[PATCH v2 net-next 3/6] flow_dissector: Add protocol specific flow dissection offload
From: Tom Herbert <hidden>
Date: 2017-08-29 23:27:38
Subsystem:
networking drivers, networking [general], networking [ipv4/ipv6], the rest · Maintainers:
Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, David Ahern, Ido Schimmel, Linus Torvalds
Add offload capability for performing protocol specific flow dissection (either by EtherType or IP protocol). Specifically: - Add flow_dissect to offload callbacks - Move flow_dissect_ret enum to flow_dissector.h, cleanup names and add a couple of values - Create GOTO_BY_RESULT macro to use in the main flow dissector switch to simplify handling of functions that return flow_dissect_ret enum - In __skb_flow_dissect, add default case for switch(proto) as well as switch(ip_proto) that looks up and calls protocol specific flow dissection Signed-off-by: Tom Herbert <redacted> --- include/linux/netdevice.h | 7 +++ include/net/flow_dissector.h | 9 +++ net/core/dev.c | 14 +++++ net/core/flow_dissector.c | 132 +++++++++++++++++++++++++++++++------------ net/ipv4/route.c | 4 +- 5 files changed, 128 insertions(+), 38 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index c5475b37a631..90ccb434e127 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h@@ -2208,6 +2208,12 @@ struct offload_callbacks { struct sk_buff **(*gro_receive)(struct sk_buff **head, struct sk_buff *skb); int (*gro_complete)(struct sk_buff *skb, int nhoff); + enum flow_dissect_ret (*flow_dissect)(const struct sk_buff *skb, + struct flow_dissector_key_control *key_control, + struct flow_dissector *flow_dissector, + void *target_container, void *data, + __be16 *p_proto, u8 *p_ip_proto, int *p_nhoff, + int *p_hlen, unsigned int flags); }; struct packet_offload {
@@ -3253,6 +3259,7 @@ struct sk_buff *napi_get_frags(struct napi_struct *napi); gro_result_t napi_gro_frags(struct napi_struct *napi); struct packet_offload *gro_find_receive_by_type(__be16 type); struct packet_offload *gro_find_complete_by_type(__be16 type); +struct packet_offload *flow_dissect_find_by_type(__be16 type); static inline void napi_free_frags(struct napi_struct *napi) {
diff --git a/include/net/flow_dissector.h b/include/net/flow_dissector.h
index e2663e900b0a..ad75bbfd1c9c 100644
--- a/include/net/flow_dissector.h
+++ b/include/net/flow_dissector.h@@ -19,6 +19,14 @@ struct flow_dissector_key_control { #define FLOW_DIS_FIRST_FRAG BIT(1) #define FLOW_DIS_ENCAPSULATION BIT(2) +enum flow_dissect_ret { + FLOW_DISSECT_RET_OUT_GOOD, + FLOW_DISSECT_RET_OUT_BAD, + FLOW_DISSECT_RET_PROTO_AGAIN, + FLOW_DISSECT_RET_IPPROTO_AGAIN, + FLOW_DISSECT_RET_CONTINUE, +}; + /** * struct flow_dissector_key_basic: * @thoff: Transport header offset
@@ -205,6 +213,7 @@ enum flow_dissector_key_id { #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) +#define FLOW_DISSECTOR_F_STOP_AT_L4 BIT(4) struct flow_dissector_key { enum flow_dissector_key_id key_id;
diff --git a/net/core/dev.c b/net/core/dev.c
index 270b54754821..22ea8daa930c 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c@@ -4860,6 +4860,20 @@ struct packet_offload *gro_find_receive_by_type(__be16 type) } EXPORT_SYMBOL(gro_find_receive_by_type); +struct packet_offload *flow_dissect_find_by_type(__be16 type) +{ + struct list_head *offload_head = &offload_base; + struct packet_offload *ptype; + + list_for_each_entry_rcu(ptype, offload_head, list) { + if (ptype->type != type || !ptype->callbacks.flow_dissect) + continue; + return ptype; + } + return NULL; +} +EXPORT_SYMBOL(flow_dissect_find_by_type); + struct packet_offload *gro_find_complete_by_type(__be16 type) { struct list_head *offload_head = &offload_base;
diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 12302acdb073..6a2cf240069a 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c@@ -9,6 +9,7 @@ #include <net/ipv6.h> #include <net/gre.h> #include <net/pptp.h> +#include <net/protocol.h> #include <linux/igmp.h> #include <linux/icmp.h> #include <linux/sctp.h>
@@ -115,12 +116,6 @@ __be32 __skb_flow_get_ports(const struct sk_buff *skb, int thoff, u8 ip_proto, } EXPORT_SYMBOL(__skb_flow_get_ports); -enum flow_dissect_ret { - FLOW_DISSECT_RET_OUT_GOOD, - FLOW_DISSECT_RET_OUT_BAD, - FLOW_DISSECT_RET_OUT_PROTO_AGAIN, -}; - static enum flow_dissect_ret __skb_flow_dissect_mpls(const struct sk_buff *skb, struct flow_dissector *flow_dissector,
@@ -322,7 +317,7 @@ __skb_flow_dissect_gre(const struct sk_buff *skb, if (flags & FLOW_DISSECTOR_F_STOP_AT_ENCAP) return FLOW_DISSECT_RET_OUT_GOOD; - return FLOW_DISSECT_RET_OUT_PROTO_AGAIN; + return FLOW_DISSECT_RET_PROTO_AGAIN; } static void
