From 34d94f2102fb361030569eb983751ed80742f6e9 Mon Sep 17 00:00:00 2001
From: Lorand Jakab <lojakab@cisco.com>
Date: Mon, 3 Jun 2013 10:01:14 -0700
Subject: openvswitch: fix variable names in comment

Signed-off-by: Lorand Jakab <lojakab@cisco.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
---
 net/openvswitch/flow.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'net/openvswitch/flow.c')

diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index b15321a2228..33df0913358 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -590,10 +590,10 @@ out:
  *    - skb->network_header: just past the Ethernet header, or just past the
  *      VLAN header, to the first byte of the Ethernet payload.
  *
- *    - skb->transport_header: If key->dl_type is ETH_P_IP or ETH_P_IPV6
+ *    - skb->transport_header: If key->eth.type is ETH_P_IP or ETH_P_IPV6
  *      on output, then just past the IP header, if one is present and
  *      of a correct length, otherwise the same as skb->network_header.
- *      For other key->dl_type values it is left untouched.
+ *      For other key->eth.type values it is left untouched.
  */
 int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key,
 		 int *key_lenp)
-- 
cgit v1.2.3-70-g09d2


From b34df5e805a6e98cae0bc5bc80c1b52d9ff811de Mon Sep 17 00:00:00 2001
From: Pravin B Shelar <pshelar@nicira.com>
Date: Thu, 13 Jun 2013 11:11:44 -0700
Subject: openvswitch: make skb->csum consistent with rest of networking stack.

Following patch keeps skb->csum correct across ovs.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
---
 net/openvswitch/actions.c            | 4 ++++
 net/openvswitch/flow.c               | 3 +++
 net/openvswitch/vport-internal_dev.c | 1 +
 net/openvswitch/vport-netdev.c       | 2 ++
 net/openvswitch/vport.h              | 7 +++++++
 5 files changed, 17 insertions(+)

(limited to 'net/openvswitch/flow.c')

diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c
index 894b6cbdd92..596d6373399 100644
--- a/net/openvswitch/actions.c
+++ b/net/openvswitch/actions.c
@@ -130,9 +130,13 @@ static int set_eth_addr(struct sk_buff *skb,
 	if (unlikely(err))
 		return err;
 
+	skb_postpull_rcsum(skb, eth_hdr(skb), ETH_ALEN * 2);
+
 	memcpy(eth_hdr(skb)->h_source, eth_key->eth_src, ETH_ALEN);
 	memcpy(eth_hdr(skb)->h_dest, eth_key->eth_dst, ETH_ALEN);
 
+	ovs_skb_postpush_rcsum(skb, eth_hdr(skb), ETH_ALEN * 2);
+
 	return 0;
 }
 
diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index 33df0913358..fca483360ce 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -618,6 +618,9 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key,
 	memcpy(key->eth.dst, eth->h_dest, ETH_ALEN);
 
 	__skb_pull(skb, 2 * ETH_ALEN);
+	/* We are going to push all headers that we pull, so no need to
+	 * update skb->csum here.
+	 */
 
 	if (vlan_tx_tag_present(skb))
 		key->eth.tci = htons(skb->vlan_tci);
diff --git a/net/openvswitch/vport-internal_dev.c b/net/openvswitch/vport-internal_dev.c
index 84e0a037918..e284c7e1fec 100644
--- a/net/openvswitch/vport-internal_dev.c
+++ b/net/openvswitch/vport-internal_dev.c
@@ -221,6 +221,7 @@ static int internal_dev_recv(struct vport *vport, struct sk_buff *skb)
 	skb->dev = netdev;
 	skb->pkt_type = PACKET_HOST;
 	skb->protocol = eth_type_trans(skb, netdev);
+	skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
 
 	netif_rx(skb);
 
diff --git a/net/openvswitch/vport-netdev.c b/net/openvswitch/vport-netdev.c
index 43712217a37..40de815b421 100644
--- a/net/openvswitch/vport-netdev.c
+++ b/net/openvswitch/vport-netdev.c
@@ -49,6 +49,8 @@ static void netdev_port_receive(struct vport *vport, struct sk_buff *skb)
 		return;
 
 	skb_push(skb, ETH_HLEN);
+	ovs_skb_postpush_rcsum(skb, skb->data, ETH_HLEN);
+
 	ovs_vport_receive(vport, skb);
 	return;
 
diff --git a/net/openvswitch/vport.h b/net/openvswitch/vport.h
index 1cef5cd3be4..293278c4c2d 100644
--- a/net/openvswitch/vport.h
+++ b/net/openvswitch/vport.h
@@ -192,4 +192,11 @@ void ovs_vport_record_error(struct vport *, enum vport_err_type err_type);
 extern const struct vport_ops ovs_netdev_vport_ops;
 extern const struct vport_ops ovs_internal_vport_ops;
 
+static inline void ovs_skb_postpush_rcsum(struct sk_buff *skb,
+				      const void *start, unsigned int len)
+{
+	if (skb->ip_summed == CHECKSUM_COMPLETE)
+		skb->csum = csum_add(skb->csum, csum_partial(start, len, 0));
+}
+
 #endif /* vport.h */
-- 
cgit v1.2.3-70-g09d2


From 93d8fd1514b6862c3370ea92be3f3b4216e0bf8f Mon Sep 17 00:00:00 2001
From: Pravin B Shelar <pshelar@nicira.com>
Date: Thu, 13 Jun 2013 11:11:32 -0700
Subject: openvswitch: Simplify interface ovs_flow_metadata_from_nlattrs()

This is not functional change, this is just code cleanup.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
---
 net/openvswitch/datapath.c |  5 +----
 net/openvswitch/flow.c     | 22 ++++++++++------------
 net/openvswitch/flow.h     |  4 ++--
 3 files changed, 13 insertions(+), 18 deletions(-)

(limited to 'net/openvswitch/flow.c')

diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
index 748aa97cbfb..0f783d9fa00 100644
--- a/net/openvswitch/datapath.c
+++ b/net/openvswitch/datapath.c
@@ -739,10 +739,7 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info)
 	if (err)
 		goto err_flow_free;
 
-	err = ovs_flow_metadata_from_nlattrs(&flow->key.phy.priority,
-					     &flow->key.phy.skb_mark,
-					     &flow->key.phy.in_port,
-					     a[OVS_PACKET_ATTR_KEY]);
+	err = ovs_flow_metadata_from_nlattrs(flow, a[OVS_PACKET_ATTR_KEY]);
 	if (err)
 		goto err_flow_free;
 
diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index fca483360ce..093c191d4fc 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -1125,10 +1125,8 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
 
 /**
  * ovs_flow_metadata_from_nlattrs - parses Netlink attributes into a flow key.
- * @priority: receives the skb priority
- * @mark: receives the skb mark
- * @in_port: receives the extracted input port.
- * @key: Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink attribute
+ * @flow: Receives extracted in_port, priority, tun_key and skb_mark.
+ * @attr: Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink attribute
  * sequence.
  *
  * This parses a series of Netlink attributes that form a flow key, which must
@@ -1136,15 +1134,15 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
  * get the metadata, that is, the parts of the flow key that cannot be
  * extracted from the packet itself.
  */
-int ovs_flow_metadata_from_nlattrs(u32 *priority, u32 *mark, u16 *in_port,
-			       const struct nlattr *attr)
+int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
+				   const struct nlattr *attr)
 {
 	const struct nlattr *nla;
 	int rem;
 
