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

Pull networking changes from David S Miller:

 1) Remove the ipv4 routing cache.  Now lookups go directly into the FIB
    trie and use prebuilt routes cached there.

    No more garbage collection, no more rDOS attacks on the routing
    cache.  Instead we now get predictable and consistent performance,
    no matter what the pattern of traffic we service.

    This has been almost 2 years in the making.  Special thanks to
    Julian Anastasov, Eric Dumazet, Steffen Klassert, and others who
    have helped along the way.

    I'm sure that with a change of this magnitude there will be some
    kind of fallout, but such things ought the be simple to fix at this
    point.  Luckily I'm not European so I'll be around all of August to
    fix things :-)

    The major stages of this work here are each fronted by a forced
    merge commit whose commit message contains a top-level description
    of the motivations and implementation issues.

 2) Pre-demux of established ipv4 TCP sockets, saves a route demux on
    input.

 3) TCP SYN/ACK performance tweaks from Eric Dumazet.

 4) Add namespace support for netfilter L4 conntrack helpers, from Gao
    Feng.

 5) Add config mechanism for Energy Efficient Ethernet to ethtool, from
    Yuval Mintz.

 6) Remove quadratic behavior from /proc/net/unix, from Eric Dumazet.

 7) Support for connection tracker helpers in userspace, from Pablo
    Neira Ayuso.

 8) Allow userspace driven TX load balancing functions in TEAM driver,
    from Jiri Pirko.

 9) Kill off NLMSG_PUT and RTA_PUT macros, more gross stuff with
    embedded gotos.

10) TCP Small Queues, essentially minimize the amount of TCP data queued
    up in the packet scheduler layer.  Whereas the existing BQL (Byte
    Queue Limits) limits the pkt_sched --> netdevice queuing levels,
    this controls the TCP --> pkt_sched queueing levels.

    From Eric Dumazet.

11) Reduce the number of get_page/put_page ops done on SKB fragments,
    from Alexander Duyck.

12) Implement protection against blind resets in TCP (RFC 5961), from
    Eric Dumazet.

13) Support the client side of TCP Fast Open, basically the ability to
    send data in the SYN exchange, from Yuchung Cheng.

    Basically, the sender queues up data with a sendmsg() call using
    MSG_FASTOPEN, then they do the connect() which emits the queued up
    fastopen data.

14) Avoid all the problems we get into in TCP when timers or PMTU events
    hit a locked socket.  The TCP Small Queues changes added a
    tcp_release_cb() that allows us to queue work up to the
    release_sock() caller, and that's what we use here too.  From Eric
    Dumazet.

15) Zero copy on TX support for TUN driver, from Michael S. Tsirkin.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1870 commits)
  genetlink: define lockdep_genl_is_held() when CONFIG_LOCKDEP
  r8169: revert "add byte queue limit support".
  ipv4: Change rt->rt_iif encoding.
  net: Make skb->skb_iif always track skb->dev
  ipv4: Prepare for change of rt->rt_iif encoding.
  ipv4: Remove all RTCF_DIRECTSRC handliing.
  ipv4: Really ignore ICMP address requests/replies.
  decnet: Don't set RTCF_DIRECTSRC.
  net/ipv4/ip_vti.c: Fix __rcu warnings detected by sparse.
  ipv4: Remove redundant assignment
  rds: set correct msg_namelen
  openvswitch: potential NULL deref in sample()
  tcp: dont drop MTU reduction indications
  bnx2x: Add new 57840 device IDs
  tcp: avoid oops in tcp_metrics and reset tcpm_stamp
  niu: Change niu_rbr_fill() to use unlikely() to check niu_rbr_add_page() return value
  niu: Fix to check for dma mapping errors.
  net: Fix references to out-of-scope variables in put_cmsg_compat()
  net: ethernet: davinci_emac: add pm_runtime support
  net: ethernet: davinci_emac: Remove unnecessary #include
  ...

1 files changed, 40 insertions, 151 deletions

diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index c75efbdc71c..f2eccd53174 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -95,6 +95,7 @@
 #include <net/checksum.h>
 #include <net/xfrm.h>
 #include <net/inet_common.h>
+#include <net/ip_fib.h>
 
 /*
  *	Build xmit assembly blocks
@@ -253,10 +254,10 @@ static inline bool icmpv4_xrlim_allow(struct net *net, struct rtable *rt,
 
