bump version to v0.9.2
      plusing:
        https://patchwork.ozlabs.org/patch/1136755/
        https://patchwork.ozlabs.org/patch/1136756/
    Dependencies:
       Some fixes in libnftnl, which are in another patch.
Quoted 47 lines · Xin Long

This patchset improves quite a few places to make vxlan/erspan opts in nft_tunnel work with userspace nftables/libnftnl, and also keep consistent with the support for vxlan/erspan opts in act_tunnel_key, cls_flower and ip_tunnel_core.

Meanwhile, add support for geneve opts in nft_tunnel. One patch for nftables and one for libnftnl will be posted here for the testing. With them, nft_tunnel can be set and used by:

  # nft add table ip filter
  # nft add chain ip filter input { type filter hook input priority 0 \; }
  # nft add tunnel filter vxlan_01 { type vxlan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"ffff\"\; }
  # nft add tunnel filter erspan_01 { type erspan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:0:0\"\; }
  # nft add tunnel filter erspan_02 { type erspan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"2:0:1:1\"\; }
  # nft add tunnel filter geneve_01 { type geneve\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:1212121234567890\"\; }
  # nft add tunnel filter geneve_02 { type geneve\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:34567890,2:2:12121212,3:3:1212121234567890\"\; }
  # nft list tunnels table filter
  # nft add rule filter input ip protocol udp tunnel name geneve_02
  # nft add rule filter input meta l4proto udp tunnel id 2 drop
  # nft add rule filter input meta l4proto udp tunnel path 0 drop
  # nft list chain filter input -a

Xin Long (7):
netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8
netfilter: nft_tunnel: parse VXLAN_GBP attr as u32 in nft_tunnel
netfilter: nft_tunnel: no need to call htons() when dumping ports
netfilter: nft_tunnel: also dump ERSPAN_VERSION
netfilter: nft_tunnel: also dump OPTS_ERSPAN/VXLAN
netfilter: nft_tunnel: add the missing nla_nest_cancel()
netfilter: nft_tunnel: add support for geneve opts

Quoted 7 lines
 include/uapi/linux/netfilter/nf_tables.h |  10 ++
 net/netfilter/nft_tunnel.c               | 170 +++++++++++++++++++++++++------
 2 files changed, 151 insertions(+), 29 deletions(-)

--
2.1.0

From 2456ad53c66489d6f2d4aac7d95216d553f94b83 Mon Sep 17 00:00:00 2001
From: Xin Long <[email protected]>
Date: Fri, 6 Dec 2019 04:31:10 -0500
Subject: [PATCH] libnftnl: nft_tunnel: add support for parsing geneve opts

---
 include/libnftnl/object.h           |   3 +
 include/linux/netfilter/nf_tables.h |  10 +++
 include/obj.h                       |  13 +++-
 src/libnftnl.map                    |   1 +
 src/obj/tunnel.c                    | 109 ++++++++++++++++++++++++++--
 src/object.c                        |   4 +-
 6 files changed, 129 insertions(+), 11 deletions(-)
include/libnftnl/object.h +3 -0
index 221b15c..e17b8d5 100644
@@ -109,6 +109,9 @@ enum {
 	NFTNL_OBJ_TUNNEL_ERSPAN_V1_INDEX,
 	NFTNL_OBJ_TUNNEL_ERSPAN_V2_HWID,
 	NFTNL_OBJ_TUNNEL_ERSPAN_V2_DIR,
+	NFTNL_OBJ_TUNNEL_GENEVE_CLASS,
+	NFTNL_OBJ_TUNNEL_GENEVE_TYPE,
+	NFTNL_OBJ_TUNNEL_GENEVE_DATA,
 };
 
 enum {
include/linux/netfilter/nf_tables.h +10 -0
index bb9b049..f74b957 100644
@@ -1729,6 +1729,7 @@ enum nft_tunnel_opts_attributes {
 	NFTA_TUNNEL_KEY_OPTS_UNSPEC,
 	NFTA_TUNNEL_KEY_OPTS_VXLAN,
 	NFTA_TUNNEL_KEY_OPTS_ERSPAN,
+	NFTA_TUNNEL_KEY_OPTS_GENEVE,
 	__NFTA_TUNNEL_KEY_OPTS_MAX
 };
 #define NFTA_TUNNEL_KEY_OPTS_MAX	(__NFTA_TUNNEL_KEY_OPTS_MAX - 1)
@@ -1750,6 +1751,15 @@ enum nft_tunnel_opts_erspan_attributes {
 };
 #define NFTA_TUNNEL_KEY_ERSPAN_MAX	(__NFTA_TUNNEL_KEY_ERSPAN_MAX - 1)
 
