Merge branch 'flow-dissector-features'

Tom Herbert says:

====================
flow_dissector: Paramterize dissection and other features

This patch set adds some new capabilities to flow_dissector:

- Add flags to flow dissector functions to control dissection
  - Flag to stop dissection when L3 header is seen (don't
    dissect L4)
  - Flag to stop dissection when encapsulation is detected
  - Flag to parse first fragment of fragmented packet. This
    may provide L4 ports
- Added new reporting in key_control
  - Packet is a fragment
  - Packet is a first fragment
  - Packet has encapsulation

Also:
  - Make __skb_set_sw_hash a general function
  - Create functions to get a flow hash based on flowi4 or flowi6
    structures without an reference to an skbuff
  - Ignore flow dissector return value from ___skb_get_hash. Just
    use whatever key fields are found to make a hash

Tested:

Ran 200 netperf TCP_RR instances for IPv6 and IPv4. Did not see any
regression. Ran UDP_RR with 10000 byte request and response size
for IPv4 and IPv6, no regression observed however I did see better
performance with IPv6 flow labels due to use of flow labels for L4
hash.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/bonding/bond_main.c

@@ -3095,7 +3095,7 @@ static bool bond_flow_dissect(struct bonding *bond, struct sk_buff *skb,
 	int noff, proto = -1;
 
 	if (bond->params.xmit_policy > BOND_XMIT_POLICY_LAYER23)
-		return skb_flow_dissect_flow_keys(skb, fk);
+		return skb_flow_dissect_flow_keys(skb, fk, 0);
 
 	fk->ports.ports = 0;
 	noff = skb_network_offset(skb);

drivers/net/ethernet/cisco/enic/enic_clsf.c

@@ -177,7 +177,7 @@ int enic_rx_flow_steer(struct net_device *dev, const struct sk_buff *skb,
 	int res, i;
 
 	enic = netdev_priv(dev);
-	res = skb_flow_dissect_flow_keys(skb, &keys);
+	res = skb_flow_dissect_flow_keys(skb, &keys, 0);
 	if (!res || keys.basic.n_proto != htons(ETH_P_IP) ||
 	    (keys.basic.ip_proto != IPPROTO_TCP &&
 	     keys.basic.ip_proto != IPPROTO_UDP))

drivers/net/hyperv/netvsc_drv.c

@@ -239,7 +239,7 @@ static bool netvsc_set_hash(u32 *hash, struct sk_buff *skb)
 	struct flow_keys flow;
 	int data_len;
 
-	if (!skb_flow_dissect_flow_keys(skb, &flow) ||
+	if (!skb_flow_dissect_flow_keys(skb, &flow, 0) ||
 	    !(flow.basic.n_proto == htons(ETH_P_IP) ||
 	      flow.basic.n_proto == htons(ETH_P_IPV6)))
 		return false;

include/linux/skbuff.h

@@ -937,14 +937,90 @@ enum pkt_hash_types {
 	PKT_HASH_TYPE_L4,	/* Input: src_IP, dst_IP, src_port, dst_port */
 };
 
-static inline void
-skb_set_hash(struct sk_buff *skb, __u32 hash, enum pkt_hash_types type)
+static inline void skb_clear_hash(struct sk_buff *skb)
 {
-	skb->l4_hash = (type == PKT_HASH_TYPE_L4);
+	skb->hash = 0;
 	skb->sw_hash = 0;
+	skb->l4_hash = 0;
+}
+
+static inline void skb_clear_hash_if_not_l4(struct sk_buff *skb)
+{
+	if (!skb->l4_hash)
+		skb_clear_hash(skb);
+}
+
+static inline void
+__skb_set_hash(struct sk_buff *skb, __u32 hash, bool is_sw, bool is_l4)
+{
+	skb->l4_hash = is_l4;
+	skb->sw_hash = is_sw;
 	skb->hash = hash;
 }
 