 	/* Limit if icmp type is enabled in ratemask. */
 	if ((1 << type) & net->ipv4.sysctl_icmp_ratemask) {
-		if (!rt->peer)
-			rt_bind_peer(rt, fl4->daddr, 1);
-		rc = inet_peer_xrlim_allow(rt->peer,
+		struct inet_peer *peer = inet_getpeer_v4(net->ipv4.peers, fl4->daddr, 1);
+		rc = inet_peer_xrlim_allow(peer,
 					   net->ipv4.sysctl_icmp_ratelimit);
+		inet_putpeer(peer);
 	}
 out:
 	return rc;
@@ -334,7 +335,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
 	struct flowi4 fl4;
 	struct sock *sk;
 	struct inet_sock *inet;
-	__be32 daddr;
+	__be32 daddr, saddr;
 
 	if (ip_options_echo(&icmp_param->replyopts.opt.opt, skb))
 		return;
@@ -348,6 +349,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
 
 	inet->tos = ip_hdr(skb)->tos;
 	daddr = ipc.addr = ip_hdr(skb)->saddr;
+	saddr = fib_compute_spec_dst(skb);
 	ipc.opt = NULL;
 	ipc.tx_flags = 0;
 	if (icmp_param->replyopts.opt.opt.optlen) {
@@ -357,7 +359,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
 	}
 	memset(&fl4, 0, sizeof(fl4));
 	fl4.daddr = daddr;
-	fl4.saddr = rt->rt_spec_dst;
+	fl4.saddr = saddr;
 	fl4.flowi4_tos = RT_TOS(ip_hdr(skb)->tos);
 	fl4.flowi4_proto = IPPROTO_ICMP;
 	security_skb_classify_flow(skb, flowi4_to_flowi(&fl4));
@@ -569,7 +571,7 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
 		rcu_read_lock();
 		if (rt_is_input_route(rt) &&
 		    net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr)
-			dev = dev_get_by_index_rcu(net, rt->rt_iif);
+			dev = dev_get_by_index_rcu(net, inet_iif(skb_in));
 
 		if (dev)
 			saddr = inet_select_addr(dev, 0, RT_SCOPE_LINK);
@@ -632,6 +634,27 @@ out:;
 EXPORT_SYMBOL(icmp_send);
 
 
+static void icmp_socket_deliver(struct sk_buff *skb, u32 info)
+{
+	const struct iphdr *iph = (const struct iphdr *) skb->data;
+	const struct net_protocol *ipprot;
+	int protocol = iph->protocol;
+
+	/* Checkin full IP header plus 8 bytes of protocol to
+	 * avoid additional coding at protocol handlers.
+	 */
+	if (!pskb_may_pull(skb, iph->ihl * 4 + 8))
+		return;
+
+	raw_icmp_error(skb, protocol, info);
+
+	rcu_read_lock();
+	ipprot = rcu_dereference(inet_protos[protocol]);
+	if (ipprot && ipprot->err_handler)
+		ipprot->err_handler(skb, info);
+	rcu_read_unlock();
+}
+
 /*
  *	Handle ICMP_DEST_UNREACH, ICMP_TIME_EXCEED, and ICMP_QUENCH.
  */
@@ -640,10 +663,8 @@ static void icmp_unreach(struct sk_buff *skb)
 {
 	const struct iphdr *iph;
 	struct icmphdr *icmph;
-	int hash, protocol;
-	const struct net_protocol *ipprot;
-	u32 info = 0;
 	struct net *net;
+	u32 info = 0;
 
 	net = dev_net(skb_dst(skb)->dev);
 
@@ -674,9 +695,7 @@ static void icmp_unreach(struct sk_buff *skb)
 				LIMIT_NETDEBUG(KERN_INFO pr_fmt("%pI4: fragmentation needed and DF set\n"),
 					       &iph->daddr);
 			} else {
-				info = ip_rt_frag_needed(net, iph,
-							 ntohs(icmph->un.frag.mtu),
-							 skb->dev);
+				info = ntohs(icmph->un.frag.mtu);
 				if (!info)
 					goto out;
 			}
@@ -720,26 +739,7 @@ static void icmp_unreach(struct sk_buff *skb)
 		goto out;
 	}
 
-	/* Checkin full IP header plus 8 bytes of protocol to
-	 * avoid additional coding at protocol handlers.
-	 */
-	if (!pskb_may_pull(skb, iph->ihl * 4 + 8))
-		goto out;
-
-	iph = (const struct iphdr *)skb->data;
-	protocol = iph->protocol;
-
-	/*
-	 *	Deliver ICMP message to raw sockets. Pretty useless feature?
-	 */
-	raw_icmp_error(skb, protocol, info);
-
-	hash = protocol & (MAX_INET_PROTOS - 1);
-	rcu_read_lock();
-	ipprot = rcu_dereference(inet_protos[hash]);
-	if (ipprot && ipprot->err_handler)
-		ipprot->err_handler(skb, info);
-	rcu_read_unlock();
+	icmp_socket_deliver(skb, info);
 
 out:
 	return;
@@ -755,46 +755,15 @@ out_err:
 
 static void icmp_redirect(struct sk_buff *skb)
 {
-	const struct iphdr *iph;
-
-	if (skb->len < sizeof(struct iphdr))
-		goto out_err;
-
-	/*
-	 *	Get the copied header of the packet that caused the redirect
-	 */
-	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
-		goto out;
-
-	iph = (const struct iphdr *)skb->data;
-
-	switch (icmp_hdr(skb)->code & 7) {
-	case ICMP_REDIR_NET:
-	case ICMP_REDIR_NETTOS:
-		/*
-		 * As per RFC recommendations now handle it as a host redirect.
-		 */
-	case ICMP_REDIR_HOST:
-	case ICMP_REDIR_HOSTTOS:
-		ip_rt_redirect(ip_hdr(skb)->saddr, iph->daddr,
-			       icmp_hdr(skb)->un.gateway,
-			       iph->saddr, skb->dev);
-		break;
+	if (skb->len < sizeof(struct iphdr)) {
+		ICMP_INC_STATS_BH(dev_net(skb->dev), ICMP_MIB_INERRORS);
+		return;
 	}
 
