[PATCH net-next 00/10] Netfilter updates for net-next

[PATCH net-next 00/10] Netfilter updates for net-next

[Pablo Neira Ayuso]

Hi,

The following batch contains Netfilter updates for net-next:

1) Split flowtable workqueues per events, from Oz Shlomo.

2) fall-through warnings for clang, from Gustavo A. R. Silva

3) Remove unused declaration in conntrack, from YueHaibing.

4) Consolidate skb_try_make_writable() in flowtable datapath,
   simplify some of the existing codebase.

5) Call dst_check() to fall back to static classic forwarding path.

6) Update table flags from commit phase.

Please, pull these changes from:

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

Thanks!

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

The following changes since commit ebfbc46b35cb70b9fbd88f376d7a33b79f60adff:

  openvswitch: Warn over-mtu packets only if iface is UP. (2021-03-16 16:28:30 -0700)

are available in the Git repository at:

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

for you to fetch changes up to 0ce7cf4127f14078ca598ba9700d813178a59409:

  netfilter: nftables: update table flags from the commit phase (2021-03-18 01:35:39 +0100)

----------------------------------------------------------------
Gustavo A. R. Silva (1):
      netfilter: Fix fall-through warnings for Clang

Oz Shlomo (1):
      netfilter: flowtable: separate replace, destroy and stats to different workqueues

Pablo Neira Ayuso (7):
      netfilter: flowtable: consolidate skb_try_make_writable() call
      netfilter: flowtable: move skb_try_make_writable() before NAT in IPv4
      netfilter: flowtable: move FLOW_OFFLOAD_DIR_MAX away from enumeration
      netfilter: flowtable: fast NAT functions never fail
      netfilter: flowtable: call dst_check() to fall back to classic forwarding
      netfilter: flowtable: refresh timeout after dst and writable checks
      netfilter: nftables: update table flags from the commit phase

YueHaibing (1):
      netfilter: conntrack: Remove unused variable declaration

 include/net/netfilter/ipv6/nf_conntrack_ipv6.h |   3 -
 include/net/netfilter/nf_flow_table.h          |  14 +-
 include/net/netfilter/nf_tables.h              |   9 +-
 net/netfilter/nf_conntrack_proto_dccp.c        |   1 +
 net/netfilter/nf_flow_table_core.c             |  57 ++----
 net/netfilter/nf_flow_table_ip.c               | 231 ++++++++++---------------
 net/netfilter/nf_flow_table_offload.c          |  44 ++++-
 net/netfilter/nf_tables_api.c                  |  32 ++--
 net/netfilter/nft_ct.c                         |   1 +
 9 files changed, 174 insertions(+), 218 deletions(-)

[PATCH net-next 01/10] netfilter: flowtable: separate replace, destroy and stats to different workqueues

[Pablo Neira Ayuso]

From: Oz Shlomo <ozsh@nvidia.com>

Currently the flow table offload replace, destroy and stats work items are
executed on a single workqueue. As such, DESTROY and STATS commands may
be backloged after a burst of REPLACE work items. This scenario can bloat
up memory and may cause active connections to age.

Instatiate add, del and stats workqueues to avoid backlogs of non-dependent
actions. Provide sysfs control over the workqueue attributes, allowing
userspace applications to control the workqueue cpumask.

Signed-off-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Paul Blakey <paulb@nvidia.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/netfilter/nf_flow_table_offload.c | 44 ++++++++++++++++++++++-----
 1 file changed, 36 insertions(+), 8 deletions(-)

diff --git a/net/netfilter/nf_flow_table_offload.c b/net/netfilter/nf_flow_table_offload.c
index 2a6993fa40d7..1b979c8b3ba0 100644
--- a/net/netfilter/nf_flow_table_offload.c
+++ b/net/netfilter/nf_flow_table_offload.c
@@ -13,7 +13,9 @@
 #include <net/netfilter/nf_conntrack_core.h>
 #include <net/netfilter/nf_conntrack_tuple.h>
 
-static struct workqueue_struct *nf_flow_offload_wq;
+static struct workqueue_struct *nf_flow_offload_add_wq;
+static struct workqueue_struct *nf_flow_offload_del_wq;
+static struct workqueue_struct *nf_flow_offload_stats_wq;
 
 struct flow_offload_work {
 	struct list_head	list;
@@ -826,7 +828,12 @@ static void flow_offload_work_handler(struct work_struct *work)
 
 static void flow_offload_queue_work(struct flow_offload_work *offload)
 {
-	queue_work(nf_flow_offload_wq, &offload->work);
+	if (offload->cmd == FLOW_CLS_REPLACE)
+		queue_work(nf_flow_offload_add_wq, &offload->work);
+	else if (offload->cmd == FLOW_CLS_DESTROY)
+		queue_work(nf_flow_offload_del_wq, &offload->work);
+	else
+		queue_work(nf_flow_offload_stats_wq, &offload->work);
 }
 
 static struct flow_offload_work *
@@ -898,8 +905,11 @@ void nf_flow_offload_stats(struct nf_flowtable *flowtable,
 
 void nf_flow_table_offload_flush(struct nf_flowtable *flowtable)
 {
-	if (nf_flowtable_hw_offload(flowtable))
-		flush_workqueue(nf_flow_offload_wq);
+	if (nf_flowtable_hw_offload(flowtable)) {
+		flush_workqueue(nf_flow_offload_add_wq);
+		flush_workqueue(nf_flow_offload_del_wq);
+		flush_workqueue(nf_flow_offload_stats_wq);
+	}
 }
 
 static int nf_flow_table_block_setup(struct nf_flowtable *flowtable,
@@ -1011,15 +1021,33 @@ EXPORT_SYMBOL_GPL(nf_flow_table_offload_setup);
 
 int nf_flow_table_offload_init(void)
 {
-	nf_flow_offload_wq  = alloc_workqueue("nf_flow_table_offload",
-					      WQ_UNBOUND, 0);
-	if (!nf_flow_offload_wq)
+	nf_flow_offload_add_wq  = alloc_workqueue("nf_ft_offload_add",
+						  WQ_UNBOUND | WQ_SYSFS, 0);
+	if (!nf_flow_offload_add_wq)
 		return -ENOMEM;
 
+	nf_flow_offload_del_wq  = alloc_workqueue("nf_ft_offload_del",
+						  WQ_UNBOUND | WQ_SYSFS, 0);
+	if (!nf_flow_offload_del_wq)
+		goto err_del_wq;
+
+	nf_flow_offload_stats_wq  = alloc_workqueue("nf_ft_offload_stats",
+						    WQ_UNBOUND | WQ_SYSFS, 0);
+	if (!nf_flow_offload_stats_wq)
+		goto err_stats_wq;
+
 	return 0;
+
+err_stats_wq:
+	destroy_workqueue(nf_flow_offload_del_wq);
+err_del_wq:
+	destroy_workqueue(nf_flow_offload_add_wq);
+	return -ENOMEM;
 }
 
 void nf_flow_table_offload_exit(void)
 {
-	destroy_workqueue(nf_flow_offload_wq);
+	destroy_workqueue(nf_flow_offload_add_wq);
+	destroy_workqueue(nf_flow_offload_del_wq);
+	destroy_workqueue(nf_flow_offload_stats_wq);
 }
-- 
2.20.1

Re: [PATCH net-next 01/10] netfilter: flowtable: separate replace, destroy and stats to different workqueues

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

On Tue, 23 Mar 2021 00:56:19 +0100 you wrote:
> From: Oz Shlomo <ozsh@nvidia.com>
> 
> Currently the flow table offload replace, destroy and stats work items are
> executed on a single workqueue. As such, DESTROY and STATS commands may
> be backloged after a burst of REPLACE work items. This scenario can bloat
> up memory and may cause active connections to age.
> 
> [...]