+static inline void
+skb_set_hash(struct sk_buff *skb, __u32 hash, enum pkt_hash_types type)
+{
+	/* Used by drivers to set hash from HW */
+	__skb_set_hash(skb, hash, false, type == PKT_HASH_TYPE_L4);
+}
+
+static inline void
+__skb_set_sw_hash(struct sk_buff *skb, __u32 hash, bool is_l4)
+{
+	__skb_set_hash(skb, hash, true, is_l4);
+}
+
+void __skb_get_hash(struct sk_buff *skb);
+u32 skb_get_poff(const struct sk_buff *skb);
+u32 __skb_get_poff(const struct sk_buff *skb, void *data,
+		   const struct flow_keys *keys, int hlen);
+__be32 __skb_flow_get_ports(const struct sk_buff *skb, int thoff, u8 ip_proto,
+			    void *data, int hlen_proto);
+
+static inline __be32 skb_flow_get_ports(const struct sk_buff *skb,
+					int thoff, u8 ip_proto)
+{
+	return __skb_flow_get_ports(skb, thoff, ip_proto, NULL, 0);
+}
+
+void skb_flow_dissector_init(struct flow_dissector *flow_dissector,
+			     const struct flow_dissector_key *key,
+			     unsigned int key_count);
+
+bool __skb_flow_dissect(const struct sk_buff *skb,
+			struct flow_dissector *flow_dissector,
+			void *target_container,
+			void *data, __be16 proto, int nhoff, int hlen,
+			unsigned int flags);
+
+static inline bool skb_flow_dissect(const struct sk_buff *skb,
+				    struct flow_dissector *flow_dissector,
+				    void *target_container, unsigned int flags)
+{
+	return __skb_flow_dissect(skb, flow_dissector, target_container,
+				  NULL, 0, 0, 0, flags);
+}
+
+static inline bool skb_flow_dissect_flow_keys(const struct sk_buff *skb,
+					      struct flow_keys *flow,
+					      unsigned int flags)
+{
+	memset(flow, 0, sizeof(*flow));
+	return __skb_flow_dissect(skb, &flow_keys_dissector, flow,
+				  NULL, 0, 0, 0, flags);
+}
+
+static inline bool skb_flow_dissect_flow_keys_buf(struct flow_keys *flow,
+						  void *data, __be16 proto,
+						  int nhoff, int hlen,
+						  unsigned int flags)
+{
+	memset(flow, 0, sizeof(*flow));
+	return __skb_flow_dissect(NULL, &flow_keys_buf_dissector, flow,
+				  data, proto, nhoff, hlen, flags);
+}
+
 static inline __u32 skb_get_hash(struct sk_buff *skb)
 {
 	if (!skb->l4_hash && !skb->sw_hash)
@@ -957,8 +1033,12 @@ __u32 __skb_get_hash_flowi6(struct sk_buff *skb, struct flowi6 *fl6);
 
 static inline __u32 skb_get_hash_flowi6(struct sk_buff *skb, struct flowi6 *fl6)
 {
-	if (!skb->l4_hash && !skb->sw_hash)
-		__skb_get_hash_flowi6(skb, fl6);
+	if (!skb->l4_hash && !skb->sw_hash) {
+		struct flow_keys keys;
+
+		__skb_set_sw_hash(skb, __get_hash_from_flowi6(fl6, &keys),
+				  flow_keys_have_l4(&keys));
+	}
 
 	return skb->hash;
 }
@@ -967,8 +1047,12 @@ __u32 __skb_get_hash_flowi4(struct sk_buff *skb, struct flowi4 *fl);
 
 static inline __u32 skb_get_hash_flowi4(struct sk_buff *skb, struct flowi4 *fl4)
 {
-	if (!skb->l4_hash && !skb->sw_hash)
-		__skb_get_hash_flowi4(skb, fl4);
+	if (!skb->l4_hash && !skb->sw_hash) {
+		struct flow_keys keys;
+
+		__skb_set_sw_hash(skb, __get_hash_from_flowi4(fl4, &keys),
+				  flow_keys_have_l4(&keys));
+	}
 
 	return skb->hash;
 }
@@ -980,19 +1064,6 @@ static inline __u32 skb_get_hash_raw(const struct sk_buff *skb)
 	return skb->hash;
 }
 
