Re: [PATCH V3 net-next] LISP: Locator/Identifier Separation Protocol
From: Tom Herbert <hidden>
Date: 2014-06-19 00:08:51
+/* Compute source port for outgoing packet
+ * first choice to use L4 flow hash since it will spread
+ * better and maybe available from hardware
+ * secondary choice is to use jhash on the Ethernet header
+ */
+static inline __be16 udp_tunnel_get_src_port(__u16 port_min, __u16 port_max,
+ struct sk_buff *skb)
+{
+ unsigned int range = (port_max - port_min) + 1;
+ u32 hash;
+
+ hash = skb_get_hash(skb);
+ if (!hash)
+ hash = jhash(skb->data, 2 * ETH_ALEN,
+ (__force u32) skb->protocol);
+This is probably okay for now, but we should really be smarter here. Looks like another consumer of a TX hash in skbuf (Eric was looking to add that I believe).
quoted hunk ↗ jump to hunk
+ return htons((((u64) hash * range) >> 32) + port_min); +} + +/* Compute source UDP port for outgoing packets on UDP tunnels + */ static inline void udp_lib_close(struct sock *sk, long timeout) { sk_common_release(sk);@@ -270,7 +291,8 @@ void udp4_proc_exit(void); #endif int udpv4_offload_init(void); - +int udpv4_create_encap_sock(struct net *net, __be16 port, struct socket **psock, + bool csum); void udp_init(void); void udp_encap_enable(void);diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h index b385348..0077832 100644 --- a/include/uapi/linux/if_link.h +++ b/include/uapi/linux/if_link.h@@ -331,6 +331,23 @@ struct ifla_vxlan_port_range { __be16 high; }; +/* LISP section */ +enum { + IFLA_LISP_UNSPEC, + IFLA_LISP_IID, + IFLA_LISP_LOCAL, + IFLA_LISP_REMOTE, + IFLA_LISP_LOCAL6, + IFLA_LISP_REMOTE6, + IFLA_LISP_ENCAP_PORT, + IFLA_LISP_LISTEN_PORT, + IFLA_LISP_TOS, + IFLA_LISP_TTL, + IFLA_LISP_UDP_CSUM, + __IFLA_LISP_MAX +}; +#define IFLA_LISP_MAX (__IFLA_LISP_MAX - 1) + /* Bonding section */ enum {diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index d92f94b..c69b198 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c@@ -2524,78 +2524,3 @@ void __init udp_init(void) sysctl_udp_wmem_min = SK_MEM_QUANTUM; } -struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb, - netdev_features_t features) -{ - struct sk_buff *segs = ERR_PTR(-EINVAL); - u16 mac_offset = skb->mac_header; - int mac_len = skb->mac_len; - int tnl_hlen = skb_inner_mac_header(skb) - skb_transport_header(skb); - __be16 protocol = skb->protocol; - netdev_features_t enc_features; - int udp_offset, outer_hlen; - unsigned int oldlen; - bool need_csum; - - oldlen = (u16)~skb->len; - - if (unlikely(!pskb_may_pull(skb, tnl_hlen))) - goto out; - - skb->encapsulation = 0; - __skb_pull(skb, tnl_hlen); - skb_reset_mac_header(skb); - skb_set_network_header(skb, skb_inner_network_offset(skb)); - skb->mac_len = skb_inner_network_offset(skb); - skb->protocol = htons(ETH_P_TEB); - - need_csum = !!(skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL_CSUM); - if (need_csum) - skb->encap_hdr_csum = 1; - - /* segment inner packet. */ - enc_features = skb->dev->hw_enc_features & netif_skb_features(skb); - segs = skb_mac_gso_segment(skb, enc_features); - if (!segs || IS_ERR(segs)) { - skb_gso_error_unwind(skb, protocol, tnl_hlen, mac_offset, - mac_len); - goto out; - } - - outer_hlen = skb_tnl_header_len(skb); - udp_offset = outer_hlen - tnl_hlen; - skb = segs; - do { - struct udphdr *uh; - int len; - - skb_reset_inner_headers(skb); - skb->encapsulation = 1; - - skb->mac_len = mac_len; - - skb_push(skb, outer_hlen); - skb_reset_mac_header(skb); - skb_set_network_header(skb, mac_len); - skb_set_transport_header(skb, udp_offset); - len = skb->len - udp_offset; - uh = udp_hdr(skb); - uh->len = htons(len); - - if (need_csum) { - __be32 delta = htonl(oldlen + len); - - uh->check = ~csum_fold((__force __wsum) - ((__force u32)uh->check + - (__force u32)delta)); - uh->check = gso_make_checksum(skb, ~uh->check); - - if (uh->check == 0) - uh->check = CSUM_MANGLED_0; - } - - skb->protocol = protocol; - } while ((skb = skb->next)); -out: - return segs; -}diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c index 546d2d4..cb77404 100644 --- a/net/ipv4/udp_offload.c +++ b/net/ipv4/udp_offload.c@@ -248,3 +248,121 @@ int __init udpv4_offload_init(void) { return inet_add_offload(&udpv4_offload, IPPROTO_UDP); } + +struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb, + netdev_features_t features) +{ + struct sk_buff *segs = ERR_PTR(-EINVAL); + u16 mac_offset = skb->mac_header; + int mac_len = skb->mac_len; + int tnl_hlen = skb_inner_mac_header(skb) - skb_transport_header(skb); + __be16 protocol = skb->protocol; + netdev_features_t enc_features; + int udp_offset, outer_hlen; + unsigned int oldlen; + bool need_csum; + + oldlen = (u16)~skb->len; + + if (unlikely(!pskb_may_pull(skb, tnl_hlen))) + goto out; + + skb->encapsulation = 0; + __skb_pull(skb, tnl_hlen); + skb_reset_mac_header(skb); + skb_set_network_header(skb, skb_inner_network_offset(skb)); + skb->mac_len = skb_inner_network_offset(skb); + skb->protocol = htons(ETH_P_TEB); + + need_csum = !!(skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL_CSUM); + if (need_csum) + skb->encap_hdr_csum = 1; + + /* segment inner packet. */ + enc_features = skb->dev->hw_enc_features & netif_skb_features(skb); + segs = skb_mac_gso_segment(skb, enc_features); + if (!segs || IS_ERR(segs)) { + skb_gso_error_unwind(skb, protocol, tnl_hlen, mac_offset, + mac_len); + goto out; + } + + outer_hlen = skb_tnl_header_len(skb); + udp_offset = outer_hlen - tnl_hlen; + skb = segs; + do { + struct udphdr *uh; + int len; + + skb_reset_inner_headers(skb); + skb->encapsulation = 1; + + skb->mac_len = mac_len; + + skb_push(skb, outer_hlen); + skb_reset_mac_header(skb); + skb_set_network_header(skb, mac_len); + skb_set_transport_header(skb, udp_offset); + len = skb->len - udp_offset; + uh = udp_hdr(skb); + uh->len = htons(len); + + if (need_csum) { + __be32 delta = htonl(oldlen + len); + + uh->check = ~csum_fold((__force __wsum) + ((__force u32)uh->check + + (__force u32)delta)); + uh->check = gso_make_checksum(skb, ~uh->check); + + if (uh->check == 0) + uh->check = CSUM_MANGLED_0; + } + + skb->protocol = protocol; + } while ((skb = skb->next)); +out: + return segs; +} +
Please split out generic changes into their own patches.
+int udpv4_create_encap_sock(struct net *net, __be16 port, struct socket **psock,
+ bool csum)
+{
+ struct sock *sk;
+ struct socket *sock;
+ struct sockaddr_in lisp_addr = {
+ .sin_family = AF_INET,
+ .sin_addr.s_addr = htonl(INADDR_ANY),
+ .sin_port = port,
+ };
+ int rc;
+Still some lisp artifacts.
+ /* Create UDP socket for encapsulation receive. */
+ rc = sock_create_kern(AF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock);
+ if (rc < 0) {
+ pr_debug("UDP socket create failed\n");
+ return rc;
+ }
+
+ /* Put in proper namespace */
+ sk = sock->sk;
+ sk_change_net(sk, net);
+
+ rc = kernel_bind(sock, (struct sockaddr *)&lisp_addr,
+ sizeof(lisp_addr));
+ if (rc < 0) {
+ pr_debug("bind for UDP socket %pI4:%u (%d)\n",
+ &lisp_addr.sin_addr, ntohs(lisp_addr.sin_port), rc);
+ sk_release_kernel(sk);
+ return rc;
+ }
+
+ *psock = sock;
+ /* Disable multicast loopback */
+ inet_sk(sk)->mc_loop = 0;
+
+ if (!csum)
+ sock->sk->sk_no_check_tx = 1;
+ return 0;
+}
+EXPORT_SYMBOL(udpv4_create_encap_sock)I was actually thinking this function could be even more general. The L2TP_ENCAPTYPE_UDP case in l2tp_tunnel_sock_create looks like something we might be able to abstract out into a separate function-- it would include IPv6 support and the possibility of binding to other than INADDR_ANY and connected sockets. l2tp_tunnel_cfg could be the basis of udp_port_cfg which contains addresses, ports, and sockopts like sk_no_check_tx...
-- 1.7.10.4 -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html