Skip to contentSkip to conversation

Netfilter fixes for net

PATCH In review Pablo Neira Ayuso 2 participants 11 messages Nov 6, 2019 – Nov 7, 2019

Subsystems (4)
lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

cover Netfilter fixes for net

Hi David,

The following patchset contains Netfilter fixes for net:

1) Missing register size validation in bitwise and cmp offloads.

2) Fix error code in ip_set_sockfn_get() when copy_to_user() fails, from Dan Carpenter.

3) Oneliner to copy MAC address in IPv6 hash:ip,mac sets, from Stefano Brivio.

4) Missing policy validation in ipset with NL_VALIDATE_STRICT, from Jozsef Kadlecsik.

5) Fix unaligned access to private data area of nf_tables instructions, from Lukas Wunner.

6) Relax check for object updates, reported as a regression by Eric Garver, patch from Fernando Fernandez Mancera.

7) Crash on ebtables dnat extension when used from the output path. From Florian Westphal.

8) Fix bogus EOPNOTSUPP when updating basechain flags.

9) Fix bogus EBUSY when updating a basechain that is already offloaded.

You can pull these changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git

Thanks.

----------------------------------------------------------------

The following changes since commit 1204c70d9dcba31164f78ad5d8c88c42335d51f8:

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net (2019-11-01 17:48:11 -0700)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git HEAD

for you to fetch changes up to 774e4d34dbebc9dc441535c4712794d336a9478c:

Merge branch 'master' of git://blackhole.kfki.hu/nf (2019-11-04 20:59:00 +0100)

----------------------------------------------------------------
Dan Carpenter (1):
netfilter: ipset: Fix an error code in ip_set_sockfn_get()

Fernando Fernandez Mancera (1):
netfilter: nf_tables: fix unexpected EOPNOTSUPP error

Florian Westphal (1):
bridge: ebtables: don't crash when using dnat target in output chains

Jozsef Kadlecsik (1):
netfilter: ipset: Fix nla_policies to fully support NL_VALIDATE_STRICT

Lukas Wunner (1):
netfilter: nf_tables: Align nft_expr private data to 64-bit

Pablo Neira Ayuso (4):
netfilter: nf_tables_offload: check for register data length mismatches
netfilter: nf_tables: bogus EOPNOTSUPP on basechain update
netfilter: nf_tables_offload: skip EBUSY on chain update Merge branch 'master' of git://blackhole.kfki.hu/nf

Stefano Brivio (1):
netfilter: ipset: Copy the right MAC address in hash:ip,mac IPv6 sets

 include/net/netfilter/nf_tables.h        |  3 +-
 net/bridge/netfilter/ebt_dnat.c          | 19 ++++++++++---
 net/netfilter/ipset/ip_set_core.c        | 49 +++++++++++++++++++++-----------
 net/netfilter/ipset/ip_set_hash_ipmac.c  |  2 +-
 net/netfilter/ipset/ip_set_hash_net.c    |  1 +
 net/netfilter/ipset/ip_set_hash_netnet.c |  1 +
 net/netfilter/nf_tables_api.c            |  7 ++---
 net/netfilter/nf_tables_offload.c        |  3 +-
 net/netfilter/nft_bitwise.c              |  5 ++--
 net/netfilter/nft_cmp.c                  |  2 +-
 10 files changed, 62 insertions(+), 30 deletions(-)
author maintainer
#1 lore raw

PATCH 1/9 netfilter: nf_tables_offload: check for register data length mismatches

Make sure register data length does not mismatch immediate data length, otherwise hit EOPNOTSUPP.