+enum nft_tunnel_opts_geneve_attributes {
+	NFTA_TUNNEL_KEY_GENEVE_UNSPEC,
+	NFTA_TUNNEL_KEY_GENEVE_CLASS,
+	NFTA_TUNNEL_KEY_GENEVE_TYPE,
+	NFTA_TUNNEL_KEY_GENEVE_DATA,
+	__NFTA_TUNNEL_KEY_GENEVE_MAX
+};
+#define NFTA_TUNNEL_KEY_GENEVE_MAX	(__NFTA_TUNNEL_KEY_GENEVE_MAX - 1)
+
 enum nft_tunnel_flags {
 	NFT_TUNNEL_F_ZERO_CSUM_TX	= (1 << 0),
 	NFT_TUNNEL_F_DONT_FRAGMENT	= (1 << 1),
include/obj.h +11 -2
index 10f806c..eb77445 100644
@@ -19,7 +19,7 @@ struct nftnl_obj {
 	uint32_t		family;
 	uint32_t		use;
 
-	uint32_t		flags;
+	uint64_t		flags;
 	uint64_t		handle;
 
 	union {
@@ -78,7 +78,7 @@ struct nftnl_obj {
 					uint32_t	gbp;
 				} tun_vxlan;
 				struct {
-					uint32_t	version;
+					uint8_t	version;
 					union {
 						uint32_t	v1_index;
 						struct {
@@ -87,6 +87,15 @@ struct nftnl_obj {
 						} v2;
 					} u;
 				} tun_erspan;
+				struct {
+					struct geneve_opt {
+						uint16_t	class;
+						uint8_t		type;
+						uint8_t		*data;
+						uint8_t		data_len;
+					}	opts[64];
+					uint8_t		opt_num;
+				} tun_geneve;
 			} u;
 		} tunnel;
 		struct nftnl_obj_secmark {
src/libnftnl.map +1 -0
index 8230d15..3b1b891 100644
@@ -357,5 +357,6 @@ LIBNFTNL_12 {
 LIBNFTNL_13 {
   nftnl_set_list_lookup_byname;
   nftnl_obj_set_data;
+  nftnl_obj_get_data;
   nftnl_flowtable_set_data;
 } LIBNFTNL_12;
src/obj/tunnel.c +102 -7

187 lines not on this page · raw message

src/object.c +2 -2
index c876add..09c635a 100644
@@ -109,7 +109,7 @@ void nftnl_obj_set_data(struct nftnl_obj *obj, uint16_t attr,
 			obj->ops->set(obj, attr, data, data_len);
 		break;
 	}
-	obj->flags |= (1 << attr);
+	obj->flags |= (1lu << attr);
 }
 
 void nftnl_obj_set(struct nftnl_obj *obj, uint16_t attr, const void *data) __visible;
@@ -152,7 +152,7 @@ EXPORT_SYMBOL(nftnl_obj_get_data);
 const void *nftnl_obj_get_data(struct nftnl_obj *obj, uint16_t attr,
 			       uint32_t *data_len)
 {
-	if (!(obj->flags & (1 << attr)))
+	if (!(obj->flags & (1lu << attr)))
 		return NULL;
 
 	switch(attr) {
Signature · 1 line
2.18.1

From ac451f621050fcb80f023084fb890363f8646776 Mon Sep 17 00:00:00 2001
From: Xin Long <[email protected]>
Date: Fri, 6 Dec 2019 04:19:53 -0500
Subject: [PATCH] nftables: nft_tunnel: add support for dumping
vxlan/erspan/geneve opts

Based on:

git://git.netfilter.org/nftables: d42e9a1 (tag: v0.9.2) build: Bump version to v0.9.2 plusing:

    https://patchwork.ozlabs.org/patch/1136755/
    https://patchwork.ozlabs.org/patch/1136756/

Dependencies:

Some fixes in libnftnl, which are in another patch.