-	*in_port = DP_MAX_PORTS;
-	*priority = 0;
-	*mark = 0;
+	flow->key.phy.in_port = DP_MAX_PORTS;
+	flow->key.phy.priority = 0;
+	flow->key.phy.skb_mark = 0;
 
 	nla_for_each_nested(nla, attr, rem) {
 		int type = nla_type(nla);
@@ -1155,17 +1153,17 @@ int ovs_flow_metadata_from_nlattrs(u32 *priority, u32 *mark, u16 *in_port,
 
 			switch (type) {
 			case OVS_KEY_ATTR_PRIORITY:
-				*priority = nla_get_u32(nla);
+				flow->key.phy.priority = nla_get_u32(nla);
 				break;
 
 			case OVS_KEY_ATTR_IN_PORT:
 				if (nla_get_u32(nla) >= DP_MAX_PORTS)
 					return -EINVAL;
-				*in_port = nla_get_u32(nla);
+				flow->key.phy.in_port = nla_get_u32(nla);
 				break;
 
 			case OVS_KEY_ATTR_SKB_MARK:
-				*mark = nla_get_u32(nla);
+				flow->key.phy.skb_mark = nla_get_u32(nla);
 				break;
 			}
 		}
diff --git a/net/openvswitch/flow.h b/net/openvswitch/flow.h
index 0875fde65b9..2a83e2141f0 100644
--- a/net/openvswitch/flow.h
+++ b/net/openvswitch/flow.h
@@ -141,8 +141,8 @@ u64 ovs_flow_used_time(unsigned long flow_jiffies);
 int ovs_flow_to_nlattrs(const struct sw_flow_key *, struct sk_buff *);
 int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
 		      const struct nlattr *);
-int ovs_flow_metadata_from_nlattrs(u32 *priority, u32 *mark, u16 *in_port,
-			       const struct nlattr *);
+int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
+				  const struct nlattr *attr);
 
 #define MAX_ACTIONS_BUFSIZE    (16 * 1024)
 #define TBL_MIN_BUCKETS		1024
-- 
cgit v1.2.3-70-g09d2


From 74f84a5726c7d08c27745305e67474b8645c541d Mon Sep 17 00:00:00 2001
From: Pravin B Shelar <pshelar@nicira.com>
Date: Mon, 17 Jun 2013 17:50:12 -0700
Subject: openvswitch: Copy individual actions.

Rather than validating actions and then copying all actiaons
in one block, following patch does same operation in single pass.
This validate and copy action one by one. This is required for
ovs tunneling patch.

This patch does not change any functionality.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 net/openvswitch/datapath.c | 271 ++++++++++++++++++++++++++++++++++++---------
 net/openvswitch/flow.c     |  10 +-
 net/openvswitch/flow.h     |   2 +-
 3 files changed, 225 insertions(+), 58 deletions(-)

(limited to 'net/openvswitch/flow.c')

diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
index 0f783d9fa00..f14816b80b8 100644
--- a/net/openvswitch/datapath.c
+++ b/net/openvswitch/datapath.c
@@ -464,16 +464,89 @@ static int flush_flows(struct datapath *dp)
 	return 0;
 }
 
-static int validate_actions(const struct nlattr *attr,
-				const struct sw_flow_key *key, int depth);
+static struct nlattr *reserve_sfa_size(struct sw_flow_actions **sfa, int attr_len)
+{
+
+	struct sw_flow_actions *acts;
+	int new_acts_size;
+	int req_size = NLA_ALIGN(attr_len);
+	int next_offset = offsetof(struct sw_flow_actions, actions) +
+					(*sfa)->actions_len;
+
+	if (req_size <= (ksize(*sfa) - next_offset))
+		goto out;
+
+	new_acts_size = ksize(*sfa) * 2;
+
+	if (new_acts_size > MAX_ACTIONS_BUFSIZE) {
+		if ((MAX_ACTIONS_BUFSIZE - next_offset) < req_size)
+			return ERR_PTR(-EMSGSIZE);
+		new_acts_size = MAX_ACTIONS_BUFSIZE;
+	}
+
+	acts = ovs_flow_actions_alloc(new_acts_size);
+	if (IS_ERR(acts))
+		return (void *)acts;
+
+	memcpy(acts->actions, (*sfa)->actions, (*sfa)->actions_len);
+	acts->actions_len = (*sfa)->actions_len;
+	kfree(*sfa);
+	*sfa = acts;
+
+out:
+	(*sfa)->actions_len += req_size;
+	return  (struct nlattr *) ((unsigned char *)(*sfa) + next_offset);
+}
+
+static int add_action(struct sw_flow_actions **sfa, int attrtype, void *data, int len)
+{
+	struct nlattr *a;
+
+	a = reserve_sfa_size(sfa, nla_attr_size(len));
+	if (IS_ERR(a))
+		return PTR_ERR(a);
+
+	a->nla_type = attrtype;
+	a->nla_len = nla_attr_size(len);
+
+	if (data)
+		memcpy(nla_data(a), data, len);
+	memset((unsigned char *) a + a->nla_len, 0, nla_padlen(len));
+
+	return 0;
+}
+
+static inline int add_nested_action_start(struct sw_flow_actions **sfa, int attrtype)
+{
+	int used = (*sfa)->actions_len;
+	int err;
+
+	err = add_action(sfa, attrtype, NULL, 0);
+	if (err)
+		return err;
+
+	return used;
+}
 
-static int validate_sample(const struct nlattr *attr,
-				const struct sw_flow_key *key, int depth)
+static inline void add_nested_action_end(struct sw_flow_actions *sfa, int st_offset)
+{
+	struct nlattr *a = (struct nlattr *) ((unsigned char *)sfa->actions + st_offset);
+
+	a->nla_len = sfa->actions_len - st_offset;
+}
+
+static int validate_and_copy_actions(const struct nlattr *attr,
+				     const struct sw_flow_key *key, int depth,
+				     struct sw_flow_actions **sfa);
+
+static int validate_and_copy_sample(const struct nlattr *attr,
+				    const struct sw_flow_key *key, int depth,
+				    struct sw_flow_actions **sfa)
 {
 	const struct nlattr *attrs[OVS_SAMPLE_ATTR_MAX + 1];
 	const struct nlattr *probability, *actions;
 	const struct nlattr *a;
-	int rem;
+	int rem, start, err, st_acts;
 
 	memset(attrs, 0, sizeof(attrs));
 	nla_for_each_nested(a, attr, rem) {
@@ -492,7 +565,26 @@ static int validate_sample(const struct nlattr *attr,
 	actions = attrs[OVS_SAMPLE_ATTR_ACTIONS];
 	if (!actions || (nla_len(actions) && nla_len(actions) < NLA_HDRLEN))
 		return -EINVAL;
-	return validate_actions(actions, key, depth + 1);
+
+	/* validation done, copy sample action. */
+	start = add_nested_action_start(sfa, OVS_ACTION_ATTR_SAMPLE);
+	if (start < 0)
+		return start;
+	err = add_action(sfa, OVS_SAMPLE_ATTR_PROBABILITY, nla_data(probability), sizeof(u32));
+	if (err)
+		return err;
+	st_acts = add_nested_action_start(sfa, OVS_SAMPLE_ATTR_ACTIONS);
+	if (st_acts < 0)
+		return st_acts;
+
+	err = validate_and_copy_actions(actions, key, depth + 1, sfa);
+	if (err)
+		return err;
+
+	add_nested_action_end(*sfa, st_acts);
+	add_nested_action_end(*sfa, start);
+
+	return 0;
 }
 
 static int validate_tp_port(const struct sw_flow_key *flow_key)
@@ -606,8 +698,24 @@ static int validate_userspace(const struct nlattr *attr)
 	return 0;
 }
 