Fixes c9626a2cbdb2 ("netfilter: nf_tables: add hardware offload support")
---
 net/netfilter/nft_bitwise.c | 5 +++--
 net/netfilter/nft_cmp.c     | 2 +-
 2 files changed, 4 insertions(+), 3 deletions(-)
net/netfilter/nft_bitwise.c +3 -2
index 974300178fa9..02afa752dd2e 100644
@@ -134,12 +134,13 @@ static int nft_bitwise_offload(struct nft_offload_ctx *ctx,
                                const struct nft_expr *expr)
 {
 	const struct nft_bitwise *priv = nft_expr_priv(expr);
+	struct nft_offload_reg *reg = &ctx->regs[priv->dreg];
 
 	if (memcmp(&priv->xor, &zero, sizeof(priv->xor)) ||
-	    priv->sreg != priv->dreg)
+	    priv->sreg != priv->dreg || priv->len != reg->len)
 		return -EOPNOTSUPP;
 
-	memcpy(&ctx->regs[priv->dreg].mask, &priv->mask, sizeof(priv->mask));
+	memcpy(&reg->mask, &priv->mask, sizeof(priv->mask));
 
 	return 0;
 }
net/netfilter/nft_cmp.c +1 -1
index bd173b1824c6..0744b2bb46da 100644
@@ -116,7 +116,7 @@ static int __nft_cmp_offload(struct nft_offload_ctx *ctx,
 	u8 *mask = (u8 *)&flow->match.mask;
 	u8 *key = (u8 *)&flow->match.key;
 
-	if (priv->op != NFT_CMP_EQ)
+	if (priv->op != NFT_CMP_EQ || reg->len != priv->len)
 		return -EOPNOTSUPP;
 
 	memcpy(key + reg->offset, &priv->data, priv->len);
Signature · 1 line
2.11.0
author maintainer
#2 lore raw

PATCH 2/9 netfilter: ipset: Fix an error code in ip_set_sockfn_get()

In reply to Pablo Neira Ayuso

From: Dan Carpenter <[email protected]>

The copy_to_user() function returns the number of bytes remaining to be copied. In this code, that positive return is checked at the end of the function and we return zero/success. What we should do instead is return -EFAULT.

Fixes a7b4f989a629 ("netfilter: ipset: IP set core support")
---
 net/netfilter/ipset/ip_set_core.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
net/netfilter/ipset/ip_set_core.c +5 -3
index e64d5f9a89dd..e7288eab7512 100644
@@ -2069,8 +2069,9 @@ ip_set_sockfn_get(struct sock *sk, int optval, void __user *user, int *len)
 		}
 
 		req_version->version = IPSET_PROTOCOL;
-		ret = copy_to_user(user, req_version,
-				   sizeof(struct ip_set_req_version));
+		if (copy_to_user(user, req_version,
+				 sizeof(struct ip_set_req_version)))
+			ret = -EFAULT;
 		goto done;
 	}
 	case IP_SET_OP_GET_BYNAME: {
@@ -2129,7 +2130,8 @@ ip_set_sockfn_get(struct sock *sk, int optval, void __user *user, int *len)
 	}	/* end of switch(op) */
 
 copy:
-	ret = copy_to_user(user, data, copylen);
+	if (copy_to_user(user, data, copylen))
+		ret = -EFAULT;
 
 done:
 	vfree(data);
Signature · 1 line
2.11.0
author maintainer
#3 lore raw

PATCH 3/9 netfilter: ipset: Copy the right MAC address in hash:ip,mac IPv6 sets

In reply to Pablo Neira Ayuso

From: Stefano Brivio <[email protected]>

Same as commit 1b4a75108d5b ("netfilter: ipset: Copy the right MAC address in bitmap:ip,mac and hash:ip,mac sets"), another copy and paste went wrong in commit 8cc4ccf58379 ("netfilter: ipset: Allow matching on destination MAC address for mac and ipmac sets").

When I fixed this for IPv4 in 1b4a75108d5b, I didn't realise that hash:ip,mac sets also support IPv6 as family, and this is covered by a separate function, hash_ipmac6_kadt().

In hash:ip,mac sets, the first dimension is the IP address, and the second dimension is the MAC address: check the IPSET_DIM_TWO_SRC flag in flags while deciding which MAC address to copy, destination or source.