Here is the summary with links:
  - [net-next,01/10] netfilter: flowtable: separate replace, destroy and stats to different workqueues
    https://git.kernel.org/netdev/net-next/c/2ed37183abb7
  - [net-next,02/10] netfilter: Fix fall-through warnings for Clang
    https://git.kernel.org/netdev/net-next/c/c2168e6bd7ec
  - [net-next,03/10] netfilter: conntrack: Remove unused variable declaration
    https://git.kernel.org/netdev/net-next/c/d4a96be65423
  - [net-next,04/10] netfilter: flowtable: consolidate skb_try_make_writable() call
    https://git.kernel.org/netdev/net-next/c/2fc11745c3ff
  - [net-next,05/10] netfilter: flowtable: move skb_try_make_writable() before NAT in IPv4
    https://git.kernel.org/netdev/net-next/c/2babb46c8c82
  - [net-next,06/10] netfilter: flowtable: move FLOW_OFFLOAD_DIR_MAX away from enumeration
    https://git.kernel.org/netdev/net-next/c/4f08f173d08c
  - [net-next,07/10] netfilter: flowtable: fast NAT functions never fail
    https://git.kernel.org/netdev/net-next/c/f4401262b927
  - [net-next,08/10] netfilter: flowtable: call dst_check() to fall back to classic forwarding
    https://git.kernel.org/netdev/net-next/c/e5075c0badaa
  - [net-next,09/10] netfilter: flowtable: refresh timeout after dst and writable checks
    https://git.kernel.org/netdev/net-next/c/1b9cd7690a1e
  - [net-next,10/10] netfilter: nftables: update table flags from the commit phase
    https://git.kernel.org/netdev/net-next/c/0ce7cf4127f1

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

[PATCH net-next 02/10] netfilter: Fix fall-through warnings for Clang

[Pablo Neira Ayuso]

From: "Gustavo A. R. Silva" <gustavoars@kernel.org>

In preparation to enable -Wimplicit-fallthrough for Clang, fix multiple
warnings by explicitly adding multiple break statements instead of just
letting the code fall through to the next case.

Link: https://github.com/KSPP/linux/issues/115
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/netfilter/nf_conntrack_proto_dccp.c | 1 +
 net/netfilter/nf_tables_api.c           | 1 +
 net/netfilter/nft_ct.c                  | 1 +
 3 files changed, 3 insertions(+)

diff --git a/net/netfilter/nf_conntrack_proto_dccp.c b/net/netfilter/nf_conntrack_proto_dccp.c
index db7479db8512..4f33307fa3cf 100644
--- a/net/netfilter/nf_conntrack_proto_dccp.c
+++ b/net/netfilter/nf_conntrack_proto_dccp.c
@@ -397,6 +397,7 @@ dccp_new(struct nf_conn *ct, const struct sk_buff *skb,
 			msg = "not picking up existing connection ";
 			goto out_invalid;
 		}
+		break;
 	case CT_DCCP_REQUEST:
 		break;
 	case CT_DCCP_INVALID:
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 224c8e537cb3..083c112bee0b 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -8557,6 +8557,7 @@ static int nf_tables_check_loops(const struct nft_ctx *ctx,
 							data->verdict.chain);
 				if (err < 0)
 					return err;
+				break;
 			default:
 				break;
 			}
diff --git a/net/netfilter/nft_ct.c b/net/netfilter/nft_ct.c
index 882fe8648653..0592a9456084 100644
--- a/net/netfilter/nft_ct.c
+++ b/net/netfilter/nft_ct.c
@@ -527,6 +527,7 @@ static void __nft_ct_set_destroy(const struct nft_ctx *ctx, struct nft_ct *priv)
 	case NFT_CT_ZONE:
 		if (--nft_ct_pcpu_template_refcnt == 0)
 			nft_ct_tmpl_put_pcpu();
+		break;
 #endif
 	default:
 		break;
-- 
2.20.1

[PATCH net-next 03/10] netfilter: conntrack: Remove unused variable declaration

[Pablo Neira Ayuso]

From: YueHaibing <yuehaibing@huawei.com>

commit e97c3e278e95 ("tproxy: split off ipv6 defragmentation to a separate
module") left behind this.

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 include/net/netfilter/ipv6/nf_conntrack_ipv6.h | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/include/net/netfilter/ipv6/nf_conntrack_ipv6.h b/include/net/netfilter/ipv6/nf_conntrack_ipv6.h
index 7b3c873f8839..e95483192d1b 100644
--- a/include/net/netfilter/ipv6/nf_conntrack_ipv6.h
+++ b/include/net/netfilter/ipv6/nf_conntrack_ipv6.h
@@ -4,7 +4,4 @@
 
 extern const struct nf_conntrack_l4proto nf_conntrack_l4proto_icmpv6;
 
-#include <linux/sysctl.h>
-extern struct ctl_table nf_ct_ipv6_sysctl_table[];
-
 #endif /* _NF_CONNTRACK_IPV6_H*/
-- 
2.20.1

[PATCH net-next 04/10] netfilter: flowtable: consolidate skb_try_make_writable() call

[Pablo Neira Ayuso]

Fetch the layer 4 header size to be mangled by NAT when building the
tuple, then use it to make writable the network and the transport
headers. After this update, the NAT routines now assumes that the skbuff
area is writable. Do the pointer refetch only after the single
skb_try_make_writable() call.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/netfilter/nf_flow_table_core.c | 12 -----
 net/netfilter/nf_flow_table_ip.c   | 74 +++++++++++++-----------------
 2 files changed, 31 insertions(+), 55 deletions(-)

diff --git a/net/netfilter/nf_flow_table_core.c b/net/netfilter/nf_flow_table_core.c
index 5fa657b8e03d..ff5d94b644ac 100644
--- a/net/netfilter/nf_flow_table_core.c
+++ b/net/netfilter/nf_flow_table_core.c
@@ -395,9 +395,6 @@ static int nf_flow_nat_port_tcp(struct sk_buff *skb, unsigned int thoff,
 {
 	struct tcphdr *tcph;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*tcph)))
-		return -1;
-
 	tcph = (void *)(skb_network_header(skb) + thoff);
 	inet_proto_csum_replace2(&tcph->check, skb, port, new_port, false);
 
