Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next

Pull networking updates from David Miller:
 "Highlights:

   1) Steady transitioning of the BPF instructure to a generic spot so
      all kernel subsystems can make use of it, from Alexei Starovoitov.

   2) SFC driver supports busy polling, from Alexandre Rames.

   3) Take advantage of hash table in UDP multicast delivery, from David
      Held.

   4) Lighten locking, in particular by getting rid of the LRU lists, in
      inet frag handling.  From Florian Westphal.

   5) Add support for various RFC6458 control messages in SCTP, from
      Geir Ola Vaagland.

   6) Allow to filter bridge forwarding database dumps by device, from
      Jamal Hadi Salim.

   7) virtio-net also now supports busy polling, from Jason Wang.

   8) Some low level optimization tweaks in pktgen from Jesper Dangaard
      Brouer.

   9) Add support for ipv6 address generation modes, so that userland
      can have some input into the process.  From Jiri Pirko.

  10) Consolidate common TCP connection request code in ipv4 and ipv6,
      from Octavian Purdila.

  11) New ARP packet logger in netfilter, from Pablo Neira Ayuso.

  12) Generic resizable RCU hash table, with intial users in netlink and
      nftables.  From Thomas Graf.

  13) Maintain a name assignment type so that userspace can see where a
      network device name came from (enumerated by kernel, assigned
      explicitly by userspace, etc.) From Tom Gundersen.

  14) Automatic flow label generation on transmit in ipv6, from Tom
      Herbert.

  15) New packet timestamping facilities from Willem de Bruijn, meant to
      assist in measuring latencies going into/out-of the packet
      scheduler, latency from TCP data transmission to ACK, etc"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1536 commits)
  cxgb4 : Disable recursive mailbox commands when enabling vi
  net: reduce USB network driver config options.
  tg3: Modify tg3_tso_bug() to handle multiple TX rings
  amd-xgbe: Perform phy connect/disconnect at dev open/stop
  amd-xgbe: Use dma_set_mask_and_coherent to set DMA mask
  net: sun4i-emac: fix memory leak on bad packet
  sctp: fix possible seqlock seadlock in sctp_packet_transmit()
  Revert "net: phy: Set the driver when registering an MDIO bus device"
  cxgb4vf: Turn off SGE RX/TX Callback Timers and interrupts in PCI shutdown routine
  team: Simplify return path of team_newlink
  bridge: Update outdated comment on promiscuous mode
  net-timestamp: ACK timestamp for bytestreams
  net-timestamp: TCP timestamping
  net-timestamp: SCHED timestamp on entering packet scheduler
  net-timestamp: add key to disambiguate concurrent datagrams
  net-timestamp: move timestamp flags out of sk_flags
  net-timestamp: extend SCM_TIMESTAMPING ancillary data struct
  cxgb4i : Move stray CPL definitions to cxgb4 driver
  tcp: reduce spurious retransmits due to transient SACK reneging
  qlcnic: Initialize dcbnl_ops before register_netdev
  ...

1 files changed, 55 insertions, 154 deletions

diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 229239ad96b1..22055b098428 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -198,6 +198,8 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
 	sk->sk_v6_daddr = usin->sin6_addr;
 	np->flow_label = fl6.flowlabel;
 
+	ip6_set_txhash(sk);
+
 	/*
 	 *	TCP over IPv4
 	 */
@@ -470,13 +472,14 @@ out:
 
 
 static int tcp_v6_send_synack(struct sock *sk, struct dst_entry *dst,
-			      struct flowi6 *fl6,
+			      struct flowi *fl,
 			      struct request_sock *req,
 			      u16 queue_mapping,
 			      struct tcp_fastopen_cookie *foc)
 {
 	struct inet_request_sock *ireq = inet_rsk(req);
 	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct flowi6 *fl6 = &fl->u.ip6;
 	struct sk_buff *skb;
 	int err = -ENOMEM;
 
@@ -503,18 +506,6 @@ done:
 	return err;
 }
 
-static int tcp_v6_rtx_synack(struct sock *sk, struct request_sock *req)
-{
-	struct flowi6 fl6;
-	int res;
-
-	res = tcp_v6_send_synack(sk, NULL, &fl6, req, 0, NULL);
-	if (!res) {
-		TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_RETRANSSEGS);
-		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSYNRETRANS);
-	}
-	return res;
-}
 
 static void tcp_v6_reqsk_destructor(struct request_sock *req)
 {
@@ -718,22 +709,66 @@ static int tcp_v6_inbound_md5_hash(struct sock *sk, const struct sk_buff *skb)
 }
 #endif
 