This way, mixing source and destination matches for the two dimensions of ip,mac hash type works as expected, also for IPv6. With this setup:

ip netns add A
ip link add veth1 type veth peer name veth2 netns A ip addr add 2001:db8::1/64 dev veth1
ip -net A addr add 2001:db8::2/64 dev veth2 ip link set veth1 up
ip -net A link set veth2 up

  dst=$(ip netns exec A cat /sys/class/net/veth2/address)

  ip netns exec A ipset create test_hash hash:ip,mac family inet6
  ip netns exec A ipset add test_hash 2001:db8::1,${dst}
  ip netns exec A ip6tables -A INPUT -p icmpv6 --icmpv6-type 135 -j ACCEPT
  ip netns exec A ip6tables -A INPUT -m set ! --match-set test_hash src,dst -j DROP

ipset now correctly matches a test packet:

  # ping -c1 2001:db8::2 >/dev/null
  # echo $?
  0
Fixes 8cc4ccf58379 ("netfilter: ipset: Allow matching on destination MAC address for mac and ipmac sets")
---
 net/netfilter/ipset/ip_set_hash_ipmac.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/netfilter/ipset/ip_set_hash_ipmac.c +1 -1
index 24d8f4df4230..4ce563eb927d 100644
@@ -209,7 +209,7 @@ hash_ipmac6_kadt(struct ip_set *set, const struct sk_buff *skb,
 	    (skb_mac_header(skb) + ETH_HLEN) > skb->data)
 		return -EINVAL;
 
-	if (opt->flags & IPSET_DIM_ONE_SRC)
+	if (opt->flags & IPSET_DIM_TWO_SRC)
 		ether_addr_copy(e.ether, eth_hdr(skb)->h_source);
 	else
 		ether_addr_copy(e.ether, eth_hdr(skb)->h_dest);
Signature · 1 line
2.11.0
author maintainer
#4 lore raw

PATCH 4/9 netfilter: ipset: Fix nla_policies to fully support NL_VALIDATE_STRICT

In reply to Pablo Neira Ayuso

From: Jozsef Kadlecsik <[email protected]>

Since v5.2 (commit "netlink: re-add parse/validate functions in strict mode") NL_VALIDATE_STRICT is enabled. Fix the ipset nla_policies which did not support strict mode and convert from deprecated parsings to verified ones.

---
 net/netfilter/ipset/ip_set_core.c        | 41 ++++++++++++++++++++++----------
 net/netfilter/ipset/ip_set_hash_net.c    |  1 +
 net/netfilter/ipset/ip_set_hash_netnet.c |  1 +
 3 files changed, 30 insertions(+), 13 deletions(-)
net/netfilter/ipset/ip_set_core.c +28 -13
index e7288eab7512..d73d1828216a 100644
@@ -296,7 +296,8 @@ ip_set_get_ipaddr4(struct nlattr *nla,  __be32 *ipaddr)
 
 	if (unlikely(!flag_nested(nla)))
 		return -IPSET_ERR_PROTOCOL;
-	if (nla_parse_nested_deprecated(tb, IPSET_ATTR_IPADDR_MAX, nla, ipaddr_policy, NULL))
+	if (nla_parse_nested(tb, IPSET_ATTR_IPADDR_MAX, nla,
+			     ipaddr_policy, NULL))
 		return -IPSET_ERR_PROTOCOL;
 	if (unlikely(!ip_set_attr_netorder(tb, IPSET_ATTR_IPADDR_IPV4)))
 		return -IPSET_ERR_PROTOCOL;
@@ -314,7 +315,8 @@ ip_set_get_ipaddr6(struct nlattr *nla, union nf_inet_addr *ipaddr)
 	if (unlikely(!flag_nested(nla)))
 		return -IPSET_ERR_PROTOCOL;
 
-	if (nla_parse_nested_deprecated(tb, IPSET_ATTR_IPADDR_MAX, nla, ipaddr_policy, NULL))
+	if (nla_parse_nested(tb, IPSET_ATTR_IPADDR_MAX, nla,
+			     ipaddr_policy, NULL))
 		return -IPSET_ERR_PROTOCOL;
 	if (unlikely(!ip_set_attr_netorder(tb, IPSET_ATTR_IPADDR_IPV6)))
 		return -IPSET_ERR_PROTOCOL;
@@ -934,7 +936,8 @@ static int ip_set_create(struct net *net, struct sock *ctnl,
 