@@ -409,9 +406,6 @@ static int nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
 {
 	struct udphdr *udph;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*udph)))
-		return -1;
-
 	udph = (void *)(skb_network_header(skb) + thoff);
 	if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
 		inet_proto_csum_replace2(&udph->check, skb, port,
@@ -447,9 +441,6 @@ int nf_flow_snat_port(const struct flow_offload *flow,
 	struct flow_ports *hdr;
 	__be16 port, new_port;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*hdr)))
-		return -1;
-
 	hdr = (void *)(skb_network_header(skb) + thoff);
 
 	switch (dir) {
@@ -478,9 +469,6 @@ int nf_flow_dnat_port(const struct flow_offload *flow,
 	struct flow_ports *hdr;
 	__be16 port, new_port;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*hdr)))
-		return -1;
-
 	hdr = (void *)(skb_network_header(skb) + thoff);
 
 	switch (dir) {
diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c
index a698dbe28ef5..2b8ee5dcef64 100644
--- a/net/netfilter/nf_flow_table_ip.c
+++ b/net/netfilter/nf_flow_table_ip.c
@@ -39,9 +39,6 @@ static int nf_flow_nat_ip_tcp(struct sk_buff *skb, unsigned int thoff,
 {
 	struct tcphdr *tcph;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*tcph)))
-		return -1;
-
 	tcph = (void *)(skb_network_header(skb) + thoff);
 	inet_proto_csum_replace4(&tcph->check, skb, addr, new_addr, true);
 
@@ -53,9 +50,6 @@ static int nf_flow_nat_ip_udp(struct sk_buff *skb, unsigned int thoff,
 {
 	struct udphdr *udph;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*udph)))
-		return -1;
-
 	udph = (void *)(skb_network_header(skb) + thoff);
 	if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
 		inet_proto_csum_replace4(&udph->check, skb, addr,
@@ -136,19 +130,17 @@ static int nf_flow_dnat_ip(const struct flow_offload *flow, struct sk_buff *skb,
 }
 
 static int nf_flow_nat_ip(const struct flow_offload *flow, struct sk_buff *skb,
-			  unsigned int thoff, enum flow_offload_tuple_dir dir)
+			  unsigned int thoff, enum flow_offload_tuple_dir dir,
+			  struct iphdr *iph)
 {
-	struct iphdr *iph = ip_hdr(skb);
-
 	if (test_bit(NF_FLOW_SNAT, &flow->flags) &&
 	    (nf_flow_snat_port(flow, skb, thoff, iph->protocol, dir) < 0 ||
-	     nf_flow_snat_ip(flow, skb, ip_hdr(skb), thoff, dir) < 0))
+	     nf_flow_snat_ip(flow, skb, iph, thoff, dir) < 0))
 		return -1;
 
-	iph = ip_hdr(skb);
 	if (test_bit(NF_FLOW_DNAT, &flow->flags) &&
 	    (nf_flow_dnat_port(flow, skb, thoff, iph->protocol, dir) < 0 ||
-	     nf_flow_dnat_ip(flow, skb, ip_hdr(skb), thoff, dir) < 0))
+	     nf_flow_dnat_ip(flow, skb, iph, thoff, dir) < 0))
 		return -1;
 
 	return 0;
@@ -160,10 +152,10 @@ static bool ip_has_options(unsigned int thoff)
 }
 
 static int nf_flow_tuple_ip(struct sk_buff *skb, const struct net_device *dev,
-			    struct flow_offload_tuple *tuple)
+			    struct flow_offload_tuple *tuple, u32 *hdrsize)
 {
-	unsigned int thoff, hdrsize;
 	struct flow_ports *ports;
+	unsigned int thoff;
 	struct iphdr *iph;
 
 	if (!pskb_may_pull(skb, sizeof(*iph)))
@@ -178,10 +170,10 @@ static int nf_flow_tuple_ip(struct sk_buff *skb, const struct net_device *dev,
 
 	switch (iph->protocol) {
 	case IPPROTO_TCP:
-		hdrsize = sizeof(struct tcphdr);
+		*hdrsize = sizeof(struct tcphdr);
 		break;
 	case IPPROTO_UDP:
-		hdrsize = sizeof(struct udphdr);
+		*hdrsize = sizeof(struct udphdr);
 		break;
 	default:
 		return -1;
@@ -191,7 +183,7 @@ static int nf_flow_tuple_ip(struct sk_buff *skb, const struct net_device *dev,
 		return -1;
 
 	thoff = iph->ihl * 4;
-	if (!pskb_may_pull(skb, thoff + hdrsize))
+	if (!pskb_may_pull(skb, thoff + *hdrsize))
 		return -1;
 
 	iph = ip_hdr(skb);
@@ -252,11 +244,12 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 	unsigned int thoff;
 	struct iphdr *iph;
 	__be32 nexthop;
+	u32 hdrsize;
 
 	if (skb->protocol != htons(ETH_P_IP))
 		return NF_ACCEPT;
 
-	if (nf_flow_tuple_ip(skb, state->in, &tuple) < 0)
+	if (nf_flow_tuple_ip(skb, state->in, &tuple, &hdrsize) < 0)
 		return NF_ACCEPT;
 
 	tuplehash = flow_offload_lookup(flow_table, &tuple);
@@ -271,11 +264,13 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 	if (unlikely(nf_flow_exceeds_mtu(skb, flow->tuplehash[dir].tuple.mtu)))
 		return NF_ACCEPT;
 
-	if (skb_try_make_writable(skb, sizeof(*iph)))
+	iph = ip_hdr(skb);
+	thoff = iph->ihl * 4;
+	if (skb_try_make_writable(skb, thoff + hdrsize))
 		return NF_DROP;
 
-	thoff = ip_hdr(skb)->ihl * 4;
-	if (nf_flow_state_check(flow, ip_hdr(skb)->protocol, skb, thoff))
+	iph = ip_hdr(skb);
+	if (nf_flow_state_check(flow, iph->protocol, skb, thoff))
 		return NF_ACCEPT;
 
 	flow_offload_refresh(flow_table, flow);
@@ -285,10 +280,9 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 		return NF_ACCEPT;
 	}
 
-	if (nf_flow_nat_ip(flow, skb, thoff, dir) < 0)
+	if (nf_flow_nat_ip(flow, skb, thoff, dir, iph) < 0)
 		return NF_DROP;
 
-	iph = ip_hdr(skb);
 	ip_decrease_ttl(iph);
 	skb->tstamp = 0;
 