-	/* Ping wants to see redirects.
-         * Let's pretend they are errors of sorts... */
-	if (iph->protocol == IPPROTO_ICMP &&
-	    iph->ihl >= 5 &&
-	    pskb_may_pull(skb, (iph->ihl<<2)+8)) {
-		ping_err(skb, icmp_hdr(skb)->un.gateway);
-	}
+	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
+		return;
 
-out:
-	return;
-out_err:
-	ICMP_INC_STATS_BH(dev_net(skb->dev), ICMP_MIB_INERRORS);
-	goto out;
+	icmp_socket_deliver(skb, icmp_hdr(skb)->un.gateway);
 }
 
 /*
@@ -868,86 +837,6 @@ out_err:
 	goto out;
 }
 
-
-/*
- *	Handle ICMP_ADDRESS_MASK requests.  (RFC950)
- *
- * RFC1122 (3.2.2.9).  A host MUST only send replies to
- * ADDRESS_MASK requests if it's been configured as an address mask
- * agent.  Receiving a request doesn't constitute implicit permission to
- * act as one. Of course, implementing this correctly requires (SHOULD)
- * a way to turn the functionality on and off.  Another one for sysctl(),
- * I guess. -- MS
- *
- * RFC1812 (4.3.3.9).	A router MUST implement it.
- *			A router SHOULD have switch turning it on/off.
- *		      	This switch MUST be ON by default.
- *
- * Gratuitous replies, zero-source replies are not implemented,
- * that complies with RFC. DO NOT implement them!!! All the idea
- * of broadcast addrmask replies as specified in RFC950 is broken.
- * The problem is that it is not uncommon to have several prefixes
- * on one physical interface. Moreover, addrmask agent can even be
- * not aware of existing another prefixes.
- * If source is zero, addrmask agent cannot choose correct prefix.
- * Gratuitous mask announcements suffer from the same problem.
- * RFC1812 explains it, but still allows to use ADDRMASK,
- * that is pretty silly. --ANK
- *
- * All these rules are so bizarre, that I removed kernel addrmask
- * support at all. It is wrong, it is obsolete, nobody uses it in
- * any case. --ANK
- *
- * Furthermore you can do it with a usermode address agent program
- * anyway...
- */
-
-static void icmp_address(struct sk_buff *skb)
-{
-#if 0
-	net_dbg_ratelimited("a guy asks for address mask. Who is it?\n");
-#endif
-}
-
-/*
- * RFC1812 (4.3.3.9).	A router SHOULD listen all replies, and complain
- *			loudly if an inconsistency is found.
- * called with rcu_read_lock()
- */
-
-static void icmp_address_reply(struct sk_buff *skb)
-{
-	struct rtable *rt = skb_rtable(skb);
-	struct net_device *dev = skb->dev;
-	struct in_device *in_dev;
-	struct in_ifaddr *ifa;
-
-	if (skb->len < 4 || !(rt->rt_flags&RTCF_DIRECTSRC))
-		return;
-
-	in_dev = __in_dev_get_rcu(dev);
-	if (!in_dev)
-		return;
-
-	if (in_dev->ifa_list &&
-	    IN_DEV_LOG_MARTIANS(in_dev) &&
-	    IN_DEV_FORWARD(in_dev)) {
-		__be32 _mask, *mp;
-
-		mp = skb_header_pointer(skb, 0, sizeof(_mask), &_mask);
-		BUG_ON(mp == NULL);
-		for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
-			if (*mp == ifa->ifa_mask &&
-			    inet_ifa_match(ip_hdr(skb)->saddr, ifa))
-				break;
-		}
-		if (!ifa)
-			net_info_ratelimited("Wrong address mask %pI4 from %s/%pI4\n",
-					     mp,
-					     dev->name, &ip_hdr(skb)->saddr);
-	}
-}
-
 static void icmp_discard(struct sk_buff *skb)
 {
 }
@@ -1111,10 +1000,10 @@ static const struct icmp_control icmp_pointers[NR_ICMP_TYPES + 1] = {
 		.handler = icmp_discard,
 	},
 	[ICMP_ADDRESS] = {
-		.handler = icmp_address,
+		.handler = icmp_discard,
 	},
 	[ICMP_ADDRESSREPLY] = {
-		.handler = icmp_address_reply,
+		.handler = icmp_discard,
 	},
 };