-static int validate_actions(const struct nlattr *attr,
-				const struct sw_flow_key *key,  int depth)
+static int copy_action(const struct nlattr *from,
+		       struct sw_flow_actions **sfa)
+{
+	int totlen = NLA_ALIGN(from->nla_len);
+	struct nlattr *to;
+
+	to = reserve_sfa_size(sfa, from->nla_len);
+	if (IS_ERR(to))
+		return PTR_ERR(to);
+
+	memcpy(to, from, totlen);
+	return 0;
+}
+
+static int validate_and_copy_actions(const struct nlattr *attr,
+				     const struct sw_flow_key *key,
+				     int depth,
+				     struct sw_flow_actions **sfa)
 {
 	const struct nlattr *a;
 	int rem, err;
@@ -627,12 +735,14 @@ static int validate_actions(const struct nlattr *attr,
 		};
 		const struct ovs_action_push_vlan *vlan;
 		int type = nla_type(a);
+		bool skip_copy;
 
 		if (type > OVS_ACTION_ATTR_MAX ||
 		    (action_lens[type] != nla_len(a) &&
 		     action_lens[type] != (u32)-1))
 			return -EINVAL;
 
+		skip_copy = false;
 		switch (type) {
 		case OVS_ACTION_ATTR_UNSPEC:
 			return -EINVAL;
@@ -667,14 +777,20 @@ static int validate_actions(const struct nlattr *attr,
 			break;
 
 		case OVS_ACTION_ATTR_SAMPLE:
-			err = validate_sample(a, key, depth);
+			err = validate_and_copy_sample(a, key, depth, sfa);
 			if (err)
 				return err;
+			skip_copy = true;
 			break;
 
 		default:
 			return -EINVAL;
 		}
+		if (!skip_copy) {
+			err = copy_action(a, sfa);
+			if (err)
+				return err;
+		}
 	}
 
 	if (rem > 0)
@@ -742,18 +858,16 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info)
 	err = ovs_flow_metadata_from_nlattrs(flow, a[OVS_PACKET_ATTR_KEY]);
 	if (err)
 		goto err_flow_free;
-
-	err = validate_actions(a[OVS_PACKET_ATTR_ACTIONS], &flow->key, 0);
-	if (err)
-		goto err_flow_free;
-
 	flow->hash = ovs_flow_hash(&flow->key, key_len);
-
-	acts = ovs_flow_actions_alloc(a[OVS_PACKET_ATTR_ACTIONS]);
+	acts = ovs_flow_actions_alloc(nla_len(a[OVS_PACKET_ATTR_ACTIONS]));
 	err = PTR_ERR(acts);
 	if (IS_ERR(acts))
 		goto err_flow_free;
+
+	err = validate_and_copy_actions(a[OVS_PACKET_ATTR_ACTIONS], &flow->key, 0, &acts);
 	rcu_assign_pointer(flow->sf_acts, acts);
+	if (err)
+		goto err_flow_free;
 
 	OVS_CB(packet)->flow = flow;
 	packet->priority = flow->key.phy.priority;
@@ -843,6 +957,66 @@ static struct genl_multicast_group ovs_dp_flow_multicast_group = {
 	.name = OVS_FLOW_MCGROUP
 };
 
+static int actions_to_attr(const struct nlattr *attr, int len, struct sk_buff *skb);
+static int sample_action_to_attr(const struct nlattr *attr, struct sk_buff *skb)
+{
+	const struct nlattr *a;
+	struct nlattr *start;
+	int err = 0, rem;
+
+	start = nla_nest_start(skb, OVS_ACTION_ATTR_SAMPLE);
+	if (!start)
+		return -EMSGSIZE;
+
+	nla_for_each_nested(a, attr, rem) {
+		int type = nla_type(a);
+		struct nlattr *st_sample;
+
+		switch (type) {
+		case OVS_SAMPLE_ATTR_PROBABILITY:
+			if (nla_put(skb, OVS_SAMPLE_ATTR_PROBABILITY, sizeof(u32), nla_data(a)))
+				return -EMSGSIZE;
+			break;
+		case OVS_SAMPLE_ATTR_ACTIONS:
+			st_sample = nla_nest_start(skb, OVS_SAMPLE_ATTR_ACTIONS);
+			if (!st_sample)
+				return -EMSGSIZE;
+			err = actions_to_attr(nla_data(a), nla_len(a), skb);
+			if (err)
+				return err;
+			nla_nest_end(skb, st_sample);
+			break;
+		}
+	}
+
+	nla_nest_end(skb, start);
+	return err;
+}
+
+static int actions_to_attr(const struct nlattr *attr, int len, struct sk_buff *skb)
+{
+	const struct nlattr *a;
+	int rem, err;
+
+	nla_for_each_attr(a, attr, len, rem) {
+		int type = nla_type(a);
+
+		switch (type) {
+		case OVS_ACTION_ATTR_SAMPLE:
+			err = sample_action_to_attr(a, skb);
+			if (err)
+				return err;
+			break;
+		default:
+			if (nla_put(skb, type, nla_len(a), nla_data(a)))
+				return -EMSGSIZE;
+			break;
+		}
+	}
+
+	return 0;
+}
+
 static size_t ovs_flow_cmd_msg_size(const struct sw_flow_actions *acts)
 {
 	return NLMSG_ALIGN(sizeof(struct ovs_header))
@@ -860,6 +1034,7 @@ static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp,
 {
 	const int skb_orig_len = skb->len;
 	const struct sw_flow_actions *sf_acts;
+	struct nlattr *start;
 	struct ovs_flow_stats stats;
 	struct ovs_header *ovs_header;
 	struct nlattr *nla;
@@ -913,10 +1088,19 @@ static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp,
 	 * This can only fail for dump operations because the skb is always
 	 * properly sized for single flows.
 	 */
-	err = nla_put(skb, OVS_FLOW_ATTR_ACTIONS, sf_acts->actions_len,
-		      sf_acts->actions);
-	if (err < 0 && skb_orig_len)
-		goto error;
+	start = nla_nest_start(skb, OVS_FLOW_ATTR_ACTIONS);
+	if (start) {
+		err = actions_to_attr(sf_acts->actions, sf_acts->actions_len, skb);
+		if (!err)
+			nla_nest_end(skb, start);
+		else {
+			if (skb_orig_len)
+				goto error;
+
+			nla_nest_cancel(skb, start);
+		}
+	} else if (skb_orig_len)
+		goto nla_put_failure;
 
 	return genlmsg_end(skb, ovs_header);
 
@@ -961,6 +1145,7 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 	struct sk_buff *reply;
 	struct datapath *dp;
 	struct flow_table *table;
+	struct sw_flow_actions *acts = NULL;
 	int error;
 	int key_len;
 