@@ -317,9 +311,6 @@ static int nf_flow_nat_ipv6_tcp(struct sk_buff *skb, unsigned int thoff,
 {
 	struct tcphdr *tcph;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*tcph)))
-		return -1;
-
 	tcph = (void *)(skb_network_header(skb) + thoff);
 	inet_proto_csum_replace16(&tcph->check, skb, addr->s6_addr32,
 				  new_addr->s6_addr32, true);
@@ -333,9 +324,6 @@ static int nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
 {
 	struct udphdr *udph;
 
-	if (skb_try_make_writable(skb, thoff + sizeof(*udph)))
-		return -1;
-
 	udph = (void *)(skb_network_header(skb) + thoff);
 	if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
 		inet_proto_csum_replace16(&udph->check, skb, addr->s6_addr32,
@@ -417,31 +405,30 @@ static int nf_flow_dnat_ipv6(const struct flow_offload *flow,
 
 static int nf_flow_nat_ipv6(const struct flow_offload *flow,
 			    struct sk_buff *skb,
-			    enum flow_offload_tuple_dir dir)
+			    enum flow_offload_tuple_dir dir,
+			    struct ipv6hdr *ip6h)
 {
-	struct ipv6hdr *ip6h = ipv6_hdr(skb);
 	unsigned int thoff = sizeof(*ip6h);
 
 	if (test_bit(NF_FLOW_SNAT, &flow->flags) &&
 	    (nf_flow_snat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
-	     nf_flow_snat_ipv6(flow, skb, ipv6_hdr(skb), thoff, dir) < 0))
+	     nf_flow_snat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
 		return -1;
 
-	ip6h = ipv6_hdr(skb);
 	if (test_bit(NF_FLOW_DNAT, &flow->flags) &&
 	    (nf_flow_dnat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
-	     nf_flow_dnat_ipv6(flow, skb, ipv6_hdr(skb), thoff, dir) < 0))
+	     nf_flow_dnat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
 		return -1;
 
 	return 0;
 }
 
 static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev,
-			      struct flow_offload_tuple *tuple)
+			      struct flow_offload_tuple *tuple, u32 *hdrsize)
 {
-	unsigned int thoff, hdrsize;
 	struct flow_ports *ports;
 	struct ipv6hdr *ip6h;
+	unsigned int thoff;
 
 	if (!pskb_may_pull(skb, sizeof(*ip6h)))
 		return -1;
@@ -450,10 +437,10 @@ static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev,
 
 	switch (ip6h->nexthdr) {
 	case IPPROTO_TCP:
-		hdrsize = sizeof(struct tcphdr);
+		*hdrsize = sizeof(struct tcphdr);
 		break;
 	case IPPROTO_UDP:
-		hdrsize = sizeof(struct udphdr);
+		*hdrsize = sizeof(struct udphdr);
 		break;
 	default:
 		return -1;
@@ -463,7 +450,7 @@ static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev,
 		return -1;
 
 	thoff = sizeof(*ip6h);
-	if (!pskb_may_pull(skb, thoff + hdrsize))
+	if (!pskb_may_pull(skb, thoff + *hdrsize))
 		return -1;
 
 	ip6h = ipv6_hdr(skb);
@@ -493,11 +480,12 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
 	struct net_device *outdev;
 	struct ipv6hdr *ip6h;
 	struct rt6_info *rt;
+	u32 hdrsize;
 
 	if (skb->protocol != htons(ETH_P_IPV6))
 		return NF_ACCEPT;
 
-	if (nf_flow_tuple_ipv6(skb, state->in, &tuple) < 0)
+	if (nf_flow_tuple_ipv6(skb, state->in, &tuple, &hdrsize) < 0)
 		return NF_ACCEPT;
 
 	tuplehash = flow_offload_lookup(flow_table, &tuple);
@@ -523,13 +511,13 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
 		return NF_ACCEPT;
 	}
 
-	if (skb_try_make_writable(skb, sizeof(*ip6h)))
+	if (skb_try_make_writable(skb, sizeof(*ip6h) + hdrsize))
 		return NF_DROP;
 
-	if (nf_flow_nat_ipv6(flow, skb, dir) < 0)
+	ip6h = ipv6_hdr(skb);
+	if (nf_flow_nat_ipv6(flow, skb, dir, ip6h) < 0)
 		return NF_DROP;
 
-	ip6h = ipv6_hdr(skb);
 	ip6h->hop_limit--;
 	skb->tstamp = 0;
 
-- 
2.20.1

[PATCH net-next 05/10] netfilter: flowtable: move skb_try_make_writable() before NAT in IPv4

[Pablo Neira Ayuso]

For consistency with the IPv6 flowtable datapath and to make sure the
skbuff is writable right before the NAT header updates.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/netfilter/nf_flow_table_ip.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c
index 2b8ee5dcef64..95adf74515ea 100644
--- a/net/netfilter/nf_flow_table_ip.c
+++ b/net/netfilter/nf_flow_table_ip.c
@@ -266,10 +266,6 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 
 	iph = ip_hdr(skb);
 	thoff = iph->ihl * 4;
-	if (skb_try_make_writable(skb, thoff + hdrsize))
-		return NF_DROP;
-
-	iph = ip_hdr(skb);
 	if (nf_flow_state_check(flow, iph->protocol, skb, thoff))
 		return NF_ACCEPT;
 
@@ -280,6 +276,10 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 		return NF_ACCEPT;
 	}
 
+	if (skb_try_make_writable(skb, thoff + hdrsize))
+		return NF_DROP;
+
+	iph = ip_hdr(skb);
 	if (nf_flow_nat_ip(flow, skb, thoff, dir, iph) < 0)
 		return NF_DROP;
 
-- 
2.20.1

[PATCH net-next 06/10] netfilter: flowtable: move FLOW_OFFLOAD_DIR_MAX away from enumeration

[Pablo Neira Ayuso]

This allows to remove the default case which should not ever happen and
that was added to avoid gcc warnings on unhandled FLOW_OFFLOAD_DIR_MAX
enumeration case.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 include/net/netfilter/nf_flow_table.h | 2 +-
 net/netfilter/nf_flow_table_core.c    | 4 ----
 net/netfilter/nf_flow_table_ip.c      | 8 --------
 3 files changed, 1 insertion(+), 13 deletions(-)

diff --git a/include/net/netfilter/nf_flow_table.h b/include/net/netfilter/nf_flow_table.h
index 54c4d5c908a5..ce507251b3d8 100644
--- a/include/net/netfilter/nf_flow_table.h
+++ b/include/net/netfilter/nf_flow_table.h
@@ -86,8 +86,8 @@ static inline bool nf_flowtable_hw_offload(struct nf_flowtable *flowtable)
 enum flow_offload_tuple_dir {
 	FLOW_OFFLOAD_DIR_ORIGINAL = IP_CT_DIR_ORIGINAL,
 	FLOW_OFFLOAD_DIR_REPLY = IP_CT_DIR_REPLY,
-	FLOW_OFFLOAD_DIR_MAX = IP_CT_DIR_MAX
 };
+#define FLOW_OFFLOAD_DIR_MAX	IP_CT_DIR_MAX
 
 struct flow_offload_tuple {
 	union {
diff --git a/net/netfilter/nf_flow_table_core.c b/net/netfilter/nf_flow_table_core.c
index ff5d94b644ac..3bdbd962a084 100644
--- a/net/netfilter/nf_flow_table_core.c
+++ b/net/netfilter/nf_flow_table_core.c
@@ -454,8 +454,6 @@ int nf_flow_snat_port(const struct flow_offload *flow,
 		new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_port;
 		hdr->dest = new_port;
 		break;
-	default:
-		return -1;
 	}
 