-static inline void skb_clear_hash(struct sk_buff *skb)
-{
-	skb->hash = 0;
-	skb->sw_hash = 0;
-	skb->l4_hash = 0;
-}
-
-static inline void skb_clear_hash_if_not_l4(struct sk_buff *skb)
-{
-	if (!skb->l4_hash)
-		skb_clear_hash(skb);
-}
-
 static inline void skb_copy_hash(struct sk_buff *to, const struct sk_buff *from)
 {
 	to->hash = from->hash;
@@ -1978,7 +2049,7 @@ static inline void skb_probe_transport_header(struct sk_buff *skb,
 
 	if (skb_transport_header_was_set(skb))
 		return;
-	else if (skb_flow_dissect_flow_keys(skb, &keys))
+	else if (skb_flow_dissect_flow_keys(skb, &keys, 0))
 		skb_set_transport_header(skb, keys.control.thoff);
 	else
 		skb_set_transport_header(skb, offset_hint);

include/net/flow.h

@@ -10,6 +10,7 @@
 #include <linux/socket.h>
 #include <linux/in6.h>
 #include <linux/atomic.h>
+#include <net/flow_dissector.h>
 
 /*
  * ifindex generation is per-net namespace, and loopback is
@@ -243,4 +244,22 @@ void flow_cache_flush(struct net *net);
 void flow_cache_flush_deferred(struct net *net);
 extern atomic_t flow_cache_genid;
 
+__u32 __get_hash_from_flowi6(struct flowi6 *fl6, struct flow_keys *keys);
+
+static inline __u32 get_hash_from_flowi6(struct flowi6 *fl6)
+{
+	struct flow_keys keys;
+
+	return __get_hash_from_flowi6(fl6, &keys);
+}
+
+__u32 __get_hash_from_flowi4(struct flowi4 *fl4, struct flow_keys *keys);
+
+static inline __u32 get_hash_from_flowi4(struct flowi4 *fl4)
+{
+	struct flow_keys keys;
+
+	return __get_hash_from_flowi4(fl4, &keys);
+}
+
 #endif

include/net/flow_dissector.h

@@ -2,7 +2,6 @@
 #define _NET_FLOW_DISSECTOR_H
 
 #include <linux/types.h>
-#include <linux/skbuff.h>
 #include <linux/in6.h>
 #include <uapi/linux/if_ether.h>
 
@@ -13,6 +12,9 @@
 struct flow_dissector_key_control {
 	u16	thoff;
 	u16	addr_type;
+	u32	is_fragment:1;
+	u32	first_frag:1;
+	u32	encapsulation:1;
 };
 
 /**
@@ -123,6 +125,11 @@ enum flow_dissector_key_id {
 	FLOW_DISSECTOR_KEY_MAX,
 };
 
+#define FLOW_DISSECTOR_F_PARSE_1ST_FRAG		BIT(0)
+#define FLOW_DISSECTOR_F_STOP_AT_L3		BIT(1)
+#define FLOW_DISSECTOR_F_STOP_AT_FLOW_LABEL	BIT(2)
+#define FLOW_DISSECTOR_F_STOP_AT_ENCAP		BIT(3)
+
 struct flow_dissector_key {
 	enum flow_dissector_key_id key_id;
 	size_t offset; /* offset of struct flow_dissector_key_*
@@ -134,23 +141,6 @@ struct flow_dissector {
 	unsigned short int offset[FLOW_DISSECTOR_KEY_MAX];
 };
 
