Skip to contentSkip to conversation

netmap support for nftables

PATCH New Pablo Neira Ayuso 1 participant 6 messages Apr 24, 2020

Subsystems (2)
lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

cover netmap support for nftables

Hi,

The following patchset adds netmap support for nftables.

Patch #1 Remove the 128-bit limitation on the set element data area.
Rise it up to 64 bytes maximum.

Patch #2 Return EOPNOTSUPP in case NAT type or flags are not supported.

Patch #3 Initialize NAT flags from control plane.

Patch #4 Add helper functions to set up NAT address and protocol

Patch #5 Add netmap support.

The following example enables source netmap using the 192.168.3.0/24 network address:

 table ip x {
            chain y {
                    type nat hook postrouting priority srcnat; policy accept;
                    snat ip prefix to 192.168.3.0/24
            }

}

You can also combine it with maps:

 table ip x {
            chain y {
                    type nat hook postrouting priority srcnat; policy accept;
                    snat ip prefix to ip saddr map { 192.168.2.0/24 : 192.168.3.0/24 }
            }

}

Comments welcome.

Thanks.

Pablo Neira Ayuso (5):
netfilter: nf_tables: allow up to 64 bytes in the set element data area
netfilter: nft_nat: return EOPNOTSUPP if type or flags are not supported
netfilter: nft_nat: set flags from initialization path
netfilter: nft_nat: add helper function to set up NAT address and protocol
netfilter: nft_nat: add netmap support

 include/net/netfilter/nf_tables.h     |   4 +
 include/uapi/linux/netfilter/nf_nat.h |   4 +-
 net/netfilter/nf_tables_api.c         |  38 ++++++---
 net/netfilter/nft_nat.c               | 110 ++++++++++++++++++++------
 4 files changed, 117 insertions(+), 39 deletions(-)

--
2.20.1

author maintainer
#1 lore raw

PATCH 1/5 netfilter: nf_tables: allow up to 64 bytes in the set element data area

So far, the set elements could store up to 128-bits in the data area.

---
 include/net/netfilter/nf_tables.h |  4 ++++
 net/netfilter/nf_tables_api.c     | 38 +++++++++++++++++++++----------
 2 files changed, 30 insertions(+), 12 deletions(-)
include/net/netfilter/nf_tables.h +4 -0
index 4ff7c81e6717..d4e29c952c40 100644
@@ -243,6 +243,10 @@ struct nft_set_elem {
 		u32		buf[NFT_DATA_VALUE_MAXLEN / sizeof(u32)];
 		struct nft_data	val;
 	} key_end;
+	union {
+		u32		buf[NFT_DATA_VALUE_MAXLEN / sizeof(u32)];
+		struct nft_data val;
+	} data;
 	void			*priv;
 };
 
net/netfilter/nf_tables_api.c +26 -12
index 9780bd93b7e4..3558e76e2733 100644
@@ -4669,6 +4669,25 @@ static int nft_setelem_parse_key(struct nft_ctx *ctx, struct nft_set *set,
 	return 0;
 }
 
+static int nft_setelem_parse_data(struct nft_ctx *ctx, struct nft_set *set,
+				  struct nft_data_desc *desc,
+				  struct nft_data *data,
+				  struct nlattr *attr)
+{
+	int err;
+
+	err = nft_data_init(ctx, data, NFT_DATA_VALUE_MAXLEN, desc, attr);
+	if (err < 0)
+		return err;
+
+	if (desc->type != NFT_DATA_VERDICT && desc->len != set->dlen) {
+		nft_data_release(data, desc->type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int nft_get_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 			    const struct nlattr *attr)
 {
@@ -4946,7 +4965,6 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 	struct nft_expr *expr = NULL;
 	struct nft_userdata *udata;
 	struct nft_data_desc desc;
-	struct nft_data data;
 	enum nft_registers dreg;
 	struct nft_trans *trans;
 	u32 flags = 0;
@@ -5072,15 +5090,11 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 	}
 
 	if (nla[NFTA_SET_ELEM_DATA] != NULL) {
-		err = nft_data_init(ctx, &data, sizeof(data), &desc,
-				    nla[NFTA_SET_ELEM_DATA]);
+		err = nft_setelem_parse_data(ctx, set, &desc, &elem.data.val,
+					     nla[NFTA_SET_ELEM_DATA]);
 		if (err < 0)
 			goto err_parse_key_end;
 