 	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
@@ -482,8 +480,6 @@ int nf_flow_dnat_port(const struct flow_offload *flow,
 		new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_port;
 		hdr->source = new_port;
 		break;
-	default:
-		return -1;
 	}
 
 	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c
index 95adf74515ea..0579e15c4968 100644
--- a/net/netfilter/nf_flow_table_ip.c
+++ b/net/netfilter/nf_flow_table_ip.c
@@ -96,8 +96,6 @@ static int nf_flow_snat_ip(const struct flow_offload *flow, struct sk_buff *skb,
 		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v4.s_addr;
 		iph->daddr = new_addr;
 		break;
-	default:
-		return -1;
 	}
 	csum_replace4(&iph->check, addr, new_addr);
 
@@ -121,8 +119,6 @@ static int nf_flow_dnat_ip(const struct flow_offload *flow, struct sk_buff *skb,
 		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v4.s_addr;
 		iph->saddr = new_addr;
 		break;
-	default:
-		return -1;
 	}
 	csum_replace4(&iph->check, addr, new_addr);
 
@@ -371,8 +367,6 @@ static int nf_flow_snat_ipv6(const struct flow_offload *flow,
 		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v6;
 		ip6h->daddr = new_addr;
 		break;
-	default:
-		return -1;
 	}
 
 	return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
@@ -396,8 +390,6 @@ static int nf_flow_dnat_ipv6(const struct flow_offload *flow,
 		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v6;
 		ip6h->saddr = new_addr;
 		break;
-	default:
-		return -1;
 	}
 
 	return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
-- 
2.20.1

[PATCH net-next 07/10] netfilter: flowtable: fast NAT functions never fail

[Pablo Neira Ayuso]

Simplify existing fast NAT routines by returning void. After the
skb_try_make_writable() call consolidation, these routines cannot ever
fail.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 include/net/netfilter/nf_flow_table.h |  12 +--
 net/netfilter/nf_flow_table_core.c    |  41 +++----
 net/netfilter/nf_flow_table_ip.c      | 147 +++++++++++---------------
 3 files changed, 84 insertions(+), 116 deletions(-)

diff --git a/include/net/netfilter/nf_flow_table.h b/include/net/netfilter/nf_flow_table.h
index ce507251b3d8..fb165697c8a1 100644
--- a/include/net/netfilter/nf_flow_table.h
+++ b/include/net/netfilter/nf_flow_table.h
@@ -229,12 +229,12 @@ void nf_flow_table_free(struct nf_flowtable *flow_table);
 
 void flow_offload_teardown(struct flow_offload *flow);
 
-int nf_flow_snat_port(const struct flow_offload *flow,
-		      struct sk_buff *skb, unsigned int thoff,
-		      u8 protocol, enum flow_offload_tuple_dir dir);
-int nf_flow_dnat_port(const struct flow_offload *flow,
-		      struct sk_buff *skb, unsigned int thoff,
-		      u8 protocol, enum flow_offload_tuple_dir dir);
+void nf_flow_snat_port(const struct flow_offload *flow,
+		       struct sk_buff *skb, unsigned int thoff,
+		       u8 protocol, enum flow_offload_tuple_dir dir);
+void nf_flow_dnat_port(const struct flow_offload *flow,
+		       struct sk_buff *skb, unsigned int thoff,
+		       u8 protocol, enum flow_offload_tuple_dir dir);
 
 struct flow_ports {
 	__be16 source, dest;
diff --git a/net/netfilter/nf_flow_table_core.c b/net/netfilter/nf_flow_table_core.c
index 3bdbd962a084..8ffd3f3c288c 100644
--- a/net/netfilter/nf_flow_table_core.c
+++ b/net/netfilter/nf_flow_table_core.c
@@ -389,20 +389,17 @@ static void nf_flow_offload_work_gc(struct work_struct *work)
 	queue_delayed_work(system_power_efficient_wq, &flow_table->gc_work, HZ);
 }
 
-
-static int nf_flow_nat_port_tcp(struct sk_buff *skb, unsigned int thoff,
-				__be16 port, __be16 new_port)
+static void nf_flow_nat_port_tcp(struct sk_buff *skb, unsigned int thoff,
+				 __be16 port, __be16 new_port)
 {
 	struct tcphdr *tcph;
 
 	tcph = (void *)(skb_network_header(skb) + thoff);
 	inet_proto_csum_replace2(&tcph->check, skb, port, new_port, false);
-
-	return 0;
 }
 
-static int nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
-				__be16 port, __be16 new_port)
+static void nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
+				 __be16 port, __be16 new_port)
 {
 	struct udphdr *udph;
 
@@ -413,30 +410,24 @@ static int nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
 		if (!udph->check)
 			udph->check = CSUM_MANGLED_0;
 	}
-
-	return 0;
 }
 
-static int nf_flow_nat_port(struct sk_buff *skb, unsigned int thoff,
-			    u8 protocol, __be16 port, __be16 new_port)
+static void nf_flow_nat_port(struct sk_buff *skb, unsigned int thoff,
+			     u8 protocol, __be16 port, __be16 new_port)
 {
 	switch (protocol) {
 	case IPPROTO_TCP:
-		if (nf_flow_nat_port_tcp(skb, thoff, port, new_port) < 0)
-			return NF_DROP;
+		nf_flow_nat_port_tcp(skb, thoff, port, new_port);
 		break;
 	case IPPROTO_UDP:
-		if (nf_flow_nat_port_udp(skb, thoff, port, new_port) < 0)
-			return NF_DROP;
+		nf_flow_nat_port_udp(skb, thoff, port, new_port);
 		break;
 	}
-
-	return 0;
 }
 
-int nf_flow_snat_port(const struct flow_offload *flow,
-		      struct sk_buff *skb, unsigned int thoff,
-		      u8 protocol, enum flow_offload_tuple_dir dir)
+void nf_flow_snat_port(const struct flow_offload *flow,
+		       struct sk_buff *skb, unsigned int thoff,
+		       u8 protocol, enum flow_offload_tuple_dir dir)
 {
 	struct flow_ports *hdr;
 	__be16 port, new_port;
@@ -456,13 +447,13 @@ int nf_flow_snat_port(const struct flow_offload *flow,
 		break;
 	}
 