Examples:

  # nft add table ip filter
  # nft add chain ip filter input { type filter hook input priority 0 \; }
  # nft add tunnel filter vxlan_01  { type vxlan\;  id 2\; ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; opts \"ffff\"\; }
  # nft add tunnel filter erspan_01 { type erspan\; id 2\; ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; opts \"1:1:0:0\"\; }
  # nft add tunnel filter erspan_02 { type erspan\; id 2\; ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; opts \"2:0:1:1\"\; }
  # nft add tunnel filter geneve_01 { type geneve\; id 2\; ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; opts \"1:1:1212121234567890\"\; }
  # nft add tunnel filter geneve_02 { type geneve\; id 2\; ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; opts \"1:1:1212121234567890,2:2:1212121234567890,3:3:1212121234567890\"\; }
  # nft list tunnels table filter
  # nft add rule filter input ip protocol udp tunnel name geneve_02
  # nft add rule filter input meta l4proto udp tunnel id 2 drop
  # nft add rule filter input meta l4proto udp tunnel path 0 drop
  # nft list chain filter input -a
---
 include/rule.h     |   3 +-
 src/evaluate.c     |   4 ++
 src/json.c         |   4 ++
 src/mnl.c          | 108 +++++++++++++++++++++++++++++++++++++++++++++
 src/netlink.c      |  72 ++++++++++++++++++++++++++++++
 src/parser_bison.y |  27 +++++++++++-
 src/rule.c         |  21 +++++++++
 src/scanner.l      |   1 +
 8 files changed, 237 insertions(+), 3 deletions(-)
include/rule.h +2 -1
index cc221d6..6241760 100644
@@ -408,11 +408,12 @@ struct tunnel {
 	uint32_t	id;
 	struct expr	*src;
 	struct expr	*dst;
-	uint32_t	label;
+	uint32_t	flags;
 	uint16_t	sport;
 	uint16_t	dport;
 	uint8_t		tos;
 	uint8_t		ttl;
+	const char	*opts;
 };
 
 /**
src/evaluate.c +4 -0
index 62c5f52..74fa739 100755
@@ -3717,6 +3717,7 @@ static int cmd_evaluate_delete(struct eval_ctx *ctx, struct cmd *cmd)
 	case CMD_OBJ_CT_HELPER:
 	case CMD_OBJ_CT_TIMEOUT:
 	case CMD_OBJ_LIMIT:
+	case CMD_OBJ_TUNNEL:
 	case CMD_OBJ_SECMARK:
 	case CMD_OBJ_CT_EXPECT:
 		return 0;
@@ -3860,6 +3861,8 @@ static int cmd_evaluate_list(struct eval_ctx *ctx, struct cmd *cmd)
 		return cmd_evaluate_list_obj(ctx, cmd, NFT_OBJECT_CT_TIMEOUT);
 	case CMD_OBJ_LIMIT:
 		return cmd_evaluate_list_obj(ctx, cmd, NFT_OBJECT_LIMIT);
+	case CMD_OBJ_TUNNEL:
+		return cmd_evaluate_list_obj(ctx, cmd, NFT_OBJECT_TUNNEL);
 	case CMD_OBJ_SECMARK:
 		return cmd_evaluate_list_obj(ctx, cmd, NFT_OBJECT_SECMARK);
 	case CMD_OBJ_CT_EXPECT:
@@ -3868,6 +3871,7 @@ static int cmd_evaluate_list(struct eval_ctx *ctx, struct cmd *cmd)
 	case CMD_OBJ_QUOTAS:
 	case CMD_OBJ_CT_HELPERS:
 	case CMD_OBJ_LIMITS:
+	case CMD_OBJ_TUNNELS:
 	case CMD_OBJ_SETS:
 	case CMD_OBJ_FLOWTABLES:
 	case CMD_OBJ_SECMARKS:
src/json.c +4 -0
index 55ce053..3ad7aba 100644
@@ -1772,6 +1772,10 @@ int do_command_list_json(struct netlink_ctx *ctx, struct cmd *cmd)
 	case CMD_OBJ_COUNTERS:
 		root = do_list_obj_json(ctx, cmd, NFT_OBJECT_COUNTER);
 		break;
+	case CMD_OBJ_TUNNEL:
+	case CMD_OBJ_TUNNELS:
+		root = do_list_obj_json(ctx, cmd, NFT_OBJECT_TUNNEL);
+		break;
 	case CMD_OBJ_QUOTA:
 	case CMD_OBJ_QUOTAS:
 		root = do_list_obj_json(ctx, cmd, NFT_OBJECT_QUOTA);
src/mnl.c +108 -0

126 lines not on this page · raw message

src/netlink.c +72 -0

90 lines not on this page · raw message

src/parser_bison.y +25 -2
index 4eade0e..5429fd9 100644
@@ -546,6 +546,7 @@ int nft_lex(void *, void *, void *);
 
 %token IN			"in"
 %token OUT			"out"
+%token OPTS			"opts"
 