-		err = -EINVAL;
-		if (set->dtype != NFT_DATA_VERDICT && desc.len != set->dlen)
-			goto err_parse_data;
-
 		dreg = nft_type_to_reg(set->dtype);
 		list_for_each_entry(binding, &set->bindings, list) {
 			struct nft_ctx bind_ctx = {
@@ -5094,14 +5108,14 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 				continue;
 
 			err = nft_validate_register_store(&bind_ctx, dreg,
-							  &data,
+							  &elem.data.val,
 							  desc.type, desc.len);
 			if (err < 0)
 				goto err_parse_data;
 
 			if (desc.type == NFT_DATA_VERDICT &&
-			    (data.verdict.code == NFT_GOTO ||
-			     data.verdict.code == NFT_JUMP))
+			    (elem.data.val.verdict.code == NFT_GOTO ||
+			     elem.data.val.verdict.code == NFT_JUMP))
 				nft_validate_state_update(ctx->net,
 							  NFT_VALIDATE_NEED);
 		}
@@ -5123,7 +5137,7 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 
 	err = -ENOMEM;
 	elem.priv = nft_set_elem_init(set, &tmpl, elem.key.val.data,
-				      elem.key_end.val.data, data.data,
+				      elem.key_end.val.data, elem.data.val.data,
 				      timeout, expiration, GFP_KERNEL);
 	if (elem.priv == NULL)
 		goto err_parse_data;
@@ -5201,7 +5215,7 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 	nf_tables_set_elem_destroy(ctx, set, elem.priv);
 err_parse_data:
 	if (nla[NFTA_SET_ELEM_DATA] != NULL)
-		nft_data_release(&data, desc.type);
+		nft_data_release(&elem.data.val, desc.type);
 err_parse_key_end:
 	nft_data_release(&elem.key_end.val, NFT_DATA_VALUE);
 err_parse_key:
Signature · 1 line
2.20.1
author maintainer
#2 lore raw

PATCH 2/5 netfilter: nft_nat: return EOPNOTSUPP if type or flags are not supported

In reply to Pablo Neira Ayuso

Instead of EINVAL which should be used for malformed netlink messages.

---
 net/netfilter/nft_nat.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/netfilter/nft_nat.c +2 -2
index 8b44a4de5329..bb49a217635e 100644
@@ -129,7 +129,7 @@ static int nft_nat_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 		priv->type = NF_NAT_MANIP_DST;
 		break;
 	default:
-		return -EINVAL;
+		return -EOPNOTSUPP;
 	}
 
 	if (tb[NFTA_NAT_FAMILY] == NULL)
@@ -196,7 +196,7 @@ static int nft_nat_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 	if (tb[NFTA_NAT_FLAGS]) {
 		priv->flags = ntohl(nla_get_be32(tb[NFTA_NAT_FLAGS]));
 		if (priv->flags & ~NF_NAT_RANGE_MASK)
-			return -EINVAL;
+			return -EOPNOTSUPP;
 	}
 
 	return nf_ct_netns_get(ctx->net, family);
Signature · 1 line
2.20.1
author maintainer
#3 lore raw

PATCH 3/5 netfilter: nft_nat: set flags from initialization path

In reply to Pablo Neira Ayuso

This patch sets the NAT flags from the control plane path.

---
 net/netfilter/nft_nat.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
net/netfilter/nft_nat.c +6 -4
index bb49a217635e..5c7ff213c030 100644
@@ -55,7 +55,6 @@ static void nft_nat_eval(const struct nft_expr *expr,
 			       &regs->data[priv->sreg_addr_max],
 			       sizeof(range.max_addr.ip6));
 		}
-		range.flags |= NF_NAT_RANGE_MAP_IPS;
 	}
 
 	if (priv->sreg_proto_min) {
@@ -63,10 +62,9 @@ static void nft_nat_eval(const struct nft_expr *expr,
 			&regs->data[priv->sreg_proto_min]);
 		range.max_proto.all = (__force __be16)nft_reg_load16(
 			&regs->data[priv->sreg_proto_max]);
-		range.flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
 	}
 
-	range.flags |= priv->flags;
+	range.flags = priv->flags;
 
 	regs->verdict.code = nf_nat_setup_info(ct, &range, priv->type);
 }
@@ -169,6 +167,8 @@ static int nft_nat_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 		} else {
 			priv->sreg_addr_max = priv->sreg_addr_min;
 		}
+
+		priv->flags |= NF_NAT_RANGE_MAP_IPS;
 	}
 
 	plen = sizeof_field(struct nf_nat_range, min_addr.all);
@@ -191,10 +191,12 @@ static int nft_nat_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 		} else {
 			priv->sreg_proto_max = priv->sreg_proto_min;
 		}
+
+		priv->flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
 	}
 
 	if (tb[NFTA_NAT_FLAGS]) {
-		priv->flags = ntohl(nla_get_be32(tb[NFTA_NAT_FLAGS]));
+		priv->flags |= ntohl(nla_get_be32(tb[NFTA_NAT_FLAGS]));
 		if (priv->flags & ~NF_NAT_RANGE_MASK)
 			return -EOPNOTSUPP;
 	}
Signature · 1 line
2.20.1
author maintainer
#4 lore raw

PATCH 4/5 netfilter: nft_nat: add helper function to set up NAT address and protocol

In reply to Pablo Neira Ayuso

This patch add nft_nat_setup_addr() and nft_nat_setup_proto() to set up the NAT mangling.