-void skb_flow_dissector_init(struct flow_dissector *flow_dissector,
-			     const struct flow_dissector_key *key,
-			     unsigned int key_count);
-
-bool __skb_flow_dissect(const struct sk_buff *skb,
-			struct flow_dissector *flow_dissector,
-			void *target_container,
-			void *data, __be16 proto, int nhoff, int hlen);
-
-static inline bool skb_flow_dissect(const struct sk_buff *skb,
-				    struct flow_dissector *flow_dissector,
-				    void *target_container)
-{
-	return __skb_flow_dissect(skb, flow_dissector, target_container,
-				  NULL, 0, 0, 0);
-}
-
 struct flow_keys {
 	struct flow_dissector_key_control control;
 #define FLOW_KEYS_HASH_START_FIELD basic
@@ -170,38 +160,6 @@ __be32 flow_get_u32_dst(const struct flow_keys *flow);
 extern struct flow_dissector flow_keys_dissector;
 extern struct flow_dissector flow_keys_buf_dissector;
 
-static inline bool skb_flow_dissect_flow_keys(const struct sk_buff *skb,
-					      struct flow_keys *flow)
-{
-	memset(flow, 0, sizeof(*flow));
-	return __skb_flow_dissect(skb, &flow_keys_dissector, flow,
-				  NULL, 0, 0, 0);
-}
-
-static inline bool skb_flow_dissect_flow_keys_buf(struct flow_keys *flow,
-						  void *data, __be16 proto,
-						  int nhoff, int hlen)
-{
-	memset(flow, 0, sizeof(*flow));
-	return __skb_flow_dissect(NULL, &flow_keys_buf_dissector, flow,
-				  data, proto, nhoff, hlen);
-}
-
-__be32 __skb_flow_get_ports(const struct sk_buff *skb, int thoff, u8 ip_proto,
-			    void *data, int hlen_proto);
-
-static inline __be32 skb_flow_get_ports(const struct sk_buff *skb,
-					int thoff, u8 ip_proto)
-{
-	return __skb_flow_get_ports(skb, thoff, ip_proto, NULL, 0);
-}
-
-u32 flow_hash_from_keys(struct flow_keys *keys);
-void __skb_get_hash(struct sk_buff *skb);
-u32 skb_get_poff(const struct sk_buff *skb);
-u32 __skb_get_poff(const struct sk_buff *skb, void *data,
-		   const struct flow_keys *keys, int hlen);
-
 /* struct flow_keys_digest:
  *
  * This structure is used to hold a digest of the full flow keys. This is a
@@ -217,4 +175,11 @@ struct flow_keys_digest {
 void make_flow_keys_digest(struct flow_keys_digest *digest,
 			   const struct flow_keys *flow);
 
+static inline bool flow_keys_have_l4(struct flow_keys *keys)
+{
+	return (keys->ports.ports || keys->tags.flow_label);
+}
+
+u32 flow_hash_from_keys(struct flow_keys *keys);
+
 #endif

net/core/flow.c

@@ -22,6 +22,7 @@
 #include <linux/cpumask.h>
 #include <linux/mutex.h>
 #include <net/flow.h>
+#include <net/flow_dissector.h>
 #include <linux/atomic.h>
 #include <linux/security.h>
 #include <net/net_namespace.h>
@@ -509,3 +510,38 @@ void flow_cache_fini(struct net *net)
 	fc->percpu = NULL;
 }
 EXPORT_SYMBOL(flow_cache_fini);
+
+__u32 __get_hash_from_flowi6(struct flowi6 *fl6, struct flow_keys *keys)
+{
+	memset(keys, 0, sizeof(*keys));
+
+	memcpy(&keys->addrs.v6addrs.src, &fl6->saddr,
+	    sizeof(keys->addrs.v6addrs.src));
+	memcpy(&keys->addrs.v6addrs.dst, &fl6->daddr,
+	    sizeof(keys->addrs.v6addrs.dst));
+	keys->control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
+	keys->ports.src = fl6->fl6_sport;
+	keys->ports.dst = fl6->fl6_dport;
+	keys->keyid.keyid = fl6->fl6_gre_key;
+	keys->tags.flow_label = (__force u32)fl6->flowlabel;
+	keys->basic.ip_proto = fl6->flowi6_proto;
+
+	return flow_hash_from_keys(keys);
+}
+EXPORT_SYMBOL(__get_hash_from_flowi6);
+
+__u32 __get_hash_from_flowi4(struct flowi4 *fl4, struct flow_keys *keys)
+{
+	memset(keys, 0, sizeof(*keys));
+
+	keys->addrs.v4addrs.src = fl4->saddr;
+	keys->addrs.v4addrs.dst = fl4->daddr;
+	keys->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
+	keys->ports.src = fl4->fl4_sport;
+	keys->ports.dst = fl4->fl4_dport;
+	keys->keyid.keyid = fl4->fl4_gre_key;
+	keys->basic.ip_proto = fl4->flowi4_proto;
+
+	return flow_hash_from_keys(keys);
+}
+EXPORT_SYMBOL(__get_hash_from_flowi4);