 %type <string>			identifier type_identifier string comment_spec
 %destructor { xfree($$); }	identifier type_identifier string comment_spec
@@ -1016,7 +1017,7 @@ add_cmd			:	TABLE		table_spec
 			}
 			|	TUNNEL		obj_spec	tunnel_obj	'{' tunnel_block '}'	stmt_separator
 			{
-				$$ = cmd_alloc_obj_ct(CMD_ADD, CMD_OBJ_TUNNEL, &$2, &@$, $3);
+				$$ = cmd_alloc(CMD_ADD, CMD_OBJ_TUNNEL, &$2, &@$, $3);
 			}
 			;
 
@@ -1113,7 +1114,7 @@ create_cmd		:	TABLE		table_spec
 			}
 			|	TUNNEL		obj_spec	tunnel_obj	'{' tunnel_block '}'	stmt_separator
 			{
-				$$ = cmd_alloc_obj_ct(CMD_CREATE, CMD_OBJ_TUNNEL, &$2, &@$, $3);
+				$$ = cmd_alloc(CMD_CREATE, CMD_OBJ_TUNNEL, &$2, &@$, $3);
 			}
 			;
 
@@ -1199,6 +1200,14 @@ delete_cmd		:	TABLE		table_spec
 			{
 				$$ = cmd_alloc(CMD_DELETE, CMD_OBJ_SECMARK, &$2, &@$, NULL);
 			}
+			| 	TUNNEL		obj_spec
+			{
+				$$ = cmd_alloc(CMD_DELETE, CMD_OBJ_TUNNEL, &$2, &@$, NULL);
+			}
+			| 	TUNNEL 		objid_spec
+			{
+				$$ = cmd_alloc(CMD_DELETE, CMD_OBJ_TUNNEL, &$2, &@$, NULL);
+			}
 			;
 
 get_cmd			:	ELEMENT		set_spec	set_block_expr
@@ -1271,6 +1280,18 @@ list_cmd		:	TABLE		table_spec
 			{
 				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_LIMIT, &$2, &@$, NULL);
 			}
+			|	TUNNELS		ruleset_spec
+			{
+				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_TUNNELS, &$2, &@$, NULL);
+			}
+			|	TUNNELS		TABLE	table_spec
+			{
+				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_TUNNELS, &$3, &@$, NULL);
+			}
+			|	TUNNEL		obj_spec
+			{
+				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_TUNNEL, &$2, &@$, NULL);
+			}
 			|	SECMARKS	ruleset_spec
 			{
 				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_SECMARKS, &$2, &@$, NULL);
@@ -3774,6 +3795,8 @@ tunnel_config		:	TYPE	string		stmt_separator	{	$<obj>0->tunnel.type = $2;	}
 			|	DPORT	NUM		stmt_separator	{	$<obj>0->tunnel.dport = $2;	}
 			|	DSCP	NUM		stmt_separator	{	$<obj>0->tunnel.tos = $2 << 2;	}
 			|	TTL	NUM		stmt_separator	{	$<obj>0->tunnel.ttl = $2;	}
+			|	FLAGS	NUM		stmt_separator	{	$<obj>0->tunnel.flags = $2;	}
+			|	OPTS	string		stmt_separator	{	$<obj>0->tunnel.opts = $2;	}
 			;
 
 tunnel_block		:	/* empty */	{ $$ = $<obj>-1; }
src/rule.c +21 -0
index 3b52f63..a50bc85 100644
@@ -1921,10 +1921,31 @@ static void obj_print_data(const struct obj *obj,
 		nft_print(octx, "%s%sid %u%s",
 			  opts->tab, opts->tab, obj->tunnel.id, opts->nl);
 
+		if (obj->tunnel.src) {
+			nft_print(octx, "%s%sip saddr ",
+				  opts->tab, opts->tab);
+			expr_print(obj->tunnel.src, octx);
+			nft_print(octx, "%s", opts->nl);
+		}
 		if (obj->tunnel.dst) {
 			nft_print(octx, "%s%sip daddr ",
 				  opts->tab, opts->tab);
 			expr_print(obj->tunnel.dst, octx);
+			nft_print(octx, "%s", opts->nl);
+		}
+		nft_print(octx, "%s%ssport %d%s", opts->tab, opts->tab,
+			  obj->tunnel.sport, opts->nl);
+		nft_print(octx, "%s%sdport %d%s", opts->tab, opts->tab,
+			  obj->tunnel.dport, opts->nl);
+		nft_print(octx, "%s%stos %d%s", opts->tab, opts->tab,
+			  obj->tunnel.tos, opts->nl);
+		nft_print(octx, "%s%sttl %d%s", opts->tab, opts->tab,
+			  obj->tunnel.ttl, opts->nl);
+		nft_print(octx, "%s%sflags %x%s", opts->tab, opts->tab,
+			  obj->tunnel.flags, opts->nl);
+		if (obj->tunnel.opts) {
+			nft_print(octx, "%s%s%s opts %s%s", opts->tab, opts->tab,
+				  obj->tunnel.type, obj->tunnel.opts, opts->nl);
 		}
 		break;
 	default:
src/scanner.l +1 -0
index 71bf60e..72733a1 100644
@@ -388,6 +388,7 @@ addrstring	({macaddr}|{ip4addr}|{ip6addr})
 "saddr"			{ return SADDR; }
 "daddr"			{ return DADDR; }
 "type"			{ return TYPE; }
+"opts"			{ return OPTS; }
 