@@ -383,6 +378,27 @@ __skb_flow_dissect_ipv6(const struct sk_buff *skb, key_ip->ttl = iph->hop_limit; } +#define GOTO_BY_RESULT(ret) do { \ + switch (ret) { \ + case FLOW_DISSECT_RET_OUT_GOOD: \ + goto out_good; \ + case FLOW_DISSECT_RET_PROTO_AGAIN: \ + goto proto_again; \ + case FLOW_DISSECT_RET_IPPROTO_AGAIN: \ + goto ip_proto_again; \ + case FLOW_DISSECT_RET_OUT_BAD: \ + default: \ + goto out_bad; \ + } \ +} while (0) + +#define GOTO_OR_CONT_BY_RESULT(ret) do { \ + enum flow_dissect_ret __ret = (ret); \ + \ + if (__ret != FLOW_DISSECT_RET_CONTINUE) \ + GOTO_BY_RESULT(__ret); \ +} while (0) + /** * __skb_flow_dissect - extract the flow_keys struct and return it * @skb: sk_buff to extract the flow from, can be NULL if the rest are specified
@@ -659,15 +675,10 @@ bool __skb_flow_dissect(const struct sk_buff *skb, case htons(ETH_P_MPLS_UC): case htons(ETH_P_MPLS_MC): mpls: - switch (__skb_flow_dissect_mpls(skb, flow_dissector, - target_container, data, - nhoff, hlen)) { - case FLOW_DISSECT_RET_OUT_GOOD: - goto out_good; - case FLOW_DISSECT_RET_OUT_BAD: - default: - goto out_bad; - } + GOTO_BY_RESULT(__skb_flow_dissect_mpls(skb, flow_dissector, + target_container, data, + nhoff, hlen)); + case htons(ETH_P_FCOE): if ((hlen - nhoff) < FCOE_HEADER_LEN) goto out_bad;
@@ -677,32 +688,44 @@ bool __skb_flow_dissect(const struct sk_buff *skb, case htons(ETH_P_ARP): case htons(ETH_P_RARP): - switch (__skb_flow_dissect_arp(skb, flow_dissector, - target_container, data, - nhoff, hlen)) { - case FLOW_DISSECT_RET_OUT_GOOD: - goto out_good; - case FLOW_DISSECT_RET_OUT_BAD: - default: - goto out_bad; + GOTO_BY_RESULT(__skb_flow_dissect_arp(skb, flow_dissector, + target_container, data, + nhoff, hlen)); + + default: { + struct packet_offload *ptype; + enum flow_dissect_ret ret; + + rcu_read_lock(); + + ptype = flow_dissect_find_by_type(proto); + + if (ptype) { + ret = ptype->callbacks.flow_dissect(skb, key_control, + flow_dissector, + target_container, + data, &proto, &ip_proto, &nhoff, + &hlen, flags); + rcu_read_unlock(); + + GOTO_BY_RESULT(ret); + } else { + rcu_read_unlock(); } - default: + goto out_bad; } + } ip_proto_again: switch (ip_proto) { case IPPROTO_GRE: - switch (__skb_flow_dissect_gre(skb, key_control, flow_dissector, - target_container, data, - &proto, &nhoff, &hlen, flags)) { - case FLOW_DISSECT_RET_OUT_GOOD: - goto out_good; - case FLOW_DISSECT_RET_OUT_BAD: - goto out_bad; - case FLOW_DISSECT_RET_OUT_PROTO_AGAIN: - goto proto_again; - } + GOTO_BY_RESULT(__skb_flow_dissect_gre(skb, key_control, + flow_dissector, + target_container, data, + &proto, &nhoff, &hlen, + flags)); + case NEXTHDR_HOP: case NEXTHDR_ROUTING: case NEXTHDR_DEST: {
@@ -768,9 +791,43 @@ bool __skb_flow_dissect(const struct sk_buff *skb, __skb_flow_dissect_tcp(skb, flow_dissector, target_container, data, nhoff, hlen); break; - default: + default: { + const struct net_offload *ops = NULL; + + if (flags & FLOW_DISSECTOR_F_STOP_AT_L4) + break; + + rcu_read_lock(); + + switch (proto) { + case htons(ETH_P_IP): + ops = rcu_dereference(inet_offloads[ip_proto]); + break; + case htons(ETH_P_IPV6): + ops = rcu_dereference(inet6_offloads[ip_proto]); + break; + default: + break; + } + + if (ops && ops->callbacks.flow_dissect) { + enum flow_dissect_ret ret; + + ret = ops->callbacks.flow_dissect(skb, key_control, + flow_dissector, + target_container, + data, &proto, &ip_proto, &nhoff, + &hlen, flags); + rcu_read_unlock(); + + GOTO_OR_CONT_BY_RESULT(ret); + } else { + rcu_read_unlock(); + } + break; } + } if (dissector_uses_key(flow_dissector, FLOW_DISSECTOR_KEY_PORTS)) {
@@ -935,7 +992,8 @@ static inline u32 ___skb_get_hash(const struct sk_buff *skb, struct flow_keys *keys, u32 keyval) { skb_flow_dissect_flow_keys(skb, keys, - FLOW_DISSECTOR_F_STOP_AT_FLOW_LABEL); + FLOW_DISSECTOR_F_STOP_AT_FLOW_LABEL | + FLOW_DISSECTOR_F_STOP_AT_L4); return __flow_hash_from_keys(keys, keyval); }
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 94d4cd2d5ea4..85f12b8e0b7f 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c@@ -1811,7 +1811,9 @@ int fib_multipath_hash(const struct fib_info *fi, const struct flowi4 *fl4, case 1: /* skb is currently provided only when forwarding */ if (skb) { - unsigned int flag = FLOW_DISSECTOR_F_STOP_AT_ENCAP; + unsigned int flag = FLOW_DISSECTOR_F_STOP_AT_ENCAP | + FLOW_DISSECTOR_F_STOP_AT_L4; +; struct flow_keys keys; /* short-circuit if we already have L4 hash present */
--
2.11.0