net/core/flow_dissector.c

@@ -121,7 +121,8 @@ EXPORT_SYMBOL(__skb_flow_get_ports);
 bool __skb_flow_dissect(const struct sk_buff *skb,
 			struct flow_dissector *flow_dissector,
 			void *target_container,
-			void *data, __be16 proto, int nhoff, int hlen)
+			void *data, __be16 proto, int nhoff, int hlen,
+			unsigned int flags)
 {
 	struct flow_dissector_key_control *key_control;
 	struct flow_dissector_key_basic *key_basic;
@@ -130,6 +131,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 	struct flow_dissector_key_tags *key_tags;
 	struct flow_dissector_key_keyid *key_keyid;
 	u8 ip_proto = 0;
+	bool ret = false;
 
 	if (!data) {
 		data = skb->data;
@@ -171,12 +173,10 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 ip:
 		iph = __skb_header_pointer(skb, nhoff, sizeof(_iph), data, hlen, &_iph);
 		if (!iph || iph->ihl < 5)
-			return false;
+			goto out_bad;
 		nhoff += iph->ihl * 4;
 
 		ip_proto = iph->protocol;
-		if (ip_is_fragment(iph))
-			ip_proto = 0;
 
 		if (!skb_flow_dissector_uses_key(flow_dissector,
 						 FLOW_DISSECTOR_KEY_IPV4_ADDRS))
@@ -187,6 +187,22 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 		memcpy(&key_addrs->v4addrs, &iph->saddr,
 		       sizeof(key_addrs->v4addrs));
 		key_control->addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
+
+		if (ip_is_fragment(iph)) {
+			key_control->is_fragment = 1;
+
+			if (iph->frag_off & htons(IP_OFFSET)) {
+				goto out_good;
+			} else {
+				key_control->first_frag = 1;
+				if (!(flags & FLOW_DISSECTOR_F_PARSE_1ST_FRAG))
+					goto out_good;
+			}
+		}
+
+		if (flags & FLOW_DISSECTOR_F_STOP_AT_L3)
+			goto out_good;
+
 		break;
 	}
 	case htons(ETH_P_IPV6): {
@@ -197,7 +213,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 ipv6:
 		iph = __skb_header_pointer(skb, nhoff, sizeof(_iph), data, hlen, &_iph);
 		if (!iph)
-			return false;
+			goto out_bad;
 
 		ip_proto = iph->nexthdr;
 		nhoff += sizeof(struct ipv6hdr);
@@ -223,8 +239,13 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 								     target_container);
 				key_tags->flow_label = ntohl(flow_label);
 			}
+			if (flags & FLOW_DISSECTOR_F_STOP_AT_FLOW_LABEL)
+				goto out_good;
 		}
 