@@ -974,9 +1159,14 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 
 	/* Validate actions. */
 	if (a[OVS_FLOW_ATTR_ACTIONS]) {
-		error = validate_actions(a[OVS_FLOW_ATTR_ACTIONS], &key,  0);
-		if (error)
+		acts = ovs_flow_actions_alloc(nla_len(a[OVS_FLOW_ATTR_ACTIONS]));
+		error = PTR_ERR(acts);
+		if (IS_ERR(acts))
 			goto error;
+
+		error = validate_and_copy_actions(a[OVS_FLOW_ATTR_ACTIONS], &key,  0, &acts);
+		if (error)
+			goto err_kfree;
 	} else if (info->genlhdr->cmd == OVS_FLOW_CMD_NEW) {
 		error = -EINVAL;
 		goto error;
@@ -991,8 +1181,6 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 	table = ovsl_dereference(dp->table);
 	flow = ovs_flow_tbl_lookup(table, &key, key_len);
 	if (!flow) {
-		struct sw_flow_actions *acts;
-
 		/* Bail out if we're not allowed to create a new flow. */
 		error = -ENOENT;
 		if (info->genlhdr->cmd == OVS_FLOW_CMD_SET)
@@ -1019,11 +1207,6 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 		flow->key = key;
 		clear_stats(flow);
 
-		/* Obtain actions. */
-		acts = ovs_flow_actions_alloc(a[OVS_FLOW_ATTR_ACTIONS]);
-		error = PTR_ERR(acts);
-		if (IS_ERR(acts))
-			goto error_free_flow;
 		rcu_assign_pointer(flow->sf_acts, acts);
 
 		/* Put flow in bucket. */
@@ -1036,7 +1219,6 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 	} else {
 		/* We found a matching flow. */
 		struct sw_flow_actions *old_acts;
-		struct nlattr *acts_attrs;
 
 		/* Bail out if we're not allowed to modify an existing flow.
 		 * We accept NLM_F_CREATE in place of the intended NLM_F_EXCL
@@ -1051,21 +1233,8 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 
 		/* Update actions. */
 		old_acts = ovsl_dereference(flow->sf_acts);
-		acts_attrs = a[OVS_FLOW_ATTR_ACTIONS];
-		if (acts_attrs &&
-		   (old_acts->actions_len != nla_len(acts_attrs) ||
-		   memcmp(old_acts->actions, nla_data(acts_attrs),
-			  old_acts->actions_len))) {
-			struct sw_flow_actions *new_acts;
-
-			new_acts = ovs_flow_actions_alloc(acts_attrs);
-			error = PTR_ERR(new_acts);
-			if (IS_ERR(new_acts))
-				goto err_unlock_ovs;
-
-			rcu_assign_pointer(flow->sf_acts, new_acts);
-			ovs_flow_deferred_free_acts(old_acts);
-		}
+		rcu_assign_pointer(flow->sf_acts, acts);
+		ovs_flow_deferred_free_acts(old_acts);
 
 		reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
 					       info->snd_seq, OVS_FLOW_CMD_NEW);
@@ -1086,10 +1255,10 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 				ovs_dp_flow_multicast_group.id, PTR_ERR(reply));
 	return 0;
 
-error_free_flow:
-	ovs_flow_free(flow);
 err_unlock_ovs:
 	ovs_unlock();
+err_kfree:
+	kfree(acts);
 error:
 	return error;
 }
@@ -1866,8 +2035,8 @@ static int ovs_vport_cmd_del(struct sk_buff *skb, struct genl_info *info)
 		goto exit_unlock;
 	}
 
-	reply = ovs_vport_cmd_build_info(vport, info->snd_portid, info->snd_seq,
-					 OVS_VPORT_CMD_DEL);
+	reply = ovs_vport_cmd_build_info(vport, info->snd_portid,
+					 info->snd_seq, OVS_VPORT_CMD_DEL);
 	err = PTR_ERR(reply);
 	if (IS_ERR(reply))
 		goto exit_unlock;
@@ -1896,8 +2065,8 @@ static int ovs_vport_cmd_get(struct sk_buff *skb, struct genl_info *info)
 	if (IS_ERR(vport))
 		goto exit_unlock;
 
-	reply = ovs_vport_cmd_build_info(vport, info->snd_portid, info->snd_seq,
-					 OVS_VPORT_CMD_NEW);
+	reply = ovs_vport_cmd_build_info(vport, info->snd_portid,
+					 info->snd_seq, OVS_VPORT_CMD_NEW);
 	err = PTR_ERR(reply);
 	if (IS_ERR(reply))
 		goto exit_unlock;
diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index 093c191d4fc..940d4b803ff 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -198,20 +198,18 @@ void ovs_flow_used(struct sw_flow *flow, struct sk_buff *skb)
 	spin_unlock(&flow->lock);
 }
 
-struct sw_flow_actions *ovs_flow_actions_alloc(const struct nlattr *actions)
+struct sw_flow_actions *ovs_flow_actions_alloc(int size)
 {
-	int actions_len = nla_len(actions);
 	struct sw_flow_actions *sfa;
 
-	if (actions_len > MAX_ACTIONS_BUFSIZE)
+	if (size > MAX_ACTIONS_BUFSIZE)
 		return ERR_PTR(-EINVAL);
 
-	sfa = kmalloc(sizeof(*sfa) + actions_len, GFP_KERNEL);
+	sfa = kmalloc(sizeof(*sfa) + size, GFP_KERNEL);
 	if (!sfa)
 		return ERR_PTR(-ENOMEM);
 
-	sfa->actions_len = actions_len;
-	nla_memcpy(sfa->actions, actions, actions_len);
+	sfa->actions_len = 0;
 	return sfa;
 }
 
diff --git a/net/openvswitch/flow.h b/net/openvswitch/flow.h
index 2a83e2141f0..e370f6246ee 100644
--- a/net/openvswitch/flow.h
+++ b/net/openvswitch/flow.h
@@ -130,7 +130,7 @@ struct sw_flow *ovs_flow_alloc(void);
 void ovs_flow_deferred_free(struct sw_flow *);
 void ovs_flow_free(struct sw_flow *flow);
 
-struct sw_flow_actions *ovs_flow_actions_alloc(const struct nlattr *);
+struct sw_flow_actions *ovs_flow_actions_alloc(int actions_len);
 void ovs_flow_deferred_free_acts(struct sw_flow_actions *);
 
 int ovs_flow_extract(struct sk_buff *, u16 in_port, struct sw_flow_key *,
-- 
cgit v1.2.3-70-g09d2


From 7d5437c709ded4f152cb8b305d17972d6707f20c Mon Sep 17 00:00:00 2001
From: Pravin B Shelar <pshelar@nicira.com>
Date: Mon, 17 Jun 2013 17:50:18 -0700
Subject: openvswitch: Add tunneling interface.

Add ovs tunnel interface for set tunnel action for userspace.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 include/uapi/linux/openvswitch.h     |  18 +++++
 net/openvswitch/actions.c            |   4 ++
 net/openvswitch/datapath.c           |  78 +++++++++++++++++++++-
 net/openvswitch/datapath.h           |   3 +
 net/openvswitch/flow.c               | 125 +++++++++++++++++++++++++++++++++++
 net/openvswitch/flow.h               |  19 ++++++
 net/openvswitch/vport-internal_dev.c |   2 +-
 net/openvswitch/vport-netdev.c       |   2 +-
 net/openvswitch/vport.c              |   4 +-
 net/openvswitch/vport.h              |   3 +-
 10 files changed, 251 insertions(+), 7 deletions(-)