---
 net/netfilter/nft_nat.c | 56 +++++++++++++++++++++++++----------------
 1 file changed, 34 insertions(+), 22 deletions(-)
net/netfilter/nft_nat.c +34 -22
index 5c7ff213c030..7442aa8b1555 100644
@@ -30,6 +30,36 @@ struct nft_nat {
 	u16			flags;
 };
 
+static void nft_nat_setup_addr(struct nf_nat_range2 *range,
+			       const struct nft_regs *regs,
+			       const struct nft_nat *priv)
+{
+	switch (priv->family) {
+	case AF_INET:
+		range->min_addr.ip = (__force __be32)
+				regs->data[priv->sreg_addr_min];
+		range->max_addr.ip = (__force __be32)
+				regs->data[priv->sreg_addr_max];
+		break;
+	case AF_INET6:
+		memcpy(range->min_addr.ip6, &regs->data[priv->sreg_addr_min],
+		       sizeof(range->min_addr.ip6));
+		memcpy(range->max_addr.ip6, &regs->data[priv->sreg_addr_max],
+		       sizeof(range->max_addr.ip6));
+		break;
+	}
+}
+
+static void nft_nat_setup_proto(struct nf_nat_range2 *range,
+				const struct nft_regs *regs,
+				const struct nft_nat *priv)
+{
+	range->min_proto.all = (__force __be16)
+		nft_reg_load16(&regs->data[priv->sreg_proto_min]);
+	range->max_proto.all = (__force __be16)
+		nft_reg_load16(&regs->data[priv->sreg_proto_max]);
+}
+
 static void nft_nat_eval(const struct nft_expr *expr,
 			 struct nft_regs *regs,
 			 const struct nft_pktinfo *pkt)
@@ -40,29 +70,11 @@ static void nft_nat_eval(const struct nft_expr *expr,
 	struct nf_nat_range2 range;
 
 	memset(&range, 0, sizeof(range));
-	if (priv->sreg_addr_min) {
-		if (priv->family == AF_INET) {
-			range.min_addr.ip = (__force __be32)
-					regs->data[priv->sreg_addr_min];
-			range.max_addr.ip = (__force __be32)
-					regs->data[priv->sreg_addr_max];
-
-		} else {
-			memcpy(range.min_addr.ip6,
-			       &regs->data[priv->sreg_addr_min],
-			       sizeof(range.min_addr.ip6));
-			memcpy(range.max_addr.ip6,
-			       &regs->data[priv->sreg_addr_max],
-			       sizeof(range.max_addr.ip6));
-		}
-	}
+	if (priv->sreg_addr_min)
+		nft_nat_setup_addr(&range, regs, priv);
 
-	if (priv->sreg_proto_min) {
-		range.min_proto.all = (__force __be16)nft_reg_load16(
-			&regs->data[priv->sreg_proto_min]);
-		range.max_proto.all = (__force __be16)nft_reg_load16(
-			&regs->data[priv->sreg_proto_max]);
-	}
+	if (priv->sreg_proto_min)
+		nft_nat_setup_proto(&range, regs, priv);
 
 	range.flags = priv->flags;
 
Signature · 1 line
2.20.1
author maintainer
#5 lore raw

PATCH 5/5 netfilter: nft_nat: add netmap support

In reply to Pablo Neira Ayuso

This patch allows you to NAT the network address prefix onto another network address prefix, a.k.a. netmapping.

Userspace must specify the NF_NAT_RANGE_NETMAP flag and the prefix address through the NFTA_NAT_REG_ADDR_MIN and NFTA_NAT_REG_ADDR_MAX netlink attributes.

---
 include/uapi/linux/netfilter/nf_nat.h |  4 ++-
 net/netfilter/nft_nat.c               | 46 ++++++++++++++++++++++++++-
 2 files changed, 48 insertions(+), 2 deletions(-)
include/uapi/linux/netfilter/nf_nat.h +3 -1
index 4a95c0db14d4..a64586e77b24 100644
@@ -11,6 +11,7 @@
 #define NF_NAT_RANGE_PERSISTENT			(1 << 3)
 #define NF_NAT_RANGE_PROTO_RANDOM_FULLY		(1 << 4)
 #define NF_NAT_RANGE_PROTO_OFFSET		(1 << 5)
+#define NF_NAT_RANGE_NETMAP			(1 << 6)
 
 #define NF_NAT_RANGE_PROTO_RANDOM_ALL		\
 	(NF_NAT_RANGE_PROTO_RANDOM | NF_NAT_RANGE_PROTO_RANDOM_FULLY)
@@ -18,7 +19,8 @@
 #define NF_NAT_RANGE_MASK					\
 	(NF_NAT_RANGE_MAP_IPS | NF_NAT_RANGE_PROTO_SPECIFIED |	\
 	 NF_NAT_RANGE_PROTO_RANDOM | NF_NAT_RANGE_PERSISTENT |	\
-	 NF_NAT_RANGE_PROTO_RANDOM_FULLY | NF_NAT_RANGE_PROTO_OFFSET)
+	 NF_NAT_RANGE_PROTO_RANDOM_FULLY | NF_NAT_RANGE_PROTO_OFFSET | \
+	 NF_NAT_RANGE_NETMAP)
 
 struct nf_nat_ipv4_range {
 	unsigned int			flags;
net/netfilter/nft_nat.c +45 -1

59 lines not on this page · raw message

Signature · 1 line
2.20.1