 "vlan"			{ return VLAN; }
 "id"			{ return ID; }
Signature · 1 line
2.18.1
#16 lore raw

Re: [PATCH nf-next 1/7] netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

In reply to Xin Long

Hi Xin,

Quoted 25 lines · Xin Long

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -1
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+	[NFTA_TUNNEL_KEY_ERSPAN_VERSION]	= { .type = NLA_U8 },
 	[NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]	= { .type = NLA_U32 },
-	[NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]	= { .type = NLA_U8 },
+	[NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]		= { .type = NLA_U8 },
 	[NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]	= { .type = NLA_U8 },
 };
 

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_erspan_init(const struct nlattr *attr, · line 270

Quoted 5 lines
net/netfilter/nft_tunnel.c +1 -1
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
 	if (err < 0)
 		return err;
 
-	version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+	version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);

I have concerns about this change and backwards-compatibility with existing users of this UAPI. Likewise, with other changes to the encoding of existing attributes elsewhere in this series.

Quoted 6 lines
net/netfilter/nft_tunnel.c +0 -0
 	switch (version) {
 	case ERSPAN_VERSION:
 		if (!tb[NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX])
Signature · 1 line
2.1.0
author
#17 lore raw

Re: [PATCH nf-next 1/7] netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

Quoted 32 lines · Simon Horman

Hi Xin,

Quoted 28 lines · Xin Long

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -1
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+     [NFTA_TUNNEL_KEY_ERSPAN_VERSION]        = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]       = { .type = NLA_U32 },
-     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR] = { .type = NLA_U8 },
+     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]         = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]        = { .type = NLA_U8 },
 };
 
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
      if (err < 0)
              return err;
 
Quoted 6 lines

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_erspan_init(const struct nlattr *attr, · line 270

Quoted 2 lines
net/netfilter/nft_tunnel.c +1 -1
-     version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+     version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);

I have concerns about this change and backwards-compatibility with existing users of this UAPI. Likewise, with other changes to the encoding of existing attributes elsewhere in this series.

userspace(nftables/libnftnl) is not ready for nft_tunnel, I don't think there will be
any backwards-compatibility issue.

Pablo?

Quoted 7 lines
Quoted 7 lines
net/netfilter/nft_tunnel.c +0 -1
      switch (version) {
      case ERSPAN_VERSION:
              if (!tb[NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX])
--

2.1.0

#20 lore raw

Re: [PATCH nf-next 1/7] netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

In reply to Xin Long

Hi,

Quoted 25 lines · Xin Long

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -1
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+	[NFTA_TUNNEL_KEY_ERSPAN_VERSION]	= { .type = NLA_U8 },
 	[NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]	= { .type = NLA_U32 },
-	[NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]	= { .type = NLA_U8 },
+	[NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]		= { .type = NLA_U8 },
 	[NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]	= { .type = NLA_U8 },
 };
 

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_erspan_init(const struct nlattr *attr, · line 270

Quoted 5 lines
net/netfilter/nft_tunnel.c +1 -1
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
 	if (err < 0)
 		return err;
 
-	version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+	version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);

I think NFTA_TUNNEL_KEY_ERSPAN_VERSION as 32-bit is just fine.

Netlink will be adding the padding anyway for u8.

I would suggest you leave this as is.

Thanks.

#15 lore raw

Re: [PATCH nf-next 2/7] netfilter: nft_tunnel: parse VXLAN_GBP attr as u32 in nft_tunnel

In reply to Xin Long
Quoted 13 lines · Xin Long

Both user and kernel sides want VXLAN_GBP opt as u32, so there's no need to convert it on each side.