+static void tcp_v6_init_req(struct request_sock *req, struct sock *sk,
+			    struct sk_buff *skb)
+{
+	struct inet_request_sock *ireq = inet_rsk(req);
+	struct ipv6_pinfo *np = inet6_sk(sk);
+
+	ireq->ir_v6_rmt_addr = ipv6_hdr(skb)->saddr;
+	ireq->ir_v6_loc_addr = ipv6_hdr(skb)->daddr;
+
+	ireq->ir_iif = sk->sk_bound_dev_if;
+
+	/* So that link locals have meaning */
+	if (!sk->sk_bound_dev_if &&
+	    ipv6_addr_type(&ireq->ir_v6_rmt_addr) & IPV6_ADDR_LINKLOCAL)
+		ireq->ir_iif = inet6_iif(skb);
+
+	if (!TCP_SKB_CB(skb)->when &&
+	    (ipv6_opt_accepted(sk, skb) || np->rxopt.bits.rxinfo ||
+	     np->rxopt.bits.rxoinfo || np->rxopt.bits.rxhlim ||
+	     np->rxopt.bits.rxohlim || np->repflow)) {
+		atomic_inc(&skb->users);
+		ireq->pktopts = skb;
+	}
+}
+
+static struct dst_entry *tcp_v6_route_req(struct sock *sk, struct flowi *fl,
+					  const struct request_sock *req,
+					  bool *strict)
+{
+	if (strict)
+		*strict = true;
+	return inet6_csk_route_req(sk, &fl->u.ip6, req);
+}
+
 struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
 	.family		=	AF_INET6,
 	.obj_size	=	sizeof(struct tcp6_request_sock),
-	.rtx_syn_ack	=	tcp_v6_rtx_synack,
+	.rtx_syn_ack	=	tcp_rtx_synack,
 	.send_ack	=	tcp_v6_reqsk_send_ack,
 	.destructor	=	tcp_v6_reqsk_destructor,
 	.send_reset	=	tcp_v6_send_reset,
 	.syn_ack_timeout =	tcp_syn_ack_timeout,
 };
 
-#ifdef CONFIG_TCP_MD5SIG
 static const struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
+	.mss_clamp	=	IPV6_MIN_MTU - sizeof(struct tcphdr) -
+				sizeof(struct ipv6hdr),
+#ifdef CONFIG_TCP_MD5SIG
 	.md5_lookup	=	tcp_v6_reqsk_md5_lookup,
 	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
-};
 #endif
+	.init_req	=	tcp_v6_init_req,
+#ifdef CONFIG_SYN_COOKIES
+	.cookie_init_seq =	cookie_v6_init_sequence,
+#endif
+	.route_req	=	tcp_v6_route_req,
+	.init_seq	=	tcp_v6_init_sequence,
+	.send_synack	=	tcp_v6_send_synack,
+	.queue_hash_add =	inet6_csk_reqsk_queue_hash_add,
+};
 
 static void tcp_v6_send_response(struct sk_buff *skb, u32 seq, u32 ack, u32 win,
 				 u32 tsval, u32 tsecr, int oif,
@@ -973,153 +1008,17 @@ static struct sock *tcp_v6_hnd_req(struct sock *sk, struct sk_buff *skb)
 	return sk;
 }
 
-/* FIXME: this is substantially similar to the ipv4 code.
- * Can some kind of merge be done? -- erics
- */
 static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
 {
-	struct tcp_options_received tmp_opt;
-	struct request_sock *req;
-	struct inet_request_sock *ireq;
-	struct ipv6_pinfo *np = inet6_sk(sk);
-	struct tcp_sock *tp = tcp_sk(sk);
-	__u32 isn = TCP_SKB_CB(skb)->when;
-	struct dst_entry *dst = NULL;
-	struct tcp_fastopen_cookie foc = { .len = -1 };
-	bool want_cookie = false, fastopen;
-	struct flowi6 fl6;
-	int err;
-
 	if (skb->protocol == htons(ETH_P_IP))
 		return tcp_v4_conn_request(sk, skb);
 
 	if (!ipv6_unicast_destination(skb))
 		goto drop;
 
-	if ((sysctl_tcp_syncookies == 2 ||
-	     inet_csk_reqsk_queue_is_full(sk)) && !isn) {
-		want_cookie = tcp_syn_flood_action(sk, skb, "TCPv6");
-		if (!want_cookie)
-			goto drop;
-	}
-
-	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1) {
-		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
-		goto drop;
-	}
-
-	req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
-	if (req == NULL)
-		goto drop;
-
-#ifdef CONFIG_TCP_MD5SIG
-	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv6_ops;
-#endif
-
-	tcp_clear_options(&tmp_opt);
-	tmp_opt.mss_clamp = IPV6_MIN_MTU - sizeof(struct tcphdr) - sizeof(struct ipv6hdr);
-	tmp_opt.user_mss = tp->rx_opt.user_mss;
-	tcp_parse_options(skb, &tmp_opt, 0, want_cookie ? NULL : &foc);
+	return tcp_conn_request(&tcp6_request_sock_ops,
+				&tcp_request_sock_ipv6_ops, sk, skb);
 