(limited to 'net/openvswitch/flow.c')

diff --git a/include/uapi/linux/openvswitch.h b/include/uapi/linux/openvswitch.h
index 424672db7f1..b15a445927d 100644
--- a/include/uapi/linux/openvswitch.h
+++ b/include/uapi/linux/openvswitch.h
@@ -246,11 +246,29 @@ enum ovs_key_attr {
 	OVS_KEY_ATTR_ARP,       /* struct ovs_key_arp */
 	OVS_KEY_ATTR_ND,        /* struct ovs_key_nd */
 	OVS_KEY_ATTR_SKB_MARK,  /* u32 skb mark */
+	OVS_KEY_ATTR_TUNNEL,    /* Nested set of ovs_tunnel attributes */
+
+#ifdef __KERNEL__
+	OVS_KEY_ATTR_IPV4_TUNNEL,  /* struct ovs_key_ipv4_tunnel */
+#endif
 	__OVS_KEY_ATTR_MAX
 };
 
 #define OVS_KEY_ATTR_MAX (__OVS_KEY_ATTR_MAX - 1)
 
+enum ovs_tunnel_key_attr {
+	OVS_TUNNEL_KEY_ATTR_ID,                 /* be64 Tunnel ID */
+	OVS_TUNNEL_KEY_ATTR_IPV4_SRC,           /* be32 src IP address. */
+	OVS_TUNNEL_KEY_ATTR_IPV4_DST,           /* be32 dst IP address. */
+	OVS_TUNNEL_KEY_ATTR_TOS,                /* u8 Tunnel IP ToS. */
+	OVS_TUNNEL_KEY_ATTR_TTL,                /* u8 Tunnel IP TTL. */
+	OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT,      /* No argument, set DF. */
+	OVS_TUNNEL_KEY_ATTR_CSUM,               /* No argument. CSUM packet. */
+	__OVS_TUNNEL_KEY_ATTR_MAX
+};
+
+#define OVS_TUNNEL_KEY_ATTR_MAX (__OVS_TUNNEL_KEY_ATTR_MAX - 1)
+
 /**
  * enum ovs_frag_type - IPv4 and IPv6 fragment type
  * @OVS_FRAG_TYPE_NONE: Packet is not a fragment.
diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c
index 596d6373399..22c5f399f1c 100644
--- a/net/openvswitch/actions.c
+++ b/net/openvswitch/actions.c
@@ -436,6 +436,10 @@ static int execute_set_action(struct sk_buff *skb,
 		skb->mark = nla_get_u32(nested_attr);
 		break;
 
+	case OVS_KEY_ATTR_IPV4_TUNNEL:
+		OVS_CB(skb)->tun_key = nla_data(nested_attr);
+		break;
+
 	case OVS_KEY_ATTR_ETHERNET:
 		err = set_eth_addr(skb, nla_data(nested_attr));
 		break;
diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
index f14816b80b8..bbd310646bc 100644
--- a/net/openvswitch/datapath.c
+++ b/net/openvswitch/datapath.c
@@ -362,6 +362,14 @@ static int queue_gso_packets(struct net *net, int dp_ifindex,
 static size_t key_attr_size(void)
 {
 	return    nla_total_size(4)   /* OVS_KEY_ATTR_PRIORITY */
+		+ nla_total_size(0)   /* OVS_KEY_ATTR_TUNNEL */
+		  + nla_total_size(8)   /* OVS_TUNNEL_KEY_ATTR_ID */
+		  + nla_total_size(4)   /* OVS_TUNNEL_KEY_ATTR_IPV4_SRC */
+		  + nla_total_size(4)   /* OVS_TUNNEL_KEY_ATTR_IPV4_DST */
+		  + nla_total_size(1)   /* OVS_TUNNEL_KEY_ATTR_TOS */
+		  + nla_total_size(1)   /* OVS_TUNNEL_KEY_ATTR_TTL */
+		  + nla_total_size(0)   /* OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT */
+		  + nla_total_size(0)   /* OVS_TUNNEL_KEY_ATTR_CSUM */
 		+ nla_total_size(4)   /* OVS_KEY_ATTR_IN_PORT */
 		+ nla_total_size(4)   /* OVS_KEY_ATTR_SKB_MARK */
 		+ nla_total_size(12)  /* OVS_KEY_ATTR_ETHERNET */
@@ -600,8 +608,30 @@ static int validate_tp_port(const struct sw_flow_key *flow_key)
 	return -EINVAL;
 }
 
+static int validate_and_copy_set_tun(const struct nlattr *attr,
+				     struct sw_flow_actions **sfa)
+{
+	struct ovs_key_ipv4_tunnel tun_key;
+	int err, start;
+
+	err = ovs_ipv4_tun_from_nlattr(nla_data(attr), &tun_key);
+	if (err)
+		return err;
+
+	start = add_nested_action_start(sfa, OVS_ACTION_ATTR_SET);
+	if (start < 0)
+		return start;
+
+	err = add_action(sfa, OVS_KEY_ATTR_IPV4_TUNNEL, &tun_key, sizeof(tun_key));
+	add_nested_action_end(*sfa, start);
+
+	return err;
+}
+
 static int validate_set(const struct nlattr *a,
-			const struct sw_flow_key *flow_key)
+			const struct sw_flow_key *flow_key,
+			struct sw_flow_actions **sfa,
+			bool *set_tun)
 {
 	const struct nlattr *ovs_key = nla_data(a);
 	int key_type = nla_type(ovs_key);
@@ -611,18 +641,27 @@ static int validate_set(const struct nlattr *a,
 		return -EINVAL;
 
 	if (key_type > OVS_KEY_ATTR_MAX ||
-	    nla_len(ovs_key) != ovs_key_lens[key_type])
+	   (ovs_key_lens[key_type] != nla_len(ovs_key) &&
+	    ovs_key_lens[key_type] != -1))
 		return -EINVAL;
 
 	switch (key_type) {
 	const struct ovs_key_ipv4 *ipv4_key;
 	const struct ovs_key_ipv6 *ipv6_key;
+	int err;
 
 	case OVS_KEY_ATTR_PRIORITY:
 	case OVS_KEY_ATTR_SKB_MARK:
 	case OVS_KEY_ATTR_ETHERNET:
 		break;
 
+	case OVS_KEY_ATTR_TUNNEL:
+		*set_tun = true;
+		err = validate_and_copy_set_tun(a, sfa);
+		if (err)
+			return err;
+		break;
+
 	case OVS_KEY_ATTR_IPV4:
 		if (flow_key->eth.type != htons(ETH_P_IP))
 			return -EINVAL;
@@ -771,7 +810,7 @@ static int validate_and_copy_actions(const struct nlattr *attr,
 			break;
 
 		case OVS_ACTION_ATTR_SET:
-			err = validate_set(a, key);
+			err = validate_set(a, key, sfa, &skip_copy);
 			if (err)
 				return err;
 			break;
@@ -993,6 +1032,33 @@ static int sample_action_to_attr(const struct nlattr *attr, struct sk_buff *skb)
 	return err;
 }
 