---
 net/netfilter/nft_tunnel.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
net/netfilter/nft_tunnel.c +0 -0
index f76cd7d..d9d6c0d 100644

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr, · line 242

Quoted 5 lines
net/netfilter/nft_tunnel.c +1 -1
@@ -239,7 +239,7 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 	if (!tb[NFTA_TUNNEL_KEY_VXLAN_GBP])
 		return -EINVAL;
 
-	opts->u.vxlan.gbp = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_VXLAN_GBP]));
+	opts->u.vxlan.gbp = nla_get_u32(tb[NFTA_TUNNEL_KEY_VXLAN_GBP]);

In netfilter, attributes go in network byte order to make it easier to send them over the wire. The only remaining part that needs to be converted to network byte order is the netlink header.

Please, leave this one as is.

#11 lore raw

Re: [PATCH nf-next 3/7] netfilter: nft_tunnel: no need to call htons() when dumping ports

In reply to Xin Long
Quoted 13 lines · Xin Long

info->key.tp_src and tp_dst are __be16, when using nla_put_be16() to dump them, htons() is not needed, so remove it in this patch.

---
 net/netfilter/nft_tunnel.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +0 -0
index d9d6c0d..e1184fa 100644

net/netfilter/nft_tunnel.c static int nft_tunnel_opts_dump(struct sk_buff *skb, · line 506

Quoted 7 lines
net/netfilter/nft_tunnel.c +2 -2
@@ -502,8 +502,8 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 static int nft_tunnel_ports_dump(struct sk_buff *skb,
 				 struct ip_tunnel_info *info)
 {
-	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, htons(info->key.tp_src)) < 0 ||
-	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, htons(info->key.tp_dst)) < 0)
+	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, info->key.tp_src) < 0 ||
+	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, info->key.tp_dst) < 0)

info->key.tp_src is already in __be16 as you describe. So I would take this as a consistency fix. I would take this as a fix.

#12 lore raw

Re: [PATCH nf-next 3/7] netfilter: nft_tunnel: no need to call htons() when dumping ports

In reply to Pablo Neira Ayuso
Quoted 9 lines · Pablo Neira Ayuso
Quoted 8 lines · Xin Long

info->key.tp_src and tp_dst are __be16, when using nla_put_be16() to dump them, htons() is not needed, so remove it in this patch.

---
 net/netfilter/nft_tunnel.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Quoted 15 lines

net/netfilter/nft_tunnel.c static int nft_tunnel_opts_dump(struct sk_buff *skb, · line 506

Quoted 12 lines
net/netfilter/nft_tunnel.c +2 -2
index d9d6c0d..e1184fa 100644
@@ -502,8 +502,8 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 static int nft_tunnel_ports_dump(struct sk_buff *skb,
 				 struct ip_tunnel_info *info)
 {
-	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, htons(info->key.tp_src)) < 0 ||
-	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, htons(info->key.tp_dst)) < 0)
+	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, info->key.tp_src) < 0 ||
+	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, info->key.tp_dst) < 0)

info->key.tp_src is already in __be16 as you describe. So I would take this as a consistency fix. I would take this as a fix.

I mean, I think this is worth fixing it indeed.

maintainer Acked-by
#13 lore raw

Re: [PATCH nf-next 3/7] netfilter: nft_tunnel: no need to call htons() when dumping ports

Quoted 26 lines · Pablo Neira Ayuso
Quoted 9 lines · Pablo Neira Ayuso
Quoted 8 lines · Xin Long

info->key.tp_src and tp_dst are __be16, when using nla_put_be16() to dump them, htons() is not needed, so remove it in this patch.

---
 net/netfilter/nft_tunnel.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Quoted 15 lines

net/netfilter/nft_tunnel.c static int nft_tunnel_opts_dump(struct sk_buff *skb, · line 506

Quoted 12 lines
net/netfilter/nft_tunnel.c +2 -2
index d9d6c0d..e1184fa 100644
@@ -502,8 +502,8 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 static int nft_tunnel_ports_dump(struct sk_buff *skb,
 				 struct ip_tunnel_info *info)
 {
-	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, htons(info->key.tp_src)) < 0 ||
-	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, htons(info->key.tp_dst)) < 0)
+	if (nla_put_be16(skb, NFTA_TUNNEL_KEY_SPORT, info->key.tp_src) < 0 ||
+	    nla_put_be16(skb, NFTA_TUNNEL_KEY_DPORT, info->key.tp_dst) < 0)

info->key.tp_src is already in __be16 as you describe. So I would take this as a consistency fix. I would take this as a fix.