+		if (flags & FLOW_DISSECTOR_F_STOP_AT_L3)
+			goto out_good;
+
 		break;
 	}
 	case htons(ETH_P_8021AD):
@@ -234,7 +255,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 
 		vlan = __skb_header_pointer(skb, nhoff, sizeof(_vlan), data, hlen, &_vlan);
 		if (!vlan)
-			return false;
+			goto out_bad;
 
 		if (skb_flow_dissector_uses_key(flow_dissector,
 						FLOW_DISSECTOR_KEY_VLANID)) {
@@ -256,7 +277,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 		} *hdr, _hdr;
 		hdr = __skb_header_pointer(skb, nhoff, sizeof(_hdr), data, hlen, &_hdr);
 		if (!hdr)
-			return false;
+			goto out_bad;
 		proto = hdr->proto;
 		nhoff += PPPOE_SES_HLEN;
 		switch (proto) {
@@ -265,7 +286,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 		case htons(PPP_IPV6):
 			goto ipv6;
 		default:
-			return false;
+			goto out_bad;
 		}
 	}
 	case htons(ETH_P_TIPC): {
@@ -275,9 +296,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 		} *hdr, _hdr;
 		hdr = __skb_header_pointer(skb, nhoff, sizeof(_hdr), data, hlen, &_hdr);
 		if (!hdr)
-			return false;
-		key_basic->n_proto = proto;
-		key_control->thoff = (u16)nhoff;
+			goto out_bad;
 
 		if (skb_flow_dissector_uses_key(flow_dissector,
 						FLOW_DISSECTOR_KEY_TIPC_ADDRS)) {
@@ -287,7 +306,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 			key_addrs->tipcaddrs.srcnode = hdr->srcnode;
 			key_control->addr_type = FLOW_DISSECTOR_KEY_TIPC_ADDRS;
 		}
-		return true;
+		goto out_good;
 	}
 
 	case htons(ETH_P_MPLS_UC):
@@ -297,7 +316,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 		hdr = __skb_header_pointer(skb, nhoff, sizeof(_hdr), data,
 					   hlen, &_hdr);
 		if (!hdr)
-			return false;
+			goto out_bad;
 
 		if ((ntohl(hdr[0].entry) & MPLS_LS_LABEL_MASK) >>
 		     MPLS_LS_LABEL_SHIFT == MPLS_LABEL_ENTROPY) {
@@ -310,21 +329,17 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 					htonl(MPLS_LS_LABEL_MASK);
 			}
 
-			key_basic->n_proto = proto;
-			key_basic->ip_proto = ip_proto;
-			key_control->thoff = (u16)nhoff;
-
-			return true;
+			goto out_good;
 		}
 
-		return true;
+		goto out_good;
 	}
 
 	case htons(ETH_P_FCOE):
 		key_control->thoff = (u16)(nhoff + FCOE_HEADER_LEN);
 		/* fall through */
 	default:
-		return false;
+		goto out_bad;
 	}
 
 ip_proto_again:
@@ -337,7 +352,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 
 		hdr = __skb_header_pointer(skb, nhoff, sizeof(_hdr), data, hlen, &_hdr);
 		if (!hdr)
-			return false;
+			goto out_bad;
 		/*
 		 * Only look inside GRE if version zero and no
 		 * routing
@@ -357,7 +372,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 						     data, hlen, &_keyid);
 
 			if (!keyid)
-				return false;
+				goto out_bad;
 
 			if (skb_flow_dissector_uses_key(flow_dissector,
 							FLOW_DISSECTOR_KEY_GRE_KEYID)) {
@@ -378,10 +393,15 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 						   sizeof(_eth),
 						   data, hlen, &_eth);
 			if (!eth)
-				return false;
+				goto out_bad;
 			proto = eth->h_proto;
 			nhoff += sizeof(*eth);
 		}
+
+		key_control->encapsulation = 1;
+		if (flags & FLOW_DISSECTOR_F_STOP_AT_ENCAP)
+			goto out_good;
+
 		goto again;
 	}
 	case NEXTHDR_HOP:
@@ -395,18 +415,53 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 		opthdr = __skb_header_pointer(skb, nhoff, sizeof(_opthdr),
 					      data, hlen, &_opthdr);
 		if (!opthdr)
-			return false;
+			goto out_bad;
 
 		ip_proto = opthdr[0];
 		nhoff += (opthdr[1] + 1) << 3;
 