+static int set_action_to_attr(const struct nlattr *a, struct sk_buff *skb)
+{
+	const struct nlattr *ovs_key = nla_data(a);
+	int key_type = nla_type(ovs_key);
+	struct nlattr *start;
+	int err;
+
+	switch (key_type) {
+	case OVS_KEY_ATTR_IPV4_TUNNEL:
+		start = nla_nest_start(skb, OVS_ACTION_ATTR_SET);
+		if (!start)
+			return -EMSGSIZE;
+
+		err = ovs_ipv4_tun_to_nlattr(skb, nla_data(ovs_key));
+		if (err)
+			return err;
+		nla_nest_end(skb, start);
+		break;
+	default:
+		if (nla_put(skb, OVS_ACTION_ATTR_SET, nla_len(a), ovs_key))
+			return -EMSGSIZE;
+		break;
+	}
+
+	return 0;
+}
+
 static int actions_to_attr(const struct nlattr *attr, int len, struct sk_buff *skb)
 {
 	const struct nlattr *a;
@@ -1002,6 +1068,12 @@ static int actions_to_attr(const struct nlattr *attr, int len, struct sk_buff *s
 		int type = nla_type(a);
 
 		switch (type) {
+		case OVS_ACTION_ATTR_SET:
+			err = set_action_to_attr(a, skb);
+			if (err)
+				return err;
+			break;
+
 		case OVS_ACTION_ATTR_SAMPLE:
 			err = sample_action_to_attr(a, skb);
 			if (err)
diff --git a/net/openvswitch/datapath.h b/net/openvswitch/datapath.h
index 16b84069521..e88ebc2f1c5 100644
--- a/net/openvswitch/datapath.h
+++ b/net/openvswitch/datapath.h
@@ -88,9 +88,12 @@ struct datapath {
 /**
  * struct ovs_skb_cb - OVS data in skb CB
  * @flow: The flow associated with this packet.  May be %NULL if no flow.
+ * @tun_key: Key for the tunnel that encapsulated this packet. NULL if the
+ * packet is not being tunneled.
  */
 struct ovs_skb_cb {
 	struct sw_flow		*flow;
+	struct ovs_key_ipv4_tunnel  *tun_key;
 };
 #define OVS_CB(skb) ((struct ovs_skb_cb *)(skb)->cb)
 
diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index 940d4b803ff..976a8b766a6 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -40,6 +40,7 @@
 #include <linux/icmpv6.h>
 #include <linux/rculist.h>
 #include <net/ip.h>
+#include <net/ip_tunnels.h>
 #include <net/ipv6.h>
 #include <net/ndisc.h>
 
@@ -603,6 +604,8 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key,
 	memset(key, 0, sizeof(*key));
 
 	key->phy.priority = skb->priority;
+	if (OVS_CB(skb)->tun_key)
+		memcpy(&key->tun_key, OVS_CB(skb)->tun_key, sizeof(key->tun_key));
 	key->phy.in_port = in_port;
 	key->phy.skb_mark = skb->mark;
 
@@ -818,6 +821,7 @@ const int ovs_key_lens[OVS_KEY_ATTR_MAX + 1] = {
 	[OVS_KEY_ATTR_ICMPV6] = sizeof(struct ovs_key_icmpv6),
 	[OVS_KEY_ATTR_ARP] = sizeof(struct ovs_key_arp),
 	[OVS_KEY_ATTR_ND] = sizeof(struct ovs_key_nd),
+	[OVS_KEY_ATTR_TUNNEL] = -1,
 };
 
 static int ipv4_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_len,
@@ -955,6 +959,105 @@ static int parse_flow_nlattrs(const struct nlattr *attr,
 	return 0;
 }
 
+int ovs_ipv4_tun_from_nlattr(const struct nlattr *attr,
+			     struct ovs_key_ipv4_tunnel *tun_key)
+{
+	struct nlattr *a;
+	int rem;
+	bool ttl = false;
+
+	memset(tun_key, 0, sizeof(*tun_key));
+
+	nla_for_each_nested(a, attr, rem) {
+		int type = nla_type(a);
+		static const u32 ovs_tunnel_key_lens[OVS_TUNNEL_KEY_ATTR_MAX + 1] = {
+			[OVS_TUNNEL_KEY_ATTR_ID] = sizeof(u64),
+			[OVS_TUNNEL_KEY_ATTR_IPV4_SRC] = sizeof(u32),
+			[OVS_TUNNEL_KEY_ATTR_IPV4_DST] = sizeof(u32),
+			[OVS_TUNNEL_KEY_ATTR_TOS] = 1,
+			[OVS_TUNNEL_KEY_ATTR_TTL] = 1,
+			[OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT] = 0,
+			[OVS_TUNNEL_KEY_ATTR_CSUM] = 0,
+		};
+
+		if (type > OVS_TUNNEL_KEY_ATTR_MAX ||
+			ovs_tunnel_key_lens[type] != nla_len(a))
+			return -EINVAL;
+
+		switch (type) {
+		case OVS_TUNNEL_KEY_ATTR_ID:
+			tun_key->tun_id = nla_get_be64(a);
+			tun_key->tun_flags |= TUNNEL_KEY;
+			break;
+		case OVS_TUNNEL_KEY_ATTR_IPV4_SRC:
+			tun_key->ipv4_src = nla_get_be32(a);
+			break;
+		case OVS_TUNNEL_KEY_ATTR_IPV4_DST:
+			tun_key->ipv4_dst = nla_get_be32(a);
+			break;
+		case OVS_TUNNEL_KEY_ATTR_TOS:
+			tun_key->ipv4_tos = nla_get_u8(a);
+			break;
+		case OVS_TUNNEL_KEY_ATTR_TTL:
+			tun_key->ipv4_ttl = nla_get_u8(a);
+			ttl = true;
+			break;
+		case OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT:
+			tun_key->tun_flags |= TUNNEL_DONT_FRAGMENT;
+			break;
+		case OVS_TUNNEL_KEY_ATTR_CSUM:
+			tun_key->tun_flags |= TUNNEL_CSUM;
+			break;
+		default:
+			return -EINVAL;
+
+		}
+	}
+	if (rem > 0)
+		return -EINVAL;
+
+	if (!tun_key->ipv4_dst)
+		return -EINVAL;
+
+	if (!ttl)
+		return -EINVAL;
+
+	return 0;
+}
+
+int ovs_ipv4_tun_to_nlattr(struct sk_buff *skb,
+			   const struct ovs_key_ipv4_tunnel *tun_key)
+{
+	struct nlattr *nla;
+
+	nla = nla_nest_start(skb, OVS_KEY_ATTR_TUNNEL);
+	if (!nla)
+		return -EMSGSIZE;
+
+	if (tun_key->tun_flags & TUNNEL_KEY &&
+	    nla_put_be64(skb, OVS_TUNNEL_KEY_ATTR_ID, tun_key->tun_id))
+		return -EMSGSIZE;
+	if (tun_key->ipv4_src &&
+	    nla_put_be32(skb, OVS_TUNNEL_KEY_ATTR_IPV4_SRC, tun_key->ipv4_src))
+		return -EMSGSIZE;
+	if (nla_put_be32(skb, OVS_TUNNEL_KEY_ATTR_IPV4_DST, tun_key->ipv4_dst))
+		return -EMSGSIZE;
+	if (tun_key->ipv4_tos &&
+	    nla_put_u8(skb, OVS_TUNNEL_KEY_ATTR_TOS, tun_key->ipv4_tos))
+		return -EMSGSIZE;
+	if (nla_put_u8(skb, OVS_TUNNEL_KEY_ATTR_TTL, tun_key->ipv4_ttl))
+		return -EMSGSIZE;
+	if ((tun_key->tun_flags & TUNNEL_DONT_FRAGMENT) &&
+		nla_put_flag(skb, OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT))
+		return -EMSGSIZE;
+	if ((tun_key->tun_flags & TUNNEL_CSUM) &&
+		nla_put_flag(skb, OVS_TUNNEL_KEY_ATTR_CSUM))
+		return -EMSGSIZE;
+
+	nla_nest_end(skb, nla);
+	return 0;
+}
+
 /**
  * ovs_flow_from_nlattrs - parses Netlink attributes into a flow key.
  * @swkey: receives the extracted flow key.
@@ -997,6 +1100,14 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
 		attrs &= ~(1 << OVS_KEY_ATTR_SKB_MARK);
 	}
 