-	if (want_cookie && !tmp_opt.saw_tstamp)
-		tcp_clear_options(&tmp_opt);
-
-	tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
-	tcp_openreq_init(req, &tmp_opt, skb);
-
-	ireq = inet_rsk(req);
-	ireq->ir_v6_rmt_addr = ipv6_hdr(skb)->saddr;
-	ireq->ir_v6_loc_addr = ipv6_hdr(skb)->daddr;
-	if (!want_cookie || tmp_opt.tstamp_ok)
-		TCP_ECN_create_request(req, skb, sock_net(sk));
-
-	ireq->ir_iif = sk->sk_bound_dev_if;
-	ireq->ir_mark = inet_request_mark(sk, skb);
-
-	/* So that link locals have meaning */
-	if (!sk->sk_bound_dev_if &&
-	    ipv6_addr_type(&ireq->ir_v6_rmt_addr) & IPV6_ADDR_LINKLOCAL)
-		ireq->ir_iif = inet6_iif(skb);
-
-	if (!isn) {
-		if (ipv6_opt_accepted(sk, skb) ||
-		    np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo ||
-		    np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim ||
-		    np->repflow) {
-			atomic_inc(&skb->users);
-			ireq->pktopts = skb;
-		}
-
-		if (want_cookie) {
-			isn = cookie_v6_init_sequence(sk, skb, &req->mss);
-			req->cookie_ts = tmp_opt.tstamp_ok;
-			goto have_isn;
-		}
-
-		/* VJ's idea. We save last timestamp seen
-		 * from the destination in peer table, when entering
-		 * state TIME-WAIT, and check against it before
-		 * accepting new connection request.
-		 *
-		 * If "isn" is not zero, this request hit alive
-		 * timewait bucket, so that all the necessary checks
-		 * are made in the function processing timewait state.
-		 */
-		if (tmp_opt.saw_tstamp &&
-		    tcp_death_row.sysctl_tw_recycle &&
-		    (dst = inet6_csk_route_req(sk, &fl6, req)) != NULL) {
-			if (!tcp_peer_is_proven(req, dst, true)) {
-				NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSPASSIVEREJECTED);
-				goto drop_and_release;
-			}
-		}
-		/* Kill the following clause, if you dislike this way. */
-		else if (!sysctl_tcp_syncookies &&
-			 (sysctl_max_syn_backlog - inet_csk_reqsk_queue_len(sk) <
-			  (sysctl_max_syn_backlog >> 2)) &&
-			 !tcp_peer_is_proven(req, dst, false)) {
-			/* Without syncookies last quarter of
-			 * backlog is filled with destinations,
-			 * proven to be alive.
-			 * It means that we continue to communicate
-			 * to destinations, already remembered
-			 * to the moment of synflood.
-			 */
-			LIMIT_NETDEBUG(KERN_DEBUG "TCP: drop open request from %pI6/%u\n",
-				       &ireq->ir_v6_rmt_addr, ntohs(tcp_hdr(skb)->source));
-			goto drop_and_release;
-		}
-
-		isn = tcp_v6_init_sequence(skb);
-	}
-have_isn:
-
-	if (security_inet_conn_request(sk, skb, req))
-		goto drop_and_release;
-
-	if (!dst && (dst = inet6_csk_route_req(sk, &fl6, req)) == NULL)
-		goto drop_and_free;
-
-	tcp_rsk(req)->snt_isn = isn;
-	tcp_rsk(req)->snt_synack = tcp_time_stamp;
-	tcp_openreq_init_rwin(req, sk, dst);
-	fastopen = !want_cookie &&
-		   tcp_try_fastopen(sk, skb, req, &foc, dst);
-	err = tcp_v6_send_synack(sk, dst, &fl6, req,
-				 skb_get_queue_mapping(skb), &foc);
-	if (!fastopen) {
-		if (err || want_cookie)
-			goto drop_and_free;
-
-		tcp_rsk(req)->listener = NULL;
-		inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
-	}
-	return 0;
-
-drop_and_release:
-	dst_release(dst);
-drop_and_free:
-	reqsk_free(req);
 drop:
 	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
 	return 0; /* don't send reset */
@@ -1235,6 +1134,8 @@ static struct sock *tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
 	newsk->sk_v6_rcv_saddr = ireq->ir_v6_loc_addr;
 	newsk->sk_bound_dev_if = ireq->ir_iif;
 
+	ip6_set_txhash(newsk);
+
 	/* Now IPv6 options...
 
 	   First: no IPv4 options.