Quoted 1 line

I mean, I think this is worth fixing it indeed.

author
#24 lore raw
In reply to Xin Long
Quoted 29 lines · Xin Long

This patchset improves quite a few places to make vxlan/erspan opts in nft_tunnel work with userspace nftables/libnftnl, and also keep consistent with the support for vxlan/erspan opts in act_tunnel_key, cls_flower and ip_tunnel_core.

Meanwhile, add support for geneve opts in nft_tunnel. One patch for nftables and one for libnftnl will be posted here for the testing. With them, nft_tunnel can be set and used by:

  # nft add table ip filter
  # nft add chain ip filter input { type filter hook input priority 0 \; }
  # nft add tunnel filter vxlan_01 { type vxlan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"ffff\"\; }
  # nft add tunnel filter erspan_01 { type erspan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:0:0\"\; }
  # nft add tunnel filter erspan_02 { type erspan\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"2:0:1:1\"\; }
  # nft add tunnel filter geneve_01 { type geneve\; id 2\; \
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:1212121234567890\"\; }
  # nft add tunnel filter geneve_02 { type geneve\; id 2\; \
Quoted 8 lines
    ip saddr 192.168.1.1\; ip daddr 192.168.1.2\; \
    sport 9000\; dport 9001\; dscp 1234\; ttl 64\; flags 1\; \
    opts \"1:1:34567890,2:2:12121212,3:3:1212121234567890\"\; }
  # nft list tunnels table filter
  # nft add rule filter input ip protocol udp tunnel name geneve_02
  # nft add rule filter input meta l4proto udp tunnel id 2 drop
  # nft add rule filter input meta l4proto udp tunnel path 0 drop
  # nft list chain filter input -a

Hi, Pablo
as you commented on other patches, I will post v2 and

Quoted 4 lines

Xin Long (7):
netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8
netfilter: nft_tunnel: parse VXLAN_GBP attr as u32 in nft_tunnel

drop these two patches

Quoted 1 line

netfilter: nft_tunnel: no need to call htons() when dumping ports

move this one to nf.git

Quoted 3 lines

netfilter: nft_tunnel: also dump ERSPAN_VERSION
netfilter: nft_tunnel: also dump OPTS_ERSPAN/VXLAN
netfilter: nft_tunnel: add the missing nla_nest_cancel()

adjust these three for nf-next.git

Quoted 1 line

netfilter: nft_tunnel: add support for geneve opts

will you also check this one before my posting v2?

Thanks.

author
#21 lore raw

Re: [PATCH nf-next 1/7] netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

In reply to Pablo Neira Ayuso
Quoted 32 lines · Pablo Neira Ayuso

Hi,

Quoted 28 lines · Xin Long

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -1
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+     [NFTA_TUNNEL_KEY_ERSPAN_VERSION]        = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]       = { .type = NLA_U32 },
-     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR] = { .type = NLA_U8 },
+     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]         = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]        = { .type = NLA_U8 },
 };
 
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
      if (err < 0)
              return err;
 
Quoted 8 lines

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_erspan_init(const struct nlattr *attr, · line 270

Quoted 2 lines
net/netfilter/nft_tunnel.c +1 -1
-     version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+     version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);

I think NFTA_TUNNEL_KEY_ERSPAN_VERSION as 32-bit is just fine.

Netlink will be adding the padding anyway for u8.

I would suggest you leave this as is.

okay.

do you think I should prepare another patch/fix for the missing nla_policy part?
[NFTA_TUNNEL_KEY_ERSPAN_VERSION] = { .type = NLA_U32 },

Thanks.

#22 lore raw

Re: [PATCH nf-next 1/7] netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

Quoted 38 lines · Xin Long
Quoted 37 lines · Pablo Neira Ayuso

Hi,

Quoted 25 lines · Xin Long

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -1
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+     [NFTA_TUNNEL_KEY_ERSPAN_VERSION]        = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]       = { .type = NLA_U32 },
-     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR] = { .type = NLA_U8 },
+     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]         = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]        = { .type = NLA_U8 },
 };
 

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_erspan_init(const struct nlattr *attr, · line 270

Quoted 5 lines
net/netfilter/nft_tunnel.c +1 -1
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
      if (err < 0)
              return err;
 
-     version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+     version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);

I think NFTA_TUNNEL_KEY_ERSPAN_VERSION as 32-bit is just fine.

Quoted 7 lines
Quoted 3 lines

Netlink will be adding the padding anyway for u8.