+	if (attrs & (1 << OVS_KEY_ATTR_TUNNEL)) {
+		err = ovs_ipv4_tun_from_nlattr(a[OVS_KEY_ATTR_TUNNEL], &swkey->tun_key);
+		if (err)
+			return err;
+
+		attrs &= ~(1 << OVS_KEY_ATTR_TUNNEL);
+	}
+
 	/* Data attributes. */
 	if (!(attrs & (1 << OVS_KEY_ATTR_ETHERNET)))
 		return -EINVAL;
@@ -1135,17 +1246,21 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
 int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
 				   const struct nlattr *attr)
 {
+	struct ovs_key_ipv4_tunnel *tun_key = &flow->key.tun_key;
 	const struct nlattr *nla;
 	int rem;
 
 	flow->key.phy.in_port = DP_MAX_PORTS;
 	flow->key.phy.priority = 0;
 	flow->key.phy.skb_mark = 0;
+	memset(tun_key, 0, sizeof(flow->key.tun_key));
 
 	nla_for_each_nested(nla, attr, rem) {
 		int type = nla_type(nla);
 
 		if (type <= OVS_KEY_ATTR_MAX && ovs_key_lens[type] > 0) {
+			int err;
+
 			if (nla_len(nla) != ovs_key_lens[type])
 				return -EINVAL;
 
@@ -1154,6 +1269,12 @@ int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
 				flow->key.phy.priority = nla_get_u32(nla);
 				break;
 
+			case OVS_KEY_ATTR_TUNNEL:
+				err = ovs_ipv4_tun_from_nlattr(nla, tun_key);
+				if (err)
+					return err;
+				break;
+
 			case OVS_KEY_ATTR_IN_PORT:
 				if (nla_get_u32(nla) >= DP_MAX_PORTS)
 					return -EINVAL;
@@ -1180,6 +1301,10 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb)
 	    nla_put_u32(skb, OVS_KEY_ATTR_PRIORITY, swkey->phy.priority))
 		goto nla_put_failure;
 
+	if (swkey->tun_key.ipv4_dst &&
+	    ovs_ipv4_tun_to_nlattr(skb, &swkey->tun_key))
+		goto nla_put_failure;
+
 	if (swkey->phy.in_port != DP_MAX_PORTS &&
 	    nla_put_u32(skb, OVS_KEY_ATTR_IN_PORT, swkey->phy.in_port))
 		goto nla_put_failure;
diff --git a/net/openvswitch/flow.h b/net/openvswitch/flow.h
index e370f6246ee..aec5e43f690 100644
--- a/net/openvswitch/flow.h
+++ b/net/openvswitch/flow.h
@@ -40,7 +40,22 @@ struct sw_flow_actions {
 	struct nlattr actions[];
 };
 
+/* Used to memset ovs_key_ipv4_tunnel padding. */
+#define OVS_TUNNEL_KEY_SIZE					\
+	(offsetof(struct ovs_key_ipv4_tunnel, ipv4_ttl) +	\
+	FIELD_SIZEOF(struct ovs_key_ipv4_tunnel, ipv4_ttl))
+
+struct ovs_key_ipv4_tunnel {
+	__be64 tun_id;
+	__be32 ipv4_src;
+	__be32 ipv4_dst;
+	u16  tun_flags;
+	u8   ipv4_tos;
+	u8   ipv4_ttl;
+};
+
 struct sw_flow_key {
+	struct ovs_key_ipv4_tunnel tun_key;  /* Encapsulating tunnel key. */
 	struct {
 		u32	priority;	/* Packet QoS priority. */
 		u32	skb_mark;	/* SKB mark. */
@@ -179,5 +194,9 @@ u32 ovs_flow_hash(const struct sw_flow_key *key, int key_len);
 
 struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *idx);
 extern const int ovs_key_lens[OVS_KEY_ATTR_MAX + 1];
+int ovs_ipv4_tun_from_nlattr(const struct nlattr *attr,
+			 struct ovs_key_ipv4_tunnel *tun_key);
+int ovs_ipv4_tun_to_nlattr(struct sk_buff *skb,
+			const struct ovs_key_ipv4_tunnel *tun_key);
 
 #endif /* flow.h */
diff --git a/net/openvswitch/vport-internal_dev.c b/net/openvswitch/vport-internal_dev.c
index e284c7e1fec..98d3edbbc23 100644
--- a/net/openvswitch/vport-internal_dev.c
+++ b/net/openvswitch/vport-internal_dev.c
@@ -67,7 +67,7 @@ static struct rtnl_link_stats64 *internal_dev_get_stats(struct net_device *netde
 static int internal_dev_xmit(struct sk_buff *skb, struct net_device *netdev)
 {
 	rcu_read_lock();
-	ovs_vport_receive(internal_dev_priv(netdev)->vport, skb);
+	ovs_vport_receive(internal_dev_priv(netdev)->vport, skb, NULL);
 	rcu_read_unlock();
 	return 0;
 }
diff --git a/net/openvswitch/vport-netdev.c b/net/openvswitch/vport-netdev.c
index 40de815b421..5982f3f6283 100644
--- a/net/openvswitch/vport-netdev.c
+++ b/net/openvswitch/vport-netdev.c
@@ -51,7 +51,7 @@ static void netdev_port_receive(struct vport *vport, struct sk_buff *skb)
 	skb_push(skb, ETH_HLEN);
 	ovs_skb_postpush_rcsum(skb, skb->data, ETH_HLEN);
 
-	ovs_vport_receive(vport, skb);
+	ovs_vport_receive(vport, skb, NULL);
 	return;
 
 error:
diff --git a/net/openvswitch/vport.c b/net/openvswitch/vport.c
index 176d449351e..413287a1877 100644
--- a/net/openvswitch/vport.c
+++ b/net/openvswitch/vport.c
@@ -325,7 +325,8 @@ int ovs_vport_get_options(const struct vport *vport, struct sk_buff *skb)
  * Must be called with rcu_read_lock.  The packet cannot be shared and
  * skb->data should point to the Ethernet header.
  */
-void ovs_vport_receive(struct vport *vport, struct sk_buff *skb)
+void ovs_vport_receive(struct vport *vport, struct sk_buff *skb,
+		       struct ovs_key_ipv4_tunnel *tun_key)
 {
 	struct pcpu_tstats *stats;
 
@@ -335,6 +336,7 @@ void ovs_vport_receive(struct vport *vport, struct sk_buff *skb)
 	stats->rx_bytes += skb->len;
 	u64_stats_update_end(&stats->syncp);
 
+	OVS_CB(skb)->tun_key = tun_key;
 	ovs_dp_process_received_packet(vport, skb);
 }
 
diff --git a/net/openvswitch/vport.h b/net/openvswitch/vport.h
index 293278c4c2d..2d961aedd71 100644
--- a/net/openvswitch/vport.h
+++ b/net/openvswitch/vport.h
@@ -184,7 +184,8 @@ static inline struct vport *vport_from_priv(const void *priv)
 	return (struct vport *)(priv - ALIGN(sizeof(struct vport), VPORT_ALIGN));
 }
 
-void ovs_vport_receive(struct vport *, struct sk_buff *);
+void ovs_vport_receive(struct vport *, struct sk_buff *,
+		       struct ovs_key_ipv4_tunnel *);
 void ovs_vport_record_error(struct vport *, enum vport_err_type err_type);
 
 /* List of statically compiled vport implementations.  Don't forget to also
-- 
cgit v1.2.3-70-g09d2


From a3e82996a8874c4cfe8c7f1be4d552018d8cba7e Mon Sep 17 00:00:00 2001
From: Pravin B Shelar <pshelar@nicira.com>
Date: Mon, 17 Jun 2013 17:50:28 -0700
Subject: openvswitch: Optimize flow key match for non tunnel flows.