 	/* Without holding any locks, create private part. */
 	if (attr[IPSET_ATTR_DATA] &&
-	    nla_parse_nested_deprecated(tb, IPSET_ATTR_CREATE_MAX, attr[IPSET_ATTR_DATA], set->type->create_policy, NULL)) {
+	    nla_parse_nested(tb, IPSET_ATTR_CREATE_MAX, attr[IPSET_ATTR_DATA],
+			     set->type->create_policy, NULL)) {
 		ret = -IPSET_ERR_PROTOCOL;
 		goto put_out;
 	}
@@ -1281,6 +1284,14 @@ dump_attrs(struct nlmsghdr *nlh)
 	}
 }
 
+static const struct nla_policy
+ip_set_dump_policy[IPSET_ATTR_CMD_MAX + 1] = {
+	[IPSET_ATTR_PROTOCOL]	= { .type = NLA_U8 },
+	[IPSET_ATTR_SETNAME]	= { .type = NLA_NUL_STRING,
+				    .len = IPSET_MAXNAMELEN - 1 },
+	[IPSET_ATTR_FLAGS]	= { .type = NLA_U32 },
+};
+
 static int
 dump_init(struct netlink_callback *cb, struct ip_set_net *inst)
 {
@@ -1292,9 +1303,9 @@ dump_init(struct netlink_callback *cb, struct ip_set_net *inst)
 	ip_set_id_t index;
 	int ret;
 
-	ret = nla_parse_deprecated(cda, IPSET_ATTR_CMD_MAX, attr,
-				   nlh->nlmsg_len - min_len,
-				   ip_set_setname_policy, NULL);
+	ret = nla_parse(cda, IPSET_ATTR_CMD_MAX, attr,
+			nlh->nlmsg_len - min_len,
+			ip_set_dump_policy, NULL);
 	if (ret)
 		return ret;
 
@@ -1543,9 +1554,9 @@ call_ad(struct sock *ctnl, struct sk_buff *skb, struct ip_set *set,
 		memcpy(&errmsg->msg, nlh, nlh->nlmsg_len);
 		cmdattr = (void *)&errmsg->msg + min_len;
 
-		ret = nla_parse_deprecated(cda, IPSET_ATTR_CMD_MAX, cmdattr,
-					   nlh->nlmsg_len - min_len,
-					   ip_set_adt_policy, NULL);
+		ret = nla_parse(cda, IPSET_ATTR_CMD_MAX, cmdattr,
+				nlh->nlmsg_len - min_len, ip_set_adt_policy,
+				NULL);
 
 		if (ret) {
 			nlmsg_free(skb2);
@@ -1596,7 +1607,9 @@ static int ip_set_ad(struct net *net, struct sock *ctnl,
 
 	use_lineno = !!attr[IPSET_ATTR_LINENO];
 	if (attr[IPSET_ATTR_DATA]) {
-		if (nla_parse_nested_deprecated(tb, IPSET_ATTR_ADT_MAX, attr[IPSET_ATTR_DATA], set->type->adt_policy, NULL))
+		if (nla_parse_nested(tb, IPSET_ATTR_ADT_MAX,
+				     attr[IPSET_ATTR_DATA],
+				     set->type->adt_policy, NULL))
 			return -IPSET_ERR_PROTOCOL;
 		ret = call_ad(ctnl, skb, set, tb, adt, flags,
 			      use_lineno);
@@ -1606,7 +1619,8 @@ static int ip_set_ad(struct net *net, struct sock *ctnl,
 		nla_for_each_nested(nla, attr[IPSET_ATTR_ADT], nla_rem) {
 			if (nla_type(nla) != IPSET_ATTR_DATA ||
 			    !flag_nested(nla) ||
-			    nla_parse_nested_deprecated(tb, IPSET_ATTR_ADT_MAX, nla, set->type->adt_policy, NULL))
+			    nla_parse_nested(tb, IPSET_ATTR_ADT_MAX, nla,
+					     set->type->adt_policy, NULL))
 				return -IPSET_ERR_PROTOCOL;
 			ret = call_ad(ctnl, skb, set, tb, adt,
 				      flags, use_lineno);
@@ -1655,7 +1669,8 @@ static int ip_set_utest(struct net *net, struct sock *ctnl, struct sk_buff *skb,
 	if (!set)
 		return -ENOENT;
 