I would suggest you leave this as is.

okay.

do you think I should prepare another patch/fix for the missing nla_policy part?
[NFTA_TUNNEL_KEY_ERSPAN_VERSION] = { .type = NLA_U32 },

Yes, please, thanks.

#25 lore raw
In reply to Xin Long

On Thu, Dec 12, 2019 at 11:02:19AM +0800, Xin Long wrote:
[...]

Quoted 7 lines

Hi, Pablo
as you commented on other patches, I will post v2 and

Quoted 4 lines

Xin Long (7):
netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8
netfilter: nft_tunnel: parse VXLAN_GBP attr as u32 in nft_tunnel

drop these two patches

Yes, you will still need the netlink policy validation for ERSPAN_VERSION which is missing, so at least one patch will be needed for this one.

Quoted 2 lines
Quoted 1 line

netfilter: nft_tunnel: no need to call htons() when dumping ports

move this one to nf.git

Given that nft_tunnel really needs care and that there is no upstream userspace code using this extension, I think using nf-next.git in this case is fine.

Thanks.

#18 lore raw

Re: [PATCH nf-next 1/7] netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

In reply to Xin Long
Quoted 39 lines · Xin Long
Quoted 32 lines · Simon Horman

Hi Xin,

Quoted 28 lines · Xin Long

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -1
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+     [NFTA_TUNNEL_KEY_ERSPAN_VERSION]        = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]       = { .type = NLA_U32 },
-     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR] = { .type = NLA_U8 },
+     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]         = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]        = { .type = NLA_U8 },
 };
 
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
      if (err < 0)
              return err;
 
Quoted 6 lines

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_erspan_init(const struct nlattr *attr, · line 270

Quoted 2 lines
net/netfilter/nft_tunnel.c +1 -1
-     version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+     version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);

I have concerns about this change and backwards-compatibility with existing users of this UAPI. Likewise, with other changes to the encoding of existing attributes elsewhere in this series.

Quoted 5 lines

userspace(nftables/libnftnl) is not ready for nft_tunnel, I don't think there will be
any backwards-compatibility issue.

Pablo?

Thanks, I'm happy to defer to Pablo on this question.

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
net/netfilter/nft_tunnel.c +0 -1
      switch (version) {
      case ERSPAN_VERSION:
              if (!tb[NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX])
--

2.1.0

#19 lore raw

Re: [PATCH nf-next 1/7] netfilter: nft_tunnel: parse ERSPAN_VERSION attr as u8

Hi Simon,

Quoted 41 lines · Simon Horman
Quoted 40 lines · Xin Long
Quoted 38 lines · Simon Horman

Hi Xin,

Quoted 25 lines · Xin Long

To keep consistent with ipgre_policy, it's better to parse ERSPAN_VERSION attr as u8, as it does in act_tunnel_key, cls_flower and ip_tunnel_core.

---
 net/netfilter/nft_tunnel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/netfilter/nft_tunnel.c +2 -1
index 3d4c2ae..f76cd7d 100644
@@ -248,8 +248,9 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 }
 
 static const struct nla_policy nft_tunnel_opts_erspan_policy[NFTA_TUNNEL_KEY_ERSPAN_MAX + 1] = {
+     [NFTA_TUNNEL_KEY_ERSPAN_VERSION]        = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V1_INDEX]       = { .type = NLA_U32 },
-     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR] = { .type = NLA_U8 },
+     [NFTA_TUNNEL_KEY_ERSPAN_V2_DIR]         = { .type = NLA_U8 },
      [NFTA_TUNNEL_KEY_ERSPAN_V2_HWID]        = { .type = NLA_U8 },
 };
 

net/netfilter/nft_tunnel.c static int nft_tunnel_obj_erspan_init(const struct nlattr *attr, · line 270

Quoted 5 lines
net/netfilter/nft_tunnel.c +1 -1
@@ -266,7 +267,7 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
      if (err < 0)
              return err;
 
-     version = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]));
+     version = nla_get_u8(tb[NFTA_TUNNEL_KEY_ERSPAN_VERSION]);

I have concerns about this change and backwards-compatibility with existing users of this UAPI. Likewise, with other changes to the encoding of existing attributes elsewhere in this series.

Quoted 6 lines
Quoted 4 lines

userspace(nftables/libnftnl) is not ready for nft_tunnel, I don't think there will be any backwards-compatibility issue.

Pablo?

Thanks, I'm happy to defer to Pablo on this question.

I agree with Xin. This uapi is not in good shape and there is no upstream userspace code for this, no nftables support for this yet. In this particular case I'm inclined to fix uapi, better sooner than never.

Thanks.