 		goto ip_proto_again;
 	}
+	case NEXTHDR_FRAGMENT: {
+		struct frag_hdr _fh, *fh;
+
+		if (proto != htons(ETH_P_IPV6))
+			break;
+
+		fh = __skb_header_pointer(skb, nhoff, sizeof(_fh),
+					  data, hlen, &_fh);
+
+		if (!fh)
+			goto out_bad;
+
+		key_control->is_fragment = 1;
+
+		nhoff += sizeof(_fh);
+
+		if (!(fh->frag_off & htons(IP6_OFFSET))) {
+			key_control->first_frag = 1;
+			if (flags & FLOW_DISSECTOR_F_PARSE_1ST_FRAG) {
+				ip_proto = fh->nexthdr;
+				goto ip_proto_again;
+			}
+		}
+		goto out_good;
+	}
 	case IPPROTO_IPIP:
 		proto = htons(ETH_P_IP);
+
+		key_control->encapsulation = 1;
+		if (flags & FLOW_DISSECTOR_F_STOP_AT_ENCAP)
+			goto out_good;
+
 		goto ip;
 	case IPPROTO_IPV6:
 		proto = htons(ETH_P_IPV6);
+
+		key_control->encapsulation = 1;
+		if (flags & FLOW_DISSECTOR_F_STOP_AT_ENCAP)
+			goto out_good;
+
 		goto ipv6;
 	case IPPROTO_MPLS:
 		proto = htons(ETH_P_MPLS_UC);
@@ -415,10 +470,6 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 		break;
 	}
 
-	key_basic->n_proto = proto;
-	key_basic->ip_proto = ip_proto;
-	key_control->thoff = (u16)nhoff;
-
 	if (skb_flow_dissector_uses_key(flow_dissector,
 					FLOW_DISSECTOR_KEY_PORTS)) {
 		key_ports = skb_flow_dissector_target(flow_dissector,
@@ -428,7 +479,15 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 							data, hlen);
 	}
 
-	return true;
+out_good:
+	ret = true;
+
+out_bad:
+	key_basic->n_proto = proto;
+	key_basic->ip_proto = ip_proto;
+	key_control->thoff = (u16)nhoff;
+
+	return ret;
 }
 EXPORT_SYMBOL(__skb_flow_dissect);
 
@@ -557,8 +616,8 @@ EXPORT_SYMBOL(flow_hash_from_keys);
 static inline u32 ___skb_get_hash(const struct sk_buff *skb,
 				  struct flow_keys *keys, u32 keyval)
 {
-	if (!skb_flow_dissect_flow_keys(skb, keys))
-		return 0;
+	skb_flow_dissect_flow_keys(skb, keys,
+				   FLOW_DISSECTOR_F_STOP_AT_FLOW_LABEL);
 
 	return __flow_hash_from_keys(keys, keyval);
 }
@@ -590,15 +649,6 @@ void make_flow_keys_digest(struct flow_keys_digest *digest,
 }
 EXPORT_SYMBOL(make_flow_keys_digest);
 
-static inline void __skb_set_sw_hash(struct sk_buff *skb, u32 hash,
-				     struct flow_keys *keys)
-{
-	if (keys->ports.ports)
-		skb->l4_hash = 1;
-	skb->sw_hash = 1;
-	skb->hash = hash;
-}
-
 /**
  * __skb_get_hash: calculate a flow hash
  * @skb: sk_buff to calculate flow hash from
@@ -611,15 +661,11 @@ static inline void __skb_set_sw_hash(struct sk_buff *skb, u32 hash,
 void __skb_get_hash(struct sk_buff *skb)
 {
 	struct flow_keys keys;
-	u32 hash;
 
 	__flow_hash_secret_init();
 
-	hash = ___skb_get_hash(skb, &keys, hashrnd);
-	if (!hash)
-		return;
-
-	__skb_set_sw_hash(skb, hash, &keys);
+	__skb_set_sw_hash(skb, ___skb_get_hash(skb, &keys, hashrnd),
+			  flow_keys_have_l4(&keys));
 }
 EXPORT_SYMBOL(__skb_get_hash);
 