-	if (nla_parse_nested_deprecated(tb, IPSET_ATTR_ADT_MAX, attr[IPSET_ATTR_DATA], set->type->adt_policy, NULL))
+	if (nla_parse_nested(tb, IPSET_ATTR_ADT_MAX, attr[IPSET_ATTR_DATA],
+			     set->type->adt_policy, NULL))
 		return -IPSET_ERR_PROTOCOL;
 
 	rcu_read_lock_bh();
@@ -1961,7 +1976,7 @@ static const struct nfnl_callback ip_set_netlink_subsys_cb[IPSET_MSG_MAX] = {
 	[IPSET_CMD_LIST]	= {
 		.call		= ip_set_dump,
 		.attr_count	= IPSET_ATTR_CMD_MAX,
-		.policy		= ip_set_setname_policy,
+		.policy		= ip_set_dump_policy,
 	},
 	[IPSET_CMD_SAVE]	= {
 		.call		= ip_set_dump,
net/netfilter/ipset/ip_set_hash_net.c +1 -0
index c259cbc3ef45..3d932de0ad29 100644
@@ -368,6 +368,7 @@ static struct ip_set_type hash_net_type __read_mostly = {
 		[IPSET_ATTR_IP_TO]	= { .type = NLA_NESTED },
 		[IPSET_ATTR_CIDR]	= { .type = NLA_U8 },
 		[IPSET_ATTR_TIMEOUT]	= { .type = NLA_U32 },
+		[IPSET_ATTR_LINENO]	= { .type = NLA_U32 },
 		[IPSET_ATTR_CADT_FLAGS]	= { .type = NLA_U32 },
 		[IPSET_ATTR_BYTES]	= { .type = NLA_U64 },
 		[IPSET_ATTR_PACKETS]	= { .type = NLA_U64 },
net/netfilter/ipset/ip_set_hash_netnet.c +1 -0
index a3ae69bfee66..4398322fad59 100644
@@ -476,6 +476,7 @@ static struct ip_set_type hash_netnet_type __read_mostly = {
 		[IPSET_ATTR_CIDR]	= { .type = NLA_U8 },
 		[IPSET_ATTR_CIDR2]	= { .type = NLA_U8 },
 		[IPSET_ATTR_TIMEOUT]	= { .type = NLA_U32 },
+		[IPSET_ATTR_LINENO]	= { .type = NLA_U32 },
 		[IPSET_ATTR_CADT_FLAGS]	= { .type = NLA_U32 },
 		[IPSET_ATTR_BYTES]	= { .type = NLA_U64 },
 		[IPSET_ATTR_PACKETS]	= { .type = NLA_U64 },
Signature · 1 line
2.11.0
author maintainer
#5 lore raw

PATCH 5/9 netfilter: nf_tables: Align nft_expr private data to 64-bit

In reply to Pablo Neira Ayuso

From: Lukas Wunner <[email protected]>

Invoking the following commands on a 32-bit architecture with strict alignment requirements (such as an ARMv7-based Raspberry Pi) results in an alignment exception:

 # nft add table ip test-ip4
 # nft add chain ip test-ip4 output { type filter hook output priority 0; }
 # nft add rule  ip test-ip4 output quota 1025 bytes

Alignment trap: not handling instruction e1b26f9f at [<7f4473f8>] Unhandled fault: alignment exception (0x001) at 0xb832e824