-	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
+	nf_flow_nat_port(skb, thoff, protocol, port, new_port);
 }
 EXPORT_SYMBOL_GPL(nf_flow_snat_port);
 
-int nf_flow_dnat_port(const struct flow_offload *flow,
-		      struct sk_buff *skb, unsigned int thoff,
-		      u8 protocol, enum flow_offload_tuple_dir dir)
+void nf_flow_dnat_port(const struct flow_offload *flow, struct sk_buff *skb,
+		       unsigned int thoff, u8 protocol,
+		       enum flow_offload_tuple_dir dir)
 {
 	struct flow_ports *hdr;
 	__be16 port, new_port;
@@ -482,7 +473,7 @@ int nf_flow_dnat_port(const struct flow_offload *flow,
 		break;
 	}
 
-	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
+	nf_flow_nat_port(skb, thoff, protocol, port, new_port);
 }
 EXPORT_SYMBOL_GPL(nf_flow_dnat_port);
 
diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c
index 0579e15c4968..714dc083f093 100644
--- a/net/netfilter/nf_flow_table_ip.c
+++ b/net/netfilter/nf_flow_table_ip.c
@@ -34,19 +34,17 @@ static int nf_flow_state_check(struct flow_offload *flow, int proto,
 	return 0;
 }
 
-static int nf_flow_nat_ip_tcp(struct sk_buff *skb, unsigned int thoff,
-			      __be32 addr, __be32 new_addr)
+static void nf_flow_nat_ip_tcp(struct sk_buff *skb, unsigned int thoff,
+			       __be32 addr, __be32 new_addr)
 {
 	struct tcphdr *tcph;
 
 	tcph = (void *)(skb_network_header(skb) + thoff);
 	inet_proto_csum_replace4(&tcph->check, skb, addr, new_addr, true);
-
-	return 0;
 }
 
-static int nf_flow_nat_ip_udp(struct sk_buff *skb, unsigned int thoff,
-			      __be32 addr, __be32 new_addr)
+static void nf_flow_nat_ip_udp(struct sk_buff *skb, unsigned int thoff,
+			       __be32 addr, __be32 new_addr)
 {
 	struct udphdr *udph;
 
@@ -57,31 +55,25 @@ static int nf_flow_nat_ip_udp(struct sk_buff *skb, unsigned int thoff,
 		if (!udph->check)
 			udph->check = CSUM_MANGLED_0;
 	}
-
-	return 0;
 }
 
-static int nf_flow_nat_ip_l4proto(struct sk_buff *skb, struct iphdr *iph,
-				  unsigned int thoff, __be32 addr,
-				  __be32 new_addr)
+static void nf_flow_nat_ip_l4proto(struct sk_buff *skb, struct iphdr *iph,
+				   unsigned int thoff, __be32 addr,
+				   __be32 new_addr)
 {
 	switch (iph->protocol) {
 	case IPPROTO_TCP:
-		if (nf_flow_nat_ip_tcp(skb, thoff, addr, new_addr) < 0)
-			return NF_DROP;
+		nf_flow_nat_ip_tcp(skb, thoff, addr, new_addr);
 		break;
 	case IPPROTO_UDP:
-		if (nf_flow_nat_ip_udp(skb, thoff, addr, new_addr) < 0)
-			return NF_DROP;
+		nf_flow_nat_ip_udp(skb, thoff, addr, new_addr);
 		break;
 	}
-
-	return 0;
 }
 
-static int nf_flow_snat_ip(const struct flow_offload *flow, struct sk_buff *skb,
-			   struct iphdr *iph, unsigned int thoff,
-			   enum flow_offload_tuple_dir dir)
+static void nf_flow_snat_ip(const struct flow_offload *flow,
+			    struct sk_buff *skb, struct iphdr *iph,
+			    unsigned int thoff, enum flow_offload_tuple_dir dir)
 {
 	__be32 addr, new_addr;
 
@@ -99,12 +91,12 @@ static int nf_flow_snat_ip(const struct flow_offload *flow, struct sk_buff *skb,
 	}
 	csum_replace4(&iph->check, addr, new_addr);
 
-	return nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
+	nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
 }
 
-static int nf_flow_dnat_ip(const struct flow_offload *flow, struct sk_buff *skb,
-			   struct iphdr *iph, unsigned int thoff,
-			   enum flow_offload_tuple_dir dir)
+static void nf_flow_dnat_ip(const struct flow_offload *flow,
+			    struct sk_buff *skb, struct iphdr *iph,
+			    unsigned int thoff, enum flow_offload_tuple_dir dir)
 {
 	__be32 addr, new_addr;
 
@@ -122,24 +114,21 @@ static int nf_flow_dnat_ip(const struct flow_offload *flow, struct sk_buff *skb,
 	}
 	csum_replace4(&iph->check, addr, new_addr);
 
-	return nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
+	nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
 }
 
-static int nf_flow_nat_ip(const struct flow_offload *flow, struct sk_buff *skb,
+static void nf_flow_nat_ip(const struct flow_offload *flow, struct sk_buff *skb,
 			  unsigned int thoff, enum flow_offload_tuple_dir dir,
 			  struct iphdr *iph)
 {
-	if (test_bit(NF_FLOW_SNAT, &flow->flags) &&
-	    (nf_flow_snat_port(flow, skb, thoff, iph->protocol, dir) < 0 ||
-	     nf_flow_snat_ip(flow, skb, iph, thoff, dir) < 0))
-		return -1;
-
-	if (test_bit(NF_FLOW_DNAT, &flow->flags) &&
-	    (nf_flow_dnat_port(flow, skb, thoff, iph->protocol, dir) < 0 ||
-	     nf_flow_dnat_ip(flow, skb, iph, thoff, dir) < 0))
-		return -1;
-
-	return 0;
+	if (test_bit(NF_FLOW_SNAT, &flow->flags)) {
+		nf_flow_snat_port(flow, skb, thoff, iph->protocol, dir);
+		nf_flow_snat_ip(flow, skb, iph, thoff, dir);
+	}
+	if (test_bit(NF_FLOW_DNAT, &flow->flags)) {
+		nf_flow_dnat_port(flow, skb, thoff, iph->protocol, dir);
+		nf_flow_dnat_ip(flow, skb, iph, thoff, dir);
+	}
 }
 
 static bool ip_has_options(unsigned int thoff)
@@ -276,8 +265,7 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 		return NF_DROP;
 
 	iph = ip_hdr(skb);
-	if (nf_flow_nat_ip(flow, skb, thoff, dir, iph) < 0)
-		return NF_DROP;
+	nf_flow_nat_ip(flow, skb, thoff, dir, iph);
 
 	ip_decrease_ttl(iph);
 	skb->tstamp = 0;
@@ -301,22 +289,21 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 }
 EXPORT_SYMBOL_GPL(nf_flow_offload_ip_hook);
 