@@ -648,7 +694,8 @@ __u32 __skb_get_hash_flowi6(struct sk_buff *skb, struct flowi6 *fl6)
 	keys.tags.flow_label = (__force u32)fl6->flowlabel;
 	keys.basic.ip_proto = fl6->flowi6_proto;
 
-	__skb_set_sw_hash(skb, flow_hash_from_keys(&keys), &keys);
+	__skb_set_sw_hash(skb, flow_hash_from_keys(&keys),
+			  flow_keys_have_l4(&keys));
 
 	return skb->hash;
 }
@@ -668,7 +715,8 @@ __u32 __skb_get_hash_flowi4(struct sk_buff *skb, struct flowi4 *fl4)
 	keys.keyid.keyid = fl4->fl4_gre_key;
 	keys.basic.ip_proto = fl4->flowi4_proto;
 
-	__skb_set_sw_hash(skb, flow_hash_from_keys(&keys), &keys);
+	__skb_set_sw_hash(skb, flow_hash_from_keys(&keys),
+			  flow_keys_have_l4(&keys));
 
 	return skb->hash;
 }
@@ -733,7 +781,7 @@ u32 skb_get_poff(const struct sk_buff *skb)
 {
 	struct flow_keys keys;
 
-	if (!skb_flow_dissect_flow_keys(skb, &keys))
+	if (!skb_flow_dissect_flow_keys(skb, &keys, 0))
 		return 0;
 
 	return __skb_get_poff(skb, skb->data, &keys, skb_headlen(skb));

net/ethernet/eth.c

@@ -132,7 +132,7 @@ u32 eth_get_headlen(void *data, unsigned int len)
 
 	/* parse any remaining L2/L3 headers, check for L4 */
 	if (!skb_flow_dissect_flow_keys_buf(&keys, data, eth->h_proto,
-					    sizeof(*eth), len))
+					    sizeof(*eth), len, 0))
 		return max_t(u32, keys.control.thoff, sizeof(*eth));
 
 	/* parse for any L4 headers */

net/sched/cls_flow.c

@@ -301,7 +301,7 @@ static int flow_classify(struct sk_buff *skb, const struct tcf_proto *tp,
 
 		keymask = f->keymask;
 		if (keymask & FLOW_KEYS_NEEDED)
-			skb_flow_dissect_flow_keys(skb, &flow_keys);
+			skb_flow_dissect_flow_keys(skb, &flow_keys, 0);
 
 		for (n = 0; n < f->nkeys; n++) {
 			key = ffs(keymask) - 1;

net/sched/cls_flower.c

@@ -129,7 +129,7 @@ static int fl_classify(struct sk_buff *skb, const struct tcf_proto *tp,
 	 * so do it rather here.
 	 */
 	skb_key.basic.n_proto = skb->protocol;
-	skb_flow_dissect(skb, &head->dissector, &skb_key);
+	skb_flow_dissect(skb, &head->dissector, &skb_key, 0);
 
 	fl_set_masked_key(&skb_mkey, &skb_key, &head->mask);
 

net/sched/sch_choke.c

@@ -170,13 +170,13 @@ static bool choke_match_flow(struct sk_buff *skb1,
 
 	if (!choke_skb_cb(skb1)->keys_valid) {
 		choke_skb_cb(skb1)->keys_valid = 1;
-		skb_flow_dissect_flow_keys(skb1, &temp);
+		skb_flow_dissect_flow_keys(skb1, &temp, 0);
 		make_flow_keys_digest(&choke_skb_cb(skb1)->keys, &temp);
 	}
 
 	if (!choke_skb_cb(skb2)->keys_valid) {
 		choke_skb_cb(skb2)->keys_valid = 1;
-		skb_flow_dissect_flow_keys(skb2, &temp);
+		skb_flow_dissect_flow_keys(skb2, &temp, 0);
 		make_flow_keys_digest(&choke_skb_cb(skb2)->keys, &temp);
 	}