Internal error: : 1 [#1] PREEMPT SMP ARM
Hardware name: BCM2835

[<7f4473fc>] (nft_quota_do_init [nft_quota])
[<7f447448>] (nft_quota_init [nft_quota])
[<7f4260d0>] (nf_tables_newrule [nf_tables])
[<7f4168dc>] (nfnetlink_rcv_batch [nfnetlink])
[<7f416bd0>] (nfnetlink_rcv [nfnetlink])
[<8078b334>] (netlink_unicast)
[<8078b664>] (netlink_sendmsg)
[<8071b47c>] (sock_sendmsg)
[<8071bd18>] (___sys_sendmsg)
[<8071ce3c>] (__sys_sendmsg)
[<8071ce94>] (sys_sendmsg)

The reason is that nft_quota_do_init() calls atomic64_set() on an atomic64_t which is only aligned to 32-bit, not 64-bit, because it succeeds struct nft_expr in memory which only contains a 32-bit pointer. Fix by aligning the nft_expr private data to 64-bit.

Fixes 96518518cc41 ("netfilter: add nftables")
Cc [email protected] # v3.13+
---
 include/net/netfilter/nf_tables.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
include/net/netfilter/nf_tables.h +2 -1
index 001d294edf57..2d0275f13bbf 100644
@@ -820,7 +820,8 @@ struct nft_expr_ops {
  */
 struct nft_expr {
 	const struct nft_expr_ops	*ops;
-	unsigned char			data[];
+	unsigned char			data[]
+		__attribute__((aligned(__alignof__(u64))));
 };
 
 static inline void *nft_expr_priv(const struct nft_expr *expr)
Signature · 1 line
2.11.0
author maintainer
#6 lore raw

PATCH 6/9 netfilter: nf_tables: fix unexpected EOPNOTSUPP error

In reply to Pablo Neira Ayuso

From: Fernando Fernandez Mancera <[email protected]>

If the object type doesn't implement an update operation and the user tries to update it will silently ignore the update operation.

Fixes aa4095a156b5 ("netfilter: nf_tables: fix possible null-pointer dereference in object update")
---
 net/netfilter/nf_tables_api.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
net/netfilter/nf_tables_api.c +2 -4
index d481f9baca2f..aa26841ad9a1 100644
@@ -5143,9 +5143,6 @@ static int nf_tables_updobj(const struct nft_ctx *ctx,
 	struct nft_trans *trans;
 	int err;
 
-	if (!obj->ops->update)
-		return -EOPNOTSUPP;
-
 	trans = nft_trans_alloc(ctx, NFT_MSG_NEWOBJ,
 				sizeof(struct nft_trans_obj));
 	if (!trans)
@@ -6499,7 +6496,8 @@ static void nft_obj_commit_update(struct nft_trans *trans)
 	obj = nft_trans_obj(trans);
 	newobj = nft_trans_obj_newobj(trans);
 
-	obj->ops->update(obj, newobj);
+	if (obj->ops->update)
+		obj->ops->update(obj, newobj);
 
 	kfree(newobj);
 }
Signature · 1 line
2.11.0
author maintainer
#7 lore raw

PATCH 7/9 bridge: ebtables: don't crash when using dnat target in output chains

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

xt_in() returns NULL in the output hook, skip the pkt_type change for that case, redirection only makes sense in broute/prerouting hooks.

Cc Linus Lüssing <[email protected]>
Fixes cf3cb246e277d ("bridge: ebtables: fix reception of frames DNAT-ed to bridge device/port")
---
 net/bridge/netfilter/ebt_dnat.c | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)
net/bridge/netfilter/ebt_dnat.c +15 -4
index ed91ea31978a..12a4f4d93681 100644
@@ -20,7 +20,6 @@ static unsigned int
 ebt_dnat_tg(struct sk_buff *skb, const struct xt_action_param *par)
 {
 	const struct ebt_nat_info *info = par->targinfo;
-	struct net_device *dev;
 