-static int nf_flow_nat_ipv6_tcp(struct sk_buff *skb, unsigned int thoff,
-				struct in6_addr *addr,
-				struct in6_addr *new_addr)
+static void nf_flow_nat_ipv6_tcp(struct sk_buff *skb, unsigned int thoff,
+				 struct in6_addr *addr,
+				 struct in6_addr *new_addr,
+				 struct ipv6hdr *ip6h)
 {
 	struct tcphdr *tcph;
 
 	tcph = (void *)(skb_network_header(skb) + thoff);
 	inet_proto_csum_replace16(&tcph->check, skb, addr->s6_addr32,
 				  new_addr->s6_addr32, true);
-
-	return 0;
 }
 
-static int nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
-				struct in6_addr *addr,
-				struct in6_addr *new_addr)
+static void nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
+				 struct in6_addr *addr,
+				 struct in6_addr *new_addr)
 {
 	struct udphdr *udph;
 
@@ -327,32 +314,26 @@ static int nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
 		if (!udph->check)
 			udph->check = CSUM_MANGLED_0;
 	}
-
-	return 0;
 }
 
-static int nf_flow_nat_ipv6_l4proto(struct sk_buff *skb, struct ipv6hdr *ip6h,
-				    unsigned int thoff, struct in6_addr *addr,
-				    struct in6_addr *new_addr)
+static void nf_flow_nat_ipv6_l4proto(struct sk_buff *skb, struct ipv6hdr *ip6h,
+				     unsigned int thoff, struct in6_addr *addr,
+				     struct in6_addr *new_addr)
 {
 	switch (ip6h->nexthdr) {
 	case IPPROTO_TCP:
-		if (nf_flow_nat_ipv6_tcp(skb, thoff, addr, new_addr) < 0)
-			return NF_DROP;
+		nf_flow_nat_ipv6_tcp(skb, thoff, addr, new_addr, ip6h);
 		break;
 	case IPPROTO_UDP:
-		if (nf_flow_nat_ipv6_udp(skb, thoff, addr, new_addr) < 0)
-			return NF_DROP;
+		nf_flow_nat_ipv6_udp(skb, thoff, addr, new_addr);
 		break;
 	}
-
-	return 0;
 }
 
-static int nf_flow_snat_ipv6(const struct flow_offload *flow,
-			     struct sk_buff *skb, struct ipv6hdr *ip6h,
-			     unsigned int thoff,
-			     enum flow_offload_tuple_dir dir)
+static void nf_flow_snat_ipv6(const struct flow_offload *flow,
+			      struct sk_buff *skb, struct ipv6hdr *ip6h,
+			      unsigned int thoff,
+			      enum flow_offload_tuple_dir dir)
 {
 	struct in6_addr addr, new_addr;
 
@@ -369,13 +350,13 @@ static int nf_flow_snat_ipv6(const struct flow_offload *flow,
 		break;
 	}
 
-	return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
+	nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
 }
 
-static int nf_flow_dnat_ipv6(const struct flow_offload *flow,
-			     struct sk_buff *skb, struct ipv6hdr *ip6h,
-			     unsigned int thoff,
-			     enum flow_offload_tuple_dir dir)
+static void nf_flow_dnat_ipv6(const struct flow_offload *flow,
+			      struct sk_buff *skb, struct ipv6hdr *ip6h,
+			      unsigned int thoff,
+			      enum flow_offload_tuple_dir dir)
 {
 	struct in6_addr addr, new_addr;
 
@@ -392,27 +373,24 @@ static int nf_flow_dnat_ipv6(const struct flow_offload *flow,
 		break;
 	}
 
-	return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
+	nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
 }
 
-static int nf_flow_nat_ipv6(const struct flow_offload *flow,
-			    struct sk_buff *skb,
-			    enum flow_offload_tuple_dir dir,
-			    struct ipv6hdr *ip6h)
+static void nf_flow_nat_ipv6(const struct flow_offload *flow,
+			     struct sk_buff *skb,
+			     enum flow_offload_tuple_dir dir,
+			     struct ipv6hdr *ip6h)
 {
 	unsigned int thoff = sizeof(*ip6h);
 
-	if (test_bit(NF_FLOW_SNAT, &flow->flags) &&
-	    (nf_flow_snat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
-	     nf_flow_snat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
-		return -1;
-
-	if (test_bit(NF_FLOW_DNAT, &flow->flags) &&
-	    (nf_flow_dnat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
-	     nf_flow_dnat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
-		return -1;
-
-	return 0;
+	if (test_bit(NF_FLOW_SNAT, &flow->flags)) {
+		nf_flow_snat_port(flow, skb, thoff, ip6h->nexthdr, dir);
+		nf_flow_snat_ipv6(flow, skb, ip6h, thoff, dir);
+	}
+	if (test_bit(NF_FLOW_DNAT, &flow->flags)) {
+		nf_flow_dnat_port(flow, skb, thoff, ip6h->nexthdr, dir);
+		nf_flow_dnat_ipv6(flow, skb, ip6h, thoff, dir);
+	}
 }
 
 static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev,
@@ -507,8 +485,7 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
 		return NF_DROP;
 
 	ip6h = ipv6_hdr(skb);
-	if (nf_flow_nat_ipv6(flow, skb, dir, ip6h) < 0)
-		return NF_DROP;
+	nf_flow_nat_ipv6(flow, skb, dir, ip6h);
 
 	ip6h->hop_limit--;
 	skb->tstamp = 0;
-- 
2.20.1

[PATCH net-next 08/10] netfilter: flowtable: call dst_check() to fall back to classic forwarding

[Pablo Neira Ayuso]

In case the route is stale, pass up the packet to the classic forwarding
path for re-evaluation and schedule this flow entry for removal.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/netfilter/nf_flow_table_ip.c | 12 ++----------
 1 file changed, 2 insertions(+), 10 deletions(-)

diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c
index 714dc083f093..3a8423899def 100644
--- a/net/netfilter/nf_flow_table_ip.c
+++ b/net/netfilter/nf_flow_table_ip.c
@@ -197,14 +197,6 @@ static bool nf_flow_exceeds_mtu(const struct sk_buff *skb, unsigned int mtu)
 	return true;
 }
 
-static int nf_flow_offload_dst_check(struct dst_entry *dst)
-{
-	if (unlikely(dst_xfrm(dst)))
-		return dst_check(dst, 0) ? 0 : -1;
-
-	return 0;
-}
-
 static unsigned int nf_flow_xmit_xfrm(struct sk_buff *skb,
 				      const struct nf_hook_state *state,
 				      struct dst_entry *dst)
@@ -256,7 +248,7 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 
 	flow_offload_refresh(flow_table, flow);
 