Following patch adds start offset for sw_flow-key, so that we can
skip tunneling information in key for non-tunnel flows.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 net/openvswitch/datapath.c |  7 ++-----
 net/openvswitch/flow.c     | 49 ++++++++++++++++++++++++++++++++++------------
 net/openvswitch/flow.h     |  6 +++---
 3 files changed, 42 insertions(+), 20 deletions(-)

(limited to 'net/openvswitch/flow.c')

diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
index bbd310646bc..f7e3a0d84c4 100644
--- a/net/openvswitch/datapath.c
+++ b/net/openvswitch/datapath.c
@@ -894,10 +894,9 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info)
 	if (err)
 		goto err_flow_free;
 
-	err = ovs_flow_metadata_from_nlattrs(flow, a[OVS_PACKET_ATTR_KEY]);
+	err = ovs_flow_metadata_from_nlattrs(flow, key_len, a[OVS_PACKET_ATTR_KEY]);
 	if (err)
 		goto err_flow_free;
-	flow->hash = ovs_flow_hash(&flow->key, key_len);
 	acts = ovs_flow_actions_alloc(nla_len(a[OVS_PACKET_ATTR_ACTIONS]));
 	err = PTR_ERR(acts);
 	if (IS_ERR(acts))
@@ -1276,14 +1275,12 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
 			error = PTR_ERR(flow);
 			goto err_unlock_ovs;
 		}
-		flow->key = key;
 		clear_stats(flow);
 
 		rcu_assign_pointer(flow->sf_acts, acts);
 
 		/* Put flow in bucket. */
-		flow->hash = ovs_flow_hash(&key, key_len);
-		ovs_flow_tbl_insert(table, flow);
+		ovs_flow_tbl_insert(table, flow, &key, key_len);
 
 		reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
 						info->snd_seq,
diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index 976a8b766a6..5c519b121e1 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -353,6 +353,14 @@ struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *la
 	return NULL;
 }
 
+static void __flow_tbl_insert(struct flow_table *table, struct sw_flow *flow)
+{
+	struct hlist_head *head;
+	head = find_bucket(table, flow->hash);
+	hlist_add_head_rcu(&flow->hash_node[table->node_ver], head);
+	table->count++;
+}
+
 static void flow_table_copy_flows(struct flow_table *old, struct flow_table *new)
 {
 	int old_ver;
@@ -369,7 +377,7 @@ static void flow_table_copy_flows(struct flow_table *old, struct flow_table *new
 		head = flex_array_get(old->buckets, i);
 
 		hlist_for_each_entry(flow, head, hash_node[old_ver])
-			ovs_flow_tbl_insert(new, flow);
+			__flow_tbl_insert(new, flow);
 	}
 	old->keep_flows = true;
 }
@@ -763,9 +771,18 @@ out:
 	return error;
 }
 
-u32 ovs_flow_hash(const struct sw_flow_key *key, int key_len)
+static u32 ovs_flow_hash(const struct sw_flow_key *key, int key_start, int key_len)
+{
+	return jhash2((u32 *)((u8 *)key + key_start),
+		      DIV_ROUND_UP(key_len - key_start, sizeof(u32)), 0);
+}
+
+static int flow_key_start(struct sw_flow_key *key)
 {
-	return jhash2((u32 *)key, DIV_ROUND_UP(key_len, sizeof(u32)), 0);
+	if (key->tun_key.ipv4_dst)
+		return 0;
+	else
+		return offsetof(struct sw_flow_key, phy);
 }
 
 struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *table,
@@ -773,28 +790,31 @@ struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *table,
 {
 	struct sw_flow *flow;
 	struct hlist_head *head;
+	u8 *_key;
+	int key_start;
 	u32 hash;
 
-	hash = ovs_flow_hash(key, key_len);
+	key_start = flow_key_start(key);
+	hash = ovs_flow_hash(key, key_start, key_len);
 
+	_key = (u8 *) key + key_start;
 	head = find_bucket(table, hash);
 	hlist_for_each_entry_rcu(flow, head, hash_node[table->node_ver]) {
 
 		if (flow->hash == hash &&
-		    !memcmp(&flow->key, key, key_len)) {
+		    !memcmp((u8 *)&flow->key + key_start, _key, key_len - key_start)) {
 			return flow;
 		}
 	}
 	return NULL;
 }
 
-void ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow)
+void ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow,
+			 struct sw_flow_key *key, int key_len)
 {
-	struct hlist_head *head;
-
-	head = find_bucket(table, flow->hash);
-	hlist_add_head_rcu(&flow->hash_node[table->node_ver], head);
-	table->count++;
+	flow->hash = ovs_flow_hash(key, flow_key_start(key), key_len);
+	memcpy(&flow->key, key, sizeof(flow->key));
+	__flow_tbl_insert(table, flow);
 }
 
 void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow)
@@ -1235,6 +1255,7 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
 /**
  * ovs_flow_metadata_from_nlattrs - parses Netlink attributes into a flow key.
  * @flow: Receives extracted in_port, priority, tun_key and skb_mark.
+ * @key_len: Length of key in @flow.  Used for calculating flow hash.
  * @attr: Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink attribute
  * sequence.
  *
@@ -1243,7 +1264,7 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
  * get the metadata, that is, the parts of the flow key that cannot be
  * extracted from the packet itself.
  */
-int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
+int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow, int key_len,
 				   const struct nlattr *attr)
 {
 	struct ovs_key_ipv4_tunnel *tun_key = &flow->key.tun_key;
@@ -1289,6 +1310,10 @@ int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
 	}
 	if (rem)
 		return -EINVAL;
+
+	flow->hash = ovs_flow_hash(&flow->key,
+				   flow_key_start(&flow->key), key_len);
+
 	return 0;
 }
 
diff --git a/net/openvswitch/flow.h b/net/openvswitch/flow.h
index bfe80b96075..999842f247a 100644
--- a/net/openvswitch/flow.h
+++ b/net/openvswitch/flow.h
@@ -156,7 +156,7 @@ u64 ovs_flow_used_time(unsigned long flow_jiffies);
 int ovs_flow_to_nlattrs(const struct sw_flow_key *, struct sk_buff *);
 int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
 		      const struct nlattr *);
-int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
+int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow, int key_len,
 				  const struct nlattr *attr);
 
 #define MAX_ACTIONS_BUFSIZE    (32 * 1024)
@@ -188,9 +188,9 @@ void ovs_flow_tbl_deferred_destroy(struct flow_table *table);
 struct flow_table *ovs_flow_tbl_alloc(int new_size);
 struct flow_table *ovs_flow_tbl_expand(struct flow_table *table);
 struct flow_table *ovs_flow_tbl_rehash(struct flow_table *table);
-void ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow);
+void ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow,
+			 struct sw_flow_key *key, int key_len);
 void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow);
-u32 ovs_flow_hash(const struct sw_flow_key *key, int key_len);
 
 struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *idx);
 extern const int ovs_key_lens[OVS_KEY_ATTR_MAX + 1];
-- 
cgit v1.2.3-70-g09d2