 	if (skb_ensure_writable(skb, ETH_ALEN))
 		return EBT_DROP;
@@ -33,10 +32,22 @@ ebt_dnat_tg(struct sk_buff *skb, const struct xt_action_param *par)
 		else
 			skb->pkt_type = PACKET_MULTICAST;
 	} else {
-		if (xt_hooknum(par) != NF_BR_BROUTING)
-			dev = br_port_get_rcu(xt_in(par))->br->dev;
-		else
+		const struct net_device *dev;
+
+		switch (xt_hooknum(par)) {
+		case NF_BR_BROUTING:
 			dev = xt_in(par);
+			break;
+		case NF_BR_PRE_ROUTING:
+			dev = br_port_get_rcu(xt_in(par))->br->dev;
+			break;
+		default:
+			dev = NULL;
+			break;
+		}
+
+		if (!dev) /* NF_BR_LOCAL_OUT */
+			return info->target;
 
 		if (ether_addr_equal(info->mac, dev->dev_addr))
 			skb->pkt_type = PACKET_HOST;
Signature · 1 line
2.11.0
author maintainer
#8 lore raw

PATCH 8/9 netfilter: nf_tables: bogus EOPNOTSUPP on basechain update

In reply to Pablo Neira Ayuso

Userspace never includes the NFT_BASE_CHAIN flag, this flag is inferred from the NFTA_CHAIN_HOOK atribute. The chain update path does not allow to update flags at this stage, the existing sanity check bogusly hits EOPNOTSUPP in the basechain case if the offload flag is set on.

---
 net/netfilter/nf_tables_api.c | 1 +
 1 file changed, 1 insertion(+)
net/netfilter/nf_tables_api.c +1 -0
index aa26841ad9a1..712a428509ad 100644
@@ -1922,6 +1922,7 @@ static int nf_tables_newchain(struct net *net, struct sock *nlsk,
 		if (nlh->nlmsg_flags & NLM_F_REPLACE)
 			return -EOPNOTSUPP;
 
+		flags |= chain->flags & NFT_BASE_CHAIN;
 		return nf_tables_updchain(&ctx, genmask, policy, flags);
 	}
 
Signature · 1 line
2.11.0
author maintainer
#9 lore raw

PATCH 9/9 netfilter: nf_tables_offload: skip EBUSY on chain update

In reply to Pablo Neira Ayuso

Do not try to bind a chain again if it exists, otherwise the driver returns EBUSY.

Fixes c9626a2cbdb2 ("netfilter: nf_tables: add hardware offload support")
---
 net/netfilter/nf_tables_offload.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
net/netfilter/nf_tables_offload.c +2 -1
index ad783f4840ef..e25dab8128db 100644
@@ -334,7 +334,8 @@ int nft_flow_rule_offload_commit(struct net *net)
 
 		switch (trans->msg_type) {
 		case NFT_MSG_NEWCHAIN:
-			if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD))
+			if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD) ||
+			    nft_trans_chain_update(trans))
 				continue;
 
 			policy = nft_trans_chain_policy(trans);
Signature · 1 line
2.11.0
maintainer
#10 lore raw
In reply to Pablo Neira Ayuso

From: Pablo Neira Ayuso <[email protected]>
Date: Wed, 6 Nov 2019 12:12:28 +0100

Quoted 22 lines

The following patchset contains Netfilter fixes for net:

1) Missing register size validation in bitwise and cmp offloads.

2) Fix error code in ip_set_sockfn_get() when copy_to_user() fails, from Dan Carpenter.

3) Oneliner to copy MAC address in IPv6 hash:ip,mac sets, from Stefano Brivio.

4) Missing policy validation in ipset with NL_VALIDATE_STRICT, from Jozsef Kadlecsik.

5) Fix unaligned access to private data area of nf_tables instructions, from Lukas Wunner.

6) Relax check for object updates, reported as a regression by Eric Garver, patch from Fernando Fernandez Mancera.

7) Crash on ebtables dnat extension when used from the output path. From Florian Westphal.

Quoted 7 lines

8) Fix bogus EOPNOTSUPP when updating basechain flags.

9) Fix bogus EBUSY when updating a basechain that is already offloaded.

You can pull these changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git

Pulled, thanks Pablo.