-	if (nf_flow_offload_dst_check(&rt->dst)) {
+	if (!dst_check(&rt->dst, 0)) {
 		flow_offload_teardown(flow);
 		return NF_ACCEPT;
 	}
@@ -476,7 +468,7 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
 
 	flow_offload_refresh(flow_table, flow);
 
-	if (nf_flow_offload_dst_check(&rt->dst)) {
+	if (!dst_check(&rt->dst, 0)) {
 		flow_offload_teardown(flow);
 		return NF_ACCEPT;
 	}
-- 
2.20.1

[PATCH net-next 09/10] netfilter: flowtable: refresh timeout after dst and writable checks

[Pablo Neira Ayuso]

Refresh the timeout (and retry hardware offload) once the skbuff dst
is confirmed to be current and after the skbuff is made writable.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/netfilter/nf_flow_table_ip.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c
index 3a8423899def..3be58b6d60af 100644
--- a/net/netfilter/nf_flow_table_ip.c
+++ b/net/netfilter/nf_flow_table_ip.c
@@ -246,8 +246,6 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 	if (nf_flow_state_check(flow, iph->protocol, skb, thoff))
 		return NF_ACCEPT;
 
-	flow_offload_refresh(flow_table, flow);
-
 	if (!dst_check(&rt->dst, 0)) {
 		flow_offload_teardown(flow);
 		return NF_ACCEPT;
@@ -256,6 +254,8 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
 	if (skb_try_make_writable(skb, thoff + hdrsize))
 		return NF_DROP;
 
+	flow_offload_refresh(flow_table, flow);
+
 	iph = ip_hdr(skb);
 	nf_flow_nat_ip(flow, skb, thoff, dir, iph);
 
@@ -466,8 +466,6 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
 				sizeof(*ip6h)))
 		return NF_ACCEPT;
 
-	flow_offload_refresh(flow_table, flow);
-
 	if (!dst_check(&rt->dst, 0)) {
 		flow_offload_teardown(flow);
 		return NF_ACCEPT;
@@ -476,6 +474,8 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
 	if (skb_try_make_writable(skb, sizeof(*ip6h) + hdrsize))
 		return NF_DROP;
 
+	flow_offload_refresh(flow_table, flow);
+
 	ip6h = ipv6_hdr(skb);
 	nf_flow_nat_ipv6(flow, skb, dir, ip6h);
 
-- 
2.20.1

[PATCH net-next 10/10] netfilter: nftables: update table flags from the commit phase

[Pablo Neira Ayuso]

Do not update table flags from the preparation phase. Store the flags
update into the transaction, then update the flags from the commit
phase.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 include/net/netfilter/nf_tables.h |  9 ++++++---
 net/netfilter/nf_tables_api.c     | 31 ++++++++++++++++---------------
 2 files changed, 22 insertions(+), 18 deletions(-)

diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index fdec57d862b7..67bc36f7f4fb 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1498,13 +1498,16 @@ struct nft_trans_chain {
 
 struct nft_trans_table {
 	bool				update;
-	bool				enable;
+	u8				state;
+	u32				flags;
 };
 
 #define nft_trans_table_update(trans)	\
 	(((struct nft_trans_table *)trans->data)->update)
-#define nft_trans_table_enable(trans)	\
-	(((struct nft_trans_table *)trans->data)->enable)
+#define nft_trans_table_state(trans)	\
+	(((struct nft_trans_table *)trans->data)->state)
+#define nft_trans_table_flags(trans)	\
+	(((struct nft_trans_table *)trans->data)->flags)
 
 struct nft_trans_elem {
 	struct nft_set			*set;
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 083c112bee0b..bd5e8122ea5e 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -900,6 +900,12 @@ static void nf_tables_table_disable(struct net *net, struct nft_table *table)
 	nft_table_disable(net, table, 0);
 }
 
+enum {
+	NFT_TABLE_STATE_UNCHANGED	= 0,
+	NFT_TABLE_STATE_DORMANT,
+	NFT_TABLE_STATE_WAKEUP
+};
+
 static int nf_tables_updtable(struct nft_ctx *ctx)
 {
 	struct nft_trans *trans;
@@ -929,19 +935,17 @@ static int nf_tables_updtable(struct nft_ctx *ctx)
 
 	if ((flags & NFT_TABLE_F_DORMANT) &&
 	    !(ctx->table->flags & NFT_TABLE_F_DORMANT)) {
-		nft_trans_table_enable(trans) = false;
+		nft_trans_table_state(trans) = NFT_TABLE_STATE_DORMANT;
 	} else if (!(flags & NFT_TABLE_F_DORMANT) &&
 		   ctx->table->flags & NFT_TABLE_F_DORMANT) {
-		ctx->table->flags &= ~NFT_TABLE_F_DORMANT;
 		ret = nf_tables_table_enable(ctx->net, ctx->table);
 		if (ret >= 0)
-			nft_trans_table_enable(trans) = true;
-		else
-			ctx->table->flags |= NFT_TABLE_F_DORMANT;
+			nft_trans_table_state(trans) = NFT_TABLE_STATE_WAKEUP;
 	}
 	if (ret < 0)
 		goto err;
 
+	nft_trans_table_flags(trans) = flags;
 	nft_trans_table_update(trans) = true;
 	list_add_tail(&trans->list, &ctx->net->nft.commit_list);
 	return 0;
@@ -8068,11 +8072,10 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
 		switch (trans->msg_type) {
 		case NFT_MSG_NEWTABLE:
 			if (nft_trans_table_update(trans)) {
-				if (!nft_trans_table_enable(trans)) {
-					nf_tables_table_disable(net,
-								trans->ctx.table);
-					trans->ctx.table->flags |= NFT_TABLE_F_DORMANT;
-				}
+				if (nft_trans_table_state(trans) == NFT_TABLE_STATE_DORMANT)
+					nf_tables_table_disable(net, trans->ctx.table);
+
+				trans->ctx.table->flags = nft_trans_table_flags(trans);
 			} else {
 				nft_clear(net, trans->ctx.table);
 			}
@@ -8283,11 +8286,9 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
 		switch (trans->msg_type) {
 		case NFT_MSG_NEWTABLE:
 			if (nft_trans_table_update(trans)) {
-				if (nft_trans_table_enable(trans)) {
-					nf_tables_table_disable(net,
-								trans->ctx.table);
-					trans->ctx.table->flags |= NFT_TABLE_F_DORMANT;
-				}
+				if (nft_trans_table_state(trans) == NFT_TABLE_STATE_WAKEUP)
+					nf_tables_table_disable(net, trans->ctx.table);
+
 				nft_trans_destroy(trans);
 			} else {
 				list_del_rcu(&trans->ctx.table->list);
-- 
2.20.1